The Journal of Pediatrics

Transcription

1 The Journal of Pediatrics Copyright 2005 Mosby, Inc. All rights reserved Volume 146, Issue 1, Pages (January 2005) 1. Providing the best to a global audience more quickly MISCELLANEOUS Page A1 Alice M. Landwehr 2. Using publicly available data to identify comprehensive pediatric hospitals SHORT COMMUNICATION Page A1 Robert W. Wilmott 3. The gestational age variable SHORT COMMUNICATION Page A1 Alan H. Jobe 4. Maternal engraftment of T-cells in children with severe combined immune deficiency can lead to long term survival SHORT COMMUNICATION Page A2 Robert W. Wilmott 5. Papaverine for arterial catheters in newborns SHORT COMMUNICATION Page A2 Alan H. Jobe 6. Does increased added sugar lead to decreased diet quality SHORT COMMUNICATION Page A2 Stephen R. Daniels 7. Use and safety of budesonide aerosol in infants with asthma SHORT COMMUNICATION Page A2 Robert W. Wilmott

2 8. Treatment of clinical seizures can improve behavior SHORT COMMUNICATION Page A3 Robert W. Wilmott 9. Children with undiagnosed asthma benefit from identification SHORT COMMUNICATION Page A3 Robert W. Wilmott 10. Lung function in BPD at term SHORT COMMUNICATION Page A3 Alan H. Jobe 11. Diarrheagenic E. coli in children SHORT COMMUNICATION Page A3 Sarah S. Long 12. Contents CONTENTS LIST Pages A4-A7 13. Editorial Board EDITORIAL BOARD Page A8 14. Authors' Guide MISCELLANEOUS Pages A9-A Information for Readers MISCELLANEOUS Page A Announcements NEWS Page A Fellowships MISCELLANEOUS Pages A15-A The morbidity and mortality conference in university-affiliated pediatric departments in Canada EDITORIAL Pages 1-2 Jeremy N. Friedman, Marie S. Pinard and Ronald M. Laxer

3 19. Increasing utilization of general pediatricians and pediatric subspecialists: Can the workforce meet the need? EDITORIAL Pages 3-5 Debra H. Fiser 20. Hyperglycemia and mortality EDITORIAL Pages 5-7 Roberto Izquierdo 21. A collaborative effort to define the epidemiology of pediatric inflammatory bowel disease: What can we learn from children with early-onset disease? EDITORIAL Pages 7-8 M. Susan Moyer 22. Safety of adult nerve agent autoinjectors in children EDITORIAL Pages 8-10 Cynthia Aaron 23. A tale of two centimeters EDITORIAL Pages Ron G. Rosenfeld 24. Confirmation of an old adage: You find what you seek EDITORIAL Pages Philip M. Sherman and Phillip I. Tarr 25. Changes in the proportion and volume of care provided to children by generalists and subspecialists ARTICLE Pages Gary L. Freed, Tammie A. Nahra, Patricia J. Venus, Stephanie D. Schech, John R.C. Wheeler and The Research Advisory Committee of the American Board of Pediatrics 26. Comparing the supply of pediatric subspecialists and child neurologists ARTICLE Pages Rachel M. Werner and Daniel Polsky 27. Criteria for identification of comprehensive pediatric hospitals and referral regions ARTICLE Pages Robert K. Kanter and Franklin Dexter

4 28. Persistent hyperglycemia in critically ill children ARTICLE Pages Edward Vincent Faustino and Michael Apkon 29. Children with early-onset inflammatory bowel disease (IBD): Analysis of a pediatric IBD consortium registry ARTICLE Pages Melvin B. Heyman, Barbara S. Kirschner, Benjamin D. Gold, George Ferry, Robert Baldassano, Stanley A. Cohen, Harland S. Winter, Patricia Fain, Chris King, Terry Smith and Hashem B. El-Serag 30. Pediatric poisoning from trimedoxime (TMB4) and atropine automatic injectors ARTICLE Pages Eran Kozer, Amnon Mordel, Shmuael Bar Haim, Mordechei Bulkowstein, Matitiahu Berkovitch and Yedidia Bentur 31. Growth hormone (GH) treatment to final height in children with idiopathic short stature: Evidence for a dose effect ARTICLE Pages Jan M. Wit, Lyset T.M. Rekers-Mombarg, Gordon B. Cutler, Jr., Brenda Crowe, Tracy J. Beck, Kristen Roberts, Anne Gill, Jean-Louis Chaussain, Herwig Frisch et al. 32. Pneumatosis cystoides intestinalis in infancy SHORT COMMUNICATION Page 53 Barbara Warner 33. Prevalence of diarrheagenic Escherichia coli in acute childhood enteritis: A prospective controlled study ARTICLE Pages Mitchell B. Cohen, James P. Nataro, David I. Bernstein, Jennifer Hawkins, Nancy Roberts and Mary A. Staat 34. Papaverine prolongs patency of peripheral arterial catheters in neonates ARTICLE Pages M. Pamela Griffin and Mir S. Siadaty 35. The Parkland Memorial Hospital experience in ensuring compliance with Universal Newborn Hearing Screening follow-up ARTICLE Pages Angela G. Shoup, Kris E. Owen, Greg Jackson and Abbot Laptook

5 36. Can clinical signs identify newborns with neuromuscular disorders? ARTICLE Pages Isabella Vasta, Maria Kinali, Sonia Messina, Andrea Guzzetta, Olga Kapellou, Adnan Manzur, Frances Cowan, Francesco Muntoni and Eugenio Mercuri 37. Motor function at school age in children with a preschool diagnosis of developmental language impairment ARTICLE Pages Richard I. Webster, Annette Majnemer, Robert W. Platt and Michael I. Shevell 38. Correction ERRATUM Page Lung function at term reflects severity of bronchopulmonary dysplasia ARTICLE Pages Ola Hjalmarson and Kenneth L. Sandberg 40. Safety of budesonide inhalation suspension in infants aged six to twelve months with mild to moderate persistent asthma or recurrent wheeze ARTICLE Pages William E. Berger, Paul Y. Qaqundah, Kathryn Blake, Jose Rodriguez- Santana, Anne-Marie Irani, John Xu and Mitchell Goldman 41. Effect of asthma intervention on children with undiagnosed asthma ARTICLE Pages Christine L.M. Joseph, Suzanne Havstad, Elizabeth W. Anderson, Randall Brown, Christine Cole Johnson and Noreen M. Clark 42. On the move? MISCELLANEOUS Page Adverse effect of high added sugar consumption on dietary intake in American preschoolers ARTICLE Pages Sibylle Kranz, Helen Smiciklas-Wright, Anna Maria Siega-Riz and Diane Mitchell

6 44. Some peculiarities of amino acid metabolism in infancy and early childhood SHORT COMMUNICATION Page 111 Paul M. Fernhoff 45. Treatment of interictal epileptiform discharges can improve behavior in children with behavioral problems and epilepsy ARTICLE Pages Ronit M. Pressler, Richard O. Robinson, Goigia A. Wilson and Colin D. Binnie 46. The correlation of deficits in IQ with midface and digit hypoplasia in children exposed in utero to anticonvulsant drugs ARTICLE Pages Lewis B. Holmes, Brent A. Coull, Jon Dorfman and Peter B. Rosenberger 47. The wealth of information conveyed by gestational age ARTICLE Pages Alan Leviton, Eve Blair, Olaf Dammann and Elizabeth Allred 48. Homozygous Factor X gene mutations Gly380Arg and Tyr163delAT are associated with perinatal intracranial hemorrhage ARTICLE Pages Falko H. Herrmann, Marta Navarette, Lizbeth Salazar-Sanchez, Juan M. Carillo, Guenter Auerswald and Karin Wulff 49. Permanent neonatal diabetes in an Asian infant ARTICLE Pages J.R. Porter, N.J. Shaw, T.G. Barrett, A.T. Hattersley, S. Ellard and A.L. Gloyn 50. Pediatric hypereosinophilic syndrome (HES) differs from adult HES ARTICLE Pages Hary T. Katz, S. Jaharul Haque and Fred H. Hsieh 51. Long-term survival in severe combined immune deficiency: The role of persistent maternal engraftment ARTICLE Pages Ilhan Tezcan, Fugen Ersoy, Ozden Sanal, Tuba Turul, Duygu Uckan, Sevim Balci, Gonul Hicsonmez, M. Prieur, S. Caillat-Zucmann, Francois Le Deist and Genevieve de Saint Basile

7 52. Proximal focal femoral deficiency A rare entity in the sonographic differential diagnosis of developmental dysplasia of the hip ARTICLE Page 141 Ralph Kayser, Konrad Mahlfeld, Henning Grasshoff and Harry Rudolph Merk 53. Efficacy and safety of acetaminophen versus ibuprofen for treating children's pain or fever: A meta-analysis ABSTRACT Pages Ran D. Goldman 54. Evaluation of a method to reduce over-the-counter medication dosing error ABSTRACT Page 143 Jill S. Halterman 55. Efficacy and safety of statin therapy in children with familial hypercholesterolemia: A randomized controlled trial ABSTRACT Pages Antonio M. Gotto, Jr. 56. Fluoxetine, cognitive-behavioral therapy, and their combination for adolescents with depression: Treatment for Adolescents with Depression Study (TADS) randomized controlled trial ABSTRACT Page 145 Richard M. Glass 57. Outcomes of allergy to insect stings in children, with and without venom immunotherapy ABSTRACT Pages Patrick J. DeMarco and Richard F. Lockey 58. Effect of dextromethorphan, diphenhydramine, and placebo on nocturnal cough and sleep quality for coughing children and their parents ABSTRACT Pages Knut Schroeder 59. Contagiousness of varicella in vaccinated cases: A household contact study ABSTRACT Page 147 Anne Gershon

8 60. The kidney in Fabry disease CORRESPONDENCE Page 148 Annick Raas-Rothschild, Michael M. Friedlaender, Galina Pizov and Rebecca Backenroth 61. Low birth weight and blood pressure: The role of neonatal factors in the small-baby-syndrome CORRESPONDENCE Pages Thomas Harder and Andreas Plagemann 62. Reply CORRESPONDENCE Page 149 Matthew W. Gillman, Janet W. Rich-Edwards, Sheryl L. Rifas-Shiman, Ken P. Kleinman and Steven E. Lipshultz 63. Early intravenous gamma-globulin treatment for Kawasaki disease: The nationwide surveys in Japan CORRESPONDENCE Pages Suthep Wanitkun 64. Reply CORRESPONDENCE Pages Hiromi Muta and Masahiro Ishii 65. Age-dependent differences in Nordic children with ITP CORRESPONDENCE Pages S. Rosthøj, I. Hedlund-Treutiger, J. Rajantie, B. Zeller, O.G. Jonsson, J.I. Henter and on behalf of the NOPHO ITP Working Group 66. Gastric suction at birth CORRESPONDENCE Pages Laurie S. Conklin, Jeffrey M. Perlman and Myra H. Wyckoff 67. Reply CORRESPONDENCE Page 153 K.J.S. Anand, Cynthia R. Rovnaghi, Richard W. Hall and Bertil Jacobson

9 This Month in THE JOURNAL OF PEDIATRICS January 2005 Volume 146 Number 1 Copyright Ó 2005 by Elsevier Inc. THE EDITORS PERSPECTIVES Providing the best to a global audience more quickly Since 1932, The Journal of Pediatrics has served as a practical guide for the continuing education of physicians who diagnose and treat disorders in infants, children and adolescents. We are dedicated to providing original, peer-reviewed articles oriented toward clinical practice. The reputation of The Journal is based upon the quality of the work of those who submit the best of their research to The Journal, relying on us to convey their work to an audience interested in staying abreast of the latest developments in pediatric medicine and improving the care that they provide to their patients. In order to make that work available more quickly, in 2003, The Journal implemented an online submission and review process. Submissions to The Journal have increased by more than 30%. We are grateful to our reviewers for sharing their time and expertise to help us provide the best work to our readers; most reviewers respond to the invitation to review a manuscript within one day and return their evaluations within 14 days of being invited. Most submissions now receive an editorial decision within 35 days of submission, a decrease of 10 days from the previous turnaround time. In addition, articles published in The Journal have been covered by news media such as Reuters Health, USA Today, the New York Times, Newsweek, CNN, ABC News, National Public Radio, the BBC, Medical News Today, Parenting Magazine, and others. 4,500 people receive the newest issue s table of contents through notifications. From September 2003 to September 2004, there were more than 460,000 full-text downloads of The Journal s articles from The Journal s website and ScienceDirect. The Journal is proud to carry the tradition of providing clinically relevant research to pediatricians and their patients. We thank the authors, reviewers, and readers for allowing us to communicate the best new work to a global audience. Alice M. Landwehr Managing Editor Using publicly available data to identify comprehensive pediatric hospitals The care of children with the most rare and complex diseases is best provided at comprehensive pediatric hospitals. Although there is evidence to support this statement, there has been lack of an objective method to identify such hospitals. In this issue of The Journal, Canter et al develop a method for identifying such hospitals using databases of hospital administrative data. A method was developed for analyzing pediatric discharges to identify hospitals with high clinical volumes, high diversity of diagnosis, and educational accreditation. Applying this approach in the state of New York identified 11 comprehensive pediatric hospitals widely distributed throughout the state. The article identifies the referral regions served by these facilities which collectively provided care for over 29% of all pediatric hospitalizations in the state of New York. This research method should be useful for future analysis of the impact of such hospitals on pediatric healthcare. Robert W. Wilmott, MD Page 26 The gestational age variable Most clinical studies of prematurity use birth weight or gestational age as the major variable for stage of development. Although birth weight is easier to measure with precision, it does not accurately reflect developmental stage. Gestational age estimates are much more accurate than in the past because the near universal use of early gestation ultrasound examinations permit quite accurate gestational dating. Leviton and colleagues develop the concept in a Medical Progress report that gestational age is a complex variable that may contain much more information than simply the development stage of the fetus in weeks and days. They suggest a number of approaches to improve the information derived from clinical studies in the future. Alan H. Jobe, MD, PhD Page 123 The Journal of Pediatrics (ISSN ) is published monthly by Elsevier Inc., 360 Park Avenue South, New York, NY Editorial Offices: Westline Industrial Drive, St. Louis, MO Business Offices: 170 S Independence Mall West, Suite 300 E, Philadelphia, PA Accounting and Circulation Offices: 6277 Sea Harbor Drive, Orlando, FL Periodicals postage paid at New York, NY, and additional mailing offices. POSTMASTER: Send address changes to The Journal of Pediatrics, Elsevier Periodicals Customer Service, 6277 Sea Harbor Drive, Orlando, FL The Journal of Pediatrics January A

10 Maternal engraftment of T-cells in children with severe combined immune deficiency can lead to long term survival Severe Combined Immune Deficiency (SCID) usually leads to severe infections of gastroenterology (GI) and respiratory systems starting early in life and is associated with a poor prognosis. Without stem cell transplantation these children usually die in infancy. Tezcan and colleagues from the Ihsan Dogramaci Children s Hospital in Ankara, Turkey, and the Clinical Immunology Unit at Necker Enfants Malades Hospital in Paris report two siblings where the 8-year-old child was not diagnosed until after the diagnosis of his 6-month old brother. Analysis of the lymphocytes in the 8-year-old showed the presence of materially derived cells. It is proposed that maternal engraftment occasionally occurs in children with SCID and lead to reconstitution of immune function with prolonged survival. Robert W. Wilmott, MD page 137 Papaverine for arterial catheters in newborns Peripheral arterial catheters are often used in critically ill infants either to avoid umbilical arterial lines or because the umbilical arterial lines were removed. Peripheral arteries tend to have vasospasm, and in adults papaverine has been used as a potent vasodilator for catheter related arterial vasospasm. Griffin and Siadaty report a randomized trial to evaluate papaverine for maintenance of peripheral arterial catheters in preterm infants. Such a use for papaverine is unusual and its safety or efficacy has not been evaluated in term or preterm infants. Papaverine did prolong catheter patency without increasing complications. However, the authors caution about its use soon after birth when the risk of IVH is high. Fortunately, the need for prolonged arterial access is unusual in preterm infants. However, if prolonged catheter use is contemplated, papaverine may help. Alan H. Jobe, MD page 62 Does increased added sugar lead to decreased diet quality With the ongoing epidemic of obesity there has been concern about the effect that added sugar may have on diet quality and calorie consumption. This concern includes even younger children prior to school age. In this issue of The Journal, Kranz et al report on the results of a cross sectional survey of 2-5 year old children in the United States. They found that the main sources of added sugar were soft drinks, fruit drinks, and desserts. The majority of children consumed less than 25% of energy from added sugar. Increasing added sugar in the diet was associated with decreasing nutrient and food group intakes. Children consuming 16% or more calories from added sugar often had inadequate calcium consumption. Pediatricians should counsel parents on an appropriately balanced diet and should support limits on added sugar in preschool children. Further research is needed to evaluate the impact of added sugar on weight and health status. Stephen R. Daniels, MD, PhD page 105 Use and safety of budesonide aerosol in infants with asthma Inhaled corticosteroids have become a standard first-line therapy for mild to severe persistent asthma in children of all ages. Budesonide inhalation suspension is a nebulized steroid approved for children as young as 12 months. In a study by Berger et al. in this issue of The Journal, the safety of budesonide inhalation suspension was studied in infants aged six to twelve months. The safety profile was excellent, and not significantly different to that of placebo, with no evidence of adrenal suppression. Robert W. Wilmott, MD page 91 2A January 2005 The Journal of Pediatrics

11 Treatment of clinical seizures can improve behavior The conventional wisdom in epilepsy is that children should be treated only if they have clinical seizures. In this issue of The Journal, Pressler et al test the hypothesis that treatment of subclinical EEG discharges with the lamotrigine might improve behavior in children with epilepsy. This was studied in a double-blind, placebo controlled, cross-over study comparing lamotrigine to placebo. Results show improvement in global ratings of behavior in the children who responded with a significant reduction in frequency or duration of discharges during active treatment. These results suggest that suppressing subclinical or interictal seizures can improve behavior in children with epilepsy and behavioral problems. Robert W. Wilmott, MD page 112 Children with undiagnosed asthma benefit from identification Christine Joseph, Noreen Clark, and colleagues from Henry Ford Health System and the University of Michigan have studied undiagnosed children with asthma who were identified through a community based screening program. It was shown that the children with undiagnosed asthma are generally non-atopic where children with diagnosed asthma had more allergies and seasonal symptoms. Medical intervention in the group of children with undiagnosed asthma resulted in significantly reduced days of symptoms, nights of symptoms, and days of restricted activity. These results demonstrate the impact of making the correct diagnosis in children with asthma and starting effective treatment. Robert W. Wilmott, MD page 96 Lung function in BPD at term Normal preterm infants have lung functional abnormalities at term that persist as the child grows. Infants with BPD have more abnormal lung function at term. Hjalmarson and Sandberg ask the question if the abnormalities in BPD differ in pattern from those of the normal preterm. They measured functional residual capacity, gas mixing, and lung elastic properties and found that the magnitude of the abnormalities increased with severe BPD, but qualitatively the changes were similar in pattern to the abnormalities in the preterm relative to the term infant when measured at term. The interesting speculation resulting from these measurements of lung function is that the new BPD that is common in the very low birth weight premature is simply a more severe form of the developmental abnormalities resulting from preterm birth. Alan H. Jobe, MD page 86 Diarrheagenic E. coli in children The study of Cohen et al is a Herculean prospective investigation using a panel of DNA gene probes to identify diarrheagenic E. coli in stools of well children and children with acute gastroenteritis evaluated in the Emergency Department or as inpatients at Cincinnati Children s Hospital Medical Center. The study population is well categorized and large (1327 children with acute gastroenteritis and 555 well children), evaluation of stool samples is exhaustive (routine bacterial pathogens, rotavirus, panel of DNA probes for known diarrheagenic E. coli and test of E. coli isolates adherence pattern to HEp-2 cells), and analysis of findings is complete and transparent. Remarkably, in aggregate of all evaluations performed, potential infectious etiology of gastroenteritis was confirmed in only one-third of children. Three categories of diarrheagenic E. coli were associated with clinical disease in this study, potentially accounting for more gastroenteritis than routinely identified enteropathogens such as Salmonella, Shigella, and others. The virtual superimposition of seasonal incidences of diarrheagenic E. coli with all other diarrheal diseases (non-e. coli-related) and especially rotavirus deserves further inquiry. Sarah S. Long, MD page 54 The Journal of Pediatrics January A

12 THE JOURNAL OF PEDIATRICS EDITOR William F. Balistreri Cincinnati, Ohio Copyright Ó 2005 by Elsevier Inc. ASSOCIATE EDITORS Stephen R. Daniels Cincinnati, Ohio Alan H. Jobe Cincinnati, Ohio Sarah S. Long Philadelphia, Pennsylvania Thomas R. Welch Syracuse, New York Robert W. Wilmott St. Louis, Missouri MANAGING EDITOR Alice M. Landwehr Cincinnati, Ohio Steven H. Abman Denver, Colorado Pasquale J. Accardo Richmond, Virginia James Brown East Syracuse, New York George R. Buchanan Dallas, Texas Stephen D. Cederbaum Los Angeles, California Ronald I. Clyman San Francisco, California Paul M. Fernhoff Atlanta, Georgia EDITORIAL BOARD Aaron Friedman Editor, AMSPDC Section Providence, Rhode Island John G. Frohna Editor, Current Best Evidence Ann Arbor, Michigan Thomas P. Green Chicago, Illinois M. Susan Jay Maywood, Illinois James F. Padbury Providence, Rhode Island Janet H. Silverstein Gainesville, Florida Ronald J. Sokol Denver, Colorado Virginia A. Stallings Philadelphia, Pennsylvania Charles Stanley Philadelphia, Pennsylvania F. Bruder Stapleton Seattle, Washington Reginald L. Washington Denver, Colorado Basil J. Zitelli Co-Editor, Grand Rounds Section Pittsburgh, Pennsylvania Anita Aperia Stockholm, Sweden INTERNATIONAL ADVISORY PANEL Miguel O Ryan Santiago, Chile Yong Poovorawan Bangkok, Thailand Rosalind L. Smyth Liverpool, United Kingdom 8A January 2005 The Journal of Pediatrics

13 THE JOURNAL OF PEDIATRICS Copyright Ó 2005 by Elsevier Inc. EDITOR William F. Balistreri, MD The Journal of Pediatrics Children s Hospital Medical Center 3333 Burnet Ave, MLC 3021 Cincinnati, OH Alice Landwehr, Managing Editor Phone: ; Fax: journal.pediatrics@cchmc.org PUBLISHER Elsevier Inc Westline Industrial Dr St Louis, MO Beth Weber, Issue Manager , ext 4789, or General Policies and Guide for Authors The Journal of Pediatrics publishes Original Research Articles, Clinical and Laboratory Observations, reviews of Medical Progress in pediatrics and related fields, Grand Rounds and clinicopathologic conference (CPCs), Special Articles, Association of Medical School Pediatric Department Chairs, Inc. (AMSPDC) commentaries, and Insights. Papers are accepted for review with the stipulation that they are submitted solely to The Journal of Pediatrics. The Journal will not consider for review papers that have been published elsewhere, even if in another language, or papers that are being considered by another publication, or are in press. If a paper by the same author or authors contains any data that were previously published, are in press, or are under consideration by another publication, a reprint of the previous article or a copy of the other manuscript should be submitted to the Editor, with a justification or explanation by the authors of any potential overlap or duplication. If the Editor is made aware of such overlapping or duplicate papers that have not been disclosed by the authors, a written explanation will be requested. If, in the judgment of the Editor, the explanation is inadequate, an appropriate official of the primary author s academic institution will be notified. As a condition of authorship, all authors must be willing to take responsibility for the entire manuscript. Multi-authored manuscripts should have a declaration of each author s contributions in the submission cover letter. All authors of a submitted manuscript must sign a form transferring copyright ownership of the manuscript to Elsevier. The form will be sent to the corresponding author when the Editors reach a decision that the manuscript may be potentially publishable. All accepted papers are subject to editorial revision and shortening. Authors should avoid redundancy between sections of text and between illustrations and text. The Editors may recommend that appendices and tables containing extensive data be published in the electronic version of The Journal of Pediatrics only and referenced in a footnote in the print edition. Research papers should include a statement that approval for human research was received from the appropriate board. Papers describing research involving human subjects should indicate that written informed consent was obtained from the parents or guardians of the children who served as subjects of the investigation and, when appropriate, from the subjects themselves. In the event that either the Editor or the reviewers question the propriety of the human investigation with respect to the risk to the subjects or to the means by which informed consent was obtained, The Journal of Pediatrics may request more detailed information about the safeguards employed and the procedures used to obtain informed consent. Copies of the minutes of the committees that reviewed and approved the research also may be requested. Authors should verify compliance with the Health Insurance Portability & Accountability Act of 1996 (HIPAA) prior to submission. Conflict of Interest Authors should disclose on the title page of the paper any potential, perceived, or real conflict of interest, especially any financial arrangement (e.g., grants received, advisory board memberships, share holdings) with a company whose product is discussed in the manuscript. If the article is accepted for publication, the disclosure statement may be published. Trade names of drugs and other products must not appear in the article title. The trade name may appear once in the abstract and once in the introduction or methods section; all other mention of the product must be in the form of the generic name (see Drug Nomenclature). Release to Media It is a violation of the copyright agreement to disclose the findings of an accepted manuscript to the media or the public before publication in The Journal of Pediatrics. Information in the manuscript may be announced when it is published on The Journal web site. Preparation of Manuscripts Manuscripts are to be submitted via Editorial ManagerÔ, the electronic submission website at Authors should review carefully the Authors Tutorial for Editorial ManagerÔ at The Journal of Pediatrics January A

14 account/authtutorial.pdf. Manuscripts must adhere to the standard layout and length guidelines. If electronic submission is not possible please contact Mrs. Alice Landwehr, Managing Editor, The Journal of Pediatrics, Children s Hospital Medical Center, 3333 Burnet Ave, MLC 3021, Cincinnati, OH Phone: ; Fax: , journal.pediatrics@cchmc.org. Format. Original articles should not exceed 6 published pages; do not exceed 18 manuscript pages, including the title page, references, and tables. Figures are calculated at 3 per printed page. Failure to comply with length restrictions may result in a delay in the processing of your paper. Potential Reviewers. To assist with a prompt, fair review process, authors should provide in the submission letter the names, complete addresses, fax numbers, and addresses of 5 to 7 potential reviewers with the appropriate expertise to evaluate the manuscript. Failure to provide 5 to 7 potential reviewers may result in delays in the processing of your paper. Letter of Submission. A Letter of Submission must accompany the paper and provide the following information in accordance with the Uniform Requirements for Manuscripts Submitted to Biomedical Journals: Writing and Editing for Biomedical Publication available at d Disclosure of any prior publications or submissions with any overlapping information; a copy of the work(s) must be provided; d A statement that the work is not and will not be submitted to any other journal while under consideration by The Journal of Pediatrics; d d A statement of any potential conflict of interest, real or perceived; this must also appear on the title page; A statement that all the authors listed on the manuscript take full responsibility for the manuscript; if more than 6 authors, an explanation of the contributions of each author must be provided. Title Page. The title page should include authors full names and academic degrees; departmental and institutional affiliations of each author; and sources of financial assistance or potential conflicts of interest, if any. Listed authors should include only those individuals who have made a significant, creative contribution to the manuscript; a list of more than 6 authors must be justified to the Editors in the Letter of Submission. One author must be designated as the correspondent, with complete address, business telephone number, fax number, and address. Include a list of key words not in the title and a short, running title. Proofs and order forms for reprints will be sent to the corresponding author, if the paper is published. The corresponding author is responsible for communicating with the Editorial Office and all other co-authors. Abstract. Full-length papers for the Original Articles section of The Journal of Pediatrics must include a structured abstract of 200 words or less, to appear after the title page, in the general outline described by the Ad Hoc Working Group for Critical Appraisal of the Medical Literature (Ann Intern Med 1987; 106: and 1990; 113:69-76). The abstract must contain the following headings: Objective(s), Study design, Results, and Conclusion(s). The objective(s) reflects the purpose of the study, that is, the hypothesis that is being tested. The study design should include the type of study, the setting for the study, the subjects (number and type), the treatment or intervention, principal outcomes measured, and the type of statistical analysis. The results section should include the outcome of the study and statistical significance if appropriate. The conclusion(s) states the significance of the results. In lieu of the structured abstract, Clinical and Laboratory Observations manuscripts should include a brief summation of 50 words, without headings. Laboratory Values. Laboratory values should be described in metric mass units. The International System of Units (SI units) can be provided in parentheses immediately after metric units. Conversion tables are available (see JAMA 1986; 255: or Ann Intern Med 1987; 106:114-29). Drug Nomenclature. Drugs should be described in both the United States Adopted Name (USAN) and International Nonproprietary Name (INN) nomenclature. At first usage cite the USAN with the INN in parentheses; subsequent appearances should use the USAN only. References. References must be numbered according to order of appearance in the text and use superscript or parenthesized numbers in the text. For reference style, follow the format set forth in Uniform Requirements for Manuscripts Submitted to Biomedical Journals ( with journal abbreviations according to Cumulated Index Medicus. If reference is to an abstract, letter, or editorial, place the appropriate term in brackets after the title. Examples of references. (if 6 or fewer authors or editors, list all; if 7 or more, list first 6 and add et al): For Journal articles: Kramarz P, DeStefano F, Gargiullo PM, Chen RT, Lieu TA, Davis RL, et al. Does influenza vaccination prevent asthma exacerbations in children? J Pediatr 2001; 138: Cozzi F, Morini F. Possible mechanisms of pacifier protection against SIDS [letter]. J Pediatr 2001;138:783. For books: Rosenstein BJ, Fosarelli PD. Pediatric pearls: the handbook of practical pediatrics. 3rd ed. St Louis: Mosby; Virginia Law Foundation. The medical and legal implications of AIDS. Charlottesville (VA): The Foundation; For chapters in books: Neufeld EF, Muenzer J. The mucopolysaccharidoses. In: Scriver CR, Beaudet AL, Sly WS, et al, eds. The metabolic and molecular bases of inherited diseases. New York: McGraw-Hill; p For web sites: American Medical Association [homepage on the Internet]. Chicago: The Association; c [updated 2001 Aug 23; cited 2002 Aug 12]. AMA Office of Group Practice Liaison; [about 2 screens]. Available from: pub/category/1736.html Tables. Tables are to be uploaded into Editorial ManagerÔ as separate documents. Each table must be double-spaced and be saved in Editorial ManagerÔ as a separate file. A concise title should be supplied for each. Tables should be self-explanatory and 10A January 2005 The Journal of Pediatrics

15 should supplement, not duplicate the text. If a table or any data therein have been previously published, a footnote must give full credit to the original source. (See Permissions. ) Figure Legends. Each illustration must be provided with a legend. Legends should be double-spaced on a separate page within the main document file following the references page. If an illustration has been previously published, the legend must give full credit to the original source. (See Permissions. ) Illustrations. A reasonable number of black-and-white illustrations will be reproduced at no cost to the authors, but the Editors retain the right to edit or delete illustrations and tables for the sake of brevity. Authors are expected to pay the extra cost associated with reproduction of color illustrations in the print Journal. If the Editors approve, the author can choose not to pay for color printing and the illustration can appear in color in the online journal only at no additional cost (see below). Figure legends must be separate from the figures. (See Figure Legends above.) If you are unable to upload illustrations into Editorial Manager, all black and white photographs must be sent as glossy prints, with good light-to-dark contrast; include an exactly matching diskette (see Illustrations in Electronic Format). The minimum size is inches and the maximum is inches; consistency of size within the article is preferred. Histologic figures must have a size bar and indication of magnification. Submit 4 sets of glossy prints, unmounted, numbered, and marked lightly on the back with author s name, top designated. Computer-generated illustrations must be legible with resolution of 300 dots per inch (dpi). Patterns or shadings must be distinguishable from each other and dark enough for reproduction. Lines, symbols, and letters must be smooth and complete. Each figure must be a separate file when uploading into Editorial ManagerÔ. Submit 4 laser-printed originals and a matching diskette if sent by mail; photocopies, dot matrix prints, and photographic halftones are unacceptable. Illustrations may also be original drawings in black ink with typographic lettering; typewritten or freehand lettering is unacceptable. Four sets of original illustrations are required if sent by mail; for freehand drawings, send glossy photographic prints. Do not send original artwork, imaging films, ECG strips, etc. Color illustrations are acceptable. Note that the colors must be dark enough and of sufficient contrast for reproduction. Fluorescent colors do not reproduce well. Avoid using color descriptors in the figure legends. The cost of color illustrations must be borne by the author(s). After final acceptance the publisher will contact authors with pricing and instructions for payment. If the Editors determine that color illustrations will be clear in black and white, the illustrations can be published in black and white in the print version and in color in the online version only at no cost to the authors. Illustrations in Electronic Format. For figures submitted in electronic format, including color illustrations, all images should be at least 5 inches wide. Images may be provided in a variety of formats. For complete instructions, please go to the Author Gateway at: and click on Artwork Instructions. We will need print-quality hard copies of all illustrations only if the paper is accepted for publication. Permissions. Direct quotations, tables, or illustrations that have appeared in copyrighted material must be accompanied by written permission for their use from the copyright owner and original authors along with complete information as to source. Written permission from the patient, or parent or guardian of a minor child, is required for publication of photographs that include the upper portion of the face; black bars over the eyes are not sufficient. Patient initials should not be used anywhere in the text, tables, or figures. Articles appear in both the print and online versions of The Journal of Pediatrics, and wording of the letter should specify permission in all forms and media. Failure to obtain permission rights for electronic publication may result in the images not appearing in the online version. Abbreviations. Complex terms used frequently in the manuscript may be abbreviated. The manuscript should include a list of all abbreviations used. Abbreviations are placed in parentheses at first use in the abstract, and again at first use in the text. Medical Progress Authors considering preparation of a review article for the Medical Progress section should submit a proposal letter and outline to the Editors for approval before submitting the full paper. Invited Commentaries Commentaries are generally invited only. Authors who wish to propose a commentary can submit a proposal letter and outline to the Editors for approval before submitting the full paper. Clinical and Laboratory Observations Papers in this format should fill 3 journal pages or less, the text 1000 words or less with a brief abstract of 50 words or less. A combined total of 2 illustrations and tables, and approximately 10 references, is recommended. Insights Submissions to the Insights section of The Journal should succinctly illuminate clinical problems or solutions of interest to readers and must fit on one published page. Captioned photographs, brief anecdotes or analyses, or even cartoons are welcome; however, a fresh, useful clinical insight must be offered. All material must be original. Text must not exceed 300 words and is subject to shortening if the text and figure(s) do not fit on one published page. Figure(s) and references may be placed in the online only version of The Journal if the piece exceeds one published page. Photograph(s) must be original glossy prints, and artwork must be in the original form (see Illustrations above). Original, signed, written permission from the patient, or parent or guardian of a minor child, is required for publication of recognizable photographs in all forms and media. (See Permissions above.) Contributors will be required to sign a standard copyright transfer agreement; therefore, all submissions must have a title. Submissions will undergo review by the Editors, and their decision to accept or reject will be final (figures from rejected contributions will be returned). Grand Rounds Authors for Grand Rounds should send a proposal letter before submission to ensure that their topic is not covered by a manuscript already in process. Manuscripts for the Grand Rounds section may be prepared in traditional clinicopathologic conference (CPC) style or as a didactic discussion. Editorial Correspondence Letters pertaining to papers published in The Journal within the past year or to related topics should not exceed 300 words. Letters must be double-spaced. References, including reference to The Journal of Pediatrics January A

16 the pertinent article in The Journal, should conform to style for manuscripts (see previous page). Inquiries Regarding Decisions All inquiries concerning manuscript decisions should be in writing; please see contact information above. The complete manuscript file will be forwarded to the appropriate Associate Editor for response to the inquiry. The Editors are not available for telephone calls regarding decisions. AMSPDC Section Pages of The Journal are reserved for the Association of Medical School Pediatric Department Chairs, Inc., (AMSPDC) which is solely responsible for their content. Only authors interested in this section should contact AMSPDC directly. All other papers must be submitted as detailed above. Aaron L. Friedman, M.D. Sylvia Kay Hassenfeld Professor and Chairman Department of Pediatrics Brown Medical School Hasbro Children s Hospital 593 Eddy Street Providence, RI Phone (401) News Items Announcements of scheduled meetings, symposia, or postgraduate courses of interest to the pediatric readership may be sent to the publisher for consideration at least 5 months in advance of the meeting date. News items of general interest to pediatricians and related specialists will also be considered. Books for Review The Journal of Pediatrics does not publish book reviews. Books sent to the Editor cannot be returned. d Letter of submission Checklist d Names and complete contact information for 5-7 suggested reviewers d d d d d d d d Statements that all authors take full responsibility for the paper and that the paper is not nor will be submitted to any other journal while being considered by The Journal of Pediatrics d Disclosures of any conflicts of interests and prior publications Title page d Title of article d Full name(s), academic degrees, and affiliations of authors d Name, address, address, telephone and fax numbers of corresponding author d Name of reprint request author or notation of no reprints d List of key words not in the title d Source of funding and conflict of interest statement, if applicable d Short running title (<12 words) Abstract (double-spaced), structured (200 words) for Original Article or unstructured (50 words) for Clinical and Laboratory Observations Article proper (double-spaced), including d List of abbreviations (double-spaced) d References (double-spaced), on a separate page d Figure legends (double-spaced), on a separate page Tables including title (double-spaced), each on a separate page, saved as a separate file Illustrations, each saved as a separate file; if mailed, properly labeled (4 sets of glossy prints) Letter(s) of permission to reproduce previously published material in all forms and media must be mailed Letters of permission to publish patient photographs in all forms and media must be mailed 12A January 2005 The Journal of Pediatrics

17 THE JOURNAL OF PEDIATRICS Information for Readers Correspondence. Communications regarding original articles and editorial management should be addressed to William F. Balistreri, MD, The Journal of Pediatrics, Children s Hospital Medical Center, Sabin Center, 3rd Floor, 3333 Burnet Ave, Cincinnati, OH ; telephone Customer service. (orders, claims, online, change of address): Elsevier Periodicals Customer Service, 6277 Sea Harbor Drive, Orlando, FL Tel: (800) (U.S. and Canada); (407) (outside U.S. and Canada). Fax: (800) (U.S. and Canada); (407) (outside U.S. and Canada). elspcs@elsevier.com Address changes must be submitted four weeks in advance. Yearly subscription rates. United States and possessions: Individual $173.00; Institution $455.00; Student/Resident: $ All other countries (prices include airspeed delivery): Individual $213.00; Institution $496.00; Student/Resident $ Single Issues $ To receive student/resident rate, orders must be accompanied by name of affiliated institution, date of term and the signature of program/ residency coordinator on institution letterhead. Orders will be billed at the individual rate until proof of status is received. Current prices are in effect for back volumes and back issues. Further information on this journal is available from the Publisher or from this journal s website ( Information on other Elsevier products is available through Elsevier s website ( Advertising information. Advertising orders and inquiries can be sent to: USA, Canada, and South America, M. J. Mrvica Associates, 2 West Taunton Ave, Berlin, NJ 08009; phone: ; FAX: Classified advertising orders and inquiries can be sent to Tracy Peppers, Elsevier Inc., 360 Park Ave South, New York, NY ; phone: ; FAX Europe and the rest of the world, Julie Toop; phone +44 (0) ; fax +44 (0) ; media@elsevier.com. Author inquiries. For inquiries relating to the submission of articles (including electronic submission where available), please visit Elsevier s Author Gateway at The Author Gateway also provides the facility to track accepted articles and set up alerts to inform you of when an article s status has changed, as well as detailed artwork guidelines, copyright information, frequently asked questions, and more. Please see Information for Authors for individual journal. Contact details for questions arising after acceptance of an article, especially those relating to proofs, are provided after registration of an article for publication. The paper used in this publication meets the requirements of ANSI/NISO Z (Permanence of Paper). Reprints. For queries about author offprints, authorsupport@ elsevier.com. To order 100 or more reprints for educational, commercial, or promotional use, contact the Commercial Reprints Department, Elsevier Inc., 360 Park Avenue South, New York, NY Fax: (212) ; reprints@elsevier.com. Reprints of single articles available online may be obtained by purchasing Pay-Per-View access for $30 per article on the journal website, Copyright ª 2005 by Elsevier Inc. Microform edition available from ProQuest Information and Learning, 300 N Zeeb Rd, Ann Arbor, MI ª 2005 Elsevier Inc. All rights reserved. This journal and the individual contributions contained in it are protected under copyright by Elsevier Inc., and the following terms and conditions apply to their use: Photocopying. Single photocopies of single articles may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for nonprofit educational classroom use. Permissions may be sought directly from Elsevier s Rights Department in Philadelphia, PA, USA: phone (+1) , fax (+1) , healthpermissions@elsevier.com. Requests may also be completed online via the Elsevier hoepage ( locate/permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) , fax: (+1) (978) , and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) ; fax: (+44) Other countries may have a local reprographic rights agency for payments. Derivative works. Subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. Permission of the Publisher is required for resale or distribution outside the institution. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic storage or usage. Permission of the Publisher is required to store or use electronically any material contained in this journal, including any article or part of an article. Except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier Rights Department, at the fax and addresses noted above. Notice. No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. Although all advertising material is expected to conform to ethical (medical) standards, inclusion in this publication does not constitute a guarantee or endorsement of the quality or value of such product or of the claims made of it by its manufacturer. The Journal of Pediatrics January A

18 ANNOUNCEMENTS January 2005 American Medical Seminars, Inc., Pediatrics for Practitioners and Emergency Medicine, January 24-28, Sarasota, Florida. Credit: 20 hours Category 1 AMA credit and AAP certification pending. For futher information, please contact: American Medical Seminars, Inc., Mr. D. Reese Piere, CME Director, PO Box 49947, Sarasota, FL ; phone; (toll-free); FAX: ; Web site: Miami Children s Hospital, 40th Annual Pediatric Postgraduate Course: Perspectives in Pediatrics, January 28 to February 3, 2005, Fountainebleau Hilton Resort and Towers, Miami Beach, Florida. Approximately 1,000 pediatricians from around the world will gather for the longest-running, independent continuing medical education program in the United States. Topics include: Adolescent Medicine, Cardiovascular, Asthma, ADHD, Orthopedics, Vaccines, Dermatology, Hematology, Practice Management, Billing and Coding Compliance, Board Review, and Community Pediatrics. For further information see the Web site: or call February 2005 American Heart Association s Scientific Councils on Cardiovascular Disease in the Young, Clinical Cardiology, and Cardiovascular Radiology and Intervention, and cosponsored by the Kawasaki Disease Foundation, the Japan Kawasaki Disease Research Center, and the Japan Heart Foundation, 8th International Symposium on Kawasaki Disease, February 17-20, 2005, Omni San Diego Hotel, San Diego, California. This 3-day conference intends to identify areas of progress in Kawasaki disease research, address controversies surrounding the modification of the current clinical case definition, and introduce new therapies for treatment of acute vasculitis. Physicians and researchers are encouraged to submit abstracts relating to Kawasaki disease; abstracts will be received between August 25 and October 1, For further information, please contact: Andrea Zabkar, Conference Coordinator, American Heart Association, scientificconferences@heart.org; Web site: www. myamericanheart.org. American Medical Seminars, Inc., Issues in Adult and Pediatric Emergency Medicine, February 21-25, Sarasota, Florida. Credit: 20 hours Category 1 AMA credit and 10 hours of AAP certification pending. For further information, please contact: American Medical Seminars, Inc., Mr. D. Reese Pierce, CME Director, PO Box 49947, Sarasota, FL ; phone; (toll-free); FAX: ; Web site: www. ams4cme.com. March 2005 American Medical Seminars, Inc., Review and Update of Pediatric Emergency Medicine, March 28-April 1, 2005, Sarasota, Florida. Credit: 20 hours Category 1 AMA credit and AAP certification pending. For further information, please contact: American Medical Seminars, Inc., Mr. D. Reese Pierce, CME Director, PO Box 49947, Sarasota, FL ; phone; (toll-free); FAX: ; Web site: www. ams4cme.com. April 2005 American Medical Seminars, Inc., Pediatric Emergency Medicine: A Review and Update, April 25-29, 2005, Sarasota, Florida. Credit 20 hours Category 1 AMA Credit and AMA Certification Pending. For further information, please contact: American Medical Seminars, Inc., Mr. D. Reese Pierce, CME Director, PO Box 49947, Sarasota, FL ; phone: (toll-free); FAX: ; Web site: www. ams4cme.com. 14A January 2005 The Journal of Pediatrics

19 FELLOWSHIPS Pediatric subspecialities and general academic pediatric training. Listing of fellowships is based on information supplied by the sponsoring institution and does not imply approval by an accrediting agency, if applicable. Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone ADOLESCENT MEDICINE California: Los Angeles Child Hosp/USC June 30, 2005 July No C Warf/ Colorado: Denver Univ Colo Health Sci Ctr/ Child Hosp Nov 1, 2005 July 3 Yes C Stevens-Simon/ California: San Francisco Univ Calif July 3 3 Yes C Irwin/ Florida: Miami Child Hosp Dec 30, 2005 July 3 No L Siqueira/ Indiana: Indianapolis Ind Univ School Med March 30, 2005 July Yes DP Orr/ Massachusetts: Boston Child Hosp June 1, 2005 July 3 Yes SJ Emans/ Minnesota: Minneapolis Univ Minn Med School Dec 31, 2005 Sept 2 3 Yes N Singh/ New York: Brooklyn State Univ NY Downstate Med Ctr (HSC) July No A Suss/ New York: New York Mount Sinai Hosp, Adolescent Health Ctr July 30, 2005 July 3 No S Edwards/ New York: Rochester Strong Mem Hosp/Golisano Child Hosp Strong Sept 1, 2005 July 3 3 Yes RE Kreipe/ Ohio: Cincinnati Cincinnati Child Hosp Med Ctr/Division Adolescent Med March 15, 2005 July 3 No M Spigarelli/ Pennsylvania: Philadelphia Child Hosp March 1, 2005 July 3 Yes DF Schwarz/ Pennsylvania: Pittsburgh Child Hosp/Univ School Med Sept 30, 2005 July 2-3 No P Murray/ Texas: Houston Baylor Coll Med Sept 15, 2005 July 3 3 Yes A Hergenroeder/ Texas: Houston Univ Tex Health Sci Ctr Oct 1, 2005 July 3 3 Yes WL Risser/ Washington: Seattle Univ Wash June 30, 2005 July 3 Yes M Smith/ *Permanent immigrants should contact the departments regarding eligibility. 15A

20 Category State: City Institution Application deadline Service available ADOLESCENT MEDICINE Cont d Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia Sept 1 3 No R Malleson/ ALLERGY/IMMUNOLOGY Alabama: Birmingham Univ Ala March 31, 2006 July 2 Yes TP Atkinson/ California: Los Angeles Univ Calif Dept Pediatrics Feb 1, 2006 July 3 Yes T Chatila/ California: San Francisco Univ Calif July 3 3 Yes D Wara/ Colorado: Denver Univ Colo Health Sci Ctr Nat Jewish Med Res Ctr March 30, 2005 July 2 3 AH Liu/ Connecticut: New Haven Yale Univ School Med July 3 3 Yes MK Hostetter/ Florida: St. Petersburg Univ South Fla/All Child Hosp July 2 3 No N Tangsinmankong/ Louisiana: New Orleans La State Univ Health Sci Ctr June 30, C Moore Louisiana: Shreveport La State Univ Health Sci Ctr March 31, 2005 July 2 No S Bahna/ Illinois: Chicago Child Mem Hosp/ Northwestern Univ July 2 3 Yes P Greenberger/ Massachusetts: Boston Child Hosp/Harvard Med School July 2-3 H Oettgen/ Nebraska: Omaha Creighton Univ Allergy Division Feb 2005 July 2 No TB Casale/ New York: Brooklyn Long Island Coll Hosp/Dept Allergy Immunology Dec 2006 July 2 No A Schneider/ New York: Brooklyn State Univ NY Downstate Med Ctr (HSC) July No H Moallem/ New York: New York Mount Sinai School Med Feb 1, 2005 July 2 3 No HA Sampson/ North Carolina: Durham Duke Univ Med Ctr July 3 AW Burks/ North Carolina: Winston-Salem Wake Forest Univ Baptist Med Ctr Dec 1, 2005 July 2 No S Peters/ Ohio: Cincinnati Child Hosp Med Ctr July 3 3 No AH Assa ad/ Ohio: Cleveland Cleveland Clinic Foundation July 2 Yes D Lang/ Pennsylvania: Philadelphia Child Hosp Philadelphia Dec 2005 July 3 3 Yes NA Pawlowski/ Pennsylvania: Pittsburgh Univ Pittsburgh/Child Hosp July 2-3 W Calhoun/ Tennessee: Memphis Univ Tenn Health Sci Ctr Dec 31, 2006 July 2 Yes DB Lew/ *Permanent immigrants should contact the departments regarding eligibility. Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone 16A January 2005 The Journal of Pediatrics

21 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone ALLERGY/IMMUNOLOGY Cont d Texas: Dallas Univ Tex Southwestern Med Ctr July 2 Yes D Khan/ Texas: Houston Baylor Coll Med/Tex Child Hosp Jan 31, 2005 July 2 No WT Shearer/ Virginia: Charlottesville Univ Va HS/Asthma Allergies Diseases Ctr Jan 1, 2006 July 3 3 Yes T Platts-Mills/ Virginia: Richmond Va Commonwealth Univ Health Systems July 2 No L Schwartz/ Washington: Seattle Univ Wash/Child Hosp Regional Med Ctr Oct 1, 2005 July 2 3 Yes CT Furukawa/ Wisconsin: Milwaukee Med Coll Wis Jan 01, 2005 July 2 3 Yes M Zacharisen/ Canada: Vancouver (BC) Child Women s Health Ctr Univ British Columbia July 1 3 No A Ferguson/ Canada: Halifax (NS) Dalhousie Univ May 1, 2005 July 1-2 No A Issekutz/ ANESTHESIA California: Los Angeles Child Hosp Dept Anesthesiology Critical Care Med Nov 30, 2005 July 1 Yes GM Scott/ Colorado: Denver Child Hosp Jan 1, 2006 July 1 No R Agarwal/ Ohio: Cincinnati Child Hosp Med Ctr 1 No CD Kurth/ Ohio: Columbus Child Hosp/Ohio State Univ July 1 3 No A Tingley/ Pennsylvania: Philadelphia St. Christopher s Hosp Child July 1 4 No DA Lowe/ CARDIOLOGY California: Los Angeles Child Hosp/Univ Southern Calif Jan 1, 2005 July 3 3 No AB Lewis/ California: Los Angeles David Geffen School Med UCLA/Mattel Child Hosp Dec 31, 2005 July 3 No V Allada/ California: Palo Alto Stanford Univ/Lucile Packard Child Hosp Jan 30, 2005 July 3 No G VanHare/ California: San Francisco Univ Calif July 3 3 No M Brook/ Colorado: Denver Univ Calif Health Sci Ctr/ Child Hosp July 3 No K Chan/ Connecticut: New Haven Yale Univ School Med Feb 1, 2005 July 3 3 No J Fahey/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

22 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone CARDIOLOGY Cont d Florida: Gainesville Univ Fla/Shands Hosp July 3 3 No KO Schowengerdt/ Florida: Miami Univ/Jackson Mem Hosp Feb 28, 2005 July 3 3 No GS Wolff/ Georgia: Atlanta Emory Univ School Med March 31, No DA Fyfe/ Illinois: Chicago Child Mem Hosp/McGaw Med Ctr May 1, 2005 July 3 3 No WH Franklin/ Illinois: Chicago Univ Chicago July 3 4 No R Abdulla/ Indiana: Indianapolis Ind Univ Riley Hosp Jan 31, R Hurwitz/ Iowa: Iowa City Univ Iowa April 1, 2005 July 3 Yes TD Scholz/ Louisiana: New Orleans Tulane Univ April 30, 2005 July 3 3 No N Ross-Ascuitto/ Massachusetts: Boston Child Hosp Boston/Harvard Med School March 1, 2005 July 3 3 No P Lang/ Michigan: Ann Arbor Univ Mich May 1, 2005 July 3 Yes M Dick II/ Minnesota: Minneapolis Univ Minn Med School Oct 31, 2005 July 3 3 No E Braunlin/ Minnesota: Rochester Mayo Clinic Jan 2006 July 3 3 No AK Cabalka/ Missouri: St. Louis Wash Univ/Child Hosp Feb 28, 2005 July 3 No MC Johnson/ New York: New York Mount Sinai Med Ctr Jan 1, 2005 July 3 3 No I Parness/ New York: New York NY Presbyterian Hosp/ Columbia Univ Med Ctr Feb 1, 2005 July 3 3 No WM Gersony/ New York: Rochester Strong Mem Hosp/Golisano Child Hosp Strong March 15, 2005 July 3 3 No RP Vermilion/ North Carolina: Durham Duke Univ Med Ctr Jan 31, 2005 July 3 No B Armstrong/ Ohio: Cincinnati Child Hosp Med Ctr Feb 28, 2005 July 3 3 No RA Spicer/ Ohio: Columbus Child Hosp/Ohio State Univ March 31, 2005 July No DP Chan/ Oregon: Portland Ore Health/Sci Univ May 1, 2005 July 3 3 Yes GH Burch/ Pennsylvania: Philadelphia Child Hosp Jan 1, 2005 July 3 PM Weinberg/ Pennsylvania: Philadelphia St. Christopher s Hosp for Child April 25, 2005 July 3 3 No L Vogel/ Pennsylvania: Pittsburgh Child Hosp March 1, 2005 July 3 3 No FS Sherman/ *Permanent immigrants should contact the departments regarding eligibility. 18A January 2005 The Journal of Pediatrics

23 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone CARDIOLOGY Cont d Texas: Houston Baylor Coll Med/Tex Child Hosp Feb 15, 2005 July 3 3 No SR Neish/ Ohio: Cleveland Case Western Univ/Rainbow Babies and Child Hosp Jan 31, 2005 July 3 No E Siwik/ Ohio: Cleveland The Cleveland Clinic Foundation March 1, 2005 July 3 3 No GL Rosenthal/ Texas: Dallas Univ Tex Southwestern Med Ctr Feb 15, 2005 July 3 3 No L Mahony/ Texas: Houston Mem Hermann Child Hosp/ Univ Tex Med School July 3 3 Yes PS Rao/ Utah: Salt Lake City Univ Utah April 1, 2005 July 3 Yes S Etheridge/ Virginia: Charlottesville Univ Va/Department Pediatric Cardiology Jan 1, 2006 July Yes GP Matherne/ Virginia: Richmond Va Commonwealth Univ Health System Sept 30, 2005 July 3 3 No WB Moskowitz/ Wisconsin: Milwaukee Med Coll Wis/Child Hosp March 1, 2005 July 3 3 No PC Frommelt/ Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia July No W Duncan/ Canada: Halifax (NS) IWK Health Ctr/Dalhousie Univ July 1 3 No A Warren/ CRITICAL CARE Alabama: Birmingham Univ Ala Sept 2005 July 3 No M Winkler/ Arkansas: Little Rock Univ Ark for Med Sci/Ark Child Hosp Sept 1, 2005 July 3 3 No SM Schexnayder/ California: Loma Linda Loma Linda Univ Med Ctr July 3 3 No S Abd-Allah/ California: Los Angeles Child Hosp July 2005 July 3 3 No C Newth/ California: Los Angeles Univ Calif Dept Pediatrics Sept 15, 2005 July 3 No I Weiss California: San Francisco Univ Calif July 1, 2005 July 3 No SJ Soifer/ California: Torrance Harbor-Univ Calif Med Ctr/ Child Hosp Orange County July 3 3 No RB Mink/ Colorado: Denver Child Hosp/Univ Colo July 15, 2005 July 3 Yes EL Dobyns/ Connecticut: New Haven Yale Univ School Med Oct 1, 2005 July 3 3 No CW Bogue/ Florida: Gainesville Univ Fla Coll Med Aug 15, 2005 July 3 No A Zaritsky/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

24 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone CRITICAL CARE Cont d Florida: Miami Child Hosp Dec 30, 2005 July 3 3 No B Totapally/ Florida: Miami Univ/Jackson Mem Hosp July 3 3 No GP Cantwell/ Florida: Miami Child Hosp Dec 30, 2005 July No A Rossi/ Georgia: Atlanta Emory Univ School Med July 3 3 No A Vats/ Illinois: Chicago Univ Child Hosp July 31, 2005 July 3 No J Downie/ Illinois: Chicago Northwestern Univ s Feinberg School Med/Child Mem Hosp Sept 1, 2005 July 3 3 No D Goodman/ Illinois: Oak Lawn Hope Child Hosp Christ Hosp June 30, 2006 July 3 No L Torero/ Illinois: Park Ridge, IL Advocate Lutheran General Hosp July 3 3 No S Havalad/ Iowa: Iowa City Univ Iowa Hosp Clinics Sept 1, 2005 July 3 Yes JG Moreland/ Kentucky: Louisville Univ/Kosair Child Hosp Oct 1, 2005 July 3 No A Hardin/ Maryland: Baltimore Univ Md Med Systems/ School Med July 3 No V Vaidya/ Massachusetts: Boston Child Hosp/Harvard Med School Aug 15, 2005 July 3 No JP Burns/ Michigan: Detroit Child Hosp Mich/Wayne State Univ June 2005 July 3 No MW Lieh-Lai/ Minnesota: Minneapolis Univ Minn Med School July 3 3 No DN Cornfield/ Missouri: St. Louis Wash Univ School Med Oct 1, 2005 July 3 No MI Goldsmith/ New York: New York NY Presbyterian Hosp/ Columbia Univ July 3 3 KV Biagas/ New York: New York NY-Presbyterian Hosp/Weill Cornell Med Ctr Sept 15, 2005 July 3 3 Yes HM Ushay/ New York: Rochester Strong Mem Hosp/Golisano Child Hosp Strong Sept 30, 2005 July 3 3 Yes H Connolly/ North Carolina: Durham Duke Child Hosp Sept 15, 2005 July 3 3 No C Weldon/ Ohio: Cleveland Rainbow Babies Child Hosp/ Case Western Reserve Univ July 4 3 No M Anderson/ Ohio: Columbus Child Hosp Ohio State Univ July 1, 2005 July 3 3 No JA Fitch/ *Permanent immigrants should contact the departments regarding eligibility. 20A January 2005 The Journal of Pediatrics

25 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone CRITICAL CARE Cont d Ohio: Cincinnati Child Hosp Med Ctr/Univ Cincinnati Aug 1, 2005 July 3 3 No HR Wong/ Pennsylvania: Philadelphia Child Hosp Sept 30, 2005 July 3 V Nadkarni/ Pennsylvania: Pittsburgh Child Hosp/Univ Pittsburgh Oct 2005 July 3 3 No R Clark/ Puerto Rico: San Juan Univ Puerto Rico/School Med Dec 15, 2005 July 3 No RL Garcia/ Tennessee: Memphis Univ Tenn Health Sci Ctr/Le Bonheur Child Med Ctr July 3 3 No S Storgion/ Texas: Dallas Univ Tex Southwestern Med Ctr Sept 15, 2005 July 3 3 No SG Kernie/ Texas: Houston Univ Tex Health Sci Ctr Oct 1, 2005 July 3 3 Yes B Cua/ Texas: Dallas Texas: Houston Baylor Coll Med/Tex Child Hosp Sept 2005 July 3 No ML McPherson/ Texas: San Antonio Univ Tex Health Sci Ctr July 2005 July 3 3 Yes R Taylor/ Utah: Salt Lake City Univ Utah/Primary Child Med Ctr Aug 15, 2005 July 3 3 No MK Witte/ Virginia: Charlottesville Univ Va Health System/ Pediatric Critical Care Jan 1, 2006 July No A Doctor/ Virginia: Richmond Med Coll Va/VCU Health System Oct 2005 July 3 3 No S Sreedhar/ Washington: Seattle Child Hosp and Regional Med Ctr July 3 No H Baden/ Wisconsin: Milwaukee Child Hosp Wis/Med Coll Aug 1, 2005 July No KJ Marcdante/ Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia July No N Froese/ Canada: Ottawa (Ont) Child Hosp Eastern Ontario Oct 01, 2005 July 2 K Menon/ DEVELOPMENTAL/ BEHAVIOR Arkansas: Little Rock Univ Ark Med Sci & Ark Child Hosp Dec 31, 2005 July 3 Yes EG Schulz/ California: Los Angeles Child Hosp July 1 3 Yes KK Imagawa/ California: Los Angeles Univ Calif Department Pediatrics Dec 1, 2005 July 3 3 Yes J Howard/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

26 Category State: City Institution Application deadline Service available DEVELOPMENTAL/BEHAVIOR Cont d Connecticut: New Haven Yale Univ School Med July 3 3 Yes DJ Schonfeld/ Illinois: Chicago Univ/Comer Child Hosp Oct 1, 2005 July 3 3 No ME Msall/ Indiana: Indianapolis Riley Hosp for Child/Ind Univ School Med July 6 No A Dusick/ Maryland: Baltimore Univ Md Med Systems/ School Med Jan 1, 2006 July 3 3 No L Grossman/ Massachusetts: Boston Med Ctr/Univ School Med Nov 1, 2005 July 3 Yes S Parker/ Massachusetts: Boston Child Hosp Sept 30, 2005 July 3 3 Yes L Rappaport/ Minnesota: Minneapolis Univ Minn/Fairview Univ Med Ctr Feb 1, 2006 Sept 3 3 Yes DP Kohen/ New York: Rochester Univ Golisano Child Hosp Strong Dec 1 July 3 3 Yes S Sulkes/ Ohio: Cleveland Rainbow Babies Child Hosp March 1, 2005 July 3 Yes J Kennell/ Ohio: Cincinnati Child Hosp Med Ctr May 1, 2006 July 3 4 No NE Lanphear/ Ohio: Cincinnati Child Hosp Med Ctr Dec 31, 2005 July 4 2 No MB Schapiro/ Ohio: Columbus Child Hosp Ohio State Univ Nov 1, 2005 July 3 3 No DL Coury/ Oregon: Portland Ore Health Sci Univ Oct 1, 2005 July 3 3 Yes PA Blasco/ Oregon: Portland Ore Health Sci Univ Oct 1, 2005 July Yes PA Blasco/ Pennsylvania: Philadelphia Child Hosp Nov 30, 2005 July 3 3 No M Pipan/ Rhode Island: Providence RI Hosp/Brown Med School July 3 3 No P High/ Texas: Dallas Univ Tex Southwestern Med Ctr July 3 3 Yes JL Black/ Virginia: Charlottesville Univ Va/Department Pediatrics Jan 1, 2006 July 3 3 Yes JA Blackman/ Washington: Seattle Univ Wash Oct 2005 July 3 3 Yes WO Walker, Jr/ Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia July No T Oberlander/ Canada: Calgary (Alta) Univ Calgary/Alberta Child Hosp Nov 15 July 2-3 No ME Clarke/ EMERGENCY MEDICINE Alabama: Birmingham Univ Ala Oct 15, 2005 July 3 No M Nichols/ California: Los Angeles Child Hosp Aug 31 July 1 No VJ Wang/ *Permanent immigrants should contact the departments regarding eligibility. Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone 22A January 2005 The Journal of Pediatrics

27 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone EMERGENCY MEDICINE Cont d California: Torrance Harbor-Univ Calif Med Ctr/ Univ Calif School Med Nov 1, 2005 July Yes M Gausche-Hill/ Colorado: Denver Univ Colo Health Sci Ctr Aug 19, 2005 July 3 3 No M Roback/ Connecticut: New Haven Yale Univ School Med Sept 1, 2005 July 3 3 No K Santucci/ Wilmington: Delaware Thomas Jefferson Univ/ DuPont Hosp Children Sept 1, 2005 July 3 3 No M Attia/ Florida: Miami Child Hosp Dec 30, 2005 July 3 No M Linares/ Georgia: Atlanta Emory Univ School Med Aug 31 June 3 3 No H Simon/ Illinois: Chicago Child Mem/Northwestern Univ Aug 19, 2005 July 3 3 No E Powell/ Illinois: Chicago Univ Child Hosp Sept 1, 2005 July 3 Yes MA Hostetler/ Kentucky: Louisville Kusari Child Hosp/Univ Louisville Aug 31, 2006 July 3 No IK Kim/ Massachusetts: Boston Med Ctr/Boston Univ School Med Aug 1, 2005 July 3 3 Yes E Calmar/ Massachusetts: Boston Child Hosp Aug 26, 2005 July 3 3 No R Bachur/ Michigan: Detroit Child Hosp Mich/Wayne State Univ June 30, 2005 July 2-3 No N Bhaya/ Minnesota: Minneapolis Univ Minn/Child Hosp Clinics Nov 1, 2005 July 3 No P Fink/ Missouri: St. Louis Child Hosp Sept 2005 July 3 3 Yes D Jaffe/ New York: New York NY-Presbyterian Hosp/ Columbia Campus Aug 30, 2005 July 3 3 No P Dayan/ New York: New York NY Presbyterian Hosp/Weill Cornell Med Ctr Sept 15, 2005 July 3 3 Yes S Platt/ New York: Rochester Strong Mem Hosp/Golisano Child Hosp Strong Sept 1, 2005 July Yes C Davis/ Ohio: Cincinnati Child Hosp Med Ctr Aug 31, 2005 July 3 No CM McAneney/ Ohio: Columbus Child Hosp/Ohio State Univ Sept 1, 2005 July 3 3 No MJ Bowman/ Pennsylvania: Philadelphia Drexel Univ Coll Med Oct 6, 2005 July No C Mull/ Pennsylvania: Philadelphia Child Hosp/Univ Pennsylvania Aug 15, 2005 July 3 3 Yes J Fein/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

28 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone EMERGENCY MEDICINE Cont d Pennsylvania: Pittsburgh Child Hosp Sept 1, 2005 June 3 RD Pitetti/ Rhode Island: Providence RI Hosp/Hasbro Child Hosp Sept 1, 2005 July 3 No DW Steele/ Tennesse: Memphis Univ Tenn/LeBonheur Child Hosp Sept 30, 2006 July 3 4 No C Forsythe/ Texas: Dallas Univ Tex Southwestern Med Ctr July 3 S Scott/ Texas: Houston Baylor Coll Med/Tex Child Hosp Sept 1, 2005 July 3 3 No CG Macias/ Utah: Salt Lake Univ Utah School Med Aug 20, 2005 July 3 3 Yes K Rittichier/ Virginia: Norfolk Eastern VA Med School/ Child Hosp The King s Daughters Sept 1, 2005 July 3 3 No MP Poirier/ Washington: Seattle Univ Washington/Child Hosp Regional Med Ctr Sept 1, 2005 July 3 3 No EJ Klein/ Wisconsin: Milwaukee Med Coll Wis/Child Hosp Wis Sept 15, 2004 July Yes C Walsh-Kelly/ Canada: Calgary (Alta) Child Hosp April 1, 2005 July 2 3 No K Millar/ Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia July No N. Dehghani Canada: Ottawa (Ont) Univ Ottawa May 1, 2005 July 3 3 No G Neto/ ENDOCRINOLOGY/ METABOLISM Alabama: Birmingham Univ Ala Department Pediatric Endocrinology March 31, 2006 July 3 K McCormick/ Arizona: Phoenix Child Hosp/Maricopa Med Ctr Dec 31, 2005 July 3 3 No K Hasan/ Arkansas: Little Rock Univ Ark Med Sci May 1, 2005 July 3 No JP Frindik/ California: Los Angeles Child Hosp April 1, 2005 July 3 3 No ME Geffner/ California: Los Angeles Mattel Child Hosp Univ Calif March 15, 2005 July 3 Yes P Cohen/ California: Palo Alto Stanford Univ School Med Feb 1, 2005 July 3 3 No LK Bachrach/ California: San Francisco Univ Calif 3 3 Yes WL Miller/ California: Torrance Harbor-Univ Calif Med Ctr/ Univ Calif School Med July 3 3 Yes J Brasel/ *Permanent immigrants should contact the departments regarding eligibility. 24A January 2005 The Journal of Pediatrics

29 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone ENDOCRINOLOGY/METABOLISM Cont d Colorado: Denver July 3 3 Yes PS Zeitler/ Connecticut: Hartford CT Child Med Ctr/Univ CT School Med Sept 1, 2005 July 3 3 No E Estrada/ Connecticut: New Haven Yale Univ School Med July 3 3 T Carpenter/ Florida: Gainesville Univ Fla Coll Med Department Pediatrics Oct 15, 2005 July 3 3 No JH Silverstein/ Florida: Miami Univ Miami/Jackson Mem Hosp 3 3 No G Berkovitz/ Georgia: Atlanta Emory Univ School Med June 3 3 No JS Parks/ Illinois: Chicago Univ Chicago Dec 31, 2005 July 3 Yes S Radovick/ Indiana: Indianapolis Ind Univ/Riley Hosp for Child July 3 3 No J Fuqua/ Iowa: Iowa Univ Iowa Hosp Clinics July 3 No E Tsalikian/ Louisiana: New Orleans La State Univ Prog July 1-3 No S Chalew/ Maryland: Bethesda National Institutes of Health July 3 No C Stratakis/ Massachusetts: Boston Child Hosp/Joslin Diabetes Ctr Feb 1, 2005 July 3 Yes JA Majzoub/ Michigan: Ann Arbor Univ Mich Health System April 1, 2005 July No D Vazquez/ Minnesota: Minneapolis Univ Minn Med School Jan 1, 2006 July 3 3 A Moran/ Minnesota: Rochester Mayo Clinic July 3 3 No WF Schwenk III/ Missouri: St. Louis Wash Univ School Med/ Child Hosp Dec 31, 2005 July 3 No LJ Muglia/ New York: Brooklyn State Univ NY Downstate Med Ctr (HSC) July No S Castels/ New York: Mineola Winthrop Univ Hosp Dec 1, 2005 July 1 3 M Castro-Magana/ New York: New York Mount Sinai School Med April 15, 2005 July 3 3 R Rapaport/ New York: New York NY-Presbyterian Hosp/ Columbia Univ Dec 31, 2005 July 3 No SE Oberfield/ New York: New York NY-Presbyterian Hosp/Weill Cornell Med Ctr Sept 15, 2005 July 3 3 No New York: Stony Brook State Univ NY Sept 30, 2005 July 3 3 No TA Wilson/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

30 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone ENDOCRINOLOGY/METABOLISM Cont d North Carolina: Durham Duke Univ Med Ctr May 1, 2005 July 3 3 Yes M Freemark/ Ohio: Cincinnati Child Hosp Med Ctr July 3 3 No S Chernausek/ Ohio: Cleveland Rainbow Babies Child Hosp/ Case Western Reserve Univ July 3 3 M Palmert/ Oklahoma: Oklahoma Okla Univ Health Sci Ctr Sept 1, 2005 July 3 3 Yes K Copeland/ Oregon: Portland Ore Health & Sci Univ Dec 31, BA Boston/ Oregon: Portland Ore Health & Sci Univ Oct 1, 2005 July 1-2 No DM Koeller/ Pennsylvania: Philadelphia Child Hosp June 1, 2005 July 3 3 Yes CA Stanley/ Pennsylvania: Philadelphia St. Christopher s Hosp for Child July 1, 2005 July 3 3 No F DeLuca/ Rhode Island: Providence RI Hosp/Brown Univ May 1, 2005 July 3 3 No CM Boney/ Tennessee: Memphis Univ Tenn June 30, 2005 July 3 3 No G Burghen/ Texas: Dallas Univ Tex Southwestern Med Ctr March 1, 2005 July 3 3 Yes P White/ Texas: Houston Baylor Coll Med/Tex Child Hosp July 3 3 Yes Virginia: Charlottesville Univ Va Health System Jan 1, 2006 July 3 Yes WL Clarke/ Wisconsin: Madison Univ Wis Hosp and Clinics May 30, 2005 July 3 No DB Allen/ Wisconsin: Milwaukee Med Coll Wis/Child Hosp July 3 No P Palma-Sisto/ Canada: Calgary (Alta) Univ Calgary/Alberta Child Hosp March 2005 July No DK Stephure/ Canada: Vancouver (BC) British Columbia Child Hosp/ Univ British Columbia July 1 3 No AGF Davidson/ Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia March 31, 2005 July 2 3 No R Rothstein/ Canada: Ottawa (Ont) Univ Ottawa March 31, 2005 July 2 3 No S Hadjiyannakis/ GASTROENTEROLOGY/ NUTRITION California: Los Angeles Child Hosp Feb 1, 2005 July 3 3 Yes FR Sinatra/ California: Los Angeles Univ Calif Med Ctr Jan 31, 2005 June 3 3 No ME Ament/ *Permanent immigrants should contact the departments regarding eligibility. 26A January 2005 The Journal of Pediatrics

31 Category State: City Institution Application deadline Service available GASTROENTEROLOGY/NUTRITION Cont d California: San Francisco Univ Calif July No MB Heyman/ California: Stanford Univ Med Ctr July 3 3 No J Kerner/ Colorado: Denver Univ Colo HSC/Child Hosp Dec 1, 2005 July 3 3 Yes JA O Connor/ Connecticut: New Haven Yale Univ School Med Dec 31, 2005 July 3 3 F Daum/ Florida: Gainesville Shands/Univ Fla July 3 3 Yes J Andres/ Georgia: Atlanta Emory Univ School Med July 3 No R Romero/ Illinois: Chicago Child Mem Hosp/ Northwestern Univ/ Feinberg School Med Feb 1, 2005 July 3 3 No BUK Li/ Illinois: Chicago Univ Chicago Dec 1, 2005 July 3 3 No S Guandalini/ Indiana: Indianapolis Ind Univ School Med/Riley Child Hosp Dec 31, 2005 July 3 No M Pfefferkorn/ Iowa Univ Iowa July 3 3 Yes WP Bishop/ Kentucky: Louisville Kosair Child Hosp/Univ Oct 1, 2005 July 3 3 No TC Stephen/ Louisiana: New Orleans La State Univ Health Sci Ctr July 1 No E Schmidt-Sommerfeld/ Massachusetts: Boston Child Boston/MGH/HMS July 3 3 No WI Lencer/ Michigan: Ann Arbor Univ Mich/C.S. Mott Child Hosp July 3 No MJ Lopez/ Michigan: Detroit Child Hosp Mich/Wayne State Univ 3 No V Tolia/ Minnesota: Minneapolis Univ Minn Med School July 3 Yes HL Sharp/ Minnesota: Rochester Mayo Clinic Dec 31, 2005 July 3 3 No M El-Youssef/ Missouri: St. Louis Child Hosp/Wash Univ July 3 3 No PI Tarr/ New York: Brooklyn State Univ NY Downstate Med Ctr (HSC) July No S Fisher/ New York: New York Child Hosp NY-Presbyterian/ Columbia Univ Med Ctr July 3 No WR Treem/ New York: New York Mt Sinai School Med Feb 1, 2005 July 3 No K Benkov/ New York: New York NY-Presbyterian Hosp/Weill Cornell Med Ctr Oct 1, 2005 July 3 3 No R Sockolow/ New York: Rochester Univ Rochester March 31, 2005 July 3 No T Rossi/ *Permanent immigrants should contact the departments regarding eligibility. Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone The Journal of Pediatrics January A

32 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone GASTROENTEROLOGY/NUTRITION Cont d Ohio: Cincinnati Child Hosp Med Ctr July 3 3 Yes MB Cohen/ Ohio: Cleveland Clinic Foundation July 3 3 No L Mahasan/ Ohio: Cleveland Rainbow Babies Child Hosp July 3 No G Chelimsky/ Ohio: Columbus Child Hosp Ohio State Univ July 3 3 No SH Erdman/ Pennsylvania: Pittsburgh Child Hosp July 3 3 No ME Lowe/ Rhode Island: Providence RI Hosp/Hasbro Child Hosp July 3 3 No NS LeLeiko/ Texas: Dallas Univ Tex Southwestern Med Ctr Dec 15, 2005 July 3 No N Rodriguez-Baez/ Texas: Houston Baylor Coll Med/Tex Child Hosp July 3 No MA Gilger/ Wisconsin: Milwaukee Med Coll Wis Nov 1, 2005 July 3 3 No SL Werlin/ Canada: Calgary (Alta) Univ Calgary/Alberta Child Hosp March 31, 2005 July No L McKenzie Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia July 3 3 No RA Schreiber/ GENERAL PEDIATRICS California: Los Angeles Child Hosp July 1 3 Yes M Bryant/ California: San Francisco Univ Calif July Yes RH Pantell/ Colorado: Denver Univ Colo Health Sci Ctr/ School Med July 30, 2005 July 2 No A Kempe/ Illinois: Chicago Child Mem Hosp/Feinberg School Med/Northwestern Univ Sept 1, 2005 July Yes R Listernick/ Illinois: Chicago Univ Child Hosp/La Rabioa Dec 1, 2005 July 3 3 No BH Rich/ Illinois: Chicago Univ Ill Med Ctr July No LG Niederman/ Indiana: Indianapolis Riley Hosp for Child Feb 30, 2005 July 2 No A Carroll/ Massachusetts: Boston Boston Univ School Med/ Boston Med Ctr Nov 1, 2005 July Yes H Bauchner/ Massachusetts: Boston Child Hosp Sept 30, 2005 July Yes J Cox/ Minnesota: Minneapolis Univ Minn Med School Dec 31, 2005 Sept Yes I Borowsky/ Minnesota: Rochester Mayo Clinic Dec 31, 2005 July 3 No GF Asay/ *Permanent immigrants should contact the departments regarding eligibility. 28A January 2005 The Journal of Pediatrics

33 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone GENERAL PEDIATRICS Cont d New York: New York Columbia Univ July 30, 2006 July 2-3 Yes M Irigoyen/ New York: New York Mount Sinai School Med July 3 3 D Leraque/ New York: Staten Island St. Vincent Catholic Med Ctr 3 SS Rabinowitz/ New York: Rochester Strong Mem Hosp/Golisano Child Hosp Strong July Yes M Weitzman/ New York: Mineola Winthrop Univ Hosp Jan 1, 2006 July 1 3 No RV Marino/ Ohio: Cincinnati Child Hosp Med Ctr Sept 1, 2005 July 3 3 Yes BP Lanphear/ Pennsylvania: Pittsburgh Child Hosp Nov 1, 2005 July Yes A Hoberman/ Rhode Island: Providence Brown Med School/RI Hosp Dec 31, 2005 July Yes AJ Alario/ Texas: Houston Tex Child Hosp/Baylor Coll Med Dec 31, 2005 July 3 3 Yes RT Kirkland/ Texas: Houston Univ Tex Health Sci Ctr Oct 1, 2005 July 2 3 Yes L Mazur/ Washington: Seattle Univ Wash July 2 3 Yes P Lozano/ Wisconsin: Milwaukee Med Coll Wis July 2-3 Yes G Flores/ GENETICS/ DYSMORPHOLOGY California: Los Angeles Cedars-Sinai Med Ctr Jan 15, 2006 July No DL Bimoin/ California: San Francisco Univ Calif July 2 2 No S Packman/ Colorado: Denver Child Hosp/Univ Colo Health Sci Ctr July 2 2 No JA Thomas/ Florida: Miami Univ Miami/Jackson Mem Hosp April 2006 July 2 2 No LI Elsas/ Iowa: Iowa City Univ Iowa Health Care/Child Hosp Iowa July 2 Yes V Sheffield/ Maryland: Bethesda National Institutes Health July 3 No C Stratakis/ Missouri: St. Louis Wash Univ Nov 1, 2005 July 3 No R Martin/ New York: New York Child Hosp NY-Columbia Univ Jan 31, No K Anyane-Yeboa/ North Carolina: Durham Duke Univ Med Ctr July 1 2 No M McDonald/ Ohio: Cincinnati Cincinnati Child Hosp Med Ctr July 2 3 No HM Saal/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

34 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone GENETICS/DYSMORPHOLOGY Cont d Oregon: Portland Ore Health Sci Univ July 2 2 RE Moses/ Pennsylvania: Philadelphia Child Hosp/Univ Pa July Yes H Kazazian/ Texas: Houston Univ Tex Med School Yes H Northrup/ Massachusetts: Boston Harvard Med School/Child Hosp Boston July 2 M Irons/ Washington: Seattle Univ Wash and Child Hosp July 2 2 Yes PH Byers/ Pennsylvania: Pittsburgh Child Hosp/Univ Pittsburgh 3 No J Voctley/ Texas: Dallas Univ Tex Southwestern Med Ctr April 1, 2006 July 2 3 No L Waber/ Illinois: Chicago Child Mem Hosp/ Northwestern July 2 2 No BK Burton/ Indiana: Indianapolis Department Med Molecular Feuetics/Ind Univ School Med Dec 1, 2006 July 2 No G Vauce/ Utah: Salt Lake City Univ Utah Health System Ctr July 2 2 No JC Curey/ Florida: Miami Division Med Genetics/Univ Miami School Med July 1, 2005 July 2 2 Yes LJ Elsas/ Minnesota: Rochester Mayo Clinic Oct 1, 2005 July 2 2 No J Ellison/ Virginia: Richmond Va Commonwealth Univ March 1, 2006 July 2-3 No JN Bodurtha/ HEMATOLOGY/ ONCOLOGY Alabama: Birmingham Univ Ala Jan 15, 2005 July 3 No R Berkow/ California: Los Angeles Mattel Child Hosp Univ Calif March 31, 2005 July 3 SA Feig/ California: Los Angeles Child Hosp Feb 15, 2005 July 3 3 No W Laug/ California: San Francisco Univ Calif July 3 3 Yes KK Matthay/ Colorado: Denver Univ Colo/Health Sci Ctr April 1, 2005 July 3 3 No RL Garcea/ Connecticut: New Haven Yale Univ School Med April 1, 2005 July 3 3 No D Beardsley/ Florida: Gainesville Univ Fla Shands Child Hosp April 30, 2005 July 3 3 No S Hunger/ Georgia: Atlanta Emory Univ School Med July 3 3 No T Abshire/ Illinois: Chicago Child Mem Hosp/ Northwestern Univ Med School March 31, 2005 July 3 3 No D Walterhouse/ *Permanent immigrants should contact the departments regarding eligibility. 30A January 2005 The Journal of Pediatrics

35 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone HEMATOLOGY/ONCOLOGY Cont d Illinois: Chicago Univ Chicago Jan 1, 2006 July 3 No EC Beyer/ Indiana: Indianapolis Riley Hosp Child/Ind Univ March 31, 2005 July 3 3 No T Vik/ Iowa: Iowa Univ Iowa Hosp and Clinics Feb 28, 2005 July 3 No TW Loew/ Louisiana: New Orleans La State Univ Health Sci Ctr Jan No RP Warrier/ Massachusetts: Boston Child Hosp July 3 Yes SE Lux/ Michigan: Ann Arbor Univ Mich/C.S. Mott Child Hosp Jan 1, 2005 July 3 No D Wechsler/ Michigan: Detroit Wayne State Univ/Detroit Med Ctr/Child Hosp Mich Feb 28, 2005 July 3 No Y Ravindranath/ Michigan: Grand Rapids DeVos Child Hosp Jan 31, 2006 July 3 No AS Cornelius/ Minnesota: Minneapolis Univ Minn April 1, 2005 July 3 3 No KS Baker/ Minnesota: Rochester Mayo Clinic Jan 2005 July 3 3 No S Khan/ Missouri: St. Louis Wash Univ/Child Hosp March 1, 2005 July 3 3 No RJ Hayashi/ New York: Brooklyn State Univ NY Downstate Med Ctr (HSC) July No S Rao/ New York: New York NY Presbyterian/Columbia Univ July 3 2 No MS Cario/ New York: New York NY Presbyterian Hosp/Mem Sloan Kettering Cancer Ctr March 1, 2005 July 3 3 No P Giardina/ New York: New York Schnerde Child Hosp April 1, 2005 July 3 3 No JM Lipton/ New York: Rochester Strong Mem Hosp/Golisano Child Hosp Strong March 1, 2005 July 3 3 Yes AS Hinkle/ North Carolina: Durham Duke Univ Hosp Jan 16, 2005 July 3 Yes SG Kreissman/ Ohio: Cincinnati Child Hosp Med Ctr March 30, 2005 July 3 3 No DA Williams/ Ohio: Cleveland Rainbow Babies Child Hosp Jan 31, 2005 July 3 No S Alexander/ Ohio: Columbus Child Hosp Ohio State Univ April 1, 2005 July 3 3 No K Klopfenstein/ Oregon: Portland Ore Health Sci Univ Feb 1, 2005 July 3 LC Stork/ Pennsylvania: Philadelphia Child Hosp March 1, 2005 July 3 2 No JM Maris/ Pennsylvania: Philadelphia St Christopher s Hosp Child March 1, 2005 July 3 3 Yes C Dampier/ Pennsylvania: Pittsburgh Child Hosp March 1, 2005 July 3 3 No AK Ritchey/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

36 Category State: City Institution Application deadline Service available HEMATOLOGY/ONCOLOGY Cont d Rhode Island: Providence RI Hosp/Brown Med School Feb 14, 2005 July 3 3 No S Meech/ Tennessee: Memphis St. Jude Child Res Hosp April 1, 2005 July 3 3 No J Rubntiz/ Tennessee: Nashville Vanderbilt Univ Med Ctr March 15, 2005 July 3 3 Yes JA Whitlock/ Texas: Dallas UT Southwestern Med Ctr/ Child Med Ctr April 1, 2005 July 3 3 No GR Buchanan/ Texas: Houston Baylor Coll Med/Tex Child Hosp July 4 3 No DG Poplack/ Texas: Houston M.D. Anderson Cancer Ctr Feb 28, 2005 July 3 3 No LL Worth/ Washington: Seattle Univ Wash/Child Hosp and Regional Med Ctr Feb 20, 2005 July 3 3 No ID Bernstein/ Wisconsin: Milwaukee Child Hosp Wis/Med Coll Wis April 15, 2005 July 3 3 Yes JP Scott/ Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia July No D Dix/ INFECTIOUS DISEASE Alabama: Birmingham Univ Ala Jan 31, 2006 July 3 No S Boppana/ Arkansas: Little Rock Univ Ark Med Sci April 2005 July 3 GE Schutze/ California: Los Angeles Cedars-Sinai Med Ctr Jan 15, 2006 July 3 3 Yes D Lehman/ California: Los Angeles Child Hosp Jan 1, 2006 July 3 No W Mason/ California: Los Angeles Univ Calif Med Ctr/Dept Pediatrics March 30, 2006 July 3 No P Krogstad/ California: San Francisco Univ Calif July No P Weintrub/ California: Palo Alto Stanford Univ School Med July 3 No Y Maldonado/ California: Torrance Harbor-Univ Calif Los Angeles/Univ Calif School Med July 3 3 No M Keller/ Colorado: Denver Univ Colo/SOM/Child Hosp May 15, 2005 July 3 3 No MJ Levin/ Connecticut: New Haven Yale Univ School Med April 1, 2006 July 3 3 Yes G Miller/ Florida: Miami Univ/Jackson Mem Hosp July 3 3 No C Mitchell/ Georgia: Atlanta Emory Univ July 3 3 No SR Nesheim/ Illinois: Chicago Child Mem Hosp/ Northwestern Univ/ Feinberg School Med June 3 3 Yes ST Shulman/ *Permanent immigrants should contact the departments regarding eligibility. Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone 32A January 2005 The Journal of Pediatrics

37 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone INFECTIOUS DISEASE Cont d Illinois: Chicago Univ Chicago July 3 No RS Daum/ Louisiana: New Orleans Tulane Univ July 3 3 No J Robinson/ Maryland: Baltimore Univ Md School Med March 31, 2005 July 3 Yes J Nataro/ Massachusetts: Boston Boston Med Ctr/Boston Univ School Med Oct 1, 2005 July 3 3 No SJ Pelton/ Massachusetts: Boston Child Hosp July 3 No RN Husson/ Michigan: Ann Arbor Univ Mich July 3 Yes JR Gilsdorf/ Michigan: Detroit Child Hosp Mich/Wayne State Univ July 31, 2005 July 3 3 No BI Asmar/ Minnesota: Minneapolis Univ Minn Med School July 3 3 No P Ferrieri/ Minnesota: Rochester Mayo Clinic March 15, 2005 July 3 3 No TG Boyce/ Missouri: St. Louis Wash Univ/Child Hosp July No JW St. Geme/ New York: Brooklyn State Univ NY Downstate Med Ctr (HSC) July No M Hammerschlag/ New York: New York Columbia Univ Coll Physicians and Surgeons July 3 No L Saiman/ New York: New York Mount Sinai School Med March 1, 2005 July 3 No R Posada/ New York: New York NY-Presbyterian Hosp-Weill Cornell Med Ctr Dec 31, 2005 July 3 3 No J Stavola/ New York: Rochester Strong Mem Hosp/Golisano Child Hosp Strong May 1, 2006 July Yes MT Caserta/ New York: Stony Brook Suny Stony Brook July 3 No L Pontrelli/ New York: Syracuse SUNY Upstate Med Univ Nov 1, 2005 July 3 3 No LB Weiner/ Ohio: Cincinnati Child Hosp Med Ctr July 3 3 No BL Connelly/ Ohio: Cleveland Clinic Lerner Coll Med Case Western Res 3 3 No J Goldfarb/ Ohio: Cleveland Rainbow Babies Child Hosp July 3 3 No G McComsey/ Pennsylvania: Philadelphia Child Hosp Dec 1, 2005 July 3 3 No PA Offit/ Pennsylvania: Philadelphia St. Christopher s Hosp Child July 3 3 No S Long/ Pennsylvania: Pittsburgh Univ/Child Hosp June 30, 2005 July 3 No M Michaels/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

38 Category State: City Institution Application deadline Service available INFECTIOUS DISEASE Cont d Rhode Island: Providence RI Hosp/Brown Med School Sept 2005 July 3 3 No D Pugatch/ Tennessee: Memphis LeBonheur Child Med Ctr/ Univ Tenn/St. Jude Child Res Hosp May 1, 2006 July 3 3 No J Shenep/ Texas: Dallas UT Southwestern Med Ctr April 2005 July 3 3 No H Jafri/ Texas: Houston Baylor Coll Med July 3 3 No JR Campbell/ Texas: Houston Univ Tex/Health Sci Ctr July 3 3 No TG Cleary/ Utah: Salt Lake Univ Utah Feb 1, 2006 July 3 No AT Pavia/ Washington: Seattle Child Hosp and Regional Med Ctr/Univ Wash May 31, 2005 July 3 3 Yes CE Rubens/ Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia July 2 No S Dobson/ Canada: Halifax (NS) IWK Health Ctr and Dalhousie Univ July No J Langley/ Canada: Toronto (Ont) Hosp Sick Child Apr 15, 2005 July 1-3 No L Ford-Jones MEDICAL TOXICOLOGY Mass: Boston Massachusetts Child Hosp Sept 1, 2005 July 2 3 Yes M Burnsewald/ Michigan: Detroit Wayne State Univ July 2 No S White/ Ohio: Cleveland Rainbow Babies Child Hosp/ Case Western Reserve Univ July 2 No M Reed/ NEONATAL/PERINATAL MEDICINE Alabama: Birmingham Univ Ala Dec 31, 2005 July 1-3 No W Carlo/ Arkansas: Little Rock Univ Ark Med Sci Dec 31, 2005 July 3 No W Hall/ California: Loma Linda Univ Child Hosp June 30, 2005 July 3 3 No A Hopper/ California: Los Angeles Cedars-Sinai Med Ctr Jan 15, 2006 July 3 3 No C Simmons/ California: Los Angeles David Geffen School Med Univ Calif July 3-4 SU Devaskar/ California: Orange Univ Calif Irvine Med Ctr July 3 3 No H Modanlou/ California: San Francisco Univ Calif July No JA Kitterman/ California: Palo Alto Stanford Univ School Med July 3 2 No LP Halamek/ California: Stanford Univ School Med July 3 2 No LP Halamek/ *Permanent immigrants should contact the departments regarding eligibility. Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone 34A January 2005 The Journal of Pediatrics

39 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone NEONATAL/PERINATAL MEDICINE Cont d California: Torrance Harbor-Univ Calif Los Angeles Med Ctr/Univ Calif School Med Nov 1, 2005 July 3 3 Yes U Raj/ Colorado: Aurora Univ Colo Health Sci Ctr July 3 3 No TA Parker/ Connecticut: Farmington Univ Conn Health Ctr Sept, 2005 July 3 No N Hussain/ Connecticut: New Haven Yale Univ School Med July 3 3 Yes I Gross/ Florida: Gainesville Univ Fla Oct 2005 July 3 3 J Neu/ Florida: Miami Univ/Jackson Mem Hosp April 1, 2005 July 3 No E Bancalari/ Florida: Tampa Univ South Fla Coll Med Sept 1, 2005 July 3 No M Saste/ Georgia: Atlanta Emory Univ School Med July 3 3 No L Black/ Illinois: Chicago Child Mem Hosp/ Northwestern Univ Feinberg School Med July 3 3 No R-A deregnier/ Illinois: Chicago Univ Chicago/Univ Hosp July 3 3 No K Lee/ Illinois: Chicago Univ Ill Med Ctr July 1 3 No D Vidyasagar/ Illinois: Park Ridge Advocate Lutheran General Hosp July 3 3 No B Puppala/ Indiana: Indianapolis Ind Univ School Med May 1 July 3 1 No DW Boyle/ Iowa: Iowa Univ Iowa Health Care/Child Hosp July 3 3 No JA Widness/ Kentucky: Louisville Univ Louisville July 3 3 No DH Adamkin/ Louisiana: New Orleans La State Univ/Tulane Univ Dec 31, No D Penn/ Louisiana: Shreveport La State Univ Health Sci Ctr 3 No R Dhanireddy/ Maryland: Baltimore Univ Md Med Systems/ School Med March 15, 2005 July 3 3 No RM Viscardi/ Massachusetts: Boston Child Hosp July 3 No S Kourembanas/ Massachusetts: Boston Floating Hosp for Child Tufts/Tufts Univ School Med July 3 3 No SK Michael/ Massachusetts: Worcester Univ Mass Mem Jan 1, 2006 July 3 3 No F Bednarek/ Michigan: Ann Arbor Univ Mich/C.S. Mott Child Hosp July 3 No R Schumacher/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

40 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone NEONATAL/PERINATAL MEDICINE Cont d Michigan: Detroit Wayne State Univ/Detroit Med Ctr/Child Hosp Mich July 3 No YR Johnson/ Michigan: Lansing Mich State Univ/Sparrow Hosp March 31, 2005 July 3 No I Gewolb/ Minnesota: Minneapolis Univ Minn Med School July 3 3 No CM Rendel/ Missouri: Kansas City Child Mercy Hosp/Univ Mo July 3 No WE Truog/ Missouri: St. Louis Wash Univ School Med/ Child Hosp Jan 1, 2006 July 3 3 No B Hackett/ New Hampshire: Lebanon Dartmouth Hitchcock Med Ctr July 4 No R Darnall/ New Jersey: New Brunswick UMDNJ-Robert Wood Johnson Med School March 15, 2006 July 3 3 No T Hegyi/ New Mexico: Albuquerque Univ NM July 3 3 No RK Ohls/ New York: Albany Albany Med Ctr March 31, 2006 July 3 3 No J Pinheiro/ New York: New York Mount Sinai School Med July 3 3 No IR Holzman/ New York: New York NY Presbyterian Hosp/ Columbia March 31, 2005 July 3 3 No SD Rubenstein/ New York: New York NY Presbyterian Hosp/Weill Cornell Med Ctr Dec 31, 2005 July 3 3 No J Perlman/ New York: Rochester Strong Mem Hosp/Golisano Child Hosp Strong July 3 3 No R Guillet/ New York: Stony Brook Stony Brook Univ Hosp March 1, 2006 July 3 3 No J DeCristofaro/ New York: Valhalla, New York Westchester Med Ctr Oct 3 3 No EF Lagamma/ North Carolina: Durham Duke Univ Med Ctr Feb 1, 2006 July 3 No R Goldberg/ North Carolina: Greenville East Carolina Univ/Brody School Med June 30, 2005 July 3 2 No JJ Cummings/ North Carolina: Winston-Salem NC Baptist Hosp/Wake Forest Univ School Med July 3 3 No TM O Shea/ Ohio: Cincinnati Child Hosp Med Ctr July 3 2 No WR Rice/ Ohio: Cleveland Metrohealth Med Ctr/Case Western Reserve Univ July 3 3 No JJ Moore/ Ohio: Cleveland CWRU/Clinic Foundation/ Metrohealth Med Ctr July 3 3 No JJ Moore/ *Permanent immigrants should contact the departments regarding eligibility. 36A January 2005 The Journal of Pediatrics

41 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone NEONATAL/PERINATAL MEDICINE Cont d Ohio: Cleveland Rainbow Babies Child Hosp 3 No C Bearer/ Ohio: Columbus Child Hosp Ohio State Univ July 3 3 No S Welty/ Oklahoma: Oklahoma Okla Univ Med Ctr July 3 No MB Escobedo/ Oregon: Portland Ore Health Sci Univ Jan 1, 2005 July 3 Yes CT McEvoy/ Pennsylvania: Hershey M.S. Hershey Med Ctr/Penn State Child Hosp March 1, 2005 July 3 3 No JW Kendig/ Pennsylvania: Philadelphia Child Hosp/Univ Pa School Med June 30, 2005 July 3 3 No PA Dennery/ Pennsylvania: Philadelphia St. Christopher s Hosp for Child July 3 3 No AB Zubrow/ Pennsylvania: Philadelphia Thomas Jefferson Univ Hosp July 3 3 No SA Pearlman/ Pennsylvania: Pittsburgh Univ Med Ctr July 3 3 No GA Silverman/ Rhode Island: Providence Brown Med School/Women Infants Hosp April 1, No BS Stonestroet/ Ext Texas: Dallas Univ Tex Southwestern Med Ctr July 3 3 No CR Rosenfeld/ Texas: Houston Baylor Coll Med July No AR Stark/ Texas: Houston Univ Tex/Houston Med School July 3 3 Yes JW Sparks/ Texas: LackLand AFB Wilford Hall Med Ctr Sept 1, 2005 July 3 R DiGeronimo/ Texas: San Antonio Univ Tex Health Sci Ctr July 3 3 Yes R Castro/ Utah: Salt Lake City Univ Utah School Med July 1, Yes RH Lane/ Virginia: Charlottesville Univ Va Jan 1, 2006 July 3 3 Yes J Kattwinkel/ Virginia: Richmond VCU Health Systems/ Virginia Commonwealth Univ Dec 31, 2005 July 3 3 No HJ Rozycki/ Washington: District Columbia Georgetown Univ Hosp July 3 3 No S Subramanian/ Washington: Seattle Univ Wash Med Ctr July 3 Yes S Juul/ Wisconsin: Milwaukee Med Coll WI/Child Hosp Wis July 3 3 No GG Konduri/ Canada: Calgary (Alta) Foothills Hosp 2-3 No D Clark/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

42 Category State: City Institution Application deadline Service available NEONATAL/PERINATAL MEDICINE Cont d Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia No H Osiovich/ Canada: Montreal Ste-Justine Hosp/Univ Montreal July No C Lachance/ Canada: Halifax (NS) IWK Health Ctr/Dalhousie Univ March 31, 2005 July 1-3 No A Howlett/ Canada: Hamilton (Ont) McMaster Univ Sept 1, 2005 July or Jan No K Lee/ Ext NEPHROLOGY California: Los Angeles Cedars-Sinal Med Ctr Jan 15, 2006 July 3 3 No E Kamil/ California: Los Angeles David Geffen School Med Dec 20, 2005 July 3 Yes O Yadin/ California: San Francisco Univ Calif July 3 3 Yes A Portule/ Connecticut: New Haven Yale Univ School Med Jan 31, 2006 July 3 3 N Siegel/ Florida: Gainesville Univ Fla/Coll Med April 30, 2005 July 3 3 No V Dharnidharka/ Florida: Miami Univ/Jackson Mem Hosp Sept 30, 2005 July No G Zilleruelo/ Illinois: Chicago Northwestern Univ/Feinberg School Med Sept 30, 2005 July 3 3 No CB Langman/ Massachusetts: Boston Child Hosp/Harvard Med School March 31, 2005 July 3 No WE Harmon/ Michigan: Detroit Child Hosp Mich July 3 No TK Mattoo/ Minnesota: Minneapolis Univ Minn Med School March 31, 2006 July 3 3 Yes C Kashtan/ Missouri: Kansas City Child Mercy Hosp/Univ Mo July 31, 2005 July 3 3 No US Alon/ New York: Brooklyn State Univ NY Downstate Med Ctr (HSC) July No M Schoeneman/ New York: New York Mount Sinai School Med Dec 1, 2005 July 3 3 Yes LM Sattlin/ New York: New York City NY-Presbyterian Hosp/Weill Cornell Med Ctr Dec 31, 2005 July 3 3 No V Johnson/ New York: Rochester Strong Mem Hosp/Golisano Child Hosp Strong July 3 Yes WS Varade/ Ohio: Cincinnati Univ Cincinnati/Child Hosp Med Ctr July 3 3 No JJ Bissler/ Ohio: Cleveland CWRU Med School/Univ Hosp March 1, 2006 July 3 3 Yes I Davis/ *Permanent immigrants should contact the departments regarding eligibility. Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone 38A January 2005 The Journal of Pediatrics

43 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone NEPHROLOGY Cont d Ohio: Columbus Child Hosp/Ohio State Univ Dec 1, 2005 July 3 3 No JD Mahan/ Louisiana: New Orleans Tulane Univ School Med Sept 1, 2005 July 3 3 No FG Boineau/ Michigan: Ann Arbor Univ Mich July 3 No D Kershaw/ Missouri: St. Louis Wash Univ/Child Hosp July 1, 2005 July 3 3 No SP Hmiel/ Pennsylvania: Philadelphia Child Hosp March 31, 2005 July 3 3 No K Meyers/ Tennessee: Memphis Univ Tenn Health Sci Ctr Dec 31, No RJ Wyatt/ Texas: Dallas Univ Tex/Southwestern Med Ctr Dec 2005 July 3 No M Baum/ Texas: Houston Baylor Coll Med/Tex Child Hosp April 1, 2006 July 3 3 Yes ED Brewer/ Texas: Houston Univ Tex/Health Sci Ctr July 1, 2005 July No RJ Portman/ Virginia: Charlottesville Univ Va/Department Pediatric Nephrology Jan 1, 2006 July 3 No VF Norwood/ Washington: Seattle Univ Wash/Child Hosp Regional Med Ctr July 3 3 Yes A Eddy/ Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia Sept 30, 2005 July No RM Hurley/ Canada: Halifax (NS) Dalhousie Univ July 1, 2005 July 1-3 No P Acott/ NEUROLOGY Alabama: Birmingham Univ Ala July 3 2 No L Dure/ Arkansas: Little Rock Univ Ark for Med Sci July 3 2 No M Griebel/ California: Loma Linda Loma Linda Univ MC 3 Yes S Ashwal/ California: Los Angeles Child Hosp July 3 2 No WG Mitchell/ California: Torrance Harbor Univ Calif Med Ctr/ School Med July 3 2 No KR Huff/ Colorado: Denver Child Hosp/UCHSC July 3 2 No T Benke/ Georgia: Atlanta Emory Univ School Med July 3 No JT Sladky/ Kentucky: Lexington Univ Ky/Department Neurology Dec 10, 2005 July 3 2 No RJ Baumann/ Illinois: Chicago Child Mem Hosp July 3 2 No J Coldstein/ Illinois: Chicago Univ Chicago Nov 1, 2005 July No K Silver/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

44 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone NEUROLOGY Cont d Indiana: Indianapolis Ind Univ School Med Nov 1, 2005 July 3 2 No L Walsh/ Iowa: Iowa Univ Iowa Carver Coll Med/ Univ Hosp & Clinics July 1 No K Mathews/ Michigan: Ann Arbor Univ Mich July 3 No F Silverstein/ Minnesota: Minneapolis Univ Minn Med School Oct 1, 2005 July 3 2 No L Charnas/ Mississippi: Jackson Univ Ms Med Ctr Nov 2005 July 3 No V Vedanarayanan/ Missouri: St Louis Wash Univ/Child Hosp Nov 1, 2005 July 3 2 No E Trevathan/ New York: Brooklyn State Univ NY Downstate Med Ctr (HSC) July No J Cracco/ Ohio: Cincinnati Child Hosp Med Ctr Dec 31, 2005 July 3 2 No MB Schapiro/ Ohio: Cleveland Cleveland Clinic Foundation Nov 30, 2005 July 3 No NR Friedman/ Ohio: Columbus Child Hosp/Ohio State Univ July No W Lo/ Ohio: Columbus Child Hosp/Ohio State Univ July 1 3 No JM Paolicchi/ Pennsylvania: Philadelphia St. Christopher s Hosp for Child April 1, 2006 July 3 3 No SV Kothare/ Pennsylvania: Philadelphia Child Hosp July 1, 2005 July No D Dlugor/ Washington: Seattle Univ Wash/Child Hosp Oct 15, 2005 July No S Gospe/ Wisconsin: Milwaukee Child Hosp Wis/Med Coll Wis Feb 2005 July 3 2 No ML Zupanc/ Canada: Vancouver (BC) Child Women s Health Ctr/ Univ British Columbia July 1-2 No A Hill/ NEUROLOGICAL SURGERY Ohio: Cincinnati Child Hosp Med Ctr July 1 No KR Crone/ NUTRITION Colorado: Denver Child Hosp/Univ Colo Health Sci Ctr July 3 3 Yes NF Krebs/ OTOLARYNGOLOGY Ohio: Cincinnati Child Hosp Med Ctr April 30, No JP Willging/ PATHOLOGY Ohio: Cincinnati Child Hosp Med Ctr July 1 3 No MH Collins/ *Permanent immigrants should contact the departments regarding eligibility. 40A January 2005 The Journal of Pediatrics

45 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone PATHOLOGY Cont d Ohio: Columbus Child Hosp/Ohio State Univ July 1 4 No S Kahwash/ Pennsylvania: Philadelphia St. Christopher s Hosp for Child July No JC Pascasio/ PSYCHIATRY Ohio: Cincinnati Child Hosp Med Ctr Jan 30, 2005 July 2 No JP Daniels/ Ohio: Cincinnati Child Hosp Med Ctr March 31, 2005 July 1-2 No D Rose/ PULMONOLOGY Alabama: Birmingham Univ Ala Dec 31, 2005 July 3 No H Gutierrez/ Arkansas: Little Rock Ark Child Hosp Dec 31, No JL Carroll/ California: Long Beach, California Univ Calif Irvine/Miller Child Hosp Oct 1, 2005 July 3 3 No E Nussbaum/ California: Los Angeles Child Hosp March 1, 2005 July 3 3 No TG Keens/ California: San Francisco Univ Calif Dec 1, 2005 July 3 No DW Nielson/ Colorado: Denver Univ Colo Health Sci Ctr Feb 1, 2005 July 3 3 Yes RR Deterding/ Connecticut: New Haven Yale Univ School Med March 31, 2006 July 3 3 No A Barry-Asaad/ Florida: Gainesville Univ Fla July 30, 2005 July 3 3 No S Chesrown/ Florida: Miami Univ/Jackson Mem Hosp Dec 30, 2005 July 1-3 No H Gelband/ Illinois: Chicago Child Mem Hosp/ Northwestern Univ July 3 3 No O Lakser/ Indiana: Indianapolis Ind Univ May 1, 2005 July 3 3 Yes MS Howenstine/ Iowa: Iowa Univ Iowa Aug 1, 2005 July 3 3 No M Weinberger/ Louisiana: New Orleans Tulane Univ School Med July 3 3 No R Beckerman/ Massachusetts: Boston Child Hosp July 3 3 No D Boyer/ Michigan: Ann Arbor Univ Mich/C.S. Mott Child Hosp March 1, 2005 July 3 No M Hershenson/ Minnesota: Minneapolis Univ Minn Med School July 3 3 No WE Regelmann/ Missouri: St. Louis Wash Univ School Med/ Child Hosp Jan 1, 2006 July 3 3 No T Ferkol/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

46 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone PULMONOLOGY Cont d New York: Brooklyn State Univ NY Downstate Med Ctr (HSC) July No M Rao/ New York: New York Mount Sinai School Med Sept 2005 July 3 No M Kattan/ New York: New York NY Presbyterian Hosp/ Columbia Univ Dec 1, 2005 July 3 3 Yes MR Bye/ New York: Rochester Univ Rochester/Strong Mem Hosp Dec 31, 2005 July 3 3 No CL Ren/ North Carolina: Durham Duke Univ Med Ctr June 30, 2006 July 3 3 No JA Voynow/ Ohio: Cincinnati Child Hosp Med Ctr April 30, 2005 July 3 Yes BA Chini/ Ohio: Cleveland Case Western Reserve Univ March 31, 2006 July 3 3 Yes J Chmiel/ Ohio: Columbus Child Hosp/Ohio State Univ March 31, 2005 July 3 3 No KS Mccoy/ Oregon: Portland Ore Health & Sci Univ Jan 1, 2006 July 3 3 No MR Powers/ Pennsylvania: Philadelphia Child Hosp March 31, 2006 July 3 3 No H Panitch/ Pennsylvania: Philadelphia St. Christopher s Hosp for Child Aug 1, 2005 July 3 3 No J Steinfeld/ Pennsylvania: Pittsburgh Child Hosp/Univ Pittsburg Jan 1, 2005 July 3 No G Kurland/ Texas: Houston Baylor Coll Med Jan 1, 2005 July 3 3 No C Oermann/ Texas: Houston Univ Tex Health Sci Ctr Oct 1, 2005 July 3 3 Yes GN Colasurdo/ Washington: Seattle Univ Wash School Med Nov 2005 July 3 3 Yes GJ Redding/ Wisconsin: Madison Univ Wis Child Hosp July 3 Yes ME Kawiec/ Wisconsin: Milwaukee Med Coll Wis Oct 1, 2005 July 3 3 No WM Gershan/ Canada: Calgary (Alta) Univ Calgary/Alberta Child/Hosp May 1, No C Bjornson/ RADIOLOGY California: Los Angeles Child Hosp Jan 1, 2006 July 1 Nelson/ Colorado: Denver Child Hosp Dec 31, 2005 July 1 4 No JD Strain/ Florida: Miami Child Hosp Dec 30, 2005 July 4 No R Restrepo/ Indiana: Indianapolis Ind Univ Jan 1, 2006 July 1 or 2 No R Gendirman/ Ohio: Cincinnati Child Hosp Med Ctr March 29, 2005 July 1 No LF Donnelly/ Ohio: Columbus Child Hosp/Ohio State Univ July 1 4 No WE Shiels II/ *Permanent immigrants should contact the departments regarding eligibility. 42A January 2005 The Journal of Pediatrics

47 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone RADIOLOGY Cont d Pennsylvania: Philadelphia St. Christopher s Hosp for Child July 1 5 No E Smergel/ RHEUMATOLOGY California: Los Angeles Child Hosp July 30, 2005 July 3 No B Shaham/ California: Los Angeles Univ Calif Dept Pediatrics Feb 1, 2006 July 3 Yes T Chatila/ California: San Francisco Univ Calif July No E Scheven/ Illinois: Chicago Child Mem Hosp/ Northwestern Univ July 3 Yes Marisa/ Illinois: Chicago Univ Chicago/La Rabida July 3 No K Onel/ Massachusetts: Boston Child Hosp/Harvard Med School July 3 H Oettgen/ Michigan: Ann Arbor Univ Mich July 3 No B Adams/ Missouri: St. Louis Wash Univ School Med/ Child Hosp July 3 No AJ White/ New York: New York Columbia Univ Child Hosp Nov 2005 July 3 3 No L Imundo/ New York: New York Hosp Special Surgery/NYH Jan 1, 2006 July 3 3 Yes T Lehman/ North Carolina: Durham Duke Univ Med Ctr/Univ NC Chapel Hill Sept 2005 July 3 3 No L Schanberg/ Ohio: Cincinnati Child Hosp Med Ctr April 2006 July 3 Yes MH Passo/ Pennsylvania: Philadelphia Child Hosp July 3 2 Yes DD Sherry/ Pennsylvania: Pittsburgh Child Hosp July 3 No R Hirsch/ Tennessee: Memphis Univ Tenn Dec 30, 2005 July 3 3 L Myers/ Texas: Dallas Univ Tenn Southwestern Med Ctr July 3 3 No M Punaro/ Texas: Houston Baylor Coll Med/Tex Child Hosp July 3 No MD Perez/ Washington: Seattle Univ Wash School Med Feb 28, 2005 July 3 Yes H Emery/ Wisconsin: Milwaukee Child Hosp Wis/Med Coll Wis July 3 3 Yes CB Williams/ Canada: Vancouver (BC) Child Women s Health Ctr, Univ British Columbia July 1 3 No RE Petty/ *Permanent immigrants should contact the departments regarding eligibility. The Journal of Pediatrics January A

48 Category State: City Institution Application deadline Service available Duration of appointment (y) Postgraduate training required (y) US citizenship required* Director of training program/phone SPORTS-MEDICINE Ohio: Cincinnati Child Hosp Med Ctr Nov 1, 2005 Aug 1 Yes J Divine/ Texas: Houston Baylor Coll Med Sept 15, 2005 July 1 3 Yes A Hergenroeder/ Texas: San Antonio Univ Tex Health Sci Ctr Oct 31, 2005 July 1 3 No JE Gomez/ SURGERY Ohio: Cincinnati Child Hosp Med Ctr July Yes BW Warner/ UROLOGY Florida: Miami Child Hosp Dec 30, 2005 July 2 No R Gosalbez/ Ohio: Cincinnati Child Hosp Med Ctr July 2 No PP Reddy/ Ohio: Columbus Child Hosp/Ohio State Univ July 1 5 No SA Koff/ *Permanent immigrants should contact the departments regarding eligibility. 44A January 2005 The Journal of Pediatrics

49 NOTES FROM THE ASSOCIATION OF MEDICAL SCHOOL PEDIATRIC DEPARTMENT CHAIRS, INC. THE MORBIDITY AND MORTALITY CONFERENCE IN UNIVERSITY-AFFILIATED PEDIATRIC DEPARTMENTS IN CANADA JEREMY N. FRIEDMAN, MBCHB, FRCPC, MARIE S. PINARD, RN, BSCN, AND RONALD M. LAXER, FRCP Little is known about the examination of morbidity and mortality (M & M) in the hospital setting. The published literature related to M & M reviews pertains almost exclusively to clinical review of deaths, particularly in surgical and anesthesia services. Our hope is that our study on how M & M is done in the pediatric teaching hospital setting in Canada will provoke discussion that may lead to a study into the effectiveness of the M & M review. The objective of this study was to describe the M & M review process in Canadian major teaching hospitals that are fully affiliated with university departments of pediatrics. A secondary objective of this study was to examine the goals and perceived effectiveness of M & M rounds and the barriers to achieving these goals. METHODS An extensive literature search was conducted via Medline for English and French titles on this subject published from 1989 onward. Relevant references cited within articles resulting from the literature search were also reviewed. Based on the findings of the literature review, a standardized questionnaire consisting of 17 items was developed outlining the key areas of interest to the investigators. The survey tool was structured to solicit descriptive information in the following domains: (1) mortality reviews, (2) morbidity reviews, (3) M & M process, and (4) perceived purpose, effectiveness, and barriers. The questionnaire was mailed to all 16 of the Chairs of Canadian university-affiliated pediatric departments. 1 In addition, a slightly modified questionnaire was sent to the Division Chiefs of the Neonatal Intensive Care Units (NICUs) and general pediatric divisions at the 16 institutions. RESULTS The response rate to the questionnaire was 28 of 39 (72%). At some institutions the same individual acted as both Department Chair and Division Chief in the NICU or General Pediatrics, and only 1 survey was sent to those individuals. There was broad national representation among the respondents. The pediatric departments ranged in size from 26 to 438 pediatric hospital beds. M & M NICU Morbidity and mortality Neonatal Intensive Care Unit In the domain of mortality reviews: 89% of respondents reviewed all deaths occurring in the hospital (including the emergency department) at M & M rounds; 37% responded that all deaths were classified as preventable/nonpreventable, whereas 33% believed that this was done for selected deaths only (Table I; available online at In the domain of morbidity reviews, 63% of respondents presented morbidity at M & M rounds and 11% responded that their organization had criteria for defining morbidity events (Table II; available online at As regards the M & M process, 78% of respondents believed that there was an individual or group responsible for overseeing M & M rounds, but only 11% indicated that there was any specific training or guidance for those individuals leading M & M; 52% of respondents answered that issues and recommendations were shared with referring physicians, with a further 30% indicating that this occurred sometimes (Table III; available online at The primary purpose of M & M was most commonly believed to be teaching/educational and improvement of patient care. The greatest barriers to an effective M & M were thought to be a lack of time and resources (Table IV; available online at DISCUSSION The results of our survey suggest that most universityaffiliated pediatric departments in Canada, as well as their divisions of NICU and General Pediatrics, are collecting data on and reviewing all (or selected) mortalities. It appears that in the vast majority of cases, the criteria for classification of morbidity events is not clear, although a majority of institutions do present some morbidity at M & M. Divisions of Pediatric Medicine (Pediatric Outcomes Research Team) and Rheumatology; the Department of Pediatrics and Medicine, the Hospital for Sick Children and the University of Toronto; and the Department of Quality and Risk Management, the Hospital for Sick Children, Toronto, Ontario, Canada. Reprint requests: Jeremy N. Friedman, MBChB, FRCPC, The Hospital for Sick Children, 555 University Ave, Toronto, Ontario M5G 1X8, Canada. J Pediatr 2005;146: /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds A M S P D C 1

50 A M S P D C There was marked variability in the process and standards for M & M reviews. It is alarming that only 11% believed that there was any specific training or guidance for those individuals entrusted with the responsibility of leading M & M rounds. It is also somewhat surprising that the issues/recommendations from M & M were sometimes not shared with referring physicians. It appears that the purpose of M & M reviews may have evolved from the historic forum for peer review to a way of improving patient care by means of teaching/educational sessions. The two biggest barriers to an effective M & M are a perceived lack of time and resources. This, together with the observation that there is almost no formalized training or guidance for those in charge of M & M and no clear definition of the standards required, begs the question whether the M & M process has assumed a high enough priority in most of these institutions/departments. The two biggest barriers to an effective M & M are a perceived lack of time and resources. The M & M review has been one of the most common traditional methods used for evaluating care in many teaching medical centers. Rutstein et al 2 suggest that unexpected deaths are important sentinel events and that analysis of their circumstances is a more simple and efficient method of identifying health determinants than that of looking at entire populations. Quality assurance activities should be planned, systematic, and ongoing. In many cases, the M & M process does not conform to this ideal. 3 One example is that only minimal attempts are made to ensure the complete identification and reporting of complications. Voluntary submission of complications is not comprehensive, and physicians are often not present when their complications are being discussed. 4,5 Physicians believe that the give-and-take of the discussions at M & M constitutes effective peer review and that they are an adequate means to change practice when needed. Often there is no documentation of conclusions and actions, so the ability to track effectiveness of the actions or to detect trends is lost. 3 We noted a lack of standards both in the requirement for documentation and the presence of review by a departmental committee with broad-based representation. There was also inconsistent sharing of issues and recommendations with the referring physicians. These factors tend to decrease the educational potential of learning from the experience of others as well as detecting trends in M & M or quality assurance issues. A review of the organization and approach of centralized Death Analysis Committees was published in France in The authors recommended the development of a national guide to the methodology and structure of death review committees as well as a structured research method for identifying causes of death. 5 We would echo this need to develop national standards and stricter definitions for the M & M process to ensure that hospitals are aware of the expectations and to make this a priority to maintain their accreditation. Strengths and Limitations This is the first study examining the M & M process in academic pediatric departments. The response to the questionnaire was very good (72%), with broad national representation. The consistency of the data among the three groups (Pediatric Departments, NICUs, and General Pediatrics) strengthens the reliability of the data. Respondents in general seemed eager for guidance and information to help improve the M & M process. The 16 university-affiliated pediatric programs are a diverse group, differing markedly in size and structure. Some are based in stand-alone pediatric hospitals, whereas others are accommodated in general hospitals or are involved in a more regionalized type of care, which affects the process of the M & M review. Terminology used and the responsibilities of different committees and individuals will again vary in different parts of the country and with different-sized departments. Some morbidity issues are likely to be dealt with by quality improvement or quality assurance programs that may be supported by the hospital administration as opposed to the academic faculty, making it very difficult to compare the M & M process between institutions. Issues such as measurement of educational outcomes and of the effectiveness of quality improvement processes were not dealt with but should form the basis of future work in this area. CONCLUSIONS The process and standards for M & M in academic pediatric centers in Canada are inconsistent. We believe that devoting increased priority and resources, together with the standardization of the M & M process and the provision of better training and guidance of the physicians responsible, could alleviate many of these deficiencies and problems. A national interdisciplinary committee to develop standards and stricter definitions for M & M probably would facilitate improving the process at all levels. At the hospital level, we would recommend appointing designated individuals, with sufficient time and resources at their disposal, to be responsible for ensuring appropriate M & M identification and review. These individuals should be given instruction with a clear set of guidelines and standards on which to base their reviews, which should be monitored by a hospital-wide committee with broad-based representation. It is likely that a study into the effectiveness of the M & M review needs to be done before this additional resource allocation will be forthcoming. Reference available online at jpeds We acknowledge the contribution of Dr Hugh O Brodovich, Professor and Chairman, Department of Pediatrics, University of Toronto, and Dr Colin MacArthur, Director of Pediatric Outcomes Research Team (PORT), Division of Pediatric Medicine, University of Toronto, for reviewing the manuscript. 2 Friedman, Pinard, and Laxer The Journal of Pediatrics January 2005

51 EDITORIALS INCREASING UTILIZATION OF GENERAL PEDIATRICIANS AND PEDIATRIC SUBSPECIALISTS: CAN THE WORKFORCE MEET THE NEED? The status of the workforce in pediatrics, particularly the subspecialty workforce, has been hotly debated in recent years Some studies suggested an impending oversupply 2,11 ; yet many fields have experienced or are predicted to experience acute shortages. 12,13 The causes of shortages have also been argued. Are fewer pediatricians entering subspecialty training? Have they been discouraged by managed care? 2 Are they inhibited by their growing debt burden? 2,14 Are they disillusioned with academic practice? 2 Are general pediatricians simply referring more patients to subspecialists? Are patients seeking more of their care from pediatric providers? Is the pediatric population requiring more medical care? 10,15,16 Is the shortage due to the increase in women in pediatrics who are not entering subspecialty training in proportion to their male colleages? 17,18 Are training programs accepting too many international graduates who may be unable to remain in this country after training? 3,18,19 Are younger generations of physicians working fewer hours in full-time positions or are more assuming part-time positions? 15,20-22 Or is an aging physician population cutting back? Are subspecialists just burning out? 23 Is the shortage relative to the increasing diversity of the population? 24,25 The studies by Freed et al 26 and Werner et al 27 in this issue of The Journal begin to shed light on some of these questions. Freed et al 26 demonstrated growth in the office visit practices for both general pediatricians and pediatric subspecialists with a decline in pediatric visits to other medical generalists and medical subspecialists from 1993 to For common diagnoses such as constipation and headache, there was growth in number of visits per member per month, and the growth was shared by general pediatricians and pediatric subspecialists. Another common diagnosis, asthma, was found to have fewer visits per member per month over time, with a greater proportion of the visits being seen by pediatric and other medical generalists compared with pediatric or other medical subspecialists. Werner et al 27 report that child neurologists are more likely than other specialists to report that the complexity of severity of patients being referred to them is less than it should be and that the number of patients being referred to them has increased. Projections of the future workforce reveal that by 2022, pediatric subspecialists and pediatricians will increase substantially, to almost 180% and 150% of the current workforce, respectively. Among child neurologists and adult neurologists, practicing physicians will be only 109% and 105% of the current workforce in They highlight the need to recruit more providers into child neurology and to address the high prevalence of inappropriately low acuity referrals. Ironically, as the existing pediatric neurologists become busier and busier, they will have less time to teach residents. The result may be that future generations of pediatricians will be increasingly less comfortable treating pediatric neurology patients themselves. Hence, this cycle of low acuity referrals will be difficult to interrupt. These studies raise several interesting questions. First of all, Freed s study centered on five commercial HMO and point-of-service plans in a large multistate health care organization. It would be important to know what incentives might have governed referrals to subspecialists in this particular organization. Such incentives could potentially influence referral patterns and the results seen. It would also be important to know how much of the referral from pediatric generalists to pediatric subspecialists was driven by overloaded schedules for the generalists. Also, one might ask whether referrals to subspecialists would have been greater in certain areas if appointments were readily available. In other words, did lack of timely appointment options reduce the potential for growth in subspecialty referrals? Multiple factors may contribute to physicians referral practice, including a patient s clinical status and the physician s beliefs of inadequacy of training and inability to stay up to date. Referral patterns for pediatricians and family physicians may differ substantially. 28 Assessment of physician specialty is a challenging undertaking. As shown by Brotherton, 29 pediatric generalists and subspecialists often adopt integrative practices; thus, it is necessary to look beyond training and certification to characterize practices accurately. The results of both the Freed and Werner studies hinge on an accurate assessment of physician specialty. In both studies, physician specialty was determined from the American Medical Association (AMA) Physician Masterfile. Some authors have questioned the reliability of this methodology, which uses self-reported primary and secondary specialty. Grumbach 30 found that excluding physicians with secondary practices in specialist fields from the generalist physician See related articles, p 14, and p 20. Reprint requests: Debra H. Fiser, MD, 800 Marshall, Little Rock, AR J Pediatr 2005;146: /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds Editorials 3

52 definition resulted in a 25% underestimate of generalist physicians. An Ohio study of family physicians discovered that there were undercounts and overcounts in certain groups using this methodology. Two groups of family physicians contributed to counting biases: family physicians who fulfill shortterm goals by part-time practice in several locations and family physicians who restrict their practice to a limited medical content area. However, because of nearly equal offsetting of overcounting (incorrect inclusions) and under counting (incorrect omissions), the 1990 Ohio family physician AMA Masterfile data were found to be adequate for workforce projections and policy studies when the county data were aggregated at the state level. 31 In one study of internists, there was a strong association between the internists primary and secondary specialties reported in the Masterfile and measures of the primary care content of physicians practices (P <.0001). 32 Other studies have also confirmed the reliability of the AMA Masterfile data. 33,34 The estimate of the number of actively practicing child neurologists in the United States in Werner s study is of some concern. They identified 604 from surveys but estimated another 300 from nonresponders, for a total of 904. This estimated figure, which represents approximately a third of the total, could introduce substantial error into the calculations and underestimate the severity of the shortage. Of note, the AMA Physician Masterfile identified 692. The Werner study also used the proportion of residency slots filled as an indicator of the magnitude of the shortage; however, this parameter may not be an adequate reflection of the shortage, since the number of slots is increasing. One of the greatest challenges in estimating workforce requirements is the need to project demand. Most studies assume constant demand. However, as Stockman notes, the ratio of 6000 children per pediatrician was thought to be appropriate in the early 1970s, and 4000:1 was thought to be appropriate in By the late 1990s, the ratio was 2000:1, and this ratio was also thought to be appropriate. 35 The AAP Committee on Workforces report indicated that the workforce was balanced but that was based on how care was provided then. 3 Clearly, referral patterns and utilization can and do change, as these studies have nicely demonstrated. Such changes must be accounted for in workforce projections. The FOPE II Pediatric Workforce Group suggested that a ratio of 1200 to 1400:1 might be ideal for the future. 15 It was somewhat surprising to see that the total number of visits per member per month for asthma declined over time in Freed s study. Available data suggest that asthma prevalence increased in the United States in the 1990s until There was a small decline from 1997 to 1999, followed by an increase again through 2001, which was the last year of the Freed study. 39 It was also surprising that pediatric subspecialists were not seeing more of these patients. Perhaps they were prevented from doing so because of a shortage of pediatric pulmonologists or allergists. Perhaps the National Heart, Lung, and Blood Institute guidelines have made caring for this chronic disorder easier, so that primary care physicians are able to handle the care efficiently and comfortably. If patients are well educated on the guidelines and have home care plans, they may have learned to avoid many physician office visits. It is less obvious why there were more visits for constipation and headache. Perhaps patient expectations for subspecialty care have increased over time with more self-referral. This study did not address self-referral versus physician referral. Similarly, in the Werner study, it is not known whether the increase in low acuity neurology referrals represents physician referral or self-referral. This question deserves further study. Freed s results call into question others conclusions that the demand for general pediatricians is decreasing or that there will be an oversupply in the future. 7,14 These studies instead suggest that utilization of general pediatricians and pediatric subspecialty providers is on the rise. Because utilization patterns are a moving target, and because the lifestyle needs and work patterns of providers change over time, prediction of workforce needs in the future will always be subject to error. Reports of predictions themselves may change the outcome. However, barring new barriers in access to pediatric providers, this trend of increased utilization may logically be expected to continue. The implications are that training programs will need to keep up with the increasing demand for training generalists and also prepare these general pediatricians to deal with lower acuity subspecialty patients, especially in fields such as child neurology, where shortages are critical and likely to be longstanding. As others have noted, informing the market of subspecialty shortage areas seems to promote some correction. 35 Whether this will be sufficient to attract trainees into all of the shortage areas or whether more active interventions will be required, remains to be seen. Debra H. Fiser, MD Professor and Chair Department of Pediatrics University of Arkansas for Medical Sciences College of Medicine Arkansas Children s Hospital Little Rock, AR REFERENCES 1. Freed GL, Nahra TA, Wheeler JR. Which physicians are providing health care to America s children? Trends and changes during the past 20 years. Arch Pediatr Adolesc Med 2004;158: Gruskin A, Williams RG, McCabe ER, Stein F, Strickler J, Chesney RW, et al. Final report of the FOPE II Pediatric Subspecialists of the Future Workgroup. Pediatrics 2000;106: Pediatric Workforce statement. American Academy of Pediatrics. Committee on Pediatric Workforce. Pediatrics 1998;102: Chevalier RL. Estimating workforce and training requirements for nephrologists through the year 2010: pediatric perspectives. J Am Soc Nephrol 1997;8: Pollack LD, Ratner IM, Lund GC. United States neonatology practice survey: personnel, practice, hospital, and neonatal intensive care unit characteristics. Pediatrics 1998;101: Redding GJ, Cloutier MM, Dorkin HL, Brotherton SE, Mulvey HJ. Practice of pediatric pulmonology: results of the Future of Pediatric Education Project (FOPE). Pediatr Pulmonol 2000;30: Shipman SA, Lurie JD, Goodman DC. The general pediatrician: projecting future workforce supply and requirements. Pediatrics 2004;113: Editorials The Journal of Pediatrics January 2005

53 8. Felice ME, Dunston FJ, Jones MD, Chesney R, Stapleton FB, Fiser DH. Pediatric workforce projections and unintended consequences. J Pediatr 2004;145: Stapleton FB, Andreoli S, Ettenger R, Kamil E, Sedman A, Chesney R. Future workforce needs for pediatric nephrology: an analysis of the nephrology workforce and training requirements by the Workforce Committee of the American Society of Pediatric Nephrology. J Am Soc Nephrol 1997;8(Suppl 5):S Stoddard JJ, Cull WL, Jewett EA, Brotherton SE, Mulvey HJ, Alden ER. Providing pediatric subspecialty care: a workforce analysis. AAP Committee on Pediatric Workforce Subcommittee on Subspecialty Workforce. Pediatrics 2000;106: Colletti RB, Winter HS, Sokol RJ, Suchy FJ, Klish WJ, Durie PR. A position paper of the North American Society for Pediatric Gastroenterology and Nutrition. Pediatric Gastroenterology Workforce Survey and future supply and demand. J Pediatr Gastroenterol Nutr 1998;26: Cooper RA, Getzen TE, McKee HJ, Laud P. Economic and demographic trends signal an impending physician shortage. Health Aff (Millwood) 2002;21: Kelly DP, Cull WL, Jewett EA, Brotherton SE, Roizen NJ, Berkowitz CD, et al. Developmental and behavioral pediatric practice patterns and implications for the workforce: results of the Future of Pediatric Education II Survey of Sections Project. J Dev Behav Pediatr 2003;24: Cull WL, Yudkowsky BK, Shipman SA, Pan RJ. Pediatric training and job market trends: results from the American Academy of Pediatrics third-year resident survey, Pediatrics 2003;112: DeAngelis C, Feigin R, DeWitt T, First LR, Jewett EA, Kelch R, et al. Final report of the FOPE II Pediatric Workforce Workgroup. Pediatrics 2000;106: Kletke PR. The changing supply of renal physicians. Am J Kidney Dis 1997;29: Tunnessen WW Jr, Guerin RO, Stockman JA III. Pediatric workforce: data from the American Board of Pediatrics. J Pediatr 2001;139: Larson WL, Holloway RG, Keran CM. Employment-seeking experiences of residents in 1996: a window into the neurology marketplace. Neurology 2000;54: Colletti RB. Pediatric gastroenterology workforce update. J Pediatr Gastroenterol Nutr 1999;28: Cull WL, Mulvey HJ, O Connor KG, Sowell DR, Berkowitz CD, Britton CV. Pediatricians working part-time: past, present, and future. Pediatrics 2002;109: Worley LL, Cooper GJ, Fiser DH. Generational evolution and the future of pediatrics. J Pediatr 2004;145: Cooper RA. Perspectives on the physician workforce to the year JAMA 1995;274: Shugerman R, Linzer M, Nelson K, Douglas J, Williams R, Konrad R. Pediatric generalists and subspecialists: determinants of career satisfaction. Pediatrics 2001;108:E American Academy of Pediatrics. Committee on Pediatric Workforce. Enhancing the racial and ethnic diversity of the pediatric workforce. Pediatrics 2000;105: Stoddard JJ, Back MR, Brotherton SE. The respective racial and ethnic diversity of US pediatricians and American children. Pediatrics 2000;105: Freed GL, Nahra TA, Venus PJ, Schech SD, Wheeler JRC, Pediatrics The Research Advisory Committee of the American Board of Pediatrics. Changes in the proportion and volume of care provided to children by generalists and subspecialists. J Pediatr 2004;146: Werner RM, Polsky D. Comparing the supply of pediatric subspecialists and child neurologists. J Pediatr 2004;146: Freed GL, Jee S, Stein L, Spera L, Clark SJ. Comparing the self-reported referral and management preferences of pediatricians and family physicians for children with juvenile rheumatoid arthritis. J Rheumatol 2003; 30: Brotherton SE. Pediatric subspecialty training, certification, and practice: who s doing what. Pediatrics 1994;94: Grumbach K, Becker SH, Osborn EH, Bindman AB. The challenge of defining and counting generalist physicians: an analysis of Physician Masterfile data. Am J Public Health 1995;85: Williams PT, Whitcomb M, Kessler J. Quality of the family physician component of AMA Masterfile. J Am Board Fam Pract 1996;9: Shea JA, Kletke PR, Wozniak GD, Polsky D, Escarce JJ. Self-reported physician specialties and the primary care content of medical practice: a study of the AMA physician Masterfile: American Medical Association. Med Care 1999;37: Baldwin LM, Adamache W, Klabunde CN, Kenward K, Dahlman C, L Warren J. Linking physician characteristics and medicare claims data: issues in data availability, quality, and measurement. Med Care 2002; 40(Suppl 8): IV Cherkin D, Lawrence D. An evaluation of the American Medical Association s Physician Masterfile as a data source: one state s experience. Med Care 1977;15: Stockman JA. Workforce issues in general pediatrics and the pediatric subspecialties: Pediatric Diplomates. A Newsletter from the American Board of Pediatrics Fall 2004: Shafazand S, Colice G. Asthma: the epidemic has ended, or has it? Chest 2004;125: Woodruff TJ, Axelrad DA, Kyle AD, Nweke O, Miller GG, Hurley BJ. Trends in environmentally related childhood illnesses. Pediatrics 2004; 113(Suppl 4): Masoli M, Fabian D, Holt S, Beasley R. The global burden of asthma: executive summary of the GINA Dissemination Committee report. Allergy 2004;59: National Health Interview Survey Atlanta: Centers for Disease Control and Prevention. HYPERGLYCEMIA AND MORTALITY Diabetes mellitus is a common diagnosis in hospitalized patients presently 9.5% of all hospital discharges. It causes a 2-fold to 4-fold increase in rates of hospitalizations and prolongs hospital length of stay by 1 to 3 days. 1 Over the past several years, a series of studies has demonstrated that hyperglycemia increases in-hospital morbidity and mortality, particularly in critically ill patients. These studies PICU Pediatric Intensive Care Unit have appeared in the adult literature. Stress-induced hyperglycemia, as defined by blood glucose >110 mg/dl, has a reported prevalence of 3% to 71%. 2 It is primarily caused by insulin resistance in liver and skeletal muscle. 2 In critically ill patients, there is lack of suppression of See related article, p 30. Reprint requests: Roberto Izquierdo, MD, Medicine and Pediatrics, SUNY Upstate Medical University, 90 Presidential Plaza, Syracuse, NY J Pediatr 2005;146: /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds Editorials 5

54 gluconeogenesis in the liver despite elevated blood glucose levels. In skeletal muscle, glucose uptake is significantly reduced. Capes et al 3 reported that hyperglycemia (blood glucose >110 mg/dl) increased in-hospital mortality and congestive heart failure in patients admitted with acute myocardial infarction. Stroke victims have higher mortality rates and greater functional disability with blood glucose levels >110 mg/ dl. 4 Furthermore, Umpierrez et al 5 noted that patients in medical and surgical units with newly discovered hyperglycemia (fasting glucose >126 mg/dl and/or random glucose >200 mg/dl on two occasions) had a higher in-patient mortality rate (16%) compared with that of patients with a prior history of diabetes (3%) or subjects with normoglycemia (1.7%). In addition, newly hyperglycemic patients had a longer length of hospital stay, a higher admission rate to an intensive care unit, and were less likely to be discharged to home, frequently requiring transfer to a transitional care unit or a nursing home. 5 These studies and others have conclusively demonstrated that there is an association between hyperglycemia and increased mortality and morbidity. Several studies have demonstrated that correction of hyperglycemia to achieve blood glucose levels as close to normal as possible diminishes the increased in-hospital mortality and morbidity. Intensive glycemic control during labor and delivery in pregnancies complicated by diabetes significantly improves fetal outcomes. 6 The Diabetes and Insulin-Glucose Infusion in Acute Myocardial Infarction (DIGAMI) study revealed that initial intensive glycemic control with an insulin-glucose infusion for at least 24 hours after an acute myocardial infarction, which was then followed by long-term treatment with subcutaneously administered insulin, improved long-term survival by 25%. 7 More recently, Van den Berghe et al 8 reported the results of a randomized, prospective intervention trial in which hyperglycemic patients in a surgical unit were treated with intravenous insulin to keep blood glucose levels <110 mg/dl versus conventional treatment, which consisted of keeping glucose levels between 180 and 200 mg/dl. The intensive insulin treatment decreased in-hospital mortality rates by 34%, blood stream infections by 46%, acute renal failure requiring dialysis or hemofiltration by 41%, the number of red cell transfusion by 50%, and critical illness associated polyneuropathy by 44%. Patients receiving intensive therapy were less likely to require prolonged mechanical ventilation and intensive care. The beneficial effects of tight glycemic control in these studies may be secondary to several physiologic processes. In an animal model of prolonged critical illness, the maintenance of normoglycemia with insulin improved immune function by augmenting phagocytosis by white blood cells. 2,9 This improved bactericidal activity may contribute to host defense and reduce the risk of severe infections. A strong inflammatory response, which is present in critical illness from trauma or sepsis, is thought to contribute to the pathogenesis of multiple organ failure. 2 Insulin and tight glycemic control appear to inhibit proinflammatory cytokines, adhesion molecules, and acute phase reactants, 1 which may reduce this inflammatory response in critical illness. Insulin also inhibits lipolysis and reduces free fatty acids, which have been associated with poor clinical outcomes. 1 The arterial vasodilatory capacity of insulin through stimulation of endothelial nitric oxide synthase may also be beneficial during cardiovascular events. 1 In this issue of The Journal of Pediatrics, Faustino and Apkon 10 reported that hyperglycemia occurs frequently among critically ill nondiabetic children and is correlated with a greater in-hospital mortality rate and longer length of stay in the Pediatric Intensive Care Unit (PICU). This is the first such report in the pediatric literature. This retrospective study involved 942 nondiabetic patients admitted to a PICU over a 3-year period. Point-of-service glucose measurements were analyzed at admission to the PICU, highest value within 24 hours of the initial value, and highest value measured during PICU stay up to 10 days after the first measurement. Three cutoff glucose values were used: 120 mg/dl, 150 mg/dl, and 200 mg/dl. The prevalence of hyperglycemia and relative risks with 95% CI were computed to compare the mortality rates above and below these three cut-off values. The prevalence of hyperglycemia within 24 hours of admission to the PICU was 70.4% in patients with a glucose value >120 mg/dl, 44.5% in patients with a glucose value >150 mg/dl, and 22.3% with a value >200 mg/dl. The mean blood glucose within 24 hours of PICU admission was 166 ± 75 mg/dl, with a range of 66 to 495 mg/dl. There was a 2.5-fold increased risk of dying if the maximum glucose obtained within 24 hours of admission to the PICU was >150 mg/dl and a 5.68-fold increased mortality risk if the maximum glucose obtained within 10 days of admission to the PICU was >120 mg/dl. Length of stay was increased if the maximum glucose obtained within 10 days of admission to the PICU was >120 mg/dl. These results are in concordance with those cited in the adult literature and previous studies. Although the study has several limitations, which are acknowledged and discussed by the authors, the two main conclusions of the study are well supported. Namely, hyperglycemia is common in PICU patients and is associated with higher in-hospital mortality rates. The next step will be to conduct intervention studies in children with critical illness to see if maintaining normoglycemia (blood glucose levels <110 mg/dl) with intensive insulin treatment has the same beneficial effects that we have seen in adults, that is, decreased mortality and morbidity. Roberto Izquierdo, MD Department of Medicine and Pediatrics SUNY Upstate Medical University Syracuse, New York REFERENCES 1. American College of Endocrinology Position Statement on Inpatient Diabetes and Metabolic Control. Endocrine Practice 2004;10(suppl 2);4: Van den Berghe GH. Role of intravenous insulin therapy in critically ill patients. Endocrine Practice 2004;10(suppl 2): Capes SE, Hunt D, Malmberg K, Gerstein HC. Stress hyperglycaemia and increased risk of death after myocardial infarction in patients with and without diabetes: a systematic overview. Lancet 2000;355: Capes SE, Hund D, Malmberg K, Pathak P, Gerstein HC. Stress hyperglycemia and prognosis of stroke in non-diabetic and diabetic patients: a systematic overview. Stroke 2001;32: Editorials The Journal of Pediatrics January 2005

55 5. Umpierrez GE, Isaacs SD, Bazargan N, You X, Thaler LM, Kitabchi AE. Hyperglycemia: an independent marker of in-hospital mortality in patients with undiagnosed diabetes. J Clin Endocrinol Metab 2002;87: Jovanovic L, Peterson CM. Insulin and glucose requirements during the first stage of labor in insulin-dependent diabetic women. Am J Med 1983;75: Malmberg K. Role of insulin-glucose infusion in outcomes after acute myocardial infarction: the diabetes and insulin-glucose infusion in acute myocardial infarction (DIGAMI) study. Endocrine Practice 2004;10(suppl 2): Van den Berghe G, Wouters P, Weekers F, Varwaest C, Bruyninckx F, Schetz M, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2001;345: Weekers F, Giuletti AP, Michalack M, Coopmans W, Van Herck E, Mathiueu C, et al. Metabolic, endocrine, and immune effects of stress hyperglycemia in a rabbit model of prolonged critical illness. Endocrinology 2003;144: Faustino EV, Apkon M. Persistent hyperglycemia in critically ill children. J Pediatr 2004;146:30-4. A COLLABORATIVE EFFORT TO DEFINE THE EPIDEMIOLOGY OF PEDIATRIC INFLAMMATORY BOWEL DISEASE: WHAT CAN WE LEARN FROM CHILDREN WITH EARLY-ONSET DISEASE? The incidence of inflammatory bowel disease (IBD) in general and Crohn disease (CD) specifically has increased over the past two decades in both children and adults. 1-3 These are pediatric disorders. As many as 25% of the patients diagnosed with IBD will present before the age of 20 years. Although less prevalent in infants and very young children, a number of case reports and small series document the onset of IBD at a very early age. 4-7 Despite significant recent advances in our understanding of the cause and pathogenesis of these diseases, many questions regarding the contribution of genetic predisposition and specific environmental triggers remain unanswered. The potential insight into these questions to be gained from rigorous epidemiologic studies based on large cohorts of patients with IBD is substantial. Also well recognized are the challenges inherent in developing a sizeable but accurate database for these analyses. In a recent review of the etiopathogenesis of IBD, Oliva- Hemker and Fiocchi 8 emphasize the importance of the pediatric perspective and, at the outset, stress the need for further pediatric studies based on reliable data collection and both regional and national databases. Until recently, studies in the pediatric literature have either been relatively small or geographically defined. In this issue of The Journal, Heyman et al 9 present the results of analysis of data from a large multicenter registry of children and adolescents with IBD that was developed to test specific hypothesis-driven epidemiologic questions. The cohort is demographically diverse and represents both incident and prevalent cases enrolled over approximately a 2-year period. This report describes the epidemiologic features of these children and adolescents including demographic details (age, sex, race/ethnicity), type [CD, ulcerative colitis (UC), indeterminate colitis (IC)] and anatomic distribution of disease, and family history of IBD. After an initial analysis by age group distribution, a distinct cutoff was noted at 8 years of age. This age was then used to define early-onset IBD and the group of children younger than 8 years of age was CD IBD IC UC Crohn disease Inflammatory bowel disease Indeterminate colitis Ulcerative colitis subsequently used for multivariate analysis of these epidemiologic variables. The authors highlight the significant differences in type and distribution of disease as well as in family history in the children with early-onset IBD. Results of the analysis of the disease type in the entire cohort were in agreement with several recent epidemiologic reports of pediatric patients with IBD. 2,3 The majority of children carried the diagnosis of CD (58%). Approximately half as many were diagnosed with UC (29%) and 13% were categorized as IC. Marked differences became apparent in IBD type, disease distribution, and family history when these characteristics were analyzed by age group. Both IC and UC were more prevalent at the younger ages. By multivariate analysis, children younger than 8 years of age were 2.5 times more likely to have UC and 3.5 times more likely to be diagnosed with IC than those older than 8 years of age. In addition, colonic involvement with or without disease at other sites was more prevalent in younger patients. The prevalence of colonic CD declined with increasing age, along with the proportion of patients carrying the diagnosis of UC and IC. Finally, patients with early-onset IBD were more likely to report a family history of IBD, particularly in the very young patients with UC. This striking difference in the younger patients has been suggested in earlier reports. 4-6 Although each study defined early onset at a different age, the results of all these studies lend credence to the concept that, at least from an epidemiologic standpoint, IBD is different in younger children. In the largest of these series, Mamula et al 4 reported the clinical characteristics of 82 children younger than 5 years of age with IBD. The disease distribution by type was CD, 33%; UC, 44%; and IC, 23%. This distribution is very similar to that reported in the Heyman study for early-onset IBD (younger than 8 years of age) and virtually mirrors the percentages when the 211 children younger than 5 years of age in See related article, p 35. Reprint requests: M. Susan Moyer, MD, University of Cincinnati College of Medicine, 3333 Burnet Ave, Cincinnati, OH J Pediatr 2005;146: /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds Editorials 7

56 the current report are compared (CD, 36%; UC, 40%; IC, 24%). The predominance of colonic disease is also apparent. Eighty-nine percent of the children with CD in the Mamula study had large-bowel involvement. In the series by Marx et al 5 in children younger than 2 years of age with CD, all 7 had Crohn colitis, and 27 of 38 (71%) children younger than 10 years of age in the Gryboski 6 series had disease involving the colon. One of the strengths of the current study was the size of the database, which provided the power to verify these important age-related differences and allowed statistical comparison with the entire cohort. There are limitations to this study, which the authors acknowledge. Inclusion of both prevalent and incident cases could impose bias despite the centers attempts to enroll consecutive patients presenting for clinic visits. This could be particularly problematic for patients with IC, in whom the initial diagnosis (incident cases) may change over time (prevalent cases). Some of the shortcomings, however, are inherent to the field of IBD in general. The classification of IBD by type remains a clinical one. Although the authors have tried to narrowly define CD, UC, and IC, there is currently no accepted standardized approach to the diagnosis. For example, not all patients being evaluated for IBD undergo an upper endoscopy. 10,11 In those who have had an endoscopic examination of their upper gastrointestinal tract, histologic and endoscopic findings can be equivocal. Even the addition of IBD serologies to our diagnostic armamentarium has not solved this problem. The patients with IC are particularly challenging. Whether the significantly higher proportion of IC in the patients with early-onset IBD represents our inability to differentiate Crohn colitis from UC, the inflammatory process in evolution, or a totally different subset of IBD is still in question. Although a number of the patients with IC in this cohort were eventually reclassified (38 of 179), approximately 10% retained the diagnosis of IC. Following these young patents with IC prospectively may shed considerable light on this question. In summary, this study has helped to define the epidemiology of IBD in children and has underscored the differences in early-onset disease. The authors are to be applauded for their efforts to work collaboratively to establish this registry that we hope will yield additional useful epidemiologic observations in the future. Identifying and validating differences among ages in disease phenotype and family history will help to lay the foundation from which questions regarding genetic predisposition and environmental impact on disease expression may be more effectively addressed. M. Susan Moyer, MD University of Cincinnati College of Medicine Schubert-Martin Inflammatory Bowel Disease Center Cincinnati Children s Hospital Medical Center Cincinnati, Ohio REFERENCES 1. Loftus EV, Sandborn WJ. Epidemiology of inflammatory bowel disease. Gastroenterol Clin North Am 2002;31: Kugathasan S, Judd RH, Hoffman RG, Heikenen J, Telega G, Khan FL, et al. Epidemiologic and clinical characteristics of children with newly diagnosed inflammatory bowel disease in Wisconsin: a statewide population-based study. J Pediatr 2003;1423: Lindberg E, Lindquist B, Holmquist L, Hildebrand H. Inflammatory bowel disease in children and adolescents in Sweden, J Pediatr Gastroenterol Nutr 2000;30: Mamula P, Telega GW, Markowitz JE, Brown KA, Russo PA, Piccoli DA, et al. Inflammatory bowel disease in children 5 years of age and younger. Am J Gastroenterol 2002;97: Marx G, Seidman EG, Martin SR, Deslandres C. Outcome of Crohn s disease diagnosed before two years of age. J Pediatr 2002;140: Gryboski JD. Crohn s disease in children 10 years old and younger: comparison with ulcerative colitis. J Pediatr Gastroenterol Nutr 1994;18: Gryboski JD. Ulcerative colitis in children 10 years old or younger. J Pediatr Gastroenterol Nutr 1993;17: Oliva-Hemker M, Fiocchi C. Etiopathogenesis of inflammatory bowel disease: the importance of the pediatric perspective. Inflamm Bowel Dis 2002; 8: Heyman M, Kirschner BS, Gold BD, Ferry G, Baldassano R, et al. Children with early onset inflammatory bowel disease (IBD): analysis of a pediatric IBD consortium registry. J Pediatr 2004;146: Kundhal PS, Storman MO, Zachos M, Critch JN, Cutz E, Griffiths AM. Gastrol antral biopsies in the differentiation of pediatric colitides. Am J Gastroenterol 2003;98: Abdullah BA, Gupta SK, Croffie JM, Pfefferkorn MD, Corkins MR, Fitzgerald JF. The role of esophagogastroduodenoscopy in the initial evaluation of childhood inflammatory bowel disease: a 7-year study. J Pediatr Gastroenterol Nutr 2002;35: SAFETY OF ADULT NERVE AGENT AUTOINJECTORS IN CHILDREN In this age of increased terrorism awareness, a great deal of attention and money is being placed on protecting the civilian population. Nerve agent poisoning is one form of terrorism that has already been used on the civilian population. 1 This knowledge has prompted US federal agencies to encourage the development of standardized treatment protocols, many of which use preloaded drug autoinjectors. Since the available autoinjectors were developed by the military, they contain drug dosages predicated on an average healthy adult, leading to appropriate concern among practitioners who treat children. Until last year, the only commercially available US nerve agent autoinjector was the Mark I (Meridian Medical Technologies, Columbia, Md) unit, containing two autoinjectors, one with 600 mg pralidoxime and the other with 2 mg atropine sulfate. The needles are designed for adult muscle mass and are large, hollow-bore needles See related article, p 41. Reprint requests: Cynthia Aaron, MD, Poison Control Center, Harper Professional Office Building, 4160 John R, Suite 616, Detroit, MI J Pediatr 2005;146: /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds Editorials The Journal of Pediatrics January 2005

57 made to inject in a fan-like fashion under pressure over approximately 10 seconds. Concern has been raised over both the appropriateness of the dose and the size of the needle in children. Recently, the FDA approved 2 a pediatric-dose atropine autoinjector (Atropen, Meridan Medical Technologies, Columbia, Md). The Atropen is available in two dosages: 0.5 mg atropine sulfate and 1.0 mg atropine sulfate on a smaller needle. The availability of the Atropen should help alleviate dosage concerns. However, since many smaller communities have limited resources for the purchase of variable-dose Atropen and training of prehospital personnel in appropriate pediatric dosing, this alternative may still lead to limited effectiveness. Addressing the dosing issue, Amitai, in 1992, 3 reviewed 240 instances of accidental pediatric atropine injection with adult dose based autoinjectors and found a low incidence of toxicity. They concluded that the home distribution of autoinjectors was sufficiently low risk (no seizures, arrhythmias, or death) versus the high benefit of having antidotal therapy immediately available in the Israeli home. Subsequently, several pediatric guidelines have suggested that adult-dose atropine and pralidoxime autoinjectors can be safely used in children larger than 13 kg and inserted to 0.8 inches. 4-7 The study in this issue of The Journal of Pediatrics (Kozer E, Mordel A, Bar Haim S, et al) reviewed data on 126 cases of accidental injection of atropine and trimedoxime (TMB-4) from new fixed-dose autoinjectors. The authors report that 22 of 126 (15%) of the children received atropine doses higher than the recommended 0.05 mg/kg, with low resulting toxicity, and 42 of 126 (30%) showed signs of atropinization. Data are not provided to show if these two groups overlap, although the authors do state that those children receiving a higher dose of atropine had a higher atropinization score. Regardless of dose or symptomatology, outcome was good and treatment was supportive in all cases. This is the first review of trimedoxime use in children, and, as such, we have no pediatric dose, pharmacokinetics, or toxicity data. The doses used in the autoinjectors were extrapolated from animal data. Because of this, there is no way to tease out if the doses received by the children were appropriate or truly considered an overdose. The contribution of trimedoxime to the recorded signs and symptoms of atropine toxicity is unclear. However, I believe that we can make good use of this information. All the available safety data are based on accidental dosing in healthy unexposed and asymptomatic children. In a nerve agent scenario, this will not be the case, although some children will inadvertently be administered autoinjectors without having been poisoned. Exposed children receiving antidotal therapy in a real scenario will have signs and symptoms of cholinergic excess. Multiple studies of symptomatic organophosphorous insecticide poisoned patients have demonstrated that once acetylcholinesterase is severely inhibited, the drug doses found in autoinjectors are more consistent with initial and not total dosing Indeed, the World Health Organization recommendations for treatment of organophosphorous insecticide poisoning advocate using a higher dosing strategy than is currently recommended in US texts for atropine and pralidoxime (30 mg/kg pralidoxime IV load followed by 8 mg/kg per hour or 4 mg/kg obidoxime load followed by 0.5 mg/kg per hour). 14 At least in adults, atropine needs appear to be higher than expected with insecticide poisoning. 4,6,9,10,12,13,15-17 The data suggest that rather than overdosing a child with an adult autoinjector, we may actually need additional drug to have an antidotal effect. I believe that the real issue in preparing for a mass casualty incident involving children exposed to a nerve agent lies in recognizing that the adult protocols currently in place do not uniformly fit the pediatric and other special populations. Children will have different physiologic and behavioral responses to chemical weapons, and, when poisoned by nerve agents, may not show the expected signs and symptoms. Children exposed to organophosphorous insecticides tend to have predominant central nervous system and pulmonary findings without muscarinic signs. 4,6,12,13 Consequently, titrating atropine to a heart rate >60 bpm and drying of pulmonary secretions may be less appropriate than titrating to the relief of bronchoconstriction. We need to educate first responders and general practitioners in the specifics of pediatric nerve agent exposure and treatment. Although the use of adult autoinjectors is not ideal, they can still be used without the fear that we will invariably overdose our patients. Cynthia Aaron, MD University of Massachusetts Medical Center Emergency Department Paxton, Massachusetts REFERENCES 1. Okumura T, Takasu N, Ishimatsu S, Miyanoki S, Mitsuhashi A, Kumada K, et al. Report on 640 victims of the Tokyo subway sarin attack. Ann Emerg Med 1996;28: FDA. FDA approves pediatric doses of Atropen. gov/bbs/topics/answers/2003/ans01232.html. Accessed December 14, Amitai Y, Almog S, Singer R, Hammer R, Bentur Y, Danon YL. Atropine poisoning in children during the Persian Gulf crisis: a national survey in Israel. JAMA 1992;268: Rotenberg JS, Newmark J. Nerve agent attacks on children: diagnosis and management. Pediatrics 2003;112: Markenson D, Redlener I. Pediatric preparedness for disasters and terrorism: a national consensus conference, executive summary, in: pediatric preparedness for disasters and terrorism, a national consensus conference. New York, New York: AHRQ DHHS EMSC Program of the Maternal and Child Health Resources Services Administration; Rotenberg JS. Diagnosis and management of nerve agent exposure. Pediatr Ann 2003;32: Henretig FM, Cieslak TJ, Eitzen E Jr. Biological and chemical terrorism. J Pediatr 2002;141: Eddleston M, Szinicz L, Eyer P, Buckley N. Oximes in acute organophosphorous pesticide poisoning: a stystematic review of clinical trials. QJ Med 2002;95: Eddleston M, Dawson A, Karalliedde L, Buckley N. Speed of initial atropinisation in significant organophosphorus pesticide poisoning-a systematic comparison of recommended regimens. J Tox Clin Toxicol 2004;42: Eddleston M, Singh S, Buckley N. Acute organophosphorus poisoning. British Medical Journal 2003;9:1542. Editorials 9

58 11. Eddleston M, Roberts D, Buckley N. Management of severe organophosphorus pesticide poisoning. Crit Care 2002;6: Lifshitz M, Rotenberg M, Sofer S, Tamiri T, Shahak E, Almog S. Carbamate poisoning and oxime treatment in children: a clinical and laboratory study. Pediatrics 1994;93: Lifshitz M, Shahak E, Sofer S. Carbamate and organophosphate poisoning in young children. Pediatr Emerg Care 1999;15: Feldmann RJ, Szajewski J. Cholinergic syndrome, in IPIC INTOX Databank; Golsousidis H, Kokkas V. Use of mg of atropine during 24 days of treatment, after a case of unusually severe parathion poisoning. Hum Toxicol 1985;4: McDononough JH Jr, Jaax NK, Crowley RA, Mays MZ, Modrow HE. Atropine and/or diazepam therapy protects against soman-induced neural and cardiac pathology. Fundam Appl Toxicol 1989;13: Thiermann H, Szinicz L, Eyer F, Worek F, Eyer P, Felgenhauer N, et al. Modern strategies in therapy of organophosphate poisoning. Toxicol Lett 1999;107: A TALE OF TWO CENTIMETERS It has been said, without too great a stretch of the imagination, that a rock will grow if one gives it enough growth hormone (GH). 1 Indeed, experiments of nature (eg, pituitary gigantism and acromegaly) demonstrate convincingly that most individuals with open epiphyses, in the presence of high serum concentrations of GH and its hormonal mediator insulin-like growth factor (IGF)-I, are capable of growing well beyond their genetic endowments. In a variety of clinical conditions characterized by growth failure, such as Turner syndrome 2 and intrauterine growth retardation, 3 for example, increasing pharmacologic dosages of GH have resulted in accelerated growth and adult heights well beyond those anticipated in untreated individuals. Not surprisingly, investigations using higher dosages of GH have demonstrated that a dose-response effect on growth exists, although, generally, the slope is relatively shallow. In children identified as GH deficient, for example, an increase in GH dosage from to 0.05 mg/kg per day resulted in a sustained increase in growth velocity, although no additional benefit was observed with a further increase to 0.1 mg/kg per day. 4 In recent studies in Turner syndrome, a clear doseresponse was demonstrated, although, once again, with only limited additional benefit at the highest GH dosage. 2 The current paper by Wit et al 5 follows on the heels of an earlier report by Leschek et al 6 on the use of GH in children with idiopathic short stature (ISS), a heterogeneous entity that includes children with familial short stature, constitutional delay, and normal GH production in the face of normal or decreased serum concentrations of IGF-I. The latter study, which constituted part of the basis for FDA approval of GH in ISS, demonstrated a statistically significant, albeit modest, effect of GH on adult heights of children with ISS, with an increase in adult height of 3.7 cm in GH-treated versus placebo-treated subjects. 6 It was recognized, however, that a potentially greater benefit of GH therapy may well have been masked by the relatively advanced age of the subjects and the modest GH dosage of mg/kg three times per week (0.22 mg/kg per week). Given this background, it is not particularly surprising that the current study by Wit et al 5 demonstrates a doseresponse effect of GH in patients with ISS. The investigations, although carefully conducted and analyzed, are characterized by GH IGF ISS Growth hormone Insulin-like growth factor Idiopathic short stature a number of the short-comings inevitable to GH studies designed over a decade ago: (1) selection of a rather limited dose range for GH (0.24 vs 0.37 mg/kg per week); (2) use of a stepwise increase in GH dosage for one arm rather than simply starting subjects at the highest dosage; (3) a high dropout rate; (4) absence of either an observational control or placebo control group; and (5) inadequate surveillance of serum concentrations of IGF-I, as a marker of both GH responsiveness and safety. Despite these caveats, the study did demonstrate a statistically significant difference in the growth response of the two extremes of GH dosage tested (ie, 0.24 vs 0.37 mg/kg per week), with mean overall height gains over the baseline predicted adult heights of 5.4 and 7.2 cm, respectively. Although this difference in adult height is less than 2 cm, it is of note that 94% of subjects receiving the high dose of GH achieved adult heights within the normal range. Even in the absence of a control group, this is an impressive accomplishment and may be compared favorably with the report by Leschek et al, 6 using the low GH dosage of 0.22 mg/kg per week. Indeed, the investigations of Leschek et al 6 and Wit et al 5 should be viewed in the same context, even though the studies were independent in both their design and execution. Taken together, they provide strong (if not conclusive) evidence that GH is capable of accelerating growth and improving adult height in most if not all children with ISS. It would appear that if initiated at a sufficiently early age and at an appropriate dosage, GH can lead to an adult height within the normal range in the majority of children with ISS. As research continues into the molecular basis for the heterogeneous conditions that constitute ISS, 7 it is likely that our therapeutic acumen will be sharpened, permitting further optimization of cost-benefit ratios for GH use in this population as well as an assessment of the potential use of IGF-I therapy in children who demonstrate some degree of GH insensitivity as an etiologic factor in their growth failure. The fact that we can treat a condition does not in and of itself imply that we should, and it is this issue that remains most germane to the management of idiopathic short stature. Critics of pharmacologic intervention have pointed out, with some legitimacy, that there are few See related article, p 45. Reprint requests: Ron G. Rosenfeld, MD, 770 Welch Road, Suite 350, Palo Alto, CA J Pediatr 2005;146: /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds Editorials The Journal of Pediatrics January 2005

59 studies that have demonstrated significant psychosocial morbidity in otherwise normal short children and that no data exist supporting a psychologic benefit to growth-promoting therapy in such patients. 8 Although such investigations have tended to focus on cross-sectional evaluations of school children rather than on the more appropriate population of children who are actually referred for pediatric endocrine evaluation, the issues raised do require our attention. On the other hand, if GH can be shown to shown to be effective in a significant number of children now labeled as having ISS, as in the current study by Wit et al, 5 it becomes difficult to justify treatment of children carrying a diagnosis of GHD while excluding those with ISS. 9 This matter is complicated further by the acknowledged difficulty in making a firm diagnosis of GHD in many cases and by the recognition that at least 50% of cases diagnosed and treated as GHD probably do not have genuine pituitary pathology. Furthermore, in many other situations such as Turner syndrome, chronic renal failure, and intrauterine growth retardation, FDA approval of GH treatment was predicated on the management of the growth failure per se rather than an underlying endocrine deficiency. A summary of this underlying conundrum would dictate that we are limited to only two logically defensible positions: (1) restrict GH therapy to replacement for children with unequivocal GHD (and acknowledge that only a minority of children diagnosed and treated as having GHD are truly deficient in GH), or (2) recognize that the effectiveness of growth-promoting therapy should dictate access and that a significantly short child who might benefit from treatment deserves consideration of treatment, no matter what the underlying diagnosis might be. 9 The latter option is appealing but mandates that the following aspects are demonstrable: (1) that short stature represents a disability to the child and that it is not amenable to more conservative approaches such as counseling and reassurance; (2) that therapy be demonstrated to be effective (and safe) at promoting both short-term growth and adult height; (3) that growth augmentation provide psychosocial benefit to the child; and (4) that a meaningful cost-benefit assessment be made in the light of the issues listed above. The studies by Leschek et al and Wit et al provide strong support for the efficacy of GH therapy. It is to be hoped that future investigations will address the perhaps more complex issues of genuine benefit to the patient. Ron G. Rosenfeld, MD Senior Vice-President for Medical Affairs Lucile Packard Foundation for Children s Health Palo Alto, CA Professor of Pediatrics Stanford University Stanford, CA Professor of Pediatrics Oregon Health and Science University Portland, OR REFERENCES 1. Rosenfeld RG. Ten axioms in the evaluation of short stature. Endocrinologist 1997;7: Sas TC, de Muinck Keizer-Schrama SMPE, Stijnen T, Jansen M, Otten BJ, Hoorweg-Nijman JJG, et al. Drop SLS: normalization of height in girls with Turner syndrome after long-term growth hormone treatment: results of a randomized dose-response trial. J Clin Endocrinol Metab 1999;84: Sas T, de Waal W, Mulder P, Houdijk M, Jansen M, Reeser M, et al. Growth hormone treatment in children with short stature born small for gestational age: 5-year results of a randomized, double-blind, dose-response trial. J Clin Endocrinol Metab 1999;84: Cohen P, Bright GM, Rogol AD, Kappelgaard A-M, Rosenfeld RG. Effects of dose and gender on the growth and growth factor response to GH in GH-deficient children: implications for efficacy and safety. J Clin Endocrinol Metab 2002;87: Wit JM, Rekers-Mombarg LTM, Cutler GB Jr, Crowe B, Beck TJ, Roberts K, et al. Growth hormone (GH) treatment to final height in children with idiopathic short stature: evidence for a dose effect. J Pediatr 2004;146: Leschek EW, Rose SR, Yanovski JA, Troendle JF, Quigley CA, Chipman JJ, et al. Effect of growth hormone treatment on adult height in peripubertal children with idiopathic short stature: a randomized, doubleblind, placebo-controlled trial. J Clin Endocrinol Metab 2004;89: Rosenfeld RG, Hwa V. Toward a molecular diagnosis for idiopathic short stature. J Clin Endocrinol Metab 2004;89: Sandberg DE, Bukowski WM, Fung CM, Noll RB. Height and social adjustment: are extremes a cause for concern and action? Pediatrics 2004;114: Allen DB, Fost N. hgh for short stature: ethical issues raised by expanded access. J Pediatr 2004;144: CONFIRMATION OF AN OLD ADAGE: YOU FIND WHAT YOU SEEK The article by Dr Mitchell Cohen and his colleagues published in this issue of The Journal 1 is best considered in the context of the global problem of diarrhea in children, which continues to plague humanity. Infants and young children are particularly prone to diarrhea and are especially vulnerable to the consequences of volume depletion. However, extra-intestinal and delayed consequences of enteric infections, such as Campylobacter jejuni related and C coli related Guillain-Barre syndrome, 2 postinfectious irritable bowel syndrome, 3 and hemolytic uremic syndrome caused by Shiga toxin producing Escherichia coli and Shigella DAEC EAEC Diffusely adherent Escherichia coli Enteroaggregative Escherichia coli dysenteriae, type 1, 4 are now emerging as complications of diarrheal infections. Indeed, in a recent study undertaken in California of bacterial infections of the gut, sequelae such as protracted diarrhea, irritable bowel syndrome, and arthritis were self-reported in 27% of 571 respondents. 5 The majority of cases of acute diarrhea in children are caused by viruses. Rotavirus is the most common cause, 6 but See related article, p 54. Dr Sherman is the recipient of a Canada Research Chair in Gastrointestinal Disease. Reprint requests: Dr Philip M. Sherman, Gastroenterology and Nutrition, Room 8409, Hospital for Sick Children, 555 University Ave, Toronto, Ontario, M5G 1X8, Canada. J Pediatr 2005;146: /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds Editorials 11

60 enteric adenoviruses, caliciviruses, and astroviruses are increasingly recognized as causative agents in both outbreaks and sporadic cases of illness. Diverse bacteria and parasites also can cause acute enteric illnesses, and, in the past 20 years, a number of organisms have been identified as etiologic agents in both sporadic cases and outbreaks of diarrheal diseases. 7 Accumulating evidence regarding the transmission of such agents has raised concern and increasing awareness of the importance for careful monitoring and surveillance to ensure safe food and water supplies. Should health care providers care about the precise cause of a patient s diarrhea? In many cases, the answer is no. Primary treatment is independent of cause and should continue to focus on interventions aimed at maintaining intravascular volume through the use of appropriate rehydration solutions. 8 Usually rehydration is administered orally, an intervention that has saved millions of lives worldwide, and should continue to be the cornerstone of all intervention strategies. 9 The futility of finding specific causes of diarrhea is well known to health care practitioners. Even when applied, broadbased diagnostic approaches fail to find a cause, either because the patient has been cleared of causative agent 10 or we do not know what organisms to seek and how precisely to seek them. Financial considerations of diagnostic laboratory testing further limit the spectrum of agents that can be considered. When one considers the breadth of organisms that have been proposed to cause diarrhea and the diverse and expensive technologies frequently required to identify them accurately, focused evaluations or empiric therapies remain reasonable approaches in the clinical setting. However, the assumption that diagnostic evaluations of the cause of diarrhea are necessarily futile or have little aggregative value can be counterproductive, both in individual cases and for society at large. Identifying a specific bacterium or parasite as a pathogen in a stool sample sometimes can have considerable value for the individual patient; for example, if the organism identified can be treated with an appropriate antimicrobial agent and such an intervention has a documented benefit on either the severity or duration of illness. Establishing the diagnosis of a reportable enteric infection is also critical for disease control authorities. 11 For instance, it is important to know whether affected patients have acquired the infection from a source continuing to transmit the agent to others. This might include, for example, a contaminated recreational or municipal water supply or a food vehicle that is still being distributed and consumed. 12 Beyond providing guidance for specific therapy and public health interventions, there is another important reason to seek to identify the causes of diarrhea. Knowledge of the cause provides an underpinning to efforts to find preventive approaches (for example, vaccines) and treatment strategies for specific illnesses. In addition, the identification of new pathogens leads to a better understanding of reservoirs and environmental niches and thereby methods to limit acquisition. For example, the hazards of eating undercooked ground beef remained unrecognized until Shiga toxin producing E coli, including those of the serotype O157:H7, was identified as a human pathogen in the early 1980s. 4,13 Because of the expense and effort required, investigations of the cause of diarrhea probably are best reserved for focused research projects. The article by Cohen et al 1 from the Children s Hospital in Cincinnati, Ohio, is one such contribution to our knowledge base. Cohen and colleagues 1 confirm the primacy of rotavirus as an enteric pathogen in children with diarrhea. In addition, using a control group, they provide evidence of human pathogenicity of enteroaggregative E coli (EAEC), diffusely adherent E coli (DAEC), and atypical enteropathogenic E coli. EAEC is so named because of its adherence pattern to cultured epithelial cells. Accumulating evidence suggests that EAEC is associated with traveler s diarrhea and persistent diarrhea. 13 As discussed by Cohen et al, 1 specialized DNA probe tests are required for optimal detection of EAEC. DAEC have been found more frequently in cases than in control subjects in some but not all clinical studies of diarrhea in children. 13 Cohen et al 1 provide additional case-control data supporting a causative role of DAEC in childhood diarrhea. Future studies will require infection of appropriate animal models and human volunteers to fulfill Koch s postulates and confirm the virulence of DAEC. It should be remembered, however, that the DAEC probably is a heterogeneous group that could differ in virulence and pathogenesis. Thus, detailed analyses of adequately powered studies will be needed to either confirm or refute the hypothesis that at least a subset of DAEC isolates cause diarrhea. Cohen et al 1 also present important negative data. They found a surprisingly low rate of recovery of bacterial species that were previously identified as enteric pathogens in human beings. One possible explanation for these findings is that stool culture testing was applied broadly and not just to subjects with bloody diarrhea, in whom the chances of identifying agents such as C jejuni, Yersinia enterocolitica, nontyphoid Salmonella, and Shigella species would be higher. It is also possible that delays in presentation to the emergency department of a referral center could reduce the rate of recovery of bacterial enteropathogens from subjects stools. Several important enteric pathogens were not sought in the study by Cohen et al. 1 Attempts to detect viruses other than rotavirus were not made. Many practitioners find Clostridium difficile and its toxins in stool samples obtained from children with community-acquired diarrhea, even in the absence of history of recent antibiotic use. A future study using a study design comparable with the one used by Cohen et al 1 could prove instructive in demonstrating the frequency of identifying such organisms in patients versus age-matched community control subjects. What should now be done to advance the field and improve the care of children with diarrhea? First, future studies should be performed in additional well-defined populations with the use of appropriately matched unaffected control subjects and with sufficient resources provided to seek a broader range of pathogens and candidate pathogens. Second, the diagnostic tests for enteric pathogens must be made more 12 Editorials The Journal of Pediatrics January 2005

61 economical. Adjunctive studies, for example, testing for calprotectin or lactoferrin in stool samples, 14 might prove helpful in targeting those subjects in whom a full diagnostic evaluation should be undertaken. Third, technological advances are urgently required so that obtaining microbiological test results on feces can become as rapid, economical, and accurate as performing an automated blood count. The knowledge that the consequences of enteric infections can go beyond dehydration provides an impetus to improve current diagnostic capabilities, especially if it can be shown in future studies that these sequelae can be averted by interventions that are based on knowledge of specific causative agents. 15 The ability to make a specific diagnosis in the clinical setting depends on an improvement in the current understanding of the full spectrum of potential etiologic agents. 16 Cohen and his associates provide a model for such investigations. Philip M. Sherman, MD, FRCPC Department of Pediatrics Research Institute, Hospital for Sick Children University of Toronto, Toronto, Ontario, Canada Phillip I. Tarr, MD The Department of Pediatrics Washington University School of Medicine, St Louis, Missouri REFERENCES 1. Cohen MB, Nataro JP, Berstein DI, Hawkins J, Roberts N, Staat MA. Prevalence of diarrheogenic Escherichia coli in acute childhood enteritis: a prospective controlled study. J Pediatr 2004;146: Moran AP, Prendergast MM. Molecular mimicry in Campylobacter jejuni and Helicobacter pylori lipopolysaccharides: contribution of gastrointestinal infections to autoimmunity. J Autoimmun 2001;16: Wang L-H, Fang X-C, Pan G-Z. Bacillary dysentery as a causative factor of irritable bowel syndrome and its pathogenesis. Gut 2004;53: Blaser MJ. Bacteria and diseases of unknown cause: hemolytic-uremic syndrome. J Infect Dis 2004;189: Rees JR, Pannier MA, McNees A, Shallow S, Angulo FJ, Vugia DJ. Persistent diarrhea, arthritis, and other complications of enteric infections: a pilot survey based on California FoodNet surveillance, Clin Infect Dis 2004;38(suppl 3):S Parashar UD, Hummelman EG, Bresee JS, Miller MA, Glass RI. Global illness and deaths caused by rotavirus disease in children. Emerg Infect Dis 2003;9: Blaser MJ, Smith PD, Ravdin JI, Greenberg HB, Guerrant RL, editors. Infections of the gastrointestinal tract. 2nd edition. Philadelphia: LWW; Thielman NM, Guerrant RL. Acute infectious diarrhea. N Engl J Med 2004;350: Casburn-Jones AC, Farthing MJG. Management of infectious diarrhoea. Gut 2004;53: Tarr PI, Neill MA, Clausen CR, Watkins SL, Christie DL, Hickman RO. Escherichia coli O157:H7 and the hemolytic uremic syndrome: importance of early cultures in establishing the etiology. J Infect Dis 1990; 162: Nestle M, editor. Safe food: bacteria, biotechnology, and bioterrorism. Berkley, CA: University of California Press; Acheson DWK, Fiore AE. Preventing foodborne disease: what clinicians can do. N Engl J Med 2004;350: Kaper JB, Nataro JP, Mobley HLT. Pathogenic Escherichia coli. Nature Rev Microbiol 2004;2: Tibble JA, Sigthorsson G, Foster R, Forgacs I, Bjarnason I. Use of surrogate markers of inflammation and Rome criteria to distinguish organic from nonorganic intestinal disease. Gastroenterology 2002;123: Frenzen PD. Deaths due to unknown foodborne agents. Emerg Infect Dis 2004;10: Salazar-Lindo E, Allen S, Brewster DR, Elliott EJ, Fasano A, Phillips AD, et al. Intestinal infections and environmental enteropathy: working group report of the second World Congress of Pediatric Gastroenterology, Hepatology, and Nutrition. J Pediatr Gastroenterol Nutr 2004;39:S Editorials 13

62 ORIGINAL ARTICLES CHANGES IN THE PROPORTION AND VOLUME OF CARE PROVIDED TO CHILDREN BY GENERALISTS AND SUBSPECIALISTS GARY L. FREED, MD, MPH, TAMMIE A. NAHRA, PHD, PATRICIA J. VENUS, MA, STEPHANIE D. SCHECH, MPH, JOHN R. C. WHEELER, PHD, AND THE RESEARCH ADVISORY COMMITTEE OF THE AMERICAN BOARD OF PEDIATRICS Objective To assess whether primary care physicians, via referrals or other mechanisms, are now providing proportionally less care for children with specific common diagnoses, thus driving greater demand for specialist services. Study design Secondary data analysis ( ) from one of the largest commercial healthcare organizations in the United States. Evaluation and management (E/M) common procedural terminology (CPT) visit codes and International Classification of Diseases (ICD) codes pertaining to asthma, constipation, headache, and heart murmurs were selected. Visits were then assigned to the specialty of physician providing care. Significant differences between and among categories of physicians were tested using logistic regression. Results Overall, pediatrician generalists and specialists provided a greater proportion of E/M visits to children in 2001 than in 1993, compared with nonpediatrician providers. However, although the absolute increase in the proportion of all E/M visits by children <18 years of age to pediatrician generalists was greater than that of pediatrician subspecialists (4.77% vs 0.69%; P <.0001), the relative increase was much smaller for the generalists (8.9% vs 19.7%; P <.0001). Findings were consistent for most of the specific diagnoses examined. Conclusions The increases in both the proportion and number of visits made to specialists has not been accompanied by a decrease in visits to generalists. (J Pediatr 2005;146:14-9) As recently as the year 2000, published reports suggested a diminishing market for pediatric subspecialty care. In one large survey, a majority of pediatric subspecialists in 15 of 17 subspecialties believed there would be no need in their communities over the next 3 to 5 years for additional subspecialists in their disciplines. 1 This work followed other studies that claimed to demonstrate an aggregate oversupply of subspecialists but did not separate pediatric from adult physicians. 2,3 However, more recently, both the medical literature and the popular press have published reports describing the scarcity of specialists in general, and pediatric specialists in particular. 4,5 Many also have noted long wait times to see subspecialists as well as the difficulty in recruiting enough subspecialists to fulfill clinical needs at major institutions. 6 This perceived scarcity has occurred while the number of pediatric specialists has risen considerably over the past two decades. 7 Why the long queues for subspecialty care? Several factors have been raised as possibilities including prolonged survival of children with chronic diseases and of those born with low birth weight. 7,8 An additional hypothesis in pediatrics is the possibility that primary care physicians, via referrals or other mechanisms, are changing their patterns of care and are now providing proportionally less care for children with specific common diagnoses, thus driving greater demand for specialist services. To test this hypothesis, we examined both the absolute and the relative changes in the proportion of care provided to children by pediatric and nonpediatric generalists and specialists, overall and for specific common diagnoses, over a 9-year period. Three of these diagnoses, asthma, constipation, and headache, are common conditions that are frequently AMA CPT American Medical Association Common procedure terminology E/M ICD Evaluation and management International Classification of Diseases See editorial, p 3, and related article, p 20. From the Child Health Evaluation and Research (CHEAR) Unit, the Division of General Pediatrics, and the Department of Health Management and Policy, University of Michigan, Ann Arbor, Michigan; the Center for United Health Care Policy and Evaluation, Minneapolis, Minnesota; and The American Board of Pediatrics, Chapel Hill, North Carolina. Supported by a grant from the American Board of Pediatrics Foundation. Submitted for publication Jun 3, 2004; last revision received Jul 28, 2004; accepted Aug 23, Reprint Requests: Gary L. Freed, MD, MPH, University of Michigan, 300 NIB 6E08, Ann Arbor, MI gfreed@med.umich.edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

63 Table I. Percent visit by physician for curative visits Medical generalist 30.4% 31.3% 31.5% 31.1% 30.2% 28.8% 28.4% 27.6% 26.8% Medical specialist 18.1% 17.3% 16.5% 17.0% 16.9% 16.8% 16.4% 16.5% 16.2% Pediatric medical generalist 48.7% 48.7% 49.3% 48.7% 49.1% 50.9% 51.8% 52.6% 53.5% Pediatric medical specialist 2.8% 2.8% 2.7% 3.3% 3.8% 3.5% 3.4% 3.3% 3.5% Table II. Number of visits per 1000 annualized member months by physician for curative visits Medical generalist Medical specialist Pediatric medical generalist , ,024 Pediatric medical specialist managed in primary care but have been anecdotally reported to be increasingly seen in subspecialty clinics. We hypothesized that, because of the generally lower acuity of these three conditions, many visits to subspecialists for these conditions may be discretionary. Conversely, for the fourth diagnosis chosen, heart murmurs, we hypothesized to reflect a greater potential for acuity and thus less discretion in subspecialty utilization. As such, subspecialty utilization for the diagnosis of heart murmurs would likely remain constant over time, whereas utilization for other diagnoses may change to reflect new patterns of discretionary subspecialty utilization. METHODS Data for this project are for the years 1993 to 2001 and originate from a large multi-state healthcare organization in the United States. Five commercial HMO and pointof-service plans within the organization were selected based on plan maturity, regional diversity, and size of membership enrollment. Located in the Southeast, Midwest, Northeast, and West, total combined enrollment of these plans exceeded 1.7 million clients for Enrollment in these plans increased approximately 100% during the 9-year study period. To analyze care patterns for pediatric outpatient visits, focus was placed on the pediatric population, <18 years of age, for each study year of interest. The number of annualized pediatric member months from 1993 to 2001 ranged from 188,991 to 407,742. Visits of all enrollees, regardless of the duration of enrollment in a plan, were selected based on either common procedural terminology (CPT) codes of interest or particular International Classification of Diseases (ICD) codes. For this study, evaluation and management (E/M) CPT visit codes and ICD codes pertaining to asthma, constipation, headache, and heart murmur were selected. E/M CPT codes included 94150, , , , and For asthma, the ICD grouping consisted of all 493.xx. For constipation, the ICD groupings were 564.0x and For headache, ICD grouping included 784.xx, , and 346.xx. For heart murmur, ICD groupings 424.xx and were selected. More than 5 million E/M visits and almost 500,000 visits for the diagnoses of interest met these criteria and were reviewed over the 9-year period. To obtain a greater level of physician specialty and subspecialty detail than what had been available historically, the commercial health organization merged its practicing physician records with data from the American Medical Association (AMA) Physician Masterfile. Containing current and historical data, the AMA Physician Masterfile is one of the most comprehensive listings of physicians in the United States and includes both members and nonmembers of the AMA. Using the more detailed information from the AMA, medical physicians were grouped into one of the following categories: (1) pediatrician generalist; (2) pediatrician specialist; (3) nonpediatrician generalist; and (4) nonpediatrician specialist. We summarized visit activity across the 9 years of data in two ways: (1) percent of total visits by physician specialty category and (2) number of visits by physician specialty category per 1000 annualized children member months. Significant differences between and among categories for the first and last years of our study period were tested using x 2 statistics. We used time series regression analyses to measure changes over time for the two visit activity measures. For the first visit activity measure, the dependent variable was defined as the total number of visits by physician specialty category divided by total number of visits in a given year. For the second visit activity measure, the dependent variable was defined as the total number of visits for each physician specialty divided by the annualized number of member months for children <18 years of age. To assess changes over time, the independent variable in the time series regressions was a categorical variable representing year of visit. Separate time series analyses were completed for the two dependent variables for each physician specialty category, respectively, for the five visit groups: E/M, asthma, constipation, headache, and heart murmur. Because the dependent variable measures were represented by proportional Changes In The Proportion And Volume Of Care Provided To Children By Generalists And Subspecialists 15

64 Figure 1. A, Percent of visits by physician specialty category for asthma. B, Number of visits by physician specialty category per 1000 annualized children member months for asthma. measures, the time series trend was analyzed using the logistical form of generalized linear models. RESULTS Overall, pediatrician generalists and pediatrician specialists provided a greater proportion of E/M visits for all causes to children in 2001 compared with 1993, concomitant with similar decreases in the proportion of care provided by nonpediatrician providers (Table I). Further, although the absolute increase in the proportion of all E/M visits by children <18 years of age to pediatrician generalists was greater than that of pediatrician subspecialists (4.77 percentage points vs 0.69 percentage points; P <.0001), the relative increase was much smaller for the generalists (8.9% vs 19.7%; P <.0001). To control for changes in enrollment, we then examined these same trends by the number of visits per year per 1000 members <18 years of age and found similar results (Table II). Again, the absolute increase from 1993 to 2001 in the number of visits per 1000 members to pediatrician generalists was greater than that to pediatrician subspecialists (30 vs 9; P <.0001) but smaller when calculated as relative increases (3% vs 13.5%; P <.0001). Analysis of the E/M visit group time trends (using data from each year of study) for the two dependent variables (Tables I and II) show that per year pediatric medical generalists significantly increased the proportion of care (OR = 1.027; CI = , ) and the number of visits per member (OR = 1.042; CI = 1.039, 1.045) during the Figure 2. A, Percent of visits by physician specialty category for constipation. B, Number of visits by physician specialty category per 1000 annualized children member months for constipation. 9-year period. Pediatric medical specialists exhibited similar increases per year for visit proportion (OR = 1.029; CI = , ) and visits per member (OR = 1.025; CI = , ). During this same period, both medical generalists and medical specialist experienced significant declines in E/M care provision. Figure 1A demonstrates the changes in the proportion of care provided for visits for asthma. When comparing the year 1993 with the year 2001, increases were seen in the proportion of visits made to pediatrician generalists (30% vs 44%; P <.0001) and to nonpediatrician generalists (17% vs 22%; P <.0001), whereas decreases occurred in the proportion of visits to pediatrician specialists (11% vs 6%; P <.0001) and to nonpediatrician specialists (42% vs 28%; P <.0001) for the care of this condition. Figure 1B shows that although the overall number of visits per 1000 members <18 years of age for asthma has decreased from 1993 to 2001, those decreases were experienced most by pediatrician specialists (10.5 vs 4.9; P <.0001) and by nonpediatrician specialists (40.2 vs 21.7; P <.0001). Conversely, the number of visits per 1000 members to pediatrician generalists increased (20 vs 33.6; P <.0001). Results of the time trend analysis over the 9-year period demonstrated an increase in the proportion of asthma visits to medical generalists (OR = 1.034; CI = , ) and to pediatric medical generalists (OR = 1.088; CI = , ). Visits per member also significantly increased for these two specialist groups. 16 Freed et al The Journal of Pediatrics January 2005

65 With regard to visits for constipation (Figure 2A), comparing the year 1993 with the year 2001, we found the absolute increase in the proportion of visits to pediatrician generalists was greater than that to pediatrician specialists (including pediatric gastroenterologists) (7% vs 4%; P <.0001). However, the relative increase for the specialists in the proportion of care provided between 1993 and 2001 was 200% (from 2% to 6%; P <.0001). Changes in the number of visits for constipation per 1000 members <18 years of age (Figure 2B) demonstrate marked absolute increases for pediatrician generalists and striking relative increases for both pediatrician generalists (137%) and pediatrician specialists (609%). Results of the time trend analysis revealed pediatric generalists (OR = 1.049; CI = , ) and pediatric specialists (OR = 1.083, CI = , ) provided on average an increasing percentage of visits per year for constipation, whereas the nonpediatricians provided proportionally less care. All four of the physician groups, however, did provide additional constipation visits per member at a significant level. Pediatric medical specialists, with an OR = 1.19 (CI = , ) experienced the greatest relative increase per member. For headache, Figure 3A demonstrates an increase in the proportion of visits provided by pediatrician generalists but little change for specialists. When comparing the year 1993 with the 2001, for the number of visits per 1000 members <18 years of age, a different pattern emerges (Figure 3B). For the pediatrician generalists, there is a clear increase in the absolute number of visits per 1000 members when comparing 1993 with 2001 (6.5 vs 11.2; P <.0001), less so for both groups of specialists. However, there is a significant rise in the relative number of visits per 1000 members to both pediatrician generalists (91%) and pediatrician specialists (50%) for this condition. Results of the time trend analysis for headache are similar to constipation visits. Both pediatric generalists (OR = 1.058; CI = , ) and pediatric specialists (OR = 1.016; CI = , ), averaged annual increases in the proportion of care provided. Again, all the physician groups showed a significant increase over time in the number of visits per member. Pediatric generalists (OR = 1.104; CI = , ) and pediatric specialists (OR = 1.075; CI = , ) had rates of increase that outpaced the medical generalist (OR = 1.036; CI = , ) and medical specialist (OR = ; C I= , ). Examination of patterns of visits for heart murmurs showed no notable changes in either proportion of visits to specific groups of generalist or specialist physicians, or visits per member. DISCUSSION Our findings demonstrate that for E/M visits overall, and for patients with two of the four conditions examined (constipation or headache), there has been an increase in the proportion of care provided by pediatrician generalists and pediatric specialists as well as an increase in the number of Figure 3. A, Percent of visits by physician specialty category for headache. B, Number of visits by physician specialty category per 1000 annualized children member months for headache. visits per 1000 members <18 years of age to these physicians. Some of the change appears as a result of growth in the market share of pediatricians (both generalists and specialists) relative to nonpediatrician providers. However, there has also been a strong increase in the overall number of physician visits per 1000 members for constipation and headache. These results indicate that primary care physicians are not providing less care for any of the diagnoses studied, but that specialists are indeed providing a greater volume of visits for patients with constipation and headache. Hence, the increases in both the proportion and number of visits made to specialists have not been accompanied by a decrease in visits to generalists. Forrest et al found that referral occurs in approximately 1 in 40 pediatrician office visits and that only 27.9% of these were to medical, as opposed to surgical, specialists. 9 As such, it is unlikely that a change in referral patterns has been the predominant factor in the increased visits to specialists. The role of self-referral, however, is unknown for pediatric patients but is believed to be quite high for adult patients, especially in relationship to other countries. 10 Although the absolute increase in the number of visits to specialists is quite small relative to the total number of E/M visits overall, the magnitude of the proportion of increase for specialists is noteworthy, especially for constipation and headache specifically. Because there are so few pediatric specialists, even small absolute increases in patient volume can have a major impact on availability of appointments. As such, even the relatively small absolute increases found in this study have probably had a dramatic effect on both the perceived and Changes In The Proportion And Volume Of Care Provided To Children By Generalists And Subspecialists 17

66 actual availability of pediatric specialty care. Which children are accounting for these increases is unknown, but Kuhlthau et al found that, in general, children most likely to receive specialty care were younger, lived in urban areas, were white, and had higher disease severity. 11 The finding of increases in visits per 1000 members for relatively common conditions (constipation and headache) warrants further examination. Our study did not address the reasons why such growth has occurred. It is possible that the rise in visits per 1000 members may reflect changes in the composition of the member pool or secular trends in the healthcare-seeking behavior of parents. It also could represent an increase in the frequency of follow-up care recommended by pediatrician generalists and specialists. The decreases in the volume of specialty care delivered for asthma over the past 9 years was juxtaposed against the finding of increased visits to both pediatrician and nonpediatrician primary care providers for this condition. The implications of this finding are unclear as previous work has found subspecialty care for asthma to be dependent on a number of factors including patient age, severity of illness, and insurance status. 12 It is possible that the publication of Clinical Practice Guidelines from the National Heart, Lung, and Blood Institute 13 and others has resulted in a greater confidence among primary care physicians in their ability to treat this condition. Whether this pattern of care delivery has had a positive or negative impact on the health of children with asthma is unknown. The fact that the proportions and number of visits for cardiac murmur were relatively unchanged over the period of this study highlights the difference between this and the other diagnoses. Murmurs have not appreciably changed in their incidence over the past 10 years. Further, the presence of murmurs, or the decision to seek care for them, is likely not as subjective as that for more symptom-based conditions such as constipation and headache. Previous studies that have examined primary or generalist care delivery in the aggregate miss the dynamics within each specific primary care discipline. 13,14 Failure to examine the primary care specialties individually can have significant implications as each has its own unique factors, especially when relating to workforce issues. 15 Also, aggregating generalists misses important changes in market share that has significant implications. 16 The increase in market share for both general and specialty pediatrician providers is a trend that has been described previously. 17 Most of the change in market share has occurred over the past 5 years. Continuation of this trend will have a marked impact on the availability of subspecialty care for children. It is unclear at this time whether this shift is occurring as a result of patient preferences, or the efforts of nonpediatrician practitioners to concentrate their practices on adult patients. As demographic changes in this country are resulting in a greater proportion of the US population being >18 years of age, market pressures may now be placed on nonpediatrician providers to concentrate their practices in the adult age group. Economic considerations also may play a role in nonpediatrician physician preferences for adult patients. Most visits and procedures for adults are reimbursed more generously than for children, even when effort is equivalent. Further, relative to Medicare physician fees, Medicaid fees fell by 14.3% between 1993 and Limitations of the study include the potential for misattribution of physician specialty in the AMA Masterfile. Although this is the most comprehensive list of all physicians in the United States and includes both members and nonmembers of the AMA, specialty designation is selfreported. Another limitation is that our results include only outpatient care; emergency department visits are not included in these analyses. An additional limitation is that as this is an analysis of claims data, no independent verification of disease coding was performed. As such, changes over time in the pattern of coding for certain conditions may have occurred and influenced our results. Finally, the data presented come from one commercial health organization. Although this organization operates in multiple states, its benefit plans and network arrangements may differ from that of other plans and public programs. We conclude that marked changes have occurred over the past 10 years with regard to both the volume and proportion of care provided by pediatrician and nonpediatrician generalists and specialists. However, the increases in patient volume seen among the specialists are accompanied by increases, not decreases, in volume for primary care providers. Further, examination of visit patterns by the specific specialty of the physician is essential to appreciate changes in the source of healthcare delivery to our nation s children. Additional studies are required to determine whether these changes have had a positive or negative impact on children s health. REFERENCES 1. Stoddard JJ, Cull WL, Jewett EA, Brotherton SE, Mulvey HJ, Alden ER. Providing pediatric subspecialty care: a workforce analysis. AAP Committee on Pediatric Workforce Subcommittee on Subspecialty Workforce. Pediatrics 2000;106: Donelan K, Blendon RJ, Lundberg GD, Calkins DR, Newhouse JP, Leape LL, et al. The new medical marketplace: physicians views. Health Aff (Millwood) 1997;16: Schroeder SA. How can we tell whether there are too many or too few physicians? the case for benchmarking. JAMA 1996;276: Landro L. As specialists grow scarcer, families turn to telemedicine. Wall Street Journal Online. Available at: 0.SB ,00.html. 5. Cooper RA. There s a shortage of specialists: is anyone listening? Acad Med 2002;77: Glabman M. Too many patients, too few physicians. Manag Care 2001; 10:20-2, McClimon PJ, Hansen TN. Why are children s hospitals so busy? J Pediatr 2003;142: Tough SC, Newburn-Cook C, Johnston DW, Svenson LW, Rose S, Belik J. Delayed childbearing and its impact on population rate changes in lower birth weight, multiple birth, and preterm delivery. Pediatrics 2002;109: Forrest CB, Glade GB, Baker AE, Bocian AB, Kang M, Starfield B. The pediatric primary-specialty care interface: how pediatricians refer children 18 Freed et al The Journal of Pediatrics January 2005

67 and adolescents to specialty care. Arch Pediatr Adolesc Med 1999;153: Forrest CB, Majeed A, Weiner JP, Carroll K, Bindman AB. Comparison of specialty referral rates in the United Kingdom and the United States: retrospective cohort analysis. BMJ 2002;325: Kuhlthau K, Ferris TG, Beal AC, Gortmaker SL, Perrin JM. Who cares for Medicaid-enrolled children with chronic conditions? Pediatrics 2001;108: Cabana M, Bruckman D, Rushton JL, Bratton SL, Green L. Receipt of asthma subspecialty care by children in a managed care organization. Amb Pediatr 2002;2: National Heart, Lung, and Blood Institute. Clinical practice guidelines. Available at Accessed on Sept. 24, Forrest CB, Whelan EM. Primary care safety-net delivery sites in the United States: a comparison of community health centers, hospital outpatient departments, and physicians offices. JAMA 2000;284: Forrest CB, Reid RJ. Prevalence of health problems and primary care physicians specialty referral decisions. J Fam Pract 2001;50: Senf JH, Campos-Outcalt D, Kutob RM. Lessons not learned from the generalist initiatives. Acad Med 2002;77: Lurie JD, Goodman DC, Wennberg JE. Benchmarking the future generalist workforce. Eff Clin Pract 2002;5: Freed GL, Nahra TA, Wheeler JRC. Which physicians are providing health care to America s children? trends and changes over the last 20 years. Arch Pediatr Adol Med 2004;158: Norton S, Zuckerman S. Trends in Medicaid physician fees, Health Aff 2000;19: APPENDIX The members of The Research Advisory Committee of the American Board of Pediatrics: William F. Balistreri, MD, Cincinnati, Ohio Thomas F. Boat, MD, Cincinnati, Ohio Russell W. Chesney, MD, Memphis, Tenn Thomas P. Gessner, MD, Latrobe, Pa David M. Jaffe, MD, St. Louis, Mo Kevin B. Johnson, MD, Nashville, Tenn George Lister, MD, Dallas, Tex Julia A. McMillan, MD, Baltimore, Md Scott A. Shipman, MD, Portland, Ore Robert O. Guerin, PhD, Chapel Hill, NC Hazen P. Ham, PhD, Chapel Hill, NC Gail A. McGuinness, MD, Chapel Hill, NC Paul V. Miles, MD, Chapel Hill, NC James A. Stockman, III, MD, Chapel Hill, NC Michele J. Wall, Chapel Hill, NC Changes In The Proportion And Volume Of Care Provided To Children By Generalists And Subspecialists 19

68 COMPARING THE SUPPLY OF PEDIATRIC SUBSPECIALISTS AND CHILD NEUROLOGISTS RACHEL M. WERNER, MD, AND DANIEL POLSKY, PHD Objective To examine physician workforce characteristics and workforce projections of one vulnerable pediatric subspecialty, child neurology, and compare this subspecialty with other pediatric subspecialties, general pediatrics, and adult neurology. Study design National survey of child neurologists compared with other pediatric specialties in the Community Tracking Study Physician Survey (CTS) and the American Medical Association (AMA) Masterfile, and input-output workforce projections. Results Child neurologists are more likely than other specialists to report that the complexity or severity of patients conditions at the time of referral is less than it should be and that the number of patients being referred to them has increased. Projections of the future workforce reveal that by 2022 the number of pediatric subspecialists and pediatricians will increase substantially, to almost 180% and 150% of the current workforce, respectively. Among child neurologists and adult neurologists, practicing physicians will be only 109% and 105% of the current workforce by Conclusions Child neurologists are more likely to face future workforce shortages than other pediatric subspecialists. To reduce future shortages, recruitment efforts need to be concentrated on the at-risk subspecialties such as child neurology. In the short-term, addressing the high prevalence of inappropriate low-acuity referrals to child neurologists may ease the gap between supply and demand. (J Pediatr 2005;146:20-5) Despite numerous predictions of a surplus of physicians, 1 particularly of pediatricians, 2 there has been recent concern that the supply of pediatricians and pediatric subspecialists does not meet the demand. The American Academy of Pediatrics most recent pediatric workforce statement suggests that the need for pediatricians has increased and will continue to expand. 3 A survey of residency program directors found no evidence of an oversupply of pediatricians, with only 2% of recently trained pediatricians having difficulty finding positions. 4 Recently, the Future of Pediatric Education II Project reported that the supply of pediatric subspecialists will not meet the demand in the future. 5 Not all pediatric subspecialties seem equally susceptible to workforce shortages. A workforce survey of pediatric gastroenterologists predicted a significant surplus of gastroenterologists by the year Similarly, studies of pediatric cardiologists and neonatologists do not predict workforce shortages in their subspecialties. 7,8 However, other pediatric subspecialists have been singled out as being particularly susceptible to future workforce shortages. The Future of Pediatric Education II Project suggested that pediatric genetics, neurology, nephrology, and rheumatology may experience workforce shortages based on the excess number of faculty positions and the declining number of people planning to start careers in these subspecialties. 5 A survey found that pediatric geneticists, neurologists, and developmental-behaviorists were most likely to predict that they would need additional physicians in the future. 9 In 1998, a Workforce Task Force of the American Academy of Neurology reported a total of 1080 active child neurologists, which translated into 819 full-time equivalent patient-care child neurologists. 10 The Task Force found that staffing was 20% below the demand for child neurology services, a shortage projected to remain unchanged through AMA CNS American Medical Association Child Neurology Society CTS IMG Community Tracking Study Physician Survey International medical graduate See editorial, p 3, and related article, p 14. From the Division of General Internal Medicine; the Department of Health Care Systems, The Wharton School; and the Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia. This research has been supported by a grant from the Child Neurology Society. Submitted for publication Mar 24, 2004; last revision received Jun 24, 2004; accepted Aug 23, Reprint requests: Dr Rachel M. Werner, Division of General Internal Medicine, The University of Pennsylvania, 1208 Blockley Hall, 423 Guardian Drive, Philadelphia, PA rwerner@wharton. upenn.edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

69 We examined physician workforce characteristics and workforce projections of one vulnerable pediatric subspecialty, child neurology, and compared this subspecialty with other pediatric subspecialties, general pediatrics, and adult neurology. METHODS Data Source for Child Neurologists We surveyed all child neurologists in the United States in This population was identified as physicians who were members of the Child Neurology Society (CNS) (n = 1051) and non-cns member physicians in the American Medical Association (AMA) Masterfile who listed child neurology as their primary or secondary specialty (n = 433). In March of 2002 a survey was sent to all identified child neurologists (n = 1484). Three waves of surveys were mailed to increase the response rate. Of the 1484 surveys mailed, 52 were undeliverable because of incorrect addresses. We received surveys from 936 physicians, giving an overall response rate of 65%. Of the 936 completed surveys, 604 met the eligibility criteria for the study (65%). Exclusions included physicians who were practicing outside of the United States (2), were semi-retired or retired (20), were deceased (1), were not finished with residency or fellowship (30), worked <20 hours per week in patient care (120), did not spend the majority of their time in child neurology (149), and returned incomplete surveys (10). To estimate the current total number of active patientcare child neurologists from the survey, we add an estimate of the number of active patient-care child neurologists among the nonresponders to the 604 responders. To estimate the number of active patient-care child neurologists among the nonresponders, we estimated the likelihood of active status among the responders. We used this estimate to predict the likelihood of active status among the nonresponders using a logistic regression based on variables known for both responders and nonresponders (available on the AMA Masterfile.) These variables are international medical graduate (IMG) status, CNS membership, age, and whether child neurology was listed as a secondary specialty only. We determined 300 of the 548 nonresponders were active in patient care. The proportion active in patient care among the nonresponders (55%) was lower than among the responders (65%) because of the high prevalence of characteristics associated with ineligibility: IMGs, non-cns members, over the age of 65, and those with child neurology as a secondary specialty. Thus, we estimate that 904 child neurologists currently provide patient care in the United States. Data Sources for Comparison Specialties WORKFORCE CHARACTERISTICS. Responses to questions in the child neurology survey were compared with the responses of pediatric subspecialists, pediatricians, and adult neurologists to the same questions in the Community Tracking Study Physician Survey (CTS). The survey, designed to provide an in-depth look at the issues and challenges that physicians face, covers questions over a range of topics, including financial incentives, care management, and income, and it includes questions identical to those contained in the child neurology survey. The CTS was designed to be representative of physicians providing direct patient care in the United States, and it followed a complex design drawn from 60 sites as well as a small, independently drawn national sample. 11 The sample included active office- and hospital-based physicians, not employed by the federal government, who spent at least 20 hours per week in direct patient care in the continental United States. 12,13 The first two rounds of telephone interviews of physicians for the CTS were included in our analysis. The first round took place between August 1996 and August A sample of 23,096 physicians was obtained, of whom 18,947 were estimated to be eligible for the survey. Physicians interviewed by telephone numbered 12,385, representing a response rate of 65%. 12 The second round of CTS took place between August 1998 and November A sample of 25,627 physicians was obtained, of whom 20,205 were estimated to be eligible for the survey. For round 2 of the survey, 12,304 physicians completed the telephone interview, giving an overall response rate of 60.9%. 13 The findings reported in our study are based on all physicians in round 1 or round 2 whose primary specialty was defined as primary care pediatrician (n = 3232), adult neurologist (n = 251), or all other pediatric subspecialists (allergy and immunology, cardiology, critical care, endocrinology, gastroenterology, hematology and oncology, infectious disease, neonatology, pulmonology, or rheumatology; n = 280). Primary specialty was defined as the specialty where the physician spent the majority of time. Item nonresponse was very low in the CTS, typically less than 3%. Missing values for some of the variables were imputed. Additional information on the survey is available. 14,15 NUMBER OF ACTIVE PATIENT-CARE PHYSICIANS. Data on the number and characteristics of active patient-care pediatric subspecialists, pediatricians, and neurologists came from the December 2001 AMA Masterfile. The AMA Masterfile defines primary specialty as the specialty where the physician spends the majority of time. Physicians were excluded from the sample if their main practice activity was not patient care. Because retirement is not accurately recorded on the AMA Masterfile, 16 we code all physicians over 65 as retired and therefore not in patient care. We identified 6515 pediatric subspecialists, 34,839 pediatricians, and 7379 neurologists. The number of active patient-care child neurologists was estimated at 904 based on the child neurology survey, as outlined earlier. Because of concerns about the methodological differences between the AMA Masterfile and our survey of child neurologists, we identified child neurologists from the AMA Masterfile as well. We found 692 child neurologists in the AMA Masterfile using the criteria outlined earlier. Because the AMA Masterfile provided an underestimate of the number of child neurologists identified in our survey, we estimated the number of active patient-care child neurologists Comparing The Supply Of Pediatric Subspecialists And Child Neurologists 21

70 Table I. Characteristics of physician specialties Child neurologists n=604 Pediatric subspecialists n=280 Pediatricians n=3232 Adult neurologists n=251 Female (%) *** 13.6 *** White (%) *** 87.1 IMG (%) Age (mean) Years of experience (mean) * * Board Certified in subspecialty (%) * Income (mean) ** *** ** Hours per week on patient care (mean) *** *p<.05 compared to child neurologists. **p<.01 compared to child neurologists. ***p<.001 compared to child neurologists. to be 904 because in this case a larger number represents a more conservative estimate of the size of the workforce. Trends in number and characteristics of physicians entering each specialty were further characterized using data published annually on the characteristics of graduate medical education These data were used to examine the degree to which training programs in each specialty have complete enrollment and the percentage of IMGs enrolled in each program. Statistical Analyses First, we compared physician characteristics and attitudes toward their practice environment between specialty groups using responses to the child neurology survey and the CTS. We compared the physician characteristics of sex, race, IMG status, age, experience, board certification, income, and hours in patient care. For these descriptive analyses, x 2 tests were performed for categorical variables and one-way analysis of variance was performed for continuous variables. We hypothesized that child neurologists were more likely than other pediatric subspecialists and adult neurologists to feel that too many patients were being referred to them because of the limited supply of child neurologists, and that child neurologists would feel that the complexity or severity of patients conditions at the time of referral is less than it should be. To test this hypothesis, we compared responses to attitudinal questions about rates and types of referral. Participants in the child neurology survey and the CTS were asked to rate the complexity or severity of patients conditions at the time of referral overall, the complexity or severity of patients conditions during the past 2 years, and the number of patients referred to them during the past 2 years on a 5-point Likert scale (ie, In general, would you say that the complexity or severity of patients conditions at the time of referral to you by primary care physicians is: much greater than it should be, somewhat greater than it should be, about right, somewhat less than it should be, or much less than it should be ). For the descriptive analysis, these responses were collapsed into three categories (ie, greater than it should be, about right, or less than it should be ) and differences in responses across subspecialties were tested using x 2 tests. Second, we conducted an input-output analysis to project the number of physicians in active patient-care in each specialty group over the next 20 years. For this analysis we assumed that the rate of input was equal to the number of physicians who completed residency and who then became active in patient care in the United States. We used the average of the number of physicians completing residency annually between 1996 and 1998 in each specialty who were active in patient care in 2001 according to the AMA Masterfile (33 child neurologists, 405 pediatric subspecialists, 1686 pediatricians, and 247 neurologists). We display the change in the number of physicians in each specialty by indexing the supply at 1.0 as of 2002 and showing the change in that index over the next 20 years. RESULTS Workforce Characteristics and Environment Overall, practicing child neurologists in the United States were less diverse than pediatricians, with a higher proportion of men and white physicians (P <.0001) but were similar in demographics to other pediatric subspecialists (Table I). Child neurologists had a similar proportion of IMGs and were of similar age compared with the other specialties. Child neurologists had 2 more years of experience than other pediatric subspecialists (P =.02) and adult neurologists (P =.03). The average income of child neurologists was close to $18,000 lower than that of other pediatric subspecialists (P =.01) and over $21,000 lower than that of adult neurologists (P =.006). On average, child neurologists spent similar hours per week on patient care compared with other pediatric subspecialists and pediatricians but significantly less than adult neurologists (P <.001). Child neurologists were much more likely to report having adequate time to spend with their patients than the other specialties (P <.001 for all comparisons) (Table II). Over three times as many child neurologists felt that in general the complexity or severity of patients conditions at the 22 Werner and Polsky The Journal of Pediatrics January 2005

71 Table II. Career satisfaction and scope of care among specialists (%) Child neurologists n=604 Pediatric subspecialists n=280 Pediatricians n=3232 Adult neurologists n=251 I spend adequte time with my patients Agree *** 67.3 *** 75.7 *** Neutral Disagree In general, the complexity or severity of patients conditions at the time of referral is: Greater than it should be *** y 34.7 *** About right y 50.0 Less than it should be y 7.4 In the last 2 years, the number of patients referred to you has: Increased *** y 38.2 *** Stayed the same y 31.7 Decreased y 30.1 ***p<.001 compared to child neurologists. yquestion not asked to primary care physicians. time of referral was less than it should be compared with the other specialties (Table II). Of child neurologists, 24.5% reported that complexity or severity of referrals was less than it should be compared with 7.4% of pediatric subspecialists and 7.4% of neurologists (P <.001 for both comparisons). A larger proportion of child neurologists also felt that the number of patients referred to them had increased in the last 2 years compared with the other specialists (65% of child neurologists vs 45.6% of pediatric subspecialists vs 38.2% of adult neurologists; P <.001 for both comparisons). Number of Physicians Entering the Workforce Trends in graduate medical education suggest that there are declining numbers of physicians entering child neurology, pediatric subspecialties, and adult neurology. Child neurology experienced the largest decline in the proportion of residency slots that were filled. A decade ago close to 70% of residency slots for child neurology were filled. In 2001, only 55% of residency slots were filled, despite a relatively steady number of positions being offered (Figure 1A) Pediatric subspecialists and adult neurologists have followed a similar downward trend in the proportion of filled residency slots. However, both have remained relatively high compared with child neurology, at 73% and 79% respectively. The proportion of residency slots that are being filled by non-u.s. medical graduates has been increasing in all pediatric subspecialties, child neurology, and adult neurology over the past decade (to 37%, 50%, and 40%, respectively). At the same time, the number of non-u.s. medical graduates entering pediatric residency programs has declined to 19% (Figure 1B). Workforce Projections Projections of the number of physicians entering specialties show stable growth in pediatric subspecialists and pediatrics (Figure 2). Over the next 20 years, the number of practicing pediatric subspecialists will increase substantially, growing to almost 180% of the current supply. Pediatricians as a group also will experience growth, reaching almost 150% of the current supply in This contrasts sharply with the projected growth in child neurology and neurology. Among child neurologists and adult neurologists, the number of practicing physicians initially grows a small amount, peaking in 2011 at 112% of the current workforce. However, both specialties will then experience a decline in their workforce. By 2022 there will be only 9% more child neurologists than the number that exist today and only 5% more adult neurologists. DISCUSSION Although there has been overall consensus that pediatric subspecialists will experience a workforce shortage in the future, we are aware of no prior research that has compared workforce projections across subspecialties. Our research suggests that not all pediatric subspecialties will experience a declining workforce. As a group, the workforce of pediatric subspecialists and pediatricians will grow substantially over the next 20 years. However, our research suggests that child neurology will not experience the same growth as the entire cohort of pediatric subspecialties. The number of child neurologists is expected to increase only a small amount. Our research also suggests some clues as to why child neurology differs from other subspecialties. With respect to income and hours worked, child neurologists have a lower income than other comparably trained subspecialists, yet they do not work fewer hours. Effectively this translates to less income per hour worked. Although we were unable to compare rates of reimbursement as a potential source of this wage difference, a different potential source did emerge. Child neurologists are more likely to report spending adequate time Comparing The Supply Of Pediatric Subspecialists And Child Neurologists 23

72 Figure 2. Projections of the number of physicians active in patient care in each specialty: child neurology, pediatric subspecialty, pediatrics, and adult neurology. Each specialty s workforce is indexed to the size of their workforce in 2002, showing relative changes in workforce size over the next 20 years. Figure 1. (A) Percentage of residency slots filled between 1992 and 2002 among child neurology, pediatric subspecialty, and pediatric and adult neurology training programs. (B) Percentage of IMGs in residency programs between 1992 and 2002 among child neurology, pediatric subspecialty, and pediatric and adult neurology training programs. with their patients. Given that reimbursement is typically tied to the number of office visits and not adjusted for the length of visit, lower incomes may be related to longer visits. The ability to maintain adequate time with their patients may change given that almost two-thirds of child neurologists feel that the number of patients being referred to them has increased, whereas less than half of other pediatric subspecialists report such an increase. If referrals to child neurologists are increasing at a faster rate than in other pediatric subspecialties and the number of physicians choosing the specialty is not growing, the burden on existing child neurologists will increase. This will cause the number of hours worked to increase, the time with each patient to decrease, and possibly more job positions for child neurologists to become available or remain open. If the labor market responds to this pressure, the income of child neurologists should increase to reflect the increased work. Given the high debt burden faced by many medical students, if child neurology offered higher incomes, it may attract prospective medical students and young pediatricians to enter the specialty Curiously, however, this upward pressure on income and number of hours worked and the downward pressure on time with patients is not evident now. Given that reimbursement rates for office visits often are not negotiated separately for each subspecialty, the market may not be responding to changes in supply and demand occurring in a particular subspecialty. As a result, projections of shortfalls require action rather than waiting for market forces to correct the shortfall. We also found that child neurologists were more likely than other pediatric subspecialists to feel that the complexity or severity of patients conditions at the time of referral is less than it should be. A possible explanation for the difference in the appropriateness of referrals to child neurologists compared with other pediatric subspecialists may be related to the distinct relationship the specialty of child neurology has with the specialty of pediatrics. Most departments of child neurology are in departments of adult neurology rather than pediatrics. Similarly, unlike most other pediatric subspecialties, board certification in child neurology is not offered by the American Board of Pediatrics. Rather, board certification is offered by neurology. 5 Finally, most people who enter the field of child neurology choose to do so during medical school, rather than during pediatric residency. 30 This may cause the coordination between child neurologists and pediatricians to be below that of other pediatric subspecialties. As a result, referral patters to child neurology may differ from other pediatric subspecialties, with more patients with low-complexity conditions being managed by child neurologists rather than by pediatricians. The increased market power of health maintenance organizations might also be related to child neurologists seeing more patients and patients with inappropriately low-acuity conditions. Health maintenance organizations have been associated with physicians feeling more pressured for time. 31 Primary care physicians who once cared for patients with complex conditions who require longer visits might be increasingly referring such patients to specialists in an attempt to decompress their schedule for routine patients. This would increase the number of referrals to child neurologists, predominantly through more referrals of patients with lowcomplexity conditions. Several limitations should be considered in interpreting our findings. First, we compare responses of two different surveys in this study, the child neurology survey and the CTS. These two surveys had different sampling methods and also 24 Werner and Polsky The Journal of Pediatrics January 2005

73 were conducted in different time periods the child neurology survey took place in 2002 and the CTS spanned 1996 to Although this increases the possibility of measurement error, other studies have documented that responses to job satisfaction surveys during this time period remained stable. 32 Additionally, the question on career satisfaction was worded differently between the two surveys, with the child neurology survey asking about overall career satisfaction and the CTS asking about current career satisfaction. This makes comparisons between groups subject to error. A second limitation to our comparisons is that we compare one pediatric subspecialty, child neurology, with all other pediatric subspecialties. It is possible that pediatric subspecialties differ from each other in important ways and, by averaging the responses, differences between the specialties are lost. It is possible that other subspecialties that have been identified as facing future labor shortages would closely resemble child neurology. Because we are limited by the sample size of pediatric subspecialists included in the CTS, we are unable to do further analyses. By comparing responses of child neurologists with those of both pediatric subspecialists and adult neurologists, we attempt to capture the important comparison groups for child neurologists. Finally, we do not predict demand for physician services, and our workforce predictions could be inaccurate if the demand for physician services changes differentially between subspecialties over the next two decades. However, based on other demand predictions 3,5 and an increase in the pediatric population predicted by the U.S. Census, it is reasonable to assume that the demand for all subspecialty services will increase with time. In this setting, a commensurate increase in the child neurology workforce would be needed. The future of pediatric subspecialties depends on attracting new physicians to the field. The good news is that pediatric subspecialties in general can expect a robust growth in the number of physicians entering their fields, alleviating concerns about future workforce shortages. The bad new is that some pediatric subspecialists, such as child neurologists, appear not to benefit from the growth in other fields. Research on the factors that attract medical students to a given subspecialty is only beginning. 30 These factors need to be further investigated and efforts to attract physicians to pediatric subspecialties need to be concentrated in the subspecialties that expect to face workforce shortages, such as child neurology. In the short-term, however, addressing the high prevalence of inappropriate low-acuity referrals to child neurologists may ease the gap between supply and demand. REFERENCES 1. Graduate Medical Education Advisory Committee. Summary report of the Graduate Medical Education National Advisory Committee to the secretary, Department of Health and Human Services. Rockville, Md: US Department of Health and Human Services, Public Health Service, Health Resources and Services Administration; American Academy of Pediatrics. Pediatric manpower recommendations. Elk Grove Village, Ill: American Academy of Pediatrics; American Academy of Pediatrics. Pediatric workforce statement. Pediatrics 1998;102: Brotherton SE. Career plans of new pediatricians: results from a survey of residency program directors. Pediatrics 1991;88: Gruskin A, Williams RG, McCabe ERB, Stein F, Strickler J. Final report of the FOPE II pediatric subspecialists of the future workgroup. Pediatrics 2000;106: Colletti RB, Winter HS, Sokol RJ, Suchy FJ, Klish WJ, Durie PR. Medical position paper: a position paper of the North American Society for Pediatric Gastroenterology and Nutrition: pediatric gastroenterology workforce survey and future supply and demand. J Pediatr Gastroenterol Nutr 1998;26: Williams RG, Kennedy TL, Moller JH. Pediatric cardiology in the 1990s. J Am Coll Cardiol 1994;23: Pollack LD, Ratner IM, Lund GC. United States neonatology practice survey: personnel, practice, hospital and neonatal intensive care unit characteristics. Pediatrics 1998;101: Stoddard JJ, Cull WL, Jewett EAB, Brotherton SE, Mulvey HJ, Alden ER. Providing pediatric subspecialty care: a workforce analysis. Pediatrics 2000;106: Bradley WG. Neurology in the next two decades. Neurology 2000;54: Metcalf CE, Kemper P, Kohn LT, Pickreign JD. Site definition and sample design for the Community Tracking Study. Washington, DC: Center for Studying Health System Change; Technical publication Keil L, Chattopadhyay F, Potter F, Reed MC. Community Tracking Study physician survey, round 1. Survey methodology report. Washington, DC: Center for Studying Health System Change; Technical publication Potter F, Strouse R, Sinclair M, Williams S, Ellrich M, Tourangeau R. Report on survey methods for the community tracking study s round 2 physician survey. Washington, DC: Center for Studying Health System Change; Technical publication Anonymous. Community Tracking Study physicians survey restricted use file: user s guide. Round 2, release 1. Washington, DC: Center for Studying Health System Change; Technical publication Anonymous. Community Tracking Study physician survey restricted use file: user s guide. Round 1, release 2. Washington, DC: Center for Studying Health System Change; Technical publication Kletke PR, Polsky D, Wazniak GD, Escarce JJ. The effect of HMO penetration on physician retirement. Health Serv Res [serial online] 2004 [cited ]; 35: Available from: URL: hospitalconnect.com/hsr/alicehersh/04-kle1.pdf. 17. Anonymous. Graduate medical education. JAMA 1993;270: Anonymous. Graduate medical education. JAMA 1994;272: Anonymous. Graduate medical education. JAMA 1995;274: Anonymous. Graduate medical education. JAMA 1996;276: Anonymous. Graduate medical education. JAMA 1997;278: Anonymous. Graduate medical education. JAMA 1998;280: Anonymous. Graduate medical education. JAMA 1999;282: Anonymous. Graduate medical education. JAMA 2000;284: Anonymous. Graduate medical education. JAMA 2001;286: Anonymous. Graduate medical education. JAMA 2002;288: Thorton J, Esposto F. How important are economic factors in choice of medical specialty? Health Econ 2003;12: Thorton J. Physician choice of medical specialty: do economic incentives matter? Appl Econ 2000;32: Fox M. Medical student indebtedness and the propensity to enter academic medicine. Health Econ 2003;12: Polsky D, Werner RM. The future of child neurology: a profile of child neurology residents. J Child Neurol 2004;19: Hadley J, Mitchell JM, Sulmasy DP, Bloche MG. Perceived financial incentives, HMO market penetration, and physicians practice styles and satisfaction. Health Serv Res 1999;34: Landon BE, Reschovsky J, Blumenthal D. Changes in career satisfaction among primary care and specialist physicians, JAMA 2003;289: Comparing The Supply Of Pediatric Subspecialists And Child Neurologists 25

74 CRITERIA FOR IDENTIFICATION OF COMPREHENSIVE PEDIATRIC HOSPITALS AND REFERRAL REGIONS ROBERT K. KANTER, MD, AND FRANKLIN DEXTER, MD, PHD Objective To identify comprehensive pediatric hospitals on the basis of publicly available data. Study design We developed identification criteria for comprehensive pediatric hospitals, then evaluated the number of hospitals meeting these selection criteria. Criteria for a comprehensive pediatric hospital included pediatric residency accreditation, pediatric inpatient volume, and diversity of pediatric disorders at each hospital. New York State hospital administrative discharge data were analyzed for patients 0 to 14 years of age, excluding neonatal diagnoses. Results Infants and children (n = 125,588) with 375 different diagnosis-related groups were discharged from 230 hospitals in Through the use of higher selective criteria (educational accreditation plus both high volume and diversity in the top decile), 11 comprehensive pediatric hospitals were identified. These hospitals serve populations of 1.7 ± 0.3 million (mean ± SD) each, with 8 referral regions throughout the state, collectively providing care for 29% of all pediatric statewide hospitalizations. Conclusions Comprehensive pediatric hospitals serve the population of New York widely and evenly. The ability to identify pediatric hospitals will permit evaluation of the relative quality of care and suggest appropriate regulatory interventions to improve pediatric hospital utilization. (J Pediatr 2005;146:26-9) Optimal care of children with severe illness and injury requires appropriate organization of hospital resources. Care of the most severe and complex problems is best provided at comprehensive pediatric hospitals. 1,2 Referral to a limited number of designated comprehensive hospitals would provide sufficient clinical volume to achieve proficiency in care of high-risk conditions, 3 to generate sufficient revenue to support costly services, 4 and to avoid redundancy of these resources. 5 The lack of an objective method to identify comprehensive pediatric hospitals results in an incomplete understanding of the existing system and may interfere with the use of existing resources. 6 Pediatric hospitals identified by national organizations 7,8 are based on self-report. Inclusion on such lists may require a membership fee and may reflect hospitals marketing interests. Lacking objective definitions, marketing claims about what constitutes a real children s hospital can deteriorate into legal disputes. 9 With pediatric hospitals lacking formal status, business interests may threaten patient access to unique regional resources. 10 Federal goals for national health care improvement include an increase in the number of states with guidelines that categorize acute care facilities with respect to pediatric personnel and resources. 11 One study indicated that accreditation of hospitals for pediatric graduate medical education purposes identified facilities that each tend to care for larger numbers of children than nonaccredited hospitals. 12 However, the numerous hospitals identified in this way exceed the local need for comprehensive pediatric hospitals in some regions. We carried out the current study to develop a research method to identify comprehensive pediatric hospitals based on publicly available data rather than by self-report. METHODS Characteristics of Hospitals Each hospital s pediatric inpatient volume was expressed as the annual number of hospital discharges. Diversity of disorders at each hospital was expressed as the probability DRG PHRR Diagnosis-related group Pediatric hospital referral region SPARCS Statewide Planning and Research Cooperative System From the Department of Pediatrics, Upstate Medical University, Syracuse, New York; and the Department of Anesthesia, University of Iowa, Iowa City, Iowa. Submitted for publication Dec 16, 2003; last revision received May 12, 2004; accepted Aug 24, Reprint requests: Robert K. Kanter, MD, Department of Pediatrics, Upstate Medical University, 750 East Adams St, Syracuse, NY kanterr@upstate.edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

75 that two patients selected at random will have the same federal diagnosis-related group (DRG) and was calculated as the sum of squares of proportions of each DRG. 13 Diversity increases as the probability decreases from a maximum of 1 to 0. Educational accreditation of hospitals was defined as sponsorship or major participation in a pediatric residency. 14 Patients Data were analyzed for patients 14 years of age and younger. Older patients were excluded because data indicating the survival benefit of pediatric hospital care pertain to younger ages. 15,16 Patients with neonatal DRGs were excluded from the study because care of newborn infants is often provided in separate facilities and may be regulated independently from other pediatric care. Data Data were obtained from the New York Statewide Planning and Research Cooperative System (SPARCS 17 ) for The data are collected by hospitals for administrative purposes and reported to the state. The data are publicly available, with no individual patient identifiers. Hospitals are required to report the information, and the data are audited according to New York State law. This analysis of publicly available data was considered exempt from institutional review board review. Analysis Hospitals reporting any pediatric discharges were studied. Diversity of disorders was evaluated for its consistency with other markers of comprehensive pediatric hospitals. Correlation of high diversity with high clinical volume was expressed as the Spearman correlation coefficient. The probability that the slope of the line for this relation differed from 0 was determined. The diversity of disorders was compared for hospitals with and without educational accreditation by the Mann-Whitney test (P <.05 was considered to represent a statistically significant difference). Criteria were considered for identification of comprehensive pediatric hospitals. The criteria included educational accreditation, clinical volume, and diversity of disorders. Residency accreditation is the only existing tool to evaluate hospitals according to uniform national standards for a range of medical and surgical services. 6 Volume is an indirect marker consistent with quality. 3 Diversity distinguishes comprehensive from specialty hospitals. 13 Numbers of hospitals meeting percentile levels of selectivity for combinations of criteria were determined. Pediatric hospital referral regions (PHRRs) were determined on the basis of patient referral patterns identified in this study. The PHRRs were formulated after identifying comprehensive pediatric hospitals by a single set of selective criteria. The county of residency was determined for each child hospitalized at a comprehensive pediatric hospital. For each county, we identified the comprehensive pediatric hospital serving the largest number of children residing in that county. Counties referring patients to a single comprehensive hospital were considered to form a PHRR. If more than one comprehensive hospital was located in a single county, all the counties referring patients to those comprehensive hospitals were considered to form a PHRR. If a comprehensive pediatric hospital did not attract the largest number of patients from within its own county, then that county was included in a PHRR with the other county serving the largest number of its patients. RESULTS A total of 125,588 infants and children with 375 different DRGs were hospitalized at 230 hospitals, statewide in Greater diversity of disorders was correlated with greater numbers of hospital discharges (Figure 1; Spearman r = 20.61, P <.001). The diversity of disorders was greater at hospitals with accredited pediatric residencies than at nonaccredited institutions (Figure 1; median = 0.11 vs 0.58, respectively, P <.05). Among all hospitals, the annual number of pediatric discharges was 1715, 469, and 172 for institutions at the 90 th, 75 th, and 50 th percentiles, respectively. Diversity of diagnoses (the probability that two patients had the same DRG) was 0.053, 0.073, and 0.105, corresponding to 158, 75, and 38 different DRGs per hospital, for institutions at the 90 th,75 th, and 50 th percentiles, respectively. The numbers of hospitals satisfying both volume and diversity criteria at 50 th,75 th, and 90 th percentile levels are shown in Figure 2. Forty-two hospitals had educational accreditation. Few hospitals that lack residency accreditation satisfied high volume or diversity criteria, and none met both volume and diversity criteria in the top decile. Eleven hospitals with residency educational accreditation have both volume and diversity in the top decile. These are considered to represent the comprehensive pediatric hospitals for purposes of describing existing regional services. The 11 comprehensive pediatric hospitals identified in this study provided care for 36,119 infants and children (29% of pediatric hospitalizations) in These hospitals were widely distributed throughout the state. The existing patterns of referrals of children to comprehensive hospitals formed 8 PHRRs (Figure 3). Each hospital served populations of 1.7 ± 0.3 million (mean ± SD). Each region was served by at least one pediatric hospital, with a maximum of three in the New York City region. One percent of hospitalized children resided outside the state. DISCUSSION Proposed selective criteria for identification of comprehensive pediatric hospitals include educational accreditation and both pediatric volume and diagnostic diversity in the top decile. These criteria identified existing facilities that are widely and evenly distributed in relation to the statewide population while eliminating redundancy in any region. Validity of the proposed criteria can be evaluated in the following ways. Content validity includes face validity (the subjective impression that criteria identify comprehensive Criteria For Identification Of Comprehensive Pediatric Hospitals And Referral Regions 27

76 Figure 1. Relation between diversity of disorders and number of discharges at each hospital. Diversity is the probability that two patients chosen at random have the same DRG (= sum of squares of proportions of each DRG). Diversity increases as probability decreases from 1 to 0. Hospitals with (:) and without (s) pediatric residency educational accreditation are indicated. For this relation, Spearman r = 20.61, P <.001. Diversity of disorders was greater at accredited hospitals (P <.05). (Figure 1 is available in color online at Figure 2. Number of hospitals in New York State meeting both volume and diversity criteria at indicated percentile levels. Results are shown separately for hospitals with (:) and without (s) residency accreditation. (Figure 2 is available in color online at elsevierhealth.com/jpeds) pediatric hospitals in a reasonable way) and sampling validity (the inclusion of pertinent aspects of the phenomenon being measured). These are both achieved by the study criteria; all 11 hospitals identified are large teaching hospitals with diverse patient populations. It might be argued that some large teaching hospitals were excluded unreasonably. Sampling validity would be stronger if hospital characteristics and personnel necessary for comprehensive services could be better identified from publicly available data. Construct validity involves observations that are consistent with theoretical concepts about the phenomenon being studied. Construct validity is supported by the homogeneity of regional population size served by each of the 11 hospitals (coefficient of variation for regional population served per comprehensive hospital = 17%). Construct validity is also supported by the similarity of regional populations per hospital determined in this study and the population previously found to be necessary to financially support pediatric hospitals. 4 Concurrent validity is the extent to which a measure is consistent with an independent measure of the same phenomenon. The lack of a gold standard to identify pediatric hospitals is the reason for undertaking this study. Among the 11 comprehensive pediatric hospitals identified in this study, 9 (Reference 7) or 11 (Reference 6) were independently identified as having a pediatric intensive care unit. One (Reference 7) or 4 (Reference 8) were independently identified as children s hospitals. Thus, other published identification lists of pediatric hospitals may be incomplete or may involve even more selective criteria than ours. New York State only had one free-standing children s hospital exclusively caring for Figure 3. Map indicates counties (light lines), pediatric hospital referral regions (bold lines), and comprehensive pediatric hospitals (d) by selective criteria. (Figure 3 is available in color online at children. Although the 11 identified hospitals are the busiest pediatric facilities in the state, children only accounted for an average of 16% of their discharges, the remainder being adult medical, surgical, obstetrics, psychiatry, and neonatal patients. Thus, exclusive care of children cannot be used as a criterion for pediatric hospital identification in New York. The single free-standing children s hospital in the state was identified by our criteria as one of the 11 comprehensive pediatric hospitals. Empirical investigation will be needed to determine how well the proposed method to identify comprehensive pediatric hospitals can be generalized to other states as a health services 28 Kanter and Dexter The Journal of Pediatrics January 2005

77 research tool. New York has a uniquely large population, with a disproportionately large number of medical schools, only one free-standing children s hospital, and no for-profit hospitals. Some states cannot satisfy the requirement for an accredited pediatric residency. Rural areas with extremely large distances are lacking in this state sample. Regulatory environments may vary among states. In other respects, New York shares features representative of other states, with a range of large metropolitan, small urban, suburban, and rural areas. The following limitations of the study are recognized. The study analyzed data collected for administrative rather than clinical purposes. The quality of hospital care was not assessed. Specific hospital resources, services, and personnel cannot be identified in hospital discharge databases. Nevertheless, high volume is associated with high quality, and high volume may serve as an indirect marker of hospitals with high capabilities until better measures of quality are available. 18 No SPARCS data are available regarding children residing in New York who are admitted to hospitals outside the state. PHRRs, determined from patterns of referral to the subgroup of facilities identified as comprehensive pediatric hospitals, are defined functionally as those geographic subunits surrounding a hospital used most by the population in those areas. 19 Previous descriptions of hospital referral regions have been related to complete samples of hospitals and the small areas (towns or zip codes) from which they draw patients. 19,20 For some marketing research purposes, competition in overlapping small areas has been a focus of study. 21 Our method differs from previous reports in that only a subset of hospitals is considered in formulating regions and counties rather than smaller areas are considered the smallest unit of analysis. Geographic units of analysis smaller than counties would result in many unclassifiable areas with no patients referred to a comprehensive hospital. Referral regions for the subset of comprehensive hospitals, as determined in this study, serve larger populations and areas than the referral regions determined for each facility in a complete sample of all hospitals. The PHRRs determined in this study are generally similar to health service areas defined by the State of New York without regard to age (compared with SPARCS health service areas, 17 only 8 of 62 counties are rearranged to describe PHRRs, with the same total number of statewide regions). This study proposes the identification of a subset of existing hospitals as comprehensive regional pediatric resources, based on simple and selective administrative criteria. Further investigation to evaluate the proposed criteria in a broader national sample is warranted. The ability to identify comprehensive pediatric hospitals will permit evaluation of the relative quality of care at such facilities and might suggest regulatory interventions to improve their appropriate utilization. REFERENCES 1. American Academy of Pediatrics, Committee on Pediatric Emergency Medicine, American College of Critical Care Medicine, Society of Critical Care Medicine, Pediatric Section. Consensus report for regionalization of services for critically ill or injured children. Pediatrics 2000;105: American Academy of Pediatrics, Committee on Pediatric Emergency Medicine. Guidelines for pediatric emergency care facilities. Pediatrics 1995; 96: Birkmeyer JD, Siewers AE, Finlayson EVA, Stukel TA, Lucas FL, Batista I, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346: Billi JE, Wise CG, Bills EA, Mitchell RL. Potential effects of managed care on specialty practice at a university medical center. N Engl J Med 1995; 333: Richardson DK, Reed K, Cutler C, Boardman RC, Goodman K, Moynihan T, et al. Perinatal regionalization versus hospital competition: the Hartford example. Pediatrics 1995;96: Kanter RK. Regional variation in child mortality at hospitals lacking a pediatric intensive care unit. Crit Care Med 2002;30: American Hospital Association. AHA Guide , Health Forum, One North Franklin, Chicago, Ill, National Association for Children s Hospitals and Related Interests. Children s Hospital Profiles. Alexandria, Va, Internet access: 10/21/03. shospital.net/index.htm. 9. Bellandi D. Stand-alone vs unit: a battle is on over what constitutes a children s hospital. Mod Healthcare 1997;27: Becker C. Antitrust action averted: agreement with Pennsylvania attorney general allows UPMC merger to proceed. Mod Healthcare 2001; 31: US Department of Health and Human Services. Healthy People 2010: understanding and improving health. 2nd edition. Washington, DC; US Government Printing Office, November Goal 1-14b. p to Kanter RK, Egan M. Utilization of pediatric hospitals in New York State. Pediatrics 2003;111: Dayhoff DA, Cromwell J. Measuring differences and similarities in hospital caseloads: a conceptual and empirical analysis. Health Serv Res 1993; 28: American Medical Association. Graduate Medical Education Directory, Chicago, Ill: American Medical Association. 15. Hannan EL, Racz M, Kavey RE, Quaegebeur JM, Williams R. Pediatric cardiac surgery: the effect of hospital and surgeon volume on inhospital mortality. Pediatrics 1998;101: Tilford JM, Simpson PM, Green JW, Lensing S, Fiser DH. Volume outcome relationship in pediatric intensive care units. Pediatrics 2000;106: New York State Department of Health. New York Statewide Planning and Research Cooperative System, Albany, Epstein AM. Volume and outcome: it is time to move ahead. N Engl J Med 2002;346: Wennberg J, Gittelsohn A. Small area variations in health care delivery. Science 1973;182: Guagliardo MF, Jablonski KA, Joseph JG, Goodman DC. Do pediatric hospitalizations have a unique geography? BMC Health Serv Res 2004;4: Sohn MW. A relational approach to measuring competition among hospitals. Health Serv Res 2002;37: Criteria For Identification Of Comprehensive Pediatric Hospitals And Referral Regions 29

78 PERSISTENT HYPERGLYCEMIA IN CRITICALLY ILL CHILDREN EDWARD VINCENT FAUSTINO, MD, AND MICHAEL APKON, MD, PHD, MBA Objectives To determine the prevalence and prognostic significance of hyperglycemia among critically ill nondiabetic children. Study design We performed a retrospective cohort study using point-of-care blood glucose measurements, hospital administrative databases, and a computerized information system; 942 nondiabetic patients admitted to our Pediatric Intensive Care Unit (PICU) from October 2000 to September 2003 were included. The prevalence of hyperglycemia was based on initial PICU glucose measurement, highest value within 24 hours, and highest value measured during PICU stay up to 10 days after the first measurement. Primary outcome was in-hospital death with PICU lengths of stay (LOS) as secondary outcome. Results Through the use of three cutoff values (120 mg/dl, 150 mg/dl, and 200 mg/dl), the prevalence of hyperglycemia was 16.7% to 75.0%. The relative risk (RR) for dying increased for maximum glucose within 24 hours >150 mg/dl (RR, 2.50; 95% confidence interval (CI), 1.26 to 4.93) and highest glucose within 10 days >120 mg/dl (RR, 5.68; 95% CI, 1.38 to 23.47). LOS was decreased for admission glucose >120 mg/dl and 150 mg/dl but increased for all threshold values for maximum glucose within 10 days. Conclusions Hyperglycemia occurs frequently among critically ill nondiabetic children and is correlated with a greater in-hospital mortality rate and longer LOS. (J Pediatr 2005;146:30-4) Hyperglycemia occurs frequently among critically ill adults, with prevalence rates reported from 3% to 71%. 1 During the acutely stressed state, hyperglycemia is thought to be advantageous, 2 providing the glucose-dependent organs such as the brain and blood cells adequate supply for their energy needs. 3,4 Hyperglycemia has also been postulated to compensate for volume loss by promoting the movement of cellular fluid into the intravascular compartment or liberating water bound to glycogen. 4 Despite potential positive effects, prolonged hyperglycemia in critically ill adults has been shown to be associated with a number of deleterious consequences 5 contributing to greater risks of morbidity and mortality, even in the absence of preexisting diabetes mellitus. 1,6,7 Elevated glucose concentrations have been associated with increased risks of congestive heart failure, 1 cardiogenic shock, 1 and poor functional recovery after stroke 6 as well as increased risks of dying after myocardial infarction 1 and ischemic stroke 6 among nondiabetic patients. Even among non critically ill adult patients admitted to general patient care units, patients with newly diagnosed hyperglycemia had a significantly higher mortality rate and a lower functional outcome compared with known diabetic patients or normoglycemic patients. 7 Hyperglycemia may be less prevalent among children because diabetes mellitus is much less common in this age group. 8,9 Furthermore, hyperglycemia in the pediatric population may have different effects on morbidity and mortality compared with adults as a consequence of different metabolic demands, 10 differences in comorbid conditions, 11 or age-dependent factors. 12 It is not known how often hyperglycemia occurs in critically ill children or what effect hyperglycemia has on outcome in younger patients. The objective of this study is to determine the prevalence of hyperglycemia among critically ill children without diabetes mellitus. Moreover, it aims to correlate hyperglycemia with death and length of stay among survivors in the pediatric intensive care unit. METHODS We conducted a retrospective cohort study by using data collected in the normal course of patient care. The study was reviewed and approved by our institution s Human Investigation Committee, which waived the need to obtain informed consent. The study examined patients cared for in the 11-bed multidisciplinary Pediatric Intensive Care Unit (PICU) within a not-for-profit private academic teaching hospital. DRG ICU Diagnosis-related group Intensive Care Unit IQR PICU Interquartile range Pediatric Intensive Care Unit See editorial, p 5. From the Section of Critical Care and Applied Physiology, Department of Pediatrics, Yale University School of Medicine, New Haven, Connecticut. Submitted for publication May 21, 2004; last revision received Jul 14, 2004; accepted Aug 27, Reprint requests: Dr Michael Apkon, Department of Pediatrics, Yale University School of Medicine, 333 Cedar St, PO Box , New Haven, CT Michael.Apkon@ yale.edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

79 Subjects All patients admitted to the PICU from October 2000 to September 2003 were eligible to be included in the study if they had at least one blood glucose concentration measured by point-of-care testing in the PICU. Patients were excluded if their diagnosis-related group (DRG) assigned at discharge corresponded to diabetes mellitus (DRG 295). Patients admitted in December 2000 were excluded because of missing glucose measurements. Sources of Data Data were obtained from three different electronic databases linked together by means of the patients unique medical record numbers. The date, time, and value for each blood glucose measurement were recorded in a computerized log maintained as part of the laboratory quality assurance process for point-of-care testing. All glucose measurements were obtained by use of the SureStep Flexx Meter (LifeScan, Inc, Milpitas, Calif), which records patient-identifying information with each measurement. The hospital administrative cost-accounting database contained information describing all patient admissions, including the dates of admission and discharge, the number of days spent hospitalized on each patient care unit, and the DRG code assigned by certified medical coders. Information describing patient demographics and outcomes were recorded in the PICU clinical information system (Clinfosys). 13 Data tables were exported from each of these data systems, imported into a relational database (Microsoft Access XP, Microsoft Corp, Redmond, Wash), and linked by means of the patients unique medical record numbers for analysis. Measurements and Analysis The independent variables for the study were initial glucose, maximum glucose within 24 hours of the first value, and the highest glucose observed during the PICU stay up to the 10th day from the initial measurement. The primary outcome measure was in-hospital death. The secondary outcome was PICU length of stay among survivors without tracheostomy. Patients with tracheostomy (DRG codes 482 and 483) were excluded in the computation for PICU length of stay because these patients may stay longer than medically necessary in the PICU as the result of an inability to establish sufficient resources for care in the community. The prevalence of hyperglycemia was computed on the basis of initial blood glucose and the highest glucose within 24 hours and within 10 days of the first value. To facilitate comparison with previous studies, 7,14,15 three cutoff values were used: 120 mg/dl, 150 mg/dl, and 200 mg/dl. Relative risks with 95% confidence intervals (CI) 16 were computed to compare the mortality rates above and below these three cutoff values. Among the survivors, the Mann-Whitney test was used to compare the lengths of PICU stay above and below the hyperglycemic thresholds. A value of P <.05 was considered significant for all analyses. Glucose values are summarized by mean ± SD. Because of the skewness of the data, the lengths of stay are presented by using mean ± SD together with the median and interquartile range (IQR, 1 st to 3 rd quartile). RESULTS During the study period, 1927 patients were admitted to the PICU; 44 patients had a diagnosis of diabetes mellitus. Of the patients without diabetes mellitus, 942 had at least one glucose measurement. There was a median of 2 glucose measurements per patient (IQR, 1 to 3 measurements) for the first 24 hours and a median of 1 glucose measurement per patient per day (IQR, 0.7 to 2 measurements) for the next 10 days after the first determination. Twenty-eight study patients had incomplete demographic data and were not included for computation of baseline patient profile. Male subjects accounted for 52.6% of the study population. The median age was 3.2 years, with IQR of 0.3 to 10.8 years. Respiratory failure (10.8%) was the most common diagnostic category among the medical patients (Table I). Patients with cardiovascular procedures (24.1%) comprised the largest group of surgical patients. Thirty-six of the 942 patients died, yielding a mortality rate of 3.8% (Table I). The mortality rate for all nondiabetic admissions during the study period was 57 in 1883 (3.0%). There were 27 patients with tracheostomy among the survivors in the study population. Excluding these patients, the survivors had a mean PICU length of stay of 5.0 ± 8.0 days, or a median of 2 days with IQR of 1 to 5 days. Initial blood glucose measurements ranged from 5 to 495 mg/dl, with a mean of 150 ± 64 mg/dl. The Figure (A) shows the frequency and cumulative frequency distribution at different glucose levels. The initial glucose measurement was >120 mg/dl in 61.8%, >150 mg/dl in 36.6%, and >200 mg/dl in 16.7% of patients. The maximum glucose values within 24 hours of the admission measurement were higher than the initial values: 166 ± 75 mg/dl, with a range of 66 to 495 mg/dl. The distribution of maximum glucose values is shown in the Figure (B). Maximum values were >120 mg/dl in 70.4%, >150 mg/dl in 44.5%, and >200 mg/dl in 22.3%. In the Figure, panel C shows the distribution of the highest glucose within 10 days of the initial value. The mean value was 172 ± 78 mg/dl and ranged from 66 to 497 mg/dl. The highest glucose values were >120 mg/dl in 75.0%, >150 mg/dl in 50.1%, and >200 mg/dl in 26.3%. The association between the different glucose levels and risk of death is shown in Table II. The risk of dying is significantly increased at maximum glucose within 24 hours of >150 mg/dl and at all cutoff values for maximum glucose within 10 days of the first. Moreover, at maximum glucose within 24 hours and within 10 days of the initial value, mortality rates increase at higher glucose values. For maximum glucose within 24 hours, mortality rates are as follows: 2.2% at 0 to 120 mg/dl, 2.5% at >120 to 150 mg/dl, 4.8% at >150 to 200 mg/dl, and 6.7% at >200 mg/dl. Mortality rates for maximum glucose within 10 days are 0.9%, 2.1%, 4.0%, and 8.1% at glucose values of 0 to 120 mg/dl, >120 to 150 mg/dl, >150 to 200 mg/dl, and >200 mg/dl, respectively. No trend was seen with the initial glucose value. A correlation between maximum glucose value within 10 days of the initial measurement and PICU length of stay was also noted in the study population (Table III). Persistent Hyperglycemia In Critically Ill Children 31

80 Table I. List of primary diagnoses Diagnoses No. of patients (%) No. of deaths Medical disorders Pulmonary diseases 158 (16.8) Respiratory failure 102 (10.8) 5 Other pulmonary diseases 56 (5.9) 0 Diseases of the neonate 55 (5.8) 3 Neurologic diseases 40 (4.2) 1 Hematologic diseases 30 (3.2) 3 Cardiovascular diseases 29 (3.1) 4 Gastrointestinal diseases 27 (2.9) 2 Sepsis 12 (1.3) 4 Miscellaneous diseases 60 (6.4) 0 Surgical procedures Cardiovascular surgery 227 (24.1) Major cardiovascular procedures 88 (9.3) 1 Other cardiovascular procedures 139 (14.8) 4 Neurosurgical procedures 99 (10.5) 2 Spinal surgery 44 (4.7) 0 Orthopedic surgery 39 (4.1) 0 Trauma 39 (4.1) 2 Tracheostomy 28 (3.0) 1 Gastrointestinal surgery 12 (1.3) 0 Miscellaneous surgical procedures 43 (4.6) 4 Figure. Distribution (n) and cumulative frequencies (-n-) of blood glucose values based on (A) initial glucose and (B) maximum glucose within 24 hours and (C) 10 days of first measurement. DISCUSSION The finding that maximum glucose values on average exceed the initial glucose values suggests that estimates of the prevalence of hyperglycemia based on initial or admission glucose values may underestimate the likelihood that patients become hyperglycemic. In a study on the prevalence of hyperglycemia in critically ill children, Ruiz Magro et al 17 reported that 50% of 353 patients without diabetes mellitus had initial glucose values >120 mg/dl. The prevalence of hyperglycemia reported in this study is greater than that reported in other studies in children even when considering higher glucose values as the cutoff. For children seen in the emergency department, the prevalence of hyperglycemia is reported to range from 3.8% to 5.0% (based on the initial blood glucose being >150 mg/dl). 11,14 However, the setting of care is different in those studies and the level of illness may be assumed to be, on average, lower in the emergency department compared with the PICU. There are no definite criteria for diagnosing hyperglycemia among patients without diabetes mellitus. Among diabetics, however, hyperglycemia is defined by the World Health Organization as a fasting blood glucose >126 mg/dl or a random blood glucose of >200 mg/dl. 18 In the present report, three cutoff values were used. The lower two values have been previously used as threshold values in studies in children. 14,15,17 The 200 mg/dl cutoff, however, is based on the WHO criteria and adult studies. 1,6,7,18 Another problem with defining the prevalence of hyperglycemia is determining which of the different glucose levels is most predictive of outcome. Most of the studies in both adults and children used initial or admission hyperglycemia. However, Krinsley 19 demonstrated that among adult Intensive Care Unit (ICU) patients, the mean glucose value and the maximum glucose value obtained during the ICU stay were more predictive of outcome than was the initial glucose value. We found no association between initial blood glucose and risk for death. This is in contrast to the findings of Ruiz Magro et al, 17 who reported that the initial glucose level was significantly higher in patients who died compared with the survivors. Our findings also conflict with results in adult subjects in whom initial blood glucose is correlated with death among both critically ill and non critically ill patients, although the overall mortality rate in adult studies is higher than we report for children. 19,20 The mortality rate of 3.0% for all patients without diabetes admitted during the study period is not significantly different from the pooled multicenter mortality rate of 2.2% to 16.4% reported by Pollack et al. 21 In contrast to initial glucose values, higher maximum glucose values measured within 24 hours and within 10 days of the initial value were correlated with higher mortality rates. This is in agreement with the study by Krinsley 19 in adult patients. The increase in mortality risk was demonstrated beginning at maximum glucose >150 mg/dl within 24 hours of the first value. At a lower threshold of 120 mg/dl, the highest 32 Vincent Faustino and Apkon The Journal of Pediatrics January 2005

81 Table II. Risk of death at different levels of hyperglycemia Cutoff value Mortality rate above cutoff value Mortality rate at or below cutoff value Relative risk (95% CI) Initial glucose 120 mg/dl 24/582 (4.1 %) 12/360 (3.3 %) 1.24 (0.63, 2.44) 150 mg/dl 13/345 (3.8 %) 23/597 (3.9 %) 0.98 (0.50, 1.91) 200 mg/dl 7/157 (4.5 %) 29/785 (3.7 %) 1.21 (0.54, 2.71) Maximum glucose within 24 h of initial glucose 120 mg/dl 30/663 (4.5 %) 6/279 (2.2 %) 2.10 (0.89, 5.00) 150 mg/dl 24/419 (5.7 %) 12/523 (2.3 %) 2.50 (1.26, 4.93) 200 mg/dl 14/210 (6.7 %) 22/732 (3.0 %) 2.22 (1.16, 4.26) Maximum glucose within 10 d of initial glucose 120 mg/dl 34/706 (4.8 %) 2/236 (0.8 %) 5.68 (1.38, 23.47) 150 mg/dl 29/472 (6.1 %) 7/470 (1.5 %) 4.13 (1.83, 9.32) 200 mg/dl 20/248 (8.1%) 16/694 (2.3%) 3.50 (1.84, 6.64) Table III. Length of PICU stay among survivors at different levels of hyperglycemia [mean ± SD and median (1 st to 3 rd quartiles) in days] Cutoff value Length of stay above cutoff value Length of stay at or below cutoff value P value Initial glucose 120 mg/dl 4.5 ± (1 4.5) 150 mg/dl 4.7 ± (1 5) 200 mg/dl 4.7 ± (1 5) Maximum glucose within 24 h of initial glucose 120 mg/dl 5.0 ± (1 5) 150 mg/dl 5.1 ± (1 5) 200 mg/dl 5.6 ± (1 6) Maximum glucose within 10 d of initial glucose 120 mg/dl 5.4 ± (1 6) 150 mg/dl 6.1 ± (1 7) 200 mg/dl 6.9 ± (1 8) 5.7 ± (1 8) 5.2 ± (1 6) 5.1 ± (1 5) 5.0 ± (1 6) 5.0 ± (1 5.5) 4.9 ± (1 5) 4.0 ± (1 4) 4.0 ± (1 4) 4.4 ± (1 5), ,.001 glucose level within 10 days begins to correlate with higher risks of dying. Higher initial glucose values was correlated with shorter PICU stays, whereas higher maximal glucose values was correlated with longer PICU stays. Our findings differ somewhat from a study of adult patients in whom admission hyperglycemia has been correlated with longer hospital stays. 7 It is tempting to hypothesize that initial hyperglycemia may indicate a patient s ability to respond to physiologic stress more adaptively, whereas hyperglycemia of longer duration may lead to more deleterious consequences. One limitation of this study is the potential for selection bias, given the number of patients who were excluded due to lack of glucose determinations. Our institution does not have any specific policies prescribing which patients will have their blood glucose level determined, although blood glucose is Persistent Hyperglycemia In Critically Ill Children 33

82 typically measured in all infants, postoperative patients, and patients who are not eating or receiving nutritional supplementation. Therefore, blood glucose may preferentially have been taken from patients who were expected to have problems with glucose control, whether hypoglycemia or hyperglycemia. In fact, the mortality rate in our study population is slightly higher than the mortality rate for all the nondiabetic admissions during the same time period. It is possible that blood glucose values in those patients without measurements could alter our estimates for the prevalence of hyperglycemia. However, when we look at the potential error, we find that even if none of the excluded patients were hyperglycemic, the overall prevalence of hyperglycemia would still be half of our estimates (8.3% to 37.5%, depending on the glucose cutoff), which we believe to be of significant concern. If this were true, it would also be the case that hyperglycemia would be associated with an even higher risk of death. Although it is also possible that hyperglycemia in excluded survivors coupled with normoglycemia in excluded nonsurvivors would invalidate our finding that hyperglycemia is correlated with death, this is unlikely; this would contradict findings in adult patients as well as in our studied cohort. Moreover, it would suggest an unusual type of selection bias for deciding for which patients to measure glucose levels. The timing of the initial glucose may also cause minor concerns. While termed initial, it does not necessarily mean it was taken on the first PICU day. This may somewhat complicate comparisons with other studies. This study explored the correlation between hyperglycemia and outcome. Our data do not allow us to evaluate a causal link between the two. It is possible that hyperglycemia and risk of dying are both correlated with a third factor, such as severity of illness. Data to evaluate adequately the severity of illness is not available to us. Even if both hyperglycemia and death were correlated with a third factor such as elevated risk scores, we do not believe that it would be possible to say whether or not high glucose had an independent effect in raising the risk of dying. There is support for such an effect. Van den Berghe et al 20 have shown that by strictly regulating blood glucose between 80 and 110 mg/dl among mechanically ventilated patients admitted in the surgical ICU, mortality rate decreases from 8.0% to 4.6%. Post hoc analysis of the same study reported that the decrease in the mortality rate was due to glucose control and not the amount of insulin given. 22 Similar findings were reported by Finney et al 23 in their study among critically ill adult patients admitted in the ICU. We have demonstrated that hyperglycemia occurs commonly in critically ill children and that there is a correlation between the risk of dying and the likelihood of having a glucose value >150 mg/dl. The importance of our work rests on the possibility that glycemic control may confer a survival advantage in children, as it does in adults. That possibility, as yet untested, does not depend on the precise frequency of hyperglycemia or on the association of hyperglycemia with another factor such as a greater severity of illness. Given these findings and the evidence that glycemic control can enhance outcome in adult patients, prospective, randomized, controlled trials comparing tight glucose control versus standard of care or control of glucose at different levels are needed to show definitively that hyperglycemia is associated with worse clinical outcome and that treatment with insulin would improve mortality rates and length of hospital stay. REFERENCES 1. Capes SE, Hunt D, Malmberg K, Gerstein HC. Stress hyperglycemia and increased risk of death after myocardial infarction in patients with and without diabetes: a systematic overview. Lancet 2000;355: Preiser JC, Devos P, Van den Berghe G. Tight control of glycaemia in critically ill patients. Curr Opin Clin Nutr Metab Care 2002;5: Gore DC, Chinkes D, Heggers J, Herndon DN, Wolf SE, Desai M. Association of hyperglycemia with increased mortality after severe burn injury. J Trauma 2001;51: Mizock BA. Alterations in fuel metabolism in critical illness: hyperglycaemia. Best Pract Res Clin Endocrinol Metab 2001;15: McCowen KC, Malhotra A, Bistrian BR. Stress-induced hyperglycemia. Crit Care Clin 2001;17: Capes SE, Hunt D, Malmberg K, Pathak P, Gerstein HC. Stress hyperglycemia and prognosis of stroke in nondiabetic and diabetic patients: a systematic overview. Stroke 2001;32: Umpierrez GE, Isaacs SD, Bazargan N, You X, Thaler LM, Kitabchi AE. Hyperglycemia: an independent marker of in-hospital mortality in patients with undiagnosed diabetes. J Clin Endocrinol Metab 2002;87: Type 2 diabetes in children and adolescents: American Diabetes Association. Pediatrics 2000;105: Green A, Christian Hirsch N, Pramming SK. The changing world demography of type 2 diabetes. Diabetes Metab Res Rev 2003;19: Agus MS, Jaksic T. Nutritional support of the critically ill child. Curr Opin Pediatr 2002;14: Valerio G, Franzese A, Carlin E, Pecile P, Perini R, Tenore A. High prevalence of stress hyperglycaemia in children with febrile seizures and traumatic injuries. Acta Paediatr 2001;90: Weise K, Zaritsky A. Endocrine manifestations of critical illness in the child. Pediatr Clin North Am 1987;34: Apkon M, Singhaviranon P. Impact of an electronic information system on physician workflow and data collection in the intensive care unit. Intensive Care Med 2001;27: Bhisitkul DM, Morrow AL, Vinik AI, Shults J, Layland JC, Rohn R. Prevalence of stress hyperglycemia among patients attending a pediatric emergency department. J Pediatr 1994;124: Gupta P, Natarajan G, Agarwal KN. Transient hyperglycemia in acute childhood illnesses: to attend or ignore? Indian J Pediatr 1997;64: Guyatt GH, Rennie D. User s guides to the medical literature. JAMA 1993;270: Ruiz Magro P, Aparicio Lopez C, Lopez-Herce Cid J, Martinez Campos M, Sancho Perez L. [Metabolic changes in critically ill children]. An Esp Pediatr 1999;51: Mesotten D, Van den Berghe G. Clinical potential of insulin therapy in critically ill patients. Drugs 2003;63: Krinsley JS. Association between hyperglycemia and increased hospital mortality in a heterogeneous population of critically ill patients. Mayo Clin Proc 2003;78: Van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, et al. Intensive insulin therapy in the critically ill patients. N Engl J Med 2001;345: Pollack MM, Patel KM, Ruttimann UE. The Pediatric Risk of Mortality III Acute Physiology Score (PRISM III-APS): a method of assessing physiologic instability for pediatric intensive care unit patients. J Pediatr 1997;131: Van den Berghe G, Wouters PJ, Bouillon R, Weekers F, Verwaest C, Schetz M, et al. Outcome benefit of intensive insulin therapy in the critically ill: insulin dose versus glycemic control. Crit Care Med 2003;31: Finney SJ, Zekveld C, Elia A, Evans TW. Glucose control and mortality in critically ill patients. JAMA 2003;290: Vincent Faustino and Apkon The Journal of Pediatrics January 2005

83 CHILDREN WITH EARLY-ONSET INFLAMMATORY BOWEL DISEASE (IBD): ANALYSIS OF A PEDIATRIC IBD CONSORTIUM REGISTRY MELVIN B. HEYMAN, MD, MPH, BARBARA S. KIRSCHNER, MD, BENJAMIN D. GOLD, MD, GEORGE FERRY, MD, ROBERT BALDASSANO, MD, STANLEY A. COHEN, MD, HARLAND S. WINTER, MD, PATRICIA FAIN, BSN, RN, CHRIS KING, RN, MSN, FNP, TERRY SMITH, MS, AND HASHEM B. EL-SERAG, MD, MPH Objective To determine the characteristics of inflammatory bowel disease (IBD) in young patients. Study design Uniform data were collected from a cohort of patients with IBD who were enrolled from January 2000 to November 2002 at six pediatric centers (Pediatric IBD Consortium). Results Of 1370 children in the registry, the mean age at IBD diagnosis was 10.3 ± 4.4 years; 54% were male, and 86% were white. Diagnosis was confirmed in 87 (6.1%) under 3 years of age, 211 (15.4%) before 6 years, 654 (47.7%) at 6 to 12 years, and 505 (36.9%) at 13 to 17 years. More than 63% of children younger than 8 years of age had isolated colonic disease, whether Crohn disease, ulcerative colitis (UC), or indeterminate colitis. Conversely, only 35% of those 8 years of age or older had isolated colonic disease (P <.0001). Overall, 29% had one or more family members with IBD. The subgroup of children younger than 3 years of age with UC had the highest prevalence of first-degree relatives with IBD (44%). Conclusions This demographically diverse pediatric IBD cohort revealed age-related variation in the distribution of IBD phenotype, with a high prevalence of isolated colonic See editorial, p 7. disease in young children. Positive family history was especially common in young patients with UC. (J Pediatr 2005;146:35-40) Inflammatory bowel disease (IBD) has recently been cited as one of the five most prevalent gastrointestinal disease burdens in the United States, with an overall health care cost of more than $1.7 billion. 1 A large proportion of individuals affected by IBD have symptoms before 30 years of age and have lifelong disease. 2-5 Although populationbased studies are lacking, an estimated 20% to 30% of patients with IBD have the onset of their symptoms, and 10% to 15% are diagnosed with IBD under the age of 18 years. 2,6-9 Young children (ie, #6 years old) with IBD represent a unique cohort of patients to investigate the initial host immune response, early intervention with immunomodulatory agents, characterization of genotype-phenotype relations, natural history of the disease, and environmental influences of disease development. 3,10-14 Understanding the factors that contribute to early age of onset could potentially facilitate intervention strategy development, thereby influencing disease outcome and risk of long-term sequelae. In 2000, the Pediatric Inflammatory Bowel Disease Consortium (PediIBDC) registry was established at six pediatric medical centers in different geographical areas of the United States. Uniformly collected data were entered into a central, computerized repository, using a standardized protocol to test a number of hypotheses-driven questions regarding the epidemiology of pediatric IBD. In this report, information in the PediIBDC registry was analyzed to examine epidemiologic features of IBD in children including demographic details, disease distribution at diagnosis, and family history. CD IBD IC Crohn disease Inflammatory bowel disease Indeterminate colitis PediIBDC UC Pediatric Inflammatory Bowel Disease Consortium Ulcerative colitis From UCSF Children s Hospital, University of California, San Francisco; The University of Chicago Children s Hospital, Chicago, Illinois; Emory University School of Medicine, Atlanta, Georgia; Texas Children s Hospital, Baylor College of Medicine, Houston, Texas; Children s Hospital of Philadelphia, Philadelphia, Pennsylvania; Children s Center for Digestive Healthcare, Atlanta, Georgia; MassGeneral Hospital for Children, Boston, Massachusetts; and Gastroenterology and Health Services Research, Houston VA Medical Center, and Baylor College of Medicine, Houston, Texas. Between April 2000 and March 2003, the Crohn s and Colitis Foundation of America (CCFA) was a major supporter of the consortium and supplemented development of the database. This support would not have been possible without the vision and commitment to pediatric IBD of Joel Cutler. Investigators were also supported in part by grants from the NIH (DK [MBH], DK and DK [BDG]), from the VA HSR & D CDA Awardee RCD [HES] and from The Nathan Cummings Foundation, Chicago, Illinois [BSK]. Submitted for publication Feb 26, 2004; last revision received Jul 14, 2004; accepted Aug 17, Reprint requests: Dr Melvin B. Heyman, 500 Parnassus Ave, MU 4-East, Room 406, University of California, San Francisco, CA mheyman@peds.ucsf.edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

84 METHODS Investigators from six regional IBD centers of the PediIBDC established a registry to collect demographic, clinical, and epidemiologic data from pediatric patients with IBD. The six sites are Texas Children s Hospital, Baylor College of Medicine; MassGeneral Hospital for Children; Emory University School of Medicine, Egleston Children s Hospital and Scottish Rite Children s Hospitals of Children s Healthcare of Atlanta; Children s Hospital of Philadelphia; UCSF Children s Hospital at the University of California, San Francisco; and the University of Chicago Children s Hospital. Institutional review board approval for the registry protocol and the informed consent and assent forms were obtained at each site before subject enrollment and data collection. The registry was established in compliance with the Health Insurance Portability and Accountability Act (HIPAA). Signed parental and patient informed consent and signed youth assent when appropriate were required from all patients enrolled. Trained abstractors at each center obtained information from the medical records (electronic and paper charts). Standardized information was entered into the registry. Data were submitted quarterly to the central repository at Baylor College of Medicine for scrutiny of data quality, maintenance of data integrity, data storage, and subsequent data analyses. Enrollment into the registry began in January Data for patients entered through November 1, 2002, were analyzed in this report. Eligible subjects included all patients with any form of IBD (ulcerative colitis [UC], Crohn disease [CD], and indeterminate colitis [IC]) who were diagnosed before 18 years of age and actively followed in the pediatric gastroenterology clinics at the six participating sites. Both prevalent (existing) and incident (new) IBD cases were prospectively enrolled. Although no rigorous enrollment criteria were applied, patients were not preselected on the basis of any specific designated clinical information but were enrolled when consented in the clinics. Patient data were entered consistent with the most recent diagnosis; thus, if an initial diagnosis was IC but subsequently determined to be CD or UC, the latter was considered to be the most accurate diagnosis of the child. Diagnosis of IBD was based on clinical history, physical examination, endoscopic appearance, histologic findings, and radiologic studies. CD was defined by (1) evidence of a granuloma in any one biopsy from upper and/or lower endoscopy; (2) in the absence of a granuloma, presence of skip lesions on colonoscopy and/or presence mucosal abnormality with microscopic focal chronic inflammatory changes in upper endoscopy or segmental small intestinal radiologic findings consistent with CD; (3) presence of perianal disease (ie, abscesses, fistulae, large skin tags); and/or (4) presence of transmural inflammation such as stricturing or fistulizing disease. Gastroduodenal findings in patients with CD were classified as normal, abnormal compatible with CD, or abnormal, not IBD. UC was defined as continuous disease confined to the colon and no evidence of small intestinal disease (other than backwash ileitis) on biopsy or by radiology. IC was defined as colitis that could not definitively be declared as CD or UC, based on the above criteria. Patients with infectious, eosinophilic, or other specific underlying cause for enteritis or colitis were excluded. The information retrieved for the purpose of this study included demographic features (age, sex, race/ethnicity), family history of IBD, IBD type (CD, UC, IC), and disease distribution. Race/ethnicity was self-reported and categorized as white (includes Hispanic and Jewish, which were listed independently), black, or other. Family history on IBD was based on reports or records provided by patients regarding IBD in siblings, parents, grandparents, or extended family. The subjects disease distribution was stratified as follows: UC pancolitis, left-sided colitis, or proctitis; and CD gastroduodenal, ileal, ileocolonic, right colon, or diffuse colitis. Statistical Analysis Differences were examined among patients in the following age categories, adapted from FDA age group distribution, at the time of IBD diagnosis: 0 to 2 years, 3 to 5 years, 6 to 12 years, and 13 to 17 years. 15 Univariate comparisons were conducted among these age groups with regard to IBD categories (CD, UC, IC), age of diagnosis of IBD, sex, race/ethnicity, disease distribution, and family history of IBD. We used x 2 tests to compare categoric variables, 2-tailed t tests to compare continuous variables, and analysis of variance tests to compare continuous variables between more than two groups. Multivariate logistic regression analysis was conducted to examine the possible determinants of early onset IBD. Preliminary analysis of the complete data set along the selected age strata revealed a distinct cutoff point at 8 years of age. Subsequent analyses were performed defining of early onset as those younger than 8 years of age and later onset as those 8 years of age and older. Covariates in the model included sex, race, family history of IBD, IBD disease category, and disease distribution in the gastrointestinal tract. Separate models were constructed for CD and UC. Parameter estimates and standard errors were used to calculate adjusted odds ratios and accompanying 95% confidence intervals. Data are expressed as mean ± SD. RESULTS Data from 1370 patients were entered into the registry; 798 (58%) had CD, 393 (29%) had UC, and 179 (13%) had IC (Table I). Of the entire group, 54% were male, but the male predominance was most significant (63%) in the youngest group (0 to 2 years) (Table II). Self-reported racial/ethnic distribution was white in 86% (including Hispanic [1%] and Jewish [9%] of the entire patient population), black in 10%, and other in 4%. Newly diagnosed or incident cases of IBD (n = 404, 29% of the total group) presented during the 34-month enrollment period. Of the new cases, 262 (65%) were CD, 92 (23%) were UC, and 50 (12%) were IC. The remaining cases were existing or prevalent cases. Approximately 10% of the patients carried a diagnosis of IBD for more than 10 years, and 36 Heyman et al The Journal of Pediatrics January 2005

85 Table I. Age at diagnosis and disease distribution of children with IBD enrolled in the PediIBDC between January 1, 2000, and November 1, 2002 (n = 1370) Age of diagnosis (y) Disease type Total Crohn disease 31 (36%) 44 (35%) 392 (60%) 331 (66%) 798 (58%) Ulcerative colitis 27 (31%) 58 (47%) 181 (28%) 127 (25%) 393 (29%) Indeterminate colitis 29 (33%) 22 (18%) 81 (12%) 47 (9%) 179 (13%) P value.814,.0001,.0001,.0001 Totals 87 (100%) 124 (100%) 654 (100%) 505 (100%) 1370 Percentages are of the total in each age group. P values are for x 2 tests comparing proportions of patients with each disease type within each age category. Table II. Demographic characteristics of patients with IBD enrolled in the PediIBDC registry between January 1, 2000, and November 1, 2002 Age of diagnosis (y) Variable 0 2 (n = 87) 3 5 (n = 124) 6 12 (n = 654) (n = 505) Total (n = 1370) P value Sex: Male (%) 55 (63%) 68 (55%) 344 (53%) 278 (55%) 745 (54%).412 Race/ethnicity White 75 (86%) 114 (92%) 568 (87%) 414 (82%) 1171 (86%) * Hispanic (1%) Jewish (9%) Black 5 (6%) 6 (5%) 60 (9%) 67 (13%) 138 (10%) * Other 7 (8%) 4 (3%) 26 (4%) 24 (5%) 61 (4%).957 Percentages are of the total in each age group. P values are for x 2 tests, comparing proportions of patients with each variable within each age category. *P for Cochrane-Armitage test for linear trend was also significant. 80% were diagnosed and actively followed at one of the centers for at least 2 years. The mean duration of disease to the time of patient enrollment in the registry was 5.4 ± 4.0 years. IBD Type by Age Group Patients diagnosed with IBD in the youngest age group (0 to 2 years) were equally affected with UC, CD, and IC (Table I). IBD-specific symptoms were identified within the first 12 months of age in 37 infants (2.7% of our registry cohort); of these, 12 (1% of the entire registry) had a definitive IBD diagnosis before 1 year of age. UC was the most prevalent disease type (47%) among those in the 3- to 5-year age group, compared with CD (35%). Conversely, CD comprised 60% of IBD in older children (6 to 12 years) and 66% of adolescents (13 to 17 years), whereas subjects with UC were present in lower percentages, 28% and 25%, respectively, in these age categories. IC was most prevalent among the 0- to 2-year-olds (33%) and declined progressively with increasing age, present in 9% of the 13- to 17-year-old patients. Age-related differences in sex distribution were not statistically significant. Incident cases showed a similar age distribution among the three IBD categories; no differences were present in distribution of patients between UC, CD, and IC among the younger (0- to 2-year and 3- to 5-year) groups, whereas the older (6- to 12-year and 13- to 17-year) groups had a significantly increased prevalence of CD compared with UC and IC (P <.0001 for each older age group). A significantly greater proportion of whites and a smaller proportion of blacks were found among younger compared with older children (Table II). Anatomic Distribution of Disease (UC, IC and CD) by Age Group Colonic involvement (with or without other disease sites) with IBD was significantly more frequent among younger children compared with older children. IBD isolated to the colon (UC, IC, colonic CD) was recorded in 60 of the 87 (69%) children diagnosed with IBD in the 0- to 2-year-olds, compared with 232 of the 505 (46%) children in the 13- to 17-year-olds (P <.0001). The proportion of children diagnosed with UC, IC, and isolated colonic CD declined with age, whereas those diagnosed with other forms of CD increased with age. Among patients with UC, a trend toward fewer cases with proctitis (4%) was observed among young patients (0 to 2 years and 3 to 5 years), compared with older children between 6 and 12 years (11%) and adolescents 13 to 17 years of age (12%) (P =.08 for both comparisons). Similarly, CD distribution tended to differ according to the age of diagnosis. Patients with isolated colonic CD comprised 24 of 75 (32%) of the children Children With Early-Onset Inflammatory Bowel Disease (IBD): Analysis Of A Pediatric IBD Consortium Registry 37

86 Figure. In patients younger than 8 years of age at the time of diagnosis, ulcerative colitis predominates compared with Crohn disease. More than twice as many patients are diagnosed with indeterminate colitis before 8 years of age compared with older children. Crohn disease is the predominant form of pediatric IBD (P <.0001 for all diseases) in patients 8 years of age and older. younger than 6 years, 80 of 392 (20%) of children 6 to 12 years, and 80 of 331 (24%) of adolescents (13 to 17 years) (P = 0.08). The proportion of patients with gastroduodenal CD increased slightly with age, present in 5% of 0- to 5-year-olds, 13% of 13- to 17-year-olds, and 10% of 6- to 12-year-olds. Analysis by year of age revealed a distinct contrast in the distribution of colonic disease between younger and older patients at approximately 8 years of age. Among patients younger than 8 years of age, 70% had a diagnosis of disease limited to the colon, compared with 46% in patients older than age 8 years (P <.05). Notable differences in the distribution of UC, CD, and IC (Figure) were observed at 8 years of age, with older patients more commonly diagnosed with CD compared with UC (65% vs 25%, P <.0001). Furthermore, IC was diagnosed in 22% of patients younger than 8 years of age but only in 10% of those 8 years of age and older (P <.0001). Of 179 patients with IC, 27 (15%) were reclassified to CD and 11 (6%) were reclassified to UC during a mean follow-up of 2 years, 3 months. Follow-up of patients with UC demonstrated reclassification to CD in 6% of cases and to IC in 2%. Subjects diagnosed initially with CD were rarely changed later to UC or IC (0.3% and 0.9%, respectively). Family History of IBD by Age Groups Overall, a positive family history of IBD was found in 398 (29%) of patients. Family history of IBD was reported by 3% in siblings, 9% in parents, and 22% in extended family members (grandparents and others). The proportion of patients with affected first-degree relatives (parents or siblings) was significantly greater among younger (0 to 2 years) than older patients with IBD. The youngest group (0 to 2 years) with UC trended toward a greater prevalence of any positive family history for IBD than older children (44% vs 28%, P =.06), and this difference was particularly apparent when family history of first-degree relatives alone was considered (44% vs 19%, P =.0002). A positive family history was not more frequent among the other disease categories and age groups. Multivariate Analysis Table III shows the results of multiple logistic regression analysis conducted to examine the presence of IBD among young children (younger than 8 years) compared with IBD in patients between 8 and 17 years of age. The type of IBD (UC, CD, or IC) and history of IBD in first-degree relatives were independent risk factors for early-onset IBD. Compared with 38 Heyman et al The Journal of Pediatrics January 2005

87 CD, UC was approximately 2.5 times and IC was 3.5 times more likely to affect children with IBD who are younger than 8 years of age. Furthermore, a positive family history of IBD in siblings or parents was associated with a 75% increase in the risk of early onset IBD. DISCUSSION We describe the epidemiologic characteristics of a single largest cohort of pediatric patients with IBD. Our report also includes a large cohort of very young children (younger than 3 years of age; n = 87; 6.4% of the total) with confirmed IBD. This information underscores increasing numbers of singlecenter reports that IBD can start in infancy and affect children younger than 3 years of age. 16 Moreover, we describe novel observations of the epidemiologic features of IBD in children. In particular, we showed that the predominance of colonic disease (UC, IC, and Crohn colitis) and the greater prevalence of positive family history of IBD were independent characteristics of younger children with IBD. This study has some limitations. Our registry is based on patient data from six large regional referral centers, which may not allow extrapolation to the entire US pediatric population with IBD. However, pediatric patients with chronic disease tend to be referred to centers such as these, and, unlike adults with IBD, children are rarely treated solely by a primary care provider or pediatrician. Conversely, centers that care for adult patients may lose a significant portion of IBD patients who are followed by internists. Since most patients were enrolled as they came for clinic visits at each center, a bias may exist for including those who have more chronic or severe disease, requiring more visits. This might preferentially affect younger patients because older teenagers tend to be referred directly to adult gastroenterologists. This potential bias might have inflated the proportions of younger children in our cohort but is unlikely to affect their clinical and epidemiologic characteristics. The epidemiology of pediatric IBD has been described in population-based studies from European centers In contrast, in North America, most available information originates primarily from single-center studies. 16,22-26 In addition, the characteristic features of early-onset IBD (ie, #6 years old) and its natural history have only recently been reported in small case series. 3,23 Previous studies in the United States and Canada have used local hospital registries to examine features and complications of pediatric IBD, including extraintestinal manifestations of IBD, perianal disease, and outcomes of medical and surgical therapies However, these studies did not examine age-related differences in disease phenotype or had insufficient sample size and power to identify such differences. In particular, the specific type of IBD and the disease distribution in different age groups have not been well characterized. It remains unknown whether family history of IBD (ie, a surrogate marker for genetic susceptibility) differs among age groups according to type of IBD (UC, CD, or IC). 22,31,32 In our study population, 37 (2.7% of our registry cohort) had IBD-specific symptoms, and 12 of these (1% of the Table III. Independent associations of early-onset IBD (<8 years of age) examined in a multiple logistic regression analysis (n = 1370) Variable Parameter estimate Adjusted risk ratio (95% CI) P value Type of IBD UC (vs CD) ( ),.0001 IC (vs CD) ( ),.0001 Race Black (vs white) ( ) Other (vs white) ( ) Female (vs male) ( ).9589 History of IBD in first-degree relatives ( ).0030 registry patients) had a definitive IBD diagnosis before 1 year of age. Additionally, 211 (15.4%) were diagnosed under the age of 6 years. These data provide compelling evidence that IBD can begin very early in life. No reports of sufficient sample size exist that describe the natural history studies in this extremely young age group. 2,3,7,13,22,33,34 The long-term outcome of pediatric patients with IBD is especially important because the relative risk of development of cancer with either UC or CD increases in proportion to the number of years from initial diagnosis. Thus, prospective, longitudinal studies of very young children with IBD, including surveillance for potential surrogate markers for cancer and the cumulative effects of immunomodulatory agents, are needed. We observed that IC is diagnosed most frequently among the youngest group of children with IBD, constituting approximately one third of those diagnosed in the first 2 years of life. This proportion of IC is significantly higher than that found in older children and adolescents in this study. The proportion of patients with IC is also greater than that previously reported in other pediatric series or in adults in whom IC accounts for 10% to 15% of newly diagnosed patients with IBD. 21,35,36 We speculate that IC might represent an evolving form of IBD that presents before definitive disease phenotype, especially in the very young. Alternatively, current classification criteria to differentiate UC from CD might not be applicable in young children. In our study, children younger than 8 years of age were much more likely to have isolated colonic disease (70% vs 46%, P <.05). This finding could reflect either a higher prevalence of UC and IC or more frequent isolated colonic involvement in patients with CD in this group. The finding of more frequent isolated colonic disease in young patients with CD is similar to the report by Mamula et al. 16 A positive family history of IBD was reported in 29% of our study population. Familial associations with both UC and CD have been recognized for many years A single-center US study published 30 years ago described a positive family history of IBD in 29% of 336 children diagnosed with UC Children With Early-Onset Inflammatory Bowel Disease (IBD): Analysis Of A Pediatric IBD Consortium Registry 39

88 younger than 21 years of age. 40 Conversely, the frequency of a positive family history in first- and second-degree relatives of children with newly diagnosed IBD was only 10% in Norway and 11% in Wisconsin. 9,17 In our study, age-related differences were significant only when first-degree relatives were considered, and especially in UC. Specifically, children younger than 3 years with UC had more frequent first-degree relatives with IBD than patients in older age groups (44% vs 12% to 24%). Subsequent studies in this population to investigate potential combinations of genetic and environmental influences in disease onset are therefore needed. The registry was supported by the tremendous and indispensable efforts of a group of local study coordinators: Amy York, Jennifer Cooper, Rachel Kreh, Catherine Geraci, and Jacqueline Madden. Advice and guidance pertaining to database development was provided by Traci Clemens, PhD, Emmes Corporation, Rockville, Md. We also thank our colleagues who have referred patients for enrollment in the registry. REFERENCES 1. Sandler RS, Everhart JE, Donowitz M, Adams E, Cronin K, Goodman C, et al. The burden of selected digestive diseases in the United States. Gastroenterology 2002;122: Baldassano RN, Piccoli DA. Inflammatory bowel disease in pediatric and adolescent patients. Gastroenterol Clin North Am 1999;28: Marx G, Seidman EG, Martin SR, Deslandres C. Outcome of Crohn s disease diagnosed before two years of age. J Pediatr 2002;140: Hyams JS. Crohn s disease in children. Pediatr Clin North Am 1996;43: Hyams JS, Davis P, Grancher K, Lerer T, Justinich CJ, Markowitz J. Clinical outcome of ulcerative colitis in children. J Pediatr 1996;129: Mendeloff AI, Calkins BM. The epidemiology of idiopathic inflammatory bowel disease. Philadelphia: Lea & Febiger; Bernstein CN, Blanchard JF. The epidemiology of Crohn s disease. Gastroenterology 1999;116: Farrokhyar F, Swarbrick ER. A critical review of epidemiological studies in inflammatory bowel disease. Scand J Gastroenterol 2001;36: Kugathasan S, Judd RH, Hoffmann RG, Heikenen J, Telega G, Khan F, et al. Epidemiologic and clinical characteristics of children with newly diagnosed inflammatory bowel disease in Wisconsin: a statewide populationbased study. J Pediatr 2003;143: Andus T, Gross V. Etiology and pathophysiology of inflammatory bowel disease: environmental factors. Hepatogastroenterology 2000;47: Dominitz JA, Young JC, Boyko EG. Outcomes of infants born to mothers with inflammatory bowel disease: a population-based cohort study. Am J Gastroenterol 2002;97: Fiocchi C. Inflammatory bowel disease: etiology and pathogenesis. Gastroenterology 1998;115: Mahmud N, Weir DG. The urban diet and crohn s disease: is there a relationship? Eur J Gastroenterol Jepatol 2001;13: Oliva-Hemker M, Fiocchi C. Etiopathogenesis of inflammatory bowel disease: the importance of the pediatric perspective. Inflamm Bowel Dis 2002; 8: Food and Drug Administration, Department of Health and Human Services, Final Rule, 21 CFR Part 201, Federal Registry 1994;59(No. 238): Mamula P, Telega GW, Markowitz JE, Brown KA, Russo PA, Piccoli DA, et al. Inflammatory bowel disease in children 5 years of age and younger. Am J Gastroenterol 2002;97: Olafsdottir EJ, Fluge G, Haug K. Chronic inflammatory bowel disease in children in western Norway. J Pediatr Gastroenterol Nutr 1989;8: Langholz E, Munkholm P, Krasilnikoff PA, Binder V. Inflammatory bowel diseases with onset in childhood: clinical features, morbidity, and mortality in a regional cohort. Scand J Gastroenterol 1997; 32: Sawczenko A, Sandhu BK, Logan RF, Jenkins H, Taylor CJ, Mian S, et al. Prospective survey of childhood inflammatory bowel disease in the British Isles. Lancet 2001;357: Moum B, Ekbom A, Vatn MH, Aadland E, Sauar J, Lygren I, et al. Inflammatory bowel disease: re-evaluation of the diagnosis in a prospective population based study in south eastern Norway. Gut 1997;40: Lindberg E, Lindquist B, Holmquist L, Hildebrand H. Inflammatory bowel disease in children and adolescents in Sweden, J Pediatr Gastroenterol Nutr 2000;30: Buller H, Chin S, Kirschner B, Kohn J, Markowitz J, Moore D, et al. Inflammatory bowel disease in children and adolescents: working group report of the first world congress of pediatric gastroenterology, hepatology, and nutrition. J Pediatr Gastroenterol Nutr 2002;35:S Farmer RG, Michener WM. Prognosis of Crohn s disease with onset in childhood or adolescence. Dig Dis Sci 1979;24: Michener WM, Farmer RG, Mortimer EA. Gastroenterology: longterm prognosis of ulcerative colitis with onset in childhood or adolescence. J Clin Gastroenterol 1979;1: Gryboski JD. Crohn s disease in children 10 years old and younger: comparison with ulcerative colitis. J Pediatr Gastroenterol Nutr 1994;18: Gryboski JD. Ulcerative colitis in children 10 years old or younger. J Pediatr Gastroenterol Nutr 1993;17: Kirschner BS. Safety of azathioprine and 6-mercaptopurine in pediatric patients with inflammatory bowel disease. Gastroenterology 1998;115: Griffiths AM, Wesson DE, Shandling B, Corey M, Sherman PM. Factors influencing postoperative recurrence of crohn s disease in childhood. Gut 1991;32: Hyams JS. Extraintestinal manifestations of inflammatory bowel disease in children. J Pediatr Gastroenterol Nutr 1994;19: Markowitz K, Grancher K, Rosa J, Simpser E, Aiges H, Daum F. Highly destructive perianal disease in children with crohn s disease. J Pediatr Gastroenterol Nutr 1995;21: Baldassano RN, Han PD, Jeshion WC, Berlin JA, Piccoli DA, Lautenbach E, et al. Pediatric Crohn s disease: risk factors for postoperative recurrence. Am J Gastroenterol 2001;96: Faybush EM, Blanchard JF, Rawsthorne P, Bernstein CN. Generational differences in the age at diagnosis with IBD: genetic anticipation, bias, or temporal effects. Am J Gastroenterol 2002;97: Alic M. Inflammatory bowel diseases are diseases of higher socioeconomic status: dogma or reality? Am J Gastroenterol 2000;95: Lapidus A. The changing epidemiology of inflammatory bowel diseases. Acta Gastroenterol Belg 2001;64: Heikenen JB, Werlin SL, Brown CW, Balint JP. Presenting symptoms and diagnostic lag in children with inflammatory bowel disease. Inflamm Bowel Dis 1999;5: Meucci G, Bortoli A, Riccioli FA, Girelli CM, Radaelli F, Rivolta R, et al. Frequency and clinical evolution of indeterminate colitis: a retrospective multi-centre study in northern Italy. GSMII (Gruppo di Studio per le Malattie Infiammatorie Intestinali). Eur J Gastroenterol Hepatol 1999;11: Orholm M, Fonager K, Sorensen HT. Risk of ulcerative colitis and crohn s disease among offspring of patients with chronic inflammatory bowel disease. Am J Gastroenterol 1999;94: Tysk C, Lindberg E, Jarnerot G, Floderus-Myrhed B. Ulcerative colitis and Crohn s disease in an unselected population of monozygotic and dizygotic twins: a study of heritability and the influence of smoking. Gut 1988;29: Bayless TM, Tokayer AZ, Polito JM, Quaskey SA, Mellits ED, Harris ML. Crohn s disease: concordance for site and clinical type in affected family members: potential hereditary influences. Gastroenterology 1996;111: Michener WM, Farmer RG, Mortimer EA. Long-term prognosis of ulcerative colitis with onset in childhood or adolescence. J Clin Gastroenterol 1979;1: Heyman et al The Journal of Pediatrics January 2005

89 PEDIATRIC POISONING FROM TRIMEDOXIME (TMB4) AND ATROPINE AUTOMATIC INJECTORS ERAN KOZER,MD,AMNON MORDEL,MD,SHMUAEL BAR HAIM,MD,MORDECHEI BULKOWSTEIN,MD, MATITIAHU BERKOVITCH,MD,AND YEDIDIA BENTUR,MD Objective To describe the effects of combined trimedoxime (TMB4) and atropine poisoning from automatic injectors (AI) in children. Study design Data was collected from two sources: calls to the Israel Poison Information Center (IPIC) during a 1-year period and a cohort of children who presented to pediatric emergency departments (EDs) after unintentional injection of an AI. Demographic data and data regarding the type of AI, site and time of injection, and the clinical manifestations were abstracted. Results Data were available for 142 patients. The median age was 8.5 years (range years). The dose of atropine and TMB4 was higher than the recommended dose for age in 22 (15.5%) cases. There were few side effects attributable to atropine: dilated pupils (26.7%), dryness of mucous membranes (24.6%), and tachycardia (22.5%). Compared with children injected with an age-appropriate dose, children injected with an AI that contained a dose that exceeds the recommended one were more likely to be symptomatic (P =.029). There were no side effects characteristic to oximes, and no specific medical intervention was required. Conclusions Unintentional pediatric atropine and TMB4 injection, even an adult dose in a small child, does not cause significant side effects. (J Pediatr 2005;146:41-4) Nerve agents are organophosphates that inhibit acetylcholinesterase activity in the synapses. Pharmacological treatments for organophosphate intoxication are based on competitive inhibition of acetylcholine by agents such as atropine and by an attempt to restore acetylcholine esterase activity in the synapses by the use of oximes. In anticipation of a possible attack with nerve agents, chemical warfare defense kits were distributed to the Israeli population in 1991 during the Persian Gulf War; kits contained an automatic injector (AI) of atropine. Since then, the atropine AI was replaced by an AI that contains atropine and trimedoxime (TMB4), a cholinesterase reactivator oxime. The strategy held by the Medical Corps, Israel Defense Force, and the Israeli Ministry of Health is that mass casualty events can be best handled if standardized guidelines (including therapeutic regimens) are implemented. For this reason, AIs with three fixed doses of atropine sulphate and TMB4 were distributed for: (1) children up to 3 years of age (AI containing 0.5 mg atropine and 20 mg TMB4); (2) children 3 to 8 years of age and adults >60 years of age (1mg atropine and 40 mg TMB4); and (3) persons >8 years of age (2 mg atropine and 80 mg TMB4). The pediatric atropine dose for organophosphate poisoning is 0.05mg/kg/dose (range mg/kg). The two pediatric AIs contain atropine in an amount that will provide doses in this range. The doses of TMB4 were largely extrapolated from animal data. Atropine overdose may be associated with mortality and morbidity. Death in children after receiving relatively low doses atropine has been described. 1,2 However, a national survey conducted in Israel during the 1991 Gulf War found no life-threatening events in 268 cases of pediatric atropine injections. 3 More than 50% of the children in that series had minimal or no signs of atropinization, and only 8% had severe signs of atropinization. AI ED Automatic injectors Emergency department IPIC TMB4 Israel Poison Information Center Trimedoxime See editorial, p 8. From the Department of Emergency Medicine, and the Clinical Pharmacology and Toxicology Unit, Assaf Harofeh Medical Center, Sakler Faculty of Medicine, Tel Aviv University; and the Israel Poison Information Centre, The Bruce Rappaport Faculty of Medicine, Rambam Medical Center, Technion, Israel. Presented in part at the North American Congress of Clinical Toxicology, Chicago, Illinois, September 2003, and at the second Mediterranean Emergency Medicine Congress, Sitges/Barcelona, Spain, September Submitted for publication Mar 28, 2004; last revision Jul 2, 2004; accepted Aug 23, Reprint requests: Eran Kozer, MD, Assaf Harofeh Medical Center, Pediatric Emergency Services, Zerifin 70300, Israel. erank@asaf. health.gov.il /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

90 Table I. Atropinization score (adopted from Amitai et al 3 ) Score Pupils Normal 1 Dilated 2 Skin Normal 1 Flushed 2 Membranes Moist 1 Dry 2 Temperature C, Neurological state Normal 1 Abnormal (hallucinations, confusion, lethargy, agitation) 2 Heart rate (beats/min) Age,6 HR,140 or age.6 HR,120 1 Age,6 HR or age.6 HR Arrhythmia or HR.160 (all ages) 3 HR, heart rate. The objective of the current study was to describe the effects of combined TMB4 and atropine poisoning in children after unintentional injection of an AI. METHODS Data were collected from two sources: from the Israel Poison Information Center (IPIC) and from patients seen in emergency departments (Eds) in Israel. Calls to the IPIC regarding unintentional injections of AIs during the 1-year period before, during, and after the war in Iraq (August 1, 2002 to July 31, 2003) were prospectively collected. The IPIC is the national poison center of Israel and the only one that serves both the general public and the health system 24 hours a day. Data are recorded on a comprehensive structured form, which includes caller and patient demographic details; route, site, and circumstances of exposure; substance involved; and time elapsed to consultation. It also includes clinical manifestations in a detailed system-oriented approach, management, and follow-up recommendations. In addition, we conducted a retrospective chart review of children who presented to the EDs of two hospitals in Israel (Assaf Harofeh Medical Center and Ma aynei-hayeshua Hospital) during the same period as a result of unintentional injection from an AI that contained atropine and TMB4. Patients seen in these hospitals were not reported to the IPIC. Assaf Harofeh is a large regional Hospital near Tel Aviv with an active pediatric ED that treats approximately 17,000 patients per year. Ma aynei-hayeshua Hospital is a local hospital near Tel Aviv that treats a mainly orthodox Jewish population. Table II. Symptoms and signs observed in children after self-injection of atropine and TMB4 AI Symptom No. (%) Dilated pupils 38 (26.7) Local reactions 37 (26) Dry mucus membranes 35 (24.6) Tachycardia 32 (22.5) Flushed skin 21 (14.8) Abnormal neurological state 4 (2.8) Other * 5 (3.5) *Includes 2 patients with vomiting, 1 patient with headache and sore throat, 1 patient with fever, and 1 patient with abdominal pain. Patients were seen in the ED by the pediatrician on call, and a telephone consult with a toxicologist was offered when needed. Demographic data and data regarding the type of AI, site and time of injection, clinical manifestations, and treatment measures were abstracted from patients records. The severity of atropinization was based on the score suggested by Amitai et al 3 (Table I). Scores of 6 or 7 were considered as normal, 8 to 10 were considered mild, and scores of 11 to 13 were considered severe atropinization. The proportion of symptomatic children, ie, those with an atropinization score of $8, was compared using the x 2 test between children injected with an age-appropriate AI and those who received a dose that exceeds the recommended dose. Our institutional review board approved the study. RESULTS Injection from AI was identified in 148 patients; 122 patients were reported to the IPIC, and 26 were identified at the two hospitals studied (22 patients at Assaf Harofeh and additional 4 patients at Ma aynei-hayeshua), of which 25 charts were available for review. Five of the patients reported to the IPIC were excluded because of insufficient data. Demographic Data Of the 142 patients included, 101 (71%) were males. Age ranged from 1.25 years to 18 years (median 8.5, mean 8.9 ± 4.4 years). In 135 patients, the site of injection was in the upper limb (mainly fingers and palm). One patient was injected in the abdomen, and in 6 patients the site of injection was not reported; 96% of the injections occurred between 7 AM and 11 PM, with most cases occurring in the early afternoon. The median time from the injection until contact with the IPIC was 35 minutes (mean 70 ± 76 minutes, range 5 minutes to 5 hours). Patients were seen in the ED within 5 minutes to 5 hours after the injection (median 59 minutes, mean 63 ± 33 minutes). Atropine and TMB4 Dosing In all cases, the history and local findings confirmed that the AI needle pricked the patients. In several cases, however, the patients or parents mentioned that they saw some liquid coming out of the needle, which indicates that not all the syringe content was injected into the patient. 42 Kozer et al The Journal of Pediatrics January 2005

91 The dose in the AI was reported in 126 (89%) cases. The dose of atropine and TMB4, assuming complete injection of the syringe, was higher than the recommended dose for age in 22 (15.5%) patients. Clinical Effects The observed clinical manifestations are presented in Table II. Local reactions such as severe pain or significant swelling were observed in 37 (26%) cases. At least two signs of atropinization (score 8-10) were found in 42 (30%) of the patients. One patient had signs of severe atropinization (score 11-13). Compared with children injected with an age-appropriate dose, children injected with a dose that exceeds the recommended one were more likely to have an atropinization score $8 (9/22 vs 33/104, respectively; P =.029). None of the patients who were asymptomatic on arrival to the ED developed symptoms while in the ED. In 11 patients with electrocardiographic recordings, there were no abnormalities other than sinus tachycardia. Four patients had neurological abnormalities (2 patients with confusion, 1 patient with headache and agitation, and 1 patient with lethargy). There were no side effects characteristic to oximes: no patient demonstrated dizziness, nausea, or muscular weakness. Treatment In the ED, 79 patients were treated in ED s; 50 children were treated in ambulatory clinics, and 13 children were observed at home. Patients were observed in the ED for 2 to 6 hours after exposure if they were asymptomatic. For the 25 patients who were treated at the two hospitals studied, the median time from exposure to discharge was 3.5 hours (range hours). One patient, a 15-year-old boy who was confused on arrival, was observed in the ED overnight, and 1 patient was admitted for observation. None of the patients was treated with physostigmine. Four patients (all of them seen at one hospital) were treated with intravenous fluids. DISCUSSION We report unintentional pediatric exposure to atropine and TMB4; in the 142 cases reported, there were no serious adverse events. All the patients, except one, were treated as outpatients or were discharged home from the ED. Oximes are compounds that contain the radical R2C=N.OH derived from condensation of aldehydes or ketones with hydroxylamine. In organophosphate poisoning, oximes are used because of their ability to reactivate acetylcholinesterase at nicotinic receptors and, to a lesser degree, at muscarinic sites. 4,5 Other effects of oximes are inhibition of Na1 conductance, inhibition of various synaptic receptors, and reversible inhibition of acetylcholinesterase. 6 Members of the oxime group include pralidoxime, obidoxime, the H oximes HI-6 and Hlo-7, methoxime, and TMB4. With some organophosphate nerve agents, the binding of the warfare agent to acetylcholinesterase becomes irreversible (ageing) within a few minutes. If oximes were given after the ageing process had occurred, they would not be effective. Incorporating an oxime into an AI will increase the likelihood of reactivating acetylcholinesterase. The only oxime approved by the Food and Drug Administration for use in the United States is pralidoxime. The methanesulfonate salt of pralidoxime (P2S) is the standard oxime in the United Kingdom. Obidoxime (toxogonin) is used in Israel and in some European countries. In Israel, TMB4 is not in routine use for cases of organophosphate intoxication, but it has been incorporated into the AI found in personal chemical warfare defense kits. TMB4 is very stable in injectable solutions 7 and is therefore suitable for AI. Studies in animals 8-10 and clinical experience in humans 11,12 demonstrated that adding oximes to atropine may improve the outcome of poisoning from acetylcholinesterase inhibitors. Oximes may have different efficacy after intoxications from different organophosphates. 13,14 Adverse effects of oximes are rarely reported. 15,16 Dizziness or blurred vision and increased diastolic blood pressure were reported after rapid infusion of pralidoxime. 15 Obidoxime has been reported to be associated with pallor, nausea, headache, generalized weakness, sore throat, and paresthesia of the face muscles. 17 Arrhythmias and liver dysfunction were described in patients who received high doses of obidoxime and atropine. 18 It has been suggested, however, that the liver dysfunction in these cases is because of organophosphate intoxication not because of the treatment with obidoxime. 19 Compared with atropine alone, adding TMB4 to atropine increased survival rates in mice exposed to physostigmine. 9 The LD 50 for TMB4 in the presence of atropine, however, was much lower than the LD 50 of TMB4 alone, suggesting that the combination of atropine and TMB4 is more toxic. Data on adverse effects of TMB4 in humans are limited. Described adverse effects include heavy headedness, tension headache, facial flushing, dry mouth, and numbness in the face and limbs. 20 These adverse effects, which are in part similar to the effect of atropine, occurred when 150 mg of TMB4 was given intravenously. Because of these effects and despite being more effective than other oximes such as pralidoxime and pyridine-aldoxime-methylchloride in cases of organophosphate poisoning, some advocate not using TMB4 for routine treatment. 20 Our findings do not support these opinions. None of the children who accidentally received an AI that contained atropine and TMB4 developed severe adverse reaction. Even in the 22 case patients who received a higher dose there were no serious adverse effects. The current report also demonstrates the effects of mass distribution of AIs to the general population. Henretig et al 21 demonstrated that it is possible to discharge an AI containing atropine and pralidoxime into a small, sterile vial. The vial contents could than be used for intramuscular injection of these antidotes to small children on a milligram per kilogram basis. This approach might be considered when AI-containing pediatric doses are not available. The current study has some limitations. In some cases we could not determine the exact dose injected by the AI. Because Pediatric Poisoning From Trimedoxime (TMB4) And Atropine Automatic Injectors 43

92 data were reported to the IPIC from many sources and because some of the data were collected retrospectively not all patients underwent a standard evaluation. The reports may have been affected by different interpretations by physicians. The treatment also was not standardized, as is reflected by the fact that all 4 patients in one of the hospitals were treated with intravenous fluids, whereas none of the 22 patients in the other hospital required fluid therapy. The IPIC management recommendations do not include routine intravenous fluids but rather oral hydration, as needed. The relatively small size of the series does not allow us to draw a definite conclusion about the outcome of combined TMB4 and atropine injection in children. The authors wish to thank Mrs Vered Steiner, Israel Poison Information Center, Rambam Medical Center, for her assistance in data collection. REFERENCES 1. Morton HG. Atropine intoxication: its manifestations in infants and children. J Pediatr 1939;14: Heath WE. Death from atropine poisoning. BMJ 1950;2: Amitai Y, Almog S, Singer R, Hammer R, Bentur Y, Danon YL. Atropine poisoning in children during the Persian Gulf crisis: a national survey in Israel. JAMA 1992;268: Willems JL, De Bisschop HC, Verstraete AG, Declerck C, Christiaens Y, Vanscheeuwyck P, et al. Cholinesterase reactivation in organophosphorus poisoned patients depends on the plasma concentrations of the oxime pralidoxime methylsulphate and of the organophosphate. Arch Toxicol 1993; 67: Worek F, Backer M, Thiermann H, Szinicz L, Mast U, Klimmek R, et al. Reappraisal of indications and limitations of oxime therapy in organophosphate poisoning. Hum Exp Toxicol 1997;16: Fossier P, Tauc L, Baux G. Side effects of phosphorylated acetylcholinesterase reactivators on neuronal membrane and synaptic transmission. Pflugers Arch 1983;396: Rubnov S, Amisar S, Levy D, Muchtar S, Schneider H. Stability of trimedoxime in concentrated acidic injectable solutions. Mil Med 1999;164: Jokanovic M, Maksimovic M. A comparison of trimedoxime, obidoxime, pralidoxime and HI-6 in the treatment of oral organophosphorus insecticide poisoning in the rat. Arch Toxicol 1995;70: Klemm WR. Efficacy and toxicity of drug combinations in treatment of physostigmine toxicosis. Toxicology 1983;27: Harris LW, Stitcher DL, Heyl WC, Lieske CN, Lowe JR, Clark JH, et al. The effects of atropine-oxime therapy on cholinesterase activity and survival of animals intoxicated with p-nitrophenyl di-n-butylphosphinate. Toxicol Appl Pharmacol 1979;49: Lifshitz M, Rotenberg M, Sofer S, Tamiri T, Shahak E, Almog S. Carbamate poisoning and oxime treatment in children: a clinical and laboratory study. Pediatrics 1994;93: Kusic R, Jovanovic D, Randjelovic S, Joksovic D, Todorovic V, Boskovic B, et al. HI-6 in man: efficacy of the oxime in poisoning by organophosphorus insecticides. Hum Exp Toxicol 1991;10: Genovese RF, Doctor BP. Behavioral comparison of the oximes TMB- 4, 2-PAM, and HI-6 in rats using operant conditioning. Pharmacol Biochem Behav 1997;56: Worek F, Reiter G, Eyer P, Szinicz L. Reactivation kinetics of acetylcholinesterase from different species inhibited by highly toxic organophosphates. Arch Toxicol 2002;76: Medicis JJ, Stork CM, Howland MA, Hoffman RS, Goldfrank LR. Pharmacokinetics following a loading plus a continuous infusion of pralidoxime compared with the traditional short infusion regimen in human volunteers. J Toxicol Clin Toxicol 1996;34: Scott RJ. Repeated asystole following PAM in organophosphate selfpoisoning. Anaesth Intensive Care 1986;14: Simon GA, Tirosh MS, Edery H. Administration of obidoxime tablets to man: plasma levels and side reactions. Arch Toxicol 1976;36: Finkelstein Y, Kushnir A, Raikhlin-Eisenkraft B, Taitelman U. Antidotal therapy of severe acute organophosphate poisoning: a multihospital study Neurotoxicol Teratol 1989;11: Bentur Y, Raikhlin-Eisenkraft B, Singer P. Beneficial late administration of obidoxime in malathion poisoning. Vet Hum Toxicol 2003;45: Xue SZ, Ding XJ, Ding Y. Clinical observation and comparison of the effectiveness of several oxime cholinesterase reactivators. Scand J Work Environ Health 1985;11(suppl 4): Henretig FM, Mechem C, Jew R. Potential use of autoinjectorpackaged antidotes for treatment of pediatric nerve agent toxicity. Ann Emerg Med 2002;40: Kozer et al The Journal of Pediatrics January 2005

93 GROWTH HORMONE (GH) TREATMENT TO FINAL HEIGHT IN CHILDREN WITH IDIOPATHIC SHORT STATURE: EVIDENCE FOR A DOSE EFFECT JAN M. WIT, MD, PHD, LYSET T. M. REKERS-MOMBARG, PHD, GORDON B. CUTLER, JR, MD, BRENDA CROWE, PHD, TRACY J. BECK, PHD, KRISTEN ROBERTS,MSC, ANNE GILL,MSC,JEAN-LOUIS CHAUSSAIN, MD, HERWIG FRISCH,MD,PHD, RAFAEL YTURRIAGA,MD, AND ANDREA F. ATTANASIO, MD, ON BEHALF OF THE EUROPEAN IDIOPATHIC SHORT STATURE STUDY GROUP Objectives To investigate in an open-label randomized study, the effect of two doses of growth hormone (GH) on final height and height velocity during the first 2 years of treatment of children with idiopathic short stature (mean baseline height standard deviation score [SDS] 23.2). Study design Patients were treated with GH at 0.24 mg/kg/week, 0.24 mg/kg/week for the first year and at 0.37 mg/kg/week thereafter (0.24!0.37), or 0.37 mg/kg/week. Final height was evaluated in 50 patients at study completion (mean treatment duration, 6.5 years). Results Patients who received 0.37 mg/kg/week (n = 72) experienced a significantly greater increase in height velocity than those who received 0.24 mg/kg/week (n = 70) (treatment difference = 0.8 cm/year; P =.003) or 0.24!0.37 mg/kg/week (n = 67) (treatment difference = 0.9 cm/year; P =.001). For the 50 patients for whom final height measurements were available, mean height SDS increased by 1.55, 1.52, and 1.85 SDS, respectively, for the three dose groups. For the primary comparison between the 0.37 mg/kg/week and 0.24 mg/kg/week dose groups, the mean treatment difference (adjusted for differences in baseline predicted height SDS) was 0.57 SDS (3.6 cm; P =.025). Mean overall height gains (final height minus baseline predicted height) were 7.2 cm and 5.4 cm for the 0.37 mg/kg/week and 0.24 mg/kg/week dose groups, respectively, without dose effects on safety parameters. Final height measurements were within the normal adult height range for 94% of patients randomized to 0.37 mg/kg/week who continued to final height. Conclusion GH treatment dose-dependently increases height velocity and final height in children with idiopathic short stature. (J Pediatr 2005;146:45-53) See editorial, p 10. Many children have marked short stature without a defined etiology. They have normal growth hormone (GH) responses to stimulation testing and are said to have idiopathic short stature, 1-3 also called non-gh-deficient short stature. Studies of GH treatment of idiopathic short stature have shown a significant increase in height velocity during the first year, with lesser increases in subsequent years. 2-4 Doses of GH have varied from 0.17 to 0.70 mg/kg/week, and increases in height velocity appear to have been influenced by the dose and frequency of administration. 3-6 Although more than 30 studies have found significant GH treatment effects on height velocity in children with idiopathic short stature, fewer studies have followed patients to final height, and final height results have varied In a meta-analysis of studies published between 1985 and 2000, GH apparently increased adult height by 5 to 6 cm in four controlled studies and by approximately 4 cm in eight uncontrolled studies. 24 In the one study with randomized controls, GH-treated girls gained about 3.5 cm above predicted height and attained a mean adult height about 7.5 cm taller than that of untreated controls. 17 This randomized, dose-response study began in 1988 and included children from 10 countries. The dose-response effect of GH on height velocity and height standard deviation score (SDS) in the first 2 to 4 years of treatment have been reported for most GH SDS Growth hormone Standard deviation score TW2 Tanner Whitehouse II From Leiden University Medical Center, Leiden, the Netherlands; St Vincent de Paul Hospital, Paris, France; Vienna University Hospital, Vienna, Austria; Hospital Ramon and Cajal, Madrid, Spain; Eli Lilly and Company, Sesto Fiorentino, Florence, Italy, and Indianapolis, Indiana. G. B. Cutler, B. Crowe, T. J. Beck, K. Roberts, and A. Gill are employees and shareholders of Eli Lilly and Company, which sponsored this study. A. Attanasio is a consultant to and shareholder of Eli Lilly and Company. J. M. Wit, J.-L. Chaussain, H. Frisch, and R. Yturriaga were all principal investigators for this study. Submitted for publication Jan 6, 2004; last revision received Jul 12, 2004; accepted Aug 23, Reprint requests: Jan M. Wit, MD, Department of Pediatrics, Leiden University Medical Center, PO Box 9600, 2300 RC Leiden, the Netherlands. j.m.wit@lumc.nl /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

94 patients, 25 as have the final height data for the Dutch patients. 22 Additionally, studies of immunologic function, 26 pubertal timing and growth, 27 and quality-of-life 28 have been reported for small subsets of patients in this study. Here we present comprehensive long-term efficacy and safety data from this study, demonstrating that with appropriate dose and duration of GH treatment, most of the measured final height values for children with idiopathic short stature fell within the normal adult range. METHODS Patients This multi-center study enrolled 239 children from 10 countries (Austria, Finland, France, Germany, Greece, Israel, the Netherlands, Norway, Spain, and the United Kingdom). All children had idiopathic short stature, defined as: height more than 2 standard deviations (SD) below the population mean according to British standards, 29 bone age <10 years (female) or 12 years (male) according to the Tanner Whitehouse II (TW2) method, 30 and height velocity below the 25th percentile for age. British height standards were used because they were among the highest quality growth standards available when the study was designed, and they were generally representative of the European population. All children were prepubertal (Tanner growth stage 1) 31,32 but at least 5 years of age. Maximal plasma GH responses to standard stimulation tests, measured at each site s local laboratory, had to be >20 mu/l (approximately 10 mg/l). Patients had to have normal thyroid function or adequate replacement therapy and no organic cause of growth failure, primary bone disease, chronic illness, or dysmorphic syndrome. However, patients with a history of intrauterine growth retardation (n = 33) or Russell- Silver syndrome (n = 4) were not specifically excluded. The protocol was approved by appropriate ethical review boards. Before conducting any study-related procedure, the investigators obtained informed consent from the patient, a parent, or a guardian. Study Design The study consisted of a 2-year, three-arm, randomized, open-label, parallel, dose-response, height velocity phase and a final height extension phase. At study entry, 239 patients were randomized to receive one of three doses of recombinant human GH (HumatropeÒ, Eli Lilly and Company, Indianapolis, Ind), administered subcutaneously 6 days per week: (1) 0.24 mg/kg/week (n = 78); (2) 0.24 mg/kg/week for the first year and 0.37 mg/kg/week thereafter (0.24/0.37) (n = 78); or (3) 0.37 mg/kg/week (n = 83). The height velocity phase was a 2-year treatment period, during which we assessed changes in height, height velocity, and bone age. 29 After completion of the height velocity phase, 173 patients entered the extension phase, continuing on the same GH dose that they had received during the second year of the height velocity phase. Of these, 50 patients continued on the extension phase to final height. Exposure to study drug was assessed as the number of years from start of treatment until the last recorded on-study visit. Height was determined using Harpenden stadiometers (Holtain, Crymych, Great Britain), with each height determination being the average of four measurements made by a single observer at each clinic visit. A single central reader (Institute of Child Health, London, UK) performed bone age assessments according to TW2. 30 To calculate predicted final height, baseline TW2 bone ages were converted to Greulich- Pyle 33 estimations. TW2 bone age was determined for the Greulich-Pyle reference radiographs by a single reader (Dr G. Gilli, Olgahospital, Stuttgart, Germany), and linear regressions were then calculated for boys (Greulich- Pyle = TW ; r 2 = 0.997) and girls (Greulich-Pyle = TW ; r 2 = 0.993). Others have shown similar strong correlations between the two methods for boys and girls 0.5 to 18 years of age. 34 We estimated baseline predicted height by the method of Bayley- Pinneau 35 for those patients for whom bone age was within the range necessary to make such a prediction. Safety was assessed from reporting of any undesirable experience that occurred after informed consent was obtained, regardless of possible relationship to GH. In addition, fasting blood samples were obtained at each clinic visit for glucose and glycosylated hemoglobin measurements. Statistical Analyses For the purpose of analysis, we defined the following populations: (1) All Randomized Patients population (n = 239): all patients who were randomized to one of the three dose groups, regardless of whether or not they received any treatment; (2) Two-Year Height Velocity population (n = 209): all patients who were randomized and completed the 2-year height velocity phase of the protocol; and (3) Final Height population (n = 50): those patients who had received the study drug and had final height measured after height velocity had decreased to <2.0 cm/year. Target height (sex-adjusted midparental height) for each patient was estimated from the heights of the parents by standard methods. 36 Height SDS was calculated on the basis of the Prader growth standards. 37 These standards were used because they were considered the most recent standards of high quality that were also reasonably representative of Europe as a whole. We analyzed within-group changes from baseline by t tests of the null hypothesis that the change was equal to zero. Between-group differences were assessed based on type 3 sums of squares from the relevant statistical model. The primary efficacy analysis was an analysis of covariance of the difference in height velocity increase between the 0.24 mg/kg/week and 0.37 mg/kg/week dose groups in the Two-Year Height Velocity population. We analyzed the final height SDS treatment difference among groups for patients in the Final Height population using a general linear model, which adjusted comparisons for baseline predicted height SDS in those patients with baseline 46 Wit et al The Journal of Pediatrics January 2005

95 Table I. Baseline characteristics of All Randomized Patients and Final Height patient populations Patient baseline characteristics and duration of GH treatment GH dose group (mg/kg/wk) / P-value All Randomized Patients Number of patients Male/Female (%) 63 / / / Age (y) 9.43 ± ± ± Height (cm) ± ± ± Height SDS ± ± ± Predicted height SDS 22.7 ± ± ± Bone age (y) 7.40 ± ± ± Height velocity (cm/y) 4.3 ± ± ± Duration of treatment (y) 4.7 ± ± ± Final Height Number of patients Male/Female (%) 65 / / / Age (y) ± ± ± Height (cm) ± ± ± Height SDS ± ± ± Predicted height SDS 22.5 ± ± ± Bone age (y) 8.47 ± ± ± Height velocity (cm/y) 4.7 ± ± ± Duration of treatment (y) 6.1 ± ± ± Except for patient number and sex distribution, data are mean ± SD. predicted height available (n = 26), and which made unadjusted comparisons in those without this value (n = 8). To address the issue of potential dropout bias, modified intent-to-treat analyses (for the Two-Year Height Velocity population) and intent-to-treat analyses (for the All Randomized Patients population) of final or last observed height SDS were performed, using the general linear model specified earlier. Additionally, for the All Randomized Patients population, we performed two nonparametric analyses of last observed height SDS: a rank analysis of covariance with baseline predicted height SDS as the covariate and a generalized Wilcoxon-Mann-Whitney test. Mean final height SDS also was estimated by fitting a repeated measures model to the available height SDS values at ages 10 to 18 years for the Two-Year Height Velocity population. This model allowed for different growth patterns in each dose group. The model included categorical terms for sex and age group (age rounded to the nearest year) and continuous terms for baseline height SDS and baseline predicted height SDS. Interaction terms for baseline age 3 treatment, sex 3 age group, and treatment 3 age group were included. The treatment effect on final height was obtained from the difference in least squares mean height SDS at 18 years of age, between the 0.24 mg/kg/week and the 0.37 mg/ kg/week dose groups. Onset of puberty was analyzed for males and females separately, using survival analysis (Kaplan-Meier) methods. For the analysis of pubertal pace, patients were included if they became pubertal (attained Tanner growth stage 2) during the study. We included data from the first 4 years after onset of puberty. Pubertal pace was assessed by repeated-measures analysis. Correlation between observations on the same patient was assumed to decline with increasing time between the observations. Response variables assessed were Tanner stage, bone age, and testicular volume. We analyzed all safety data for the All Randomized Patients population. A comparison of adverse event frequencies among the dose groups was made using Fisher s exact test. Between-group comparisons of glucose and glycosylated hemoglobin were performed using t testing. Except where otherwise stated, results are expressed as mean ± standard error. The significance level was set at.05 (two-sided) for all analyses. Where possible, analyses included all three dose groups. However, as the comparison between the 0.37 mg/kg/week and the 0.24 mg/kg/week dose groups was considered primary, some comparisons involving the intermediate dose group have been omitted for clarity of presentation. RESULTS Patient Disposition Of the 239 patients who started GH treatment, 209 completed the initial 2-year height velocity phase, and 50 were followed to final height. The main reason for discontinuation was decision of the patient, investigator, or sponsor (n = 134, Growth Hormone (GH) Treatment To Final Height In Children With Idiopathic Short Stature: Evidence For A Dose Effect 47

96 Figure 1. Mean height SDS for patients with adult height measurements (Final Height population) and for patients without adult height measurements (Non-Final Height subgroup). The number of patients at each time point is indicated for each dose group (D1: 0.24 mg/kg/week; D2: 0.24/0.37 mg/kg/week; D3: 0.37 mg/kg/week). Symbols are 0.24 mg/kg/week (d), 0.24/0.37 mg/kg/week (:), and 0.37 mg/kg/week (r). 56%). Nine patients (3.8%) discontinued because of perceived lack of efficacy, and 3 patients (1.3%) discontinued because of adverse events. Other reasons for discontinuation were: lost to follow-up, protocol violations, or lack of compliance with treatment. There were no significant differences among dose groups in the reasons for discontinuation and no significant difference among dose groups in the time to discontinuation. Baseline Data The baseline characteristics for the All Randomized Patients and Final Height populations did not differ significantly among the three dose groups (Table I), except for height SDS and predicted height SDS within the All Randomized Patients population (and also within the Two- Year Height Velocity population, data not shown). Because baseline predicted height is a strong predictor of final height, most of the statistical analyses adjusted for baseline predicted height SDS. Height Velocity Data At completion of the 2-year height velocity phase, the patient group that received 0.37 mg/kg/week had experienced a significantly greater increase in height velocity than the groups that had received either 0.24 mg/kg/week or 0.24/0.37 mg/kg/week. The mean height velocity of the patients who received 0.37 mg/kg/week (n = 71) increased from 4.4 ± 0.1 to 8.4 ± 0.2 cm/year from baseline to the 2-year time point, a significantly greater increase than that for patients who received 0.24 mg/kg/week (n = 68) (4.2 ± 0.1 to 7.5 ± 0.2 cm/year) (treatment difference = 0.8 cm/year; P =.003) or 0.24/0.37 mg/kg/week (n = 66) (4.5 ± 0.1 to 7.6 ± 0.2 cm/year) (treatment difference = 0.9 cm/year; P =.001). Final Height Data There were no significant differences among dose groups for duration of treatment (Table I). The mean ages of the Final Height population dose groups, at final height measurement, were: 18.1 ± 2.1, 17.7 ± 1.5, and 17.8 ± 1.7 years for the 0.24 mg/kg/week, 0.24/0.37 mg/kg/week, and 0.37 mg/kg/week dose groups, respectively. After treatment for a mean period of 6.5 years, and at a mean age of 18 years, the mean between-dose effect (0.24 mg/kg/week vs 0.37 mg/kg/ week) on final height SDS was 0.57 ± 0.25 SDS (3.6 cm; P =.025; Table II). There were no significant differences 48 Wit et al The Journal of Pediatrics January 2005

97 Table II. Dose effects (0.24 mg/kg/week vs 0.37 mg/kg/week) of GH treatment for idiopathic short stature on final height SDS in the Final Height population, and on last observed height SDS in the Two-Year Height Velocity and All Randomized Patients populations Final or last observed height SDS (least squares mean ± SE) GH dose group (mg/kg/wk) Dose effect y Population and analysis N (n) * / SDS cm P value z Final Height General linear regression 50 (34) ± ± ± ± Two-Year Height Velocity General linear regression 209 (142) ± ± ± ± Repeated measures 138 (86) ± ± ± ± All Randomized Patients General linear regression 239 (161) ± ± ± ± *N=number of patients in analysis for all three dose groups; n = number of patients for 0.24 mg/kg/wk and 0.37 mg/kg/wk dose groups. yvalue represents the difference in final or last observed height SDS between the 0.24 mg/kg/wk and 0.37 mg/kg/wk dose groups, with the corresponding value in cm. zp value for 0.24 mg/kg/wk vs 0.37 mg/kg/wk dose groups. Repeated measures linear model of height SDS at 18 years of age. The repeated measures analysis, which required baseline predicted height SDS as a covariate, included fewer patients because some patients had baseline bone ages too young for Bayley-Pinneau height prediction. among the three dose groups in the rate of bone age progression or in the onset or pace of puberty (data not shown). Because many patients discontinued before attaining final height, additional analyses were performed in the Two- Year Height Velocity and the All Randomized Patients populations to confirm the robustness of the dose effect (Table II). Two nonparametric intent-to-treat analyses in the All Randomized Patients population also confirmed the significantly greater last observed height SDS in the 0.37 mg/kg/ week dose group (P =.001, rank analysis of covariance; and P =.0004, generalized Wilcoxon-Mann-Whitney test). To examine further whether dropout bias contributed to doserelated height gains observed in the Final Height population, we compared yearly mean height SDS for the Final Height population and for patients without a final height measurement (Non-Final Height subgroup) (Figure 1). At each year of study, the differences in height SDS between the 0.24 mg/ kg/week and 0.37 mg/kg/week dose groups for the Non-Final Height subgroup were similar to or greater than those observed in the Final Height population, confirming that the dose-related treatment difference observed in the Final Height population cannot be explained by a dropout bias. Moreover, values of mean height SDS for the 0.37 mg/kg/ week dose groups were similar at years 3 through 6 for the Final Height population and the Non-Final Height subgroup. Within-group analyses of the gain in height SDS (final height SDS minus baseline height SDS) showed highly significant increases in height SDS from baseline to final height for each dose group: 1.55 ± 0.14 SDS, 1.52 ± 0.27 SDS, and 1.85 ± 0.20 SDS for the 0.24 mg/kg/week, 0.24/0.37 mg/kg/week, and 0.37 mg/kg/week dose groups, respectively (all P <.001 for within-group t test of null hypothesis that mean value equals zero). Furthermore, the patients who received the 0.37 mg/kg/week dose came close to achieving their genetic height potential, as the difference between final height and target (midparental) height was not significant ( 1.3 ± 1.2 cm; P =.288). As a measure of overall within-group treatment effect, the difference between final height and baseline predicted height was 5.4 ± 0.9 cm (P #.001) for the 0.24 mg/kg/week dose group and 7.2 ± 1.7 cm (P #.001) for the 0.37 mg/kg/ week dose group (Figure 2A). At 0.24 mg/kg/week, 71% of final height SDS values were within the normal range (22.0 to 12.0), whereas at 0.37 mg/kg/week, 94% of final height SDS values (all but one) were within the normal range (Figure 2B). The one patient whose final height SDS was below normal had a gain in height SDS of approximately 1 SDS during treatment. Safety Adverse events were reported for 162 patients (Table III), with no significant differences among dose groups. The majority of adverse events represented common childhood conditions, with the most frequent being infection (18.0%), pharyngitis (14.2%), and flu syndrome (10.5%). Serious adverse events occurred in 31 patients, but most were isolated incidences such as routine pediatric surgical procedures (eg, appendectomy) and hospitalizations for injuries. Three patients discontinued because of unexpected adverse events (2 serious) that were classified by the investigator as possibly related to study treatment. The events were: (1) hospitalization for a slipped capital femoral epiphysis following a fall during an epileptic seizure in a 16-year-old male with a history of epilepsy who had been taking GH at 0.37 mg/kg/week for 5.3 years; (2) impaired glucose tolerance in a 13-year-old female after 8.4 years of GH treatment at 0.24 Growth Hormone (GH) Treatment To Final Height In Children With Idiopathic Short Stature: Evidence For A Dose Effect 49

98 Figure 3. Individual and mean ± SE fasting glucose levels at baseline and endpoint for patients in the 0.24 mg/kg/week (n), 0.24/0.37 mg/kg/week (;), and 0.37 mg/kg/week (r) dose groups showed no GH dose effect. Figure 2. (A) Mean ± SE final height minus baseline predicted height (cm) in the Final Height population for patients in the 0.24 mg/kg/week, 0.24/0.37 mg/kg/week, and 0.37 mg/kg/week dose groups who had bone ages old enough for Bayley-Pinneau height prediction. P values are for within-group t test of the null hypothesis that the mean value equals zero. (B) Individual and mean ± SE final height SDS values for patients in the 0.24 mg/kg/week (n), 0.24/0.37 mg/kg/week (;), and 0.37 mg/kg/week (r) dose groups showed that most GH-treated patients for whom final height measurements were available achieved a normal final height (22.0 to 12.0 SDS). The dotted line represents the lower limit of the normal range. The ranges for final height SDS are shown for each dose group. Prader standards, dependent on age and sex, are available for the conversion of SDS to cm. To convert height SDS at age 18 years to cm, for males, multiply height SDS by 6.7 cm and add cm; for females, multiply height SDS by 5.8 cm and add cm. 37 mg/kg/week, with glycosylated hemoglobin of 6.3% (normal reference: 2.0%-6.0%) and blood glucose at 2 hours in an oral glucose tolerance test of >11 mmol/l (>198 mg/dl) (glycosylated hemoglobin was normal when measured 1 year later); and (3) an intraabdominal desmoplastic small-cell tumor with triple (muscular, epithelial, and neuroendocrine) differentiation in a 12-year-old-male, after 6.5 years of GH treatment at 0.24 mg/kg/week. Despite surgery and chemotherapy, this patient died 4 years after discontinuation from the study. Genetic analysis of tumor tissue revealed the characteristic chromosomal translocation associated with this rare tumor (46, XY, t[11;22][p13;q12]). There were no significant changes in mean fasting blood glucose from baseline to study endpoint for any dose group (Figure 3). Fasting glucose concentrations above the upper limit of normal (7 mmol/l; 126 mg/dl) were seen in 9 patients among the different GH doses (0.24 mg/kg/week: n = 2; 0.24/0.37 mg/kg/week: n = 3; 0.37 mg/kg/week: n = 4); in each case this occurred on a single occasion and with normal levels of glycosylated hemoglobin. Mean glycosylated hemoglobin concentration also did not change significantly from baseline to endpoint for any dose group (data not shown). In a single patient, glycosylated hemoglobin increased from below 6% to 7.4% to 10.3% during three visits in the second year of study. Fasting blood glucose levels for this patient were normal; the patient was not followed after stopping treatment at completion of the 2-year protocol. DISCUSSION The present study establishes a dose-response relationship for the effect of GH treatment on height velocity and final height in children with idiopathic short stature. The dose of 0.37 mg/kg/week produces a greater increase in height velocity and a greater final height (by approximately 3 cm) than the dose of 0.24 mg/kg/week. 25 The robustness of the dose effect on final height is supported by a similar effect on last observed height among all randomized patients and among patients who completed at least 2 years of treatment. The overall effect of the 0.37 mg/kg/week dose, which was estimated by comparing final height with the height that patients were predicted to achieve without treatment, averaged 7.2 cm. Because studies of untreated patients with idiopathic short stature have shown that, on average, they fail to achieve their baseline predicted height, 38,39 the estimate of 7.2 cm is likely to be conservative. As a result of this increase, 94% of final heights for patients receiving the 0.37 mg/kg/week dose were within the normal adult range, compared with 71% for patients receiving the 0.24 mg/kg/week dose, and 56% for 50 Wit et al The Journal of Pediatrics January 2005

99 Table III. Number and percent of randomized patients with idiopathic short stature who experienced at least one adverse event at some time during GH treatment GH dose group (mg/kg/wk) Adverse event category 0.24 n = /0.37 n = n = 83 Total N = 239 Death 0 * Treatment-emergent adverse event y 47 (60.3) 57 (73.1) 58 (69.9) 162 (67.8) Discontinuation because of an adverse event 2 (2.6) 0 1 (1.2) 3 (1.3) Nonserious clinically significant adverse event z 8 (10.3) 14 (17.9) 14 (16.9) 36 (15.1) Serious adverse event 11 (14.1) 4 (5.1) 16 (19.3) 31 (13.0) Serious, possibly drug-related adverse event 1 (1.3) 0 1 (1.2) 2 (0.8) *One patient with a desmoplastic small round cell tumor died 4 years after discontinuation from this study. ya treatment-emergent adverse event is any adverse event that was not present at baseline or any pre-existing condition or event present at baseline that increased in severity during the study. za nonserious clinically significant adverse event is a treatment-emergent adverse event that has been reported in association with GH exposure, although not necessarily in a causal manner. A serious adverse event is an event that results in one of the following: death, initial or prolonged inpatient hospitalization, life-threatening consequences, severe or permanent disability, cancer, or other significant consequence. patients who changed to the higher dose after 1 year at the lower dose. Based on these findings, initiation of treatment at the dose of 0.37 mg/kg/week is recommended to increase the likelihood of achieving a final height within the normal range. A potential limitation of this study is the lack of final height measurements for many randomized patients. Although not surprising for a study involving long-term, daily injections in children, this limitation could lead to dropout bias if the dose effect on height SDS differed between patients with and without final height measurements. To address this issue, intent-to-treat and modified intent-to-treat analyses of final or last observed height SDS were performed. All analyses supported the greater effect of the 0.37 mg/kg/ week dose compared with the 0.24 mg/kg/week dose, and the magnitude of the dose effect on final or last observed height SDS was similar in each analysis population. These observations provide strong evidence against an important dropout bias in the GH dose effect on final height SDS. The height gains relative to baseline predicted height, 17,18,24 and the increases in height SDS from baseline to final height 6,7,17,24 were consistent with those reported in other studies. Additionally, the dose effect on final height appears consistent with what would be predicted from the dose effect on initial height velocity, suggesting that the initial treatment effect may be important for an optimal final height outcome. Furthermore, the initial brisk growth rate is important in allowing short children treated with GH the opportunity to catch up with their peers during childhood. Predicting the long-term response to GH treatment with sufficient accuracy to be useful clinically has thus far proven difficult. In the Dutch patients of this study, GH dose and bone age delay were predictors of overall final height gain (final height minus baseline predicted height), but the prediction model explained only half of the between-subject variability (r 2 = 0.50). 22 Other investigators have had similar difficulty in predicting long-term response to treatment. 18,20 Despite the difficulty predicting individual responses, the outcome of treatment with the 0.37 mg/kg/week dose was excellent, with 94% of final height values falling within the normal adult range. The safety profile throughout the study was generally consistent with what would be expected from a pediatric population and the known rare adverse events associated with GH treatment, and there were no apparent dose-related effects. 40,41 One male patient developed an intraabdominal desmoplastic small round cell tumor while receiving the lower GH dose. This malignancy occurs most often in adolescent males and is associated with a specific chromosomal translocation, encoding a highly oncogenic transcription factor believed to be responsible for tumor development. 42 The tumor from this patient exhibited this characteristic translocation. In this study, the effect of GH on carbohydrate metabolism was similar to that in previous reports of patients with idiopathic short stature 43 or patients with Ullrich-Turner syndrome. 44,45 In the present study, mean fasting glucose and glycosylated hemoglobin concentrations did not change during treatment, and there were no cases of diabetes. One patient, who had received the lower dose for 8.4 years, discontinued the study because of impaired glucose tolerance; however, a glycosylated hemoglobin measurement performed 1 year after study discontinuation was normal. Serum insulin was not measured in this study. Although GH has a physiologic anti-insulin effect with respect to carbohydrate metabolism, reports have indicated that GH has no clinically relevant effect on fasting insulin levels in children with idiopathic short stature. 23,43 In particular, Leschek et al demonstrated that mean fasting insulin concentrations were not significantly different between GH- and placebo-treated groups across the duration of a randomized, double-blind, placebo-controlled study of patients with idiopathic short stature. 23 Present findings indicate that with the GH dose of 0.37 mg/kg/week, most final height values for children having Growth Hormone (GH) Treatment To Final Height In Children With Idiopathic Short Stature: Evidence For A Dose Effect 51

100 idiopathic short stature will fall within the normal adult range. The overall safety profile for GH treatment of these patients is similar to that for other pediatric indications. The authors would like to thank Dr Peter C. Bates, Cambridge Medical Writing Services, Ickleton, United Kingdom, Dr Christopher S. Konkoy, and Dr Anne M. Wolka for help in preparing the manuscript. The other principle investigators of the European Idiopathic Short Stature Study Group are: W. Stoegmann, MD, W. Tulzer, MD, Austria; M.-J. Kaar, MD, H. L. Lenko, MD, Finland; J.-C. Job, MD, PhD, F. Girard, MD, J. Roger, MD, M. L. A. Pierson, MD, P. Chatelain, MD, PhD, M. Colle, MD, France; U. Irle, MD, M. B. Ranke, MD, PhD, O. Butenandt, MD, PhD, R. Muehlenberg, MD, H. J. Boehles, MD, G. von Kalckreuth, MD, W. G. Sippell, MD, PhD, Germany; C. Dacou-Voutetakis, MD, PhD, Greece; E. Leiberman, MD, Israel; D. Aarskog, MD, PhD, Norway; M. A. Preece, MD, J. M. H. Buckler, MD, P. H. W. Rayner, MD, D. A. Price, MD, United Kingdom. REFERENCES 1. Ranke MB. Towards a consensus on the definition of idiopathic short stature: summary. Horm Res 1996;45(suppl 2): Drug and Therapeutics Committee of the Lawson Wilkins Pediatric Endocrine Society: guidelines for the use of growth hormone in children with short stature. J Pediatr 1995;127: Guyda HJ. Growth hormone treatment of non-growth hormone deficient subjects: the International Task Force Report. Clin Pediatr Endocrinol 1996;5(suppl 7): Hindmarsh PC, Pringle PJ, Di Silvio L, Brook CG. Effects of 3 years of growth hormone therapy in short normal children. Acta Paediatr Scand 1990; 66(suppl 3): Barton JS, Gardiner HM, Cullen S, Hindmarsh PC, Brook CGD, Preece MA. The growth and cardiovascular effects of high dose growth hormone therapy in idiopathic short stature. Clin Endocrinol 1995;42: Hopwood NJ, Hintz RL, Gertner JM, Attie KM, Johanson AJ, Baptista J, et al. Growth response of children with non-growth-hormone deficiency and marked short stature during three years of growth hormone therapy. J Pediatr 1993;123: Guyda HJ. Four decades of growth hormone therapy for short children: what have we achieved? J Clin Endocrinol Metab 1999;84: Zadik Z, Mira U, Landau H. Final height after growth hormone therapy in peripubertal boys with a subnormal integrated concentration of growth hormone. Horm Res 1992;37: Loche S, Cambiaso P, Setzu S, Carta D, Marini R, Borrelli P, et al. Final height after growth hormone therapy in non-growth hormone-deficient children with short stature. J Pediatr 1994;125: Hindmarsh PC, Brook CGD. Final height of short normal children treated with growth hormone. Lancet 1996;348: Lopez-Siguero JP, Martinez-Aedo MJ, Moreno-Molina JA. Final height after growth hormone therapy in children with idiopathic short stature and a subnormal growth rate. Acta Paediatr Suppl 1996;417:121; Abstract. 12. Zadik Z, Segal N, Limony Y. Final height prediction models for pubertal boys. Acta Paediatr Suppl 1996;417: Bernasconi S, Street ME, Volta C, Mazzardo G, Italian Multicentre Study Group. Final height in non-growth hormone-deficient children treated with growth hormone. Clin Endocrinol 1997;47: Schmitt K, Blumel P, Waldhor T, Lassi M, Tulzer G, Frisch H. Shortand long-term (final height) data in children with normal variant short stature treated with growth hormone. Eur J Pediatr 1997;156: Zadik Z, Zung A. Final height after growth hormone therapy in short children: correlation with siblings height. Horm Res 1997;48: Buchlis JG, Irizarry L, Crotzer BC, Shine BJ, Allen L, MacGillivray MH. Comparison of final heights of growth hormone-treated vs untreated children with idiopathic growth failure. J Clin Endocrinol Metab 1998;83: McCaughey ES, Mulligan J, Voss LD, Betts PR. Randomised trial of growth hormone in short normal girls. Lancet 1998;351: Hintz RL, Attie KM, Baptista J, Roche A. Effect of growth hormone treatment on adult height in children with idiopathic short stature. N Engl J Med 1999;340: Lanes R. Growth velocity, final height and bone mineral metabolism of short children treated long-term with growth hormone. Curr Pharm Biotechnol 2000;1: Lopez-Siguero JP, Garcia-Garcia E, Carralero I, Martinez-Aedo MJ. Adult height in children with idiopathic short stature treated with growth hormone. J Pediatr Endocrinol Metab 2000;13: Pasquino AM, Pucarelli I, Roggini M, Segni M. Adult height in short normal girls treated with gonadotropin-releasing hormone analogs and growth hormone. J Clin Endocrinol Metab 2000;85: Wit JM, Rekers-Mombarg LT. Dutch Growth Hormone Advisory Group. Final height gain by GH therapy in children with idiopathic short stature is dose dependent. J Clin Endocrinol Metab 2002;87: Leschek EW, Rose SR, Yanovski JA, Troendle JF, Quigley CA, Chipman JJ, et al. Effect of growth hormone treatment on adult height in peripubertal children with idiopathic short stature: A randomized, doubleblind, placebo-controlled trial. J Clin Endocrinol Metab 2004;89: Finkelstein BS, Imperiale TF, Speroff T, Marrero U, Radcliffe DJ, Cuttler L. Effect of growth hormone therapy on height in children with idiopathic short stature: a meta-analysis. Arch Pediatr Adolesc Med 2002;156: Rekers-Mombarg LTM, Massa GG, Wit JM, Matranga AMC, Bucker JMH, Butenandt O, et al. Growth hormone therapy with three dosage regimens in children with idiopathic short stature. J Pediatr 1998;132: Rekers-Mombarg LTM, Rijkers GT, Massa GG, Wit JM. Dutch Growth Hormone Working Group: immunologic studies in children with idiopathic short stature before and during growth hormone therapy. Horm Res 1995;44: Rekers-Mombarg LTM, Kamp GA, Massa GG, Wit JM. Dutch Growth Hormone Working Group: influence of growth hormone treatment on pubertal timing and pubertal growth in children with idiopathic short stature. J Pediatr Endocrinol 1999;12: Rekers-Mombarg LTM, Busschbach J, Massa GG, Dicke J, Wit JM. Dutch Growth Hormone Working Group: quality of life of young adults with idiopathic short stature: effect of growth hormone treatment. Acta Paediatr 1998;87: Tanner JM, Whitehouse RH, Takaishi M. Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965 part II. Arch Dis Child 1966;41: Tanner JM, Whitehouse RH, Cameron N, Marshall WA, Healy MJR, Goldstein H. Assessment of Skeletal Maturity and Prediction of Adult Height (TW2-method). 2nd ed. London: Academic Press; Marshall WA, Tanner JM. Variations in pattern of pubertal changes in girls. Arch Dis Child 1969;44: Marshall WA, Tanner JM. Variations in the pattern of pubertal changes in boys. Arch Dis Child 1970;45: Greulich WW, Pyle SI. Radiographic Atlas of Skeletal Development of the Hand and Wrist. 2nd ed. Stanford, Calf: Stanford University Press; Ruiz-Echam M, Rueda C, Labena C, Labarta J, Mayayo E, Ferrandez Longas A. Spanish longitudinal study of growth and development: bone maturation according to Greulich-Pyle and Tanner-Whitehouse 2-RUS methods from 6 months until 18 years of age. Presented at the 41st meeting of the European Society for Paediatric Endocrinology; September Bayley N, Pinneau SR. Tables for predicting adult height from skeletal age: revised for use with the Greulich-Pyle hand standards. J Pediatr 1952;40: Tanner JM, Whitehouse RH, Marshall WA, Carter BS. Predictions of adult height from height, bone age, and occurrence of menarche, at ages 4 to 16 with allowance for midparent height. Arch Dis Child 1975;50: Prader A, Largo RH, Molinari L, Issler C. Physical growth of Swiss children from birth to 20 years of age: First Zurich Longitudinal Study of Growth and Development. Helv Paediatr Acta 1989;(suppl 52): Wit et al The Journal of Pediatrics January 2005

101 38. Hintz RL, Attie KM, Baptista J, Roche A. Effect of growth hormone treatment on adult height in children with idiopathic short stature. N Engl J Med 1999;340: Bramswig JH, Fasse M, Holthoff ML, von Lengerke HJ, von Petrykowski W, Schellong G. Adult height in boys and girls with untreated short stature and constitutional delay of growth and puberty: accuracy of five different methods of height prediction. J Pediatr 1990;117: Maneatis T, Baptista J, Connelly K, Blethen S. Growth hormone safety update from the National Cooperative Growth Study. J Pediatr Endocrinol 2000;13(suppl 2): Wilton P. Adverse events during GH treatment: 10 years experience in KIGS: a pharmacoepidemiological survey. In: Ranke MB, Wilton P, eds. Growth Hormone Therapy in KIGS 10 Years Experience. Heidelberg: Barth; p Gerald WL, Ladany M, de Alava E, Cuatrecasas M, Kushner BH, LaQuaglia MP, et al. Clinical, pathologic, and molecular spectrum of tumors associated with t(11;22)(p13;q12): desmoplastic small round-cell tumor and its variants. J Clin Oncol 1998;16: Saenger P, Attie KM, DiMartino-Nardi J, Hintz R, Frahm L, Frane JW. Metabolic consequences of 5-year growth hormone (GH) therapy in children treated with GH for idiopathic short stature. J Clin Endocrinol Metab 1998;83: Quigley CA, Crowe BJ, Anglin DG, Chipman JJ. Growth hormone and low dose estrogen in Turner syndrome: results of a United States multi-center trial to near-final height. J Clin Endocrinol Metab 2002;87: Ranke MB, Partsch CJ, Lindberg A, Dorr HG, Bettendorf M, Hauffa BP, et al. Adult height after GH therapy in 188 Ullrich-Turner syndrome patients: results of the German IGLU Follow-up Study Eur J Endocrinol 2002;147: Years Ago in The Journal of Pediatrics PNEUMATOSIS CYSTOIDES INTESTINALIS IN INFANCY Paris L. J Pediatr 1955;46:1-17. Despite advances in neonatal care, the incidence of necrotizing enterocolitis (NEC) has remained unchanged over the last three decades. Mortality remains high at 25% to 30% of affected infants. The radiographic and pathologic hallmark of NEC is pneumatosis intestinalis. The report from Leonard Paris in 1955 provided detailed pathologic findings from a 5- month-old infant diagnosed with pneumatosis cystoid intestinalis. He summarized epidemiologic findings from the literature at the time as they pertained to infants, and he offered a theory to NEC s pathogenesis. There are clear differences in the epidemiologic characteristics of affected infants from those we find today, reflecting improvements in neonatal care. At the same time, it is humbling to find that progress in understanding pathophysiology has not progressed in a similar fashion. The cases of pneumatosis intestinalis in Paris s report were term or near-term infants who were presented at 2 to 14 weeks of age, some with underlying pathology such as congenital megacolon and some without. Today, prematurity is the single most important risk factor in the development of pneumatosis intestinalis, with gestational age indirectly related to incidence. Term infants comprise only approximately 10% of infants with NEC. Despite this striking difference, the similarities also are apparent. No infant is born with NEC, intestinal inflammation is always present, and mucosal integrity has been altered. Clues from these similarities have been applied to understanding the pathogenesis. Paris summarized the major theories of his time regarding pneumatosis intestinalis: mechanical, related to the break in mucosal integrity; bacterial, causing mucosal ulceration, and; chemical, from defective carbohydrate metabolism. These same themes of mucosal integrity, bacterial colonization, and the presence of substrate continue at the core of current theories. Despite advances in laboratory research, we have yet to identify the unknown factor hypothesized by Paris that sets into motion a final common mechanism to produce disease. Barbara Warner, MD Division of Neonatology Cincinnati Children s Hospital Medical Center Cincinnati, OH YMPD /j.jpeds Growth Hormone (GH) Treatment To Final Height In Children With Idiopathic Short Stature: Evidence For A Dose Effect 53

102 PREVALENCE OF DIARRHEAGENIC ESCHERICHIA COLI IN ACUTE CHILDHOOD ENTERITIS: A PROSPECTIVE CONTROLLED STUDY MITCHELL B. COHEN 1 MD, JAMES P. NATARO 2 MD, PHD, DAVID I. BERNSTEIN 3 MD, JENNIFER HAWKINS 1 BA, NANCY ROBERTS 3 MS, AND MARY A. STAAT 3 MD, MPH Objective Since diarrheagenic E. coli are not identified by common clinical laboratory techniques, we hypothesized that these organisms might be an unrecognized cause of enteritis in children in the U.S. Study design 1327 children with acute gastroenteritis were identified prospectively by active surveillance in the Emergency Department (ED) and the inpatient units at Cincinnati Children s Hospital Medical Center. Stool samples were evaluated for diarrheagenic E. coli using a panel of DNA probes and adherence pattern to HEp-2 cells. Stool samples from a reference group of 555 well children were studied for comparison. Results Gene probe studies, but not HEp-2 cell adherence, demonstrated that enteroaggregative, diffusely adherent and enteropathogenic E. coli were associated with clinical illness. Each was isolated significantly more often from study subjects in the ED than controls. In children <1 year of age, enteroaggregative E. coli were isolated significantly more often from both inpatients (4.7%, Odds Ratio = 3.4, 95% confidence intervals , p <0.03) and ED patients (10.0%, Odds Ratio = 7.2, 95% confidence intervals , p <0.001) than from well children (1.4%). Conclusions Diarrheagenic E. coli, especially enteroaggregative E. coli, may be an important, unrecognized cause of childhood diarrhea in the U.S. (J Pediatr 2005;146:54-61) In the United States, most childhood diarrheal disease is caused by pathogens not recognized in routine clinical testing. We had two reasons to suspect that diarrheagenic E. coli might be an important unrecognized cause of childhood diarrhea. First, these organisms are frequently found in the food supply in developing countries and therefore with increasing global trade they may be imported in food items sold in the U.S. For example, a number of recent food-borne outbreaks of enteric infection including cyclospora (1), and hepatitis A (2), have resulted from imported food. Second, regardless of the source of infection, clinical illness caused by many diarrheagenic E. coli is indistinguishable from viral gastroenteritis. Since diarrheagenic E. coli would not be identified by common clinical laboratory techniques, we hypothesized that that these organisms represent an unrecognized cause of diarrheal illness in the U.S. The strains of Escherichia coli that cause diarrhea are traditionally classified into groupings that include enterotoxigenic (ETEC), enteroinvasive (EIEC), enterohemorrhagic (EHEC), enteropathogenic (EPEC), diffusely adherent (DAEC) and enteroaggregative (EAEC) E. coli (1, 3). Except for EHEC, e.g., Shiga-toxin (stx) producing E. coli O157:H7, diarrheagenic E. coli are not routinely sought as stool pathogens in clinical laboratories in the US. Many of these diarrheagenic E. coli have, however, been shown to be important enteropathogens in developing countries (3-8) and in some developed countries (9, 10). To better understand the importance of these organisms as etiologic agents of acute infectious childhood diarrhea, we undertook a prospective study using a comprehensive panel of DNA probes and HEp-2 cell adherence to identify diarrheagenic E. coli in children with gastroenteritis and in an age-matched reference group of well children. n EAEC infant and childhood diarrhea enteroaggregative E. coli n n DNA probes HEp-2 cell assay See editorial, p 11. Cincinnati Children s Hospital Medical Center, Divisions of Gastroenterology, Hepatology and Nutrition 1, and Infectious Diseases 3, Cincinnati, Ohio and the Center for Vaccine Development, Departments of Pediatrics, Medicine, and Microbiology & Immunology, University of Maryland School of Medicine 2, Baltimore, Maryland. Grant support: This study was supported by Awards from the National Institutes of Health N01 AI and R01 DK We thank Carolyn Gunther, Phil Chen, Lisa Brockman, Cara Kohlreiser, Tara Hamann, Lori McKenna, Karen Munson, Sara O Cull, Jolyn Phillips, Amy Pohler, Michelle Roth, and Carla Thieman and Suzanne Davis for technical and nursing assistance. Correspondence and reprint requests to: Mitchell B. Cohen, MD, Cincinnati Children s Hospital Medical Center, Gastroenterology, Hepatology and Nutrition, MLC 2010, 3333 Burnet Ave., Cincinnati, OH USA, Phone: , Fax: mitchell.cohen@cchmc. org /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

103 Table I. DNA probes used to identify diarrheogenic Escherichia coli Escherichia coli pathotype Specific for DNA probe description Reference EAEC AA plasmid pcvd432 (1 kb EcoRI-PstI) EPEC eae gene pcvd434 (1 kb SalI-KpnI) EAF plasmid pjpn16 (1 kb BamHI-SalI) DAEC daac gene PLSM852 (390 bp PstI) EIEC Invasion pps2.5 (2.5 kb HindIII) ETEC LT enterotoxin pcvd403 (1.2.kb BamH1) STp enterotoxin pcvd426 (157 bp PstI) STh enterotoxin pcvd427 (216 bp EcoRI) EHEC Adherence/hemolysin pcvd419 (3.4 kb (HindIII) Shiga toxin I pjn37-19 (1.1 kb BamHI) Shiga toxin II PNN (842 bp SmaI-PstI) EAEC, enteroaggregative Escherichia coli; EPEC, enteropathogenic E coli; DAEC, diffuse adhering E coli; EIEC, enteroinvasive E coli; ETEC, enterotoxigenic E coli; EHEC, enterohemorrhagic E coli. METHODS Study Design, Inclusion Criteria and Informed Consent The study and consent forms were approved by the Institutional Review Board (IRB) of Cincinnati Children s Hospital Medical Center (CCHMC). Two populations of children with clinical illness were studied: 1) children #5 years of age evaluated and discharged from the Emergency Department (ED), and 2) an inpatient (IP) group of children #18 years of age either hospitalized or admitted to the shortstay (23-hr admission unit) at CCHMC. For inclusion in either study arm, children were identified with diarrhea and/or vomiting who could provide an unformed stool sample. Diarrhea was defined as an episode of $ 3 loose stools in a 24 hr period; the consistency of the stools and presence of blood was judged by the caretaker. Vomiting was defined as forceful expulsion of gastric contents occurring at least once in a 24 hr period. Fever was defined by recorded temperature in the ED or inpatient ward >100.4 F. Cases were identified in the ED by real time active surveillance performed by a study nurse who was located in the ED every fourth day. Cases were identified for the IP study by active surveillance (Monday-Friday) performed by study nurses as previously described (11). Stool was collected within <48 hours of hospitalization. Children were excluded if they had vomiting but no diarrhea and a concomitant illness, e.g. a respiratory illness that could explain the vomiting or if they had received antimicrobial therapy in the 7 days prior to presentation. A common reference (control) group was recruited from well child clinics, daycare centers and community volunteers. We attempted to match reference group subjects (2 ill subjects: 1 reference group subject) by age and month of enrollment. We used the entire control group for comparisons with the IP study and the control group subset #5 years old as a reference group for the ED study. Enrollment began in September 1998 and continued until February 2000 for the IP and control groups. Enrollment began in March 1999 and continued for 12 months until February 2000 for the ED group. Microbiology For microbiologic analysis, whole stools samples were obtained from all control subjects. For inpatient subjects 95.6% of the samples were from whole stool specimens; the remainder were from swabs. For subjects evaluated in the ED, 47.1% were bulk stools, 44.1% were from swabs and 8.9% were unknown or not recorded. Stool cultures were sent to the clinical laboratory at CCHMC to identify Salmonella spp., Shigella spp., Aeromonas and Plesiomonas spp., E. coli O157, Campylobacter spp., and Yersinia enterocolitica. Rotavirus was detected by a commercial assay (RotacloneÒ; Meridian Diagnostics, Cincinnati, OH). Stool samples were also inoculated onto EMB differential culture media (Difco, Detroit, MI). After overnight incubation, 5 characteristic (lactose-fermenting, non-mucoid) colonies from each stool sample were identified and streaked onto Trypticase Soy Agar (TSA) with 5% sheep blood (TSAII, Difco). Organisms grown on TSAII were indole tested. Five colonies with characteristic morphology on EMB agar and a positive indole test were selected as E. coli organisms and inoculated onto 11 LB agar plates (Difco) in a standardized array with appropriate positive and negative controls. Cultures were incubated overnight at 37 C after which the colonies were fixed on Whatman 541 (Whatman International Ltd., Maidstone England) filters and hybridized under stringent conditions with a biotinylated DNA fragment (Gibco/BRL-Bioprime DNA Labeling System and BLUGENE Non-radioactive Nucleic Acid Detection System, Carlsbad, CA) as previously described (12, 13) using the probes shown in Table I. These probes are >90% sensitive and specific for their gene targets (3). An additional set of 5 E. coli colonies per subject was analyzed for HEp-2 cell adherence (14). The five colonies were tested separately and the person performing the adherence phenotype was unaware of the group from which the isolates Prevalence Of Diarrheagenic Escherichia Coli In Acute Childhood Enteritis: A Prospective Controlled Study 55

104 Statistical Analysis The biostatistical design of the study assumed that the total prevalence of diarrheagenic E. coli infection would be 7.4%. This is based on the identification of EPEC in 5.6% of stool samples from children with diarrhea in Seattle (15), a prevalence of 0.8% of EHEC in Cincinnati based on historical clinical laboratory data, and a conservative estimate of 0.25% for each other category of diarrheagenic E. coli (5.6% EPEC 1 0.8% EHEC % 3 4 other groups = 7.4%). We estimated the total prevalence of diarrheagenic E. coli in the control group to be 1% based on our previous studies in which the control group prevalence in adults prior to international travel was between 0% (12) and 2% (13). With an a = 0.05, the power of this study with a sample size of 550 in each group to detect a significant difference was We prospectively planned to perform a subgroup analysis of infants <1 year of age because of the possibility of an increased frequency of these organisms as etiologic agents of diarrhea in younger children (16-18). Proportions were compared by Chisquare analysis or Fisher exact test. Odds Ratios with 95% confidence intervals were determined. Nonparametric data were analyzed by Wilcoxon Rank Sum. Figure 1. Screening, enrollment and sample population in the Inpatient (A) and Emergency Department (B) study populations. were obtained. Colonies were classified as demonstrating aggregative adherence (AA), localized adherence (LA), diffuse adherence (DA) or no adherence. Testing for C. difficile toxin, parasites, and other viruses, e.g., astroviruses, adenoviruses and noroviruses, was not performed. RESULTS All children in the correct age group who presented to the ED or the IP service were screened for eligibility. A total of 1882 subjects were enrolled and had stool samples tested. This included an IP sample of 643 of whom 563 (88%) were #5 years old, an ED sample of 684, of whom all were #5 years old and a control sample of 555, of whom 486 (88%) were #5 years old. The study screening and enrollment numbers are shown in Figure 1 and the demographic characteristics and clinical features of each study group are shown in Table II. It proved more difficult to enroll control group subjects who were not in diapers. Therefore, the mean age of the control group is younger than that of the comparable ED and IP groups. The duration of diarrhea prior to study enrollment was shorter in the ED group compared with IP as would be predicted by the study design that included real time surveillance in the ED. Vomiting was more frequently seen and fever was less frequently seen in the ED group. Blood was reported more frequently in stools of hospitalized patients. Routine bacterial pathogens, i.e., Salmonella spp. (n=14), Shigella spp. (n=1), Aeromonas spp. (n=2), E. coli O157 (n=7), Campylobacter spp. (n=2), and Yersinia enterocolitica (n=1) were isolated in 2.1% of the ill study subjects. The distribution between IP and ED subjects was similar (n=14, 2.2% for IP and n=14, 2.0% for ED). Routine bacterial pathogens were identified in 0.5% of controls (n=3, all Aeromonas spp.). Rotavirus was found in 20.2% of study patients in the ED, and 20.3% of IPs. Both were significantly greater (p<0.001) than the 1.1% prevalence found in controls. There was a significantly increased prevalence of diarrheagenic E. coli in study subjects presenting to the ED (167/684=24.4% vs. control 78/555=14.1%, Odds Ratio (OR) = 1.7, 95% confidence intervals , p<0.001) but no significant difference between IP (83/643=12.9%) and controls. This also held true for the subset of IP and controls #5 years old (76/563=13.5% vs. 75/486=15.4%, respectively). The total frequency of probe positive isolates (Table II) was greater than the number of subjects with diarrheagenic E. coli infection. This is attributed to several probes identifying a single class of diarrheagenic E. coli and because of a small number of subjects with co-infection of more than one class of diarrheagenic E. coli (none in the control group, 11 (1.7%) in the IP group and 20 (2.9%) in the ED group). When analyzed by specific pathogen, EAEC, atypical (eae-positive and EAF-negative) EPEC and DAEC were isolated significantly more often from ED study subjects than from the comparable control group (OR = 2.8, 95% confidence intervals , p<0.001, OR = 1.6, 95% confidence intervals , p=0.05, and OR = 1.9, 95% confidence intervals , p <0.001, for EAEC, EPEC and DAEC respectively) (Table III). HEp-2 cell adherent E. coli were not more prevalent in the ED or IP groups, however, the large 56 Cohen et al The Journal of Pediatrics January 2005

105 Table II. Demographic and clinical characteristics of the study population and study subjects positive for enteroaggregative Escherichia coli Inpatients ED Control subjects All #5 y old EAEC positive #5 y old EAEC positive All #5 y old N Mean age (mo) 25.3 ± ± 0.6* 15.7 ± ± 0.5 * 7.8 ± 0.9 y 22.3 ± ± 0.6 Median age (mo) Age range (mo),1 213,1 60,1 84,1 60,1 36,1 225,1 58 Median days of 3 (0 69) 3 (0 84) 3 (0 17) 2 (0 60) 1 y (0 32) NA NA diarrhea at entry into study (range) Percent with vomiting 76% 74% y 72% 82% 70% z 0% Percent with fever 64% 66% y 60% 43% 54% 0% 0% Percent with history 10.4% 8.2% y 0.4% 3.7% 6.3% 0% 0% of blood in stool Percent in daycare 22.2% 23.8% 16% 25.6% 25.4% 25.6% 27.0% IP, inpatient; ED, Emergency Department; EAEC, enteroaggregative Escherichia coli; NA, not applicable. *P <.001 vs control subjects #5 years old. yp <.001 vs ED. zp =.02 vs ED. P <.01 vs ED. number of AA isolates in the control group indicates that this test is not specific for pathogenic strains (Table II). In the subset of infants <1 year of age (Table IV), probe positive EAEC were isolated more commonly in both IP and ED groups (OR = 3.4, 95% confidence intervals , p <0.03, and OR = 7.2, 95% confidence intervals , p <0.001, respectively) while DAEC were more commonly found in ED study subjects only (OR = 1.9, 95% confidence intervals , p <0.001). E. coli demonstrating AA in the HEp-2 cell assay were more prevalent in the ED group <1 year of age (OR = 1.6, 95% confidence intervals , p <0.001), although there was also a high prevalence rate (22.6%) in controls. The type of stool specimen obtained (whole stool vs. swab) did not influence the yield of E. coli pathogens. The isolation rate for EAEC from subjects evaluated in the ED was 9.6% for bulk stool samples and 9.3% from swabs. The isolation rate for EAEC from inpatients was 4.8% for whole stool samples and an identical 4.8% from swabs. A minority of subjects (14.3% in the ED cohort, and 12.1% in the inpatient cohort) were enrolled with only vomiting, i.e., no diarrhea. Stratifying the data into a group with diarrhea and/or vomiting and another group with diarrhea, i.e., eliminating those with vomiting alone, did not significantly change the results as shown in Figure 2. Similarly, stratifying by subjects who did not have another identified pathogen (predominantly rotavirus), also did not change the findings that EAEC (p=0.0002), DAEC (p=0.01) and atypical EPEC (p=0.02) (eae positive E. coli) were more frequently identified in the ED cohort as shown in Figure 2. In the IP group 22/83 (26.5%) were positive for rotavirus as well as diarrheagenic E. coli. Among IP without any diarrheagenic E. coli isolated, 111/560 (19.8%) were positive for rotavirus. The difference is not significant (p=.20). Of those samples from the ED positive for diarrheagenic E. coli, 40/167 (23.9%) were positive for rotavirus. Among ED subjects without any diarrheagenic E. coli, 92/517 (19.8%) were positive for rotavirus. The difference is not significant (p=.33). Among the 25 subjects positive for EAEC in the IP sample, 8 (32%) had co-infection with other pathogens, predominantly rotavirus. Among the 63 subjects positive for EAEC in the ED sample, 18 (29%) had co-infection with other pathogens, predominantly rotavirus. When we eliminated co-pathogens from the analysis, we found that atypical EPEC (p=0.05) and EHEC organisms (p=0.03) were also more frequent in the inpatient cohort. The demographic and clinical features of subjects with and without probe positive EAEC are shown in Table II. Only 1 subject with EAEC in the IP sample was >5 years old. Those with EAEC infection were somewhat younger (mean age 7.8 months) than other children in the ED group and had a statistically greater incidence of fever and somewhat less likelihood of vomiting. In the IP group, EAEC positive subjects were less likely to have blood in their stool. However, despite these differences, subjects with EAEC were clinically indistinguishable from non-eaec subjects in the IP or ED groups (Table II). Similarly, as shown in Figure 3, the seasonal prevalence of EAEC from both ED and IP subjects paralleled the overall prevalence of gastroenteritis with a strong seasonal peak in March-April which is also the expected time for peak rotaviral diarrhea in Cincinnati (11). In both March-April and July-September the prevalence of EAEC was approximately 10% in all study subjects. Prevalence Of Diarrheagenic Escherichia Coli In Acute Childhood Enteritis: A Prospective Controlled Study 57

106 Table III. Prevalence of diarrheogenic Escherichia coli in inpatients, Emergency Department patients, and control subjects by gene probe and HEp-2 assay Probe or assay All IP (n = 643) (% positive) IP #5 y old (n = 563) (% positive) ED #5 y (n = 684) (% positive) All control (n = 555) (% positive) Control subjects #5 y old (n = 486) (% positive) EAEC: AA probe * AA by HEp EPEC: Eae probe y EAF probe LA by HEp DAEC: DAEC probe * DA by Hep EIEC ETEC LT STp STh EHEC stx-i 1.1 z stx-ii EHEC IP, inpatients; ED, Emergency Department; EAEC, enteroaggregative Escherichia coli; EPEC, enteropathogenic E coli; DAEC, diffuse adhering E coli; EIEC, enteroinvasive E coli; ETEC, enterotoxigenic E coli; EHEC, enterohemorrhagic E coli; AA, aggregative adherence; LA, localized adherence; DA, diffusely adherent. *P <.001 vs control <5 years old. yp <.05 vs control <5 years old. zp =.04 vs control. P =.02 vs control. DISCUSSION Pediatric diarrhea is a common medical problem in the U.S. with costs of >$1 billion/year (19). As expected, in both the IP and ED groups, rotavirus was the predominant single etiologic agent identified (20.3% and 20.2% of subjects respectively). Also as expected, routine bacterial enteropathogens were identified in a relatively small number (only 2.1%) of study subjects. The putative pathogen Aeromonas species were isolated only from control subjects. Using these two common clinical tests, we identified <25% of the etiology of enteritis. Although diarrheagenic E. coli are thought to be a common cause of infant diarrhea in developing countries, the importance of these organisms as mediators of acute pediatric diarrhea in the U.S. is uncertain. Our study was designed to identify the prevalence of E. coli as potential etiologic agents of diarrheal disease in a large Children s Hospital in the midwest U.S. Previous reports in the U.S. have evaluated the prevalence of various non-stx producing E. coli pathogens as causes of childhood diarrhea (15, 20-22). However, none has been a comprehensive study of all E. coli pathogens, or evaluated the newest classes of diarrheagenic E. coli, e.g., EAEC and DAEC and/or included a control group. Although our data are derived from a single point of care, the study location and sample size suggest that diarrheagenic E. coli may be an important cause of endemic childhood diarrhea in the U.S. If confirmed, these results could have an important impact on the evaluation and possible treatment of childhood diarrhea. Only some categories of diarrheagenic E. coli were associated with intestinal infection in our study. Despite early reports that ETEC were responsible for a significant portion of severe childhood diarrheal disease in Chicago (23) and endemic illness in White Mountain Apache children (24, 25), ETEC have not been isolated in subsequent prospective studies. These include studies of urban childhood diarrhea in Boston (26), Baltimore (20) and Houston (22). In addition, ETEC were not found in ill family members studied in Charlottesville more frequently than in controls (27). We also found that ETEC were not a significant cause of sporadic diarrhea in our study population. Similarly, as expected in the absence of an outbreak, EIEC were not found in our sample groups. In a prospective point-of-care study in Seattle children, stxproducing E. coli were isolated from 2.1% of all stool cultures (28), a higher rate than we observed. However, our historical isolation rate using sorbitol-macconkey agar screening in the clinical laboratory is 0.8%, consistent with the rates of isolation in our current study. Three categories of diarrheagenic E. coli were associated with clinical illness in our study. EAEC, DAEC and atypical EPEC were isolated significantly more often from study subjects in the ED than in the control group. In children <1 year of age, EAEC were isolated more commonly in both the 58 Cohen et al The Journal of Pediatrics January 2005

107 Table IV. Prevalence of diarrheogenic Escherichia coli in inpatients, Emergency Department patients, and control patients younger than 1 year old by gene probe and HEp-2 assay Probe or assay: IP (n = 343) <1 y old (% positive) ED (n = 362) <1 y old (% positive) Control subjects (n = 363) <1 y old (% positive) EAEC probe 4.7 * 10.0 * 1.4 AA by HEp * 22.6 EPEC eae probe EAF probe LA by HEp DAEC probe y 6.3 DA by HEp EIEC ETEC LT STp STh EHEC stx-i Stx-II EHEC IP, inpatients; ED, Emergency Department; EAEC, enteroaggregative Escherichia coli; EPEC, enteropathogenic E coli; DAEC, diffuse adhering E coli; EIEC, enteroinvasive E coli; ETEC, enterotoxigenic E coli; EHEC, enterohemorrhagic E coli; AA, aggregative adherence; LA, localized adherence; DA, diffusely adherent. *P <.001 vs control. yp =.03 vs control. IP and ED groups and DAEC were found more commonly in ED study subjects. A striking finding is the high isolation rate of probe-positive EAEC in ED study subjects <1 year of age (10.0%) compared with controls (1.4%). Thus, the younger mean age of the control group did not present a bias in favor of a difference. Although the less frequent isolation of EAEC from IP subjects could theoretically be because EAEC infection results in a relatively mild illness and does not often lead to hospitalization, the slight clinical differences (Table II) we observed between the EAEC positive subjects and others in either the ED or IP groups do not explain the discrepancy between ED and IP isolation rates. EAEC and DAEC were initially categorized as part of a larger group of enteroadherent E. coli. These strains differed from classical EPEC strains in that they did not show localized adherence in the HEp-2 cell assay (29). The aggregative or stacked brick appearance of EAEC in this bioassay permitted epidemiologic investigation and EAEC were found to be associated with persistent diarrhea in developing countries (14, 30). There was uncertainty about EAEC pathogenicity because these organisms are found in apparently healthy individuals and because some epidemiologic studies failed to show an association with disease (13, 31, 32). However, evidence from volunteer studies (33, 34) and outbreaks (reviewed in (35) ) has confirmed the pathogenicity of some EAEC strains. The mechanisms by which these organisms cause disease is thought to involve adherence to the intestinal mucosa, possibly in both the small and large intestine, followed by secretion of one or more enterotoxins and/or stimulation of IL-8 release by a flagellar protein (36-38). The DNA probe used in our study was initially developed by Baudry et al (39) as an empiric tool to identify EAEC with high sensitivity and specificity (ca. 90% sensitive and 99% specific). However, more recently, Nataro and colleagues have reported that the EAEC probe encodes a putative ABC transporter apparatus, which is co-regulated with, and required for, efficient translocation of the dispersin protein, a newly identified EAEC virulence factor (40, 41). Moreover, these investigators have shown that the transporter and dispersin are part of a wider package of virulence genes under the common control of the transcriptional activator AggR; strains possessing the AggR regulon have been termed typical EAEC (42). As part of this study, we tested E. coli isolates for adherence to HEp-2 cells. The HEp-2 test for adherent E. coli appears to be poorly specific for pathogenic EAEC. Our data are consistent with the hypothesis that the EAEC probe may constitute a marker for pathogenic or typical EAEC strains. This observation is especially important in light of recent publications suggesting that adherence to formalin-fixed HEp-2 cells may constitute a convenient surrogate for the traditional assay using live cells (43, 44). Our data would suggest that molecular methods to detect typical EAEC are superior to adherence-based assays. Our study also confirms earlier work by Bokete et al (15) who identified eae-positive and EAF-negative E. coli as a frequent isolate in Seattle children whose stools were submitted for microbiologic analysis. Although these EPEC organisms were more commonly identified in children with Prevalence Of Diarrheagenic Escherichia Coli In Acute Childhood Enteritis: A Prospective Controlled Study 59

108 Figure 3. Seasonal prevalence of enteroaggregative E. coli (EAEC) infection: The combined prevalence of EAEC infection during the 12 consecutive months (March 1999-February 2000) in which active surveillance occurred in both the Emergency Department (ED) and Inpatient study. The prevalence of EAEC infection paralleled the overall prevalence of gastroenteritis with a strong seasonal peak in March-April and another increased isolation rate in July-September. During the periods March-April and July-September the isolation rate of EAEC was approximately 10% from all study subjects. n = EAEC isolates, e=total number of study subjects. Figure 2. Prevalence of probe-positive E. coli. Subjects are stratified by symptom (all subjects, diarrhea present, vomiting only) and by whether they had other co-pathogens (predominantly rotavirus) identified. Removing subjects who had vomiting alone or other co-pathogens isolated did not change the strong associations with EAEC, atypical EPEC or DAEC. diarrhea than controls in our study, our isolation from the control group shows that there is asymptomatic carriage of these organisms as well. Fewer studies have investigated the epidemiology of DAEC. In contrast to Levine et al (45), who did not observe an increased prevalence in infants <1 year of age with diarrhea, we found an association of DAEC with diarrhea in this age group. In contrast to Jallat et al (46), who identified DAEC in hospitalized infants, children and adults, we found an increased prevalence of DAEC in ED subjects but not in hospitalized patients. However, our study design did not specify criteria for hospitalization and is therefore not directly comparable. In our study there was a significant co-infection rate with rotavirus among those subjects who also had diarrheagenic E. coli isolated. The seasonal prevalence curves of rotavirus and EAEC isolation also are quite similar. While infection with either organism alone was associated with enteritis, it is possible that co-infection with rotavirus and EAEC is a risk factor for symptomatic or more highly symptomatic illness. Our study did not specifically address this issue, but now that the prevalence of EAEC infection is known, it is possible to design such a study. In summary, the current study suggests that diarrheagenic E. coli may be an important, unrecognized cause of diarrhea in children in the U.S., perhaps accounting for 10% of all acute gastroenteritis. Further study of the pathogenic mechanisms of these E. coli may facilitate the development of rapid diagnostic tests which, in turn, may have significant implications for the approach to diagnostic and treatment protocols for acute diarrhea in children. REFERENCES 1. Herwaldt BL, Ackers ML. An outbreak in 1996 of cyclosporiasis associated with imported raspberries. The Cyclospora Working Group. N Engl J Med 1997;336: Hutin YJ, Pool V, Cramer EH, Nainan OV, Weth J, Williams IT, et al. A multistate, foodborne outbreak of hepatitis A. National Hepatitis A Investigation Team. N Engl J Med 1999;340: Nataro JP, Kaper JB. Diarrheagenic Escherichia coli. Clin Microbiol Rev 1998;11: Albert MJ, Faruque SM, Faruque AS, Neogi PK, Ansaruzzaman M, Bhuiyan NA, et al. Controlled study of Escherichia coli diarrheal infections in Bangladeshi children. J Clin Microbiol 1995;33: Albert MJ, Faruque AS, Faruque SM, Sack RB, Mahalanabis D. Casecontrol study of enteropathogens associated with childhood diarrhea in Dhaka, Bangladesh. J Clin Microbiol 1999;37: Porat N, Levy A, Fraser D, Deckelbaum RJ, Dagan R. Prevalence of intestinal infections caused by diarrheagenic Escherichia coli in Bedouin infants and young children in Southern Israel. Pediatr Infect Dis J 1998;17: Scaletsky IC, Fabbricotti SH, Carvalho RL, Nunes CR, Maranhao HS, Morais MB, et al. Diffusely adherent Escherichia coli as a cause of acute diarrhea in young children in Northeast Brazil: a case-control study. J Clin Microbiol 2002;40: Scaletsky IC, Fabbricotti SH, Silva SO, Morais MB, Fagundes-Neto U. HEp-2-adherent Escherichia coli strains associated with acute infantile diarrhea, Sao Paulo, Brazil. Emerg Infect Dis 2002;8: Knutton S, Shaw R, Phillips AD, Smith HR, Willshaw GA, Watson P, et al. Phenotypic and genetic analysis of diarrhea-associated Escherichia coli isolated from children in the United Kingdom. J Pediatr Gastroenterol Nutr 2001;33: Pabst WL, Altwegg M, Kind C, Mirjanic S, Hardegger D, Nadal D. Prevalence of enteroaggregative Escherichia coli among children with and without diarrhea in Switzerland. J Clin Microbiol 2003;41: Staat MA, Azimi PH, Berke T, Roberts N, Bernstein DI, Ward RL, et al. Clinical presentations of rotavirus infection among hospitalized children. Pediatr Infect Dis J 2002;21: Cohen et al The Journal of Pediatrics January 2005

109 12. Heck JE, Staneck JL, Cohen MB, Weckbach LS, Giannella RA, Hawkins J, et al. Prevention of Travelers Diarrhea: Ciprofloxacin versus Trimethoprim/Sulfamethoxazole in Adult Volunteers Working in Latin America and the Caribbean. J Travel Med 1994;1: Cohen MB, Hawkins JA, Weckbach LS, Staneck JL, Levine MM, Heck JE. Colonization by enteroaggregative Escherichia coli in travelers with and without diarrhea. J Clin Microbiol 1993;31: Nataro JP, Kaper JB, Robins-Browne R, Prado V, Vial P, Levine MM. Patterns of adherence of diarrheagenic Escherichia coli to HEp-2 cells. Pediatr Infect Dis J 1987;6: Bokete TN, Whittam TS, Wilson RA, Clausen CR, O Callahan CM, Moseley SL, et al. Genetic and phenotypic analysis of Escherichia coli with enteropathogenic characteristics isolated from Seattle children. J Infect Dis 1997;175: Levine MM, Edelman R. Enteropathogenic Escherichia coli of classic serotypes associated with infant diarrhea: epidemiology and pathogenesis. Epidemiol Rev 1984;6: Haider K, Faruque SM, Shahid NS, Albert MJ, Nahar S, Malek A, et al. Enteroaggregative Escherichia coli infections in Bangladeshi children: clinical and microbiological features. J Diarrhoeal Dis Res 1991;9: Gonzalez R, Diaz C, Marino M, Cloralt R, Pequeneze M, Perez-Schael I. Age-specific prevalence of Escherichia coli with localized and aggregative adherence in Venezuelan infants with acute diarrhea. J Clin Microbiol 1997; 35: Avendano P, Matson DO, Long J, Whitney S, Matson CC, Pickering LK. Costs associated with office visits for diarrhea in infants and toddlers. Pediatr Infect Dis J 1993;12: Kotloff KL, Wasserman SS, Steciak JY, Tall BD, Losonsky GA, Nair P, et al. Acute diarrhea in Baltimore children attending an outpatient clinic. Pediatr Infect Dis J 1988;7: Pickering LK, Evans DJ Jr, Munoz O, DuPont HL, Coello-Ramirez P, Vollet JJ, et al. Prospective study of enteropathogens in children with diarrhea in Houston and Mexico. J Pediatr 1978;93: Caeiro JP, Mathewson JJ, Smith MA, Jiang ZD, Kaplan MA, Dupont HL. Etiology of outpatient pediatric nondysenteric diarrhea: a multicenter study in the United States. Pediatr Infect Dis J 1999;18: Gorbach SL, Khurana CM. Toxigenic Escherichia coli: a cause of infantile diarrhea in Chicago. N Engl J Med 1972;287: Sack RB, Hirschhorn N, Brownlee I, Cash RA, Woodward WE, Sack DA. Enterotoxigenic Escherichia-coli-associated diarrheal disease in Apache children. N Engl J Med 1975;292: Sack RB, Santosham M, Reid R, Black R, Croll J, Yolken R, et al. Diarrhoeal diseases in the White Mountain Apaches: clinical studies. J Diarrhoeal Dis Res 1995;13: Echeverria P, Blacklow NR, Smith DH. Role of heat-labile toxigenic Escherichia coli and Reovirus-like agent in diarrhoea in Boston children. Lancet 1975;2: Guerrant RL, Hughes JM, Lima NL, Crane J. Diarrhea in developed and developing countries: magnitude, special settings, and etiologies. Rev Infect Dis 1990;12(Suppl 1):S Klein EJ, Stapp JR, Clausen CR, Boster DR, Wells JG, Qin X, et al. Shiga toxin-producing Escherichia coli in children with diarrhea: a prospective point-of-care study. J Pediatr 2002;141: Vial PA, Mathewson JJ, DuPont HL, Guers L, Levine MM. Comparison of two assay methods for patterns of adherence to HEp-2 cells of Escherichia coli from patients with diarrhea. J Clin Microbiol 1990;28: Bhan MK, Raj P, Levine MM, Kaper JB, Bhandari N, Srivastava R, et al. Enteroaggregative Escherichia coli associated with persistent diarrhea in a cohort of rural children in India. J Infect Dis 1989;159: Gomes TA, Blake PA, Trabulsi LR. Prevalence of Escherichia coli strains with localized, diffuse, and aggregative adherence to HeLa cells in infants with diarrhea and matched controls. J Clin Microbiol 1989;27: Echeverria P, Serichantalerg O, Changchawalit S, Baudry B, Levine MM, Orskov F, et al. Tissue culture-adherent Escherichia coli in infantile diarrhea. J Infect Dis 1992;165: Mathewson JJ, Johnson PC, DuPont HL, Satterwhite TK, Winsor DK. Pathogenicity of enteroadherent Escherichia coli in adult volunteers. J Infect Dis 1986;154: Nataro JP, Deng Y, Cookson S, Cravioto A, Savarino SJ, Guers LD, et al. Heterogeneity of enteroaggregative Escherichia coli virulence demonstrated in volunteers. J Infect Dis 1995;171: Okeke IN, Nataro JP. Enteroaggregative Escherichia coli. Lancet Infect Dis 2001;1: Vial PA, Robins-Browne R, Lior H, Prado V, Kaper JB, Nataro JP, et al. Characterization of enteroadherent-aggregative Escherichia coli, a putative agent of diarrheal disease. J Infect Dis 1988;158: Savarino SJ, Fasano A, Robertson DC, Levine MM. Enteroaggregative Escherichia coli elaborate a heat-stable enterotoxin demonstrable in an in vitro rabbit intestinal model. J Clin Invest 1991;87: Steiner TS, Nataro JP, Poteet-Smith CE, Smith JA, Guerrant RL. Enteroaggregative Escherichia coli expresses a novel flagellin that causes IL-8 release from intestinal epithelial cells. J Clin Invest 2000;105: Baudry B, Savarino SJ, Vial P, Kaper JB, Levine MM. A sensitive and specific DNA probe to identify enteroaggregative Escherichia coli, a recently discovered diarrheal pathogen. J Infect Dis 1990;161: Nishi J, Sheikh J, Mizuguchi K, Luisi B, Burland V, Boutin A, et al. The export of coat protein from enteroaggregative escherichia coli by a specific ATP-binding cassette transporter system. J Biol Chem 2003: M Okeke IN, Lamikanra A, Czeczulin J, Dubovsky F, Kaper JB, Nataro JP. Heterogeneous virulence of enteroaggregative Escherichia coli strains isolated from children in Southwest Nigeria. J Infect Dis 2000;181: Nataro J. Enteroaggregative E. coli. In: Hughes J, editor. Emerging Infectious Diseases 6: ASM Press; In press. 43. Miqdady MS, Jiang ZD, Nataro JP, DuPont HL. Detection of enteroaggregative Escherichia coli with formalin-preserved HEp-2 cells. J Clin Microbiol 2002;40: Spencer J, Chart H, Smith HR, Rowe B. Improved detection of enteroaggregative Escherichia coli using formalin-fixed HEp-2 cells. Lett Appl Microbiol 1997;25: Levine MM, Ferreccio C, Prado V, Cayazzo M, Abrego P, Martinez J, et al. Epidemiologic studies of Escherichia coli diarrheal infections in a low socioeconomic level peri-urban community in Santiago, Chile. Am J Epidemiol 1993;138: Jallat C, Livrelli V, Darfeuille-Michaud A, Rich C, Joly B. Escherichia coli strains involved in diarrhea in France: high prevalence and heterogeneity of diffusely adhering strains. J Clin Microbiol 1993;31: Jerse AE, Yu J, Tall BD, Kaper JB. A genetic locus of enteropathogenic Escherichia coli necessary for the production of attaching and effacing lesions on tissue culture cells. Proc Natl Acad Sci U S A 1990;87: Nataro JP, Baldini MM, Kaper JB, Black RE, Bravo N, Levine MM. Detection of an adherence factor of enteropathogenic Escherichia coli with a DNA probe. J Infect Dis 1985;152: Bilge SS, Clausen CR, Lau W, Moseley SL. Molecular characterization of a fimbrial adhesin, F1845, mediating diffuse adherence of diarrheaassociated Escherichia coli to HEp-2 cells. J Bacteriol 1989;171: Wood PK, Morris JG Jr, Small PL, Sethabutr O, Toledo MR, Trabulsi L, et al. Comparison of DNA probes and the Sereny test for identification of invasive Shigella and Escherichia coli strains. J Clin Microbiol 1986;24: Moseley SL, Echeverria P, Seriwatana J, Tirapat C, Chaicumpa W, Sakuldaipeara T, et al. Identification of enterotoxigenic Escherichia coli by colony hybridization using three enterotoxin gene probes. J Infect Dis 1982; 145: Levine MM, Xu JG, Kaper JB, Lior H, Prado V, Tall B, et al. A DNA probe to identify enterohemorrhagic Escherichia coli of O157:H7 and other serotypes that cause hemorrhagic colitis and hemolytic uremic syndrome. J Infect Dis 1987;156: Newland JW, Neill RJ. DNA probes for Shiga-like toxins I and II and for toxin-converting bacteriophages. J Clin Microbiol 1988;26: Prevalence Of Diarrheagenic Escherichia Coli In Acute Childhood Enteritis: A Prospective Controlled Study 61

110 PAPAVERINE PROLONGS PATENCY OF PERIPHERAL ARTERIAL CATHETERS IN NEONATES M. PAMELA GRIFFIN, MD, AND MIR S. SIADATY, MD, MS Objective To test the hypothesis that the continuous infusion of papaverine-containing solutions in peripheral arterial catheters would decrease the catheter failure rate and increase the functional duration of the catheter in neonates. Study design In a prospective, randomized, placebo-controlled, masked trial, 82 catheters were placed in 70 neonates in the papaverine group and 98 catheters were placed in 71 neonates in the placebo group. Results The catheters in the papaverine group remained functional for a significantly longer duration than those in the placebo group. The median (25th%, 75th%) time before catheter failure was 16.6 (9.5, 24.3) days in the papaverine group and 12 days (6.1, 18.2) in the placebo group (P =.023; Cox proportional hazards model). There was no significant difference in the incidence of intraventricular hemorrhage (IVH) between groups, and there was no evidence of hepatic toxicity. Conclusions The continuous infusion of papaverine-containing fluids prolongs the patency of peripheral arterial catheters in neonates. In this small number of infants, we found no difference in the incidence of IVH or hepatic toxicity. (J Pediatr 2005;146:62-5) Critically ill neonates require arterial access for frequent blood sampling and for continuous monitoring of blood pressure. A peripheral arterial catheter is necessary if an umbilical arterial catheter cannot be placed in the immediate period after birth, or if the infant is older. 1,2 Peripheral arteries are small and prone to vasospasm, and this may shorten the functional duration of peripheral arterial catheters. 3 Papaverine causes the relaxation of smooth muscle, including vascular smooth muscle, which could assist in prolonging the life of the catheter. 4,5,6 In 1993, Heulitt and co-workers found that papaverine prolonged peripheral arterial catheter patency in children 3 weeks to 18 years of age. 6 The use of papaverine in infants <3 weeks of age has not been previously studied because of the potential increased risk of intracranial hemorrhage as a result of cerebral vascular dilatation with the use of papaverine. 7,8 To address this concern before undertaking our study, we conducted a 2-year chart review of neonates who had peripheral arterial lines, some of whom had papaverine added to the infusion solution, and we detected no increased risk of intraventricular hemorrhage (IVH). Thus we proceeded with this randomized study of papaverine-containing solutions in neonates. We hypothesized that the continuous infusion of papaverine-containing solutions in peripheral arterial catheters would decrease the catheter failure rate and would prolong the functional duration of catheters in neonates. METHODS Infants admitted to the University of Virginia neonatal intensive care unit between August 1995 and August 2002 were eligible for the study if they required an indwelling peripheral arterial catheter for clinical care. The Institutional Review Board approved the protocol, and informed consent was obtained before study enrollment. During the study period, 382 infants were eligible for enrollment. Infants were excluded for the following reasons. One hundred eighty-nine were not enrolled because parents were not available or were not approached, 33 parents refused to give consent, 10 infants were on extracoporeal membrane oxygenation, 4 infants had documented IVH, 2 infants were not randomized by the pharmacy, 2 infants were given papaverine because of attending physician preference, and ALT AST Alanine aminotransferase Aspartate aminotransferase IVH Intraventricular hemorrhage From the Departments of Pediatrics and Health Evaluation Sciences, University of Virginia Health System, Charlottesville, Virginia. Submitted for publication May 13, 2004; last revision received Jul 22, 2004; accepted Aug 18, Reprint requests: M. Pamela Griffin, MD, Department of Pediatrics, Box , University of Virginia, Charlottesville, VA mpg7u@virginia. edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

111 Table I. Characteristics of study patients Papaverine n=70 (82 catheters) Placebo n=71 (98 catheters) Figure 1. Flowchart of the study. 1 infant who had two catheters was excluded because of randomization to both groups. Figure 1 shows the fate of the eligible infants. Treatment assignment was randomized by the pharmacy using a table of random numbers. Study patients were randomized to receive arterial catheter infusion solutions of 0.9% or 0.45% sodium chloride with 1.0 U/mL sodium heparin with or without papaverine 30 mg/250 ml that was prepared by the pharmacy. Investigators and medical personnel caring for the infants were unaware of patient assignment. Twenty-four-gauge catheters made of FEP (fluorinated ethylene propylene) polymer (Jelco, Ethicon Endo-Surgery, Inc, Cincinnati, Ohio) were placed using aseptic technique. Catheter failure was defined as the inability to draw blood from the catheter. At the time of catheter insertion, the person who inserted the catheter recorded the date and time of insertion and the number of insertion attempts in that extremity. We recorded information about the number of draws from the catheter and the administration of dopamine or dobutamine during the time the catheter was in place. The person removing the catheter recorded the date, time, and reason for removal (catheter failure, accidental or elective removal, death). In addition, we noted birth weight, gestational age, age at catheter insertion, and the results of cranial ultrasonography. Cranial ultrasounds were obtained by the medical staff as part of the infant s clinical care. We routinely obtain cranial ultrasonography during the first week of life and again at discharge for infants with birth weight <1500 g. Larger infants do not routinely have cranial ultrasonography performed unless there is a clinical indication. The grading system of Papile and co-workers was used to describe the IVH. 9 Statistics We based our sample size determination on data from the study by Heulitt and co-workers 6 with a 22% failure rate in Birth weight (g) 952 (692, 2670) 910 (655, 2150) Gestational age (wk) 27 (25, 36) 27 (24, 34) Male (%) Age at insertion (d) 11 (3, 19) 12 (7, 25) Total number 51 (28, 93) 52 (25, 83) of draws Attempts at placement Unknown 6 12 Person who inserted catheter Attending Fellow Resident Nurse 4 7 Unknown 2 4 Dopamine Dobutamine 7 8 Values for birth weight, gestational age, age at insertion, and total draws are median (25th%, 75th%). For other variables, the number in each group is given. None of the variables are significantly different. controls and 7% failure rate in the papaverine group. For our study, we assumed a 22% failure rate in controls and believed that a reduction to 11% catheter failure in the treated group would be clinically significant. Accepting an a error of.05 and a b error of 0.2, we determined that 89 catheters for study were required in each group. For analyzing the effects of papaverine on the functional duration of catheters, we used a Cox proportional hazards survival model. This model allows the use of censored data in infants who had catheters removed for a reason independent of catheter failure. For analysis of the cranial ultrasonography results, we used a marginal longitudinal model with a Poisson random component, log link, and exchangeable correlation matrix. 10 To look for differences between the groups in demographics, effect of vasoactive drugs, and hepatic toxicity, we used a logistic model in which the outcome was the group membership. In addition, we used Mann-Whitney, Pearson s x 2, and t tests for per-variable significance testing. RESULTS We studied 82 catheters in 70 infants in the papaverine treated group and 98 catheters in 71 infants in the placebo group. The demographic characteristics of the two groups were similar (Table I). All of the catheters were placed percutaneously except one in the placebo group that was surgically placed and electively removed after 11.5 days. There Papaverine Prolongs Patency Of Peripheral Arterial Catheters In Neonates 63

112 Table III. Cranial ultrasounography results before and after treatment Papaverine n=70 Placebo n=71 Figure 2. Effect of papaverine on catheter survival. At time 0, there were 82 catheters in the papaverine group and 98 catheters in the placebo group. Cox proportional hazards model showing the proportion of catheters that remained functional over time after insertion. Papaverine significantly prolonged the duration of the catheter (P =.023). Table II. Reasons for removal of arterial catheters Papaverine n = 82 (%) Placebo n = 98 (%) Catheter failure 40 (49) 56 (57) Elective removal 34 (41) 28 (29) Accidental removal 6 (7) 8 (8) Death 2 (2) 6 (6) was no difference between groups for the number of times that blood was drawn from the catheter, the person inserting the catheter, or the number of attempts at catheter placement. The attendings and fellows inserted the majority of the catheters in both groups. There was evidence of a previous puncture in 41% of the infants receiving papaverine and 40% of those receiving placebo. The effect of papaverine on arterial catheter failure is shown in Table II. Although there were more catheter failures in the placebo group and more elective removals in the papaverine group, with our sample size the difference in reason for catheter removal was not significant (x 2 = 4.1; P = 0.25). However, papaverine significantly prolonged the duration of the catheter (P =.023, Cox proportional hazards model) (Figure 2). In the papaverine group, the median (25th%, 75th%) catheter duration was 16.6 (9.5, 24.3) days, whereas in the placebo group the median catheter duration was 12 days (6.1, 18.2). The longest catheter duration was 44 days, in the papaverine group. In the placebo group the longest catheter duration was 27.6 days. Several infants in both groups received dopamine and dobutamine at some time during the catheter # studies before treatment Status before treatment: No hemorrhage Grade Grade Grade Grade Periventricular bleed 1 0 # studies after treatment Status after treatment No hemorrhage Grade * Grade Grade 3 8 * 2 Grade Periventricular bleed 2 1 *Indicates 1 infant in the papaverine group who had periventricular hemorrhage in addition to grade 3 IVH and 1 infant in the placebo group who had periventricular hemorrhage in addition to grade 1 IVH. period, but these drugs did not have a significant effect on the functional duration of the catheters. After intervention, in the papaverine group there were 19 infants with IVH, whereas in the placebo group there were 15 infants with IVH (Table III). Before study enrollment, 4 infants in the papaverine group had grade 3 or 4 IVH, 3 infants did not have ultrasonography before initiation of treatment, 1 infant who had no hemorrhage on the initial study had a grade 3 hemorrhage after treatment, and 1 infant had a grade 2 hemorrhage initially and then a grade 3 on repeat study. The infant who had no hemorrhage initially had the study performed at 2 days of age, and the radiologist read that there was asymmetry of the choroid plexus on the left but no definite hemorrhage. Papaverine was started at 5 days of age. A repeat study at 7 days of age was read as asymmetric hyperechoic material on the left and was called a grade 3 IVH. The infant with the grade 2 IVH had a grade 1 IVH at 1 day of age. A repeat study at 3 days of age was read as bilateral grade 2 IVH. Papaverine was started at 8 days of age. Follow-up ultrasonography at 12 days of age showed a slight increase in hydrocephalus but no change in the hemorrhage. A study at 19 days of age was read as an increase in hydrocephalus and was then called grade 3 IVH. In the papaverine group, 1 infant had periventricular hemorrhage in addition to a grade 3 IVH, and 1 infant had periventricular hemorrhage without an IVH. In the placebo group, 1 infant had a periventricular hemorrhage in addition to a grade 1 IVH. There was no significant difference between the groups for the development of IVH. We found no evidence that papaverine caused hepatic toxicity. In the papaverine group, median (25th%, 75th%) 64 Griffin and Siadaty The Journal of Pediatrics January 2005

113 values were conjugated bilirubin 0.3 mg/dl (0.1, 1.55), aspartate aminotransferase (AST) 29 U/L (22, 41), alanine aminotransferase (ALT) 18 U/L (12, 29), and alkaline phosphatase 292 U/L (188, 407). In the placebo group the conjugated bilirubin was 0.2 mg/dl (0.1, 1.43), AST 30 U/L (22, 44), ALT 25 U/L (14, 38), and alkaline phosphatase 230 U/L (134, 384). DISCUSSION The main finding of our study is that papaverinecontaining infusion fluids prolonged the functional duration of peripheral arterial catheters in neonates without evidence of toxicity. Although the functional duration of the catheters in the papaverine group was prolonged, we did not find that the catheter failure rate was significantly different between groups. In pediatric patients, Heulitt and co-workers found that papaverine in arterial catheter infusion fluids prolonged patency of the catheter and decreased the catheter failure rate. 6 Our study differs from the study by Heulitt and coworkers in that we had a much higher catheter failure rate. In our study the catheter failure rate in the placebo group was 57% compared with 22% in the study by Heulitt and coworkers. One explanation for this may be that the catheters in Heulitt s study remained in place for a shorter length of time. The median time before catheter failure in infants receiving papaverine in our study was just over 16 days, whereas in the Heulitt study only 3% of catheters in the papaverine group remained in place at 16 days. There are few other studies of papaverine use in young children. Papaverine, however, has been studied in children 3 months to 5 years of age after cardiac catheterization. 11 In a study of 56 patients, one small dose of papaverine given into the femoral artery postcatheterization did not prevent the loss of pulse in the catheterized artery. In Heulitt s study of older children no adverse effects of papaverine were noted. 6 Because none of the children in that study were <3 weeks of age the concern of increased risk of IVH was minimal. In our study, although we found no significant difference between groups for the risk of IVH, we had only a small number of patients to examine. In addition, whereas our usual practice is to obtain cranial ultrasonography in premature infants at 7 days of age and at the time of discharge, our study protocol did not mandate that all infants in the study have cranial ultrasonography. Thus, we would recommend caution in using papaverine in the extremely premature infants in the immediate period after birth when these infants are at highest risk of developing an intracranial hemorrhage. Generally, our group does not add papaverine to the arterial infusion solution of premature infants until the infant is several weeks old. Another potential side effect of papaverine that has been reported in a small number of adults is hepatic toxicity with elevated transaminases. 12,13 Heulitt s study of papaverine use in children did not find any evidence of hepatic toxicity. 6 We also found no evidence of hepatic toxicity in our neonatal patients. In summary, we found that the continuous infusion of papaverine in peripheral arterial catheters prolonged the functional duration of the catheters without any harmful side effects. Common sense, however, suggests caution in using papaverine in premature infants in the immediate period after birth when they are at greatest risk of developing an intracranial hemorrhage. REFERENCES 1. Sellden H, Nilsson K, Larsson LE, Ekstrom-Jodal B. Radial arterial catheters in children and neonates: a prospective study. Crit Care Med 1987; 15: Randel SN, Tsang BH, Wung JT, Driscoll JM Jr, James LS. Experience with percutaneous indwelling peripheral arterial catheterization in neonates. Am J Dis Child 1987;141: Lemke RP, al-saedi SA, Belik J, Casiro O. Use of tolazaline to counteract vasospasm in peripheral arterial catheters in neonates. Acta Paediatr 1996;85: Hou SM, Seaber AV, Urbaniak JR. Relief of blood-induced arterial vasospasm by pharmacologic solutions. J Reconstr Microsurg 1987;3: Hillier C, Watt PA, Spyt TJ, Thurston H. Contraction and relaxation of human internal mammary artery after intraluminal administration of papaverine. Ann Thorac Surg 1992;53: Heulitt MJ, Farrington EA, O Shea TM, Stoltzman SM, Srubar NB, Levin DL. Double-blind, randomized, controlled trial of papaverinecontaining infusions to prevent failure of arterial catheters in pediatric patients. Crit Care Med 1993;21: Bevan RD, Vijayakamaran E, Gentry A, Wellman T, Bevan JA. Intrinsic tone of cerebral artery segments of human infants between 23 weeks of gestation and term. Pediatr Res 1998;43: Milburn JM, Moran CJ, Cross DT III, Diringer MN, Pilgram TK, Dacey RG Jr. Increase in diameters of vasospastic intracranial arteries by intraarterial papaverine administration. J Neurosurg 1998;88: Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal intraventricular hemorrhage: a study of infants with birth weights less than 1500 g. J Pediatr 1978;92: Diggle P, Heagerty P, Liang KY, Zeger S. Analysis of Longitudinal Data. 2nd ed. Oxford: Oxford University Press; Boris JR, Harned RK II, Logan LA, Wiggins JW Jr. The use of papaverine in arterial sheaths to prevent loss of femoral artery pulse in pediatric cardiac catheterization. Pediatr Cardiol 1998;19: Pathy MS, Reynolds AJ. Papaverine and hepatotoxicity. Postgrad Med J 1980;56: Driemen PM. Papaverine-hepatotoxic or not? J Am Geriatr Soc 1973; 21: Papaverine Prolongs Patency Of Peripheral Arterial Catheters In Neonates 65

114 THE PARKLAND MEMORIAL HOSPITAL EXPERIENCE IN ENSURING COMPLIANCE WITH UNIVERSAL NEWBORN HEARING SCREENING FOLLOW-UP ANGELA G. SHOUP, PHD, KRIS E. OWEN, MS, GREG JACKSON, MD, AND ABBOT LAPTOOK, MD Objective Reduce false-positive results and loss to follow-up through systematic modifications in Universal Newborn Hearing Screening at a large public hospital. Study design During a pilot program, neonates who failed technician-performed automated auditory brain stem response were scheduled for diagnostic evaluation. In year 1, audiologists rescreened neonates who failed, and those who did not pass were screened as outpatients. For years 2 through 4, neonates who failed were rescreened by technicians before inpatient audiology rescreening. Results For the pilot, 3759 neonates were screened; 1% (n = 43) failed and 44% (n = 19) were lost to follow-up. In year 1, 15,297 neonates were screened and 2% (n = 365) failed; audiology rescreening reduced this to <1% (n = 129). Outpatient rescreening yielded 0.5% (n = 70) who failed; 17% (n = 12) were lost to follow-up. In year 2, 16,384 neonates were screened, 3% (n = 456) failed, and 1% (n = 167) failed after technician rescreen; audiology rescreening reduced inpatient fails to 0.6% (n = 108), and 0.4% (n = 61) failed outpatient rescreening; 11% (n = 7) were lost to follow-up. Results for years 3 and 4 were similar to year 2, with further reduction in loss to follow-up to 11% (n = 6) and 1.7% (n = 1). Conclusions Successful Universal Newborn Hearing Screening with reduced false-positive results and loss to follow-up can be accomplished with a planned schedule of inpatient rescreens and outpatient rescreening at the birthing facility. (J Pediatr 2005;146:66-72) Universal newborn hearing screening (UNHS) is mandated in many states, with the goal of identification of and intervention for neonates with auditory impairment. Appropriate early intervention can reduce deleterious effects of hearing loss on language development, educational achievement, and future vocational options. 1-3 Although there is widespread support for UNHS, the needs to reduce false-positive results and increase return for follow-up continue to be cited as requiring improvement. 4 Only 2% to 7% of neonates who do not pass inpatient newborn hearing screenings will have sensorineural hearing loss. 4,5 Reduction of false-positive rates decreases undue parental anxiety and reduces the number of neonates referred for more costly diagnostic audiologic evaluations. A critical step in decreasing false-positive rates is to diminish the overall fail rate for inpatient hearing screening. This is especially important in view of the percentage of neonates who fail UNHS and are lost to follow-up, which ranges from 9% to 52%. 4 Parkland Memorial Hospital (PMH), in Dallas, Texas, is the largest single-site birthing hospital in the United States, with a current birth rate of ~17,000 infants annually. Selective hearing screening for high-risk and neonatal intensive care unit (NICU) neonates has been conducted since Ensuring that all neonates are screened and appropriate follow-up services are initiated in large urban institutions poses unique challenges as the result of patient volume, early hospital discharge, establishment of an outpatient medical home, and a mobile population. To overcome these challenges, implementation of UNHS at PMH required development of strategies that may be applicable to other health care systems. Specific performance goals were adopted to establish the UNHS program, including achieving hearing screening in >95% of newborn infants before discharge, 6 minimizing AAP NICU PMH American Academy of Pediatrics Neonatal intensive care unit Parkland Memorial Hospital PPV SNHL UNHS Positive predictive value Sensorineural hearing loss Universal newborn hearing screening From the Department of Otolaryngology and the Department of Pediatrics, University of Texas Southwestern Medical Center, Dallas, Texas; and the Department of Pediatrics, Brown Medical School, Providence, Rhode Island. Submitted for publication Jan 22, 2004; last revision received Jul 07, 2004; accepted Aug 20, Reprint requests: Angela G. Shoup, PhD, UT Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX angela.shoup@ utsouthwestern.edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

115 false-positive results by achieving a refer rate of <2%, and limiting loss to follow-up to <10% for neonates with possible auditory impairment. We describe a pilot project and modifications incorporated during the first and second years of the UNHS program to achieve these objectives. Data from the third and fourth years demonstrate the stability of the findings. METHODS A multidisciplinary team from administration, neonatology, nurse management, pediatric nurse practitioners, and audiology began meeting in 1996 to plan UNHS. Coordination of the program is through audiology, and responsibilities include program oversight, planning and development, technician training, performance monitoring, quality assurance/improvement, education of house staff and other health care providers, data management, parent counseling, and coordination of follow-up services through the use of telephone and written communication. Newborn hearing screenings for the pilot project, year 1 and year 2, were conducted with Algo 2e Color (Natus Medical, San Carlos, Calif) automated auditory brainstem response systems, which provide a pass or refer result. During year 3, the screening units were upgraded to the Algo 3. The sensitivity of the Algo 2e color and the Algo 3 is 99.98% and the specificity is 96%. Screenings were completed within the newborn nursery and NICU by technicians trained in use of equipment, newborn care, and documentation. Verification of technician skills included written examination, performance skills checklist, and monitoring of performance. Data management was accomplished with a proprietary database that imports data from the screening equipment. Entry screens for additional patient information (eg, demographics, history, risk factors for hearing loss and progressive hearing loss) were designed to complement paper forms used in the program and to automate tracking of neonates. Details of three time intervals are described below and are illustrated in the Figure. Based on pilot project results, modifications in years 1 and 2 of UNHS were implemented to improve return for follow-up. Protocol for the Pilot Project (April 1999 to August 1999) To prepare for UNHS, a pilot project was designed and implemented in April of The pilot project was modeled after screening programs detailed in the literature and used by many hospital programs. 7-9 Two dedicated hearing screening technicians, with assistance from an audiologist, conducted screenings before discharge. In addition to all neonates at risk for hearing loss 10 and those in the neonatal intensive care unit, approximately 50% of neonates admitted to the newborn nursery were screened. Parents of neonates who did not pass were counseled by an audiologist before discharge about the hearing screening and the results for their baby. They were then given an appointment for pediatric otolaryngologic and audiologic evaluations at Children s Medical Center of Dallas at 1 month of age. If an infant did not return for follow-up, the parents were contacted by telephone and mail to reschedule the appointment. Protocol for Year 1 (September 1999 to August 2000) UNHS began on September 1, To maintain fulltime 24-hour coverage 7 days per week, additional technicians were trained to cover for the dedicated screeners during holidays, vacations, time off, and sick time. Currently, the screening technician pool consists of 14 nursery employees who can rotate into the screening position as needed. Based on the results of the pilot project, the protocol was modified: Neonates who did not pass inpatient screening by a technician were rescreened by an audiologist before discharge. Neonates who were referred on inpatient audiology rescreening from the newborn nursery were scheduled for outpatient screening at the nursery 10 to 12 days after discharge; those who were referred on outpatient screening were provided audiologic counseling and given appointments for otolaryngology and diagnostic audiology appointments. Patients from the NICU did not receive an outpatient rescreen but were referred directly for diagnostic evaluations after referral on inpatient audiology rescreening. Protocol for Years 2, 3, and 4 (September 2000 to August 2003) A technician rescreen was added 24 hours after an initial refer result. An audiologist would conduct a final automated auditory brainstem response before discharge if the neonate did not pass the second technician screen. Based on improvement in follow-up for neonates referred from the newborn nursery, neonates referred from the NICU were scheduled for outpatient rescreening before being referred for diagnostic evaluations. Diagnostic Evaluations for Referred Infants The audiologic evaluation included some or all of the following: click-evoked, tone burst, and bone conduction auditory brain stem response, as well as transient-evoked otoacoustic emissions, as indicated. The diagnostic test battery is designed to categorize the type and degree of hearing loss as well as the site of lesion. Per standard terminology, infants were classified as having sensorineural hearing loss (SNHL), conductive, or mixed hearing loss. Appropriate intervention services, including medical management, amplification, and/ or early intervention services, were recommended on the basis of the type and degree of hearing loss. An additional category of auditory pathology, referred to as auditory neuropathy or dys-synchrony, may also be discovered with this test battery. Patients with auditory neuropathy have present otoacoustic emissions reflecting normal function of the outer hair cells in the cochlea but abnormal auditory nervous system function, as evidenced by a characteristically abnormal auditory brain stem response and/or absent middle ear reflexes. Complicated prognoses The Parkland Memorial Hospital Experience In Ensuring Compliance With Universal Newborn Hearing Screening Follow-Up 67

116 Figure. Diagram of the screening program. Entire diagram reflects the final protocol (years 2 through 4). Portions in white, gray, and black represent components of the screening process derived from the pilot project, year 1, and years 2 through 4 of UNHS, respectively. exist for these patients who can have varied auditory perceptual capabilities despite the physiologic test results. Treatment options also vary. Traditional amplification is often not beneficial for patients with auditory neuropathy/dys-synchrony, but cochlear implantation is a viable treatment option in some. 11 Data Analysis Results are presented for referral rates and hearing loss both unilaterally and bilaterally. Tests for trends over the time periods were conducted with the use of the linear-by-linear association test statistic. 12 Comparisons of outcomes for two different years were accomplished by means of a 2-sided Fisher exact test. 13 SPSS statistical software was used. RESULTS The Parkland newborn population is 79% Hispanic and 15% black; 67.2% have Medicaid and 2.3% have commercial insurance. Maternal age is <20 years in 20% and 95% receive prenatal care. The average length of stay is 42 hours for neonates delivered vaginally without complications and 72 hours for those delivered by C-section in the newborn nursery and 18 days for neonates in the neonatal intensive care unit (8% of the population). Screening results for each year are listed in Table I. Since all neonates in the NICU were screened for all intervals, but only 55% of neonates admitted to newborn nursery were screened for the pilot project, composition of the patient populations differed between the pilot project and years 1 through 4: 11% of the population screened for the pilot project was from the NICU, compared with approximately 6% of the population screened in years 1 through 4. There were no significant differences in the proportion of failed infants between years 2 through 4 (P =.570) or in the loss to follow-up for these years (P =.118). Since the same protocol was used for years 2 through 4, analysis of effects of protocol changes focus on the pilot project, year 1 and year 2. With the addition of predischarge and outpatient rescreening, proportions of patients lost to follow-up over the time intervals significantly declined (P <.0001), primarily because of a higher lost to follow-up rate for the pilot project as compared with years 1 (P =.002) and 2 (P <.0001). The proportion of neonates that did not return for follow-up was not statistically different between years 1 and 2 (P =.458). For the pilot project, 1% of neonates from the newborn nursery failed, compared with 2.4% from the NICU. The same pattern was evident for years 1 and 2, with ~0.3% to 0.4% of neonates from newborn compared to 2% from intensive care requiring diagnostic testing. For the pilot project, lost to follow-up rate was ~40% to 45% for neonates from newborn and from intensive care. For year 1, however, lost to follow-up rate improved for neonates from the newborn nursery to 11.8%, with a smaller improvement for those from intensive care to 31%. The data for year 2 indicates a more equivalent lost to follow-up rate for the nurseries of 11.9% for newborn and 10% for intensive care. Effects of multiple screens on refer rates for outpatient diagnostic services are noted in Table II. Addition of inpatient rescreening by an audiologist reduced the number of neonates requiring outpatient rescreens by 65% for year 1 and 76% for year 2. There was a greater reduction in number of neonates requiring outpatient rescreens through multiple inpatient screens for year 2 than for year 1 (P <.0001). In addition, the audiologists only needed to screen 167 of 456 neonates with initial fail results due to technician rescreening at 24 hours, thus improving on the previous year s utilization of the audiologist (P <.0001). Addition of outpatient rescreening further reduced the number of neonates referred for medical/ audiologic evaluation by 46% for year 1 and 43% for year 2. Thus, the proportion of patients that failed all hearing screenings and were referred for diagnostic evaluations was reduced over time (P <.0001). Post hoc testing comparing two time periods indicated the proportion of referrals in the pilot phase was significantly higher than the proportions in years 1 and 2 (P <.0001); however, the proportion in year 2 did not differ from year 1 (P =.255). Final diagnoses for neonates who completed pediatric otolaryngologic and diagnostic audiologic evaluation are listed in Table III. Prevalence of bilateral SNHL in our population is approximately 1 in 1000 live births. This does not include 5 neonates who were referred bilaterally but were lost to followup for both year 1 and year 2. If it is assumed that these neonates had bilateral SNHL, prevalence estimates increase to 68 Shoup et al The Journal of Pediatrics January 2005

117 Table I. Summary of results from the pilot project and UNHS in years 1, 2, 3, and 4 Pilot project Year 1 Year 2 Year 3 Year 4 N (NBN/ NICU) u % N (NBN/ NICU) u N (NBN/ NICU) u N (NBN/ NICU) u % N (NBN/ NICU) u % 1.14% y 70 f 0.46% y 61 f 0.37% y 54 f 0.33% y 59 f 0.37% y 0.42% y % y % y % y % y 0.72% y % y % % y % y 44% z 12 17% z % z 6 11% z 1 1.7% 31% z 5 17% z 5 17% z 3 15% z 52% z 7 17% z 2 6% z 3 8.8% z 1 3% Total births , ,449 16,099 15,943 Not screened % * % * % * % * Died % * % * % * % * Transferred % * % * % * % * Discharged % * % * % * % * No. screened 3, % 15, % * 16, % * 16, % 15, % (3344/415) u (14,355/942) u (15,478/906) u (14,990/1016) u (14,918/903) u Failed on HS f 43 f (33/10) u (51/19) u (42/19) u (32/22) u (43/16) u Failed bilaterally 16 (11/5) u (21/9) u (17/12) u (11/9) u (16/10) f Failed unilaterally 27 (22/5) u (30/10) u (25/7) u (21/13) u (27/6) u Lost to follow-up 19 (15/4) u (6/6) u (5/2) u (3/3) u (1/0) u Bilateral 5 (4/1) u (1/4) u (3/2) u (2/1) u Unilateral 14 (11/3) u (5/2) u (2/0) u (1/2) u (1/0) u Failed (died) % z 1 1.6% z 4 7.4% z 1 1.7% z u NBN/NICU refers to the number of neonates derived from the newborn nursery and the neonatal intensive care unit, respectively. *Percentage based on total number of births. ypercentage based on number screened. Failed on HS f, Total failed on hearing screening for diagnostic audiologic evaluation. zpercentage based on number of category specific fails, that is, percentage lost to follow-up is computed based on total number of fails; percentage bilateral lost-to-follow-up is based on bilateral fails and percentage unilateral lost to follow-up is based on unilateral fails. approximately 1.5 in If all auditory impairments are considered, prevalence is approximately 3 in 1000 live births. Positive predictive values (PPV) of screening at each level are presented in Table IV. As noted, the populations screened differed for the pilot project and years 1 and 2. Data from years 1 and 2 reflect the total delivery population, whereas the pilot project includes all patients from the NICU and only 55% of those from the newborn nursery. The higher PPV for the pilot phase compared with years 1 and 2 presumably reflects the larger proportion of infants from intensive care with a greater disposition to hearing deficits. The higher the PPV, the lower the number of neonates referred for diagnostic testing who do not have hearing loss. For example, the derived PPV for the initial screening for years 1 and 2 predicts that 1 in 16 to 28 patients who did not pass would have bilateral SNHL. Alternatively, inclusion of all types of hearing loss for the initial screen for years 1 and 2 indicates that ~1 in 7 to 9 neonates who failed would have some auditory impairment. At discharge from the hospital, the PPV for years 1 and 2 improved, signifying that 1 in 6 to 7 neonates who failed will have bilateral SNHL and 1 in 2.5 will have some auditory impairment. Addition of outpatient rescreening yielded a PPV predicting that 1 in 3 to 4 neonates who failed will have bilateral SNHL and 1 in 1.4 will have some auditory impairment. Specificity of newborn hearing screens was uniformly high. For years 1 and 2, initial specificity was approximately 98% for bilateral SNHL and all auditory impairments. This improved to greater than 99% by discharge from the hospital. DISCUSSION Regardless of the size of the hospital, successful UNHS requires constant vigilance, maintenance, planning, and revision of the program to meet target goals. Return of patients for additional testing after discharge is challenging because of mobile populations, variation of providers across hospital and outpatient settings, financial barriers, and time/ effort needed to track neonates who do not pass. Therefore, minimizing false-positive results is of utmost importance. Repeated hearing screenings at planned intervals reduced the number of neonates requiring follow-up by 65% to 77%. The rationale for the multiple screens performed in our UNHS program reflects multiple considerations. To screen as close to 100% of newborn infants as possible, the initial hearing screen was done at approximately 4 hours after birth, The Parkland Memorial Hospital Experience In Ensuring Compliance With Universal Newborn Hearing Screening Follow-Up 69

118 Table II. Effect of inpatient and outpatient rescreening on the number of neonates referred for diagnostic evaluation for the pilot project and years 1 through 4 Pilot Year 1 Year 2 Year 3 Year 4 Failed on initial 43 (1.14) * 365 (2.39) 456 (2.78) 541 (3.38) 737(4.66) inpatient HS 1 Failed on 24-hour inpatient HS Not done Not done 167 (1.02) 184 (1.15) 239 (1.51) Failed on audiology Not done 129 (0.84) 108 (0.66) 81 (0.51) 110 (0.69) inpatient HS Failed on audiology Not done 70 (0.46) 61 (0.37) 54 (0.34) 59 (0.37) outpatient HS Scheduled for diagnostic evaluation 43 (1.14) 70 (0.46) 61 (0.37) 54 (0.34) 59 (0.37) *N (percent based on total neonates screened). HS, hearing screen. Table III. Diagnostic results for neonates seen for pediatric otolaryngologic and audiologic evaluation for the pilot project, and years 1, 2, 3, and 4 Pilot Year 1 Year 2 Year 3 Year 4 Dx outcome N % * N % * N % * N % * N % * WNL Unilateral Conductive mild Conductive moderate Sensorineural mild Sensorineural moderate Sensorineural severe Sensorineural profound Mixed severe Auditory neuropathy Bilateral Conductive mild Conductive moderate Sensorineural mild Sensorineural moderate Sensorineural severe Sensorineural profound Mixed moderate Mixed severe Auditory neuropathy Lost to follow-up Failed (died) Total *Percentage based on total number of patients who failed. Mild = 20 to 40 db; moderate = 40 to 70 db; severe = 70 to 90 db; profound = 911 db. when all infants in the newborn nursery have completed a 4-hour observation period in a central admissions nursery to ensure proper adaptation. Infants in the newborn nursery are subsequently cared for in either multiple nurseries or the mothers rooms. Early screening may contribute to a high refer rate, since infants may have debris in the external auditory canal and/or middle ear fluid. This may resolve within the first 24 hours of life and provides the rationale for a rescreen by a technician after 24 hours. Utilization of the technician for this initial rescreen reduces the number referred for screening by an audiologist. Other reasons for infants to initially refer may include incorrect earphone placement and/or electrode placement and other artifacts related to the hearing screening technology, environment, or patient state. Rescreens by the 70 Shoup et al The Journal of Pediatrics January 2005

119 Table IV. Positive predictive values (PPV %) of screenings for bilateral sensorineural hearing loss and any auditory impairment Bilateral fails only/bilateral SNHL All fails/bilateral SNHL All auditory impairments Pilot Year1 Year2 Pilot Year1 Year2 Pilot Year1 Year2 Initial inpatient screen Hour rescreen ND ND 21.6 ND ND 9.6 ND ND 28.7 Audiology inpatient screen ND ND ND Outpatient rescreen ND ND ND Percent PPV determined by dividing number of neonates with bilateral SNHL or auditory impairment who failed the screening by the total number of fails. ND, Not done. audiologist shortly before discharge ensure that such technical issues are not contributing to a false-positive result. The outpatient rescreen was incorporated into our protocol to address the issues of incomplete resolution of middle ear fluid before hospital discharge in some infants. Initial screenings and rescreenings for infants in the NICU are done on a different schedule dictated by the infant s status and gestational age. Our experience indicates that allocation of appropriate resources on the front end leads to success in reducing falsepositive results and improving compliance with follow-up, thus focusing on the smaller number of neonates truly at risk for auditory impairment. This not only decreases the burden on already strained healthcare resources but also limits the number of families that must deal with the uncertainty of hearing loss in their newborn infants. Furthermore, healthcare providers view refer results on hearing screening more seriously in a program with few false-positive results and may therefore be more effective in encouraging parents to return for services. After our program was in place, the State of Texas released recommendations/rules for UNHS. According to the guidelines, the primary care provider is responsible for ensuring return for follow-up for those infants who do not pass a newborn hearing screening. The PMH team elected to continue to track the infants from our program and coordinate follow-up. Although infants are assigned to a follow-up clinic and/or pediatrician before discharge, in some cases the parents elect to take the infant to another facility. In such cases, primary care providers may not always be aware of newborn hearing screening results. Further, some primary care providers may not be aware of resources available for followup testing and intervention services. Keeping the coordination of follow-up services for UNHS programs at the birthing facility allows for improved continuity of care. PPVs for the UNHS program at PMH surpass those reported in the literature. 7,14,15 The Colorado program reported PPVs that ranged from 5% to 19% for bilateral SNHL across 26 hospitals involved in UNHS for 1992 to The PPV for bilateral SNHL for our facility was 77% and 55% for years 1 and 2, respectively. This means that 5.3 to 20 neonates would be referred in the Colorado program for each identified with SNHL. With the repeat hearing screenings provided in our program, only 1.3 to 1.8 neonates would be referred for diagnostic testing to identify an infant with bilateral SNHL. The UNHS program exceeded guidelines suggested by the American Academy of Pediatrics task force (AAP). 6 The AAP suggested that a hospital-based UNHS program should screen a minimum of 95% of neonates with a false-positive rate of <3% and a refer rate for formal audiologic testing not to exceed 4%. At PMH, >99% of neonates were screened with a falsepositive rate of <1% by hospital discharge and refer rate for diagnostic audiologic testing of <0.5%. This was feasible even with a delivery population of approximately 16,000 births per year. For identification of bilateral SNHL, the AAP suggested that 95% should return for follow-up. Follow-up was achieved in 83% for neonates who were referred bilaterally in our population by year 2 and had improved further to 98.3% by year 4. Much of the data in the literature on follow-up rates are reported for neonates who are referred bilaterally. Unilateral hearing loss may be associated with cognitive deficits, academic difficulties, and behavioral adjustment issues Given these considerations, we follow neonates who are referred unilaterally because our patients are predominantly Hispanic, many with English as a second language, and are therefore considered a high-risk population Thus, the performance goal of the program was >90% return for follow-up for neonates with any auditory impairment, unilateral or bilateral. By year 4, that goal was reached. Our return for follow-up improved from 56% for the pilot project to 98.3% by year 4 of UNHS. Through changes in protocol, the UNHS program improved over time. Although it is not known if there would be a large cost differential between the three protocols, such an analysis is beyond the scope of the current paper due to the variety of assumptions that must be addressed. For example, cost per child identified with hearing loss, cost per screening based on different protocols, and lost opportunity cost will alter the conclusions. Experience at PMH demonstrates that UNHS can be successfully implemented in a public hospital with a larger number of annual births than many US states and territories, including Alaska, Delaware, the District of Columbia, Maine, Montana, New Hampshire, North Dakota, Rhode Island, South Dakota, Vermont, Wyoming, the Virgin Islands, and Guam. 27 With the use of repeated hearing screens, including an The Parkland Memorial Hospital Experience In Ensuring Compliance With Universal Newborn Hearing Screening Follow-Up 71

120 outpatient rescreen at the birthing facility, false-positive rates can be decreased, use of audiologists time and resources can be improved, and compliance with follow-up can be enhanced. To maximize efficiency, this is accomplished with a carefully planned program using hearing screening technicians overseen and managed by an audiologist. Factors that may be associated with improvement in return for follow-up include scheduling of patients at the birthing hospital for the initial follow-up appointment. In addition, with successful reduction of falsepositive results, other health care providers are more supportive of encouraging their patients to return for follow-up testing. Accomplishment of successful UNHS requires a vested staff and close attention to detail, a family-centered approach, comprehensive coordination of provider services, and commitment to ongoing quality assurance/improvement. We thank the rest of the Parkland Memorial Hospital UNHS planning and oversight team for their valuable contributions: Dale Talley, CPNP, Pam Ford, RN, Cheryl Aldridge, CPNP, and Janet Cover, CPNP. Our dedicated and talented newborn hearing screening technicians are appreciated for their commitment to the program. We also thank the National Council of Jewish Women and their faithful volunteers for financial assistance in establishing the high-risk and NICU screening programs in 1986 and their continued commitment in both time and funds to the program. The Hoblitzelle Foundation made the transition to UNHS possible with a challenge grant to Parkland. Successful UNHS at Parkland would not be possible without the continued support from the newborn nursery, neonatal intensive care unit, and Parkland administration. Appreciation is extended to William Frawley, PhD, for assistance with and review of data analysis. REFERENCES 1. Yoshinaga-Itano C, Coulter D, Thomson V. The Colorado Newborn Hearing Screening Project: effects on speech and language development for children with hearing loss. J Perinatol 2000;20:S Yoshinaga-Itano C, Sedey AL, Coulter DK, Mehl AL. Language of earlyand later-identified children with hearing loss. Pediatrics 1998;102: Apuzzo M-RL, Yoshinaga-Itano C. Early identification of neonates with significant hearing loss and the Minnesota child development inventory. Semin Hear 1995;16: Helfand M, Thompson DC, Davis R, McPhillips H, Homer CJ, Lieu TA. Newborn Hearing Screening. Systematic Evidence Review No. 5 (Contract to the Oregon Health and Science University Evidence-based Practice Center, Portland, Oregon). AHRQ Publication No. 02-S001. Rockville, Md: Agency for Healthcare Research and Quality. October Berg AO, Allan JD, Frame PS, Homer CJ, Johnson MS, Klein JD, et al. Newborn Hearing Screening: Recommendations and Rationale. US Preventive Services Task Force. AHRQ Publication No A. Rockville, Md: Agency for Healthcare Research and Quality; October Erenberg A, Lemons J, Sia C, Tunkel D, Ziring P, Adams M, et al. Newborn and neonate hearing loss: detection and intervention: American Academy of Pediatrics, Task Force on Newborn and Neonate Hearing, Pediatrics 1999;103: Mehl AL, Thomson V. Newborn hearing screening: the great omission. Pediatrics 1998;101: McDaniel SL, Olguin M, Horn KL. Hear early: New Mexico s Universal Newborn Hearing Screening Program. Otolaryngol Clin North Am 1999;32: Mason JA, Herrmann KR. Universal infant hearing screening by automated auditory brain stem response measurement. Pediatrics 1998;101: Joint Committee on Infant Hearing. Joint Committee on Infant Hearing 1994 Position Statement. Pediatrics 1994;95: Rapin I, Gravel J. Auditory neuropathy : physiologic and pathologic evidence calls for more diagnostic specificity. Int J Pediatr Otorhinolaryngol 2003;67: Agresti A. Categorical data analysis. New York: John Wiley & Sons; Hollander M, Wolfe DE. Nonparametric statistical methods. New York: John Wiley & Sons; Prieve B, Dalzell L, Berg A, Bradley M, Cacace A. The New York State universal newborn hearing screening demonstration project: outpatient outcome measures. Ear Hear 2000;21: Vohr BR, Carty LM, Moore PE, Letourneau K. The Rhode Island Hearing Assessment Program: experience with statewide hearing screening ( ). J Pediatr 1998;133: Culbertson JL, Gilbert LE. Children with unilateral sensorineural hearing loss: cognitive, academic and social development. Ear Hear 1986;7: Bess FH, Tharpe AM. An introduction to unilateral sensorineural hearing loss in children. Ear Hear 1986;7: Klee TM, Davis-Dansky E. A comparison of unilaterally hearingimpaired children and normal-hearing children on a battery of standardized language tests. Ear Hear 1986;7: Bess FH, Tharpe AM, Gibler AM. Auditory performance of children with unilateral sensorineural hearing loss. Ear Hear 1986;7: Bess FH, Tharpe AM. Case history data on unilaterally hearingimpaired children. Ear Hear 1986;7: Fradd SH, Correa VI. Hispanic students at risk: do we abdicate or advocate? Except Child 1989;56: Cohen OP, Fischgrund JE, Redding R. Deaf children from ethnic, linguistic and racial minority backgrounds: an overview. Am Ann Deaf 1990; 135: Cummins J. A theoretical framework for bilingual special education. Except Child 1989;56: Fischgrund JE, Cohen OP, Clarkson RL. Hearing-impaired children in black and Hispanic families. Volta Rev 1987;89: Lee DJ, Gomez-Marin O, Lee HM. Sociodemographic and educational correlates of hearing loss in Hispanic children. Paediatr Perinat Epidemiol 1997;11: Walker-Vann C. Profiling Hispanic deaf students: a first step toward solving the greater problems. Am Ann Deaf 1998;143: National Vital Statistics Report. Vol 49, No. 5, July 24, Shoup et al The Journal of Pediatrics January 2005

121 CAN CLINICAL SIGNS IDENTIFY NEWBORNS WITH NEUROMUSCULAR DISORDERS? ISABELLA VASTA,MD,MARIA KINALI,MD,SONIA MESSINA,MD,ANDREA GUZZETTA,MD,OLGA KAPELLOU,MD,ADNAN MANZUR,MD, FRANCES COWAN,MD,FRANCESCO MUNTONI,MD,AND EUGENIO MERCURI,MD Objective To evaluate retrospectively the prevalence of neuromuscular disorders in 83 newborns referred to a tertiary care center because of hypotonia and weakness and/or contractures, with a possible diagnosis of neuromuscular disorder. We also aimed to establish whether clinical signs could help to identify infants with neuromuscular disorders. Study design Sixty-six of the 83 infants who fulfilled the inclusion criteria (79.5%) had an identifiable disorder, which was a neuromuscular disorder in 39 (46.9%). Results Absent or extremely reduced antigravity movements were mainly found in infants with neuromuscular disorders (sensitivity and specificity 97.4% and 75%), whereas partial range antigravity movements were more frequent in infants with other diagnosis. Contractures were mainly found in infants with peripheral nerve or muscle involvement but also were relatively frequent in infants with genetic or metabolic syndromes (sensitivity 69.2%, specificity 61.3%). Reduced fetal movements and abnormal liquor were frequent but not present consistently in infants with neuromuscular disorders (sensitivity 46.1% and 38.4%) and were found rarely in infants with other disorders (specificity 88.6% and 75.0%). Conclusions Severe muscle weakness and contractures are the most reliable indicators of a neuromuscular disorder and should be carefully assessed in an infant with neonatal hypotonia. (J Pediatr 2005;146:73-9) Neonatal hypotonia is a nonspecific clinical sign frequently associated with both central nervous system (CNS) and peripheral nervous system abnormalities. A few studies have systematically evaluated infants with neonatal hypotonia and found that this was caused by peripheral nerve involvement in one third of subjects, whereas the remaining two thirds had evidence of CNS involvement, including metabolic and genetic causes. 1 In the early 80s, Dubowitz 2 suggested that a clinical examination could help to distinguish infants with primary neuromuscular disorders from those with CNS involvement. Although contractures or obvious signs of weakness with absent or reduced antigravity movements on stimulation of the limbs suggest a neuromuscular disorder, reduced visual alertness, convulsions, and abnormal movements are generally suggestive of more global CNS dysfunction. 2 This view led to many infants with weakness and contractures being referred to a tertiary care center for further neuromuscular investigations. The possibility of arriving at a diagnosis in a newborn who presents with these clinical signs has been enormously facilitated in the last decades by the advent of electrophysiological, imaging, metabolic, and genetic tests. As the number of identifiable neuromuscular, genetic, and metabolic conditions for which a specific test can be performed has increased, it has become all the more important to identify clinical signs that can help the clinician to target specific investigations. 3,4 The aim of this study was to evaluate retrospectively the prevalence of neuromuscular disorders in a population of newborns referred to a tertiary care center with this as a suggested diagnosis. We also aimed to establish the extent to which the presence of neonatal clinical and instrumental signs can help in the differential diagnosis and reliably identify neuromuscular disorders. CK CNS Creatine kinase Central nervous system MRI US Magnetic resonance imaging Ultrasonography From Dubowitz Neuromuscular Centre, Department of Paediatrics, Imperial College, Hammersmith Hospital, London, UK; and the Department of Pediatrics, Pediatric Neurology Unit, Catholic University, Rome, Italy. Submitted for publication Mar 30, 2004; last revision received Jun 10, 2004; accepted Aug 18, Reprint requests: Eugenio Mercuri, MD, Department of Paediatrics, Hammersmith Hospital, Du Cane Road, London W12 OHN, UK. e.mercuri@imperial.ac.uk /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

122 Table I. Specific diagnosis reached in our cohort Genetic (n = 8) Metabolic (n = 8) Antenatal injurious (n = 4) Prader-Willi syndrome (n = 4) Peroxisomal disorders (n = 3) Trauma (car accident second trimester) (n = 1) Skeletal dysplasia (n = 1) Nonketotic hyperglycinemia (n = 3) Infection (n = 1) Multiple pterigia syndrome (n = 1) Pyruvate dehydrogenase deficiency (n = 1) Perinatal injury (n = 2) Systemic hyalinosis (n = 1) Respiratory chain defect disorder (n = 1). Aicardi Goutieres syndrome (n = 1) SUBJECTS AND METHODS This was a retrospective study of infants born at or referred to a tertiary care center for neuromuscular disorders (Hammersmith and Queen Charlotte s Hospital, London) in the first weeks after birth between January 1998 and May Infants were enrolled only if they had hypotonia associated with muscle weakness (revealed by absent or reduced antigravity movements spontaneously or on stimulation) and/or contractures that suggested a neuromuscular disorder. Infants with hypotonia and clear signs of CNS involvement without weakness, contractures, or other signs of possible neuromuscular involvement were not included. Case notes were evaluated, and antenatal and perinatal information was collected. This included the presence of consanguinity and family history, poly/oligohydramnios, reduced fetal movements, Apgar scores, need for respiratory support at birth and in the neonatal period, feeding difficulties, and seizures. Details of the clinical examination, including muscle tone, movements, reflexes, and visual behavior also were noted, together with possible dysmorphic signs or abnormalities outside the nervous system. Muscle tone and reflexes were assessed using structured neonatal neurological examinations that provide age-dependant normative data. 5,6 Abnormal movements were classified as: absent, if no movement could be detected spontaneously or on stimulation; severely reduced, if only flickers of movements could be observed; and partially reduced, if the infant had some movements, spontaneously or on stimulation, but these were not full range antigravity movements. We also documented all the investigations performed. These included serum creatine kinase (CK), metabolic investigations, karyotype and other genetic tests, neurophysiology (electromyography and nerve conduction velocity studies), muscle or nerve biopsy, and neuroimaging (cranial ultrasonography [US] and, when available, brain magnetic resonance imaging [MRI]). Statistical Analysis The sensitivity and specificity of individual signs to identify neuromuscular disorders was calculated. RESULTS A total of 83 infants fulfilled the inclusion criteria. Of those 83 infants, 49 (59.0%) were male. The mean gestational age at birth was weeks (range ) Of those 83 infants, 38 (45.7%) presented with hypotonia and weakness, 6 (7.2%) with hypotonia and contractures, and 39 (46.9%) with a combination of the three. In all, diagnosis at presentation was that of neonatal hypotonia and of a possible neuromuscular disorder. Sixty-six of the 83 infants (79.5%) had an identifiable disorder, which was a neuromuscular disorder in 39 (39/ 83 = 46.9%). Genetic syndromes were found in 8 (8/ 83 = 9.6%), and specific metabolic disorders in another 8 (8/83 = 9.6%). Other diagnoses were a form of severe encephalopathy with peripheral nervous system involvement (7/83 = 8.4%) and acquired disorders following antenatal brain injury (4/83 = 4.8%) (Tables I and II). The remaining 17 infants (17/83 = 20.4%) without specific diagnosis could be classified into the following groups: prominent bulbar symptoms (4/83 = 4.8%), possible metabolic disorder (7/ 83 = 8.4%), and extrapyramidal syndrome (2/83 = 2.4%). Four infants (4/83 = 4.8%) had suspected genetic syndromes with associated dysmorphic features, but no final diagnosis could be reached. Tables I and II give details of the specific diagnosis within each group. Consanguineous Pedigree Consanguineous pedigrees were identified in 12 infants (12/83 = 14.4%) in our study population and were more frequently found in the infants with a metabolic disorder (4/8 = 50%). In the neuromuscular group, consanguineous pedigrees were found in only 3 of the 39 infants (7.6%). Family History A positive family history for occurrence of a similar condition was found in 17 families (17/83 = 20.4%). Positive family histories were found more frequently in the infants with an encephalopathy with neuropathy (5/7 = 71.4%). In the neuromuscular group, positive family history was found in only 9 of the 39 infants (23.0%). Antenatal Findings Reduced fetal movements were found in 23 (23/ 38 = 27.7%) infants and were more frequent in the infants with a neuromuscular disorder (18/39 = 46.1%). Polyhydramnios was found in 23 (23/38 = 27.7%) of pregnancies and was noted more frequently in infants with a neuromuscular disorder (14/39 = 35.8%). 74 Vasta et al The Journal of Pediatrics January 2005

123 Oligohydramnios was found in three pregnancies of syndromic infants and in one infant with a congenital myopathy. Intrauterine Growth Retardation and Head Circumference (Occipito-Frontal Circumference) Birth weight below the 10th centile was found in 16 (16/ 83 = 19.2%) infants and was more frequent in the infants with a neuromuscular disorder (9/39 = 23%) and in the infants with antenatal acquired injury (1/4 = 25%). None of the infants had birth occipito-frontal circumference (OFC) below the 3rd centile. Respiratory Problems Neonatal respiratory problems were present in 57 infants (57/83 = 68.6%). Of those 57, 12 (21.0%) required only bag and mask ventilation at birth, 11 (19.2%) required nasal continuous positive airway pressure (CPAP), and 34 (59.6%) intubation and ventilation at birth or within the first 48 hours after birth. Respiratory problems were more frequent in infants with encephalopathy and neuropathy (7/7 = 100%). In the neuromuscular group, 27 infants (27/39 = 69.2%) had respiratory problems at birth, with 18 of those infants (18/ 39 = 46.1%) requiring intubation and ventilation. Respiratory problems were also frequent in the other groups. Feeding Difficulties Feeding difficulties of sufficient severity to necessitate the use of a nasogastric tube were present in 75 infants (75/ 83 = 90.3%) and were frequent in all groups. In the neuromuscular group, 34 infants (34/39 = 85.1%) required nasogastric tube feeding. Dysmorphic Features Dysmorphic features were present in 35 (35/ 83 = 42.1%) infants. Multiple dysmorphic features were more frequent in the infants with a recognized (5/35 = 14.2%) or suggested (2/35 = 5.7%) genetic syndrome. Multiple dysmorphic features also were found in 4 infants with neuromuscular disorder (4/35 = 11.4%). Mild dysmorphic features were noted in 17 infants (17/ 35 = 48.5%) and were more frequent in infants with a diagnosis of encephalopathy with neuropathy and a neuromuscular disorder. These included long fingers, micrognathia, low-set ears, high arch palate, and absent dermatoglyphics. Contractures Contractures at birth were noted in 44 infants (44/ 83 = 53.0%) and were more frequent in infants with a neuromuscular disorder (27/39 = 69.2%) and in those with encephalopathy and neuropathy (54/7 = 57.1%). Both isolated (mainly talipes) and multiple contractures could be found equally in neuromuscular disorders and in other diseases. Visual Behavior Poor alertness and/or other signs of encephalopathy were present in 31 infants (31/83 = 37.3%). Only 6 infants (6/39 = 15.3%) with a neuromuscular disorder had poor alertness. Seizures Seizures occurred in the neonatal period in 15 infants (15/83 = 18.0%) and were more frequent in patients with metabolic diagnosis (4/8 = 50%). Only 2 infants with neuromuscular diagnosis had seizures (2/39 = 5.1%). Cranial US/MRI Cranial US and/or MRI in the neonatal period were performed in 73 infants (73/83 = 87.9%) and were abnormal in 49 (49/73 = 67%) infants. Abnormalities were found in 100% of infants who had an antenatally acquired brain injury. US or MRI abnormalities were found in 48.7% (19/39) of infants with neuromuscular disorders. Table III gives details of the cranial US/MRI abnormalities within each group. Creatine Kinase CK levels were available in all infants and were persistently raised in 5 infants. All of them had a neuromuscular disorder. Sensitivity and Specificity Absent or markedly reduced antigravity movements had the highest sensitivity and specificity for neuromuscular disorders (0.97 and 0.75, respectively). Reduced fetal movements and polyhydramnios had very high specificity (0.88 and 0.75, respectively) but lower sensitivity. Table IV shows details of sensitivity, specificity, and positive and negative predictive values for each of the clinical parameters evaluated with respect to a neuromuscular diagnosis. The Figure shows a flow chart developed according to the information achieved in our cohort. Poor Antigravity Movements Extremely reduced (in 33) or absent (in 16) antigravity movements were documented in 49 infants (49/83 = 59.0%). Partial range antigravity movements were found in 18 (18/ 83 = 22%) infants. In the neuromuscular group, all but one infant had extremely reduced or absent antigravity movements (38/39 = 97.4%). DISCUSSION The aim of this study was to establish the relative accuracy of clinical examination in the diagnosis of infants referred to a tertiary care center because of a suspected neuromuscular disorder. The main reason for referral was that these infants were not only hypotonic, but they also had contractures and/or weakness that suggested a neuromuscular disorder. It is well Can Clinical Signs Identify Newborns With Neuromuscular Disorders? 75

124 Table II. Clinical findings within each group of our cohort Consang. pedigree (12) Positive family history (17) Antenatal findings Oligo/ Polyhydr. (26) Neonatal presentation RMF (23) H+W H+C H+W+C NMD (39) CMD (8) 1 (8%) 1 (6%) 1 (4%) 6 (26%) 2 (2%) - 6 (7%) C myopathy (16) 2 (17%) 3 (18%) 8 (31%) 7 (31%) 5 (6%) - 11 (13%) C myotonic dys (6) - 4 (23%) 4 (15%) 1 (4%) 1 (1%) - 5 (6%) WWS (2) - 1 (6%) - 2 (9%) (2%) Others (7) (8%) 2 (9%) 4 (5%) 1 (1%) 2 (2%) Genetic Syndromes (8) 2 (17%) - 4 (15%) 2 (9%) 5 (6%) 3 (4%) - Metabolic (8) 4 (33%) 2 (12%) - 1 (4%) 5 (6%) 1 (1%) 2 (2%) Anten. Injurious (4) 1 (8%) (4%) 2 (2%) - 2 (2%) Encephalopathy + neuropathy (7) 2 (17%) 5 (29%) 2 (8%) - 3 (4%) - 4 (5%) OTHER (17) - Bulbar (4) (4%) 1 (4%) 3 (4%) - 1 (1%) - Suggested metabol. (7) - 1 (6%) 3 (11%) - 4 (5%) 1 (1%) 2 (2%) - Suggested genet. (4) (4%) - 2 (2%) - 2 (2%) - Extrapyramidal s. (2) (2%) - - C, contractures; CMD, congenital muscular dystrophy; H, hypotonia; IPPV, intermittent positive pressure ventilation; NCPAP, nocturnal continuous positive airway pressure; NMD, neuromuscular disorder; PRAG, partial range antigravity; RMF: reduced fetal movement; W, weakness; WWS, Walker-Warburg syndrome. Table III. Clinical and imaging findings of CNS involvement within each group known that weakness in neonates can be assessed by looking for a reduction of antigravity movements, and that poor antigravity movements are an indicator of a neuromuscular disorder. 2 In this study we tried to establish whether a more detailed observation and classification of these movements would be Brain imaging Poor alertness Seizures Mild Moderate/Severe NMD (39) - CMD (8) cystic PVL - C myopathy (16) infarct and cortical dysplasia, 1 periventricular WM changes - C myotonic dys. (6) WWS (2) lissencephaly - Others (7) Genetic syndromes (8) diffuse WM changes Metabolic (8) diffuse WM changes, 1 polymicrogyria Injurious (4) abnormal myelination 1 damage in basal ganglia 1 bilateral infarct 1 schizencephaly Encephalopathy diffuse WM changes peripheral neuropathy (7) OTHER (17) - Bulbar (4) diffuse WM changes - Suggested metabol. (7) Suggested genet. (4) pontocerebellar hypoplasia, 1 lissencephaly - Extrapyramidal syndrome (2) CMD, congenital muscular disorder; NMD, neuromuscular disorder; PVL, periventricular leucomalacia; WM, white matter; WWS, Walker-Warburg syndrome. useful. Although one cannot apply to newborns the Medical Research Council Scale or other scales that are used to assess strength in older children, we still found it possible to establish a grading of muscle power, distinguishing infants who did not have full antigravity movements but who still had some partial 76 Vasta et al The Journal of Pediatrics January 2005

125 Apgar <5 (34) Respiratory problems (57) Feeding difficulties Dysmorphism (35) Antigravity movement Mask NCPAP IPPV (75) Mild Severe Absent Poor PRAG 3 (9%) 1 (2%) 2 (3%) 1 (2%) 5 (7%) 1 (3%) - 2 (3%) 6 (9%) - 6 (17%) 2 (3%) 2 (3%) 8 (14%) 16 (21%) 4 (11%) 2 (6%) 6 (9%) 10 (15%) - 4 (12%) 1 (2%) - 5 (8%) 6 (8%) 1 (3%) - 1 (1%) 5 (7%) - 2 (6%) (3%) 2 (3%) - 2 (6%) 2 (3%) (6%) 1 (2%) - 2 (3%) 5 (7%) (1%) 5 (7%) 1 (1%) 2 (6%) 1 (2%) 2 (3%) - 5 (7%) 3 (8%) 5 (14%) - 2 (3%) 5 (7%) 2 (6%) 2 (3%) - 3 (5%) 8 (11%) 1 (3%) 1 (3%) - 1 (1%) 2 (3%) 3 (9%) - 1 (2%) 2 (3%) 4 (5%) - 2 (6%) - 2 (3%) 2 (3%) 5 (14%) (12%) 7 (9%) 3 (8%) 3 (8%) 1 (1%) 1 (1%) 3 (4%) 1 (3%) 1 (2%) 1 (2%) - 4 (5%) 1 (3%) (1%) 2 (6%) 2 (3%) 2 (3%) 2 (3%) 7 (9%) 2 (6%) 1 (3%) - 1 (1%) 3 (4%) 2 (6%) 1 (2%) 1 (2%) 2 (3%) 4 (5%) - 2 (6%) 3 (4%) (3%) 1 (3%) (1%) range movements from those in whom there were no movements or only flickers of movements even after stimulation. As the presence of reduced antigravity movements was one of our main inclusion criteria it is not surprising that they were observed in 67 out of the 83 infants included in the study. Absent or extremely reduced antigravity movements, however, were only observed in 49 infants, whereas the other 18 had partial range antigravity movements. Although absent or extremely poor antigravity movements were mainly found in infants with neuromuscular disorders (sensitivity and specificity; 97.4% and 75 %, respectively) partial range antigravity movements were more frequent in patients with other diagnosis. Contractures also were mainly found in infants with primary neuromuscular disorder, but sensitivity and specificity were lower than for weakness (sensitivity 69.2%, specificity 61.3%) as contractures also were relatively frequent in infants with genetic or metabolic syndromes or in those who suffered an injury to the brain in the first or second trimester in pregnancy. The number of joints affected by contractures did not help in differential diagnosis as both isolated (mainly talipes) and multiple contractures could be found equally in neuromuscular disorders and in other diseases Reduced fetal movements and abnormal liquor volume were frequent but not consistently present in infants with neuromuscular disorders (sensitivity 46.1% and 38.4%, respectively) but were rarely found in infants with other disorders (specificity 88.6% and 75.0%, respectively). The need for respiratory assistance and the occurrence of feeding difficulties were often observed in infants with neuromuscular disorders, but these had low specificity as they were also frequent in other diseases. Table IV. Sensitivity, specificity, and positive and negative predictive values for each of the clinical parameters evaluated with respect to a neuromuscular diagnosis Sensitivity (%) Specificity (%) PPV (%) PNV (%) Extremely reduced/ absent AG movement Partial range AG movement Contractures RFM Oligopolydramnios Apgar, Respiratory problems at birth Feeding difficulties Poor alertness Seizures Dysmorphism AG, antigravity; PPV, positive predictive value; NPV, negative predictive value; RMF, reduced fetal movement. Our findings suggest, therefore, that weakness and contractures, especially when associated with reduced fetal movements and abnormal liquor volume, are strongly indicative of a neuromuscular disorder. The recognition of these signs is even more important when they are found in infants with clinical signs of CNS involvement and brain lesions. Can Clinical Signs Identify Newborns With Neuromuscular Disorders? 77

126 Figure. Diagnostic algorythm in infants with hypotonia, weakness and/or contractures. Nineteen of the 39 infants (48%) with a neuromuscular disorder also had some abnormalities on brain imaging, and many also had seizures or poor alertness. Brain abnormalities were more frequently in the form of minor changes such as nonspecific periventricular changes, which are not uncommon in infants with weak respiratory muscles. However, in 5 infants the abnormalities were marked, and definitely structural in 3 of the 5. In these infants, a detailed clinical examination suggested a concomitant central and peripheral involvement. Two infants in our cohort had Walker Warburg syndrome, a form of congenital muscular dystrophy with structural brain changes that were recently mapped to chromosome 9. 7 Both infants had reduced alertness and/or seizures and predominant signs of CNS involvement. Brain MRI showed abnormal cortical folding (lissencephaly), white matter changes, and severe ventricular dilatation. In both infants, however, the presence of weakness and contractures raised the suspicion of a neuromuscular involvement, which was further supported by high serum CK levels. This prompted a muscle biopsy, which showed pathological and immunohistochemical changes in keeping with a diagnosis of Walker Warburg syndrome. Similarly, the presence of distal weakness and contractures in 7 infants with encephalopathy suggested a peripheral neuropathy, which was confirmed by neurophysiological studies showing a marked reduction of nerve conduction velocity. A similar phenotype with enceph- 78 Vasta et al The Journal of Pediatrics January 2005

127 alopathy and neuropathy has been previously reported, 8 but the aetiology of this form is still unknown. In our group, infants were tested for mitochondrial, lysosomal, peroxisomal, carbohydrate glycosylation disorders, and with UV radiation studies, but the results of these tests were all normal (manuscript in preparation). Our results, therefore, suggest that severe muscle weakness, alone or in combination with contractures, is a reliable indicator of a neuromuscular disorder and that CK and neurophysiological tests should be performed every time these clinical features are present even if the infant has clinical and imaging signs of CNS involvement. This simple guideline may help to increase the detection of disorders with associated neuromuscular and brain involvement. This is particularly important as achieving a diagnosis in these cases allows not only better prognostic information but also the possibility of accurate genetic counselling for the family. REFERENCES 1. Richer LP, Shevell MI, Miller SP. Diagnostic Profile of Neonatal Hypotonia: an 11-year study. Pediatr Neurol 2001;25: Dubowitz V. The Floppy Infant. Philadelphia: JB Lippincott Company; Prasad AN, Prasad C. The floppy infant: contribution of genetic and metabolic disorders. Brain Develop 2003;27: Johnston HM. The floppy weak infant revisited. Brain Develop 2003;25: Dubowitz LMS, Dubowitz V, Mercuri E. The Neurological Assessment of the Preterm and Full-term Newborn Infant. 2nd ed. London: McKeith Press; Dubowitz L, Mercuri E, Dubowitz V. An optimality score for the neurologic examination of the term newborn. J Pediatr 1998;133: Beltran-Valero de Bernabe D, Currier S, Steinbrecher A, Celli J, van Beusekom E, van der Zwaag B, et al. Mutations in the O-mannosyltransferase gene POMT1 give rise to the severe neuronal migration disorder Walker- Warburg syndrome. Am J Hum Genet 2002;71: Harding BN, Boyd SG. Intractable seizures from infancy can be associated with dentato-olivary dysplasia. J Neurol Sci 1991;104: Can Clinical Signs Identify Newborns With Neuromuscular Disorders? 79

128 MOTOR FUNCTION AT SCHOOL AGE IN CHILDREN WITH A PRESCHOOL DIAGNOSIS OF DEVELOPMENTAL LANGUAGE IMPAIRMENT RICHARD I. WEBSTER, MBBS, MSC, FRACP, ANNETTE MAJNEMER, PHD, OT, ROBERT W. PLATT, PHD, AND MICHAEL I. SHEVELL, MD, CM, FRCPC Objectives To evaluate fine motor (FM) and gross motor (GM) function shortly after school entry in children with a preschool diagnosis of developmental language impairment (DLI). Study design A cohort of children (n = 70) diagnosed at pre school age with DLI was reevaluated in elementary school. Language, cognitive, and motor outcomes were assessed through the use of the Battelle Developmental Inventory (BDI). Language was further assessed through the use of the Vineland Adaptive Behavior Scale, Peabody Picture Vocabulary, and Expressive One Word Picture Vocabulary Tests. Performance below 21.5 SD of the normative mean on any test was considered to represent impairment. Results Forty-three children (mean age, 7.4 ± 0.7 years) underwent reassessment at a mean of 3.8 ± 0.7 years after initial preschool assessment. Mean scores for BDI motor domains (FM, 78.3 ± 11.4; GM, 84.9 ± 13.3) fell below normative values. Twenty-two children (52%) had motor impairment (FM, 17 of 42; GM, 15 of 42); 35 of 43 (81%) continued to have language impairment. BDI communication raw scores correlated most strongly with FM (q = 0.73, P <.001) and GM (q = 0.58, P =.003) raw scores but showed only moderate correlations with cognitive raw scores (q = 0.41, P =.05). Conclusions Impaired motor function is an important comorbidity in DLI. Factors critical to motor performance may also contribute to language deficits in DLI. (J Pediatr 2005;146:80-5) Developmental Language Impairment (DLI) is a primary impairment in language occurring in a child with otherwise normal nonverbal cognitive function. Children with a language impairment secondary to another process, such as a neurologic or autistic spectrum disorder, global developmental disability, or hearing impairment, are usually not considered to have DLI. 1 Although the term Specific Language Impairment (SLI) has been used to describe the same group of children, 2 there is increasing evidence SLI is not specific and that children with SLI have impairments in other developmental domains. 3 Children with language impairment commonly have difficulties producing normal speech sounds (phonologic impairment). 2 Although a phonologic impairment cannot be considered sufficient in and of itself to make a diagnosis of language impairment, the common association between language and phonologic impairment seen in children with DLI 2,4 raises the possibility that factors that contribute to motor planning and sequencing may also be important for other phases of language processing. Several authors have reported an association between motor and language impairments in school-aged children Robinson 5 found that 90% of children enrolled in a residential school for children with specific speech and language disorders had evidence of motor impairment. Powell and Bishop 7 studied a group of 17 children with SLI and found that a number of measures of motor function were significantly impaired in children with SLI when compared with control children. Rintala et al 8 assessed the motor skills of a group of 76 children with communication disorders. They found that 71% of these children met criteria for definite or borderline movement difficulties through the use of the ABC BDI DLI DQ EOWPVT Assessment Battery for Children Battelle Developmental Inventory Developmental Language Impairment Developmental Quotients Expressive One Word Picture Vocabulary Test FM GM PPVT SLI VABS Fine motor Gross motor Peabody Picture Vocabulary Test Specific Language Impairment Vineland Adaptive Behavior Scale From the Departments of Neurology/ Neurosurgery and Pediatrics, McGill University, School of Physical and Occupational Therapy, Department of Epidemiology and Biostatistics, McGill University, Montreal, Quebec, Canada. Supported by Fonds de recherche en santé du Québec; the Montreal Children s Hospital Research Institute and the John Yu Scholarship (Children s Hospital at Westmead, Sydney, Australia); the Montreal Children s Hospital Foundation; and Canadian Institutes of Health Research. Submitted for publication Feb 3, 2004; last revision received Jun 8, 2004; accepted Sep 9, Reprint requests: Dr Michael Shevell, Room A-514, Montreal Children s Hospital, 2300 Tupper St, Montreal, Quebec H3h 1p3, Canada. michael.shevell@muhc.mcgill.ca /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

129 Movement Assessment Battery for Children (Movement ABC). Hill assessed the motor skills of 19 children with SLI by using the Movement ABC and tests of praxis. 9 The Movement ABC showed evidence of motor impairment in 11 of 19 (58%) children with SLI; moreover, even children with normal scores on the movement ABC showed evidence of impairments in gestural praxis. However, despite the above findings, it is unclear what the likelihood of later motor impairment is in a child with DLI. The current study had two aims: (1) to document the frequency of fine motor and gross motor impairment in school-aged children with a preschool diagnosis of DLI and (2) to investigate the association between language development and motor performance in DLI. METHODS A cohort of children diagnosed at preschool age as having DLI underwent follow-up in early elementary school. Details relating to the assembly and diagnostic evaluation of the original cohort have been published in detail In summary, a consecutive cohort of 224 preschool children referred to the Montreal Children s Hospital over an 18- month period between June 1, 1996, and November 30, 1997, for initial evaluation of suspected developmental delay underwent assessment. Children were seen by a pediatric subspecialist (either developmental pediatrics or pediatric neurology) and were referred for appropriate diagnostic tests and rehabilitation consultations. A child was considered to have DLI if they had a significant delay in language in the presence of otherwise normal development in other developmental domains. Children with a hearing impairment, a preexisting neurologic disorder, or an autistic spectrum disorder were excluded. On the basis of these assessments, 70 of 224 (31%) were categorized as having DLI. The current study presents developmental findings obtained at follow-up (approximately 4 years later) on this cohort of children. The hospital s institutional review board approved the study, and informed consent was obtained from the child s parent or guardian before study participation. Measures The full-scale Battelle Developmental Inventory (BDI) was used to assess motor, cognitive, social, adaptive, and language outcomes for this study. 20 The BDI is a standardized instrument that assesses development in 5 separate domains and has very good psychometric properties. 21 For certain items, the BDI does not require that a child demonstrate a particular skill but rather accepts parental report of the ability to perform the specific task. As such, the BDI can be considered to be a measure of attainment of developmental skills. The BDI was designed as an instrument to test children with developmental disability, so the original sample used for establishing normative data included both typically developing children and children with disabilities. The BDI fine motor domain (BDI FM) assesses manual dexterity and upper extremity coordination. The gross motor domain (BDI GM) assesses posture, balance, and locomotor skills. The communication domain is intended to provide a comprehensive measure of language development and as such involves tests of morphology (the structure of words), syntax (the relation between words and other units within a sentence), vocabulary, phonology (the system of sounds of a language), and discourse (the way in which sentences are combined). The BDI cognitive domain includes visuospatial tests, tests of reasoning and academic skills, conceptual development, and memory. Three further tests of language performance at school age were used: the Peabody Picture Vocabulary Test III A (PPVT), 22 the Expressive One Word Picture Vocabulary Test Revised (EOWPVT-R), 23 and the Vineland Adaptive Behavior Scale (VABS) communication domain. 24 The BDI adaptive domain was used to measure self-help and task-related skills; the BDI personal-social domain was used to measure skills that permit social interaction. The VABS daily living and socialization domains were also administered. The Montreal Children s Hospital sees approximately equal numbers of French- and English-speaking children. Children were assessed in the language in which they were most proficient. For French-speaking children, a French language version of the PPVT-III R (Échelle de vocabulaire en images Peabody A [EVIP-A] 25 ) and a translated version of the EOWPVT-R, both with French-Canadian norms, were used. A translated version of the BDI was also administered to French-speaking children; however, the BDI communication domain was not used because of its heavy reliance on items testing morphosyntax and discourse specific to the English language. Parents of French-speaking children were administered a translated version of the VABS. In the absence of French-Canadian normative data for the BDI and the VABS, published North American data sets 21,24 were used for both French- and English-speaking families. Standard scores were calculated for the VABS and the PPVT-R/EVIP. Developmental Quotients (DQ) were calculated from raw scores for BDI scores. The translated version of the EOWPVT provides only the percentile values. At the time of data analysis, it became clear that a number of children had BDI raw scores for individual domains that fell below the floor level for calculating DQs (1 st centile). For this reason, BDI raw scores were used to assess the correlations within subjects between scores in separate domains. A child was considered to have an impairment in a developmental domain if their score fell more than 1.5 SD below the mean as recommended by the BDI manual 21 (ie, a standard score #77). A child was considered to have a language impairment if performance on any language test fell more than 1.5 SD below the mean. The same threshold was applied to identify impairment in separate domains of the VABS. Statistical Analysis Statistical analysis was performed with the use of the Statistical Package for Social Sciences software. 26 t tests were used to compare the distributions of results in children Motor Function At School Age In Children With A Preschool Diagnosis Of Developmental Language Impairment 81

130 Table I. Values of motor, cognitive, adaptive, social, and language variables Mean ± SD (n) Below 21.5 SD (impaired/total) French-speaking subgroup (n) English-speaking subgroup (n) P value BDI (FM) 78.3 ± 11.4 (42) 17/42 (41%) 79.5 ± 13.2 (18) 77.4 ± 10.1 (24).56 BDI (GM) 84.9 ± 13.3 (42) 15/42 (36%) 83.6 ± 13.8 (18) 86.0 ± 13.1 (24).57 BDI (communication) 69.5 ± 8.9 (24) 20/24 (83%) Not administered 69.5 ± 8.9 (24). BDI (cognitive) 80.0 ± 14.2 (42) 16/42 (38%) 80.9 ± 15.5 (18) 79.2 ± 13.4 (24).70 BDI (personal/social) 75.4 ± 11.8 (42) 24/42 (57%) 73.1 ± 12.1 (18) 77.0 ± 11.6 (24).29 BDI (adaptive) 77.3 ± 13.1 (42) 22/42 (52%) 75.7 ± 12.4 (18) 78.5 ± 13.7 (24).5 VABS (communication) 80.6 ± 16.9 (42) 20/42 (48%) 79.6 ± 15.4 (18) 81.3 ± 18.3 (24).74 VABS (daily living) 83.1 ± 19.1 (42) 19/42 (45%) 78.1 ± 15.5 (18) 87.0 ± 20.8 (24).13 VABS (social) 87.1 ± 13.0 (42) 8/42 (19%) 82.2 ± 9.5 (18) 90.8 ± 14.2 (24).03 PPVT 87.7 ± 16.3 (32) 8/32 (25%) 88.7 ± 19.1 (15) 86.7 ± 13.8 (17).75 EOWPVT (%) * 23.9 ± 25.5 (33) 11/33 (33%) 13.2 ± 22.3 (16) 33.9 ± 24.7 (17).02 *Percentile ranks were available only for normative data for the EOWPVT; hence, results of the EOWPVT are reported using percentile ranks. recruited and those not recruited and between French- and English-speaking subgroups. x 2 statistics were used to compare differences in sex and severity of delay between children recruited and those not recruited. The correlations between BDI raw scores in separate developmental domains were calculated by means of nonparametric techniques (Spearman rho [r]) because these scores were not normally distributed. These correlations were performed for the 24 English-speaking children who were tested with all domains of the BDI. A separate analysis was performed to examine the correlations between motor and cognitive domains for the entire group. A P value of <.05 was used as the threshold for statistical significance for all tests. A Bonferroni correction was not applied when assessing the correlation between separate domains because such a correction potentially masks clinically significant associations. 27 RESULTS Group Characteristics Seventy children met criteria for DLI at the time of the original study, and 43 (61%) were reassessed in the current study. The families of 13 (19%) children declined participation in the study, and 14 (20%) were lost to follow-up. On the basis of variables assessed at intake, no significant difference existed between children who completed follow-up and those who were not recruited in terms of the initial age of assessment (recruited mean age, 3.6 ± 0.7 years; not recruited, 3.6 ± 0.7 years; P =.99), age at initial parental concern (recruited, 1.9 ± 0.7 years; not recruited, 2.2 ± 0.6 years; P =.08), sex distribution (recruited boys, 36 of 43 [84%]; not-recruited boys, 22 of 27 [83%]; P = 1.0), and severity of delay at intake (moderate or severe delay recruited, 27 of 43 [64%]; not recruited, 15 of 27 [56%]; P =.55). The mean age of children at the time of their initial assessment was 3.6 ± 0.7 years, the mean age at follow-up was 7.4 ± 0.7 years, and the mean interval between initial and subsequent school age assessments was 3.8 ± 0.7 years. The majority (36 of 43 [84%]) were boys; 24 of 43 (56%) showed greatest proficiency in English and the remainder in French. One child was assessed with the EOWPVT and PPVT only and did not complete assessment with the BDI or VABS. Three children (7%) had a history of prenatal or perinatal complications (1 fetal distress, 2 delivery complications), and 2 children (5%) had a history of maternal gestational diabetes. One child was born prematurely. A family history of speech/language disorders was reported in 19 of 33 (58%) families in which data were available. Two children had microcephaly at their initial assessment, one an isolated cleft lip, and one features consistent with Opitz G syndrome. At the time of their initial pre school age assessment, 24 of 42 (57%) children had been referred to occupational therapy, and almost all (39 of 42 [93%]) had been referred for speech/language pathology. A previously published survey investigating the use of rehabilitation services by children drawn from this cohort found that 36% of children received occupational therapy services within 6 months of their assessment and 70% received speech/language pathology within the same interval. 28 At school age follow-up, only 5 of 42 (12%) were receiving regular occupational therapy; however, 29 of 42 (69%) were receiving regular speech/ language therapy and 15 of 42 (36%) were receiving additional assistance with their education beyond a regular classroom placement. The parents of 10 of 42 (24%) children considered that their child had difficulties with motor development, 7 of 42 (17%) parents considered that their child had difficulties with activities of daily living, and 30 of 42 (71%) considered that their child had ongoing difficulties with language. Prevalence of Impairment Descriptive statistics on motor, cognitive, and communication measures are presented in Table I. The mean values for French- and English-speaking children differed significantly for the EOWPVT and the VABS socialization domains. The 82 Webster et al The Journal of Pediatrics January 2005

131 Figure 1. Distribution of BDI GM DQs is shown for entire cohort. It should be noted that as a result of floor effects associated with the conversion of percentile scores to DQs, all children whose raw score falls at or below the 1 st centile score 65. Figure 2. Distribution of BDI FM DQs is shown for entire cohort. It should be noted that all children whose raw score falls at or below the 1 st percentile for age receive a DQ of 65. difference in distributions in the latter two domains may reflect cultural differences between French- and English-speaking children. The difference between scores for the EOWPVT may reflect the small size of the locally available normative dataset. Other tests showed a similar distribution in Frenchand English-speaking children, suggesting that these tests had a comparable performance in both groups. To illustrate the spectrum of motor impairment seen among children with DLI, histograms were constructed showing DQs for the BDI GM and FM domains (Figures 1 and 2). Slightly more than half of the children (22 of 42 [52%]) were impaired on at least one measure of motor function. Seventeen (17 of 42 [41%]) children had FM delay, whereas 15 of 42 (36%) children had GM delay. Ten (10 of 42 [24%]) children had evidence of both FM and GM impairment, 7 of 42 (17%) had evidence of FM impairment only, and 5 of 42 (12%) had evidence of GM impairment only. Thirty-five (35 of 43 [81%]) children were impaired on at least one language test. Three (3 of 42 [7%]) children had motor impairment despite the resolution of language impairments, and only 5 of 42 (12%) children had neither motor nor language impairments. Twelve children (12 of 25 [48%]) referred for pre school age occupational therapy assessment had motor impairment at outcome, whereas 10 of 17 (59%) not referred were noted to have motor impairment (P =.48). The association between motor, language, and cognitive impairments was examined by comparing the correlation between these measures in the group of 24 children who were assessed with all BDI domains (Table II). This showed that communication and motor scores were highly correlated and that FM scores were strongly correlated with cognitive scores. However, communication and cognitive scores showed only a modest correlation, and GM scores were poorly correlated with cognitive scores. To investigate whether the pattern of correlations noted was applicable to the entire sample (Frenchand English-speaking children), the analysis was repeated with the BDI cognitive and motor domains only. Similar correlations continued to be noted between cognitive and FM domains (r = 0.61, P <.001), and cognitive and GM domains continued to show a poor correlation (r = 0.15, P =.35). DISCUSSION This study evaluated developmental outcome at school age in a cohort of children with an initial diagnosis of DLI in the early preschool years. At the time of their original medical assessment, children were clinically identified on the basis of language impairment; however, at follow-up, motor impairment proved to be an important objectively documented comorbidity. Our study found that approximately half the children had delays in FM or GM domains. More than half the children had impairments in the BDI adaptive domain, which is weighted toward items assessing motor performance. Yet, despite the high frequency of motor impairment identified on assessment, only 5 of 42 (12%) children were receiving occupational therapy, and only 10 of 42 (24%) parents considered that their child continued to have difficulties with motor skills. In Quebec, there are no legislated thresholds for the delivery of rehabilitation services for children. Most of these children continued to receive speech/language therapy; however, in the absence of data from the children s schools, we cannot determine whether children were not receiving occupational therapy because impairments had not been identified or because services were not available. Motor impairment in children of school age is associated with academic and social comorbidity. 29 Children with language impairment often exhibit difficulties in academic performance and are at risk for social and emotional problems. 30 The presence of motor impairment, with its effects on participation in extracurricular recreational activities, is likely to have an impact on social and emotional function. Difficulties Motor Function At School Age In Children With A Preschool Diagnosis Of Developmental Language Impairment 83

132 Table II. Correlations between BDI communication, cognitive, and gross and fine motor raw scores BDI domain Cognitive Gross motor Fine motor r = 0.73 * Communication r = 0.41 r = 0.58 * P =.05 P =.003 * P,.001 * Cognitive. r = 0.16 P =.46 r = 0.61 * P =.002 * Gross motor. r = 0.57 * P =.004 * Correlations are shown between communication, cognitive, and motor measures for 24 of 43 (56%) children who were tested with all domains of the BDI. *Correlation coefficients greater than with handwriting and other FM tasks in the classroom may further disadvantage these children. 31 Although the causes of academic and social impairment in DLI are multifactorial, reflecting a range of developmental problems, it is likely that motor impairment is an added factor for some children. Given that an associated motor impairment is potentially modifiable, 8 the identification and treatment of concurrent motor deficits may decrease the component of comorbidity that results from impairments in motor skills. Most previous studies investigating motor function in children with DLI have examined children requiring special education as a result of language impairment. 5,7,9,13,15 Such samples are likely to be biased toward children with more severe impairments. Our study reassessed a community-derived cohort of preschool-aged children with DLI. In contrast to the above studies, the majority of children were educated in a normal classroom without any extra educational assistance. The prevalence of motor impairment (51%) that we identified was marginally lower than that reported in previous studies that used a standardized instrument to identify impairment (58% to 90%). 5,8,9 It is likely that this results from recruiting a cohort of children who had less severe language impairment than in previously reported studies. Dissociation between language and cognitive function with impairment of language relative to cognition is used to define SLI. Given the motor complexity of speech, we hypothesized that factors common to motor function and to communication (for example, motor planning, sequencing, and timing) might in part be associated with language impairment in our cohort of children. Consistent with this, we found that BDI communication scores were more strongly correlated with motor scores (both GM and FM) than with cognitive scores. Moreover, the significant correlations between GM and communication performance (despite poor correlations between GM and cognitive performance) suggest that factors critical to GM function may also lead to language impairment but that these factors have less effect on cognitive function. Although previous studies have reported correlations between language measures and FM measures, 10,12 correlations between communication and GM measures have not, to our knowledge, been reported. Schwartz and Regan 12 hypothesized that requirements for speed and repetitive sequencing might underlie the receptive language impairments seen in DLI. However, using discriminant function analysis, Powell and Bishop 7 found that poor performance on a number of tests of GM function were highly significant in discriminating between language-impaired and control children. The association between GM tests (requiring less complex motor sequencing than FM tasks) and communication skills suggests that factors involved in motor planning and feedback control may also play a role in DLI. Our findings suggest that a common pathogenesis may underlie both motor and language impairment in children with DLI. The children recruited for reassessment were diagnosed at pre school age as having isolated language impairment. Despite this, half of these children had motor impairment at follow-up. The pre school age diagnostic process involved assessment by physicians as well as occupational therapists (for most children) and reflected the standard evaluative process for children with suspected developmental delay at our institution. However, given that children did not systematically undergo formal testing of motor development at the time of their pre school age assessment, we do not know whether similar levels of motor impairment would have been identified at intake. It is possible (1) that motor impairments were present and not identified by the initial process of evaluation, (2) that motor deficits were not severe enough for the clinicians to consider that they were clinically significant, or (3) that some motor deficits may not be apparent until tasks requiring more complex motor skill are tested. Further studies investigating the longitudinal history of motor impairment in DLI are needed to address this issue. The children enrolled were drawn from a consecutive cohort of 224 children referred by community-based pediatricians and family physicians for assessment of developmental delay A previous survey of these physicians that investigated the reasons prompting referral for subspecialty medical assessment indicated that perceived severity of the delay was the most important factor. 19 It is likely that the group initially referred was biased toward children with greater degrees of impairment. Moreover, there was some evidence that the group who completed follow-up was biased toward children with more severe impairments. It is thus possible that these results may be generalizable only to children with more severe impairments. However, given the strength of associations and the weak differences between the group under study and those lost to follow-up, this is not likely to be a major problem. A cohort of children with DLI identified on the basis of population screening may show a lower incidence of motor impairment than was identified in the current study. Further population-based screening studies investigating motor comorbidity in DLI are needed to clarify this issue. The authors thank Vicky Stuhec, Caroline Erdos, and Rena Birnbaum, who performed developmental testing; Vicky Stuhec, Nicholas Hall, Lisa Steinbach, and Nancy Marget for assistance with data entry and coordination of this study; the evaluating physicians who participated in the original phase of this study: Drs Ronald Barr, Charles Larson, Bernard Rosenblatt, Kenneth Silver, and Gordon 84 Webster et al The Journal of Pediatrics January 2005

133 Watters; and the parents and children who participated in this longitudinal study. REFERENCES 1. Trauner D, Wulfeck B, Tallal P, Hesselink J. Neurological and MRI profiles of children with developmental language impairment. Dev Med Child Neurol 2000;42: Leonard LB. Children with specific language impairment. Cambridge, Mass: MIT Press; Ors M. Time to drop specific in specific language impairment. Acta Paediatr 2002;91: Bishop DVM, Edmundson A. Language impaired 4-year olds: distinguishing transient from persistent impairment. J Speech Hear Dis 1987;52: Robinson RJ. Causes and associations of severe and persistent specific speech and language disorders in children. Dev Med Child Neurol 1991;33: Johnston RB, Stark RE, Mellits ED, Tallal P. Neurological status of language-impaired and normal children. Ann Neurol 1981;10: Powell RP, Bishop DV. Clumsiness and perceptual problems in children with specific language impairment. Dev Med Child Neurol 1992;34: Rintala P, Pienimäki K, Ahonen T, Cantell M, Kooistra L. The effects of a psychomotor training programme on motor skill development in children with developmental language disorders. Hum Mov Sci 1998;17: Hill EL. A dyspraxic defect in specific language impairment and developmental coordination disorder? Evidence from hand and arm movements. Dev Med Child Neurol 1998;40: Bishop DVM, Edmundson A. Specific language impairment as a maturational lag: evidence from longitudinal data on language and motor development. Dev Med Child Neurol 1987;29: Bishop DVM. Motor immaturity and specific language impairment: evidence for a common genetic basis. Am J Med Genet 2001;114: Schwartz M, Regan V. Sequencing, timing and rate relationships between language and motor skill in children with receptive language delay. Dev Neuropsychol 1996;12: Preis S, Schittler P, Lenard HG. Motor performance and handedness in children with developmental language disorder. Neuropediatrics 1997;28: Hill EL. Non-specific nature of specific language impairment: a review of the literature with regard to concomitant motor impairments. Int J Lang Commun Disord 2001;36: Owen SE, McKinlay IA. Motor difficulties in children with developmental disorders of speech and language. Child Care Health Dev 1997; 23: Shevell MI, Majnemer A, Rosenbaum P, Abrahamowicz M. Etiologic determination of childhood developmental delay. Brain Dev 2001;23: Shevell MI, Majnemer A, Rosenbaum P, Abrahamowicz M. Etiologic yield of single domain developmental delay: a prospective study. J Pediatr 2000;137: Shevell MI, Majnemer A, Rosenbaum P, Abrahamowicz M. Etiologic yield of subspecialists evaluation of young children with global developmental delay. J Pediatr 2000;136: Shevell MI, Majnemer A, Rosenbaum P, Abrahamowicz M. Profile of referrals for early childhood developmental delay to ambulatory subspecialty clinics. J Child Neurol 2001;16: Newborg J, Stock JR, Wnek L, Guidabaldi J, Svinicki J. Battelle Developmental Inventory. Itasca, Ill: Riverside Publishing; Berls AT, McEwen IR. Battelle Developmental Inventory. Phys Ther 1999;79: Dunn LM, Dunn LM. Peabody Picture Vocabulary Test. 3rd edition. Circle Pines, Minn: AGS Publishing; Gardner MF. Expressive One-Word Picture Vocabulary Test, Revised. Novato, Calif: Academic Therapy Publications; Sparrow S, Balla DA, Cicchetti DV. Vineland Adaptive Behavior Scales Interview Edition. Circle Pines, Minn: American Guidance Service; Dunn LM, Thériault CM, Dunn LM. Échelle de vocabulaire en images. Circle Pines, Minn: AGS Publishing; Statistical Package for Social Sciences [computer program]. Version Chicago, Ill: SPSS Inc; Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology 1990:1; Majnemer A, Shevell MI, Rosenbaum P, Abrahamowicz M. Early rehabilitation service utilization in young children with developmental delays. Child Care Health Dev 2001;28: Dewey D, Wilson BN. Developmental coordination disorder: what is it? Phys Occup Ther Pediatr 2001;20: Aram DM, Ekelman BL, Nation JE. Preschoolers with language disorders: 10 years later. J Speech Lang Hear Res 1984;27: Losse A, Henderson SE, Elliman D, Hall D, Knight E, Jongmans M. Clumsiness in children: do they grow out of it? A 10-year follow-up study. Dev Med Child Neurol 1991;33: CORRECTION The abstract of Urinary tract infection: Is there a need for routine renal ultrasonography? (Zamir G, Sakran W, Horowitz Y, Koren A, Miron D. Arch Dis Child 2004;89:466-8) by T. J. Beattie, which appeared in the November 2004 issue of The Journal (volume 145, pages 682-3), should have included Dr. Beattie s degrees: MB, CHB, FRCP(Glas), and FRCPCH. Motor Function At School Age In Children With A Preschool Diagnosis Of Developmental Language Impairment 85

134 LUNG FUNCTION AT TERM REFLECTS SEVERITY OF BRONCHOPULMONARY DYSPLASIA OLA HJALMARSON, MD, PHD, AND KENNETH L. SANDBERG, MD, PHD Objective To test whether infants with bronchopulmonary dysplasia (BPD) express the same functional impairments at term as healthy, preterm infants, and whether clinical severity of BPD is qualitatively or quantitatively related. Study design Prospective measurements on a consecutive sample of 50 infants with BPD and 19 healthy preterm controls in a university hospital. BPD infants were classified as severe, moderate, or mild, according to their need for oxygen. A multiple-breath nitrogen wash-out method was used to assess functional residual capacity (FRC) and gas mixing efficiency. Mechanical variables were estimated by the occlusion test. Results Infants with severe BPD had lower FRC, less efficient gas mixing, and higher specific conductance than those with mild and moderate BPD, and the preterm controls. Mild and moderate BPD did not differ in any property from each other but differed from controls in the same variables. The elastic properties of the respiratory system appeared unaffected by BPD. Conclusions The ventilatory impairments in BPD were of the same nature as in healthy preterm infants when compared with term infants, but their magnitude was related to the clinical severity of the BPD. Gas mixing efficiency together with FRC appears to be useful to assess lung development in BPD. (J Pediatr 2005;146:86-90) Bronchopulmonary dysplasia (BPD) is a common chronic lung disease affecting very preterm infants. The infants have hypoxemia and often need mechanical ventilation for prolonged periods of time. The dominant lung lesions in animal models of BPD 1,2 and in humans after modern treatment regimens 3-5 are delayed development of the acinus of the lung with abnormal alveolarization, abnormal elastin deposition, 1,3,6 and disturbed vascularization. 2,7,8 The nature and the relative roles of factors acting together with immaturity to induce BPD are unknown, although some, such as hyperoxia, 9,10 volutrauma, 1,2 and inflammation, 11,12 cause lesions similar to BPD in animal models. Infants born preterm but without lung disease after birth have abnormal lung function at term compared with healthy term infants, suggesting disturbed acinar development of the lungs or peripheral airway disease. 13 The deviations in lung function in healthy preterm infants may be caused by premature exposure to air, to the act of breathing, or to factors interfering with the normal genetic schedule of lung development in late gestation. We hypothesized that BPD, in functional terms, may be regarded as a more severe form of this process. The aim of this study was to investigate whether BPD affects the same ventilatory variables of the lungs as those impaired in healthy preterm infants, and how the severity of the disease relates to these variables. We studied functional residual capacity (FRC), gas mixing efficiency, and mechanics of the respiratory system. We used the diagnostic criteria of BPD recently proposed by the NICHD/NHLBI/ORD Workshop 14 (Table I). Study Design METHODS This prospective study was designed to compare three groups of preterm infants with mild, moderate, and severe BPD, respectively, together with a sample of healthy controls, with respect to four different variables of lung development and function at a postmenstrual BPD FRC Bronchopulmonary dysplasia Functional residual capacity PMA Postmenstrual age From the Department of Pediatrics, The Sahlgrenska Academy at Göteborg University, Göteborg, Sweden. Submitted for publication Jun 4, 2004; last revision received Jul 13, 2004; accepted Aug 17, Reprint requests: Ola Hjalmarson, MD, Department of Pediatrics, The Queen Silvia Children s Hospital, SE Göteborg, Sweden. ola.hjalmarson@pediat.gu.se /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

135 Table I. Definitions of BPD for infants <32 weeks of gestational age at birth, according to the NICHD/ NHLB/ORD Workshop 14 General criterion: Additional criteria: Mild BPD Moderate BPD Severe BPD Treatment with oxygen.21% for at least 28 days Breathing room air at 36 weeks PMA or at discharge Need for,30% oxygen at 36 weeks PMA or at discharge Need for 30% oxygen or more and/or positive pressure at 36 weeks PMA or at discharge Assessment is made at 36 weeks PMA or discharge to home, whichever comes first. For detailed description, see original paper. Table II. Maturation, weight and gender of studied infants Variable Mild BPD n=19 Moderate BPD n=19 Severe BPD n=12 Controls n=19 Males/females 5/14 10/9 8/4 9/10 Birth weight (g) Gestational age at birth (wk) Weight at study (g) Postmenstrual age at study (wk) Weights and ages are given as medians and ranges. age (PMA) corresponding to term. FRC, compliance, resistance of the respiratory system, and gas mixing efficiency were assessed. The study was approved by the Board of Ethics of the Medical Faculty at Göteborg University. Informed consent was given by the parents. The hypothesis to be tested was whether the groups differed from each other and from the controls in each of the variables assessed. The size of the sample was calculated with the aim of detecting a difference between samples of at least one standard deviation (SD) of each variable at a significance level of 5% with a power of 80%. Subjects To ascertain unbiased sampling, all consecutive, surviving infants with BPD were considered for inclusion in the study. In addition, 19 healthy preterm infants without signs of respiratory disease or need for extra oxygen after 2 days of age were recruited as controls. Inclusion criteria were a gestational age <31 weeks, based on menstrual and obstetrical ultrasonography data; informed parental consent; and practical possibilities for the conduction of the lung function test. Exclusion criteria were cardiac or other significant malformations and being outside +/2 2 standard deviations for weight and length at birth. For BPD infants, the BPD severity had to be classifiable according to the criteria in Table I. 14 For the purpose of BPD classification, oxygen dependency was defined as inability to keep oxygen saturation at 93% or above for at least 12 hours a day when breathing room air. At least one of the parents was present during infant lung function assessments. Nitrogen Wash-out Method FRC and indices of gas mixing efficiency were measured by the multiple breath nitrogen wash-out method, adapted for newborn infants as described previously. 15,16 In brief, the infants were studied supine after a meal, during sleep, and without sedation. A Rendell-Baker mask size 0 (Soucek Mask, London, UK) was filled with a water-soluble putty and was gently fitted over the nose and mouth of the infant. A metal piece with a cylindrical canal, 20 mm long and 5 mm in diameter, connects the mask to the sidewall of a tube with a constant flow of breathing gas (around 150 ml/second). The equipment dead space was around 1 ml. At the outlet, the bypass flow was led through a pneumotachograph (Fleisch no. 0) (Hugo Sachs Elektronik, March-Hugstetten, Germany) with a linear response. A valve was attached at the inlet so that breathing gases could be switched instantaneously. When the mask was fitted, breathing flow was recorded as a modulation of the constant bypass flow. The probe of a infrared nitrogen meter (Hewlett-Packard 47320A, Loveland, Colo) was mounted into the metal-piece canal so that the nitrogen concentration of the gas passing into and out of the infant could be recorded continuously. Recording started when the infant was breathing regularly and was sleeping. In order to obtain a wash-out recording, the breathing gas was switched to 100% oxygen during an expiration and continued until end-expiratory nitrogen concentration was below 1/40th of the concentration at the start of the wash-out. The test was performed two to three times for each infant with at least 5 minutes between measurements. Recordings were visually monitored on a screen and were discarded for leaks and major breathing irregularities. All signals were digitized with 12 bits accuracy at a frequency of 250 Hz and were stored on disc. For the analysis, the time scale of the nitrogen signal was corrected for the measured transit time of the meter (50 ms). FRC was calculated as the total expired nitrogen volume during washout divided by end-tidal nitrogen concentration preceding the wash-out. Efficiency of gas mixing was assessed by calculating a moment ratio of the course of nitrogen elimination. 16 The zeroth and the first and second moments of end-tidal nitrogen concentration were calculated with accumulated expired volume as the independent variable, normalized by FRC, and the ratio of the first and second moments to the zeroth Lung Function At Term Reflects Severity Of Bronchopulmonary Dysplasia 87

136 Table III. Treatment received by study infants Variable Mild BPD n=19 Moderate BPD n = 19 Severe BPD n=12 Controls n=19 Prenatal steroids, number treated (%) 11 (58) 11 (58) 6 (50) 8 (42) Surfactant treatment, number treated (%) 7 (37) 19 (100) 12 (100) 0 (0) Mechanical ventilation, number treated (%) 10 (53) 16 (84) 12 (100) 0 (0) Mechanical ventilation, days; median (range) 1 (0 9) 5 (1 32) 21.5 (3.5 60) 0 0 Oxygen treatment, days; median (range) 49 (28 70) 79 (40 115) 180 (90 450) 0 (0 2) Betametasone prior to extubation, n (%) 0 (0) 6 (32) 10 (83) 0 (0) Patent ductus arteriosus, n (%) 9 (47) 9 (47) 7 (58) 0 (0) There was a significant difference in surfactant treatment and mechanical ventilation (P <.005) between mild BPD and moderate or severe BPD. Duration of mechanical ventilation differed significantly between all groups (P <.01). Differences in oxygen treatment resulted from the classification criteria. n, number. moment, M 1 /M 0 and M 1 /M 0, respectively, were used. Tidal volume and respiratory rate were calculated from the recording of flow. The reproducibility of FRC and moment ratios were reported previously. 16 Breathing Mechanics Compliance and resistance of the respiratory system were assessed by the single occlusion technique according to the European Respiratory Society/American Thoracic Society criteria. 17 A Rendell-Baker mask, attached to a Fleisch-type pneumotachograph size number 00, was used, and occlusions were performed manually, close to maximal inspiration. Pressure was measured inside the mask. Expirations with a pressure plateau of at least 0.2 seconds and a linear flowvolume plot after occlusion were selected for automatic calculation of the time constant of the respiratory system. Compliance of the respiratory system was then calculated from the volume and pressure differences at occlusions and resistance and conductance from the time constant of the respiratory system and the compliance data. The result was calculated as the mean of 4 to 10 occlusions. Statistical Analysis After graphic tests for normality, groups were compared by analysis of variance, followed by Tukey s post hoc test or Mann-Whitney U test. Linear multiple regression analysis with numerical values assigned to the disease groups was used to test for confounders. The x 2 test was used when appropriate (Statistica software, StatSoft, Inc, Tulsa, Okla). Adjusted P values <.05 were considered significant (two-sided tests). RESULTS Fifty infants with BPD were included in the study. (Table II) The BPD groups had similar gestational age at birth, birth weight, and PMA at time of study. The control group was more mature at birth, and they had about 1 week greater PMA and body weight at the time of study. There were more females in the mild BPD group and more males in the severe BPD group. However, neither gestational age nor gender contributed significantly to the variance of any of the lung function variables in multiple regression analysis. The BPD groups differed in respirator days, surfactant treatment, and duration of oxygen treatment but not in treatment with prenatal steroids or in the patent ductus arteriosus (Table III). A consistent pattern in the lung function measurements was that M = mild and moderate BPD appeared similar but differed from severe BPD as well as from controls in several aspects (Table IV). FRC, even after correction for body size, was reduced in severe BPD as compared with controls. The differences in compliance of the respiratory system between the groups were not significant after correction for lung size. Resistance of the respiratory system did not differ between the groups, but, after correction for lung size by calculating specific conductance, this variable was significantly higher in mild and moderate BPD than in controls, and it was still higher in severe BPD. Gas mixing efficiency, expressed as moment ratios, was significantly reduced in severe and mild BPD, compared with controls, and in severe compared with moderate BPD. Respiratory rate was increased in severe BPD. When the analysis was repeated with the mild and moderate BPD groups combined, the combined and severe BPD groups and differed from controls in FRC, moment ratios, specific conductance, compliance, and respiratory rate by post hoc test. DISCUSSION We measured different aspects of ventilatory function in infants with BPD, classified according to recently proposed criteria, in order to determine if the same variables are affected as are altered by uncomplicated preterm delivery. We also wanted to study whether clinical criteria, based on severity of disease, correlate with different types and degrees of lung involvement when examined close to term. Although the BPD groups had similar gestational ages at birth, the controls were more mature. It is rare for infants born at 23 to 25 weeks gestational age not to develop lung disease. Consequently, we had to use healthy controls that were more mature at birth. However, neither within the BPD group nor within the control group were there significant correlations between gestational age at birth and any of the variables assessed, and no relationship was found using multiple regression analysis. 88 Hjalmarson and Sandberg The Journal of Pediatrics January 2005

137 Table IV. Lung function data Variable Unit Severe BPD Moderate BPD Mild BPD Controls Functional residual capacity ml 33 (6.5) * 50 (12) y 52 (8.4) y 63 (13) Functional residual ml/kg 15 (3.7) * 19 (3.3) 18 (3.3) 21 (3.4) capacity per kg body weight Compliance of the respiratory system ml/cmh 2 O 1.8 (.33) * 2.6 (.85) y 2.8 (.61) y 3.2 (.46) Specific compliance cmh 2 O (.012).051 (.011).055 (.011).052 (.013) Resistance of the respiratory system cmh 2 O/L/s 87 (26) 93 (35) 80 (20) 96 (28) Specific conductance (cmh 2 O,s) (.15) *.25 (.076).26 (.087) y.18 (.06) Moment ratio M 1 /M (.16) y 2.57 (.28) 2.64 (.23) y 2.40 (.28) Moment ratio M 2 /M (1.1) yz 10.5 (1.9) 10.9 (1.5) y 9.3 (1.9) Respiratory rate min (14) * 56 (14) y 53 (11) 44 (10) Tidal volume ml/kg 6.5 (1.4) y 7.4 (1.4) 7.2 (1.2) 7.8 (1.3) Results are presented as means and SD. Differences between severe BPD and other groups are only signed in column severe BPD. There were no significant differences between mild and moderate BPD. Significant means P <.05 at post hoc test. Symbols denote significant difference from: *all other groups. ycontrols. zmoderate BPD. The same was true for gender. We therefore consider the control group to be adequate for comparisons of BPD and non-bpd states, although the effect of extreme immaturity per se on lung function is still unknown. Infants with severe BPD had lower body weights at study than the other groups, which had to be accounted for in the analysis. They also breathed at a higher rate, which was not expected to influence FRC or the mechanics measurements because the time constant of the respiratory system was shorter and a semi-static method to assess mechanics was used. Severe BPD infants were treated for longer periods with mechanical ventilation and oxygen than the other groups and also were given steroids before extubation more often. Half of the infants with mild BPD were never ventilated. The mixing of inspired air with the FRC is a physiologically important and complex process. 18 The bulk of the mixing takes place in the peripheral part of the airways. It is dependent on the structure of the acinus but also on any inhomogeneity of ventilation between the lungs or parts of the lungs. As delayed septation of acini, large primitive alveoli, and inhomogeneous lung structure are known to be present in fatal BPD, 3,4 we hypothesized that gas mixing efficiency would be affected by the severity of BPD. A multiple breath nitrogen wash-out technique, designed to have a very low dead space, was chosen to assess FRC and gas mixing efficiency. We have previously shown that moment ratios, calculated from nitrogen decay curves against ventilation, normalized by lung volume, is the best tool to express gas mixing efficiency in small infants. 16 The result showed that infants with severe BPD were more severely affected in terms of gas mixing efficiency than those with moderate or mild BPD. These two latter groups, however, did not differ significantly from each other. However, even the mildest form of BPD differed in this respect from healthy controls. The finding of reduced FRC in severe BPD is consistent with other studies FRC was most compromised in severe BPD. The results show that severe BPD affects not only the maturation of the fine structures of the acinus but also the functional lung volume. Compliance and resistance are both functions of lung size. To be able to compare these measurements between individuals or groups with different lung sizes, specific compliance and conductance are calculated by dividing compliance and the reciprocal of resistance with FRC. This is a quite straightforward procedure in healthy persons, but it requires some caution when applied in cases of lung disease. When FRC is assessed by a gas wash-out method, only the resting volume in contact with open airways is detected, and trapped gas is not measured. However, no bias is introduced when specific compliance is calculated, as compliance should be normalized by the lung volume that is able to respond by expansion and air entry. Compliance of the respiratory system was significantly lower in the BPD groups in comparison with controls, and compliance in severe BPD also was lower than in the two milder groups (Table IV). However, after compensating for differences in lung size, the differences did not persist. Consequently, we did not find any indication of changes in elastic performance of the respiratory system in BPD. This result may seem surprising in relation to studies from the 1970s and 1980s. 19,22,23 However, preterm controls were used in this study, which is important as healthy preterms have lower compliance of the respiratory system than infants born at term, 13 and BPD has changed over the last decades. 14 The interpretation of specific conductance is more problematic. Airway resistance, which constitutes one component of the resistance of the respiratory system, assessed by the occlusion method, is strongly dependent on the area of large airways, which in turn is affected by lung size. Trapped gas may contribute to lung size, which means that specific conductance can be overestimated if FRC is not measured by a method including trapped gas. Overestimations may also Lung Function At Term Reflects Severity Of Bronchopulmonary Dysplasia 89

138 occur in infants with a disturbed balance between airway and acinar development. Histological studies of animals 2 and humans 3 have shown that in contemporary BPD, in contrast to original observations, 24 airways are less involved in the early stages of disease. Our observations are in accordance with this. BPD infants did not have higher resistance than controls, and no definite trend was seen when the BPD groups were compared with one another. In contrast, specific conductance was considerably higher in BPD, and severe BPD showed the highest values. In other words, in relation to FRC, large airways offered lower resistance to flow in BPD, especially in severe BPD, than for healthy controls. Trapped gas may contribute to the result. However, the result may also indicate that acinar growth is hampered in BPD, whereas the airways grow more normally. The high specific conductance in BPD is mainly a sign of the imbalance of acinar and airway development. 3,5 The pattern that emerges is that preterm infants who still need 30% or more oxygen at 36 weeks PMA have a pronounced functional abnormality at term, which we interpret as impaired development of the lungs. Qualitatively similar but less severe abnormalities are seen in infants with less sever forms of BPD. The milder forms of BPD differed significantly from the controls at term in gas mixing efficiency, indicating that mild BPD is neither a state of slow adaptation to extrauterine life only nor a fast recovery from an initial injury but has similar characteristics of disturbed lung development as severe BPD, persisting at least up to term. This pattern implies an accentuation of the abnormalities previously found in healthy preterm infants in comparison with term infants. It can be interpreted as a sign of simplified peripheral lung structure and inhomogeneity of lung parenchyma known to be present in BPD. 3-5 Also the imbalance between development of airways and acini with no signs of impairment of resistance to breathing indicates that the at lung periphery is the target of disease at this age. The similarities of functional lung impairment of BPD and of uncomplicated preterm infants at term highlights the fact that major lung development occurs in the periphery of the lung over the last months of pregnancy, and that this process is vulnerable to disturbance. Although different mechanisms may be involved, some pathways of pathological lung development may be similar in BPD and uncomplicated prematurity. More information about lung growth is needed to guide clinical interventions to support a normal development in preterm infants with and without respiratory insufficiency. Skilful technical assistance by Kristina Hellgren and Gunilla Sten is acknowledged. REFERENCES 1. Albertine KH, Jones GP, Starcher BC, Bohnsack JF, Davis PL, Cho S, et al. Chronic lung injury in preterm lambs: disordered respiratory tract development. Am J Respir Crit Care Med 1999;159: Coalson JJ, Winther VT, Siler-Khodr T, Yoder BA. Neonatal chronic lung disease in extremely immature baboons. Am J Resp Crit Care Med 1999; 160: Margraf LM, Tomashefski JF Jr, Bruce MC, Dahms BB. Morphometric analysis of the lung in bronchopulmonary dysplasia. Am Rev Respir Dis 1991; 143: Husain AN, Siddiqui NH, Stocker JT. Pathology of arrested acinar development in postsurfactant bronchopulmonary dysplasia. Hum Pathol 1998;29: Coalson JJ. Pathology of new bronchopulmonary dysplasia. Semin Neonatol 2003;8: Thibeault D, Mabry SM, Ekekezie II, Troug WE. Lung elastic tissue maturation and pertubations during the evolution of chronic lung disease. Pediatrics 2000;106: Bhatt AJ, Pryhuber GS, Huyck H, Watkins RH, Metlay LA, Maniscalco WM. Disrupted pulmonary vasculature and decreased vascular endothelial growth factor, Flt-1, and TIE-2 in human infants dying with bronchopulmonary dysplasia. Am J Respir Crit Care Med 2001;164: Parker TA, Abman SH. The pulmonary circulation in bronchopulmonary dysplasia. Semin Neonatol 2003;8: Coalson JJ, Winter VT, delemos RA. Decreased alveolarization in baboon survivors with bronchopulmonary dysplasia. Am J Respir Crit Care Med 1995;152: Warner BB, Stuart LA, Papes RA, Wispe JR. Functional and pathological effects of prolonged hyperoxia in neonatal mice. Am J Physiol 1998;275:L Yoon BH, Romero R, Jun JK, Park KH, Park JD, Ghezzi F, et al. Amniotic fluid cytokines (interleukin-6, tumor necrosis factor-alpha, interleukin-1 beta, and interleukin-8) and the risk for the development of bronchopulmonary dysplasia. Am J Obstet Gynecol 1997;177: Speer P. Inflammation and bronchopulmonary dysplasia. Semin Neonatol 2003;8: Hjalmarson O, Sandberg K. Abnormal lung function in healthy preterm infants. Am J Respir Crit Care Med 2002;165: Jobe AH, Bancalari E. Bronchopulmonary dysplasia. NICHD/NHLBI/ ORD Workshop Summary. Am J Respir Crit Care Med 2001;163: Sandberg K, Sjöqvist BA, Hjalmarson O, Olsson T. Analysis of alveolar ventilation in the newborn. Arch Dis Child 1984;59: Shao H, Sandberg K, Sjöqvist BA, Hjalmarson O. Moment analysis of multi-breath nitrogen wash-out in healthy preterm infants. Pediatr Pulmonol 1998;25: Gappa M, Colin AA, Goetz I, Stocks J. Passive respiratory mechanics: the occlusion techniques. Eur Respir J 2001;17: Engel LA. Gas mixing within the acinus of the lung. J Appl Physiol Respir Environ Exercise Physiol 1983;54: Gerhardt T, Hehre D, Feller R, Reifenberg L, Bancalari E. Serial determination of pulmonary function in infants with chronic lung disease. J Pediatr 1987;110: Shao H, Sandberg K, Hjalmarson O. Impaired gas mixing and low lung volume in preterm infants with mild chronic lung disease. Pediatr Res 1998; 43: Kavvadia V, Greenough A, Dimitriou G. Early prediction of chronic oxygen dependency by lung function test results. Pediatr Pulmonol 2000;29: Ahlström H. Pulmonary mechanics in infants surviving severe neonatal respiratory insufficiency. Acta Paediatr Scand 1975;64: Lindroth M, Jonson B, Svenningsen NW, Mortensson W. Pulmonary mechanics, chest x-ray and lung disease after mechanical ventilation in low birth weight infants. Acta Paediatr Scand 1980;69: Northway WH, Rosan RC, Porter DY. Pulmonary disease following respirator therapy of hyaline membrane disease: bronchopulmonary dysplasia. N Engl J Med 1967;276: Hjalmarson and Sandberg The Journal of Pediatrics January 2005

139 SAFETY OF BUDESONIDE INHALATION SUSPENSION IN INFANTS AGED SIX TO TWELVE MONTHS WITH MILD TO MODERATE PERSISTENT ASTHMA OR RECURRENT WHEEZE WILLIAM E. BERGER, MD, PAUL Y. QAQUNDAH, MD, KATHRYN BLAKE, PHARMD, JOSE RODRIGUEZ-SANTANA, MD, ANNE-MARIE IRANI, MD, JOHN XU, PHD, AND MITCHELL GOLDMAN, MD, PHD Objective To compare the safety of budesonide inhalation suspension (BIS) with placebo in infants 6 to 12 months of age with mild to moderate persistent asthma or recurrent wheeze. Study design In this multicenter, randomized, double-blinded, parallel-group, placebo-controlled study, 141 patients received 0.5 mg BIS (n = 48), 1.0 mg BIS (n = 44), or placebo (n = 49) once daily for 12 weeks. The primary variable was adrenal function, based on cosyntropin-stimulated plasma cortisol levels. Spontaneous adverse events and clinical laboratory findings also were monitored. Results Overall, the types and frequencies of adverse events reported during the study were comparable across treatment groups. The response to cosyntropin stimulation was similar across treatment groups, with no significant difference between BIS treatment and placebo. Conclusions The safety profile of BIS was similar to that of placebo, with no suppressive effect on adrenal function in patients 6 to 12 months of age with mild to moderate persistent asthma or recurrent wheeze. (J Pediatr 2005;146:91-5) Inhaled corticosteroids (ICS) are preferred for the long-term treatment of all severities of persistent asthma in children of all ages. 1 Currently, budesonide inhalation suspension (BIS) (AstraZeneca LP, Wilmington, Del) is the only nebulized ICS approved by the US Food and Drug Administration (FDA) and the only ICS approved for use in asthmatic children as young as 12 months. No other ICS is approved for children younger than 4 years. Cromolyn sodium, the only other FDA-approved nebulized anti-inflammatory asthma medication, is approved for use in children 2 years of age and older. Cromolyn sodium is less efficacious than BIS 2 and is not preferred for the long-term treatment of persistent asthma in children. 1 In the United States, safety and efficacy of nebulized BIS have been demonstrated in three 12-week, randomized, double-blinded, placebo-controlled studies involving 1018 infants and young children 6 months to 8 years of age with mild to severe persistent asthma. 3-5 Long-term (52-week) open-label extensions of these studies have shown that BIS treatment (median daily doses of 0.5 to 1.0 mg) is similarly well tolerated in children 8 months to 9 years of age. 6,7 Controlled clinical studies addressing the safety and efficacy of ICS exclusively in infants are lacking; consequently, the FDA has requested that further safety and efficacy data be collected for the use of ICS in this patient population. The current study was conducted to assess the safety of once-daily administration of nebulized BIS in infants with AE ANCOVA BIS Adverse event Analysis of covariance Budesonide inhalation suspension FDA ICS Food and Drug Administration Inhaled corticosteroid(s) From Allergy and Asthma Associates of Southern California, Mission Viejo, California; Pediatric Care Medical Group, Inc, Huntington Beach, California; Nemours Children s Clinic, Jacksonville, Florida; Pediatric Pulmonary Program of San Juan, San Juan, Puerto Rico; Virginia Commonwealth University, Richmond, Virginia; and AstraZeneca LP, Wilmington, Delaware. Supported by AstraZeneca LP, Wilmington, Delaware. Dr Berger receives research grants from AstraZeneca, GlaxoSmithKline, Aventis, and Schering-Plough; he is on the speaker s bureau for AstraZeneca, GlaxoSmithKline, Aventis, and Schering- Plough. Dr Blake receives research grants from AstraZeneca, GlaxoSmith Kline, Merck, and Schering-Plough; she is on the speaker s bureau for Astra- Zeneca, GlaxoSmithKline, and Merck. Dr Irani receives research grants from AstraZeneca; she is a paid consultant and on the speaker s bureau for AstraZeneca. Dr Rodriguez-Santana has no inter-est or financial arrangement, includ-ing grants, advisory board memberships, or share holdings, with AstraZeneca. Dr Qaqundah has no potential, perceived, or real conflict of interest with the company that produces budesonide inhalation suspension (Astra- Zeneca). Each author contributed equally to the preparation and review of this manuscript. Submitted for publication Feb 19, 2004; last revision received Jun 30, 2004; accepted Aug 27, Reprint requests: William E. Berger, MD, Allergy and Asthma Associates of Southern California, Medical Center Road, Suite 244, Mission Viejo, CA Weberger@ uci.edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

140 diagnosed mild to moderate persistent asthma or symptoms of recurrent wheeze. Adrenal function was the primary variable. Although the study was not powered to assess efficacy, asthma symptoms, global health status, treatment failures, and patient withdrawals were monitored. METHODS Patients Enrolled patients were $6 but <12 months of age, with mild to moderate persistent asthma. Patients with $2 episodes of persistent or recurrent wheeze who might have benefited from inhaled anti-inflammatory treatment also were enrolled at the investigators discretion. Patients with severe asthma, a history of assisted ventilation (except at birth), or other severe chronic lung diseases were excluded. Exclusion criteria also included treatment with systemic corticosteroids within 4 weeks of study entry and treatment with oral corticosteroids or ICS, slow-release oral b 2 -adrenergic agonists, long-acting b 2 -adrenergic agonists, or 5-lipoxygenase and leukotriene antagonists within 2 weeks of random assignment. The study was performed according to the Declaration of Helsinki, and the institutional review board at each investigation site approved the study protocol. Each patient s legal guardian provided informed consent. Study Design This randomized, double-blinded, placebo-controlled, parallel-group study was conducted at 55 US centers (AstraZeneca LP study SD ). During a 2- to 3-week run-in period, patients discontinued maintenance asthma therapies (except for anticholinergics and metaproterenol). Patients with asthma symptoms on $3 of the 7 days before random assignment and who provided basal morning and 1-hour postcosyntropin stimulation plasma specimens were randomly assigned 1:1:1 to receive 0.5 or 1.0 mg Pulmicort Respules (BIS) once daily or placebo. Random assignment was stratified according to patient age (6 months to <7 months, 7 months to <8 months, 8 months to <9 months, and 9 months to <12 months). Treatments were randomly assigned in blocks of 3 within each stratum. Study medications were administered by means of a Pari LC-Plus nebulizer connected to a Pari Master compressor (Pari Respiratory Equipment, Inc, Richmond, Va) with a face mask or mouthpiece. The first dose of study medication was administered at the clinic, and guardians were instructed on the proper method for subsequent medication administration. Patients returned to the clinic 4 times after random assignment (weeks 2, 4, 8, and 12) for evaluation and data collection. Adrenal function, the primary variable, was determined on the basis of mean changes from baseline to week 12 (or time of early discontinuation) in postcosyntropin-stimulated plasma cortisol levels. Basal plasma cortisol samples were obtained no later than 8:30 AM, when possible, and poststimulation samples were obtained 1 hour after a bolus injection of mg cosyntropin (Cortrosyn; Organon Inc, West Orange, NJ). Subnormal adrenal function was defined as a post cosyntropin plasma cortisol value <500 nmol/l. 8 Secondary safety variables included the incidence and severity of adverse events (AEs) and changes from baseline in vital signs, physical examination results, oropharyngeal and nasal fungal cultures, and clinical laboratory test results. Comprehensive physical examinations and laboratory and fungal culture assessments were performed at visit 1 and study end (week 12 or time of early discontinuation). Height was measured with the use of a tape measure while patients were in a supine position. Vital signs were monitored at each visit. Efficacy variables included asthma symptoms and breakthrough medication use (reported in patient diaries), percentage of symptom-free days, number of treatment failures and patient discontinuations, and global assessment of asthma status. Physicians compared global asthma status at study end or time of discontinuation with pretreatment status, based on a scale from 1 (a great deal better) to 5 (a great deal worse). Concomitant Treatment Anticholinergics and metaproterenol were allowed during the run-in period if used on a regular basis. Shortacting b 2 -adrenergic agonists were allowed as rescue medication throughout the study. Oral prednisone or prednisolone was permitted after random assignment to treat asthma exacerbations, but use could not exceed 10 days. Other medications considered necessary for the patient s welfare were given at the discretion of the investigator. Statistical Analyses Sample size calculations for this study were not based on formal statistical calculations; the size and design of this study were based on guidance from the FDA. The difference between groups in the change from baseline to week 12 in cosyntropin-stimulated plasma cortisol levels was analyzed by means of analysis of covariance (ANCOVA), with treatment as the main effect and the baseline cosyntropin-stimulated value as a covariate. The Wilcoxon rank sum test also was used to analyze these data. A similar ANCOVA was used to analyze changes from baseline to the average over the 12-week treatment period in asthma symptom scores. The incidence of AEs and changes from baseline in physical and laboratory assessments were summarized descriptively. ANCOVA, with treatment as the main effect and baseline as the covariate, was used to analyze changes from baseline to the average over the 12-week, double-blinded treatment period in nighttime and daytime asthma symptom scores. The percentage of symptomfree days over the 12-week, double-blinded treatment period was analyzed by means of ANCOVA, with treatment as the main effect. Physicians global assessments at the last visit on treatment and the proportion of patients with treatment failures were analyzed by means of the Mantel-Haenszel test. RESULTS Patient Demographic and Baseline Characteristics Of the 216 patients enrolled in the study, 141 were randomly assigned to treatment (Table I). Patient mean age 92 Berger et al The Journal of Pediatrics January 2005

141 Table I. Patient demographics and baseline characteristics Treatment group 0.5 mg BIS (n = 48) 1.0 mg BIS (n = 44) Placebo (n = 49) Total (N = 141) Sex, n (%) Female 16 (33.3) 18 (40.9) 20 (40.8) 54 (38.3) Male 32 (66.7) 26 (59.1) 29 (59.2) 87 (61.7) Age, mo Mean ± SD 8.3 ± ± ± ± 1.7 Range Mean plasma cortisol ± SD, nmol/l Basal ± ± ± ± Cosyntropin-stimulated ± ± ± ± Age stratum, n (%) $6to,9 mo 26 (54.2) 25 (56.8) 25 (51.0) 76 (53.9) $9to,12 mo 22 (45.8) 19 (43.2) 24 (49.0) 65 (46.1) was 8.4 months. With the exception of a higher percentage of boys than girls 9 to 12 months of age in the 0.5-mg BIS group (82% vs 18%, respectively), patient demographics were similar across treatment groups and age strata. Plasma cortisol levels were comparable across treatment groups (Table I) and age strata (data not shown). Of the 141 patients randomly assigned to treatment, 117 (83%) completed the study. The most common reasons for treatment discontinuation were loss of patient to follow-up and other reasons (such as noncompliance, physician s choice, patient relocation, withdrawn consent). No patient discontinued the study because of treatment failure. Concomitant Treatment Concomitant asthma medications were used during the study by 19% and 25% of patients receiving 0.5 and 1.0 mg BIS, respectively, and 31% of patients receiving placebo. Oral corticosteroids were the most common concomitant asthma medications and were used by 7 (15%), 9 (21%), and 11 (22%) patients in the 0.5-mg BIS, 1.0-mg BIS, and placebo groups, respectively. Safety ADRENAL FUNCTION. Some patients in the intention-totreat population did not have sufficient treatment duration or adrenal function data to meet the specifications set forth in the FDA s written request. Thus, adrenal function was assessed in a subset of 76 patients who provided both basal and stimulated plasma cortisol samples at baseline and end of treatment and who remained in the study for $78 days after random assignment (evaluable population). Cosyntropin stimulation at baseline and week 12 resulted in increased mean plasma cortisol levels in all treatment groups. Mean poststimulation plasma cortisol levels at week 12 were 674, 661, and 650 nmol/l in the 0.5-mg BIS, 1.0-mg BIS, and placebo treatment groups, respectively. Changes from baseline in poststimulated plasma cortisol levels were similar for both Table II. Shift from baseline to week 12 in cosyntropin-stimulated plasma cortisol values (nmol/l) Week 12 Treatment group Baseline value $500, n (%) <500, n (%) 0.5 mg BIS $ (75.0) 4 (14.3),500 2 (7.1) 1 (3.6) Total 23 (82.1) 5 (17.9) 1.0 mg BIS $ (70.6) 2 (11.8),500 3 (17.6) 0 (0.0) Total 15 (88.2) 2 (11.8) Placebo $ (77.4) 1 (3.2),500 3 (9.7) 3 (9.7) Total 27 (87.1) 4 (12.9) active treatment groups and placebo (ANCOVA, P >.58; Wilcoxon rank sum test, P >.89). Shifts in cosyntropin-stimulated plasma cortisol levels from baseline to week 12 are summarized in Table II for the evaluable population. Shifts from normal baseline ($500 nmol/l) to subnormal on-treatment levels (<500 nmol/l) 8 were observed in 4 (14.3%), 2 (11.8%), and 1 (3.2%) of the patients in the 0.5-mg BIS, 1.0-mg BIS, and placebo groups, respectively. Of these 7 patients, only 1 received treatment with an oral corticosteroid during the study (days 63 to 68). Shifts from subnormal at baseline to normal on treatment were observed in 2 (7.1%), 3 (17.6%), and 3 (9.7%) patients in the 0.5-mg BIS, 1.0 mg-bis, and placebo groups, respectively. Results of adrenal function analyses using the intentionto-treat patient population (ie, all randomly assigned patients who received $1 dose of study medication and who had $1 Safety Of Budesonide Inhalation Suspension In Infants Aged Six To Twelve Months With Mild To Moderate Persistent Asthma Or Recurrent Wheeze 93

142 Table III. Adverse events reported by $10% of patients in any treatment group Adverse event 0.5 mg BIS (n = 48) 1.0 mg BIS (n = 44) Placebo (n = 49) Total (N = 141) Respiratory infection 22 (45.8) 22 (50.0) 23 (46.9) 67 (47.5) Otitis media 23 (47.9) 12 (27.3) 20 (40.8) 55 (39.0) Fever 12 (25.0) 10 (22.7) 17 (34.7) 39 (27.7) Rhinitis 10 (20.8) 12 (27.3) 10 (20.4) 32 (22.7) Vomiting 5 (10.4) 7 (15.9) 5 (10.2) 17 (12.1) Aggravated asthma 4 (8.3) 4 (9.1) 8 (16.3) 16 (11.3) Diarrhea 4 (8.3) 5 (11.4) 6 (12.2) 15 (10.6) Tooth disorder 6 (12.5) 7 (15.9) 2 (4.1) 15 (10.6) Sinusitis 4 (8.3) 6 (13.6) 4 (8.2) 14 (9.9) Coughing 4 (8.3) 2 (4.5) 6 (12.2) 12 (8.5) Rash 5 (10.4) 4 (9.1) 3 (6.1) 12 (8.5) Conjunctivitis 7 (14.6) 0 (0.0) 4 (8.2) 11 (7.8) Pharyngitis 7 (14.6) 0 (0.0) 2 (4.1) 9 (6.4) Rhonchi 1 (2.1) 1 (2.3) 6 (12.2) 8 (5.7) Values shown are number (%) of patients. observation taken while receiving study medication; N = 141) were consistent with those presented for the evaluable population. Also, no differences were noted between the BIS groups and placebo group in the change from baseline in plasma cortisol values within either of the 2 age strata: 6 to <9 months or 9 to <12 months (data not shown). ADVERSE EVENTS. One hundred twenty-nine of 141 patients (91.5%) had $1 AE while receiving treatment. Approximately 90%, 98%, and 88% of patients in the 0.5-mg BIS, 1.0-mg BIS, and placebo groups, respectively, reported $1 AE. The most commonly reported AEs are shown in Table III. Overall, the types and frequencies of AEs reported during the study were comparable across treatment groups. Tooth disorder (eg, teething, teething pain, teething syndrome), reported at a higher frequency with BIS treatment, was considered by the study investigator to be unrelated to study medication in each case. The frequency of AEs did not differ appreciably as a function of age strata, except in the case of rhinitis. Rhinitis was reported by about twice as many patients 6 to 9 months of age compared with patients 9 to 12 months of age in both the 0.5-mg BIS (27% vs 14%) and placebo (28% vs 13%) groups. Discontinuations caused by AEs were limited to 1 patient in the 0.5-mg BIS group who discontinued treatment early after a treatment-related rash on the face and neck and 2 patients in the 1.0-mg BIS group after serious AEs requiring hospitalization (asthma [n = 1] and pneumonia [n = 1]). Serious AEs were reported in 3 patients in the 0.5-mg BIS group (asthma, respiratory infection, viral infection) and 2 patients in the 1.0-mg BIS group (asthma, pneumonia). All of the serious AEs occurred in children 9 to <12 months of age, and none were considered related to treatment. OTHER SAFETY ASSESSMENTS. Mean and median changes in vital signs from baseline to each visit were small, with no clinically meaningful differences between the treatment groups. Mean height increased from baseline to week 12 in all treatment groups and in both age strata. Compared with placebo, mean differences in change in height from baseline of 0.2 cm (95% CI, 20.6 to 1.0) and 0.6 cm (95% CI, 20.2 to 1.4) were observed in the 0.5-mg BIS and 1.0-mg BIS groups, respectively. Positive fungal nasal cultures were rare, and there were no significant differences between the treatment groups with respect to fungal cultures at week 12 when adjusted for baseline culture results. Mean and median changes in hematology and clinical chemistry values were generally comparable among the 3 treatment groups. Efficacy Compared with patients receiving placebo, the mean percentage of symptom-free days was numerically greater for patients receiving 0.5 and 1.0 mg BIS (37.5%, 48.8%, and 43.4%, respectively). Treatment-related differences in physicians global assessments of patient health status did not reach statistical significance, but physicians rated 90% and 85% of patients in the 0.5-mg BIS and 1.0-mg BIS groups, respectively, a great deal better or somewhat better compared with 67% of patients in the placebo group. BIS treatment resulted in a greater reduction in daytime and nighttime symptoms compared with placebo. DISCUSSION The primary purpose of this 12-week study was to collect additional safety data for nebulized BIS in a patient population that is underrepresented in clinical trials of ICS. Adrenal function, the primary variable, was assessed in children with asthma 6 to 12 months of age, based on changes in cosyntropinstimulated plasma cortisol levels. The cosyntropin-stimulation test is considered an adequate measure of adrenal function Berger et al The Journal of Pediatrics January 2005

143 Consistent with previously reported studies showing no effect of short-term (12 weeks) or long-term (52 weeks) BIS treatment on adrenal function in children 6 months to 8 years of age, 3-5,7 the current study demonstrated no suppression of adrenal function with once-daily BIS treatment in infants 6 to 12 months of age. Compared with placebo, treatment with BIS was not associated with a difference in response to cosyntropin stimulation. The safety of BIS, based on other safety variables (ie, AEs, study discontinuations, and clinical laboratory test results), was similar to that of placebo and consistent with that in children older than 12 months. 3-5 The most frequently reported AEs were those common to children with asthma (respiratory infection, otitis media, fever, rhinitis). The frequency of conjunctivitis, the most commonly reported ocular AE, was low. A review of BIS safety that included studies enrolling children as young as 6 months similarly reports a low frequency of ocular AEs with BIS. 10 In general, proper administration of nebulized treatments should limit ocular deposition and any potential for ocular AEs. As in previous short-term growth studies of ICS use in children, patients receiving BIS demonstrated reduced growth velocity compared with those receiving placebo. However, changes in growth velocity are more adequately addressed in long-term studies. Long-term studies with inhaled budesonide have shown that initial reductions in growth velocity are transient 11 and that expected final adult height can be achieved with long-term treatment. 12 Assessing treatment efficacy in infants is difficult because data (eg, symptoms) are reported indirectly by the parent or guardian, and it is difficult to obtain objective lung function measurements. 13 Moreover, the current study was not powered to address efficacy. Nevertheless, asthma symptom control and global health status were somewhat improved in patients treated with BIS versus placebo. Adrenal function was assessed only in a subset of patients who provided appropriate plasma cortisol samples and remained in the study for $78 days after random assignment, so additional data should be sought for this patient population despite the difficulties involved. In conclusion, this study provides additional safety data for BIS in an understudied patient population. The results of this 12-week study, conducted exclusively in infants 6 to 12 months of age, indicate that once-daily administration of BIS, 0.5 or 1.0 mg, is safe and well-tolerated in infants with mild to moderate persistent asthma or recurrent wheeze. REFERENCES 1. National Asthma Education and Prevention Program. Expert Panel Report: Guidelines for the Diagnosis and Management of Asthma Update on Selected Topics: J Allergy Clin Immunol 2002;110(Suppl 5): S Leflein JG, Szefler SJ, Murphy KR, Fitzpatrick S, Cruz-Rivera M, Miller CJ, et al. Nebulized budesonide inhalation suspension compared with cromolyn sodium nebulizer solution for asthma in young children: results of a randomized outcomes trial. Pediatrics 2002;109: Shapiro G, Mendelson L, Kraemer MJ, Cruz-Rivera M, Walton- Bowen K, Smith JA. Efficacy and safety of budesonide inhalation suspension (Pulmicort Respules) in young children with inhaled steroid-dependent, persistent asthma. J Allergy Clin Immunol 1998;102: Kemp JP, Skoner DP, Szefler SJ, Walton-Bowen K, Cruz-Rivera M, Smith JA. Once-daily budesonide inhalation suspension for the treatment of persistent asthma in infants and young children. Ann Allergy Asthma Immunol 1999;83: Baker JW, Mellon M, Wald J, Welch M, Cruz-Rivera M, Walton- Bowen K. A multiple-dosing, placebo-controlled study of budesonide inhalation suspension given once or twice daily for treatment of persistent asthma in young children and infants. Pediatrics 1999;103: Leflein JG, Gawchik SM, Galant SP, Lyzell E, Young M, Cruz-Rivera M, et al. Safety of budesonide inhalation suspension (Pulmicort Respules) after up to 52 weeks of treatment in infants and young children with persistent asthma. Allergy Asthma Proc 2001;22: Irani A-M, Cruz-Rivera M, Fitzpatrick S, Hoag J, Smith JA. Effects of budesonide inhalation suspension on hypothalamic-pituitary-adrenal axis function in infants and young children with persistent asthma. Ann Allergy Asthma Immunol 2002;88: Grinspoon SK, Biller BMK. Clinical review 62: laboratory assessment of adrenal insufficiency. J Clin Endocrinol Metab 1994;79: Chrousos GP, Harris AG. Hypothalamic-pituitary-adrenal axis suppression and inhaled corticosteroid therapy, I: general principles. Neuroimmunomodulation 1998;5: Szefler SJ, Lyzell E, Fitzpatrick S, Cruz-Rivera M. Safety profile of budesonide inhalation suspension in the pediatric population: worldwide experience. Ann Allergy Asthma Immunol 2004;93: The Childhood Asthma Management Program Research Group. Longterm effects of budesonide or nedocromil in children with asthma. N Engl J Med 2000;343: Agertoft L, Pedersen S. Effect of long-term treatment with inhaled budesonide on adult height in children with asthma. N Engl J Med 2000;343: Kemp JP. Study designs and challenges in clinical studies conducted in infants and children with asthma. J Allergy Clin Immunol 1999;104: S Safety Of Budesonide Inhalation Suspension In Infants Aged Six To Twelve Months With Mild To Moderate Persistent Asthma Or Recurrent Wheeze 95

144 EFFECT OF ASTHMA INTERVENTION ON CHILDREN WITH UNDIAGNOSED ASTHMA CHRISTINE L. M. JOSEPH, PHD, SUZANNE HAVSTAD, MA, ELIZABETH W. ANDERSON, MPH, RANDALL BROWN, MD, CHRISTINE COLE JOHNSON, PHD, AND NOREEN M. CLARK, PHD Objective To measure the effect of an asthma intervention on the functional status and morbidity of children with undiagnosed asthma. Study design Data from a randomized trial were used to compare outcomes at baseline and follow-up for children with undiagnosed and diagnosed asthma. We studied 510 symptomatic children with diagnosed asthma (diagnosed) and 299 children with symptoms but no diagnosis (undiagnosed). Baseline functioning and morbidity were similar for undiagnosed and diagnosed patients classified as moderate-severe. Results There were fewer undiagnosed reported allergies, seasonal symptoms, and other respiratory diagnoses (all P < 0.01). Among the moderate-severe, functional status, for example, symptom-days (P =.02), symptom-nights (P <.01), and days of restricted activity (P <.01), was significantly reduced at follow-up for the undiagnosed in the intervention group but not for undiagnosed control subjects. Findings were similar for children with diagnosed asthma. Conclusions Children with undiagnosed asthma were generally nonatopic, although some had symptoms at a level comparable to children with a diagnosis. The intervention successfully improved functional status for children with undiagnosed asthma as well as for children with diagnosed asthma. These results can be applied to ongoing discussions related to case detection. (J Pediatr 2005;146:96-104) Estimates for asthma morbidity and mortality in the United States have been highest in low-income, urban communities. The unsettling statistics, along with studies showing that asthma can be controlled once appropriately treated, have led researchers and health care providers to explore screening for this condition, especially in communities considered to be at high risk. 5,6 A wide range of activities have fallen under the heading of screening. 5-7 The Centers for Disease Control and Prevention (CDC) apply this term to the use of spirometry to identify individuals who exhibit signs and symptoms of asthma. 7 Case detection is used by CDC to refer to questionnaires that identify people who report asthma symptoms. 7 One reason that providers and researchers are interested in screening or case detection for asthma is to identify those in need of intervention, including people with symptoms indicative of asthma, but with no formal diagnosis. To accomplish this task, questionnaires have an advantage over spirometry in that they are less expensive and more practical for field work; however, there are limitations to their use. For example, the probability that the children identified do not have asthma (false-positive results) may be high, and the additional resources needed to clinically confirm questionnaire findings can be substantial. 2-6 Another problem is that questionnaires are often administered at one point in time with no additional follow-up. Administering questionnaires more frequently to determine symptom patterns over time can be costly as well as logistically challenging. Increasing the certainty that resources can be used effectively and efficiently to reduce morbidity in children with asthma symptoms requires case identification methods to gather more information on the symptom experience over time, especially for children classified as having undiagnosed asthma. To date, there are few longitudinal data sets with the ability to BMI CDC ED Body mass index Centers for Disease Control Emergency Department GEE NAEPP Generalized estimating equation National Asthma Education and Prevention Program Expert Panel 1-4 From Henry Ford Health System, Department of Biostatistics and Research Epidemiology, Detroit, Michigan; the University of Michigan School of Public Health, Ann Arbor, Michigan; and the University of Michigan School of Medicine, Ann Arbor, Michigan. Supported by the National Heart, Lung, and Blood Institute. Submitted for publication Mar 2, 2004; last revision received Jul 8, 2004; accepted Sep 1, Reprint requests: Dr Christine L. M. Joseph, Henry Ford Health System, Department of Biostatistics and Research Epidemiology, 1 Ford Place, Suite 3E, Detroit, MI cjoseph1@hfhs.org /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

145 Figure. Eligibility and exclusion criteria applied to participants used in the analysis: Effect of an asthma intervention on children with undiagnosed asthma. explore the persistence of asthma symptoms reported by children classified as undiagnosed and the impact of an intervention on their functional status and morbidity. We used data from a randomized trial to describe and compare characteristics of children fulfilling study criteria for undiagnosed asthma with that of children with a physician diagnosis of asthma. 4 These data have been used previously to describe the feasibility of school-based case detection for asthma and the issues related to this activity. 4 Toward this goal, the data were used to compare changes in symptoms, parent management, grades, and school absenteeism for all children in the intervention group as opposed to the control group. 8 In this study, we present the first attempt to describe changes in functional status and morbidity for undiagnosed children exposed to an asthma intervention. In addition to comparing baseline characteristics for children with undiagnosed and diagnosed asthma, the goals of this analysis were (1) to describe symptom frequency, functional status, and asthma morbidity for children fulfilling criteria for undiagnosed asthma; (2) to examine the functional status and morbidity reported by the undiagnosed over time; and (3) to determine whether an intervention would improve functional status and reduce morbidity in children with undiagnosed asthma. All comparisons were made with children from the same population who reported a physician diagnosis of asthma. METHODS The data used in this analysis were from a randomized trial evaluating the effect of a comprehensive school-based asthma program on the functional status and asthma-related morbidity of urban elementary schoolchildren. 4 The goal of the intervention was to promote, teach, and support better self-management of asthma among participating children and their families. The setting for the trial was a geographically defined area in the Detroit Public School System. The average percentage of children <18 years of age meeting federal guidelines for poverty in the six zip codes comprising this area was 44.6%, according to the 1990 Census. The primary outcome variables included improvements in functional status (eg, symptom-days and days of restricted activity) and Effect Of Asthma Intervention On Children With Undiagnosed Asthma 97

146 Table I. Comparison of selected characteristics for participating children by diagnostic status Undiagnosed (n = 299) Diagnosed (n = 510) n % n % OR (95% CI) P value Sex of child Male 141 (47.2) 282 (55.3) 0.7 ( ).03 Female 158 (52.8) 228 (44.7) Household income.15, (58.4) 248 (52.1) 1.2 ( ).11 #15, (41.6) 228 (47.9) Caregiver high school diploma Yes 254 (87.0) 435 (86.8) 1.0 ( ).93 No 38 (13.0) 66 (13.2) $1 Smoker in household Yes 162 (54.2) 267 (52.4) 1.1 ( ).38 No 137 (45.8) 243 (47.7) Used prescribed asthma medication Yes 25 (8.4) 331 (64.9) 0.05 ( ),.001 No 274 (91.6) 179 (35.1) Other respiratory diagnoses* Yes 97 (32.4) 340 (66.7) 0.2 ( ),.001 No 202 (67.6) 170 (33.3) Allergies Yes 82 (27.4) 227 (44.5) 0.5 ( ),.001 No 217 (72.6) 283 (55.5) Seasonal variation in symptoms Yes 144 (57.8) 359 (75.7) 0.4 ( ),.001 No 129 (47.3) 115 (24.3) Usual activity level = high Yes 206 (69.1) 331 (64.9) 1.2 ( ).19 No 92 (30.9) 179 (35.1) Wheeze with exercise Yes 109 (37.6) 371 (74.2) 0.2 ( ),.001 No 181 (62.4) 129 (25.8) Cough with exercise Yes 239 (81.3) 391 (78.0) 1.2 ( ).28 No 55 (18.7) 110 (22.0) *Child has received other respiratory diagnoses, for example, bronchitis, reactive airway disease, viral bronchitis, wheezing or wheezy bronchiolitis, asthmatic bronchitis. Table II. Comparison of reported baseline functional status by severity classification for children with diagnosed and undiagnosed asthma * Undiagnosed Diagnosed n Mean (SD) n Mean (SD) P value Mild-intermittent and mild-persistent Symptom-days (10.5) (8.7).79 Symptom-nights (2.3) (2.7).01 Changed caregiver plans (1.9) (3.9),.01 Restricted activity (2.4) (9.5),.01 Missed school (2.4) (3.3),.01 Moderate-persistent and severe-persistent Symptom-days (29.5) (29.3).86 Symptom-nights (28.4) (29.2).61 Changed caregiver plans (6.6) (10.2).68 Restricted activity (16.2) (21.0).37 Missed school (5.6) (9.5).08 Functional status = symptom-days, symptom-nights, changed caregiver plans, restricted activity, school days missed. *Recall period is last 3 months. reductions in asthma-related morbidity (eg, emergency department [ED] visits and asthma-related hospitalizations). The methods for this study are described in a previous publication. 4,8 Briefly, potential participants were identified through caregiver responses to an eligibility questionnaire sent to the home. Included were questions on demographics, child s height and weight, occurrence and frequency of respiratory symptoms, prescribed medications currently being 98 Joseph et al The Journal of Pediatrics January 2005

147 Table III. Functional status and morbidity over time for undiagnosed and diagnosed by randomization group: Mild (n = 483) * Variable Undiagnosed control subjects Baseline Follow 1 Follow 2 Baseline Undiagnosed intervention Follow 1 Follow 2 Baseline Diagnosed control subjects Follow 1 Follow 2 Baseline Diagnosed Intervention Follow 1 Follow 2 Symptom-days (9.8) (17.0) (15.2) (7.0) (8.2) (11.1) (10.2) (9.5) (9.1) (7.4) P =.03 y P =.03 P,.01 P =.78 Symptom-nights (2.1) (5.7) (2.9) (2.3) (24.9) (6.6) (2.5) (12.9) (11.1) (2.7) P =.06 P =.99 P =.11 P =.76 Changed plans z (1.5) (1.0) (0.7) (1.7) (1.1) (1.5) (2.5) (4.4) (1.9) (5.6) P =.01 P =.52 P =.02 P = (15.4) 3.0 (11.1) 1.6 (10.0) 5.2 (13.3) 4.5 (16.4) 0.9 (2.3) Restricted activity (1.5) (12.6) (1.8) (3.0) (13.0) (1.9) (6.2) (7.5) (7.9) (12.0) P =.98 P =.69 P,.01 P = (19.9) 3.7 (12.3) Missed school days (1.7) (2.5) (2.9) (2.7) (3.7) (2.6) (3.5) (3.9) (2.6) (3.4) P =.88 P =.92 P,.01 P = (2.5) 1.3 (2.8) Mean ED visits (0.5) (0.5) (0.9) (1.4) (0.7) (0.8) (1.9) (1.7) (1.2) (3.6) P =.28 P =.92 P,.01 P, (2.2) 0.5 (1.6) $ 1 ED visit, n (%) (7.2) (8.4) (8.4) (10.9) (9.8) (10.9) (32.9) (19.2) (19.2) (36.2) P =.67 P =.42 P =.02 P = (13.5) 27 (19.2) Functional status = symptom-days, symptom-nights, changed caregiver plans, restricted activity, school days missed; morbidity = ED visits and hospitalizations. *Includes persons fulfilling symptom criteria for mild-intermittent and mild-persistent asthma. yall P values are the result of generalized estimating equations. Models were adjusted for age, sex, caregiver education, and household income. zcaregiver changed plans. taken for respiratory symptoms, the occurrence of respiratory symptoms or breathing problems with weather and season changes, exposure to cigarette smoke, and whether or not the child had a physician diagnosis of asthma. Children were eligible to enroll if they met study criteria for currently diagnosed or undiagnosed asthma. Baseline interviews were conducted with caregivers of all eligible children. Variables collected at baseline included education of primary caregiver, household income, and medications being currently used for asthma symptoms. Variables collected for measurement of functional status included symptom-days, symptom-nights, changes in caregiver plans, days of restricted activity, and school days missed because of asthma symptoms. The time frame for the functional status variables was the last 3 months. Caregivers were asked about the child s usual activity level (high, moderate, low). All caregivers were asked if the child had ever received the following alternative diagnoses: bronchitis, reactive airway disease, viral bronchitis, wheezing or wheezy bronchitis, and asthmatic bronchitis. Also collected was information on asthma morbidity in the last 12 months, for example, symptom exacerbations resulting in ED visits or hospitalizations. After the baseline assessment, schools were randomized to an intervention or control group. Asthma education in the form of open airways was provided to eligible children in the intervention schools, along with a series of activities for their caregivers, teachers, building personnel, and classmates. 4,8,9 The control group was asked to continue with their usual care for asthma. Follow-up telephone interviews with the caregivers of all participating children were scheduled for 12 months and 24 months after intervention. At each follow-up interview, caregivers were once again asked about a physician diagnosis of asthma, functional status, and asthma-related morbidity. Analysis Current asthma was defined as the report of a physician diagnosis of asthma with either a report of symptoms in the Effect Of Asthma Intervention On Children With Undiagnosed Asthma 99

148 Table IV. Functional status and morbidity over time for undiagnosed and diagnosed by randomization group: Moderate-severe (n = 119) * Variable Undiagnosed control subjects Baseline Follow 1 Undiagnosed intervention Diagnosed control subjects Diagnosed intervention Follow Follow Follow Follow Follow Follow Follow 2 Baseline 1 2 Baseline 1 2 Baseline 1 2 Symptomdays (34.5) (25.4) (16.3) (20.4) (5.9) (21.3) (27.7) (10.6) (15.2) (28.9) P =.35 y P =.02 P,.01 P, (29.1) 11.9 (25.6) Symptomnights (34.4) (1.3) (1.1) (26.7) (1.7) (3.5) (24.6) (5.7) (14.0) (32.0) P =.03 P,.01 P,.01 P, (23.7) 2.5 (5.3) Changed plans z (1.7) (0.4) (0.6) (8.0) (0.9) (1.0) (4.8) (3.7) (13.7) (13.6) P =.05 P =.03 P =.24 P = (17.2) 1.2 (2.5) Restricted activity (2.9) (14.0) (1.6) (1.4) (21.9) (3.3) (11.5) (20.5) P =.06 P,.01 P =.13 P = (5.9) 5.5 (16.2) Missed school days (4.0) (5.6) (5.8) (2.2) (6.5) (5.6) (9.0) (12.2) P =.07 P =.02 P =.71 P, (3.6) 2.4 (4.1) Mean ED visits $ 1 ED visit, n (%) (0.5) 0.3 (0.9) 0.2 (0.6) 1.9 (3.0) 0.9 (1.8) 1.5 (2.8) 2.1 (3.7) 1.3 (2.4) Unable to calculate P =.79 P =.02 P = (10.5) (5.3) (10.5) (55.6) (24.4) (33.3) (51.2) (23.3) Unable to calculate P =.23 P =.16 P = (3.4) 15 (34.9) Functional status = symptom-days, symptom-nights, changed caregiver plans, restricted activity, and school days missed; morbidity = ED visits and hospitalizations. *Includes persons fulfilling symptom criteria for mild-intermittent and mild-persistent asthma. yall P values are the result of generalized estimating equations. Models were adjusted for age, sex, caregiver education, and household income. zcaregiver changed plans. past 12 months or a report of having used a prescribed asthma medication in the past 12 months. Undiagnosed asthma was defined as a caregiver report of at least 3 of the following 5 symptoms in the past 12 months: persistent cough, wheeze with a cold, wheeze without a cold, wheeze accompanied by dyspnea, chest tightness or heaviness; or cough/wheeze with exercise on at least 3 occasions in the last 12 months. To classify participants by severity, the frequency of daytime and nighttime symptoms used in the asthma severity classifications of the National Asthma Education and Prevention Program Expert Panel II (NAEPP-II) were applied to caregiver reports. 4,8,9 Investigators interpreted and assigned numeric values when terms such as frequent and continual were used in the NAEPP-II criteria. When NAEPP-II criteria used the timeframe of 1 month, investigators calculated a 90-day (or 3-month) equivalent. This pattern was also followed for daily and weekly timeframes for symptom frequencies. The resulting severity classifications have been used in previous publications. 4,8 To compare characteristics, functional status, and morbidity of the undiagnosed to those of the diagnosed while adjusting for clustering, appropriate generalized linear modeling procedures were used, including SAS procedures PROC MIXED for continuous variables and PROC GENMOD for categoric variables. 10 When appropriate, odds ratios and the corresponding 95% confidence intervals were calculated to describe the association of interest. Generalized estimating equation (GEE) models were used to assess changes in functional status and morbidity over the study period. This method of analysis was selected to maximize the sample size for analysis and to account for the correlation between repeated measurements. A normal distribution was assumed for continuous data (eg, age) and a binomial distribution for binary data (eg, proportion making 100 Joseph et al The Journal of Pediatrics January 2005

149 $ 1 ED visit). All continuous data were transformed by the square root transformation before GEE analysis; however, for ease of interpretation, means and standard deviations of untransformed data are provided in tables and results. RESULTS The Figure shows the characteristics of the children whose caregiver participated in the eligibility phase of the study. Caregivers of 4576 elementary school children attending 14 urban elementary schools completed an eligibility form (83% response rate after taking into account 7% average daily absenteeism for Detroit Public Schools). Approximately 97% of children were black and 47% were male. The prevalence of physician-diagnosed asthma was 16.9%. The majority of these children reported symptoms or had taken prescribed asthma medication within the past 12 months, thus fulfilling criteria for current asthma (n = 720, or 15.7%). A total of 533 children (11.7%) reported asthma symptoms but did not report a physician diagnosis (undiagnosed). We excluded 26 children whose caregiver reported no physician diagnosis at eligibility but later reported a diagnosis at baseline. This left 507 undiagnosed children (11.1%), of whom 299 had caregivers who completed a baseline interview and at least 1 follow-up interview. Among the diagnosed, 510 children had caregivers who completed the baseline interview and at least 1 follow-up interview. Participants were classified by severity, using the NAEPP-II classifications as a guide 11 (Figure). A smaller percentage of the diagnosed fell into the mild-intermittent category than the undiagnosed (47.6% vs 60.5%, respectively; P <.001). Percentages were similar for mild-persistent (29.4% diagnosed vs 24.9% undiagnosed; P =.08) but less so for the moderate-persistent (17.1% vs 11.7%; P =.06). Despite the almost 2-fold difference in the percentage of undiagnosed children in the severe-persistent category compared with that of the diagnosed, this difference did not reach statistical significance (5.9% diagnosed vs 3.0% undiagnosed, P =.13). Baseline characteristics of children with undiagnosed asthma were compared with those of children with an asthma diagnosis. The mean age of the undiagnosed and diagnosed groups were similar, 8.1 years (SD = 1.3) and 8.1 years (SD = 1.2) respectively; P =.45. As shown in Table I, when compared with the diagnosed, undiagnosed children were significantly less likely to be male (OR = 0.7 [0.5 to 1.0], P =.03); to be taking prescribed asthma medications (OR = 0.05 [0.03 to 0.07], P <.001); and were significantly less likely to report alternative diagnoses (OR = 0.2 [0.2 to 0.3], P <.001). Undiagnosed children were also significantly less likely to report allergies (OR = 0.5 [0.4 to 0.6], P <.001) and a seasonal variation in symptoms (OR = 0.4 [0.3 to 0.5], P <.001). Significantly fewer undiagnosed children had wheeze with exercise when compared with the diagnosed (37.6% vs 74.2%, respectively; P <.01). Because body mass index (BMI) could influence symptom reports, we examined this variable by whether or not the child had an asthma diagnosis at baseline (data not shown). Approximately 49% of participants had information sufficient for calculating BMI. The mean BMI for the undiagnosed and diagnosed groups was 20.6 (SD = 6.5) and 21.2 (SD = 6.6), respectively (P =.39). Table II shows the reported baseline functional status for children in the mild-intermittent and mild-persistent categories combined. The mean number of days of impaired functioning reported for the undiagnosed was significantly lower than that reported by the diagnosed in the mild categories (all P values #.03), with the exception of symptomdays (P =.79). Reports of impaired functioning were similar for the undiagnosed and diagnosed in the moderate and severe persistent categories (all P values $.08). Using GEE, we tested the differences between undiagnosed and diagnosed children with regard to change in functional status and morbidity over time. Analyses were stratified by randomization status and by severity, such that the mild-intermittent and mild-persistent categories were collapsed to form one category, as were the moderatepersistent and the severe-persistent. Analyses were adjusted for age, sex, household income, and caregiver education. Mild (Mild-Intermittent and Mild-Persistent) UNDIAGNOSED CONTROL SUBJECTS VERSUS UNDIAGNOSED INTERVENTION. Among the undiagnosed control subjects in the mild category, significant reductions were observed only for symptom-days (P =.03) and caregiver changed plans (P =.01). Likewise, only symptom-days was significantly reduced (P =.03) for the undiagnosed intervention in the mild category (Table II). Asthma morbidity was not significantly reduced from baseline for either group. DIAGNOSED CONTROL SUBJECTS VERSUS DIAGNOSED GINTERVENTION. Symptom-days (P <.01), changed plans (P =.02), restricted activity (P <.01), and missed school days (P <.01) were significantly reduced for mild diagnosed control subjects. Only changed plans (P =.01) was significantly reduced for the diagnosed intervention group (Table III). Significant reductions were observed for both the mild diagnosed control subjects and the mild diagnosed intervention group with regard to asthma morbidity (P <.05 for mean ED visits for both groups; P =.02 and.05 for diagnosed control subjects and intervention, respectively, making $ 1 ED visit). Moderate-Severe (Moderate-Persistent and Severe-Persistent) UNDIAGNOSED CONTROL SUBJECTS VERSUS UNDIAGNOSED INTERVENTION. Among the undiagnosed control subjects in the moderate-severe group, only symptom-nights was significantly reduced over time (P =.03). In contrast, for the undiagnosed intervention group, all reports of impaired functioning were significantly reduced (all P values #.03) (Table IV). Results for asthma morbidity were not calculated because of small cell sizes for the undiagnosed control subjects and were insignificant for the undiagnosed intervention group. Effect Of Asthma Intervention On Children With Undiagnosed Asthma 101

150 DIAGNOSED CONTROL SUBJECTS VERSUS DIAGNOSED INTERVENTION. For diagnosed control subjects in the moderate-severe category, symptom-days (P <.01) and symptom-nights (P <.01) were significantly reduced. By comparison, all measurements of impaired functioning were significantly reduced for those in the diagnosed intervention group. Significant reductions in asthma morbidity were observed in the diagnosed control subjects for mean ED visits only (P =.02) and in the diagnosed intervention group for the number of persons making $ 1 ED visit (P =.03). Attrition We examined attrition by random assignment. For the first follow-up, control subjects were more likely to drop out of the study when compared with the intervention group (P =.04), but attrition was similar by second follow-up (P =.74). If examined by the presence of an asthma diagnosis at baseline, undiagnosed control subjects were more likely to drop out than were diagnosed control subjects (P =.02 and.05 for first and second follow-up, respectively), but there were no differences in attrition observed for the undiagnosed versus the diagnosed in the intervention groups. We also compared the characteristics and functional status of persons completing only a baseline interview (ie, no information for follow-up) with that of persons completing a baseline and at least 1 follow-up. The undiagnosed with no follow-up were slightly younger (0.3 years) than the undiagnosed with follow-up (P =.01), but this was the only difference observed. DISCUSSION The rationale behind case detection is that it should be possible to identify children with undiagnosed disease by asking about the symptoms we would expect children with asthma to have. The single most pressing methodologic concern is the lack of a gold standard for using case detection to identify children with undiagnosed asthma. 12,14 The process of reaching consensus on a definition should include the results of intervening on the undiagnosed once they have been identified using a specific instrument. This issue has also been discussed in previous publications. 2,4,12 Although the definition varies widely, the literature estimates the prevalence of undiagnosed asthma to be 5.3% to 20% when case detection is used in children less than 18 years. Similar estimates are reported for studies conducted in US urban cities and nonwhite communities. 2-5,12-18 Estimates up to 37.5% have been reported for individual symptoms, such as wheeze in the absence of an asthma diagnosis. 16,19-21 Wide discrepancies in prevalence also occur when clinical outcomes, eg, spirometry, are used to identify children with undiagnosed asthma, as opposed to case detection. 2,12,22 Previous studies have suggested that undiagnosed children cluster at the low end of a severity continuum. 2,12 The expenditure of resources on this group has been challenged because the benefit may be so low as to be undetectable. Information on the functional status and morbidity of the undiagnosed over time as well as at one given time point may help to estimate the return on such an investment. In comparing baseline characteristics, we found that the undiagnosed children were less likely to report allergies, a seasonal variation in symptoms, and receipt of other respiratory diagnoses when compared with the diagnosed. Together, these findings suggest a nonatopic form of disease that may have added to the difficulty in obtaining a diagnosis. In addition to the treatment and management implications this information presents to health care providers, results suggest that principles in the asthma management program were also applicable to children with nonatopic asthma. 23,24 The absence of factors related to atopic asthma has been reported in at least two other studies using a case detection approach. According to unpublished data from the National Cooperative Inner City Asthma Study (an earlier study of asthma in urban children), the undiagnosed were more likely to have a negative medical history, including the absence of atopy, as determined by skin prick tests. 23 In the Odense schoolchild study (Denmark), factors related to undiagnosed asthma included the absence of rhinitis symptoms. 15 However, in a pilot study of elementary school children in Detroit, children with undiagnosed asthma were less likely to report allergies and eczema but were more likely to have received a diagnosis of bronchitis, compared with children with an asthma diagnosis. 2 The undiagnosed also had fewer reports of wheeze with exercise, but it is unclear if this was due to being less active or fewer symptoms during activity. Evidence for the latter scenario may be that caregivers of the undiagnosed reported fewer days of restricted activity, but reports of the child s activity level among the undiagnosed were similar to that of the diagnosed. The distribution of severity, as determined by symptom type and frequency, was skewed toward milder disease for the undiagnosed. Differences in distribution were generally less pronounced as severity levels increased. Since use of prescribed medication to treat symptoms was rare among the undiagnosed, we would expect the distribution of severity to be skewed even further toward milder levels if these children were treated. The reports of symptoms appeared to persist throughout the study period, although there was less evidence of this among the undiagnosed control subjects in the moderate-severe category, probably because of the small sample size. In terms of functional status, symptoms did not appear to interfere with functioning among undiagnosed children categorized as mild-intermittent and mild-persistent. Our results suggest that the undiagnosed had symptoms of a lesser intensity but at a frequency similar to the diagnosed, since reports of symptom-days were similar for the two groups. If the symptoms were of less intensity, there may have been less motivation to seek medical attention. In contrast, children classified as moderate and severe-persistent had a level of functioning similar to that of the diagnosed in this same category. Results suggest that this latter group of undiagnosed may have been more tolerant of symptoms or were hampered 102 Joseph et al The Journal of Pediatrics January 2005

151 in terms of health care access, since they had not yet received a diagnosis of asthma. The results of our analysis suggest that the intervention was effective for the more severe groups, with the greatest benefit to functionality, as opposed to morbidity. These results must be interpreted with caution. We note that general declines in reports of impaired functioning over the study period were observed for almost all diagnostic groups, regardless of random assignment to the treatment or control group. In addition to an actual decrease in symptoms, this may be due to some study artifact that we did not adjust for or may reflect a regression to the mean. 25 It also may be due to a remission of symptoms as the children grow older. 26 For both functional status and morbidity, the undiagnosed had significantly lower baseline estimates than the diagnosed for all factors, leaving less variation for large reductions. In addition, ED visits were rare among the undiagnosed. Stratifying by severity level may have lessened the impact of these limitations on our results, but it also reduced our sample size, as did the need to adjust for random assignment. Given the above, it is difficult to determine if the levels of significance for observed reductions in functional status among the undiagnosed in the intervention group would not have been paralleled in the undiagnosed control group, but for a lack of power. Future longitudinal studies designed to assess health outcomes or morbidity patterns in children meeting some criteria for undiagnosed asthma need to be sufficiently large to have enough subjects for analysis. Attrition was a limitation in our study, as in similar asthma studies. 27 Control subjects were asked to continue with their usual care and thus received less impetus to remain in the program. Attrition could have caused our results to be underestimates (ie, a group that is predominantly mild at baseline is now predominately moderate-severe because of attrition of the former) or overestimates (ie, easier to detect differences over time in a more severe group). Both conditions would reduce the generalizability of our findings and may also have affected our ability to detect statistical differences by reducing our sample size. Nonetheless, we did not observe meaningful differences between participants and dropouts at baseline. Recall bias is a potential problem for studies relying on primary data collection such as ours. A 2-week recall period was used in the National Cooperative Inner City Asthma Study; however, other studies suggest that in describing the burden of asthma, using recall periods less than 12 months may serve to underestimate this burden. 28,29 We asked about symptoms in the last 3 months, and this recall period has been used previously. 30 Twelve-month recall for asthma ED visits and hospitalizations is often used in asthma surveillance and research. 1,2,27,29,30 In this study, children identified as having undiagnosed asthma and classified as moderate-severe, based on symptom reports, showed reductions in days of impaired functional status after an intervention. This was despite a generally nonatopic disease presentation. When using case detection to identify children with undiagnosed asthma, erring toward high specificity may result in a more effective and efficient use of resources. The authors thank the following persons for their assistance and support: Detroit Public Schools, Detroit Public Schools Office of Research, Evaluation, and Assessment, and Melissa Valerio and her team of interviewers. We also thank Manlan Lui for her assistance in preparing the data for analysis. REFERENCES 1. Centers for Disease Control and Prevention. Surveillance for asthma: United States, MMWR Surveillance Summaries 2002;51(SS01): Joseph CLM, Foxman B, Leickly FE, Peterson E, Ownby D. Prevalence of possible undiagnosed asthma and associated morbidity among urban schoolchildren. J Pediatr 1996;129: Crain EF, Weiss KB, Bijur PE, Hersh M, Westbrook L, Stein RE. An estimate of the prevalence of asthma and wheezing among inner-city children. Pediatrics 1994;94: Clark NM, Brown R, Joseph CLM, Anderson L, Liu M, Valerio M, Gong M. Issues in screening for asthma at school and estimating prevalence in a population of urban children. J Clin Epidemiol 2002;55: Yawn BP, Wollan P, Scanlon PD, Kurland M. Outcome results of a school-based screening program for undertreated asthma. Ann Allergy Asthma Immunol 2003;90: Boss LP, Wheeler LS, Williams PV, Bartholomew LK, Taggart VS, Redd SC. Population-based screening or case detection for asthma: are we ready? J Asthma 2003;40: Centers for Disease Control and Prevention. Strategies for Addressing Asthma Within a Coordinated School Health Program. Atlanta, Ga: Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion; 2002www.cdc.gov/healthyyouth/healthtopics/asthma. 8. Clark NM, Brown R, Joseph CLM, Anderson EW, Liu M, Valerio MA. Effects on symptoms, parent management, grades and absenteeism of a comprehensive school based asthma program. Chest 2004;125: American Lung Association. Open Airways for Schools. New York, NY: American Lung Association, SAS manual: SAS/STAT User s Guide, Version 8. Cary, NC: SAS Institute Inc; National Asthma Education Program; Office of Prevention, Education, and Control. Expert Panel Report 2: Guidelines for the diagnosis and management of asthma. National Heart, Lung, and Blood Institute; National Institutes of Health; Public Health Service; United States Department of Health and Human Services, Bethesda, Md; April 1997, NIH publication No Joseph CLM, Foxman B, Leickly FE, Peterson EL. Sensitivity and specificity of asthma definitions and symptoms used in a survey of childhood asthma. J Asthma 1999;36: Gerald LB, Redden D, Turner-Hensen A, Feinstein R, Hemstreet MP, Haines C, et al. A multi-stage asthma screening procedure for elementary school children. J Asthma 2002;39: Toelle BG, Peat JK, Salome CM, Mellis CM, Woolcock AJ. Toward a definition of asthma for epidemiology. Am Rev Respir Dis 1992;146: Siersted H, Boldsen J, Hansen HS, Mostgaard G, Hyldebrandt N. Population based study of risk factors for underdiagnosis of asthma in adolescence: Odense schoolchild study. Br Med J 1998;316: Christiansen SC, Martin SB, Schleicher NC, Koziol JA, Mathews KP, Zuraw BL. Current prevalence of asthma-related symptoms in San Diego s predominantly Hispanic inner-city children. J Asthma 1996;33: Feinstein RA, Hains CS, Hemstreet MP, Turner-Henson A, Redden DT, Martin B, et al. A simple step-test protocol for identifying suspected unrecognized exercise-induced asthma (EIA) in children. Allergy Asthma Proc 1999;20: Effect Of Asthma Intervention On Children With Undiagnosed Asthma 103

152 18. Nelson DA, Johnson CC, Divine GW, Strauchman C, Joseph CLM, Ownby DR. Ethnic differences in the prevalence of asthma among suburban schoolchildren. Ann Allergy Asthma Immunol 1997;78: Kaur B, Anderson HR, Austin J, Burr M, Harkins LS, Strachan DP, Warner JO. Prevalence of asthma symptoms, diagnosis, and treatment in year old children across Great Britain (International Study of Asthma and Allergies in Childhood, ISAAC UK). Br Med J 1998;316: Yeatts K, Shy C, Wiley J, Music S. Statewide adolescent asthma surveillance. J Asthma 2000;37: Hetlevik O, Ploen O, Nystad W, Magnus P. The wheezing schoolchild: an undiagnosed asthmatic: a follow-up of children with parentally reported episodes of wheeze without diagnosed asthma. Scand J Primary Health Care 2000;18: Remes ST, Pekkanen J, Remes K, Salonen RO, Korppi M. In search of childhood asthma: questionnaire, tests of bronchial hyperresponsiveness, and clinical evaluation. Thorax 2002;57: Joseph CLM, Ownby DR, Mitchell H, Visness CM, Peterson EL, Johnson CC, Leickly FE. Comparison of participants with undiagnosed and diagnosed asthma: National Cooperative Inner-City Asthma Study. Abstract. American Thoracic Society Annual Meeting, Orlanda, Fla, Oettgen HC, Geha RS. IgE in asthma and atopy: cellular and molecular connections. J Clin Invest 1999;104: Davis CE. The effect of regression to the mean in epidemiologic and clinical studies. Am J Epidemiol 1976;104: Guilbert T, Krawiec M. Natural history of asthma. Pediatr Clin North Am 2003;50: Bender BG, Ikle DN, DuHamel T, Tinkelman D. Retention of asthmatic patients in a longitudinal clinical trial. J Allergy Clin Immunol 1997;99: Mitchell H, Senturia Y, Gergen P, Baker D, Joseph C, McNiff- Mortimer K, et al. Design and methods of the National Cooperative Inner- City Asthma Study. Pediatr Pulmonol 1997;24: Fuhlbrigge AL, Adams RJ, Guilbert TW, Grant E, Lozano P, Janson SL, et al. The burden of asthma in the United States: level and distribution are dependent on interpretation of the National Asthma Education and Prevention Program Guidelines. Am J Respir Crit Care Med 2002;166: Ford ES, Mannino DM, Homa DM, Gwynn C, Redd SC, Moriarty DG, Mokdad AH. Self-reported asthma and health-related quality of life: findings from the behavioral risk factor surveillance system. Chest 2003; 123: Joseph et al The Journal of Pediatrics January 2005

153 ADVERSE EFFECT OF HIGH ADDED SUGAR CONSUMPTION ON DIETARY INTAKE IN AMERICAN PRESCHOOLERS SIBYLLE KRANZ, PHD, RD, HELEN SMICIKLAS-WRIGHT, PHD, ANNA MARIA SIEGA-RIZ, PHD, RD, AND DIANE MITCHELL, MSRD Objectives To assess the effect of Dietary Reference Intakes (DRI)-recommended added sugar intake levels on nutrient and food group consumption by preschoolers. Study design This was a cross-sectional study of 2- to 5-year-olds in the United States Department of Agriculture Continuing Survey of Food Intake by Individuals (CSFII), 1994 to 1996, and 1998 (n = 5437). Main food sources of added sugar were established. For five categories (#10% energy from added sugar, 11% to 15%, 16% to 20%, 21% to 25%, and >25%), mean nutrient and food group consumption and proportion of children not meeting the DRI were ascertained. The nonparametric test for trend indicated significance of changes. Analysis was stratified by age (2- to 3-year-olds versus 4- to 5-year-olds), and survey design corrected to maintain the nationally representative character of the data. Results The majority of children consumed less than 25% of energy from added sugar; the main sources were fruit and or soft drinks and desserts. Increasing added sugar consumption was paralleled by decreasing nutrient and food group intakes and increasing proportions of children with intakes below the DRI. Calcium intake was insufficient in large proportions of children consuming 16% or more from added sugar. Conclusions In preschoolers, the new DRI for added sugar are reason for concern. Further research is needed to investigate the effects of the DRI on diet quality, body weight, and health status in young children. (J Pediatr 2005;146:105-11) National and international groups have issued recommendations for added sugar intake. In the United States Department of Agriculture s (USDA) Food Guide Pyramid, added sugar consumption is suggested to range between 6% to 10% of total energy. 1 The World Health Organization recommends limiting added sugar intake to <10% of total energy and in the 2003 report states that a reduction of sugar-sweetened drinks for children is a strategy to reduce chronic diseases. 2 In contrast to these guidelines, the National Academy of Sciences (NAS) released the recommendations for energy and macronutrient intakes, which included a Dietary Reference Intake (DRI) of 25% or less of total energy from added sugar. This cut-point was based on evidence of potentially adverse consequences on outcomes including biomarkers (eg, blood lipids), behavior, dental caries, obesity, coronary heart disease, and nutrient intakes. 3 The differentiation between total sugar and added sugar is important 4 ; total sugars include all sugars in the diet, including those in nutrient-dense foods, such as lactose in milk or fructose in fruits. Consequently, diets high in total sugar would not necessarily be expected to result in lower diet quality. Added sugar is a component in foods that might not be related to the nutrient density of any food, for example, pastries containing low levels of micronutrients could be just as high in added sugar as a fortified cereal, which might contain not only high sugar but also have high micronutrient density. Thus, recommendations on added sugar intake in the American population should be based solely on research distinguishing those two kinds of sugar in the diet. The new DRI are of great public health importance. They are used by health professionals in providing dietary guidance or serve as guides in planning and delivering federal nutrition programs. Assessment of dietary adequacy is often based on DRI cut-points. AI CSFII DRI EAR Adequate intake Continuing Survey of Food Intakes by Individuals Dietary reference intakes Estimated average requirement NHANES USDA National Health and Nutrition Examination Survey United States Department of Agriculture From the Department of Nutritional Sciences, College of Health and Human Development, Pennsylvania State University, State College, Pennsylvania; and the Departments of Nutrition and Maternal and Child Health, School of Public Health, University of North Carolina, Chapel Hill. Supported by a Seed Grant by the College of Health and Human Development, Pennsylvania State University. Submitted for publication Jan 29, 2004; last revision received Aug 5, 2004; accepted Aug 30, Reprints are not available from the authors. Correspondence to: Dr Sibylle Kranz, Department of Nutritional Sciences, College of Health and Human Development, Pennsylvania State University, Henderson Building 5G, State College, PA sxk72@ psu.edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

154 Table I. Selected sociodemographic characteristics of the sample from the CSFII 1994 to 1996 and 1998 (in percentage of total population in each age group) Characteristics CSFII 1994 to 1996 and y (n = 2805) 4 1 5y (n = 2632) Sex Boys Ethnicity Non-Hispanic white Non-Hispanic black Non-Hispanic other Hispanic Income,130% of poverty 130% 184% of poverty 185% 350% of poverty.350% of poverty Education * Less than high school High school More than high school Employment status * Employed *Of female head of household. Policy-makers use them to design dietary guidelines, such as the Food Guide Pyramid. Thus, the recommendations are a powerful tool and have great potential to aid in the effort to limit nutrition related health problems. The purpose of this study was to investigate the effect of added sugar intake on diet quality in preschool children by using dietary data available subsequent to NHANES (National Health and Nutrition Examination Survey) III (1988 to 1992), which was used by the DRI macronutrients committee. Given the relation between chronic diseases and demographic variables, 5-7 we investigated intakes of added sugar by children from various sociodemographic groups. Finally, we examined associations between intake of added sugars and food group as well as nutrients. This approach recognizes that diets are complex and that food group analysis reflects interactions of nutrients and other dietary components. METHODS Estimates of dietary intakes were based on data for children 2 through 5 years of age who provided 2 days of dietary intake information in the Continuing Survey of Food Intakes by Individuals (CSFII), 1994 to 1996 and 1998 (n = 5437). The 1994 to 1996 USDA CSFII survey data were collected through the use of a stratified, multistage, area probability sample to obtain a nationally representative sample of noninstitutionalized persons living in households. Interviewers followed preset protocols to complete the survey. Detailed description of the study design can be found in the study by Tippet and Cypel. 8 Dietary intake data were collected through the use of a standardized multiple pass approach. Day 1 of dietary intake was collected by means of a household interview, whereas day 2, which was collected 3 to 10 days after day 1 but not on the same day of the week, could be reported during a phone interview. Interview respondents reported 24-hour recalls for children younger than age 6. In 1998, the same methods were used for a supplemental survey in children. All survey waves (CSFII, 1994 to 1996 and 1998) were combined and used in this study. The data set provides sociodemographic information as well as number of servings from the Food Guide Pyramid food groups and teaspoons of added sugar for all individuals who reported dietary intake (n = 21,662). Sociodemographic information, such as age, race, and sex of the children as well as years of education and employment status of the female head of household, was ascertained during the survey. As described by Welsh et al, 1 an added sugar category was developed for the CSFII Pyramid Servings data set to capture all caloric carbohydrate sweeteners (monosaccharides, disaccharides, and higher saccharides). Included were all sugars and sugar-containing ingredients added during processing or preparation, and sugars eaten separately (candy) or added at the table (syrups, white or brown sugar, and so forth) but excluding all naturally occurring sugars (such as fructose in fruits). 9 Descriptive analysis, means and standard errors, were ascertained to describe the sample. Main sources of added sugar consumption were acquired by using the food grouping system developed by a group of researchers at the University of North Carolina at Chapel Hill (UNC food grouping system). 10 We calculated average daily consumption contributed from each UNC food group and retained a list of the highest 10. We then ranked this list in descending order. Analysis was stratified by 5 added sugar intake levels: (1) meeting the recommendation of #10% of energy from added sugar (25% of the sample), (2) between 11% and 15% of energy (26%), (3) 16% to 20% (24%), (4) 21% to 25% (14%), and (5) in excess of the new recommendation (>25% of energy) (12%). Intakes were analyzed by income and ethnicity groups. Mean and standard error of average nutrient and food group consumption as well as nutrient density (nutrient consumption per 1000 kcal average energy intake) were calculated. The proportions of children not meeting the Adequate intake (AI) for calcium or Estimated Average Requirement (EAR) for other nutrients were ascertained. A nonparametric test for trend across ordered groups was conducted to obtain a z-score and P value to assess significance of the changes of dietary intakes with increasing levels of added sugar consumption. 11 The statistical package used for analysis was STATA (version 7.0), which allows adjustment for sample design effect 106 Kranz et al The Journal of Pediatrics January 2005

155 Table II. Main food sources * of added sugar consumption in preschoolers using the University of North Carolina food grouping system y for 2- to 3-year-olds and 4- to 5-year-olds (in percentage of total added sugar consumption within each age group) UNC Category Example foods 2- to 3-year-olds 4- to 5-year-olds % z % z Fruit drinks Lemonades, juices with 10% fruit juice High-fat desserts Ice cream, pies, cookies, cakes Regular soft drinks Cola type Sugars and jellies Candy Low-fiber, ready-to-eat cereal Cornflakes, Special K Low-fat desserts Gelatin desserts Low-fat milk 2% milk, skim milk Low-fiber other fruits Fruit salad/cocktail, applesauce Medium-fat milk Whole milk, milkshakes Total *Main food sources of added sugar are similar except for the candy and low-fiber cereal categories, which are reversed for the 4- to 5-year-olds. yunc food groups, Grouping of foods into 75 categories with similar nutrition profiles. 10 zpercentage of total added sugar consumption of children in this age group. Includes flavored milk and items prepared with low-fat milk powder. Table III. Proportion of preschoolers at different levels of added sugar intake by income and ethnicity for (a) 2- to 3-year-olds and (b) 4- to 5-year-olds (in percentage of individuals within each income or ethnicity group) Sugar consumption (% of total energy) 0% 10% 11% 15% 16% 20% 21% 25% >25% (a) Sample size n = 781 (28%) n = 743 (26%) n = 618 (22%) N = 367 (13%) n = 296 (11%) (percentage of total sample) Income level (% poverty), Ethnicity Non-Hispanic white Non-Hispanic black Hispanic Non-Hispanic other (b) n = 547 (21%) n = 682 (26%) n = 662 (25%) N = 412 (16%) N = 329 (12%) Income level (% poverty), Ethnicity Non-Hispanic white Non-Hispanic black Hispanic Non-Hispanic other Adverse Effect Of High Added Sugar Consumption On Dietary Intake In American Preschoolers 107

156 Table IV. Trends of energy, nutrient, and food group intakes (mean ± SEM) by added sugar consumption levels for 2- to 3-year-olds (a) and 4- to 5-year-olds (b) Sugar consumption (% total energy) 0% 10% 11% 15% 16% 20% 21% 25% >25% (a) (n = 781) (n = 743) (n = 618) (n = 367) (n = 296) Test for trend Mean ± SEM z score P value Macronutrients Energy (kcal) ± ± ± ± ± Carbohydrate * 52.4 ± ± ± ± ± ,.001 Protein * 15.8 ± ± ± ± ± ,.001 Fat * 33.1 ± ± ± ± ± ,.001 Fiber (g) 10.5 ± ± ± ± ± ,.001 Micronutrients Calcium (mg) ± ± ± ± ± ,.001 Iron (mg) 11.9 ± ± ± ± ± ,.001 Folate (mg) ± ± ± ± ± ,.001 Vitamin A (RAE) ± ± ± ± ± ,.001 Vitamin B 12 (mg) 3.4 ± ± ± ± ± ,.001 Vitamin C (mg) ± ± ± ± ± Food groups 2 Grains (6) 6.8 ± ± ± ± ± ,.07 Vegetable (3) 2.5 ± ± ± ± ± ,.07 Fruit (2) 3.7 ± ± ± ± ± ,.001 Dairy (2) 2.2 ± ± ± ± ± ,.001 (b) (n = 547) (n = 682) (n = 662) (n = 412) (n = 329) Macronutrients Energy (kcal) ± ± ± ± ± ,.71 Carbohydrate * 50.9 ± ± ± ± ± ,.001 Protein * 16.0 ± ± ± ± ± ,.001 Fat * 34.4 ± ± ± ± ± ,.001 Fiber (g) 12.7 ± ± ± ± ± ,.001 Micronutrients Calcium (mg) ± ± ± ± ± ,.001 Iron (mg) 13.6 ± ± ± ± ± ,.001 Folate (mg) ± ± ± ± ± ,.001 Vitamin A (RAE) ± ± ± ± ± ,.001 Vitamin B 12 (mg) 4.6 ± ± ± ± ± ,.001 Vitamin C (mg) ± ± ± ± ± Food groups y Grains (6) 5.8 ± ± ± ± ± ,.001 Vegetable (3) 2.1 ± ± ± ± ± ,.001 Fruit (2) 2.2 ± ± ± ± ± ,.001 Dairy (2) 2.3 ± ± ± ± ± ,.001 *In percentage of total energy. yin the Food Guide Pyramid for Young Children (2 to 6 years old) servings, in parenthesis are the recommended number of servings. and weighting to maintain the nationally representative character of the data. All analysis was conducted for two separate age groups of children (2- and 3-year-olds compared with 4- and 5-year-olds). RESULTS Half of the children in the sample were girls and approximately half was in each age group (Table I). Approximately one third were in the lowest income group and one quarter were in the highest income group. The majority of children was non-hispanic white and had an employed female head of household. Average added sugar intake was lower in 2- and 3- year-olds compared with 4- and 5-year-olds: 14.9% and 16.5% of total energy, respectively (equivalent to 13.5 ± 0.19 and 17.2 ± 0.24 teaspoons per day). In younger children, 108 Kranz et al The Journal of Pediatrics January 2005

157 average added sugar intake represented in teaspoons per day was 4.4 tsp in the 0% to 10% group, 11.9 tsp in the 11% to 15%, 16.5 tsp in the 16% to 20%, 21.3 tsp in the 21% to 25%, and 23.1 tsp in the highest added sugar consumption group. Consumption was higher in the older children: 5.0 tsp in the 0% to 10% group, 13.7 tsp in the 11% to 15%, 19.0 tsp in the 16% to 20%, 24.4 tsp in the 21% to 25%, and 26.4 tsp in the highest added sugar consumption group. Main sources of added sugar in preschoolers are provided in Table II. The rank order of the top-10 categories were the same in both age groups, with the exception of the candy and low-fiber cereal category. Half of the added sugar consumption was contributed by the top three items: fruit drinks, high-fat desserts, and regular soft drinks. Eleven percent and 12% of the younger and older children, respectively, consumed more than 25% of energy from added sugar. Consumption levels differed between income and ethnic groups. Proportionally fewer children at the lowest income level and those classified as non-hispanic other consumed the highest level of added sugar. More children in the younger age group than in the older age group had 10% or less of added sugar intake. Higher proportions of non-hispanic white and black children consumed more than 15% of energy from added sugar (Table III). With the exception of carbohydrates and vitamin C, average consumption of macronutrients and micronutrients significantly decreased with increasing added sugar level (Table IV). In addition to the nutrients listed in Table IV, this trend was also observed for saturated fat, zinc, magnesium, vitamin B 6, and sodium (data not shown). On average, children consuming above 25% of energy from added sugar consumed a significantly lower proportion of energy from protein, fat, and less fiber but increased percentage of energy from carbohydrates than children at lower levels of added sugar intake. Absolute intakes of micronutrients were also lower in this group, with the exception of vitamin C. Investigation of nutrient density (intake per 1000 kcal) indicated that micronutrient density significantly decreased with increasing level of added sugar intake, whereas carbohydrate density significantly increased (data not shown). The proportion of children with intakes below the AI for calcium and EAR for other nutrients increased for most nutrients with increasing level of added sugar intake (data not shown). The data were most striking for calcium. Forty percent of the 2- and 3-year-olds and approximately 70% of 4- and 5-year-olds in the highest added sugar intake group fell below the AI for calcium. Almost one quarter of the younger children and one-half of the older children were below the AI at added sugar intakes of 16% to 25%. Even at the lowest added sugar intake level, a large proportion of children, especially 4- and 5-year-olds did not meet the AI for calcium. For most nutrients, the proportion of children consuming less than the EAR was highest at sugar intakes of more than 25% of energy, but for several nutrients the percentage of older children not reaching the EAR was close to 5% for vitamins A and C and 13% for folate at an added sugar intake of 11% to 15% of total energy. Intake of servings of Food Guide Pyramid food groups changed as well. Grains, vegetables, fruits, and dairy decreased with increasing added sugar levels below the recommended intakes in the older children. The decline was steady across all added sugar categories. Especially noteworthy was the reduction in intake of fruit and dairy foods in both age groups. The number of servings of meat did not change significantly between groups (data not shown). DISCUSSION This study has shown a significant trend toward decreasing intakes of nutrients and food groups by preschool children with increasing added sugar levels. Children with the highest level of added sugar intake (25% of energy or more) had the lowest consumption of most nutrients and servings of grains, vegetables, fruits, and dairy. One exception to the decline in nutrient intake was carbohydrates, which increased with increasing added sugar intake, whereas fat consumption decreased. This reciprocal relation between sugar and fat consumption has been described previously and is thought to be one reason for the lack of consistent evidence for a parallel relation between increasing added sugar consumption and higher body weight. The dilution of nutrients with increasing added sugar intake levels as described in this study has been shown by others as well Thus, the new recommendation for added sugar intake might be a reason for concern. Decreasing micronutrient intakes with higher added sugar consumption probably are observed because the main food source of added sugar intake are foods with low micronutrient profiles, 14,19 such as sweetened soft drinks and juices. Hence, limitation of added sugar intake could result in higher nutrient dense diets. This was also demonstrated by a study limiting the amount of added sugar in the diet of federal School Breakfast Program participants to <10% of total energy. 20 Nutrient density in that study increased when added sugar intake was decreased. Calcium, an important nutrient in growing children, had been inadequate in a large proportion of children before the study. The number of servings of dairy had been below the recommended amounts; thus, the inadequate calcium levels had been paralleled by the low consumption of dairy, probably because sweetened beverages replaced milk and dairy products as beverages After reflecting on data from other studies, some researchers concluded that the change in nutrient and food group consumption, although statistically significant, was so small as to bear no clinical importance. 18 However, we believe that even a reduction of one-half serving of a food group intake in children with high added sugar levels measured in a nationally representative data set is an alarming result. In this study, we found that children with the lowest added sugar level consumed approximately one serving more of grains, fruits, and dairy compared with children in the highest added sugar intake group. This is especially alarming when the intake reduction occurs in food groups that are currently not consumed at the recommended intake levels, such as dairy, fruits, and vegetables. Adverse Effect Of High Added Sugar Consumption On Dietary Intake In American Preschoolers 109

158 A further point to consider in this discussion is the effect of food fortification. Many foods in the food supply that are high in added sugar are also fortified, such as sweetened cereals and breads. As our results indicated, the effect of added sugar intake level on nutrient consumption might be modified by the fortification of foods. Consumption of most micronutrients decreases to a much weaker extent, as could be expected with increasing added sugar intake. Although increased added sugar consumption leads to decreased nutrient density, the magnitude of this effect is lowered after controlling for fortification in the food supply. 17 Hence, fortification masks the potentially adverse effects of the high added sugar content in the diet. Dietary intake behaviors established in childhood track over time, thus establishing that a diet high in added sugar during childhood might affect the development of chronic diseases later in life. 24 Also, added sugar consumption is predicted by a number of sociodemographic variables, for example, that income and social status affect added sugar intakes 25 and that non-hispanic children consume higher levels of added sugar than Hispanic children. 26 Consequently, the recommendation for added sugar intake for the overall American population might not be appropriate for some subgroups of the population. The growth and development of children with a high likelihood of consuming increased levels of added sugar might be negatively affected by the DRI. Rather than being encouraged by health professionals and caretakers to decrease added sugar consumption, they will maintain a diet that puts them at high risk for calcium inadequacy and low intakes of fruits and vegetables. One limitation to our study was that children s diets were reported by proxy and were therefore prone to incomplete or biased information. This limitation in the data collection in preschool children has been acknowledged by the research community and will have to be tolerated until accurate measurement methods have been developed for freeliving children. 27 However, we used a large, nationally representative sample of American preschoolers to investigate the relation between added sugar intake levels and nutrient consumption. We included children with 2 days of dietary intake and are confident that the resulting mean intakes are good estimates of the average consumption in American preschoolers. NHANES data provides dietary intake information, biomarkers, and measured height and weight; however, this data set has been used in the development of the DRI. Consequently, we decided to use the other large nationally representative data set, the CSFII, which provided us with more updated information than NHANES III. Debates about sugars and health are longstanding. The Joint World Health Organization/Food and Agriculture Organization Expert Consultation on Diet, Nutrition, and Prevention of Chronic Disease recommended that intake of free sugars be no more than 10% of total energy consumed. 2 In the USDA s Food Guide Pyramid, moderate added sugar consumption was defined as 6% to 10% of total energy. 1 These ongoing debates argue for continued evaluations of the effect of sugars on dietary intakes, metabolism, and chronic diseases. In addition, the point has been raised that most added sugars are not visible in foods and are not listed on the food label, making it difficult for the consumer to estimate consumption. 28 Our results indicate reasons for concern. The DRI for added sugar intake might not be appropriate for American preschoolers. In contrast to others, 29 we found that although the most dramatic decreases in micronutrient intakes were observed at added sugar intake levels of more than 25% of total energy, food group consumption and the proportion of children meeting the AI for calcium was also very low at added sugar intakes of less than 10% of calories. More research is needed to investigate the appropriateness of the DRI with respect to overall diet quality as well as measurable health outcomes, such as body weight. Nationally representative data collecting information on added sugar consumption as well as biomarkers and body weight status will be helpful in determining a relation. Also, large, longitudinal studies examining the long-term effect of high added sugar diets in young children might help elucidate the causality of increased body weight and risk factors for chronic diseases and diet patterns. However, until more data are available, the DRI for added sugar might adversely affect young children in the long run. The consequences of high added sugar diets must be considered with reference to sociodemographic factors and the relations to overall diet quality and health status. REFERENCES 1. Welsh S, Davis C, Shaw A. USDA Food Guide: Background and Development. Hyattsville, Md: United States Department of Agriculture; Report No World Health Organization/Food and Agriculture Organization of the United Nations Expert Consultation. Diet, nutrition, and the prevention of chronic diseases. Technical Support Series 916. Geneva, Switzerland: World Health Organization; Institute of Medicine of the National Academy of Sciences. Dietary Reference Intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and amino acids (macronutrients). Report. Washington, DC: National Academy Press; Sigman-Grant M, Morita J. Defining and interpreting intakes of sugars. Am J Clin Nutr 2003;78:815S-26S. 5. Ernst ND, Obarzanek E. Child health and nutrition: obesity and high blood cholesterol. Prev Med 1994;23: Ogden CL, Flegal KM, Carroll MD, Johnson CL. Prevalence and trends in overweight among US children and adolescents, JAMA 2002;288: Must A, Strauss RS. Risks and consequences of childhood and adolescent obesity. Int J Obes Relat Metab Disord 1999;23:S Tippet KS, Cypel YS. Design and Operation: The Continuing Survey of Food Intake by Individuals and the Diet and Health Knowledge Survey Nationwide Food Surveys Report 96-1: United States Department of Agriculture, Agricultural Research Service; Report No United States Department of Agriculture Agricultural Research Service. Design and Operation: The Continuing Survey of Food Intakes by Individuals and the Diet and Health Knowledge Survey, and 1998; Report No. NSF Report No Popkin BM, Haines PS, Siega-Riz AM. Dietary patterns and trends in the United States: the UNC-CH approach. Appetite 1999;32: Cuzick J. A Wilcoxon-type test for trend. Stat Med 1985;4: Gibson SA. Associations between energy density and macronutrient composition in the diets of pre-school children: sugars vs starch. Int J Obes Relat Metab Disord 2000;24: Farris RP, Nicklas TA, Myers L, Berenson GS. Nutrient intake and food group consumption of 10-year-olds by sugar intake level: the Bogalusa Heart Study. J Am Coll Nutr 1998;17: Kranz et al The Journal of Pediatrics January 2005

159 14. Gibney M, Sigman-Grant M, Stanton JL, Keast DR. Consumption of sugars. Am J Clin Nutr 1995;62:178S-93S. 15. Lewis CJ, Youngmee PK, Behlen Dexter P, Yetley EA. Nutrient intakes and body weights of persons consuming high and moderate levels of added sugars. J Am Diet Assoc 1992;92: Murphy SP, Johnson RK. The scientific basis of recent US guidance on sugars intake. Am J Clin Nutr 2003;78:827S-33S. 17. Alexy U, Sichert-Hellert W, Kersting M. Fortification masks nutrient dilution due to added sugars in the diet of children and adolescents. J Nutr 2002;132: Forshee RA, Storey ML. The role of added sugars in the diet quality of children and adolescents. J Am Coll Nutr 2001;20: Guthrie JF, Morton JF. Food sources of added sweeteners in the diets of Americans. J Am Diet Assoc 2000;100: Worobey HS, Worobey J. Efficacy of a preschool breakfast program in reducing refined sugar intake. Int J Food Sci Nutr 1999;50: Fisher OJ, Mitchell DC, Smiciklas-Wright H, Birch LL. Maternal milk consumption predicts the tradeoff between milk and soft drinks in young girls diets. J Nutr 2000;131: Harnack L, Stang J, Story M. Soft drink consumption among US children and adolescents: nutritional consequences. J Am Diet Assoc 1999;99: Guenther PM. Beverages in the diets of American teenagers. J Am Diet Assoc 1986;86: Zive MM, Berry CC, Sallis JF, Grank GC, Nader PR. Tracking dietary intake in white and Mexican-American children from age 4 to 12. J Am Diet Assoc 2002;102: Brady LM, Lindquist CH, Herd SL, Goran MI. Comparison of children s dietary intake patterns with US dietary guidelines. Br J Nutr 2000; 84: Kranz S, Siega-Riz AM. Sociodemographic determinants of added sugar intake in preschoolers 2 to 5 years old. J Pediatr 2002;140: Serdula MK, Alexander MP, Scanlon KS, Bowman BA. What are preschool children eating? A review of dietary assessment. Annu Rev Nutr 2001;21: American Dietetic Association. Position paper: dietary guidance for healthy children ages 2 to 11. J Am Diet Assoc 2004;104: Lineback DR, Jones JM. Sugars and health workshop: summary and conclusions. Am J Clin Nutr 2003;78:893S-7S. 50 Years Ago in The Journal of Pediatrics SOME PECULIARITIES OF AMINO ACID METABOLISM IN INFANCY AND EARLY CHILDHOOD Schreier K. J Pediatr 1955:46; In January 1955, Kurt Schreier of Heidelberg, Germany, and Rochester, New York, provided a comprehensive review of normal amino acid metabolism and the relation of amino acids with growth rate during gestation and early life. He also proposed a classification of inborn errors of protein metabolism. Many of his observations remain true, whereas others have been revised or discarded. At the time, much of the renewed interest in amino acid metabolism was sparked by new technologies such as column chromatography and stable isotopes, which allowed investigators to examine normal and aberrant amino acid metabolic pathways. Still valid are his observations that the placenta and fetal cells actively concentrate amino acid from the maternal circulation and that the amino nitrogen concentration is highest in the youngest and smallest fetus. In his review of the relations of anabolic hormones with growth, he noted the work of several investigators that testosterone benefits the growth and general condition of premature infants has not been confirmed. What is the chance that those trials would be approved today? Regarding the protein-sparing action of carbohydrates and fats, he wrote For this reason, milk is a wonderful food for the growing organism, not only because it supplies proteins, but also carbohydrate and fat, simultaneously to the tissues. Mother Nature still knows best! Although most of his observations on normal amino acid metabolism still hold true, his classification of inborn errors of metabolism has undergone significant evolution. Disorders of impaired kidney function would today include the aminoacidurias secondary to renal tubular transport disorders. His second group included generalized metabolic disturbances, citing cystinosis as an example. The third group was inborn errors in which a portion or an entire protein was missing such as hemophilia, Wilson hepatolenticular degeneration, and hypophophatasia. Fourth were the idiopathic hypoproteinemias such as agammaglobulinemia or hypoalbuminemia. His final group included disorders of phenylalanine and tyrosine metabolism such as PKU, albinism, alkaptonuria, and tyrosinosis. He concluded by agreeing with basic researchers that often fatal pathological changes during adult life may have their origin in a very short period of dietary abnormality in infancy. Although this statement might still ignite vigorous discussion, clearly understanding the role of metabolic pathways in growth and development is as critical today as it was 50 years ago. Paul M. Fernhoff, MD, FAAP, FACMG Division of Medical Genetics Department of Human Genetics Emory University School of Medicine Atlanta, GA YMPD /j.jpeds Adverse Effect Of High Added Sugar Consumption On Dietary Intake In American Preschoolers 111

160 TREATMENT OF INTERICTAL EPILEPTIFORM DISCHARGES CAN IMPROVE BEHAVIOR IN CHILDREN WITH BEHAVIORAL PROBLEMS AND EPILEPSY RONIT M. PRESSLER, MD, MRCPCH, RICHARD O. ROBINSON, FRCPCH, GOIGIA A. WILSON, AND COLIN D. BINNIE, MD, FRCP Objectives It is generally agreed that children should be treated for epilepsy only if they have clinical seizures. The aim of this study was to examine whether suppressing interictal discharges can affect behavior in children with epilepsy. Study design In a double-blinded, placebo-controlled, crossover study, 61 children with well-controlled or mild epilepsy were randomly assigned to add-on therapy with either lamotrigine followed by placebo or placebo followed by lamotrigine. Ambulatory electroencephalographic recordings and behavioral scales were performed during baseline and at the end of placebo and drug phases. The primary hypothesis to be tested was that behavioral scales would improve specifically in patients with a reduction of electroencephalographic discharges during active drug treatment. Results Global rating of behavior significantly improved only in patients who showed a significant reduction in either frequency (P <.05) or duration of discharges (P <.05) during active treatment but not in patients with without a significant change in discharge rate. This improvement was mainly seen in patients with partial epilepsy (P <.005). Conclusions Our data suggest that suppressing interictal discharges can improve behavior in children with epilepsy and behavioral problems, particularly partial epilepsy. Focal discharges may be involved in the underlying mechanisms of behavioral problems in epilepsy. (J Pediatr 2005;146:112-7) Children with epilepsy are at a higher risk for development of behavioral problems and psychiatric disorders than are their healthy peers 1 or children with other chronic disease. 2 This is not only important for children with uncontrolled epilepsy and learning difficulties but also for the majority of children with epilepsy whose seizures respond well to antiepileptic drugs (AEDs) and who are educated in mainstream schools. Such children have been found to have more learning and behavioral problems in school compared with matched control children and achieve less than expected for their age and IQ. 3,4 Increased behavioral problems in children with epilepsy are a consequence of a number of interacting influences including underlying brain lesion, age of onset, AEDs, psychosocial issues, seizure type and frequency, and interictal abnormalities in the electroencephalogram. 5,6 Interictal discharges or subclinical epileptiform discharges occur in up to 80% of patients with ongoing epilepsy, although they may not be seen in every EEG recording. 7 The clinical relevance of these discharges is unclear; specifically, it is uncertain whether they are truly subclinical or cause brief disruptions of cognitive function (as described by Aarts et al 8 as transitory cognitive impairment) and behavior. 9 The only way to determine whether discharges cause cognitive and behavioral problems in children with epilepsy is by finding whether cognition and behavior improve when epileptiform discharges are suppressed. It is generally agreed by neurologists and pediatricians that patients should be treated for epilepsy only if they have clinical seizures. Treating the EEG, so-called EEG cosmetics, is generally condemned. We aimed to test this view by performing a double-blinded, placebo-controlled, crossover trial to examine whether suppressing interictal discharges can improve behavior AED Antiepileptic drug MANOVA Multivariate analysis of variance NCYPE National Centre for Young People with Epilepsy TCI EEG Transitory cognitive impairment Electroencephalography From the Department of Clinical Neurophysiology, King s College Hospital, King s College London, London, United Kingdom; and the Department of Paediatric Neurology, Guy s Hospital, King s College London, London, United Kingdom. R. Pressler and G. Wilson were supported by the Fund for Epilepsy and by Glaxo Wellcome. Lamotrigine (Lamictal) and placebo tablets were provided by Glaxo Wellcome, Stevenage, United Kingdom (now Glaxo- SmithKline). Submitted for publication Nov 2, 2003; last revision received Jul 13, 2004; accepted Aug 31, Reprint requests: Dr Ronit Pressler, Department of Clinical Neurophysiology, National Hospital for Neurology and Neurosurgery, Queen s Square, WC1N 3BG, United Kingdom. ronit.pressler@uclh.org /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

161 in children with epilepsy and by implication whether interictal discharges can cause psychosocial dysfunction. To avoid the confounding factor of changing seizure frequency on behavior, we included only patients who were seizure free or who had infrequent seizures. It was essential to exclude an independent psychotropic effect of the drug on behavior and thus we included both patients with and without interictal discharges. Consequently, it was possible to compare behavioral changes in patients with and without a reduction of interictal discharges. METHODS Patients were recruited from pediatric outpatient clinics at three study sites: Guy s Hospital, King s College Hospital, and The National Centre for Young People with Epilepsy (NCYPE), United Kingdom. Additional patients were referred to the study sites from hospitals in the South Thames Region. Patients 7 to 17 years of age were eligible if they had a diagnosis of epilepsy and were seizure free or were having occasional seizures but in whom the responsible clinician or parent believed that further adjustments to AEDs was not warranted. The inclusion criteria for occasional seizures took account of seizure severity and were defined as no more than 1 generalized tonic-clonic seizure in the last 6 months or no more than 1 complex partial seizure or 2 simple partial seizures in the last month or no more than 5 absences occurring on any 1 day within the last 3 months. Other inclusion criteria were an IQ of $70 or mental age of at least 7 years. In keeping with the requirements of the local ethics committee, some evidence of cognitive impairment or psychosocial dysfunction was required to provide ethical justification for participation, specifically, that parents were sufficiently concerned about behavior/cognition to seek help. The protocol was approved by the ethics committee at all three study sites, and written informed consent was obtained from all parents and patients 16 years of age or older and oral agreement from patients younger than 16 years of age. Patients were randomly assigned to receive lamotrigine (Lamictal, GlaxoWellcome, now GlaxoSmithKline, Stevenage, United Kingdom) followed by placebo, or placebo followed by lamotrigine, in addition to the current AED regime each for 13 weeks (Figure). The dose of lamotrigine depended on age, weight, and concomitant AEDs according to the recommendations current at the time the study was conducted. For children taking sodium valproate, lamotrigine was titrated up to 2 mg/kg per day (#12 years of age) or 150 mg/d (>12 years of age). For children not taking sodium valproate, lamotrigine was titrated up to 10 mg/kg per day (#12 years of age) or 300 mg/d (>12 years of age). During a 4-week, single-blinded phase, all subjects received placebo to familiarize patients and parents with trial procedures and to provide a reference point if the subsequent phases showed an order effect. At entry, physical and neurologic examination, history, routine and ambulatory EEG, standard biochemical tests, AED concentrations and IQ tests (Wechsler Intelligence Scale for Children, WISC-III) were performed and behavioral scales were completed by parents and teachers. Patients were assessed at the end of each treatment phase (weeks 17 and 31) when the following were recorded: physical examination, lamotrigine blood levels, ambulatory EEG, neuropsychologic tests, behavioral scales for parents and teachers, and documentation of compliance, seizures, and possible adverse events. Seizures were classified according to the criteria of the International League against Epilepsy. Ambulatory monitoring was performed for a 12- to 24- hour period with the use of the Oxford Medilog 8-channel cassette system or the digital Embla recording system. EEG recordings were analyzed visually, and epileptiform discharges were defined as spikes, sharp waves, spike wave complexes, or multiple spike discharges. A continuous run of epileptiform waveforms would be considered as one discharge if not interrupted by normal activity of more than 1 second. Discharges were considered subclinical when the available methods of clinical observation, applied under particular circumstances, failed to show any changes in the patient. 8 Discharges were quantified in each patient during the eyesopen phase of a 12- to 24-hour period as frequency of discharges (number per hour) and discharge time (duration in seconds per hour). The minimum duration allocated to any single discharge was 1 second. We assessed behavior with the Conners Rating Scales for parents and teachers. The Conners Rating Scales are factor analytically derived scales for assessing problem behavior in children. The parents rating scale consists of a list of 93 questions and the teachers rating scale of 39 questions. Raw scores are translated into t scores by sex and age. The t scores have a mean of 50 and a standard deviation of 10, and higher scores denote more serious behavior problems. 10 The parents rating scale has 8 subscales designated as (I) antisocial, (II) anxious-shy, (III) conduct disorder, (IV) hyperactive-immature, (V) learning problem, (VI) obsessivecompulsive, (VII) psychosomatic, and (VIII) restless-disorganized. The teachers rating scale has 6 subscales labeled as (I) anxious-passive, (II) asocial, (III) conduct problem, (IV) daydream-attention, (V) emotional-indulgence, and (VI) hyperactivity. Rating forms were completed by the same persons on all occasions. The primary hypothesis tested was that behavioral scales would improve specifically in those patients who had a reduction of discharges during the active drug phase. Changes in global rating of behavior were analyzed by repeated-measurement multivariate analysis of variance (MANOVA) between treatment groups (lamotrigine and placebo), with response to lamotrigine as covariant (with or without reduction of discharges). To identify the most relevant behavioral subscale, the univariate test was used. A P value of <.05 was considered significant. All statistical tests were 2-tailed. Analysis was by intention to treat. RESULTS Of the 64 patients screened, 61 were enrolled in the study and randomly assigned to receive first lamotrigine and Treatment Of Interictal Epileptiform Discharges Can Improve Behavior In Children With Behavioral Problems And Epilepsy 113

162 Figure. Trial profile. A/E, adverse events; W/D, withdrawn; P/V, protocol violation. then placebo or vice versa (Figure). Thirteen children were withdrawn from the study, including 2 children who did not enter the double-blinded treatment phase. Eight of these children had discharges. Patients in both groups had similar demographics and baseline characteristics (Table). Two thirds of patents had interictal discharges at baseline (n = 42), with a mean discharge rate of 9.7 per hour (range, 0 to 115.6/h) and mean duration of 15.4 s/h (range, 0 to 140 s/h). Patients with idiopathic partial epilepsy were more likely to have discharges (15 of 16) than patients with idiopathic generalized epilepsy (13 of 19) or with symptomatic partial epilepsy (14 of 26) (x 2 = 7.36; df =2;P <.05). Most children (77%) were seizure free during the 3 months preceding baseline. Eight children had up to 2 partial seizures per month, 2 between 3 and 5 absences per month, and 4 had more than 6 absences per month but no more than 3 absences on any 1 day. Children with seizures were more likely to have interictal discharges (12 of 14) than children without seizures (30 of 47), but this difference was not significant (x 2 = 2.40; df = 1; not significant [NS]). The seizure frequency did not change significantly during the study. Forty-seven patients were seizure free at baseline (47 of 61, 77%), 40 during the placebo phase (78%), and 38 during the lamotrigine phase (81%). In the 3 months preceding baseline, the mean seizure frequency was 3.43 (SD, 13.4) seizures per month; during the placebo phase, 3.24 (SD, 114 Pressler et al The Journal of Pediatrics January 2005

163 10.4) seizures per month; and during the lamotrigine phase, 3.28 (SD, 14.9) seizures per month. Twenty-one (44%) patients had a reduced frequency of discharges, whereas 16 (33%) patients either had no change or an increase in the frequency of discharges and 11 (23%) patients had no discharges during either lamotrigine or placebo phase. Twenty-three (48%) patients had a reduced duration of discharges, whereas 14 (29%) patients either had no change or an increase of discharge duration and 11 (23%) patients had no discharges in either lamotrigine or placebo phase. The effect of lamotrigine on discharges was similar across the types of epilepsies. The mean behavioral scores at baseline as assessed by both parents and teachers for the whole group were all within the normal range (50 ± 1 SD). Considering scores of >2 SD above the mean as abnormal in individual patients, 13 (22%) had at least 1 abnormal score in the parental assessment and 20 patients (33%) had at least 1 abnormal score in the teachers assessment. There was no difference in global rating of behavior (combining parents and teachers scale) when comparing placebo and lamotrigine for the total group of patients (MANOVA: F = 0.79; df = 14; NS). We found a significant improvement in global rating of behavior in the children who showed a reduction of discharges during the lamotrigine phase (MANOVA: frequency of discharges: F = 2.17; df = 14; P <.05; duration of discharges: F = 2.50; df =14;P <.05). This improvement was seen across all parental subscales and in 4 of 6 subscales in the teachers scale. It was significant for the parental conduct disorder subscale (P <.05) and parental psychosomatic subscale (P <.05). None of the teachers subscales individually showed a significant difference in the univariate test. This effect depended largely on whether patents had a partial or generalized epilepsy (MANOVA: F = 3.53; df = 14; P <.005). Patients with partial epilepsy were more likely to show an improvement of behavior when discharges were suppressed, whereas a change of behavioral rating in patients with generalized epilepsy was independent of the effect on discharges. This effect was the same for idiopathic partial epilepsy and symptomatic partial epilepsy. A similar difference was seen, depending on which drug the patient was receiving (MANOVA: F = 2.78; df = 28; P <.001). Patients with carbamazepine were more likely to have an improvement of behavior if discharges were suppressed than patients with sodium valproate or other drugs. However, patients with partial epilepsy showed a similar effect whether receiving carbamazepine or receiving other drugs. There was no order effect caused by random assignment (MANOVA: F=1.01; df = 14; NS) nor presence of seizures (MANOVA: F = 0.26; df = 14; NS). A subanalysis of children without seizures at baseline (n = 35) showed the same trends in the behavioral scales and similar results in the general linear modeling: a significant change in the global rating of behavior during active treatment for patients with a reduction in frequency of discharges (MANOVA: frequency of discharges: F = 4.16; df = 14; P <.05) and duration of discharges (F = 4.82; df = 14; P <.01). Table. Patient demographics and baseline characteristics Characteristic LTG/placebo (n = 31) Placebo/LTG (n = 30) Age (y) 11.4 (3.1) 11.7 (3.0) Sex (male) 20 (65%) 19 (63%) Age of onset (y) 5.6 (3.6) 7.8 (3.7) IQ (WISC-R score) 93.1 (17.2) 98.6 (22.1) Type of epilepsy Idiopathic partial epilepsy 10 (16%) 6 (10%) Idiopathic generalized epilepsy 8 (13%) 11 (18%) Symptomatic partial epilepsy 13 (21%) 13 (21%) Concomitant antiepileptic drugs None 3 4 Carbamazepine Sodium valp Other, including combinations * 6 4 Seizures/mo y No seizures 22 (71%) 25 (83%) $1 Seizure 9 (29%) 5 (17%) Discharges in initial EEG 21 (68%) 21 (70%) Frequency 10.7 (21.3) 8.6 (21.9) Duration (s) 16.4 (29.8) 14.4 (30.3) LTG, lamotrigine; WISC, Wechsler Intelligence Scale for Children. Values are shown as mean (SD) unless otherwise indicated. *Other AEDs included ethosuximide, phenytoin, and vigabatrin; combinations include carbamazepine and vigabatrin, carbamazepine and valproate, carbamazepine and acetazolamide, clobazam and ethosuximide, valproate and phenytoin. yin 3 months before study. Adverse events were evaluated for 59 patients after exclusion of two patients who were withdrawn in the singleblinded baseline phase. Apparent treatment-related adverse events were observed in 23 of 58 patients (40%) during the lamotrigine phase and in 19 of 52 patients (37%) in the placebo phase (x 2 = 3.24; df = 1; NS). Adverse events led to withdrawal from the study in 6 patients: in 5 as the result of a rash and in 1 as the result of dizziness and nausea. The latter was later found to have a high phenytoin level of 31.7 mg/ml. All other adverse events were mild and transient. DISCUSSION In this study, suppression of interictal discharges was associated with improved global rating of behavior in children with behavioral problems and epilepsy. This study presents such evidence in patients with epilepsy under controlled conditions. The controlled study design with standardized behavioral questionnaires and an appropriate number of patients avoided methodologic pitfalls. 5 Single observations and uncontrolled reports claim an improvement of cognitive functioning by suppressing discharges with AEDs in patients with epilepsy. 8,11 Our group has previously performed a preliminary study with the use of sodium valproate or clobazam add-on to suppress discharges in Treatment Of Interictal Epileptiform Discharges Can Improve Behavior In Children With Behavioral Problems And Epilepsy 115

164 10 children with uncontrolled epilepsy. A reduction of discharge rate was associated with improvement in global rating of psychosocial function in 8 of 10 children. 9 However, all but 1 patient showed an unexpected reduction in seizure frequency on active treatment, making the result difficult to interpret. To remove this confounding factor in the present trial, we included only patients who were either seizure free or were having few seizures. We found no significant differences between patients with and without seizures. In clinical practice, occasional seizures are not considered to represent any substantial seizure burden, and these patients are often regarded as well controlled. Generalized spike-and-slow-wave discharges may be accompanied by subtle clinical changes. 12 To exclude these children would involve ignoring a major group of subjects who might benefit from the treatment proposed. Even though all patients had a degree of behavioral or cognitive dysfunction reported by parents, this was not reflected by the baseline scores. Nevertheless, because of the design of the study, the conclusions necessarily apply only to children with some concern about learning or behavior. In clinical research, behavioral scales, such as the Conners Rating scale 4,9 or the Achenbach Child Behavior Checklist, 13 are often used to compare t score before and after intervention quantitatively rather than qualitatively defining what is a normal or abnormal score. 4,14,15 This may obtain a significant result that may or may not be clinically relevant, but this problem is inherent to all research using behavioral scales in children with mild problems. Our findings are in contrast to a recent study in 8 children with learning and behavioral problems whose behavior did not improve under active treatment with sodium valproate. 16 However, only 4 patients had a reduction of discharges, and, more importantly, no separate analysis was performed for patients with and without reduction of discharges. It was mainly patients with partial epilepsy who benefited from discharge suppression equally in patients with idiopathic and symptomatic partial epilepsy. In our study, lamotrigine had a similar effect on discharges in both partial and generalized epilepsy. Although more patients with idiopathic partial epilepsy had frequent discharges, many patients with symptomatic partial epilepsy had no discharges or a low discharge frequency, making a relation between baseline discharge frequency and magnitude of rating change unlikely. Patients taking carbamazepine were also more likely to show an improvement when discharges were suppressed compared with patients taking other drugs. This was due to the choice of first-line AEDs in partial and generalized epilepsy rather than due to an independent effect because carbamazepine is not associated with the improvement in partial epilepsy. It remains unclear whether the combination of lamotrigine and sodium valproate may cause more behavioral dysfunction than other combinations. It is well established that AED polytherapy itself is a risk for behavioral dysfunction in children with epilepsy. 17 By adding another drug into the current regime of our patients, it is possible that behavioral problems were accentuated in some. Nevertheless, we found a significant behavioral improvement in the lamotrigine group in the patients with a reduction of discharges. Lamotrigine is one of the few AEDs that suppresses discharges It does not appear to adversely affect cognition in epileptic patients. 19,20 In a recent study, low-dose lamotrigine had a positive effect on reaction time measurements and on 1 of 6 mood scales in healthy volunteers; however, the number of volunteers tested was small. 21 Furthermore, several uncontrolled studies reported improved cognition and behavior. 22,23 This improvement was apparently unrelated to seizure control. We could not confirm an overall effect of lamotrigine but rather an indirect effect through suppression of interictal discharges. The changes on the behavioral scale seen in our patients cannot be attributed to drug effects alone because they are confined to those subjects who showed a reduction of epileptiform activity on lamotrigine and were not seen in those who showed no reduction or had no discharges. How are interictal discharges and psychosocial disturbances related? Interictal discharges and behavioral problems could both be caused by an underlying pathology and thus could be coexisting but independent phenomena. However, in our crossover study, the patients acted as their own controls, and only those with a reduction in discharges showed an improvement of behavior. Interictal discharges may cause fragmented sleep, a well-recognized cause of cognitive and behavioral problems. 14,24 Moreover, in a recent study with lamotrigine, no improvement of nocturnal discharges nor neuropsychologic function could be found. 25 Finally, interictal discharges may cause psychosocial disturbances by directly interacting with cognitive and behavioral function. With the use of electroencephalography-linked cognitive tests, transitory cognitive impairment (TCI) has been found in 50% of patients investigated. 26 Generalized bursts lasting at least 3 seconds are most likely to produce demonstrable TCI, but they can also be found during briefer and focal discharges. 27 TCI may impair day-to-day psychosocial function. 6,9 It is well established that children with focal electroencephalographic abnormalities or complex partial seizures are particularly vulnerable to psychiatric and behavioral disturbances. 2,6,13,15 Our results provide evidence for the first time that focal discharges particularly may play a role in the underlying mechanisms of behavioral problems. These results have far-reaching implications for the treatment of children with epilepsy. Do we undertreat patients with epilepsy by aiming only to suppress clinically obvious seizures? The point at issue is not whether to treat the EEG but whether seizures, so subtle as to be recognizable only by TCI testing produce disability sufficient to justify treatment. This also questions the term subclinical epileptiform or interictal discharge : If discharges are causing cognitive changes such as TCI and behavioral changes, they are, strictly speaking, neither subclinical nor interictal. A further dilemma is that discharges seen in the electroencephalogram may or may not cause TCI or behavioral problems, depending on which area of the brain they occur. It has been shown that TCI is a specific 116 Pressler et al The Journal of Pediatrics January 2005

165 dysfunction of the brain area where the discharges occur. 27 Thus, discharges may cause a wider range of deficits than would be practical to test for in an individual patient. Our study suggests that at least in children with epilepsy and additional behavioral disturbances, treatment of interictal discharges, if present, with appropriate AEDs should be considered. We are very grateful for the support of the technical and medical staff at the Departments of Clinical Neurophysiology and Paediatrics at Guy s Hospital, King s College Hospital, and NCYPE. We thank Dr Clough, Dr Crouchman, Dr McCormack, and Prof Ross, of King s College Hospital; Dr Besag, of NCYPE; Dr El-Radhi, of Queen Mary s Hospital; Dr Evans, of Lewisham Hospital; Dr Hulse, of Maidstone Hospital; Dr Lin, Dr Pohl, and Dr Robb, of Guy s Hospital; Dr Robards, of Pembury Hospital; Dr de Silva, of Farnborough Hospital; and Dr Stern, of St Thomas Hospital, for patient referral. REFERENCES 1. Rutter M, Graham P, Yule W. A neuropsychiatric study in childhood. Clin Dev Med 1970;35/36: Hoare P. The development of psychiatric disorder among schoolchildren with epilepsy. Dev Med Child Neurol 1984;26: Seidenberg M, Beck N, Geisser M, Giordani B, Sackellares JC, Berent S, et al. Academic achievement of children with epilepsy. Epilepsia 1986;27: Aman MG, Werry JS, Turbott SH. Behavior of children with seizures: comparison with norms and effect of seizure type. J Nerv Ment Dis 1992;180: Kwan P, Brodie MJ. Neuropsychological effects of epilepsy and antiepileptic drugs. Lancet 2001;357: Besag FMC. Childhood epilepsy in relation to mental handicap and behavioural disorders. J Child Psychol Psychiatry 2002;43: Ajmone Marsan C, Zivin LZ. Factors related to the occurrence of typical paroxysmal abnormalities in the EEG records of epileptic patients. Epilepsia 1970;11: Aarts JH, Binnie CD, Smit AM, Wilkins AJ. Selective cognitive impairment during focal and generalized epileptiform EEG activity. Brain 1984;107: Marston D, Besag F, Binnie CD, Fowler M. Effects of transitory cognitive impairment on psychosocial functioning of children with epilepsy: a therapeutic trial. Dev Med Child Neurol 1993;35: Conners CK. Conners Rating Scales Manual. Instruments for use with children and adolescents. New York, NY: Multi-Health Systems, Inc; Rugland AL. Neuropsychological assessment of cognitive functioning in children with epilepsy. Epilepsia 1990;31(Suppl 4):S Binnie CD. Cognitive impairment during epileptiform discharges: is it ever justifiable to treat the EEG?. Lancet Neurol 2003;2: Weglage J, Demsky A, Pietsch M, Kurlemann G. Neuropsychological, intellectual, and behavioral findings in patients with centrotemporal spikes with and without seizures. Dev Med Child Neurol 1997;39: Stores G, Wiggs L, Campling G. Sleep disorders and their relationship to psychological disturbance in children with epilepsy. Child Care Health Dev 1998;24: Bagley C. Multiple influences on deviant behaviour in children with epilepsy. J Biosoc Sci 1973;5: Ronen GM, Richards JE, Cunningham C, Secord M, Rosenbloom D. Can sodium valproate improve learning in children with epileptiform bursts but without clinical seizures? Dev Med Child Neurol 2000;42: Bourgeois BF. Antiepileptic drugs, learning, and behavior in childhood epilepsy. Epilepsia 1998;39: Binnie CD, Van Emde BW, Kasteleijn-Nolste-Trenite DG, de Korte RA, Meijer JW, Meinardi H, et al. Acute effects of lamotrigine (BW430C) in persons with epilepsy. Epilepsia 1986;27: Besag FM. Lamotrigine in the management of subtle seizures. Rev Contemporary Pharmacother 1994;5: Eriksson AS, Knutsson E, Nergardh A. The effect of lamotrigine on epileptiform discharges in young patients with drug-resistant epilepsy. Epilepsia 2001;42: Aldenkamp AP, Arends J, Bootsma HP, Diepman L, Hulsman J, Lambrechts D, et al. Randomized double-blind parallel-group study comparing cognitive effects of a low-dose lamotrigine with valproate and placebo in healthy volunteers. Epilepsia 2002;43: Gibbs J, Appleton RE, Rosenbloom L, Yuen WC. Lamotrigine for intractable childhood epilepsy: a preliminary communication. Dev Med Child Neurol 1992;34: Buchanan N. The efficacy of lamotrigine on seizure control in 34 children, adolescents and young adults with intellectual and physical disability. Seizure 1995;4: Cortesi F, Giannotti F, Ottaviano S. Sleep problems and daytime behavior in childhood idiopathic epilepsy. Epilepsia 1999;40: Placidi F, Marciani MG, Diomedi M, Scalise A, Pauri F, Giacomini P, et al. Effects of lamotrigine on nocturnal sleep, daytime somnolence and cognitive functions in focal epilepsy. Acta Neurol Scand 2000;102: Binnie CD, Marston D. Cognitive correlates of interictal discharges. Epilepsia 1992;33(Suppl 6):S Shewmon DA, Erwin RJ. The effects of focal interictal spikes on perception and reaction time, II: neuroanatomic specificity. Electroencephalogr Clin Neurophysiol 1988;69: Treatment Of Interictal Epileptiform Discharges Can Improve Behavior In Children With Behavioral Problems And Epilepsy 117

166 THE CORRELATION OF DEFICITS IN IQ WITH MIDFACE AND DIGIT HYPOPLASIA IN CHILDREN EXPOSED IN UTERO TO ANTICONVULSANT DRUGS LEWIS B. HOLMES, MD, BRENT A. COULL, PHD, JON DORFMAN, MD, AND PETER B. ROSENBERGER, MD Objective Children exposed during pregnancy to the anticonvulsant drugs phenytoin, phenobarbital, and carbamazepine as monotherapy and polytherapy have an increased frequency of midface and digit hypoplasia. Some children also have cognitive dysfunction. The hypothesis tested is that the anticonvulsant drug-exposed child with midface and digit hypoplasia is more likely to have cognitive dysfunction. Methods Children exposed to anticonvulsant drugs (n = 80) were recruited for a follow-up evaluation, which included testing cognitive function and a physical examination for head size, height, and the presence of midface and digit hypoplasia. Microcephaly, midface and digit hypoplasia, and major malformations were correlated with full scale (FSI), performance (PIQ) and verbal (VIQ) intelligence. Results The presence of the 3 anticonvulsant-exposed children with microcephaly had a deficit of 23.7 IQ points in FSI in comparison with the other children with a normal head size. Either midface or digit hypoplasia, after excluding the persons with microcephaly and with the Bonferroni correction, correlated significantly with deficits in VIQ (212.7), PIQ (210) and FSI (212.8) IQ points (P =.0061). There was no decrease in IQ in association with major malformations. Conclusion The presence of midface and digit hypoplasia in a child exposed to anticonvulsant drugs in pregnancy is an indication for a systematic developmental evaluation. (J Pediatr 2005;146:118-22) Exposure of the fetus to anticonvulsant drugs prescribed either to prevent seizures or to treat the symptoms of mood disorders occurs in 0.4% of pregnancies 1 and is one of the most common potentially teratogenic exposures to drugs in pregnancy. In the 1970s and 1980s, clinicians reported that children who had been exposed to the anticonvulsant phenytoin alone or in combination with phenobarbital 2-4 or carbamazepine 5 had a distinctive pattern of midface hypoplasia (broad or depressed bridge of nose, short nose, long upper lip, and flattened maxilla), digit hypoplasia (an excess of arch patterns and stiff interphalangeal joints), microcephaly, growth restriction, and major malformations. This pattern of abnormalities was initially referred to as the fetal hydantoin syndrome, 3 and, later, as anticonvulsant embryopathy. Midface hypoplasia was shown in a recent cohort study 1 of newborn infants to be the most common of the features of the anticonvulsant embryopathy, occurring in 13% and 15% of infants exposed, respectively, to phenytoin and phenobarbital as monotherapy; digit hypoplasia occurred in 12% and 10% of newborn infants with these exposures, respectively. Early reports 3 suggested that the child with the anticonvulsant embryopathy had an increased risk of mental retardation, but many of the children were exposed in utero to polytherapy, which is more harmful than monotherapy. More recent studies of children and adults exposed only to an anticonvulsant as monotherapy showed that those exposed in utero to phenytoin alone (34 mother-child pairs) 6 and phenobarbital alone (33 adult men) 7 had subtle deficits of 7 to 10 IQ points. Children exposed to carbamazepine alone (36 mother-child pairs) did not show deficits in IQ in comparison with controls. 6 Although midface and digital hypoplasia are common associated findings in anticonvulsant-exposed children, the findings are less noticeable in older persons. Does their presence matter? We postulated that the child who had been exposed during FSI PIQ Full scale IQ Performance IQ VIQ Verbal IQ From the Genetics and Teratology Unit, Pediatric Service and Pediatric Neurology, Neurology Service, Massachusetts General Hospital and the Departments of Pediatrics and Neurology, Harvard Medical School, Boston, MA; and the Department of Biostatistics, Harvard School of Public Health. Support was provided for the clinical studies by Parke-Davis Inc. (now Pfizer) and, during the analyses, by the Peabody Foundation, Boston. Submitted for publication Mar 31, 2004; last revision received Jul 22, 2004; accepted Aug 18, Reprint requests: Lewis B. Holmes, MD, Genetics and Teratology Unit, Pediatric Service, Warren 801, Massachusetts General Hospital, Fruit Street, Boston, MA holmes.lewis@mgh.harvard.edu /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

167 pregnancy to anticonvulsants drugs and had midface or digit hypoplasia has a greater risk of having cognitive dysfunction than the child exposed to the same drugs but without these features. METHODS AND MATERIALS The anticonvulsant-exposed children were recruited from multiple sources: (1) the 38 exposed newborn infants identified in 1972 to 1974 in a systematic study of the physical features of 7742 newborn infants 8 ; (2) the medical records of neurologists who had treated pregnant epileptic women with anticonvulsants; (3) anticonvulsant-exposed children identified from a review of the discharge diagnoses of a health maintenance organization; and (4) children whose parents responded to advertisements about this study. A recruited child was excluded if he/she had experienced perinatal asphyxia, significant head trauma, or encephalitis in childhood, was taking phenobarbital or any other drug that can affect cognitive function, or was a twin. No child withdrew from the evaluation, once initiated. This study was approved by Massachusetts General Hospital s Human Studies Committee. The parents were asked to read and to sign the informed consent document. Children 9 years of age and older were asked to sign, indicating their assent to participating in the evaluations. The parents of the drug-exposed children were asked to bring the child to the hospital for several examinations in one session: (1) physical examination, using a protocol for recording presence/absence of specific features and measurements; the examiner was blinded as to the exposure status of the child being evaluated (controls were evaluated, as well, but those results are not presented); (2) recording dermal ridge patterns; (3) photographs of face; (4) cephalometric radiographs; (5) radiographs of the hands and feet; (6) dental molds; and (7) the Wechsler Intelligence Scale for Children. The medical history of each enrolled child was obtained, including a review of the pregnancy history for significant exposure to smoking, alcohol, and other potential teratogens. Any reported major malformations were confirmed through records obtained from the child s pediatricians and other specialists. The findings of several of these examinations were reported previously, including changes in tooth size, 9 dermal ridge patterns, 10 the epiphyses in the hands 11 and feet, 12 and the changes in cranial bone dimensions. 13,14 The correlations described were presented initially in at the annual meeting of the David W. Smith Workshop on Morphogenesis and Malformations. The presence of these outcomes was determined in this analysis: (1) the anticonvulsant face, defined as the presence of a broad or depressed bridge of nose, short nose with anteverted nostrils, and a long upper lip; (2) digit hypoplasia, defined as: (a) the presence of three or more arch patterns on the ten fingers, (b) severe stiffness of the distal interphalangeal joint in one or more fingers, or (c) marked hypoplasia of the fingernails; (3) a major malformation, defined as a structural abnormality with surgical, medical, or cosmetic importance; this definition excludes minor anomalies, birth marks, deformations, physiologic findings (much as undescended testes in a child being born prematurely) and internal anomalies detected by prenatal sonography but not by the examining pediatrician; and (4) microcephaly, defined as a head size at least 2 standard deviations below the mean values of children of the same race, sex, and age. 16 Intelligence was determined using the Wechsler Intelligence Scale for Children. This test was administered by a experienced tester who was unaware of the child s anticonvulsant drug exposure or of the presence of any associated physical abnormalities. The mean value (±1 SD) was determined for full scale IQ (FSI), performance IQ (PIQ), and verbal IQ (VIQ). The findings in all individuals tested were compared with the established norms. The distribution of IQs was expressed in these categories of function: mentally retarded <70, borderline 71 to 84, low average 85 to 95, average/high average 96 to 115, superior 116 to 129, and very superior $130. The correlation of the presence of microcephaly, midface hypoplasia, digit hypoplasia, and a major malformation was determined for FSI, PIQ, and VIQ. These correlations were made using multivariate analysis of variance. 17 Correlations were made also between FSI, PIQ, and VIQ with each anticonvulsant-drug treatment group. RESULTS Eighty children were enrolled, but only 76 had testing of intellectual function. All 76 were examined and evaluated for the presence of major malformations. Seventy-five had measurements of head size and height. Dermal ridge patterns were recorded in 68 of the 76. The question as to whether or not the infant had, overall, the features of the anticonvulsant face was added after the study began and was recorded for 41 of the children evaluated. These 76 children who had testing of cognitive function had been exposed during pregnancy to phenytoin alone (16), phenobarbital alone (17), carbamazepine alone (4), phenytoin and phenobarbitol (32), and other polytherapies (7). The six (7.9%) major malformations in these 76 children were: heart defect and severe digit hypoplasia (1), lipomenigocele (1), meatal stenosis (1), double ureter (1), ureteral obstruction (1), hydronephrosis and reflux (1), and scoliosis (1). The results of the tests of intelligence in the 76 drugexposed children are as follows: 2.6% mentally retarded, 11.8% borderline, 17.1% low average, 48.7% average/high average, 15.8% superior, and 3.9% very superior (Figure 1). After excluding the 3 persons with microcephaly, the distribution in these six groups was: 2.7%, 9.6%, 16.4%, 50.7%, 16.4%, and 4.1%, respectively. The strongest correlation of intelligence and the physical features of the anticonvulsant-exposed children was with the presence of microcephaly, a feature of 3 children, who had a deficit of IQ points in FIQ and a similar deficit in VIQ (224.3) and PIQ (217.9) (Table I) in comparison with the children with a normal head size. The The Correlation Of Deficits In IQ With Midface And Digit Hypoplasia In Children Exposed In Utero To Anticonvulsant Drugs 119

168 Figure 1. Histograms to show distribution of IQ among all 76 of anticonvulsant-exposed persons (mentally retarded <70:2.6%, borderline 71-84: 11.8%, low average 85-95: 17.1%;, average/high average : 48.7%, superior : 15.8%; very superior >130: 3.9%). 9 of 41 anticonvulsant-exposed children with midface hypoplasia ( the anticonvulsant face ) had deficits of marginal significance in FIQ (211.4 points; P =.054), PIQ (210.3 points: P =.078) and Verbal IQ (210.6 points; P =.127) in comparison with the drug-exposed persons who did not have midface hypoplasia. Similar defects in intelligence were associated with digit hypoplasia in 8 children, but these associations in FIQ (213.3 points; P =.034) and VIQ (214.7 points; P =.038) were not significant, using the Bonferroni correction. However, the correlation of intelligence test results with the presence of either midface or digit hypoplasia was significant statistically, even after excluding 2 infants with microcephaly (who had also been examined for the anticonvulsant face ): VIQ 212.7, PIQ 210, and FIQ (P =.0061). The distribution of the FIQs among the 13 children with either type of hypoplasia was: mentally retarded 0%, borderline 23%, low average 30.8%, average/high average 38.5%, superior 7.7%, very superior 0% (Figure 2, Table I). There was no decrease in IQ in association with the major malformations, which were present in 6 (7.9%) of the 76 children (Table I). The possibility of there being an association between the presence of midface hypoplasia and digit hypoplasia was evaluated using Fisher s exact test. However, there was none: 2 of 6 (33%) of the anticonvulsant drug-exposed persons with midface hypoplasia had digit hypoplasia, and 6 of 35(20%) without midface hypoplasia had digit hypoplasia (P =.578). In the group of 76 who had been exposed to several different anticonvulsant drug regimens, no significant differences, using multivariate analysis of variance, were identified in the association with FSI, VIQ, or PIQ (Table I) with the specific drug exposures in these small samples. DISCUSSION This is an analysis within a group of anticonvulsantexposed persons, without a comparison with a control group. Figure 2. Distribution of IQ among 13 persons with midface and digit hypoplasia in comparison with the 28 who did not have either feature. There were too few persons exposed to specific drugs as monotherapy to make a meaningful comparison of the effects of specific anticonvulsant drugs. However, the findings raise the question as to whether the presence of subtle craniofacial features (midfacial hypoplasia) and digit hypoplasia associated with exposure in utero to an anticonvulsant drug should be considered by the clinician as a marker of concern for the child s cognitive function. The same correlation has been suggested by Ornoy and Cohen 18 and by Mawer et al, 19 who evaluated children exposed to several different anticonvulsant drugs and reported a correlation between the presence of dysmorphic features of the face, but not major malformations, with the presence of developmental delay. The fetal effects of each anticonvulsant drug are not the same. For example, the child exposed to valproic acid has an increased risk, as high as 5%, 20 of having spina bifida, long bone deficiencies, 21 and an increased frequency of developmental delay and, possibly, autism. 19,22 Severe nail hypoplasia, stiffness of the interphalangeal joints, 11 and terminal transverse limb defects 23 have been reported only in phenytoinexposed children. Because each of these anticonvulsant drugs, ie, valproic acid, carbamazepine. and phenytoin, also causes craniofacial abnormalities, it will be important to determine whether each exposure, as monotherapy, is associated with cognitive dysfunction or not. The findings of deficits in IQ associated with the presence of subtle craniofacial and digital hypoplasia raises the question as to whether there are associated brain abnormalities in the persons with cognitive dysfunction. This association has been established in persons with fetal alcohol syndrome, who typically have subtle craniofacial abnormalities as well. 24 Magnetic resonance imaging studies of 25 persons with fetal alcohol syndrome and 25 with the less severe fetal alcohol effects showed a significant alteration in the relationship of brain stem to diencephalon and in the shape of the corpus callosum. 25 Another example is persons with isolated cleft lip/ palate or cleft palate alone who have associated subtle brain changes, eg, enlarged cavum septi pellucidi 26 and altered brain morphology 27 in association with cognitive dysfunction Holmes et al The Journal of Pediatrics January 2005

169 Table I. Comparison of IQ of individuals with specific physical features to that of the other anticonvulsantexposed individuals Physical feature VIQ PIQ FIQ 1. Microcephaly * (n = 3) (.021) (.045) (.010) * 2. Midface hypoplasia. (n = 9) a) including microcephaly (n = 9) (.127) (.078) (.054) b) without microcephaly (n = 8) 28.9 (.196) 28.9 (.125) (.095) 3. digit hypoplasia a) including child with microcephaly (n = 8) (.038) 28.8 (.166) (.034) b) excluding child with microcephaly (n = 7) (.111) 23.6 (.580) 28.9 (.161) 4. both midface and digit hypoplasia a) including all (n = 15) (.0053) * (.0041) * (.0006) * b) excluding children with microcephaly (n = 13) (.0240) (.0298) (.0061) * 5. Major malfunctions (n = 6) a) including child with microcephaly (n = 6) 9.5 (.218) 6.7 (.308) 9.6 (.158) b) excluding child with microcephaly (n = 5) 7.1 (.433) 4.5 (.554) 6.8 (.392) () = P value. *for P value to be significant, using the Bonferroni correction, it must be, Table II. Correlation of intelligence with drug treatment Anticonvulsant drug FIQ PIQ VIQ Polytherapy 1 vs monotherapy (.333) 15.6 (.111) 11.1 (.797) Phenytoin (n = 16) vs other drugs (n = 60) (.864) (.389) (.398) Phenobarbital (n = 17) vs all other drugs (n = 59) (.834) (.859) (.669) Phenytoin and Phenobarbital (n = 32) vs all other drugs (n = 44) 17.5 (.310) (.732) (.221) 1 all polytherapies: 32 phenytoin and phenobarbital; 7 other combinations. 2 all monotherapies: phenytoin (16), phenobarbital (17), and carbamazepine (4). The anticonvulsant-exposed infant with midface and digit hypoplasia may have a greater genetic susceptibility to the teratogenic effect of these drugs than the drug-exposed but unaffected child. Several underlying biochemical or physiologic deficiencies have been postulated to be the basis for this increased susceptibility: a deficiency of the metabolizing enzyme epoxide hydrolase, 29 an increased production of free radicals, 30 bradycardia and associated fetal hypoperfusion, 31 and altered expression of the HOX transcription factors. 32 To our knowledge, none of these postulated mechanisms of teratogenesis has been extended to explain the specific predilection for midface hypoplasia and digit hypoplasia in anticonvulsant-exposed persons. We suggest, as one potential example, that an altered expression of the transcription factor MSX-1 (HOX 7) could correlate with the occurrence of both midface hypoplasia and digit hypoplasia. Several observations are consistent with this hypothesis. MSX-1 controls the development of the tooth field in the mesenchyme of the developing mouth. 33 In one study, 9 children exposed in utero to phenytoin and phenobarbital had an increased frequency of agenesis of teeth. A missense mutation in MSX-1 was identified in persons with hereditary absence of teeth. 34 Knockout mice deficient in MSX-1 have craniofacial and dental anomalies. 35 Polymorphisms of MSX-1 have been associated with limb malformations. 36 Genetic differences in the MSX-1 transcription factor or in its function, when present by chance in the anticonvulsant-exposed fetus, could cause that child to have altered growth of the mid-face structures and the developing brain. These anatomic findings are visible at birth, and the associated changes in cranial bone structures of affected children persist into the teenage years. 14 This hypothesis about the teratogenesis of anticonvulsants being a result of the altered expression of the transcription factor MSX-1 has not, to our knowledge, been tested or confirmed. Future studies of the correlation of midface hypoplasia with cognitive function should focus on correlations with specific anticonvulsant drugs as monotherapy and in comparison with a control group. The potential confounders to be analyzed should include the dose and maternal blood levels of the anticonvulsant drugs and the intelligence of each parent. In addition, there can be more objectivity in the characterization of the midface hypoplasia, such as shortening of the maxilla, 13,14 by measurements of cranial structures measured in cephalometric radiographs. Studies of brain structures, using magnetic resonance imaging-based brain morphometry 37 may identify additional abnormalities, particularly in persons with midface hypoplasia and cognitive dysfunction. The Correlation Of Deficits In IQ With Midface And Digit Hypoplasia In Children Exposed In Utero To Anticonvulsant Drugs 121

170 We appreciate very much the cooperation of the families in these studies. We thank especially Gretchen Timmel, MEd, Psychologist, Learning Disabilities Unit, Pediatric Neurology, Massachusetts General Hospital for Children, who carried out testing of cognitive function. REFERENCES 1. Holmes LB, Harvey EA, Coull BA, Huntington KB, Khoshbin S, Hayes AM, et al. The teratogenicity of anticonvulsant drugs. N Engl J Med 2001;344: Hill RM, Verniaud WN, Horning MG, McCulley LB, Morgan NF. Infants exposed in utero to antiepileptic drugs. Am J Dis Child 1974;127: Hanson JW, Myrianthopoulos NC, Sedgwick Harvey MA, Smith DW. Risks of the offspring of women treated with hydantoin anticonvulsants, with emphasis on the fetal hydantoin syndrome. J Pediatr 1976;89: Hiillesmaa VK, Teramo K, Granstrom ML, Bardy AH. Fetal head growth retardation associated with maternal antiepileptic drugs. Lancet 1981; 11: Jones KL, Lacro RV, Johnson KA, Adams J. Pattern of malformations in the children of women treated with carbamazepine during pregnancy. N Engl J Med 1989;320: Scolnik D, Nulman I, Rovet J, Gladstone D, Czuchta D, Gardner HA, et al. Neurodevelopment of children exposed in utero to phenytoin and carbamazepine monotherapy. JAMA 1994;271: Reinisch JM, Sanders SA, Mortenson EL, Rubin DB. In utero exposure to phenobarbital and intelligence deficits in adult men. JAMA 1995;274: Leppig KA, Werler MM, Cann CI, Cook CA, Holmes LB. Predictive value of minor anomalies in association with major malformations. J Pediatr 1987;110: Orup HI Jr, Keith DA, Holmes LB. Prenatal anticonvulsant drug exposure: teratogenic effect on the dentition. J Craniofac Genet Dev Biol 1998;18: Bokhari A, Coull B, Holmes LB. Effect of prenatal exposure to anticonvulsant drugs on dermal ridge patterns of fingers. Teratology 2002;66: Lu MCK, Sammel MD, Cleveland RH, Ryan LM, Holmes LB. Digit effects produced by prenatal exposure to antiepileptic drugs. Teratology 2000; 61: Bokhari A, Connolly S, Harvey EA, Holmes LB. Effects on toes from prenatal exposure to anticonvulsants. Teratology 1998;66: VanLang QN, Tassinari MS, Keith DA, Holmes LB. Effect of in utero exposure to anticonvulsants on craniofacial development and growth. J Craniofacial Genet Dev Biol 1984;4: Orup HI Jr, Holmes LB, Keith DA, Coull BA. Craniofacial skeletal deviations following in utero exposure to the anticonvulsant phenytoin: monotherapy and polytherapy. Orthod Craniofacial Res 2003;6: Holmes LB, Harvey EA, Coull B, Rosenberger PB, Orup HI Jr, Bokhari A, et al. Anticonvulsant embryopathy: the interrelationship of cranial and skeletal changes in cognitive function. Proc Greenwood Ctr 1998;18: Hall JG, Froster-Iskemius UG, Allanson JE. Handbook of Normal Physical Measurements. New York: Oxford University Press; Johnson RA, Witchern DW. Applied Multivariate Statistical Analyzes. Englewood Cliffs, NJ: Prentice Hall; Ornoy A, Cohen E. Outcome of children born to epileptic mothers treated with carbamazepine during pregnancy. Arch Dis Childh 1996;75: Mawer G, Clayton-Smith J, Coyle H, Kini U. Outcome of pregnancy in women attending an out-patient epilepsy clinic: adverse features associated with higher doses of sodium valproate. Seizure 2002;11: Omtizigt JGC, Los FJ, Grobbee DE, Pijpers L, Jahoda MGJ, Brandenburg H, et al. The risk of spina bifida aperta after first-trimester exposure to valproate in a prenatal cohort. Neurol 1992;42(suppl 5): Verloes A, Frikiche A, Gremillet C, Paquay T, Decortis T, Rigo, J et al. Proximal phocomelia and radial ray aplasia in fetal valproic syndrome. Eur J Pediatr 1990;149: Moore SJ, Turnpenny P, Quinn A, Glover S, Lloyd DJ, Montgomery T, et al. A clinical study of 57 children with fetal anticonvulsant syndromes. J Med Gen 2000;37: Sabry MA, Farag TI. Hand anomalies in fetal-hydantoin syndrome: from nail/phalangeal hypoplasia to unilateral acheiria. Am J Med Genet 1996; 62: Frias JL, King GJ, Williams CA. Cephalometric assessment of selected malformation syndromes. Birth Defects: Original Article Series 1982;18: Bookstein FL, Sampson PD, Streissguth AP, Connors PD. Geometric morphometrics of corpus callosum and subeortical structures in the fetalalcohol-affected brain. Teratology 2001;64: Nopoulos P, Berg S, Van Demark D, Richman L, Canady J, Andreasen NC. Increased incidence of a midline brain anomaly in patients with nonsyndromic clefts of the lip on palate. J Neuroimaging 2001;11: Nopoulos P, Berg S, Canady J, Richman L, Van Demark D, Andreasen NC. Structural brain abnormalities in adult males with clefts of the lip and/or palate. Genet Med 2002;4: Nopoulos P, Berg S, Van Demark D, Richman L, Canady J, Andreasen NC. Cognitive dysfunction in adult males with non-syndromic clefts of the lip and/or palate. Neuropsychologia 2002;40: Buehler BA, Delimont D, Van Waes M, Finnell RH. Prenatal prediction of risk of the fetal hydantoin syndrome. N Engl J Med 1990; 322: Wells PG, Win LM. Biochemical toxicology of chemical teratogenesis. Clin Rev Biochem Mol Biol 1996;31: Danielsson BR, Azarbayjani F, Skold AC, Webster WS. Initiation of phenytoin teratogenesis: pharmacologically induced embryonic bradycardia and arrhythomia resulting in hypoxic and possible free radical damage at reoxygenation. Teratology 1997;56: Faiella A, Wernig M, Consalez GG, Hostick U, Hofmann C, Hustert E, et al. A mouse model for valproate teratogenicity: parental effects, homeotic transformations, and altered HOX expression. Hum Mol Gen 2000;9: Sharpe PT. Homeobox genes and orofacial development. Conn Tiss Res 1995;32: Vastardis H, Karimbuux N, Gerthua SW, Seidman JG, Seidman CE. A human MSX1 homeodomain missence mutation causes selective tooth agenesis. Nature Genetics 1996;13: Chen YP, Bei M, Woo I, Satokata I, Maas R. MSX1 controls inductive signaling in mammalian tooth morphogenesis. Development 1996;122: Huang S-J, Beaty TH, McIntosh I, Hefferon T, Panny SR. Association between homeobox-containing gene MSX1 and the occurrence of limb deficiency. Am J Med Gen 1998;75: Caviness VS Jr, Makris N, Lange NT, Herbert M, Kennedy DN. Advanced applications of MRI in human brain science. Keio J Med 2000;49: Holmes et al The Journal of Pediatrics January 2005

171 MEDICAL PROGRESS THE WEALTH OF INFORMATION CONVEYED BY GESTATIONAL AGE ALAN LEVITON,MD,* EVE BLAIR, PHD,* OLAF DAMMANN,MD,** AND ELIZABETH ALLRED,MS** Epidemiologic studies designed to identify the antecedents of neonatal disorders often include a variable for gestational age at delivery. The gestational age at delivery variable is a surrogate for an enormous amount of complex information. Yet, some use this variable in their analyses as if everything about it were simple and known. Our goal is to raise doubts about the wisdom of this approach and to provoke interest in a variable of utmost importance to those who search for events and characteristics that influence early human development. In this paper, we explore the variety and extent of information conveyed by the gestational age at delivery variable and offer suggestions about how to analyze data to identify the prenatal and perinatal antecedents of disease when ignoring the complexities of the gestational age variable would provide misleading information. WHAT IS GESTATIONAL AGE? Gestational age is a measure of elapsed time. Because fetal development follows a strict time schedule, the gestational age at delivery is closely associated with many other things, including the following. Mediators of Damage If delivery occurs much before term, the variable labeled gestational age at delivery carries information about the phenomena that lead to preterm birth. This includes expressions of fetal dysfunction such as severe growth restriction and unfavorable cardiotocogram stress patterns. The information carried by the gestational age at delivery variable can also include information about such pathologic disorders as preterm labor and, before labor, preterm rupture of membranes. Both are now considered to be sequelae or correlates of fetal inflammation. 1 In addition, systemic fetal inflammation appears to influence the risk of intraventricular hemorrhage, 2-4 cerebral white matter damage, 3,5,6 cerebral palsy, 7,8 and chronic lung disease. 9 Not only does gestational age convey information about potentially damaging exposures, but it also conveys information about how the fetus responds to such exposures. Fetal phenomena that vary with gestational age are considered developmentally regulated. For example, the fetal systemic inflammatory response appears to be considerably more vigorous in very preterm newborn infants than in gestationally older newborn infants Inadequacy of Protective Substances The paucity of protectors model is gaining acceptance because of increasing evidence that very low gestational age newborn infants often do not produce adequate amounts of substances needed for their own well-being. 15 In utero, the fetus s inability to produce the full amount of needed proteins is not a problem because the placenta or the gravida can provide supplementation. 16 Ex utero, not only is the newborn infant deprived of this supplementation, but the need for these proteins may be increased. In this model, the synthesis of each protein needed to stave off disease is a developmentally regulated process. One group of protectors consists of hormones and growth factors that reduce programmed cell death and necrosis in the face of adversity. 15 This group is exemplified by thyroid hormones. Circulating levels of thyroxine increase with gestational age. 17 After birth, preterm born infants with the lowest levels of thyroxine are at increased risk of intraventricular hemorrhage, 18 cerebral white matter damage, 19 cognitive limitation, 20 and cerebral palsy. 21 The beneficial effects of thyroid hormones on the development of oligodendrocytes in the brain 22 support the hypothesis that a paucity of thyroxine contributes to brain damage. Very low gestational age newborn infants do not produce adequate amounts of proteins that regulate the synthesis of substances that can have adverse effects at high concentrations. 23 For example, interleukin-10 can modulate the synthesis of inflammatory From the Neuroepidemiology Unit, Children s Hospital and Department of Neurology, Harvard Medical School, Boston, Massachusetts; the Centre for Child Health Research, University of Western Australia, West Perth, Western Australia; and Perinatal Infectious Disease Epidemiology Unit, Departments of Obstetrics and Pediatrics, Hannover Medical School, Hannover, Germany. *Contributed equally. **Contributed equally. Supported by grants from the National Institutes of Health (NS40069), National Health and Medical Research Council (No ), and Wilhelm- Hirte Stiftung, Hannover, Germany. Submitted for publication June 17, 2004; revision received Sep 9, 2004; accepted Sep 30, Reprint requests: Dr Alan Leviton, Children s Hospital, 300 Longwood Ave, Mailstop 222, Boston, MA J Pediatr 2005;146: /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

172 cytokines. Very low gestational age newborn infants produce less interleukin-10 than term newborn infants In laboratory animals, interleukin and other substances that block the synthesis of products of inflammation reduce brain damage after an insult. 23 Thus, the limited production of anti-inflammatory agents represents another set of developmentally regulated processes that might place the very low gestational age newborn infant at heightened risk. Protectors can also include substances that remove toxins. For example, free radicals have been implicated in the pathogenesis of a number of disorders relatively unique to the preterm newborn infant, 28 including chronic lung disease, 29 retinopathy of prematurity, 30 and necrotizing enterocolitis. 31 The blood of preterm newborn infants has less antioxidant activity than does the blood of term newborn infants 32 and is therefore less able to minimize the damage caused by free radicals. Structural Vulnerability Structural changes such as alveolarization of the lung 33 and the synthesis of myelin in the brain 34 are also developmentally regulated processes that may be especially sensitive to disruption at particular stages of their development. Postnatal Physiologic Instability Postnatal arterial ph, po 2, and blood pressure increase with increasing gestational age at delivery. 35 Even among infants of comparable gestational age at delivery, these measures vary, and those with lower values are at greater risk of morbidity and mortality. Maternal Factors Gestational age at delivery can convey information about maternal demographic characteristics. For example, preterm birth after a history of infertility may reflect not simply the medical condition of infertility but also socioeconomic factors that influence access to conception assistance. Other maternal characteristics are more clearly medical and involve factors on the causal path to preterm birth, including maternal indications for preterm delivery such as severe pregnancy-induced hypertension. QUALITY OF GESTATIONAL AGE INFORMATION The accuracy of gestational age calculated from the date of the last normal menstrual period leaves something to be desired, 36 especially when women are not sure of the date or whether their last period was normal. Even when they are certain of the date of the last menstrual period, gestational age calculated from the last period can be underestimated or overestimated. 37 The most valid measure of gestational age in pregnancies that are not the result of in vitro fertilization is based on early ultrasound measurements. 38 Nevertheless, gestational age estimates based on a combination of last menstrual period and early fetal ultrasound measurements can vary considerably, depending on the algorithm chosen 39 and when fetometry is performed. DEVELOPMENTALLY REGULATED PROCESSES Quality of Information about Developmentally Regulated Processes In the absence of other information about maturation, the duration of gestation can act as a surrogate for all the developmentally regulated processes that are often collectively subsumed under the term maturity. Some developmentally regulated processes cannot be assessed directly, and some we do not even know about. As our understanding of fetal development advances, more developmental processes will become measurable. Our hope is that as measurement of these processes improves, they will be included in relevant data analyses. At present, much of the information about a developmentally regulated process consists of just a few snapshots rather than the many images that constitute a motion picture. As a consequence, the information about developmentally regulated processes is limited. Variables chosen for inclusion in analysis must therefore be chosen with care, since conclusions drawn from analyses of data based on faulty measurement are likely to be correspondingly faulty. However, as understanding of fetal development advances, developmental processes will be measured more accurately, so that one day they will be suitable for including in relevant data analyses. In perinatal epidemiology, what is measured is often only an approximation of what would be measured under ideal conditions. For example, perhaps the best measure of thyroidrelated developmentally regulated processes for studies of brain disorders is the active form of the hormone in the brain, tri-iodothyronine (brain-t3). 17 What is routinely measured in many state-mandated newborn screening programs is the blood (not brain) level of the thyroid hormone with four (not three) iodine molecules (blood-t4). The relation between brain-t3 and blood-t4 is probably not linear and may even vary under different conditions. 17 Nevertheless, blood-t4 may be the best routinely available measure of the thyroid-related developmentally regulated process. Some responses of circulating cells to stimuli are currently measurable in vitro. 25,40,41 When response capability is not assessed, the assumption is that the measurement of a protein in the blood provides all the information that is needed about the developmentally regulated processes leading to the synthesis and degradation of that protein. The concentration of the protein might be the baseline, the trough, or the zenith of a response to an exogenous stimulus. Thus, serial specimens are needed. 124 Leviton et al The Journal of Pediatrics January 2005

173 Intercorrelations Among Developmentally Regulated Processes As a result of differing genetic and environmental influences, each baby is not developmentally at the same level as all other babies of the same gestational age. Although the gestational age at delivery variable conveys information about the central tendency of each developmentally regulated process (the gestation-specific population mean), it is not individualized to each of that specific baby s developmentally regulated processes. Therefore, use of the gestational age at delivery variable to adjust simultaneously for all developmentally regulated processes assumes that every baby is developmentally identical in every possible way to every other baby of the same gestational age. To the extent that this assumption is false, adjusting in this manner is suboptimal. Because all developmentally regulated processes are, by definition, highly associated with gestational age, all are likely to correlate with each other, if imperfectly. This poses a problem when trying to identify the unique contribution of an individual developmentally regulated process. Measuring Developmentally Regulated Processes Each developmentally regulated process can be described/measured/classified in many ways. Here we consider three broadly defined measures, the measured value, the gestational age at delivery-specific Z score, and a gestational age or gestational age at delivery surrogate. The measured value can be represented by two components, the gestational age specific central tendency (for which gestational age is an excellent surrogate) and the individual variation from that central tendency. For variables that are normally distributed within each gestation stratum or can be transformed to approximate a normal distribution, this variation can be expressed as Z scores, which are multiples of the standard deviation of the gestational age specific distribution for that developmentally regulated process. Each Z score measures the precocity or retardation of a particular developmentally regulated process in the individual relative to all individuals of the same gestational age. One of the conceptual advantages of Z scores is that they allow ready comparison. Thus, looking at a newborn infant s Z score for a developmentally regulated process can provide a quick impression of how close this process is to his or her gestational age target. This is similar to using a Denver Developmental Screening Test to assess multiple developmental tasks and comparing the child s performance on each task to that of a large cohort. At any particular gestational age at delivery, sicker babies can be identified by virtue of how extreme in the undesirable direction are their scores for each measure of physiologic well-being. This is the basis of illness severity scores ability to predict death. 42 Nonlinearity of Developmentally Regulated Process Disease Relations Frequently, a paper reports that adjustment was made for gestational age at delivery but does not provide additional detail. If the investigators used an untransformed continuous gestational age variable, they may have unrealistically assumed that the gestational age disease relation is linear and by extension that the relation between the relevant developmentally regulated processes and disease risk are also linear. We encourage investigators to report fully what was done and make explicit all relevant assumptions. Guidelines for Data Analysis No simple set of guidelines serves all instances. Each decision about adjusting for gestational age needs to be individualized, considering the question that is being asked, the particular set of antecedents and outcomes, and what is known about the relevant developmentally regulated processes and whether or not they are currently measurable. These very general guidelines are intended to provide some suggestions about ways to handle some problems we identify. These suggestions apply to observational studies, clinical trials, and other studies in which gestational age is associated with both outcome and exposure of interest. Because postnatal age also conveys information about development and senescence, we raise the possibility that these guidelines for gestational age may also apply to the postnatal age variable. We suggest the following steps: (1) Identify measurable developmentally regulated processes that influence the exposure-outcome relation. (2) Measure each of the developmentally regulated processes as well as possible. (3) Assess the relation between each relevant developmentally regulated process and gestational age in strata of the exposure, the outcome, and all the other relevant developmentally regulated processes. Knowledge of the nature of the relations among developmental processes and gestational age should assist in identifying the transformation that will convey the most useful information in multivariable models. Where the relations are not linear, the developmentally regulated process variables should be transformed in a way that might make them so. (4) Explore the relation between the exposure and the outcome in all gestational age at delivery strata and in strata of all relevant developmentally regulated processes. Exploring exposure-outcome relations in narrow strata defined by gestational age at delivery indicates how the exposure-outcome relation changes with gestational age (effect modification). 43 This information can help with the decision how best to include in the multivariable model, information about gestational age at delivery, which could be continuous (linear or transformed) but may be categoric. 44 For example, if different antecedent-outcome relations are seen in N different gestational age at delivery strata, then one solution is to create N-1 individual gestational age at delivery variables, with the remaining stratum representing the reference category. One-to-one matching on gestational age is often difficult to achieve, in part because the matching must be very close to be effective. Even close matching, however, The Wealth Of Information Conveyed By Gestational Age 125

174 does not eliminate the desirability of measuring all developmentally regulated processes related both to exposure and outcome and then to assess how each child s measurement departs from the central tendency. (5) Because it conveys information about all unmeasured and unidentified developmentally regulated processes, some of which may be associated with the exposure-outcome relation, always assess the effects of including a gestational age at delivery variable in all relevant risk factor analyses. Many developmentally regulated processes are not currently individually measured or even identifiable. In multivariable models, gestational age at delivery will convey information about them as well as about all gestational age related exogenous pregnancy factors that are associated with the outcome, but not in the model. When all relevant gestational age associated factors are included in the model, the gestational age at delivery will no longer convey supplemental information. (6) Evaluate the need for more than one gestational age at delivery variable. Different developmentally regulated processes can have different relations with gestational age at delivery. One might have a linear relation and another might require some transformation of the gestational age at delivery variable to make the relation linear. It is reasonable to include more than one gestational age at delivery variable, each appropriate for a different developmentally regulated process. (7) Include a variable for each measurable developmentally regulated process that confounds the exposure outcome relation. Supplementing the gestational age at delivery variable with a variable for each measured developmentally regulated process that might distort the exposure outcome relation reduces reliance on assumptions that a central tendency measure is acceptable when, in fact, specific information is preferable. Including additional variables has the potential to reduce power, which may be important when a study is already close to acceptable limits of power. However, this loss of power will tend to be offset by the increased precision afforded by the more accurate identification and measurement of the covariates affecting the outcome. Another approach is to adjust for several developmentally regulated processes with a single summary variable. 45 (8) Compare multiple multivariable models. One of our favorite strategies is to compare different multivariable models. For example, models including gestational age at delivery as a linear variable are compared with those including gestational age at delivery in other forms (eg, qualitative dummy variables) or those without any gestational age at delivery variable. Usually, we do not know how much of the risk information conveyed by the gestational age at delivery variable should be allocated to specific correlates of gestational age at delivery. For example, if two highly correlated developmentally regulated processes have similar relations with gestational age, the exposure, and the outcome, the decision about the best model may rest less on statistical considerations than on biological plausibility and prior knowledge. 46 Multivariable models that include prepregnancy, gestational, and perinatal factors should allow variables to be assessed in the temporal order in which they might influence the outcome This is especially important when preconditioning or sensitization are possible. CONCLUSION AND THE FUTURE When good information about delivery-associated exposures, vulnerabilities, response modifiers, and maternal characteristics is unavailable, adjusting for gestational age at delivery may be the only, if suboptimal, strategy. When a significant developmentally regulated confounder of the association under investigation can be reliably measured, it should be included in the analysis as Z scores together with a gestational age at delivery variable. This allows the available information to be differentiated from unidentified or unmeasured developmentally regulated processes associated with the outcome. In the near future, many developmentally regulated characteristics involved in disease production, amelioration, and prevention will be much better measured than they are today. In the more distant future, investigators will have the option of entirely replacing the gestational age at delivery variable. However, a unidimensional surrogate of these many processes will doubtless still be useful in some circumstances (eg, as a screening tool or where populations rather than individuals are the subjects as in the adjustment of comparisons of perinatal mortality between different centers). 44,45 REFERENCES 1. Goncalves LF, Chaiworapongsa T, Romero R. Intrauterine infection and prematurity. Ment Retard Dev Disabil Res Rev 2002;8: Investigators DEN. The correlation between placental pathology and intraventricular hemorrhage in the preterm infant. Pediatr Res 1998;43: De Felice C, Toti P, Laurini RN, Stumpo M, Picciolini E, Todros T, et al. Early neonatal brain injury in histologic chorioamnionitis. J Pediatr 2001;138: Heep A, Behrendt D, Nitsch P, Fimmers R, Bartmann P, Dembinski J. Increased serum levels of interleukin 6 are associated with severe intraventricular haemorrhage in extremely premature infants. Arch Dis Child Fetal Neonatal 2003;88:F Yoon BH, Romero R, Yang SH, Jun JK, Kim IO, Choi JH, et al. Interleukin-6 concentrations in umbilical cord plasma are elevated in neonates with white matter lesions associated with periventricular leukomalacia. Am J Obstet Gynecol 1996;174: Leviton A, Paneth N, Reuss ML, Susser M, Allred EN, Dammann O, et al. Maternal infection, fetal inflammatory response, and brain damage in very low birth weight infants: Developmental Epidemiology Network Investigators. Pediatr Res 1999;46: Nelson KB, Dambrosia JM, Grether JK, Phillips TM. Neonatal cytokines and coagulation factors in children with cerebral palsy. Ann Neurol 1998;44: Grether JK, Nelson KB, Dambrosia JM, Phillips TM. Interferons and cerebral palsy. J Pediatr 1999;134: Speer CP. Pre- and postnatal inflammatory mechanisms in chronic lung disease of preterm infants. Paediatr Respir Rev 2004;5:S Leviton et al The Journal of Pediatrics January 2005

175 10. Berner R, Niemeyer CM, Leititis JU, Funke A, Schwab C, Rau U, et al. Plasma levels and gene expression of granulocyte colony-stimulating factor, tumor necrosis factor-alpha, interleukin (IL)-1beta, IL-6, IL-8, and soluble intercellular adhesion molecule-1 in neonatal early onset sepsis. Pediatr Res 1998;44: Rebuck N, Gibson A, Finn A. Neutrophil adhesion molecules in term and premature infants: normal or enhanced leucocyte integrins but defective L-selectin expression and shedding. Clin Exp Immunol 1995;101: Nanthakumar NN, Fusunyan RD, Sanderson I, Walker WA. Inflammation in the developing human intestine: a possible pathophysiologic contribution to necrotizing enterocolitis. Proc Natl Acad Sci U S A 2000;97: Rozycki HJ, Comber PG, Huff TF. Cytokines and oxygen radicals after hyperoxia in preterm and term alveolar macrophages. Am J Physiol Lung Cell Mol Physiol 2002;282:L Yoon BH, Romero R, Moon J, Chaiworapongsa T, Espinoza J, Kim YM, et al. Differences in the fetal interleukin-6 response to microbial invasion of the amniotic cavity between term and preterm gestation. J Matern Fetal Neonatal Med 2003;13: Dammann O, Leviton A. Brain damage in preterm newborns: might enhancement of developmentally-regulated endogenous protection open a door for prevention? Pediatrics 1999;104: Reuss ML, Paneth N, Susser M. Does the loss of placental hormones contribute to neurodevelopmental disabilities in preterm infants? Dev Med Child Neurol 1994;36: Morreale de Escobar G, Ares S. The hypothyroxinemia of prematurity. J Clin Endocrinol Metab 1998;83: Paul DA, Leef KH, Stefano JL, Bartoshesky L. Low serum thyroxine on initial newborn screening is associated with intraventricular hemorrhage and death in very low birth weight infants. Pediatrics 1998;101: Leviton A, Paneth N, Reuss ML, Susser M, Allred EN, Dammann O, et al. Hypothyroxinemia of prematurity and the risk of cerebral white matter damage. J Pediatr 1999;134: Den Ouden AL, Kok JH, Verkerk PH, Brand R, Verloove-Vanhorick SP. The relation between neonatal thyroxine levels and neurodevelopmental outcome at age 5 and 9 years in a national cohort of very preterm and/or very low birth weight infants. Pediatr Res 1996;39: Reuss ML, Paneth N, Pinto-Martin JA, Lorenz JM, Susser M. The relation of transient hypothyroxinemia in preterm infants to neurologic development at two years of age. N Engl J Med 1996;334: Rodriguez-Pena A. Oligodendrocyte development and thyroid hormone. J Neurobiol 1999;40: Dammann O, Leviton A. Brain damage in preterm newborns: biological response modification as a strategy to reduce disabilities. J Pediatr 2000;136: Jones CA, Cayabyab RG, Kwong KY, Stotts C, Wong B, Hamdan H, et al. Undetectable interleukin (IL)-10 and persistent IL-8 expression early in hyaline membrane disease: a possible developmental basis for the predisposition to chronic lung inflammation in preterm newborns. Pediatr Res 1996;39: Chheda S, Palkowetz KH, Garofalo R, Rassin DK, Goldman AS. Decreased interleukin-10 production by neonatal monocytes and T cells: relationship to decreased production and expression of tumor necrosis factoralpha and its receptors. Pediatr Res 1996;40: Blahnik MJ, Ramanathan R, Riley CR, Minoo P. Lipopolysaccharideinduced tumor necrosis factor-alpha and IL-10 production by lung macrophages from preterm and term neonates. Pediatr Res 2001;50: Spera PA, Ellison JA, Feuerstein GZ, Barone FC. IL-10 reduces rat brain injury following focal stroke. Neurosci Lett 1998;251: Saugstad OD. Update on oxygen radical disease in neonatology. Curr Opin Obstet Gynecol 2001;13: Russell GA. Antioxidants and neonatal lung disease. Eur J Pediatr 1994; 153:S Pitkanen OM, Hallman M, Andersson SM. Correlation of free oxygen radical-induced lipid peroxidation with outcome in very low birth weight infants. J Pediatr 1990;116: Grisham MB. Oxidants and free radicals in inflammatory bowel disease. Lancet 1994;344: Rogers S, Witz G, Anwar M, Hiatt M, Hegyi T. Antioxidant capacity and oxygen radical diseases in the preterm newborn. Arch Pediatr Adolesc Med 2000;154: Hilaire G, Duron B. Maturation of the mammalian respiratory system. Physiol Rev 1999;79: Back SA, Luo NL, Borenstein NS, Levine JM, Volpe JJ, Kinney HC. Late oligodendrocyte progenitors coincide with the developmental window of vulnerability for human perinatal white matter injury. J Neurosci 2001;21: Richardson DK, Phibbs CS, Gray JE, McCormick MC, Workman- Daniels K, Goldmann DA. Birth weight and illness severity: independent predictors of neonatal mortality. Pediatrics 1993;91: Kramer MS, McLean FH, Boyd ME, Usher RH. The validity of gestational age estimation by menstrual dating in term, preterm, and postterm gestations. JAMA 1988;260: Mongelli M, Wilcox M, Gardosi J. Estimating the date of confinement: ultrasonographic biometry versus certain menstrual dates. Am J Obstet Gynecol 1996;174: Neilson JP. Ultrasound for fetal assessment in early pregnancy. Cochrane Database Syst Rev 2000: CD Blondel B, Morin I, Platt RW, Kramer MS, Usher R, Breart G. Algorithms for combining menstrual and ultrasound estimates of gestational age: consequences for rates of preterm and postterm birth. BJOG 2002;109: Schultz C, Rott C, Temming P, Schlenke P, Moller JC, Bucsky P. Enhanced interleukin-6 and interleukin-8 synthesis in term and preterm infants. Pediatr Res 2002;51: Kotiranta-Ainamo A, Rautonen J, Rautonen N. Imbalanced cytokine secretion in newborns. Biol Neonate 2004;85: Richardson DK, Corcoran JD, Escobar GJ, Lee SK. SNAP-II and SNAPPE-II: simplified newborn illness severity and mortality risk scores. J Pediatr 2001;138: Dammann O, Allred EN, Kuban KC, Van Marter LJ, Stewart JE, Pagano M, et al. Hypocarbia during the first 24 postnatal hours and white matter echolucencies in newborns </=28 weeks gestation. Pediatr Res 2001; 49: Cochran WG. The effectiveness of adjustment by subclassification in removing bias in observational studies. Biometrics 1968;24: Rubin DB. Estimating causal effects from large data sets using propensity scores. Ann Intern Med 1997;127: Hernan MA, Hernandez-Diaz S, Werler MM, Mitchell AA. Causal knowledge as a prerequisite for confounding evaluation: an application to birth defects epidemiology. Am J Epidemiol 2002;155: Leviton A, Pagano M, Kuban KC, Krishnamoorthy KS, Sullivan KF, Allred EN. The epidemiology of germinal matrix hemorrhage during the first half-day of life. Dev Med Child Neurol 1991;33: Gauvreau K, Pagano M. Interpretation of the relationships among variables in an observational study. Boston: Harvard School of Public Health; Blair E, Stanley F. When can cerebral palsy be prevented? The generation of causal hypotheses by multivariate analysis of a case-control study. Paediatr Perinat Epidemiol 1993;7: The Wealth Of Information Conveyed By Gestational Age 127

176 CLINICAL AND LABORATORY OBSERVATIONS HOMOZYGOUS FACTOR X GENE MUTATIONS GLY380ARG AND TYR163DELAT ARE ASSOCIATED WITH PERINATAL INTRACRANIAL HEMORRHAGE FALKO H. HERRMANN, PHD, MARTA NAVARETTE, MD, LIZBETH SALAZAR-SANCHEZ, PHD, JUAN M. CARILLO, MD, GUENTER AUERSWALD,MD,AND KARIN WULFF,PHD Intracranial hemorrhage (ICH) is a severe complication of Factor X (FX) deficiency. We report 6 homozygous patients with central nervous system (ICH) bleeds. Five patients are homozygous for the mutation Gly380Arg and one for the novel deletion Tyr163delAT. We describe the association of these mutations with ICH bleeding. (J Pediatr 2005;146:128-30) Factor X (FX) deficiency is a very rare hemostatic disorder inherited as an autosomal recessive trait. The FX gene (factor 10 [F10]) is located on chromosome 13q14 and contains eight exons. 1 The variable clinical severity resulting from FX deficiency correlates poorly with the laboratory phenotype. 2 Intracranial hemorrhage (ICH) at birth or in early childhood is a rare but severe complication of FX deficiency. METHODS Patients The patients described in this report came from different areas of Costa Rica. The clinical and hemostatic assessment was done at the Children s Hospital of CCSS, San Jose, Costa Rica. ICH was diagnosed by computed tomography. The patients gave informed written consent for the molecular genetic analysis, which include the storage of the DNA for further studies of the family. The study was approved by the local human investigation committee. Molecular Genetic Study Ethylenediaminetetraacidic acid (EDTA)-blood was collected from patients with FX deficiency. The molecular basis of FX deficiency was analyzed by sequencing the F10 gene (all coding regions, the exon-intron boundaries, and the 5# flanking region containing the promoter) in probands with reduced FX level. All sequences were analyzed at least twice. 3,4 The haplotypes of FX alleles were determined by the following polymorphisms: hexanucleotide (TTGTGA) deletion in nt-343 to nt-348 (symbol A, the rare allele carries the deletion), base changes in the promotor region-222c>t (symbol B) and 2220C>A (symbol C), polymorphisms in intervening regions IVS1-55ins16bp (E); IVS2-16C>T (G); IVS3198A>C (H), as well as the exon 7 polymorphism Thr224Thr (817T>C, symbol D). The frequent alleles of the studied polymorphisms are marked by 1, the rare alleles by 2. The polymorphisms were identified by sequencing, 5 or in case of deletion of TTGTGA, by heteroduplex analysis. 6 The FX haplotypes were detected by analysis of the homozygosity of the mutations. RESULTS AND DISCUSSION We report on ICH, which occurred in 6 homozygous FX deficient patients: 5 patients were homozygous for the missense mutation Gly380Arg; 1 patient was homozygous for the FX gene deletion Tyr163delAT. FX F10 GI Factor X Factor 10 gene Gastrointestinal ICH PCC Intracranial hemorrhage Prothrombin complex concentrate From the University Greifswald, Institute of Human Genetics, Greifswald, and the Professor Hess Children s Hospital, Bremen, Germany; and the Children s Hospital of CCSS, and the University of Costa Rica, CIHATA, San Jose, Costa Rica. Submitted for publication Mar 22, 2004; last revision received Jul 14, 2004; accepted Aug 18, Reprint requests: No reprints. Please address correspondence to: Falko H. Hermann, University Greifswald, University Hospital, Institute of Human Genetics, Fleischmannstr , Greifswald, Germany /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

177 Table. Characterization of homozygous patients of the Gly380Arg substitution of the F10 gene Case-Number Sex m m m f m m FX:C,1,1,1,1,2,1 FX:antigen (FX:Ag),1,1,1 Symptoms: ICH no # days after birth GI hemorrhage Hemarthrosis Epistaxis Gingival hemorrhage Easy bruising Hematomas Hematuria The clinical findings of 6 homozygous patients for the missense mutation Gly380Arg are summarized in the Table. Family histories were negative for bleeding diathesis. The heterozygous parents were asymptomatic. Five of these 6 homozygous patients had an ICH soon after birth. Further bleeding symptoms were gastrointestinal (GI) hemorrhage, epistaxis, gingival hemorrhage, easy bruising, hematomas, and hematuria. The patients were treated with F IX complex concentrates (prothrombin complex concentrate [PCC]). The sequelae of ICH in patient 1 was blindness, deafness, and learning disability. Patient 2 has partial paralysis, and he attends a special school. Patient 3 also is in a special school. Patient 4 has seizures and developmental delay. Patient 5 had hydrocephalus and died at 15 months of age. Patient 6 has not had severe problems, but he was diagnosed early and was treated with PCCs. The mutation Gly380Arg was recently described in Italy in a compound heterozygous patient without bleeding problems. 7 We have detected this mutation only in Costa Rica. 4 This is the first description of homozygosity for the Gly380Arg mutation in a patient with FX deficiency. The haplotype of Arg380 alleles is in all cases the same for the analyzed polymorphisms (A1, B1, C2, E1, G1, H1, D1), indicating that the mutation is probably of the same origin. The prevalence of 6 homozygous patients for Gly380Arg in Costa Rica seems to result from a founder effect. All these 6 homozygous patients have severe FX deficiency, and 5 had an ICH. In human FXa-protein Gly380 is located immediately after the active site 379. As shown by molecular modeling, the substitution of nonpolar Gly with the polar Arg leads to the formation of new hydrogen bonds with Ala234. The very large side chain of Arg causes alteration of the protein folding and could lead to secretion problems, 7 which would explain the phenotype. In vitro expression for the mutation is currently under investigation. The girl with the novel homozygous deletion Tyr163delAT has FX:C activity of <1% (FX:Ag <1%). Five days after delivery she had an ICH. She now has slight motor problems but attends regular school and does not have learning disabilities. At 4 years of age she had a GI hemorrhage. Now she has heavy menstrual periods for which she is treated with plasma concentrates at a high dosage. The novel deletion Tyr163delAT interrupts the reading frame and leads to a stop codon 163 (TAT CAT>TCAT) in exon 6 of the F10 gene. This premature termination of the FX protein resulted in severe FX deficiency. This mutation probably increases the risk of ICH. However, this is the only reported homozygous case of this mutation. Patients with severe FX deficiency are defined as those who have spontaneous and/or life-threatening bleeding episodes, such as hemarthrosis, muscle hematomas, GI hemorrhage, or ICH. 8 The prevalence of ICH was 9% in a study of symptomatic patients from Iran (3/32). 2 In the few described case reports of FX deficiency with ICH, 9-11 the molecular defect of the F10 gene was not determined. Phenotype-genotype correlations for ICH in FX deficiency have not been performed previously. The prevalence of perinatal ICH in 5 of 6 FX deficient patients with the homozygous mutation Gly380Arg suggests the association of this symptom to this mutation. On the basis of the correlation found in this report, genetic counseling and prenatal diagnosis in the families affected by the Gly380Arg and Tyr163delAT mutations are recommended. This work was supported in part by the German Federal Ministry for Education and Research (NBL3 program, reference ) and by the German Academic Exchange Service (DAAD). REFERENCES 1. Uprichard J, Perry DJ. Factor X deficiency. Blood Rev 2002;16: Peyvandi F, Mannucci PM, Lak M, Abdoullahi M, Zeinali S, Sharifian R, et al. Congenital factor X deficiency: spectrum of bleeding symptoms in 32 Iranian patients. Br J Hematol 1998;102: Homozygous Factor X Gene Mutations Gly380Arg And Tyr163delAT Are Associated With Perinatal Intracranial Hemorrhage 129

178 3. Wulff K, Herrmann FH. Twenty-two novel mutations of the factor VII gene in factor VII deficiency. Hum Mutat 2000;15: Herrmann FH, Wulff K, Lopaciuk S, Pollmann H. Two novel factor X mutations in severe factor X deficiency. Throm Haemost 2001;(suppl): P Millar DS, Elliston L, Deex P, Krawczak M, Wacey AI, Reynaud J, et al. Molecular analysis of the genotype-phenotype relationship in factor X deficiency. Hum Genet 2000;106: Wulff K, Ebener U, Wehnert Ch-S, Ward PA, Reuner U, Hiebsch W, et al. Direct molecular genetic diagnosis and heterozygote identification in X-linked Emery-Dreifuss muscular dystrophy by heteroduplex analysis. Dis Markers 1997;13: Vianello F, Lombardi AM, Boldrin C, Luni S, Girolami A. A new factor X defect (factor X Padua 3): a compound heterozygous between true deficiency (Gly(380)-Arg) and a abnormality (Ser(334)-Pro). Thromb Res 2001;104: Peyvandi F, Menegatti M, Santagostino E, Akhavan S, Upichard J, Perry DJ, et al. Gene mutations and three-dimensional structural analysis in families with severe factor X deficiency. Br J Haematol 2002; 117: Citak A, Ucsel R, Karabocuoglu M, Unuvar A, Uzel N. A rare cause of intracranial hemorrhage:factor X deficiency. Pediatr Emerg Care 2001;17: El Kalla S, Menon NS. Neonatal congenital Factor X deficiency. Pediatr Hematol Oncol 1991;8: Sumer T, Ahmad M, Sumer NK, Al-Mouzan MI. Severe congenital factor X deficiency with intracranial haemorrhage. Eur J Pediatr 1986;145: Herrmann et al The Journal of Pediatrics January 2005

179 PERMANENT NEONATAL DIABETES IN AN ASIAN INFANT J. R. PORTER, MBBS, N. J. SHAW, MBCHB, T. G. BARRETT, MBBS, PHD, A. T. HATTERSLEY, BMBCH, DM, S. ELLARD, PHD, AND A. L. GLOYN,D.PHIL We describe a novel homozygous missense glucokinase mutation (R397L) resulting in insulin-treated neonatal diabetes in an infant from a consanguineous Asian family. Both parents were heterozygous for R397L and had mild hyperglycemia. Glucokinase mutations should be considered in infants of all ethnic groups with neonatal diabetes and consanguinity. (J Pediatr 2005;146:131-3) Neonatal diabetes has been defined as hyperglycemia requiring insulin treatment that presents within the first 3 months of life. 1 It is a rare condition with an estimated incidence of 1 in 400,000 live births in the UK population. 1 About half of all infants with neonatal diabetes will have transient neonatal diabetes, which remits within a median of 3 months but often recurs in adolescence and is linked in the majority of cases to abnormalities in the chromosome 6q region. 2 The remainder of infants will have permanent neonatal diabetes (PNDM). In the majority of cases, there is no known cause, but approximately one third of cases may be due to KCNJ11 activating mutations, 3 and 5 cases of PNDM have been shown to be due to homozygous or compound heterozygous inactivating mutations in the glucokinase (GCK) gene (gene name GCK). 4,5 In the heterozygous form, GCK mutations cause maturity-onset diabetes of the young (GCK-MODY). 6 Four of the five GCK-associated PNDM (GCK- PNDM) cases to date have been in consanguineous Arabic or European families. CASE STUDY A growth-retarded (weight, 1.76 kg; 14 standard deviation scores weight for age and sex 7 ) female infant (see Figure, subject IV:1) was born at 36 weeks gestation to a healthy, primigravid, British-born mother (Figure, subject III:2) from a Pakistani family. There was no history of gestational diabetes. The father (Figure, subject III:1) was also a British-born Pakistani, healthy, and was the mother s first cousin. Glucose monitoring revealed hyperglycemia, with a venous glucose level of 14.7 mmol/l (265 mg/dl) on the first day of life. Hyperglycemia persisted (capillary glucose, 11 to 15 mmol/l [198 to 270 mg/dl]) without acidosis or ketosis despite endogenous insulin secretion (insulin, 82.5 pmol/l [11.5 mu/l] at glucose, 14.7 mmol/l [265 mg/dl]), but the baby failed to gain weight by day 7, and insulin (Insulatard, Novo Nordisk, Bagsvaerd, Denmark) was commenced. The baby remained insulin treated with a gradually increasing insulin requirement over the first year of life. Fructosamine at 4 months was 420 mmol/l (normal range, 200 to 285 mmol/l), and DCCT aligned HbA1c at 8 months was 10.2%, suggesting inadequate control. At 13 months, the infant is healthy and thriving (9 th centile for weight appropriate for parents), on a dose of 1 U/kg per day of Human Mixtard 20 (Novo Nordisk, Bagsvaerd, Denmark) mixed insulin with HbA1c 9.9%. Cytogenetic investigation and methylation status of chromosome 6q were normal. Islet cell antibodies were negative. There was an extensive family history of type 2 diabetes without complications (Figure), and during this time, the infant s first cousin (subject IV:3) presented at age 14 years with impaired glucose tolerance (venous fasting glucose, 5.8 mmol [104 mg/dl], 2-hour glucose, 8.4 mmol/l [151 mg/dl]). This suggested a family history of GCK- MODY. Testing of the infant s parents revealed fasting plasma glucose levels >5.5 mmol/l [99 mg/dl], consistent with GCK-MODY 8 (subject III:2, 5.8 mmol/l [104 mg/dl] capillary; subject III:1, 6.1 mmol/l [110 mg/dl] venous). The aim of this study was to identify the genetic cause of neonatal diabetes in this infant. We hypothesized that PNDM in this infant could be due to a homozygous mutation in the GCK gene. Direct sequencing MODY Maturity-onset diabetes PNDM Permanent neonatal diabetes From the Diabetes Home Care Unit, Birmingham Children s Hospital, Birmingham, United Kingdom; and the Institute of Biomedical and Clinical Science, Peninsula Medical School, Exeter, United Kingdom. Supported by Diabetes UK RD Lawrence Fellowship (ALG), a Diabetes UK Clinical Training Fellowship (JRP), and Wellchild. Submitted for publication May 7, 2004; last revision received Jul 29, 2004; accepted Sep 10, Reprint requests: Dr John Porter, Diabetes Home Care Unit, Birmingham Children s Hospital, Steelhouse Lane, Birmingham B4 6NH, UK. j.porter@bham.ac.uk /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

180 Table. Summary of clinical characteristics of published cases of GCK-PNDM GCK mutation Birth weight (kg) [centile] Gestation (wk) Age at diagnosis (d) Initial plasma glucose (mmol/l) [mg/dl] Initial plasma C-peptide (pmol/l) Maximum insulin treatment (U/kg per d) M210K/M210K 1.67 [,3] [303] Undetectable 1.1 T228M/T228M 1.65 [,3] [713] Low 1.4 A378V/A378V 1.55 [10] [241] Undetectable 0.6 IVS812/IVS [,3] [1027] N/A N/A IVS812/G264S 1.87 [,3] [216] N/A 0.9 R397L/R397L 1.76 [,3] [265] Table adapted from Njølstad et al. 5 N/A, data not available. Figure. Partial pedigree showing family with the R397L mutation. Male subjects are shown by squares and female subjects by circles, with diagonal line to indicate deceased family members. Solid black symbols represent PNDM; half-filled black symbols represent GCK-MODY. Mutation status is denoted by M for mutant allele and N for normal allele. Individuals tested were either homozygous (MM) or heterozygous (NM) for the R397L mutation. DNA was not available from the remaining family members whose affection status is shown by unfilled symbols (unaffected) or gray symbols (diabetes mellitus with the exception of individual IV:3, who has impaired glucose tolerance). of the entire coding region and intron-exon boundaries of the GCK gene on an ABI 3100 capillary sequencer (Applied Biosystems, Warrington, United Kingdom) identified a novel homozygous missense mutation (R397L) in exon 9 in the proband. Sequencing of exon 9 in both parents demonstrated that they were both heterozygous for this mutation. DISCUSSION We describe an Asian child with a homozygous mutation in the GCK gene causing neonatal diabetes. The study of this family further emphasizes that GCK mutations should be considered as a cause of neonatal diabetes in consanguineous families. GCK is a key regulatory enzyme in glycolysis in the pancreatic b-cell. Because of its uniquely low affinity for glucose and low sensitivity to feedback from its product (glucose-6-phosphate), GCK forms the rate-limiting step in glycolysis. This allows GCK to function as a glucose sensor, which, in the normal individual, triggers insulin release at a plasma glucose level of ~5 mmol/l [90 mg/dl]. Heterozygous inactivating GCK mutations increase the threshold for glucose-stimulated insulin release, but compensation by the wild-type allele limits this to a threshold of 6 to 8 mmol/l [108 to 144 mg/dl]. In homozygous cases, the increase in the threshold for glucose-stimulated insulin release will be considerably higher and proportional to the severity of the GCK mutation. 9 The R397L mutation is a novel missense mutation that is very likely to cause reduced GCK function because it cosegregates with hyperglycemia in the family, is not found in 100 normal subjects, and residue R397 is conserved across species. Three missense mutations described previously in GCK-PNDM have been close to the glucose-binding site in GCK (M210K, T228M, A378V), with a fourth patient homozygous for a splice site mutation (IVS8 12T / G) predicted to result in a truncated protein and a fifth patient having a compound heterozygous mutation (IVS8 12T / G/ G264S). 4,5 These mutations are associated with a severe phenotype with negligible insulin production at birth 5 (see the Table). In this current case, the infant had a significant insulin production shortly after birth, as shown by the measured insulin level and failure of the glycemia to deteriorate above 15 mmol/l without treatment. The present insulin dose (1 U/kg) is similar to that seen when there is full insulin replacement. This milder phenotype may indicate that the R397L mutation causes a less severe kinetic defect than the previously reported mutations. Functional studies and measurement of C-peptide in this infant are required to support this hypothesis. GCK-PNDM is not a common cause of PNDM in European populations, 10,11 and initial studies suggest that dominant activating mutations in KCNJ11, which are frequently spontaneous, are more common. 3 All 6 cases of GCK-PNDM described have been associated with some degree of consanguinity. This would fit with the increased prevalence of recessive disorders in the offspring of consanguineous marriages. Identification of these infants is important for genetic counselling in further pregnancies and for advising on treatment in heterozygous family members with 132 Porter et al The Journal of Pediatrics January 2005

181 GCK-MODY. We suggest that GCK-PNDM should be considered in an infant with PNDM, growth restriction, and a family history of consanguinity in all ethnic groups, and as a first investigation, the parents fasting glucose should be checked even if there is not a clear history of diabetes. The authors acknowledge the cooperation of the family involved in this study and the help of the staff of Birmingham Children s Hospital and City Hospital Diabetes Homecare Unit and Exeter Molecular Genetics Laboratory. REFERENCES 1. Shield JP, Gardner RJ, Wadsworth EJ, Whiteford ML, James RS, Robinson DO, et al. Aetiopathology and genetic basis of neonatal diabetes. Arch Dis Child Fetal Neonatal Ed 1997;76:F Temple IK, Shield JP. Transient neonatal diabetes, a disorder of imprinting. J Med Genet 2002;39: Gloyn AL, Pearson ER, Antcliff JF, Proks P, Bruining GJ, Slingerland AS, et al. Activating mutations in the gene encoding the ATP sensitive potassium channel subunit Kir6.2 and permanent neonatal diabetes. N Engl J Med 2004;350: Njolstad PR, Sovik O, Cuesta-Munoz A, Bjorkhaug L, Massa O, Barbetti F, et al. Neonatal diabetes mellitus due to complete glucokinase deficiency. N Engl J Med 2001;344: Njolstad PR, Sagen JV, Bjorkhaug L, Odili S, Shehadeh N, Bakry D, et al. Permanent neonatal diabetes caused by glucokinase deficiency: inborn error of the glucose-insulin signaling pathway. Diabetes 2003;52: Hattersley AT, Turner RC, Permutt MA, Patel P, Tanizawa Y, Chiu KC, et al. Linkage of type 2 diabetes to the glucokinase gene. Lancet 1992; 339: Cole TJ, Freeman JV, Preece MA. Body mass index reference curves for the UK, Arch Dis Child 1995;73: Stride A, Vaxillaire M, Tuomi T, Barbetti F, Njolstad PR, Hansen T, et al. The genetic abnormality in the beta cell determines the response to an oral glucose load. Diabetologia 2002;45: Gloyn AL. Glucokinase (GCK) mutations in hyper- and hypoglycemia: maturity-onset diabetes of the young, permanent neonatal diabetes, and hyperinsulinemia of infancy. Hum Mutat 2003;22: Vaxillaire M, Samson C, Cave H, Metz C, Froguel P, Polak M. Glucokinase gene mutations are not a common cause of permanent neonatal diabetes in France. Diabetologia 2002;45: Gloyn AL, Ellard S, Shield JP, Temple IK, Mackay DJ, Polak M, et al. Complete glucokinase deficiency is not a common cause of permanent neonatal diabetes. Diabetologia 2002;45:290. Permanent Neonatal Diabetes In An Asian Infant 133

182 PEDIATRIC HYPEREOSINOPHILIC SYNDROME (HES) DIFFERS FROM ADULT HES HARY T. KATZ, MD, S. JAHARUL HAQUE, PHD, AND FRED H. HSIEH, MD The idiopathic hypereosinophilic syndrome (HES) developed in a 15-year-old boy who presented with colitis, cough, rash, and hepatitis. Molecular analysis failed to demonstrate the Fip1-like1-Platelet Derived Growth Factor Receptor a chain (FIP1L1-PDGFRA) mutation described in adult patients with HES. There are significant clinical differences between the pediatric and adult presentations of HES. (J Pediatr 2005;146:134-6) Idiopathic HES is a rare disorder characterized by eosinophilia >1500/mm 3 for longer than 6 months, no evidence of known causes of eosinophilia, and evidence of organ involvement. 1 Virtually any organ can be involved with HES. This disorder most commonly occurs between 20 and 50 years of age and is more common among males than females. We report a case of pediatric HES, describe our patient s clinical outcome, and highlight differences between pediatric and adult HES based on a literature review. CASE REPORT A 15 year-old Caucasian male presented with abdominal pain, diarrhea, and a 10-lb weight loss. Colonoscopy revealed neutrophilic infiltration of the colonic mucosa with crypt abscesses in the epithelium and lamina propria consistent with chronic active colitis. Stool was positive for Clostridium difficile toxin, attributed to chronic minocycline use for acne, and was negative for ova and parasites. A complete blood count revealed an absolute eosinophil count of 1890 (normal <400/mm 3 ). The patient was treated with metronidazole and was asymptomatic for 6 months. He then developed a nonproductive cough, night sweats, and a diffuse pruritic, papular rash. Computed tomography of the chest showed small peripheral pulmonary nodules and ground glass opacities. A complete blood count revealed an absolute eosinophil count of 52,000/mm 3. Additional lab studies revealed an IgE level of 8561 (7-110 U/mL), alkaline phosphatase of 1149 ( U/mL), g-glutamyl transpeptidase of 193 (0-50 U/mL), serum tryptase of 4.7 ( mg/l), and B12 level of 475 ( pg/ml). Ultrasonography of the liver showed a diffusely abnormal parenchymal pattern with echogenic linear radiating bands of density and dilated bile ducts in the left lobe. Echocardiography was normal. Molecular analysis of the patient s peripheral blood for the Fip1-like1-Platelet Derived Growth Factor Receptor a chain (FIP1L1-PDGFRA) fusion tyrosine kinase associated with HES in 9 of 16 adults 2 was negative. Open lung biopsy showed patchy interstitial and intra-alveolar inflammation with a predominance of eosinophils. Skin biopsy showed acute neutrophilic folliculitis with perivascular dermatitis with eosinophils. Neither biopsy revealed evidence of vasculitis. Bone marrow biopsy demonstrated a hypercellular marrow with predominantly eosinophils and no blasts, consistent with idiopathic HES. No bone marrow mastocytosis or morphologically abnormal mast cells were noted. Karyotype was normal. Physical exam at that time was significant for erythematous, pruritic papules on the skin, a normal cardiac exam, no splenomegaly, and no neurologic abnormalities. The patient was started on prednisone 60 mg orally daily. One week later his absolute eosinophil count was 110/mm 3. Tapering prednisone to <10 mg orally daily led to consistent recurrence of eosinophilia and symptoms. Therapy with imatinib mesylate (Gleevec, Novartis, Basel, Switzerland) starting at 100 mg orally daily and increased to 300 mg orally daily was not effective in controlling eosinophilia in the absence of prednisone. FIP1L1-PDGFRA Fip1-like1-Platelet Derived Growth Factor Receptor a chain HES Hypereosinophilic syndrome From the Department of Pulmonary, Allergy, and Critical Care Medicine and the Department of Cancer Biology, Cleveland Clinic Foundation, Cleveland, Ohio. Supported by NIH grants HL , CA , and GM Submitted for publication Apr 28, 2004; last revision received Aug 27, 2004; accepted Sep 13, Reprint requests: Dr Fred H. Hsieh, Cleveland Clinic Foundation, Pulmonary, Allergy, and Critical Care Medicine, Allergy and Immunology Section, 9500 Euclid Avenue C22, Cleveland, OH hsiehf@ccf.org /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

183 Table I. Demographic information, presenting symptoms, and organ system involvement of 38 cases of pediatric HES Number DISCUSSION Percent Age Mean = 8.2 years Range = 1 16 years Sex Male Female Presenting symptoms * Fever Arthralgias Fatigue Rash Cough Neurologic symptoms Dyspnea Diarrhea Abdominal pain Vomiting Headache Sore throat Organs involved Heart Lungs Skin Nervous system Gastrointestinal Associated with ALL Chromosomal abnormality ALL, acute lymphoblastic leukemia. *Data not available for 4 patients. A search of the literature for pediatric hypereosinophilic syndrome or child hypereosinophilic syndrome revealed 38 cases reported in the English language (Table I). 1,4-23 Pediatric HES has only a slight male predominance (55.3% male vs 44.7% female), whereas adult HES is reported to be more common among males than females, in a ratio of 9 to 1. 1 In the adult literature, the frequencies of symptoms found on presentation are: fatigue (26%), cough (24%), dyspnea (16%), rash (12%), and fever (12%). 3 We found that fever (58.8%), arthralgias (23.5%), and rash (23.5%) were more common presenting symptoms in pediatric cases. Of note, headache and sore throat were not uncommon presenting symptoms (11.8% each). As with adults, involvement of the cardiovascular system is the major source of morbidity and mortality. Pediatric HES is commonly associated with chromosomal abnormalities, including trisomy 8 and a translocation involving the abl oncogene. 5-9 In almost 40% of reported cases, pediatric HES has been associated with acute leukemia, especially acute lymphoblastic leukemia. 4-6,10-12 The goal of therapy in patients with HES is to lower the eosinophil count Table II. Treatment, response, and morbidity/ mortality of 38 cases of pediatric HES Treatment * Number Percent Response to treatment z Corticosteroids /28 responded as monotherapy or in conjunction with hydroxyurea, cyclophosphamide, vincristine, methotrexate, and/or cyclosporin Vincristine /9 responded in conjunction with prednisone, hydroxyurea, and/or 6-MP 6-mercaptopurine /4 responded in conjunction with vincristine Hydroxyurea /3 responded in conjunction with prednisone and vincristine ACTH /3 responded as monotherapy Busulfan /2 responded as monotherapy Methotrexate /2 responded in conjunction with prednisone and cyclosporine Cyclophosphamide /2 responded in conjunction with prednisone Cyclosporin /2 responded in conjunction with prednisone and methotrexate Interferon-a 1 3 1/1 responded as monotherapy Outcome y Alive Expired ACTH, adrenocorticotropic hormone; 6-MP, 6-mercaptopurine. *Data not available for 5 patients. ydata not available for 2 patients. Numbers reported are at the time of publication. Of the 21 expired patients, no data were available for 3 patients in regards to survival time. The mean length of survival was 10.6 months from the time of diagnosis in the remaining 18 patients (range months). zresponse defined as patient alive at time of publication. and prevent organ dysfunction. Corticosteroids are the initial treatment of choice. For children unresponsive to steroids, agents such as hydroxyurea, vincristine, mercaptopurine, and, recently, interferon-a have been used 13,14,30 (Table II). Pediatric Hypereosinophilic Syndrome (HES) Differs From Adult HES 135

184 Monoclonal anti-il-5 antibody therapy may be an important therapeutic option, although it has not yet been tested in pediatric HES A novel tyrosine kinase created by the fusion of the FIP1L1 and PDGFRA genes has been recently described in patients with HES. 2 This fusion protein is the result of an interstitial deletion on chromosome 4 and is a target of the tyrosine kinase inhibitor imatinib mesylate. Thus, a positive therapeutic response to imatinib is likely if this fusion protein is identified, although patients have responded to imatinib even if no fusion protein is identified. Patients with the FIP1L1-PDGFRA gene are likely to have the myeloproliferative variant of HES, as opposed to the lymphoproliferative (eg, T-cell clone overproducing interleukin-5 leading to eosinophilia) or otherwise unclassified forms of the disease. 26 No pediatric case with the FIP1L1-PDGFRA fusion gene has been reported to date. Long-term prognosis has not been reported in pediatric HES. However, information was available for 36 of the 38 cases we reviewed; 15 patients were reported to be alive at the time their cases were published, whereas 21 had expired. Of these 21 patients, no data were available for 3 patients regarding survival. The mean length of survival was 10.6 months (range months) from the time of diagnosis in the remaining 18 patients. In the adult HES population, with earlier diagnosis and intensive follow-up, the survival of HES patients has improved over the years. In 1975 Chusid et al reported a mean survival of 9 months and a 3-year survival of 12%. 1 In a series of 40 patients with HES published in 1989, an 80% survival at 5 years and a 42% survival at 15 years was reported. 24 REFERENCES 1. Chusid MJ, Dale DC, West BC, Wolff SM. The hypereosinophilic syndrome: analysis of fourteen cases with review of the literature. Medicine (Baltimore) 1975;54: Cools J, DeAngelo DJ, Gotlib J, Stover EH, Legare RD, Cortes J, et al. A tyrosine kinase created by fusion of the PDGFRA and FIP1L1 genes as a therapeutic target of imatinib in idiopathic hypereosinophilic syndrome. N Engl J Med 2003;348: Fauci AS, Harley JB, Roberts WC, Ferrans VJ, Gralnick HR, Bjornson BH. The idiopathic hypereosinophilic syndrome: clinical, pathophysiologic and therapeutic considerations. Ann Intern Med 1982;97: Alfaham MA, Ferguson SD, Sihra B, Davies J. The idiopathic hypereosinophilic syndrome. Arch Dis Childhood 1987;62: Chilcote RR, Pergament E, Kretschmer R, Mikuta JC. The hypereosinophilic syndrome and lymphoblastic leukemia with extra C-group chromosome and q141 marker. J Pediatr 1982;101: Tono-oka T, Sato Y, Matsumoto T, Ueno N, Ohkawa M, Shikano T, et al. Hypereosinophilic syndrome in acute lymphoblastic leukemia with a chromosome translocation [t(5q;14q)]. Med Pediatr Oncol 1984;12: Michel G, Thuret I, Capodano AM, Scheiner C, Guitard AM, Mozziconacci MJ, et al. Myelofibrosis in a child suffering from a hypereosinophilic syndrome with trisomy 8: response to corticotherapy. Med Pediatr Oncol 1991;19: Egesten A, Hagerstrand I, Kristoffersson U, Garwicz S. Hypereosinophilic syndrome in a child mosaic for a congenital triplication of the short arm of chromosome 8. Br J Haematol 1997;96: Bakhshi S, Hamre M, Mohamed AN, Feldman G, Ravindranath Y. t(5;9)(q11;q34): a novel familial translocation involving Abelson oncogene and association with hypereosinophilia. J Pediatr Hematol Oncol 2003;25: Nelken RP, Stockman JA. The hypereosinophilic syndrome in association with acute lymphoblastic leukemia. J Pediatr 1976;89: Blatt J, Proujansky R, Horn M, Phebus C, Longworth D, Penchansky L. Idiopathic hypereosinophilic syndrome terminating in acute lymphoblastic leukemia. Pediatr Hematol Oncol 1992;9: Narayanan G, Hussain BM, Chandralekha B, Sivasankar C, Nair MK. Hypereosinophilic syndrome in acute lymphoblastic leukemia: case report and literature review. Acta Oncologica 2000;39: Marshall GM, White L. Effective therapy for a severe case of the idiopathic hypereosinophilic syndrome. Am J Pediatr Hematol Oncol 1989; 11: Sakamoto K, Erdreich-Epstein A, declerck Y, Coates T. Prolonged clinical response to vincristine treatment in two patients with idiopathic hypereosinophilic syndrome. Am J Pediatr Hematol Oncol 1992;14: Falade AG, Darbyshire PJ, Raafat F, Booth IW. Hypereosinophilic syndrome in childhood appearing as inflammatory bowel disease. J Pediatr Gastroenterol Nutr 1991;12: Kumar KA, Anjaneyulu A, Murthy JM. Idiopathic hypereosinophilic syndrome presenting as childhood hemiplegia. Postgrad Med J 1992;68: Takamizawa M, Iwata T, Watanabe K, Hayakawa H, Takahashi T, Furue M, et al. Elevated production of interleukin-4 and interleukin-5 by T cells in a child with idiopathic hypereosinophilic syndrome. J Allergy Clin Immunol 1994;93: Ishida Y, Hayashi M, Higaki A, Matsumoto K, Iikura Y, Ishikawa J, et al. Hypereosinophilic syndrome with generalized myasthenia gravis. J Pediatr 1996;128: Nadarajah S, Krafchik B, Roifman C, Horgan-Bell C. Treatment of hypereosinophilic syndrome in a child using cyclosporine: implication for a primary T-cell abnormality. Pediatrics 1997;99: Schulman H, Hertzog L, Zirkin H, Hertzanu Y. Cerebral sinovenous thrombosis in the idiopathic hypereosinophilic syndrome in childhood. Pediatr Radiol 1999;29: Eapen RS, Lemler MS, Ramaciotti C. Serial echocardiographic changes in a child with hypereosinophilia syndrome. Pediatr Cardiol 2001;22: Spiegel R, Zalman L, Gavriel H, Horovitz Y. Hypereosinophilic syndrome in a child presenting as eosinophilic pharyngitis. J Pediatr Hematol Oncol 2003;25: Scheinfeld N. Fatal fulminant hypereosinophilic syndrome with a petechial rash in a 16-year-old girl. J Pediatr Hematol Oncol 2003;25: Lefebvre C, Bletry O, Degoulet P, Guillevin L, Bentata-Pessayre M, Le Thi Huong D, et al. Prognostic factors of hypereosinophilic syndrome: study of 40 cases. Ann Med Interne (Paris) 1989;140: Griffin JH, Leung J, Bruner RJ, Caligiuri MA, Briesewitz R. Discovery of a fusion kinase in EOL-1 cells and idiopathic hypereosinophilic syndrome. Proc Natl Acad Sci USA 2003;100: Leiferman KM, Gleich GJ. Hypereosinophilic syndrome: case presentation and update. J All Clin Immunol 2004;113: Plotz SG, Simon HU, Darsow U, Simon D, Vassina E, Yousefi S, et al. Use of an anti-interleukin-5 antibody in the hypereosinophilic syndrome with eosinophilic dermatitis. N Engl J Med 2003;349: Garrett JK, Jameson SC, Thomson B, Collins MH, Wagoner LE, Freese DK, et al. Anti-interleukin-5 (mepolizumab) therapy for hypereosinophilic syndromes. J Allergy Clin Immunol 2004;113: Klion AD, Law MA, Noel P, Kim YJ, Haverty TP, Nutman TB. Safety and efficacy of the monoclonal anti-interleukin-5 antibody SCH55700 in the treatment of patients with hypereosinophilic syndrome. Blood 2004;103: Leblond P, Lepers S, Thebaud E, Mazingue F, Lambilliote A, Fournier M, et al. Idiopathic hypereosinophilic syndrome: case report in an infant. Arch Pediatr 2004;11: Katz, Haque, and Hsieh The Journal of Pediatrics January 2005

185 LONG-TERM SURVIVAL IN SEVERE COMBINED IMMUNE DEFICIENCY: THE ROLE OF PERSISTENT MATERNAL ENGRAFTMENT ILHAN TEZCAN,MD,PHD, FUGEN ERSOY,MD,OZDEN SANAL,MD,TUBA TURUL,MD,DUYGU UCKAN,MD,SEVIM BALCI,MD, GONUL HICSONMEZ,MD,M.PRIEUR,MD,S.CAILLAT-ZUCMANN,MD,FRANCOIS LE DEIST,MD,AND GENEVIEVE DE SAINT BASILE,MD Two siblings with severe combıned immune defıcıency, one with maternal engraftment and detectable immunologic functions who was alive at the age of 8 years are presented. Both patients had the same JAK3 gene mutation, suggesting that maternal engraftment may result in immune competence leading to long-term survival in patients with severe combıned immune defıcıency. (J Pediatr 2005;146:137-40) Severe combined immunodeficiency (SCID) is a heterogeneous group of primary immunodeficiency disorders characterized by deficiency of both T- and B-lymphocyte development or function. 1 SCID usually presents with severe and persistent infections of the respiratory system and diarrhea, starting within the first months of life, which ultimately lead to failure to thrive. Unless hematopoietic stem cell transplantation is performed, the patients die in infancy. Patients with SCID may be engrafted with transplacentally derived maternal T-lymphocytes, which are not rejected by the immunodeficient fetus. 2 However, since the engrafted T cells are not functionally competent, the engraftment does not usually affect the course of the disease. In this report, we present a case of a patient with atypical SCID with detectable immunologic functions as the result of maternal engraftment. The patient was diagnosed as JAK3 deficiency at the age of 8 years after the diagnosis of his 6-month-old brother with the same genetic defect but carrying the typical features of SCID. CASE REPORT An 8-year-old-boy was first referred to our hospital when he was 4 years old because of recurrent oral aphtous lesions, oral moniliasis, upper and lower respiratory tract infections, chronic diarrhea, and persistent skin lesions. He was previously hospitalized four times because of lower and upper respiratory infections since the age of 1.5 years. Each time, he responded well to antibiotic treatment. Oral moniliasis first appeared at 1 month of age and recurred frequently. The parents were first cousins. One female and one male child died in infancy of undefined causes, and a 4-month-old boy died with oral thrush and oral aphtous lesions (Figure). The patient had a weight of 19 kg (3rd centile) and a height of 116 cm (3rd to 10th centile). On physical examination, seborrheic dermatitis was noted on the scalp and punctiform hyperkeratotic lesions, defined as follicular hyperkeratotic dermatitis distributed widely on the skin. He had detectable tonsillary tissue and microlymphadenopathy. The liver was palpable 4 cm below the right costal margin. Immunologic findings are given in Table I. The expression of the Vb family by CD41 and CD81 cells revealed a restricted TCR repertoire, especially within the CD8 subset (Table II). The patient recently had development of myelodysplastic syndrome with progressive leukopenia. Dysplasia of all three cell lineages, atypical monocytosis, and bone marrow findings, together with spontaneous cell proliferation, led to the diagnosis of chronic myelomonocytic leukemia. He received hematopoietic stem cells from his mother after chemotherapy and conditioning regimen. He died at 55 days of central nervous system complications with evidence of donor engraftment. When the patient (patient 1) was 8 years old and under evaluation for possible diagnosis of immunodeficiency, his 6-month-old brother (patient 2) presented with fever, recurrent cough, moniliasis, and failure to thrive and had characteristic features of SCID GVHD Graft versus host disease SCID Severe combined immune deficiency From the Immunology Unit, the Hematology Unit, and the Medical Genetics Unit, Hacettepe University, Ihsan Dogramaci Children s Hospital, SIhhiye, Ankara, Turkey; and the Histology, Embryology, and Cytogenetic Unit, the Clinical Immunology Unit, and INSERM U 429, Necker- Enfants Malades Hospital, Paris, France. Submitted for publication Apr 9, 2003; last revision received Jul 7, 2004; accepted Sep 13, Reprint requests: Dr Ilhan Tezcan, Hacettepe University Medical School, Ihsan Dogramaci Children s Hospital, Immunology Department, 06100, SIhhiye, Ankara, Turkey /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds

186 (Table I). A diagnosis of SCID was thus entertained in both patients, with a potential for maternal engraftment in the older one. Lymphocyte chromosome analysis of patient 1 identified a 46XX karyotype. HLA typing was performed both on fibroblasts from a skin biopsy specimen and on peripheral blood lymphocytes in patient 1. Figure. Pedigree of the family. HLA Typing The HLA typing was as follows: Mother: A2, B51, DR11, DQ03/A24, B8, DR14, DQ05; Father: A2 B8, DR11, DQ03/A24, B18, DR03, DQ02; Pt 1 fibroblast: A2, B51, DR11, DQ03/A24, B18, DR11, DQ03; Pt 1 lymphocyte: A2, Table I. Immunologic findings of siblings Patient 1 Immunologic investigations Normal values Age-matched Patient 2 Normal values Age-matched Lymphocytes (/mm 3 ) ,500 IgG (mg/dl) IgA (mg/dl) IgM (mg/dl) IgE (IU/L) Antibody responses Patient 1 Patient 2 Normal values Isohemagglutinins 1/8 Not tested.1/32 Polio antibody 1.1/4.1/40 Polio antibody 2.1/4.1/40 Polio antibody 3.1/4.1/40 Lymphocyte subsets Percentage (%) Patient 1 Patient 2 Normal values * CD CD CD CD CD In vitro lymphocyte proliferation test Cpm * Patient 1 Patient 2 Control PHA 17.7 Absent ConA CD PMA 1 I *cpm, counts per minute. 138 Tezcan et al The Journal of Pediatrics January 2005

187 Table II. Vb repertoire analysis of patient 1 Phenotype among lymphocyte subsets in patient 1 CD3 CD4 % CD8 Patient Normal values Patient Normal values Patient Normal values * CD45RO CD45RA TcR a/b TcR g/d Vb ± ± 1.8 Vb ± ± 1.6 Vb ± ± 3.0 Vb ± ± 1.8 Vb ± ± 1.7 Vb ± ± 1.9 Vb ± ± 1.2 *Normal values are age-matched. B51, DR11, DQ03/A24, B8, DR14, DQ05; Pt 2 lymphocyte: A24, B18, DR03, DQ02/A24, B8, DR14, DQ05. Genetic Analysis The SCID phenotype of patient 2 and consanguinity in the family was suggestive of an autosomal recessive inheritance. Polymorphic marker analysis and JAK3 sequencing were performed on the fibroblast cell line in patient 1 and on whole blood cells in patient 2. Polymorphic marker analysis of peripheral blood lymphocytes from patient 1 showed maternal haplotype segregation. Sequencing of the JAK3 coding sequence identified a homozygous mutation C2236T leading to T714M in both patients. This missense mutation was localized in the JAK homology 2, (JH2) domain, the kinaselike domain of the protein. Polymorphic marker segregation analysis showed that both patients had the homozygous haplotype at the JAK3 locus on chromosome 19p13.2. DISCUSSION It is known that the human placenta is an incomplete barrier allowing bidirectional passage of blood cells. Maternal cells can be detected as microchimerism in 42% of cord blood samples from healthy newborn infants. 3,4 However, survival of maternal cells is usually limited by effective rejection by the immunocompetent host. Incidence of maternal engraftment was reported in 24% to 40% of patients with SCID. 5,6 In general, patients with maternally engrafted SCID present with severe infections in the early months of life and have cutaneous manifestations. Although our patient had some recurrent infections, he survived up to 8 years of age with normal levels of immunoglobulins, isohemagglutinins, and anti-polio type I antibody response. This case with JAK3-deficient SCID survived for 8 years with some immunologic functions due to maternally transferred T cells. Although further evaluation of maternal engraftment in different subpopulations could not be performed, given the normal presence of host B-cell population in JAK3-deficient patients, engraftment of maternal B cell is not expected. Similarly, maternal NK cell engraftment is unlikely in patient 1, since the number of NK cells was very low and not significantly different from the number observed in patient 2 with the classic SCID phenotype. The underlying mechanism is probably the HLA compatibility of the mother and the patient, due to parental consanguinity, which may have played an important role in the T, B, and antigen presenting cell cooperation. In patient 1, naive T cells (CD45RA) were almost undetectable in the CD4 population, whereas 10% of CD8 cells were expressing CD45RA. Even though it is possible that this CD45RA1 population may represent thymopoiesis in the patient, considering the fact that the patient does not have its own T cells and has a restricted Vb repertoir pattern, this is highly unlikely. In human beings, differentiation of naive, effector, and memory cells in the CD8 population is highly fragmentary. Absence of CD27 or CD28 on CD45RA cells is indicative of effector CD8 cells, which respond poorly to most stimuli but are actually primed, having features of past antigen stimulation, in contrast to naive CD81CD271 population. 7 Although the absence of CD27 on this CD8145RA1 population was not evaluated in patient 1, we think that this population may still correspond to effector CD8 cells instead of naive type. Also, the maternally engrafted CD8 population has been shown to display reduced proliferation to mitogens, which is also a feature of effector cells, as in our case. 5 Skin manifestation of graft versus host disease (GVHD) has been reported in 40% to 83% of engrafted patients. 5,8 Our patient had the chronic eczematous dermatitis form of skin GVHD, with a late onset at 4 years of age, but lacked other GVHD manifestations such as hepatic involvement, nephritis, Long-Term Survival In Severe Combined Immune Deficiency: The Role Of Persistent Maternal Engraftment 139

188 or hematologic manifestations. Similar to our case, Müller et al reported a higher incidence of maternal T-cell engraftment in the B1 SCID group, with a milder eczematous form of skin GVHD. 5 In conclusion, our study shows that in rare cases, maternal engraftment may provide the required immune competence and result in prolonged survival in patients with SCID. REFERENCES 1. Fischer A, Cavazzana-Calvo M, De Saint Basile G, De Villartay JP, Di Santo JP, Hivroz C, et al. Naturally occuring deficiencies of the immune system. Annu Rev Immunol 1997;15: Conley ME, Nowell PC, Henle G, Douglas SD. XX T cells and XY B cells in two patients with severe combined immunodeficiency. Clin Immunol Immunopathol 1984;31: Lo Y, Lo ESF, Watson N, Noakes L, Sargent IL, Thilaganathan B, et al. Two-way cell traffic between mother and fetus: biologic and clinical implications. Blood 1996;88: Lo Y, Lau T, Chan L, Leung T, Chang A. Quantitative analysis of the bi-directional fetomaternal transfer of nucleated cells and plasma DNA. Clin Chem 2000;48: Müller SM, Ege M, Pottharst A, Schulz AS, Schwarz K, Friedrich W. Transplacentally acquired maternal T lymphocytes in severe combined immunodeficiency: a study of 121 patients. Blood 2001;98: Thompson LF, O Connor RD, Bastian JF. Phenotype and function of engrafted maternal T cells in patients with severe combined immunodeficiency. J Immunol 1984;133: Hamann D, Baars PA, Rep MHG, Hooibrink B, Kerkhof-Garde SR, Klein MR, et al. Phenotypic and functional seperation of memory and effector human CD81 T cells. J Exp Med 1997;186: Denianke KS, Frieden IJ, Cowan MJ, Williams ML, McCalmont TH. Severe combined immunodeficiency cutenous manifestations of maternal engraftment in patients with severe combined immunodeficiency: a clinicopathologic study. Bone Marrow Transplant 2001;28: Tezcan et al The Journal of Pediatrics January 2005

189 INSIGHTS PROXIMAL FOCAL FEMORAL DEFICIENCY A RARE ENTITY IN THE SONOGRAPHIC DIFFERENTIAL DIAGNOSIS OF DEVELOPMENTAL DYSPLASIA OF THE HIP Figure 1. A, Ultrasound image of a proximal focal femoral deficiency in the left hip of a 4-month-old infant. Two key anatomical landmarks, the acetabular labrum and lower edge of the ilium, cannot be clearly identified. The cartilage-bone interface also cannot be visualized. B, Schematic representation of Figure 1A. Figure 2. Anteroposterior pelvic radiograph of the hip shown in Figures 1A and B confirms the diagnosis of KALAMACHI type II proximal focal femoral deficiency. As can be expected at this age, the femoral head cannot be visualized. Findings include: (1) clear lateralization of the femur, dislocation of the hip, and a supraacetabular bump; (2) increased diameter of the obturator foramen; (3) horizontal roof of the acetabulum with no signs of acetabular dysplasia; and (4) thickened teardrop shadow. We will describe the differential diagnosis of the rare hip anomaly proximal focal femoral deficiency (PFFD) based on an analysis of more than 10,000 ultrasound images of the hips examined in our orthopedic clinics from 1988 to The sonographic examinantion was performed according to GRAF. 1 This ultrasound procedure has become an established method for detection of developmental dysplasia of the hip (DDH). 2,3 Two cases of PFFD were detected during the investigation period. Both neonates had been referred to the orthopedic clinic for further evaluation after the initial ultrasound findings suggested an anomaly in the coxal-femoral region. Correct identification of anatomical structures in the acetabular region (acetabular labrum and lower edge of the ilium) was not possible, but the iliac line, femoral head, and greater trochanter could be reliably visualized (Figure 1, A-B). On the healthy other side the anatomic structures could be clearly identified by ultrasound examination. Because of these findings, a deformity in the coxal-femoral region was suspected and further diagnosis was recommended. The KALAMACHI type II PFFD 4 was subsequently diagnosed (Figure 2), and specific treatment was initiated. PFFD should be suspected in any neonate with suspicious clinical findings and failure of ultrasound to clearly visualize anatomical landmarks in the region of the acetabulum (acetabular labrum, lower edge of the ilium, cartilage-bone interface) despite the use of a sufficient imaging technique. However, ultrasound examination cannot confirm the specific diagnosis of PFFD. When evaluating these cases by ultrasound, another problem to be considered is the fact that PFFD can be mistaken for type IV hip dysplasia. Radiographs should then be obtained to confirm or refute the tentative diagnosis. Infants with PFFD can then receive proper treatment without unnecessary therapeutic trial and error. Reprint requests: Ralph Kayser, MD, Department of Trauma and Reconstructive Surgery, Charite - University Medicine Berlin, Campus Benjamin Franklin, Hindenburgdamm 30, D Berlin, Germany. Phone: , Fax: Kayser-Berlin@t-online.de J Pediatr 2005;146: /$ - see front matter Copyright ª 2005 Elsevier Inc. All rights reserved /j.jpeds Ralph Kayser, MD Konrad Mahlfeld, MD Henning Grasshoff, MD Orthopedic Clinic, Otto von Guericke University Magdeburg, Germany Harry Rudolph Merk, MD Orthopedic Clinic, Ernst Moritz Arndt University Greifswald, Germany References are available online at 141

190 CURRENT BEST EVIDENCE CLINICAL RESEARCH ABSTRACTS FOR PEDIATRICIANS EDITOR S NOTE: Journals reviewed for this issue: Archives of Disease in Childhood, Archives of Pediatrics and Adolescent Medicine, British Medical Journal, Journal of the American Medical Association, Journal of Pediatrics, The Lancet, New England Journal of Medicine, Pediatric Infectious Diseases Journal, and Pediatrics. Gurpreet K. Rana, BSc, MLIS, Taubman Medical Library, University of Michigan, contributed to the review and selection of this month s abstracts. John G. Frohna, MD, MPH Efficacy and safety of acetaminophen versus ibuprofen for treating children s pain or fever: A meta-analysis Perrott DA, Piira T, Goodenough B, Champion GD. Arch Pediatr Adolesc Med 2004;158: Context Acetaminophen and ibuprofen are used frequently in children, yet there is uncertainty about their relative safety and efficacy. Objectives To determine which of the two drugs is more efficacious in treating pain and which is more effective in reducing temperature in febrile children and to compare their safety. Design A meta-analysis of all studies in which single-dose acetaminophen or ibuprofen were measured in a randomly blinded way in children younger than 18 years old. Main outcome measures For pain relief: risk ratio for achieving more than 50% of maximum pain relief; for fever: effect size for febrile temperature reduction; for safety: riskratio of minor (no withdrawal) and major (withdrawal) adverse effects from treatment. Study identification Searches of online databases through May 2002 with identification of randomized controlled trials of acetaminophen and ibuprofen for pediatric pain and fever. Studies reviewed Of 127 studies identified, 17 met the criteria and were used for safety analysis; 3 were used for the pain analysis and 10 for the fever analysis. Results Ibuprofen (4 to 10 mg/kg) and acetaminophen (7 to 15 mg/kg) showed comparable efficacy (3 pain relief trials, 186 children). The risk ratio point estimates were 1.14 (95% CI, 0.82 to 1.58) at 2 hours after receiving the dose and 1.11 (95% CI, 0.89 to 1.38) at 4 hours. Ibuprofen (5 to 10 mg/kg) reduced temperature more than acetaminophen (10 to 15 mg/kg) at 2, 4, and 6 hours after treatment (respective weighted-effect sizes: 0.19 [95% CI, 0.05 to 0.33], 0.31 [95% CI, 0.19 to 0.44], and 0.33 [95% CI, 0.19 to 0.47]) (9 fever trials; 1078 children). For 10 mg/kg ibuprofen (acetaminophen, 10 to 15 mg/kg), corresponding effect sizes were 0.34 (95% CI, 0.12 o 0.56), 0.81 (95% CI, 0.56 to 1.03), and 0.66 (95% CI, 0.44 to 0.87). There was no evidence the drugs differed from each other (or placebo) in incidence of minor or major harm (17 safety trials; 1820 children). Conclusions In children, lower doses of ibuprofen (4 to 10 mg/kg) were as effective as acetaminophen (7 to 15 mg/kg) for the management of pain. Ibuprofen (5 to 10 mg/kg) was a more effective antipyretic than acetaminophen (10 to 15 mg/kg) at 2, 4, and 6 hours after treatment. There was no difference in safety between the two drugs or between each one and placebo. Comment Acetaminophen and ibuprofen are the most common over-the-counter drugs for pediatric fever and pain. Their use is driven mainly by parental fever phobia, as denoted by Schmitt, 1 and although many clinicians are suggesting not to be so vigorous about managing fever, others are recommending an antipyretic as soon as fever begins. The recommendations in the recent literature varied between the very safe acetaminophen to the slightly more effective ibuprofen, yet answering a parent s question about which drug is better was difficult. Previous systematic reviews could find no definite clinical difference between the two drugs because few prospective, methodologically sound papers were published about it. Moreover, drug dose was different in these studies, and comparing the different outcome measures was almost unpalatable. This meta-analysis uses a methodologic foundation and statistical approach that allows the authors to compare the effectiveness of the drugs, mainly to evaluate change in fever. Using a higher point estimate of the mean effect size, they conclude that 10 mg/kg ibuprofen is better than, and as safe as, 10 to 15 mg/kg acetaminophen. This recommendation appears to tilt the balance toward ibuprofen. However, these results should be taken with a grain of salt. First, the 10 papers used in the meta-analysis for fever management have a small sample size and low power. The assortment of outcome measures used in these studies could have underpowered the meta-analysis, especially since Perrott et al use standardized effect, a method with questionable reliability. Second, the variety of 142 Clinical Research Abstracts For Pediatricians The Journal of Pediatrics January 2005

191 doses used in the studies to date could also make the interpretation challenging. In many institutions, acetaminophen is used at 15 mg/kg, 50% more than administered in most studies, and a difference that could alter the findings illustrated here. Third, adverse events are traditionally under-reported, and almost all studies have an underdescription of adverse event types and methods used to collect the information. The battle between the safe camp and the efficacious camp is not over, although both drugs appear to be relatively safe. Perhaps, most importantly for readers of this well-written paper, is the fact that the analysis included only a single-dose comparison. Studies with multidose acetaminophen or ibuprofen are scarce but represent the reality show at home, in the office, or on the pediatrics ward. Is the safety of multiple doses still close to equivalent between the two drugs? Is the effectiveness of ibuprofen still slightly better? These questions are still to be answered in future, large, prospective studies and leave us somewhat uncertain what to say when parents inquire which drug is better. Ran D. Goldman, MD Division of Emergency Medicine and Clinical Pharmacology and Toxicology Department of Pediatrics, University of Toronto Toronto, On M5G1X8, Canada REFERENCE 1. Schmitt BD. Fever phobia: misconceptions of parents about fevers. Am J Dis Child 1980;134: Evaluation of a method to reduce overthe-counter medication dosing error Frush KS, Luo X, Hutchinson P, Higgins JN. Arch Pediatr Adolesc Med 2004;158: Context Many studies have evaluated methods for reducing medication errors in the hospital setting, yet little is known about strategies to reduce dosing errors in the home. Objectives To introduce a simple method of dosing overthe-counter medication in a home setting using a color-coding concept and to compare dosing deviation from recommended dosage using the color-coded method with dosing deviation using conventional package labeling. Design Setting center. Randomized controlled clinical trial. Pediatric emergency center at a tertiary care medical Participants A sample of 101 caregivers of children with nonemergent complaints were randomly assigned to the colorcoded or conventional group by alternating assignment. Interventions The conventional group was instructed to use a conventional dosing method to determine and measure a dose of acetaminophen for their child, whereas the colorcoded group was instructed to use a color-coded method. Main outcome measures For both dose determination and dose measuring, percentage of deviation from recommended acetaminophen dosage was calculated and compared between the 2 groups. Results There was no significant difference in sociodemographic characteristics between the 2 groups. However, for dose determination, the average deviation (25.8% vs 1.7%) and median deviation (1% vs 0%) from recommended dosage were both higher for the group using conventional methods compared with the group using the color-coded method. The Wilcoxon rank sum test indicated that the median deviation was significantly different between the 2 groups (P <.001). Similar results were obtained for dose measuring. The average deviation (29% vs 0.5%) and the median deviation (17.2% vs 0%) from recommended dosage were higher for the group using conventional methods compared with the group using the color-coded method. The median deviation was also significantly different between the 2 groups (P <.001). Conclusions This study suggests a marked improvement in caregivers ability to correctly determine and measure an overthe-counter medication for their child, using a color-coded method compared with conventional methods. Comment This clinical trial by Frush et al showed that a simple color-coded method of measuring acetaminophen made it less likely for parents to underdose or overdose their child with the medication. Because acetaminophen is the most widely used nonprescription medication for children, 1 a reduction in dosing errors probably would have a large impact. However, there are a few limitations of the trial. First, caregivers were assigned to their treatment group by alternating assignment rather than by a true random mechanism (thus making the assignment of subjects predictable), and neither the subjects nor data collectors were blinded to the treatment group allocation. This raises the potential for biased results. Second, it is not clear that the participants in the two groups were treated equally in all respects except for the intervention being tested. Last, the study population included parents of patients seen in a tertiary care center emergency department for nonemergent problems. This limits the ability to generalize the findings to parents using this method in the home setting. Despite these limitations, the results are promising. The intervention is simple and has the potential to improve dosing accuracy by parents. Randomized trials in other settings, particularly in the home setting, would be helpful to see if the findings from this analysis hold true. 2 Jill S. Halterman, MD, MPH Assistant Professor of Pediatrics University of Rochester School of Medicine Rochester, NY REFERENCES 1. Simon HK, Weinkle DA. Over-the-counter medications: do parents give what they intend to give? Arch Pediatr Adolesc Med 1997;151: Rand CM, Conn KM, Crittenden CN, Halterman JS. Does a colorcoded method for measuring acetaminophen doses reduce the likelihood of dosing error? Arch Pediatr Adolesc Med 2004;158: Clinical Research Abstracts For Pediatricians 143

192 Efficacy and safety of statin therapy in children with familial hypercholesterolemia: A randomized controlled trial Wiegman A, Hutten BA, de Groot E, Rodenburg J, Bakker HD, Büller HR, et al. JAMA 2004;292: Context Children with familial hypercholesterolemia have endothelial dysfunction and increased carotid intima-media thickness (IMT), which herald the premature atherosclerotic disease they have later in life. Although intervention therapy in the causal pathway of this disorder has been available for more than a decade, the long-term efficacy and safety of cholesterollowering medication have not been evaluated in children. Objectives To determine the 2-year efficacy and safety of pravastatin therapy in children with familial hypercholesterolemia. Design Randomized, double-blinded, placebo-controlled trial. Setting An academic medical referral center in the Netherlands. Participants Two hundred fourteen children with familial hypercholesterolemia, 8 to 18 years of age. Interventions After initiation of a fat-restricted diet and encouragement of regular physical activity, children were randomly assigned to receive treatment with 20 to 40 mg/d pravastatin (n = 106) or a placebo tablet (n = 108). Main outcome measures The primary efficacy outcome was the change from baseline in mean carotid IMT compared between the 2 groups over 2 years; the principal safety outcomes were growth, maturation, and hormone level measurements over 2 years as well as changes in muscle and liver enzyme levels. Results Compared with baseline, carotid IMT showed a trend toward regression with pravastatin (mean [SD], ÿ0.010 [0.048] mm; P =.049), whereas a trend toward progression was observed in the placebo group (mean [SD], [0.044] mm; P =.28). The mean (SD) change in IMT compared between the 2 groups (0.014 [0.046] mm) was significant (P =.02). Also, pravastatin significantly reduced mean LDL cholesterol levels compared with placebo (ÿ24.1% vs +0.3%, respectively; P <.001). No differences were observed for growth, muscle or liver enzymes, endocrine function parameters, Tanner staging scores, onset of menses, or testicular volume between the 2 groups. Conclusions Two years of pravastatin therapy induced a significant regression of carotid atherosclerosis in children with familial hypercholesterolemia, with no adverse effects on growth, sexual maturation, hormone levels, or liver or muscle tissue. Comment Familial hypercholesterolemia (FH), induced by a defect in the expression of hepatic LDL receptors, increases the risk for early heart disease, with coronary events occurring sometimes by the third decade of life in those afflicted. 1 Lifestyle changes alone may have a minimal impact on this genetic dyslipidemia, and the more potent options of drugs or blood apheresis to remove LDL have their own issues. 2 After a number of clinical trials that have reported significant coronary risk reductions, there can be no question that use of the lipid medications called statins must be considered as part of the treatment of high-risk adults, such as those who present with FH. 3 More debatable is whether drugs are appropriate for adolescent or younger patients. Wiegman et al report that pravastatin treatment attenuated the progression of carotid atherosclerosis in children with FH. The mean age of the children in this study was 13 years, and approximately 60% of the participants were younger than 14 years. Although the sample size was small, positive results were observed: compliance to treatment was good (84% of tablets were taken) and pravastatin was not associated with differential effects on school performance, maturation, and other developmental parameters. The one case of extreme creatine kinase elevation occurred in a placebo-allocated child. Otherwise, rates of elevations of liver and muscle enzymes were similar between the groups. These results complement those of other studies in adults with FH that reported improvement in vascular parameters with statin therapy. 4,5 As a surrogate marker of benefit, carotid intimal medial thickness in adult patients has been correlated with coronary risk, 6 but no clinical end point studies have confirmed that statins protect pediatric patients with FH from having cardiovascular events in the future. Furthermore, although pravastatin was well tolerated over the 2 years of follow-up, the study does not provide longer-term safety data. Until researchers can fill these gaps, the question remains whether delaying the start of drug therapy until after adolescence may be just as beneficial as intervening in childhood. Nevertheless, when deciding on the intensity of treatment, physicians must bear in mind the child s total risk factor profile as well as the potential to halt the establishment of atherosclerotic disease and to reduce lifetime risk with earlier initiation of lipid treatment. 7 The accelerated time course for developing atherosclerosis seen with FH makes starting prevention in childhood an important practice. Antonio M. Gotto, Jr, MD, DPhil The Stephen and Suzanne Weiss Dean Weill Medical College of Cornell University New York, NY REFERENCES 1. Gotto AM Jr. Targeting high-risk young patients for statin therapy. JAMA 2004;292: Gordon BR, Stein E, Jones P, Illingworth DR. Indications for low-density lipoprotein apheresis. Am J Cardiol 1994;74: Vaughan CJ, Gotto AM Jr. Update on statins: Circulation 17;110: Nolting PR, de Groot E, Zwinderman AH, Buirma RJ, Trip MD, Kastelein JJ. Regression of carotid and femoral artery intima-media thickness 144 Clinical Research Abstracts For Pediatricians The Journal of Pediatrics January 2005

193 in familial hypercholesterolemia: treatment with simvastatin. Arch Intern Med 2003;163: Smilde TJ, van Wissen S, Wollersheim H, Trip MD, Kastelein JJ, Stalenhoef AF. Effect of aggressive versus conventional lipid lowering on atherosclerosis progression in familial hypercholesterolaemia (ASAP): a prospective, randomised, double-blind trial. Lancet 2001;357: Chambless LE, Heiss G, Folsom AR, Rosamond W, Szklo M, Sharrett AR, Clegg LX. Association of CAD incidence with carotid arterial wall thickness and major risk factors: the Atherosclerosis Risk in Communities (ARIC) study, Am J Epidemiol 1997;146: Tonstad S. Treatment of children with familial hypercholesterolemia. Exp Rev Cardiovasc Ther 2003;1: Fluoxetine, cognitive-behavioral therapy, and their combination for adolescents with depression: Treatment for Adolescents With Depression Study (TADS) randomized controlled trial Treatment for Adolescents With Depression Study (TADS) Team. JAMA 2004;292: Context Initial treatment of major depressive disorder in adolescents may include cognitive-behavioral therapy (CBT) or a selective serotonin reuptake inhibitor. However, little is known about their relative or combined effectiveness. Objective To evaluate the effectiveness of 4 treatments among adolescents with major depressive disorder. Design Randomized controlled trial. Setting The trial was conducted at 13 US academic and community clinics between spring 2000 and summer Participants Volunteer sample of 439 patients, ages 12 to 17 years, with a primary Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, diagnosis of major depressive disorder. Interventions Twelve weeks of (1) fluoxetine alone (10 to 40 mg/d), (2) CBT alone, (3) CBT with fluoxetine (10 to 40 mg/d), or (4) placebo (equivalent to 10 to 40 mg/d). Placebo and fluoxetine alone were administered in a double-blinded fashion; CBT alone and CBT with fluoxetine were administered in unblinded fashion. Main outcome measures Children s Depression Rating Scale-Revised total score and, for responder analysis, a (dichotomized) Clinical Global Impressions improvement score. Results Compared with placebo, the combination of fluoxetine with CBT was statistically significant (P =.001) on the Children s Depression Rating Scale-Revised. Compared with fluoxetine alone (P =.02) and CBT alone (P =.01), treatment of fluoxetine with CBT was superior. Fluoxetine alone was a superior treatment to CBT alone (P =.01). Rates of response for fluoxetine with CBT were 71.0% (95% CI, 62% to 80%); fluoxetine alone, 60.6% (95% CI, 51% to 70%); CBT alone, 43.2% (95% CI, 34% to 52%); and placebo, 34.8% (95% CI, 26% to 44%). On the Clinical Global Impressions improvement responder analysis, the 2 fluoxetine-containing conditions were statistically superior to CBT and to placebo. Clinically significant suicidal thinking, which was present in 29% of the sample at baseline, improved significantly in all 4 treatment groups. Fluoxetine with CBT showed the greatest reduction (P =.02). Seven (1.6%) of 439 patients attempted suicide; there were no completed suicides. Conclusions The combination of fluoxetine with CBT offered the most favorable tradeoff between benefit and risk for adolescents with major depressive disorder. Comment Treatment of adolescents with major depression, and particularly the possibility of an increased risk of suicidal behavior associated with antidepressant drug treatment, have become issues of major controversy in recent months. Adolescent patients, their parents, and their physicians face heightened concerns in weighing the potential benefits and risks in making choices about treating major depression, a disorder marked by considerable suffering, functional impairment, and risk of death from suicide. This report of the 12-week primary results of the Treatment for Adolescents with Depression Study (TADS) provides important empiric data bearing on this clinical issue. This trial was the first to evaluate the efficacy in adolescents of combined treatment with a serotonin-selective antidepressant (fluoxetine) and CBT compared with fluoxetine alone, CBT alone, and placebo. Combined treatment was superior on most measures of response. Fluoxetine was administered in an unblinded fashion in the combined treatment group a design limitation that might have advantaged the combined treatment group due to positive expectational effects. Suicidal thinking decreased with all 4 treatments, even placebo, reflecting the importance and clinical relevance of the careful monitoring and attention offered to all of the patients and parents in this multicenter, randomized, controlled trial. The results agree with previous trials in adult patients indicating that combined treatment with an evidence-based psychotherapy and antidepressant medication is optimal for most patients with major depression. Richard M. Glass, MD University of Chicago Chicago, IL Outcomes of allergy to insect stings in children, with and without venom immunotherapy Golden DBK, Kagey-Sobotka A, Norman PS, Hamilton RG, Lichtenstein LM. N Engl J Med 2004;351: Context Children are thought to outgrow the allergy to insect stings, but there are no reports documenting the natural history of this reaction. Objectives To study the outcome of allergic reactions to insect stings in childhood 10 to 20 years afterward in patients who had not received venom immunotherapy and in those who had been treated. Clinical Research Abstracts For Pediatricians 145

194 Design Observational study. Setting Allergy clinic at a major urban health center in the United States. Participants Children (n = 1033) in whom allergic reaction to insect stings were diagnosed between 1978 and 1985, of whom 356 received venom immunotherapy. Interventions Telephone and mail survey between January 1997 and January Main outcome measures stings. Type and severity of reactions to Results Of the 1033 patients, 512 patients (50%) responded, with a mean follow-up period of 18 years, a mean duration of venom immunotherapy of 3.5 years in treated patients, and an incidence of stings of 43%. Systemic reactions occurred less frequently in patients who had received venom immunotherapy (2 of 64 patients, or 3%) than in untreated patients (19 of 111 patients, 17%; P = 0.007). Patients with a history of moderate-to-severe reactions had a higher rate of reaction if they had not been treated (7 of 22 patients, or 32%) than if they had received venom immunotherapy (2 of 43 patients, or 5%; P =.007). In patients who had been treated and who had a history of mild (cutaneous) systemic reaction (ie, one with only cutaneous manifestations), none of the 21 subjects who received stings had a systemic reaction. Conclusions A clinically important number of children do not outgrow allergic reactions to insect stings. Venom immunotherapy in children leads to a significantly lower risk of systemic reaction to stings even 10 to 20 years after treatment is stopped, and this prolonged benefit is greater than the benefit seen in adults. Comment This study is a retrospective analysis comparing outcomes in children after insect sting reactions, some of whom received venom immunotherapy and others who did not. Results demonstrate that systemic reactions occur less frequently in children for whom venom immunotherapy is indicated and who receive venom immunotherapy versus those who do not receive such therapy. In patients with a history of moderate-to-severe reactions, those who receive venom immunotherapy have a lower rate of reaction on resting than those who do not receive immunotherapy. Among treated patients with mild cutaneous reactions, none had a systemic reaction on subsequent stings. This was not significantly different than those who had not been given venom immunotherapy (P =.12). This study reveals that most children, regardless of whether or not they are treated, lose their insect sensitivity by adulthood. However, those with a history of moderate-to-severe reactions, who are at risk for a systemic reaction on subsequent stings, do statistically better with treatment even 10 to 20 years after immunotherapy is stopped. Thus, a course of venom immunotherapy for moderate-to-severe anaphylaxis is warranted to prevent resting systemic reactions even years after therapy has been stopped. Patrick J. DeMarco, MD Richard F. Lockey, MD Division of Allergy and Immunology University of South Florida College of Medicine and James A. Haley VA Hospital Tampa, FL Effect of dextromethorphan, diphenhydramine, and placebo on nocturnal cough and sleep quality for coughing children and their parents Paul IM, Yoder KE, Crowell KR, Shaffer ML, McMillan HS, Carlson LC, Dilworth DA, Berlin CM. Pediatrics 2004;114: e Context Although dextromethorphan and diphenhydramine are not recommended by the American Academy of Pediatrics for treatment of coughs, they are frequently used. Objectives To determine whether dextromethorphan and diphenhydramine are superior to placebo for the treatment of nocturnal cough and sleep difficulty associated with upper respiratory infections and to determine whether parents have improved sleep quality when their children receive the medications when compared with placebo. Design Randomized, double-blinded study. Setting Two university-affiliated practices in Hershey, Pennsylvania. Participants infections. Parents of 100 children with upper respiratory Interventions After stratification by age, patients were randomly assigned to receive either dextromethorphan (dosed based on age), diphenhydramine (dose based on weight), or placebo. Main outcome measures Frequency, severity, and bothersome nature of the nocturnal cough recorded on 2 consecutive days, initially on the day of presentation, when no medication had been given the previous evening, and then again on the subsequent day, after the intervention. Sleep quality for both the child and the parent were also assessed for both nights. Results For the entire cohort, all outcomes were significantly improved on the second night of the study when either medication or placebo was given. However, neither diphenhydramine nor dextromethorphan produced a superior benefit when compared with placebo for any of the outcomes studied. Insomnia was reported more frequently in those who were given dextromethorphan, and drowsiness was reported more commonly in those who were given diphenhydramine. Conclusions Diphenhydramine and dextromethorphan are not superior to placebo in providing nocturnal symptom relief for children with cough and sleep difficulty as a result of an upper respiratory infection. Furthermore, the medications 146 Clinical Research Abstracts For Pediatricians The Journal of Pediatrics January 2005

195 given to children do not result in improved quality of sleep for their parents when compared with placebo. Each clinician should consider these findings, the potential for adverse effects, and the individual and cumulative costs of the drugs before recommending them to families. Comment This randomized, placebo-controlled trial evaluated the effect of two common over-the-counter (OTC) medications on nighttime cough in children. In line with other recent evidence, both OTC cough medicines were no more effective than placebo. Although this study did not report details about the selection and randomization process, all groups appeared to be similar at baseline in terms of sex, age, and ethnicity, suggesting that random assignment was successful. The authors acknowledge that the study duration was short and that a no-treatment arm would have provided further data on the placebo effect, which could have been present in this study. It is reassuring that cough improved substantially in all treatment arms. The authors rightly conclude that despite the strong desire for both parents and clinicians to ease symptoms of cough in children, the use of OTC cough medicines is questionable for what largely is a self-limited condition, both in terms of cost-effectiveness and the potential for adverse effects. Knut Schroeder, MD, MSc, PhD, MRCGP, MRCP Department of Community-Based Medicine University of Bristol Bristol, Avon BS6 6JL, UK Contagiousness of varicella in vaccinated cases: A household contact study Seward JF, Zhang JX, Maupin TJ, Mascola L, Jumaan AO. JAMA 2004;292: Context Some data suggest that cases of varicella in vaccinated patients may be less contagious than those who are unvaccinated; however, these data are limited. Objectives To describe secondary attack rates within households according to disease history and vaccination status of the primary case and household contacts and to estimate varicella vaccine effectiveness. Population-based, active varicella surveillance pro- Design ject. Setting A community of approximately 320,000 in Los Angeles County, California, from 1997 to Participants Cases were reported by child care centers, private and public schools, and health care clinicians. Information on household contacts age, varicella history, and vaccination status was collected. Main outcome measures Varicella secondary attack rate among household contacts; vaccine effectiveness using secondary attack rates in unvaccinated and vaccinated children and adolescents. Results A total of 6316 varicella cases were reported. Among children and adolescents aged 1 to 14 years, secondary attack rates varied according to age and by disease and vaccination status of the primary case and exposed household contacts. Among contacts aged 1 to 14 years exposed to unvaccinated cases, the secondary attack rate was 71.5% if they were unvaccinated and 15.1% if they were vaccinated (risk ratio [RR], 0.21; 95% confidence interval [CI], ). Overall, vaccinated cases were half as contagious as unvaccinated cases. However, vaccinated cases with 50 lesions or more were similarly contagious as unvaccinated cases whereas those with fewer than 50 lesions were only one third as contagious (secondary attack rate, 23.4%; RR, 0.32 [95% CI, ]). Vaccine effectiveness for prevention of all disease was 78.9% (95% CI, 69.7%-85.3%); moderate disease, 92% ( lesions) and 100% (clinician visit); and severe disease, 100%. Conclusions Under conditions of intense exposure, varicella vaccine was highly effective in preventing moderate and severe disease and about 80% effective in preventing all disease. Breakthrough varicella cases in household settings were half as contagious as unvaccinated persons with varicella, although contagiousness varied with numbers of lesions. Comment This article indicates several important phenomena concerning varicella and varicella vaccine. It is now well known that breakthrough varicella sometimes occurs when a vaccinated person is exposed to the wild type virus. How frequently this occurs is still being evaluated. In this 5-year population-based, active-surveillance study, vaccination provided 80% complete protection against the disease after household exposure and prevented severe breakthrough varicella almost entirely. Household exposures to varicella provide the strongest test for the vaccine because reported attack rates range from 61% to 100% among susceptible individuals following this intense form of challenge. Reassuringly, transmission of varicella from children with breakthrough disease was only about one third that expected, compared to transmission by unvaccinated children. These data are in keeping with other published information and provide important confirmation about the efficacy of routine immunization against chickenpox. Interestingly, even among individuals with past natural infection, there were some cases of varicella and therefore one must question therefore whether protection approaching 100% by vaccination can ever be achieved against chickenpox. On the other hand, given the numerous outbreaks of varicella in schools and day care centers reported since 1995, when the vaccine was licensed in the US, it seems worthwhile to explore whether a second dose of varicella vaccine, given routinely to all children, would improve protection and possibly result in more lasting immunity due to an anamnestic response. The CDC is currently considering whether or not to recommend a second routine dose for children aged 1 to 12 years. Anne Gershon, MD Columbia University New York, NY Clinical Research Abstracts For Pediatricians 147

196 LETTERS The kidney in Fabry disease To the Editor: We have read with great interest the article by Desnick and Brady 1 on Anderson-Fabry disease (FD) in childhood. As reviewed by the authors, FD is an X-linked glycosphingolipidosis as a result of the deficiency of the enzyme a-galactosidase A. The accumulation of globotriaosylceramide in the lysosomes of various tissues is the underlying known pathology causing the clinical symptoms. Until recently, most disease features had been reported in adults. Hemizygous adult males present with acroparesthesias, angiokeratomas, corneal opacities, hypohydrosis, and end-organ disease affecting the kidney, heart, and brain. Interestingly, years ago glycolipid storage was reported in fetal kidney, myenteric plexus, 2 and fetal cornea. 3 Since then, pediatric observations of children with FD have been published, 4 but during the last few years, the pediatric FD phenotype has been studied and reported more systematically. 5,6 The review by Desnick and Brady 1 in The Journal emphasizes the importance of this disease in pediatric practice. We report on the daughter of a kidney-transplanted patient with FD reported by Friedlaender et al. 7 At 12 years of age, she had a kidney biopsy that demonstrated the typical changes of FD (Figure), including striking, diffuse foamy vacuolization of the glomerular epithelial cells, and at 14years of age, bilateral cornea verticillata were observed. This observation emphasizes that significant kidney involvement can be present not only in males but also in heterozygous females, even as early as 12 years of age. Complaints of acroparesthesias in children should lead to urinanalysis for proteinuria. Today, the pediatrician has to be aware of FD because early diagnosis is mandatory for accurate genetic counseling and timely enzyme replacement therapy with the hope of preventing complications of the disease. Annick Raas-Rothschild, MD Michael M. Friedlaender, MD Galina Pizov, MD Rebecca Backenroth, MD, MPH Departments of Human Genetics, Nephrology, and Pathology Hadassah Hebrew University Medical School Ein Keren Jerusalem 91120, Israel YMPD /j.jpeds REFERENCES 1. Desnick RJ, Brady RO. Fabry disease in childhood. J Pediatr 2004; 144(suppl 5):S Elleder M, Poupetova H, Kozich V. [Fetal pathology in Fabry s disease and mucopolysaccharidosis type I]. Cesk Patol 1998;34:7-12. Figure. Renal biopsy showing enlarged podocytes with foamy cytoplasm that cause glomerular atelectasis. Periodic acid-schiff stain, original magnification. (Figure available in color online at 3. Tsutsumi A, Uchida Y, Kanai T, Tsutsumi O, Satoh K, Sakamoto S. Corneal findings in a foetus with Fabry s disease. Acta Ophthalmol (Copenh) 1984;62: MacDermot KD, Holmes A, Miners AH. Anderson-Fabry disease: clinical manifestations and impact of disease in a cohort of 98 hemizygous males. J Med Genet 2001;38: Pintos Morell G. [Fabry s disease: diagnosis in the pediatric age group]. An Esp Pediatr 2002;57: Ries M, Ramaswami U, Parini R, Lindblad B, Whybra C, Willers I, et al. The early clinical phenotype of Fabry disease: a study on 35 European children and adolescents. Eur J Pediatr 2003;162: Friedlaender MM, Kopolovic J, Rubinger D, Silver J, Drukker A, Ben-Gershon Z, et al. Renal biopsy in Fabry s disease eight years after successful renal transplantation. Clin Nephrol 1987;27: Low birth weight and blood pressure: The role of neonatal factors in the small-baby-syndrome To the Editor: With regard to the phenomenological relationship observed between low birthweight and increased cardiovascular risk, Gillman et al published an important study on perinatal predictors of newborn blood pressure. 1 Interestingly, in contrast to data from long-term studies on the small-baby syndrome, low birthweight is associated with low blood pressure in newborns. Launer et al show that the direction of the relation between birthweight and blood pressure reverses after 3 months of age. 2 We suggest that this reversal in early postnatal life strongly indicates the presence of neonatal factors acting critically, factors which are able to set those 148 Letters The Journal of Pediatrics January 2005