CONGENITAL ADRENAL HYPERPLASIA (CAH) is

Transcription

1 /02/$15.00/0 The Journal of Clinical Endocrinology & Metabolism 87(9): Printed in U.S.A. Copyright 2002 by The Endocrine Society doi: /jc High Reliability of Neonatal Screening for Congenital Adrenal Hyperplasia in Switzerland MICHAEL STEIGERT, EUGEN J. SCHOENLE, ANNA BIASON-LAUBER, AND TONI TORRESANI Department of Endocrinology and Diabetology, University Children s Hospital, CH-8032 Zurich, Switzerland Newborn screening for congenital adrenal hyperplasia (CAH) is justified by the sometimes difficult clinical diagnosis and the risks associated with missed diagnosis, particularly the life-threatening salt-wasting crisis. In Switzerland, nationwide screening for CAH by measuring 17-hydroxyprogesterone levels in dried blood spots was introduced in At the Zurich University Children s Hospital, 50% of the population of Switzerland is screened. The aim of the study was to evaluate the efficiency of the Zurich screening program. Between January 1, 1993, and May 31, 2001, 333,221 newborns were screened for CAH. Thirty-one newborns had CAH (incidence, 1 in 10,749); 30 were detected through screening (sensitivity, CONGENITAL ADRENAL HYPERPLASIA (CAH) is one of the most frequent inborn endocrine disorders; it comprises autosomal recessive disorders of cortisol biosynthesis in the adrenal gland caused by various enzyme deficiencies. Of these, 21-hydroxylase deficiency (21-OHD) is caused by mutations on the CYP21 gene located on the short arm of chromosome 6 and accounts for more than 90% of all CAH cases (1 3). Comparatively, all other adrenal enzyme deficiencies leading to CAH are relatively rare. The clinical presentation of patients with CAH is heterogeneous and depends on the type of gene mutation as well as the sex of the patient (4). Some newborns with CAH may present without clinical signs or symptoms postnatally, and in these the diagnosis of CAH can obviously not be made on clinical basis. The unrecognized and thus untreated disease may entail life-threatening salt-wasting crisis in the newborn period and lead to morbidity later in life, including precocious puberty and short stature. Thus, the benefit of a neonatal screening that fulfils the criteria proposed for feasibility for neonatal screening and that can prevent delayed diagnosis of CAH and its associated morbidity and mortality is evident (5, 6). The most suitable biochemical marker for diagnosing CAH is 17-hydroxyprogesterone (17-OHP), the steroid metabolite lying just upstream of the block. Measuring this parameter can achieve early diagnosis and treatment of CAH in affected, yet symptom-free, newborns (7). Newborn screening for CAH by measuring 17-OHP levels in dried blood spots was incorporated in the Swiss neonatal screening program for metabolic and endocrine diseases at the end of It can detect 21-OHD as well as 11- hydroxylase deficiency (11 -OHD). None of the other, by far Abbreviations: CAH, Congenital adrenal hyperplasia; 21-OHD, 21- hydroxylase deficiency; 17-OHP, 17-hydroxyprogesterone. 97%). A recall for suspected CAH was performed in only 60 cases, corresponding to a very low recall rate (0.0018%). In 30 recalls CAH was confirmed (positive predictive value, 50%; specificity, 99.99%). Fifteen of 31 patients profited from screening, as CAH had not been recognized clinically. The timely availability of screening results made therapy possible within the first week of life in most cases and helped in preventing salt-wasting crisis in all patients. With a sensitivity of 97%, a specificity of 99.99%, and a positive predictive value of 50%, the Zurich neonatal screening program for CAH can be considered highly reliable. (J Clin Endocrinol Metab 87: , 2002) less frequent, enzyme deficiencies of the adrenal gland leading to CAH can be found with this screening parameter. The aim of our study was to evaluate the efficiency of the screening program for CAH at the Zurich University Children s Hospital 9 yr after its initiation. Subjects and Methods Screening for CAH is performed as a part of the Swiss Neonatal Screening Program. Samples are collected from newborns on a single filter paper card for all tests h after birth. The Neonatal Screening Protocol requires that these samples are sent daily by mail to the screening center. The Swiss neonatal screening program is organized as follows. All newborn babies in Switzerland are screened h after birth for the following disorders: phenylketonuria, galactosemia, biotinidase deficiency, congenital hypothyroidism, and CAH. Two centers are in charge of the screening, each covering approximately 50% of the births. The laboratories are located in Bern for the western part of Switzerland and in Zurich for the eastern part of Switzerland. The analytical methods and procedures for screening are the same in the two screening centers. The screening center for eastern Switzerland is located at the University Children s Hospital in Zurich. More than 99% of the newborns are reached by the screening. For CAH screening, 17-OHP is measured with an immunofluorometric assay (DELFIA, Wallac, Inc., Turku, Finland) without prior extraction. The 17-OHP measurements are interpreted by comparing them with gestational age-related cutoff values (8). The whole screening procedure is in accordance with the recently published recommendations of the Working Group on Neonatal Screening of the European Society for Pediatric Endocrinology (9). The results of the screening are available on the 6th or 7th day of life. All 17-OHP values less than 30 nmol/liter blood are considered normal. When a 17-OHP value greater than 90 nmol/liter blood in term babies or clearly above the 95th percentile in premature babies is found, a full endocrinological examination is requested. The recall is always made by telephone and supplemented by written information. In cases with borderline 17-OHP values (17-OHP between nmol/liter), a second filter paper sample is requested as a first-line follow-up measure. The result of this second sample is then used, in comparison with the first result, for deciding whether a recall is necessary. A recall is therefore defined as a full endocrinological 4106

2 Steigert et al. Screening for CAH in Switzerland J Clin Endocrinol Metab, September 2002, 87(9): examination and follows either a clearly elevated 17-OHP screening value or repeatedly and continuously mildly elevated 17-OHP values. A request for a filter paper card control without any other intervention is not considered a recall. In a first step of our evaluation we retrospectively reviewed the data for all patients with elevated 17-OHP who were recalled for suspected CAH and in whom an endocrinological work-up had been performed during the evaluation period from January 1, 1993, until May 31, In a second ascertainment we reviewed the data of all patients born during the evaluation period in whom at any time after birth CAH had been clinically suspected and laboratory testing had been performed, to identify any patient who had not been detected by screening. As screening for CAH by measuring 17-OHP levels can detect CAH caused by 21-OHD and 11- OHD, but is not sensitive for any of the other, by far less frequent, adrenal enzyme deficiencies leading to CAH, we will refer only to these two enzymatic defects as CAH in the following text unless specified otherwise. Results Population, sex distribution, and gestational age Some 333,221 newborns were screened for CAH during the evaluation period. Of these, 162,512 (48.77%) were boys, and 170,709 (51.23%) were girls. The exact gestational age at birth was reported by the clinician in approximately one third of the newborns; the majority of the remaining two thirds were referred to by their respective birth weights rather than their gestational ages at birth; in a very small proportion of the screened newborns data were provided for neither gestational age nor birth weight. From the available data we deducted a preterm rate of 5.69% (314,261 term vs. 18,960 preterm infants out of 333,221 newborns). The sex ratio in the preterm infants was 54.28% males vs % females. TABLE 1. Comparison between recalls of born at term and preterm infants All newborns Term Preterm Screened population 333, ,261 (94.4%) 18,960 (5.6%) (n) Recalls (n) Recall rate % % % CAH confirmed (n) False positive recalls (n) Positive predictive value 50% 82% 8% The positive predictive value was 50% in all newborns and 82% in infants born at term. Only six false positive recalls were performed in 314,261 infants born at term. The male to female ratio was 51.2:48.8 in all newborns vs. 54.2:45.8 in preterm infants. Recall rate and positive predictive value A recall for suspected CAH was performed in 60 cases. The recall rate in all newborns was thus %. Thirty-four of these recalls were performed in infants born at term, and 26 were performed in the preterm infant group. A comparison of the recall results of the two subgroups of the term and the preterm born infants is listed in Table 1. The recall rate in the term group was % (34 of 314,201) and thus roughly 13-fold lower than the recall rate of 0.014% in the preterm infant group (26 of 18,960). The diagnosis of CAH was confirmed in 30 of those 60 recall cases; 29 had 21-OHD; 1 had 11 -OHD (17-OHP, 37.4 nmol/liter blood on d 4). The positive predictive value of a recall in the screened population was thus 50%. By comparing the subgroups of the infants born at term to the group of preterm infants we found a positive predictive value of 82% for term infants and 8% for preterm infants. In only 6 of 314,261 term infants a false positive recall was performed (0.0019%). CAH cases found after neonatal screening, incidence, sensitivity, and specificity One false negative screening test was found during our evaluation period. In this term female infant the 17-OHP level was within the normal range in the regular screening test performed on d 4 (11.6 nmol/liter). A second test was requested by the clinician because of clitoral hypertrophy and was performed on d 11; in this second sample 17-OHP was markedly elevated (220 nmol/liter), and 21-OHD was later confirmed by urinary steroid analysis. Apart from this one case, we have no knowledge of any other case missed by the screening. However, we found three additional patients with steroidogenic enzyme deficiencies that cannot be detected by measuring 17-OHP levels in the first week of life. The first patient, a term female, had a normal screening test (17-OHP, 15.5nmol/liter on d 4); an ACTH test at age 5 yr performed because of clinical suspicion of CAH led to the diagnosis of 21-OHD late-onset CAH. Two patients had steroidogenic enzyme deficiencies other than 21-OHD or 11 -OHD; in both cases the screening yielded 17-OHP levels within the normal range, as expected. The first patient, a term female, had 3 -hydroxysteroid dehydrogenase deficiency (17-OHP, 6.8 nmol/liter blood on d 4) and was diagnosed on evaluation of intersexual genitalia at age 3 yr; the second, a term male, had 17 -hydroxylase deficiency (17-OHP, 22.7 nmol/liter blood on d 4). Altogether, 31 of 333,221 newborns had CAH; of these, 30 had 21-OHD, and 1 had 11 -OHD. The incidence of 21-OHD CAH was thus 1 in 11,108. As the screening detected 30 of those 31 CAH patients, we set the sensitivity at 97%. The rate of false positive screening tests (recalls where diagnosis of CAH was wrongly considered) was 30 of 333,221 (0.009%). Of these 30 false positive recalls, 24 were performed in preterm infants (84%). The specificity of the screening was 99.99%. Comparison of CAH-confirmed recalls with false positive recalls The recalled subjects with confirmed CAH had a comparably higher 17-OHP level than the subjects with a false positive recall, and their average gestational age was higher than that in the patient group with a false positive recall (Table 2). Salt wasting In 5 patients, 1 girl and 4 boys with confirmed CAH, laboratory signs of beginning salt wasting were reported at the time the recall reached the patient (sodium, 135 mmol/ liter; potassium, 6.5 mmol/liter). Therapy with hydrocortisone and fludrocortisone was begun immediately after the recall reached the patient (d 6 9; mean, 7.6), and consecu-

3 4108 J Clin Endocrinol Metab, September 2002, 87(9): Steigert et al. Screening for CAH in Switzerland TABLE 2. Comparison of CAH-confirmed recalls with false positive recalls All recalls CAH confirmed CAH not confirmed Patient numbers (n) 60 (100%) 30 (50%) 30 (50%) Gestational age (wk) 36.5 (26 42) 39.4 (34 42) 33.5 (26 39) Sex ratio (male/female) 36/24 13/17 23/7 17-OHP Peak (nmol/liter) 238 (35 675) 363 (56 675) 111 (35 540) The mean gestational age and the peak 17-OHP value was higher in the CAH-confirmed patient group. TABLE 3. Clinical suspicion as diagnostic tool in CAH tively severe salt wasting did not develop in any of these patients. Of those 5 patients with signs of initial salt wasting, only the affected girl, but none of the 4 boys, was clinically suspected to have CAH. Of the remaining 25 patients 15 showed no detectable signs of salt wasting up to the time of the recall, whereas 10 patients could not be evaluated for possible salt wasting because data were incomplete. At any rate no patient with acute salt-wasting crisis was reported during the entire evaluation period. Clinical suspicion as diagnostic tool in CAH Of the 30 recalled patients with confirmed CAH the diagnosis had been clinically suspected in 15 patients. In the remaining 15 patients the diagnosis had not been considered clinically and was evaluated only after a positive screening test. In the 1 patient mentioned with the false negative screening test, the diagnosis had been clinically suspected. Altogether, 15 of those 31 patients (48%) with CAH diagnosed during the evaluation period could profit from screening because the diagnosis had not been recognized clinically. In the female patient group CAH was clinically suspected in 13 patients (76%), whereas in the male patient group the diagnosis had been clinically suspected in only 2 cases (15%). The sensitivity of the clinical suspicion was thus 53% compared with a sensitivity of 97% for the screening. Discussion This is the first study that evaluates the efficiency of neonatal screening for CAH in Switzerland. With 30 patients affected by 21-OHD and 1 patient by 11 -OHD, the incidence of CAH in our screened population was 1 in 10,749 for 21-OHD CAH and 1 in 333,221 for 11 -OHD. The incidence for 21-OHD CAH in our patient groups is slightly higher than the worldwide incidence of CAH, which has been reported to be 1 in 14,199 live births (10); it lies, however, within the figures of CAH incidence recently evaluated in other European populations (11 13). A former study performed in Switzerland during the prescreening era ( ) found an incidence of 21-OHD of 1 in 15,472 (14). The evaluated incidence of CAH in Switzerland has thus increased since the onset of screening; this finding proved to be marginally significant (P 0.059, by 2 test). As we may assume that the All recall patients Male recalls Female recalls Patients (n) CAH confirmed Clinically suspected in confirmed CAH 15/30 (50%) 2/13 (15%) 13/17 (76%) Percentage detected through screening 15/30 (50%) 11/13 (85%) 4/17 (24%) Fifty percent of all CAH cases were clinically missed. In boys the rate of missed cases was higher than in girls. real incidence, despite possible bias in the cohorts caused by immigration into Switzerland throughout the last decades, has not changed significantly during the 2 evaluation periods compared, we conclude that without screening 1 of 3 cases with 21-OHD CAH remained previously undiagnosed, and screening for CAH has improved the rate of case detection in Switzerland by one third. The Zurich neonatal screening program, unlike the majority of CAH-screening programs worldwide, does not use a single screening test, but requires follow-up testing on newborns with borderline 17-OHP levels in the first screening test (15, 16). The aim of this strategy is to reduce the high rate of false positive findings, particularly in preterm infants, which is one of the major critical aspects of CAH newborn screening, because it induces at the same time both costs through necessary further evaluation and unnecessary anxiety in the families of the false positive reported newborns (8, 17). Our approach results in a very low recall rate (0.0018%) and a very high positive predictive value of 50% of the recalls compared with other recently published studies on CAH screening in Europe and the United States, as well as in those with weight-adjusted criteria for 17-OHP levels (11, 12, 17 19). Consequently, our rate of false positive recalls was lower than the international average (17). Our screening strategy thus proves to be highly efficient in all patient groups and seems to be particularly successful in reducing the number of false positive test results in the group of the preterm infants. The group of the CAH-confirmed recalls had a higher gestational age compared with the group of false positive recalls, which again reflects the high positive predictive value of our screening in the group of the term infants. At the same time the mean 17-OHP concentration was clearly higher in the patient group with confirmed CAH than in the group in which CAH was not confirmed. From these results we conclude that the positive predictive value of our screening test is directly associated with increasing gestational age and increasing 17-OHP levels. The sex distribution of CAH in our screened population was 13 boys (41%) vs. 18 girls (59%). A similar sex distribution with a lower prevalence of CAH in males has been reported in other studies performed in European popula-

4 Steigert et al. Screening for CAH in Switzerland J Clin Endocrinol Metab, September 2002, 87(9): tions; this difference in sex distribution, however, has not been documented in U.S. populations (11 13, 18, 20). The finding of a lower prevalence in males in our patient group is nearly identical to that of a former case survey-based incidence study also performed in Switzerland 20 yr ago that showed a sex distribution of 44% boys and 56% girls (14). However, our patient number in this study may not be big enough to allow interpretation of these data. In half of the patients with CAH the diagnosis had not been anticipated clinically. More than 90% of the male patient group and almost one quarter of all female CAH patients were detected through screening alone. These findings are comparable to the data reported by others, many of them recently published and based on European populations, and emphasize that with clinical skills alone about half of the cases of CAH are not detected in the newborn period (11, 12, 17, 21). The sensitivity of our screening method for CAH was 97%. The one patient with CAH who was not detected in the regular screening on the fourth day of life had a markedly elevated 17-OHP level in the same test design 5 d later; we cannot offer a satisfactory explanation for this remarkable constellation. It is in any case yet another reminder that a screening program cannot replace clinical skills and that the two diagnostic tools of clinical examination and screening should be used in combination to achieve the best possible results in the detection of CAH, as has been documented previously (12). Late-onset CAH was only diagnosed once in our screening population. The screening is not designed to detect this nonclassical form of CAH, because in some patients with this mild form of CAH the 17-OHP values at birth are not elevated, and even though in some other patients with lateonset CAH the biochemical defect might be evident in the newborn period and thus lead to mildly elevated 17-OHP values at the time of screening, the identification of this mild form of CAH would require lower cut-off values, most likely giving rise to a significant increase in unwanted false positive recalls (9). In those patients with CAH for whom we had data on the onset of pharmacological therapy, we found the average time of therapy start to be at the age of 6.7 d, with a range from 1 22 d; these figures indicate that in comparison with other published data, our screening program enables us to begin therapy early (11, 13, 17, 20). Five of the 30 patients with CAH showed the first clinical symptoms of salt wasting by the time the recall reached the patient. Salt-wasting crisis usually has its earliest onset at the end of the first week of life, but may not appear until the second month of life (1, 2, 17). This is beyond the time of our average therapy onset. Furthermore, as we have no knowledge of any case of severe salt-wasting crisis in our screened population, we conclude that the short recall time of our screening strategy succeeded in detecting all patients at risk of developing salt-wasting crisis early enough to introduce timely therapy and thus prevent adverse outcome. Two patients with steroidogenic defects other than 21- OHD or 11 -OHD (the girl with 3 -OHD and the boy with 17 -OHD) could, as expected, not be detected through our screening, because these metabolic disorders do not cause elevated 17-OHP levels. In conclusion, with a sensitivity of 97%, a positive predictive value of 50%, and a specificity of 99.99%, our data indicate that the Zurich neonatal screening program for CAH is highly effective. By using gestational age-related cut-off values, the screening has significantly reduced the number of unwanted recalls, particularly in premature infants. It efficiently distinguishes affected from nonaffected newborns. As we have no knowledge of any newborn enrolled in the screening program undergoing severe salt-wasting crisis, we conclude that our screening program was able to detect all patients with a potential risk of developing saltwasting crisis through timely and adequate therapy. Longterm follow-up studies will provide information on the proportion of cases with late-onset CAH that cannot be detected through neonatal screening because of slowly or late rising 17-OHP levels. Acknowledgments We thank the staff of our laboratory and Lissy Alig in particular for her technical assistance, and Luciano Molinari for the statistical analysis. Received January 3, Accepted May 16, Address all correspondence and requests for reprints to: Eugen J. Schoenle, M.D., Department of Endocrinology and Diabetology, University Children s Hospital, Steinwiesstrasse 75, CH-8032 Zurich, Switzerland. schoenle@kispi.unizh.ch. This work was supported in part by Swiss National Science Foundation Grant These data were first presented at the LWPES/ ESPE Sixth Joint Meeting, Montreal, Canada, July References 1. American Academy of Pediatrics 2000 Technical report: congenital adrenal hyperplasia. Pediatrics 106: Miller WL 1994 Genetics, diagnosis and management of 21-hydroxylase deficiency. J Clin Endocrinol Metab 78: Miller WL, Levine LS 1987 Molecular and clinical advances in congenital adrenal hyperplasia. J Pediatr 111: Hughes IA 1998 Congenital adrenal hyperplasia: a continuum of disorders. Lancet 352: Holtzman NA 1978 Newborn screening for inborn errors of metabolism. Pediatr Clin North Am 25: Honour JW, Torresani T 2001 Evaluation of neonatal screening for congenital adrenal hyperplasia. Horm Res 55: Pang S, Hotchkiss J, Drash AL, Levine LS, New MJ 1977 Microfilter paper method for 17-hydroxyprogesterone radioimmunoassay: its application for rapid screening for congenital adrenal hyperplasia. J Clin Endocrinol Metab 45: Torresani T, Grueters A, Scherz R, Burckhardt JJ, Harras A, Zachman M 1994 Improving the efficacy of newborn screening for congenital adrenal hyperplasia by adjusting the cut-off level of 17 -hydroxyprogesterone to gestational age. Screening 3: Working Group on Neonatal Screening of the European Society for Pediatric Endocrinology 2001 Procedure for neonatal screening of the European Society for Pediatric Endocrinology. Horm Res 55: Pang S, Wallace MA, Hofman L, Thuline HC, Dorche C, Lyon I, Dobins RH, Kling S, Fujieda K, Suwa S 1998 Worldwide experience in newborn screening for classical congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Pediatrics 81: Cartigny-Maciejewski M, Guilley N, Vanderbecken S, Gonde S, Stuckens C, Ponte C, Weill J, Farriaux JP, Paux E 1999 Le dépistage neonatal de l hyperplasie congenitale des surrénales par deficit en 21-hydroxylase: experience lilloise Arch Pédiatr 6: Balsamo A, Cacciari E, Piazzi S, Cassio A, Bozza D, Pirazzoli P, Zappulla F 1996 Congenital adrenal hyperplasia: neonatal mass screening compared with clinical diagnosis only in the Emilia-Romagna region of Italy, Pediatrics 98: Thilén A, Nordenström A, Hagenfeldt L, von Döbeln U, Guthenberg C, Larsson A 1998 Benefits of neonatal screening for congenital adrenal hyperplasia (21-hydroxylase deficiency) in Sweden. Pediatrics 101:e11

5 4110 J Clin Endocrinol Metab, September 2002, 87(9): Steigert et al. Screening for CAH in Switzerland 14. Werder EA, Siebenmann RE, Knorr-Mürset G, Zimmermann A, Sizonenko PC, Theintz G, Girard J, Zachmann M, Prader A 1979 The incidence of congenital adrenal hyperplasia in Switzerland: a survey of patients born in 1960 to Helv Pediatr Acta 35: Pang S, Shook MK 1997 Current status of neonatal screening for congenital adrenal hyperplasia. Curr Opin Pediatr 9: The National Newborn Screening and Genetics Resource Center 2001 National Newborn Screening Report Austin, Texas: NNSGRC; Pang S, Clark A 1993 Congenital adrenal hyperplasia due to 21-hydroxylase deficiency: newborn screening and its relationship to the diagnosis and treatment of the disorder. Screening 2: Allen DB, Hoffman GL, Fitzpatrick P, Laessig R, Maby S, Slyper A 1997 Improved precision of newborn screening for congenital adrenal hyperplasia using weight-adjusted criteria for 17-hydroxyprogesterone levels. J Pediatr 130: Kwon C, Farrell PM 2000 The magnitude and challenge of false positive newborn screening test results. Arch Pediatr Adolesc Med 145: Kovacs J, Votava F, Heinze G, Solyom J, Lebl J, Pribilincova S, Frisch H, Battelino T 2001 Lessons from thirty years of clinical diagnosis and treatment of congenital adrenal hyperplasia in five middle European countries. J Clin Endocrinol Metab 88: Thilén A, Larsson A 1990 Congenital adrenal hyperplasia in Sweden Prevalence, symptoms and age at diagnosis. Acta Pediatr Scand 79: