The management of cystic fibrosis (CF) has improved. clinical investigations. Factors Influencing Outcomes in Cystic Fibrosis* A Center-Based Analysis

clinical investigations Factors Influencing Outcomes in Cystic Fibrosis* A Center-Based Analysis Charles Johnson, MB, ChB; Steven M. Butler, PhD; Michael W. Konstan, MD; Wayne Morgan, MD; and Mary Ellen B. Wohl, MD on behalf of the Scientific Advisory Group and the Investigators of the Epidemiologic Study of Cystic Fibrosis Context: Guidelines for managing cystic fibrosis (CF) patients have been widely circulated, but little is known about the variations in practice between sites and their association with outcomes. Objective: To determine whether differences in lung health existed between groups of patients attending different CF care sites and to determine whether these differences are associated with differences in monitoring and intervention. Design: The analysis was conducted using data from the Epidemiologic Study of Cystic Fibrosis from 1995 through 1996. Setting: This was an observational database collecting prospective information from a large number of CF patients undergoing routine care in North America. Participants: Participating sites that had at least 50 CF patients who had each made at least one visit to a center during the 2-year study period were ranked on the basis of median values for FEV 1 within each of three age groups (6 to 12 years, 13 to 17 years, and > 18 years). Interventions: There were no prespecified interventions in this observational study. Main outcome measures: The frequency of patient monitoring and the use of therapeutic interventions were compared between sites in the upper and lower quartiles after stratification within the site for disease severity. Results: Within-site rankings tended to be consistent across the three age groups. Patients who were treated at higher ranking sites had more frequent monitoring of their clinical status, measurements of lung function, and cultures for respiratory pathogens. These patients also received more interventions, particularly IV antibiotics for pulmonary exacerbations. Conclusion: We found substantial differences in lung health across different CF care sites. We found that frequent monitoring and increased use of appropriate medications in the management of CF are associated with improved outcomes. (CHEST 2003; 123:20 27) Key words: cystic fibrosis; epidemiology; evaluation; monitoring; spirometry; respiratory tract culture; treatment Abbreviations: CF cystic fibrosis; ESCF Epidemiologic Study of Cystic Fibrosis The management of cystic fibrosis (CF) has improved dramatically in the 60 years since the early description of the disease by Anderson, 1 when the life expectancy of patients was 1 year. Currently, data from a number of registries estimate the median age of survival to be 31 years. 2,3 This *From Genentech (Drs. Johnson and Butler), South San Francisco, CA; the Department of Pediatrics (Dr. Konstan), Case Western Reserve University School of Medicine, Cleveland, OH; Arizona Health Sciences Center (Dr. Morgan), Tucson, AZ; and the Department of Pediatrics (Dr. Wohl), Harvard Medical School, Boston, MA. Drs. Johnson and Butler are employees of Genentech. Drs. Konstan, Morgan, and Wohl are consultants for the Epidemiologic Study remarkable advance has been achieved through the establishment of specialized care facilities, improved pancreatic enzyme preparations, and the development of effective antibiotics with which to treat pulmonary exacerbations. 4,5 Guidelines developed by the Cystic Fibrosis Foundation 6 have contributed of Cystic Fibrosis sponsored by Genentech. All sources of support for the Epidemiologic Study of Cystic Fibrosis in the form of grants, case report forms, and data analysis were provided by Genentech. Manuscript received November 19, 2001; revision accepted May 22, 2002. Correspondence to: Charles Johnson, MB,ChB, Genentech, 1 DNA Way, Mail Stop 59, South San Francisco, CA 94080; e-mail: johnson.charles@gene.com 20 Clinical Investigations

to the standardization of CF management. These guidelines recommend that patients have at least four clinical visits per year, that the measurement For editorial comment see page 1 of lung function be performed every 6 months, and that cultures of respiratory tract secretions be conducted annually. In assessing outcomes for CF patients, the ideal measure is mortality. However, relatively few patients die in the short term. Therefore, measurements of pulmonary function have become surrogate outcomes in most clinical trials, 7 10 as there is a strong association between lung function and mortality rates. 11 13 The Epidemiologic Study of Cystic Fibrosis (ESCF) is a multicenter, longitudinal, observational study that prospectively collects detailed clinical, therapeutic, microbiological, and lung function data from a large number of CF treatment sites in the United States and Canada. 14 16 Initiated in December 1993, the study enrolled 18,411 patients by December 31, 1995. The goal of this study was to identify whether differences in outcomes, specifically lung health as ascertained using the surrogate marker of FEV 1, existed between care sites. We sought to determine whether particular practice patterns at care sites were associated with better outcomes. In light of data linking improved survival to the frequent use of IV antibiotics, 17 an additional goal of the study was to determine whether intensive treatment with IV antibiotics was associated with better outcomes. Materials and Methods The analysis examined ESCF data collected during the 2-year period from 1995 through 1996. Initiated in December 1993, the study had enrolled 18,411 patients by December 1995, 14 which was estimated to include 80% of the population of CF patients in the United States and approximately 10% of that in Canada. The 194 sites included 90% of those accredited by the US Cystic Fibrosis Foundation. All Canadian and US sites were invited to participate if they had 10 patients. Patients were enrolled during routine care visits. Institutional review boards for each site reviewed the study. Many did not require informed consent as patient anonymity was maintained. Data were collected prospectively at each visit or hospitalization using studyspecific case report forms. Every effort was made to ensure adherence to data collection by providing patient-specific data to the sites for source verification. In addition, data comparing the site to regional and national data were provided to investigators for external validation. Data-checking programs were run by the statistical coordinating center (ClinTrials Research, Inc; Lexington, KY), and queries were sent to the sites in order to resolve discrepancies and outliers. The data included findings for every spirometry test (designated as having been obtained when sick or stable ), the results of every culture, and every antipseudomonal antibiotic intervention and/or hospitalization. Routine therapies were recorded at each visit, including the use of airway clearance techniques, oral bronchodilators, inhaled bronchodilators, oral corticosteroids, inhaled corticosteroids, inhaled cromolyn or nedocromil, inhaled dornase alfa, nonsteroidal antiinflammatory agents, insulin or oral hypoglycemic agents, nutritional supplements, pancreatic enzymes, and oral antibiotics. The start and stop dates of therapy with oral quinolones, inhaled antibiotics, and IV antibiotics also were recorded. All of these variables were examined in the analysis. A ranking was used to evaluate the use of antibiotics so that if inhaled and IV antibiotics were used simultaneously, the patient fell into the IV antibiotic group. The three age groups (6 to 12 years, 13 to 17 years, and 18 years) at each site were ranked on the basis of the last stable median FEV 1 percent predicted value. 18 The upper and lower quartiles of these groups were compared. The patients in the upper and lower quartiles of each age group were pooled (Table 1) and stratified by FEV 1 for the comparison of practices (Table 2), outcomes (Table 3), and interventions (Tables 4 and 5). The definitions of severity used the CF Foundation categories of 40% predicted of FEV 1 values for severe disease and 40 to 69% predicted for moderate disease, with a modification of 70 to 99% predicted for mild disease and 100% predicted for normal. This definition is consistent with the CF groupings, the population of which was distributed so that the severity groups had sufficient numbers for analysis. If there were 10 patients in a specific age group at an individual site, the site was excluded from the evaluation of that age group. To be included in this analysis, each center had to have at least 50 patients who had made at least one visit during the 2-year period and had at least one spirometry test value. Within each age group, the median value for the FEV 1 percent predicted in each age group at each site was calculated, and the sites were ranked from lowest to highest according to this median value. The sites then were divided into quartiles so that those Table 1 Characteristics of CF Patients at Upper and Lower Quartile Sites by Age Group Ranking Sites, No. 28 28 21 21 26 26 Patients, No. 1,028 682 458 319 805 859 FEV 1,* predicted 93 74 85 62 65 50 Weight for age,* percentile 35 31 33 20 37 30 *Mean values for patients were pooled from all sites in the quartile. www.chestjournal.org CHEST / 123 / 1/ JANUARY, 2003 21

Table 2 Mean Number of Visits, Spirometry Readings, and Cultures over a 2-Year Period for Patients at Upper and Lower Quartile Sites by Age Group and Severity of Lung Disease* Ranking FEV 1,% predicted Patients 40 20 64 21 69 149 380 40 69 112 212 80 128 331 292 70 99 511 314 245 101 261 153 100 385 92 112 21 64 34 Total 1,028 682 458 319 805 859 Visits 40 12.9 13.9 15.0 14.3 11.4 11.0 40 69 11.7 10.6 11.4 10.9 9.7 8.6 70 99 8.6 7.6 8.5 7.5 7.3 7.4 100 8.0 7.0 6.9 5.6 6.5 6.2 p value 0.001 0.001 0.003 Spirometries 40 12.5 10.6 18.5 12.0 12.9 8.6 40 69 11.1 9.3 12.6 8.9 10.5 6.9 70 99 7.2 5.9 8.5 5.8 7.3 5.5 100 5.7 5.5 6.5 3.7 6.0 4.3 p value 0.001 0.001 0.001 Cultures 40 8.0 5.7 7.1 7.0 6.5 6.5 40 69 5.7 3.6 5.3 4.4 5.3 4.7 70 99 4.3 2.6 4.0 2.9 4.1 3.8 100 3.9 2.2 4.0 2.2 4.2 3.3 p value 0.001 0.007 0.007 *p Values were calculated using stratified Wilcoxon rank sum tests comparing upper and lower quartile sites. with the highest values (upper quartile) could be compared with those having the lowest values (lower quartile). For comparisons of practices and interventions between the quartiles, which are markedly affected by disease severity, patients within each quartile were pooled and stratified by disease severity based on FEV 1 (ie, 40% of predicted, 40 to 69% of predicted, 70 to 99% of predicted, and 100% of predicted). This allowed the sites in the upper and lower quartiles to be compared based on their practices within groups of patients with similar lung health. Statistical Methods For each possible pair of the three age groups, the consistency of site rankings between the two age groups was assessed using Spearman rank correlation coefficients. Statistical comparison of upper quartile sites with lower quartile sites were done using Mantel-Haenszel tests for dichotomous variables and Wilcoxon rank sum tests for counts of events or durations of therapy, with stratification by FEV 1 (ie, 40% of predicted, 40 to 69% of predicted, 70 to 99% of predicted, and 100% of predicted) so that practices were compared within patients of a similar severity level. Stratified descriptive statistics included either proportions for dichotomous variables, means for counts of events or medians for duration of therapy. The quartiles also were compared globally on the overall mean FEV 1 percent predicted and weight for age percentile using t tests. Results Of the 194 participating sites in the ESCF, 132 with 8,125 patients met the eligibility criteria for this analysis. For the 6 to 12-year-old age group, 112 sites with 3,361 patients (mean, 30 patients per site; SD, 19 patients per site) were eligible. For the adolescent age group (13 to 17 years of age), 84 sites with 1,734 patients (mean, 21 patients per site; SD, 11 patients per site) were eligible. For the adult group ( 18 years of age), 103 sites with 3,030 patients (mean, 29 patients per site; SD, 24 patients per site) were eligible. The observation that a consistent trend of monitoring and intervention was observed across the four quartiles (ie, that the middle quartiles fell between the upper and lower extremes) allowed the subsequent analysis to be restricted to the upper and lower quartiles. For sites that qualified with at least two age groups, the rankings of sites tended to be consistent across the age groups with Spearman rank correlation coefficients of 0.30 for ages 6 to 12 years and 13 to 17 years (p 0.007; 80 patients), 0.29 for ages 6 to 12 years and 18 years (p 0.007; 84 patients), and 0.17 for ages 13 to 17 years and 18 years (p 0.17; 69 patients). Considering each pair of age groups, very few sites (six or fewer) were in the upper quartile in one age group and the lower quartile in the other age group. The observed differences between upper and lower quartile sites in disease severity, as characterized by FEV 1 percent predicted, were substantial, ranging from 15 to 23% of age points (Table 1). Comparing patients within each of the severity groups, the monitoring of events (ie, number of visits, spirometry testing, and cultures) occurred 22 Clinical Investigations

Table 3 Patients With Detection of Each Microorganism at Upper and Lower Quartile Sites by Age Group and Severity of Lung Disease* Organism FEV 1, % predicted P aeruginosa Standard 40 90.0 87.5 85.7 88.4 85.9 87.4 40 69 69.6 78.3 82.5 89.1 84.9 83.2 70 99 56.0 61.2 70.2 67.3 70.1 73.9 100 41.3 40.2 67.0 57.1 67.2 67.7 p value 0.09 0.98 0.72 Multiply resistant 40 10.0 25.0 33.3 23.2 30.2 25.5 40 69 8.0 8.5 11.3 15.6 23.9 15.1 70 99 3.3 3.8 10.2 8.9 14.9 3.9 100 2.3 2.2 8.0 0.0 14.1 5.9 p value 0.39 0.65 0.001 B cepacia 40 5.0 3.1 19.1 8.7 14.1 10.0 40 69 4.5 3.3 8.8 6.3 8.8 4.8 70 99 2.5 2.9 7.4 5.0 6.9 3.3 100 1.0 1.1 3.6 0.0 4.7 2.9 p value 0.87 0.10 0.005 S maltophilia 40 20.0 9.4 9.5 7.3 14.8 10.3 40 69 14.3 5.2 10.0 7.8 7.6 6.9 70 99 6.7 3.2 6.1 6.9 6.5 3.9 100 4.2 0.0 11.6 0.0 4.7 0.0 p value 0.001 0.38 0.07 *Values given as percentage of patients, unless otherwise indicated. p Values were calculated using stratified Mantel-Haenszel tests of independence comparing upper and lower quartile sites. more frequently at upper quartile sites (Table 2). These trends were highly statistically significant in each age group. The greatest differences in the number of cultures were seen in the youngest age group, which had 40 to 77% more cultures performed at upper quartile sites than at lower quartile sites. Certain microorganisms were reported more frequently at upper quartile sites (Table 3). Compared with the lower quartile sites, there was a higher incidence of patients with at least one culture positive for Pseudomonas aeruginosa and for at least one strain of this organism that was reported as multiply resistant. Also, compared with the lower quartile sites, there were more patients with at least one culture positive for Burkholderia cepacia, and for Stenotrophomonas maltophilia at the upper quartile sites. Some therapies were used more frequently at the upper quartile sites (Table 4). Upper quartile sites administered oral corticosteroids and inhaled cromolyn or nedocromil more frequently than did lower quartile sites consistently across all age and diseaseseverity groups. Very large differences were seen in the use of inhaled cromolyn and nedocromil, ranging from 10 to 42% (p 0.001). Oral nonquinolone antibiotics also were used more frequently in upper quartile sites. Prophylactic inhaled antibiotics (ie, aminoglycoside and quinolones) were used more frequently in young pediatric patients. Among other routine therapies, substantial and significant differences were observed only in adults. In this age group, upper quartile sites administered more inhaled bronchodilators, more inhaled corticosteroids, and more dornase alfa than lower quartile sites (Table 4). No differences were seen in the use of pancreatic enzymes and airway clearance techniques, which were used by 90% of patients in almost all age and disease-severity groups. No attempts were made to quantify airway clearance techniques. For patients who were 18 years of age, upper quartile sites tended to report more frequent treatments of exacerbations with IV antibiotics (Table 5). These differences were proportionately largest in the relatively healthy patients, particularly in the age group of 13 to 17 years. Upper quartile sites also tended to treat patients for longer periods, regardless of disease severity. This trend was particularly strong in adult patients, for whom the increased duration was highly significant (p 0.001). Upper quartile sites also reported more frequent treatments of exacerbations with inhaled antibiotics in patients 18 years, with differences once again proportionately largest in the relatively healthy patients. In contrast, in patients 18 years of age the reported www.chestjournal.org CHEST / 123 / 1/ JANUARY, 2003 23

Table 4 Use of Specific Routine Therapies over a 2-Year Period by Patients at Upper and Lower Quartile Sites by Age Group and Severity of Lung Disease* Therapy FEV 1, % predicted Prophylactic inhaled antibiotics 40 85.0 35.9 52.4 63.8 41.0 53.4 40 69 49.1 33.5 48.8 35.2 37.2 34.6 70 99 23.9 16.6 22.9 24.8 28.0 15.7 100 10.4 10.9 16.1 19.1 18.8 14.7 p value 0.001 0.59 0.71 Nonquinolone oral antibiotics 40 80.0 87.5 85.7 87.0 83.9 75.5 40 69 83.0 89.6 93.8 81.3 81.3 63.4 70 99 85.3 79.3 84.5 71.3 77.0 61.4 100 84.4 79.4 82.1 61.9 79.7 52.9 p value 0.17 0.001 0.001 Cromolyn or nedocromil 40 75.0 32.8 47.6 30.4 45.6 34.0 40 69 50.9 30.2 43.8 32.8 36.3 25.7 70 99 51.9 28.3 40.0 23.7 33.3 20.9 100 40.5 26.1 32.1 14.3 35.9 17.7 p value 0.001 0.001 0.001 Inhaled corticosteroids 40 30.0 43.8 47.6 46.4 59.7 46.6 40 69 38.4 39.6 46.3 43.0 50.1 36.0 70 99 37.8 32.5 45.3 40.6 47.1 36.6 100 41.6 33.7 31.3 28.6 37.5 17.7 p value 0.15 0.38 0.001 Oral corticosteroids 40 50.0 26.6 42.9 34.8 46.3 28.2 40 69 47.3 27.4 35.0 24.2 35.1 20.6 70 99 27.8 12.7 28.6 17.8 26.1 16.3 100 23.4 9.8 20.5 14.3 23.4 17.7 p value 0.001 0.005 0.001 Inhaled bronchodilators 40 90.0 96.9 95.2 98.6 100.0 96.3 40 69 94.6 93.4 96.3 96.9 95.2 87.3 70 99 89.6 86.0 87.4 84.2 88.1 75.2 100 80.0 80.4 77.7 81.0 85.9 64.7 p value 0.38 0.81 0.001 Dornase alfa 40 95.0 87.5 95.2 94.2 89.9 78.2 40 69 69.6 75.5 86.3 85.9 78.6 68.5 70 99 53.8 52.2 60.0 65.4 62.5 46.4 100 37.1 44.6 50.0 52.4 45.3 47.1 p value 0.54 0.46 0.001 *Values given as percentage of patients, unless otherwise indicated. p Values were calculated using stratified Mantel-Haenszel tests of independence comparing upper and lower quartile sites. use of oral quinolones for the treatment of exacerbations tended to be greater in lower quartile sites, except in the youngest, sickest patients (6 to 12 years of age, 40% predicted FEV 1 ). In adults, there was no significant difference between upper and lower quartile sites for the use of inhaled antibiotics or oral quinolones. Because of less accurate documentation of stop dates for inhaled antibiotics and oral quinolones, the duration of treatment could not be reliably assessed. Discussion This analysis takes advantage of the large number of sites and patients reporting to the ESCF. It presents an opportunity to study outcomes and practices in a comprehensive way. The approach was to compare CF care sites based on an important index of lung health (ie, the percentage of predicted FEV 1 values). Large differences in lung health across sites were demonstrated. The sites reporting the highest values for median FEV 1 monitored patients, obtained measurements of lung function, and obtained cultures for respiratory pathogens more frequently. The differences in monitoring were most striking for spirometry testing and respiratory cultures, and were less marked for visit frequency. Also at these sites, more patients of all ages were prescribed inhaled cromolyn or nedocromil and oral corticosteroids for all degrees of disease severity. Adult patients at these sites were prescribed bronchodilator agents, dornase alfa, and inhaled cortico- 24 Clinical Investigations

Table 5 Treatment of Pulmonary Exacerbations by Route over a 2-Year Period by Patients at Upper and Lower Quartile Sites by Age Group and Severity of Lung Disease* Therapy FEV 1, % predicted IV antibiotic treatments No. 40 2.4 2.4 3.4 3.2 2.4 2.8 40 69 1.8 1.4 1.8 1.5 1.7 1.6 70 99 0.5 0.5 0.9 0.5 0.9 0.9 100 0.3 0.2 0.5 0.2 0.4 0.7 p value 0.10 0.001 0.91 Median duration, d 40 15.0 21.0 25.4 19.0 21.3 17.6 40 69 18.7 15.5 17.0 16.0 17.7 16.0 70 99 15.6 12.5 15.0 15.0 16.0 15.5 100 12.0 15.0 15.0 11.0 15.0 13.5 p value 0.11 0.01 0.001 Inhaled antibiotic treatments 40 1.0 0.9 1.2 1.3 1.1 1.1 40 69 1.1 0.9 0.9 0.8 0.9 0.8 70 99 0.7 0.4 0.7 0.5 0.6 0.6 100 0.4 0.3 0.4 0.3 0.5 0.6 p value 0.001 0.07 0.18 Oral quinolone treatments 40 2.2 1.3 1.5 1.5 1.4 1.5 40 69 1.0 1.2 1.2 1.7 1.3 1.5 70 99 0.4 0.5 0.8 1.2 1.2 1.1 100 0.3 0.4 0.6 0.8 0.7 1.2 p value 0.10 0.004 0.98 *Values given as mean No. of treatments per patient, unless otherwise indicated. p Values were calculated using stratified Wilcoxon rank sum tests comparing upper and lower quartile sites. steroids more frequently than their lower quartile counterparts. At upper quartile sites, IV antibiotics were used more frequently and for longer duration. It is possible that this increased use of antibiotics may have resulted in the increase in prevalence of bacterial strains, including resistant strains that were noted at the upper quartile sites. However, we have reported previously an increase in the detection of organisms and resistance at sites that monitor more frequently. 19 Data were presented only for the upper and lower quartile sites, but the results are further supported by the fact that for the two middle quartiles the results generally lie between those for the upper and lower quartiles. There are exceptions to this generalization, but these are primarily in the severely ill patients where the numbers are relatively small. It also should be noted that the generalizability of these observations may be limited by the exclusion of sites with 50 patients. Also, the decision by a number of Canadian sites not to participate in the study may have affected the interpretation of the results. Possible explanations for the substantial differences in lung function results between the upper and lower quartile sites include differences in the patterns of practice, but they might also include genetic or other differences in the local patient populations, differences in local conditions including climate, the distance that patients live from the site, and socioeconomic status. Data from the CF Foundation Registry indicate that in 1996 only 226 of the 19,570 patients (1.4%) reported a lack of insurance coverage, so at least this aspect of socioeconomic status is not likely to be an important factor. 20 Upper quartile sites were somewhat more likely to be located in the northeastern region of the United States, and, in the pediatric age groups only, there was a tendency toward larger sites in the upper quartile compared to the lower quartile. Although the results of this study suggest that the upper quartile sites were more likely to perform special assessments and prescribe some therapies, another possible explanation that always exists in observational studies is more thorough reporting of these events at upper quartile sites. To guard against this possibility, each site received patient-specific data reports and was requested to verify the accuracy of their data. The use of nonsteroidal anti-inflammatory drugs was not considered because, at the time of this study, the ESCF did not distinguish high-dose ibuprofen therapy 9 from the analgesic use of nonsteroidal anti-inflammatory medications. Oral nonquinolone antibiotics were extensively used for treatment of patients across all age and severity groups. Those antibiotics were used significantly less often in adolescent and adult patients at www.chestjournal.org CHEST / 123 / 1/ JANUARY, 2003 25

the lower quartile centers than in those patients at the upper quartile centers. These data are consistent with those from a study undertaken in the United Kingdom 21 indicating improved outcomes for young CF patients who received long-term antistaphylococcal therapy. Regarding the use of antibiotics to treat pulmonary exacerbations, the most striking differences between upper and lower quartile sites occurred in patients 18 years of age, particularly in the adolescent age group. Adolescents from upper quartile sites received more treatment with IV and inhaled antibiotics, and less treatment with oral quinolones. For IV and inhaled antibiotics, the differences were greater for patients with milder lung disease, suggesting that these patients were treated more aggressively. The duration of treatment with IV antibiotics was also greater in upper quartile sites in adolescents, as well as in adults. More intensive use of IV and inhaled antibiotics at upper quartile sites might account for the better lung function observed in patients at these sites. The identification of a pulmonary exacerbation warranting antibiotic therapy, particularly IV therapy, may be related to the frequency of obtaining spirometry, since a fall in FEV 1 values commonly results in treatment for a pulmonary exacerbation. Whatever leads to the decision to treat, it appears that the upper quartile sites are treating patients more often, particularly those considered to have mild lung disease. One possible conclusion from these data is that the use of antibiotics contributes to the improved outcomes seen in patients at upper quartile sites. Support for intensive antibiotic therapy comes from data reported in a Danish study, 17 which have shown that a program of quarterly hospital admissions for IV antipseudomonal therapy is associated with improved survival. Further support for intensive antibiotic therapy also comes from studies of inhaled, high-dose tobramycin. 10 However, other factors in Denmark, such as attention to nutritional support and access to state-supported health care facilities, also may play a role in improved outcomes. These data suggest a greater prevalence of multiply resistant P aeruginosa, B cepacia, and S maltophilia at upper quartile sites. This could represent an important and disturbing association between increased antibiotic use and the emergence of resistant strains. It is possible that this results from an ascertainment bias due to the higher frequency of cultures in the upper quartile sites. We have shown previously 20 that increased frequency of cultures and better culture techniques result in the increased identification of resistant organisms. This observation highlights the need for continued research into improved antimicrobial agents and a better understanding of the mechanisms of colonization and infection of the lung and the development of antibiotic resistance in CF. Both inhaled cromolyn or nedocromil and oral corticosteroids were administered more often at upper quartile sites. The association between poor outcomes in CF patients and a vigorous inflammatory response was first noted by Wheeler et al 22 and led to a large multicenter trial 23 of oral corticosteroid use in CF patients. Although this study showed a clear benefit in terms of the maintenance of pulmonary function, a 1996 review 4 stated that the adverse effects on linear growth and glucose metabolism should limit the long-term use of oral corticosteroids. This analysis demonstrates the widespread use of both oral and inhaled corticosteroids but does not distinguish between long-term and sporadic use. A surprising result of this analysis is the finding of an increased administration of inhaled cromolyn or nedocromil at upper quartile sites. These agents are known to modulate the inflammatory response but are also reported to block non-cf transmembrane receptor chloride channels. 24,25 Cromolyn-mediated or nedocromil-mediated blocking of chloride channels would appear to be counterproductive. The strength of the observed association warrants investigation into possible mechanisms that might explain this paradox. In summary, this large observational study of CF demonstrates significant and clinically relevant differences across sites in lung health as measured by FEV 1. The striking finding of this analysis is the consistency with which the sites in the upper quartile, regardless of age and disease severity, monitor and treat patients more frequently than do lower quartile centers. This association is particularly strong for younger and relatively healthy CF patients. These results suggest, but do not prove, the hypothesis that frequent monitoring and more interventions, including the more frequent use of IV antibiotics for longer duration, potentially result in better outcomes for CF patients. References 1 Anderson DH. Cystic fibrosis of the pancreas and its relation to celiac disease. Am J Dis Child 1938; 344 399 2 Cystic Fibrosis Foundation Patients Registry. 1999 annual data report. Bethesda, MD: Cystic Fibrosis Foundation Patients Registry, 2000 3 Corey M, Farewell V. Determinants of mortality from cystic fibrosis in Canada, 1970 1989. Am J Epidemiol 1996; 143: 1007 1017 4 Ramsey BW. Management of pulmonary disease in patients with cystic fibrosis. N Engl J Med 1996; 335:179 188 5 Davis PB, Drumm M, Konstan MW. Cystic fibrosis. Am J Respir Crit Care Med 1996; 154:1229 1256 6 The Cystic Fibrosis Foundation Center Committee and 26 Clinical Investigations

Guidelines Subcommittee. Cystic fibrosis foundation guidelines for patient services, evaluation, and monitoring in cystic fibrosis centers. Am J Dis Child 1990; 144:1311 1312 7 Ramsey BW, Boat TF. Outcome measures for clinical trials in cystic fibrosis: summary of a cystic fibrosis foundation consensus conference. J Pediatr 1994; 124:177 192 8 Fuchs HJ, Borowitz DS, Christiansen DH, et al. Effect of aerosolized recombinant human DNase on exacerbations of respiratory symptoms and on pulmonary function in patients with cystic fibrosis. N Engl J Med 1994; 331:637 642 9 Konstan MW, Byard PJ, Hoppel CL, et al. Effect of high-dose ibuprofen in patients with cystic fibrosis. N Engl J Med 1995; 332:848 854 10 Ramsey BW, Pepe MS, Quan JM, et al. Intermittent administration of inhaled tobramycin in patients with cystic fibrosis: cystic fibrosis inhaled tobramycin study group. N Engl J Med 1999; 340:23 30 11 Kerem E, Reisman J, Corey M, et al. Prediction of mortality in patients with cystic fibrosis. N Engl J Med 1992; 326:1187 1191 12 Corey M, Edwards L, Levison H, et al. Longitudinal analysis of pulmonary function decline in patients with cystic fibrosis. J Pediatr 1997; 131:809 814 13 Milla CE, Warwick WJ. Risk of death in cystic fibrosis patients with severely compromised lung function. Chest 1998; 113:1230 1234 14 Morgan WJ, Butler SM, Johnson CA, et al. Epidemiologic study of cystic fibrosis: design and implementation of a prospective, multicenter, observational study of patients with cystic fibrosis in the US and Canada. Pediatr Pulmonol 1999; 28:231 241 15 Konstan MW, Butler SM, Schidlow DV, et al. Patterns of medical practice in cystic fibrosis: part I. Evaluation and monitoring of health status of patients. Pediatr Pulmonol 1999; 28:242 247 16 Konstan MW, Butler SM, Schidlow DV, et al. Patterns of medical practice in cystic fibrosis: part II. Use of therapies. Pediatr Pulmonol 1999; 28:248 254 17 Frederiksen B, Lanng S, Koch C, et al. Improved survival in the Danish center-treated cystic fibrosis patients: results of aggressive treatment. Pediatr Pulmonol 1996; 21:153 158 18 Knudson RJ, Lebowitz MD, Holberg CJ, et al. Changes in the normal maximal expiratory flow-volume curve with growth and aging. Am Rev Respir Dis 1983; 127:725 734 19 Shreve M, Butler S, Kaplowitz HJ, et al. Impact of microbiology practice on cumulative prevalence of respiratory tract bacteria in patients with cystic fibrosis. J Clin Microbiol 1999; 37:753 757 20 Cystic Fibrosis Foundation Patient Registry. 1996 annual data report, Bethesda, MD: Cystic Fibrosis Foundation Patient Registry, 1997 21 Weaver LT, Green MR, Nicholson K, et al. Prognosis in cystic fibrosis treated with continuous flucloxacillin from the neonatal period. Arch Dis Child 1994; 70:84 89 22 Wheeler WB, Williams M, Matthews WJ Jr, et al. Progression of cystic fibrosis lung disease as a function of serum immunoglobulin G levels: a 5-year longitudinal study. J Pediatr 1984; 104:695 699 23 Eigen H, Rosenstein BJ, FitzSimmons S, et al. A multicenter study of alternate-day prednisone therapy in patients with cystic fibrosis: Cystic Fibrosis Foundation Prednisone Trial Group. J Pediatr 1995; 126:515 523 24 Penner R, Matthews G, Neher E. Regulation of calcium influx by second messengers in rat mast cells. Nature 1988; 334:499 504 25 Paulmichl M, Li Y, Wickman K, et al. New mammalian chloride channel identified by expression cloning. Nature 1992; 356:238 241 Forthcoming Articles in CHEST Adult Toxicology in Critical Care; Part I: General Approach to the Intoxicated Patient Moklesi and coauthors Accuracy of Pleural Puncture Sites: A Prospective Comparison of Clinical Examination With Ultrasound Diacon and coworkers Editorial comment by Doelken and Strange Accelerated Decline of Lung Function in COPD Patients With Chronic Hepatitis C Virus Infection: A Preliminary Study Based on Small Numbers of Patients Kanazawa and colleagues Interferon Therapy Induces the Improvement of Lung Function by Inhaled Corticosteroid Therapy in Asthmatic Patients With Chronic Hepatitis C Virus Infection: A Preliminary Study Kanazawa and colleagues Prostacyclin Therapy Before Pulmonary Thromboendarterectomy in Patients With Chronic Thromboembolic Pulmonary Hypertension Nagava and coauthors www.chestjournal.org CHEST / 123 / 1/ JANUARY, 2003 27