Late Diagnosis Defines a Unique Population of Long-term Survivors of Cystic Fibrosis
http://www.100md.com
美国呼吸和危急护理医学 2005年第3期
Cystic Fibrosis Center, The University of Colorado Health Sciences Center
National Jewish Medical and Research Center
The Children's Hospital, Denver, Colorado
ABSTRACT
Although the median survival for patients with cystic fibrosis (CF) is 32.9 years, a small group of patients live much longer. We analyzed the genotype and phenotype of CF patients 40 years and older seen between 1992 and 2004 at the National Jewish Medical and Research Center (n = 55). These patients were divided into two groups according to age at diagnosis: an early diagnosis (ED) group, median age at diagnosis 2.0 years (range 0.1eC15 years, n = 28), and a late diagnosis (LD) group, median age of diagnosis 48.8 years (range 24eC72.8 years, n = 27). Consistent with the hypothesis that the CFTR genotype affects the age at diagnosis, CFTR F508 homozygous individuals were more common in the ED group. Although patients in the ED group were predominantly male, the majority of LD patients were female. Patients with CF diagnosed late had a significantly lower prevalence of pancreatic insufficiency and CF-related diabetes, and better lung function. Fewer patients in the LD groups were infected with Pseudomonas aeruginosa, whereas a greater percentage had cultures positive for nontuberculous mycobacteria. This is the largest cohort of older patients with CF described to date, and our findings indicate that patients diagnosed as adults differ distinctly from survivors of long-term CF diagnosed as children.
Key Words: genotype lung mutation Pseudomonas aeruginosa nontuberculous mycobacteria sweat chloride
Cystic fibrosis (CF) is the most common, lethal, inherited disease in whites, affecting 1 in 2,500 individuals (1). When CF was originally described by D. Andersen in 1936, the affected population was identified as children with severe pancreatic insufficiency, recurrent pulmonary infections, and a life expectancy of less than 2 years. Since then, improvements in treatment have extended the median age of survival to 32.9 years (2, 3). In addition, identification of the CF gene (cystic fibrosis transmembrane conductance regulator [CFTR]) and improved diagnostic testing has expanded the spectrum of recognized disease phenotypes to include milder aspects of CFTR dysfunction (4, 5).
Although certain missense mutations are associated with milder disease, phenotypic variability in CF cannot be explained on the basis of the CFTR genotype alone (6, 7). In this regard, an important study population is the cohort of individuals who survive to older age despite having typical diagnostic features of CF. In the past, individuals who survived to old age, particularly those diagnosed after the third decade of life, were considered curiosities, often the subject of individual case reports in the medical literature (8eC11). More recently, several series have described characteristics of adult patients, including those diagnosed after childhood (12eC14). Our adult CF center is located at a national respiratory disease referral hospital, the National Jewish Medical and Research Center (Denver, CO), and testing for CF is done routinely in patients with persistent airway infections. We undertook a retrospective chart review to characterize older patients with CF identified or followed up at our center. Fifty-five individuals 40 years or older were identified, falling above the 95th percentile of patients with CF currently under care at 117 CF centers in the United States (3). As a group, extensive heterogeneity in pulmonary disease severity, spectrum of lung infection, pancreatic involvement, and severity of CFTR mutations were observed. However, two discrete populations, differentiated by their age at diagnosis, were present with respect to genotype and phenotype. The series of adult patients with CF described here is considerably older than previously reported cohorts, and the subgroup of patients diagnosed in adulthood were diagnosed 20 to 30 years later than previously reported series of adults diagnosed with CF (12eC14). Our analysis demonstrates that, although older patients with CF diagnosed in adulthood often have fairly classic features of CF, they are more frequently women, typically have normal pancreatic function, have less severe pulmonary disease, and have a different spectrum of pulmonary infection. In particular, infection with nontuberculous mycobacteria (NTM) may represent a common mode of presentation in these patients. Components of this report were previously reported as an abstract at the 16th annual North American CF conference (15).
METHODS
A retrospective chart review of the clinical records of the University of Colorado adult CF clinic between 1992 and 2004, encompassing 334 individuals, was performed with all data entered into an Excel spreadsheet (Microsoft, Redmond, WA). The subset of individuals 40 years or older at the time of last contact was analyzed. All patients included in the analysis met the criteria for diagnosis of CF as proposed by the statement of the CF Foundation (CFF) consensus panel (16). These diagnostic criteria require one or more characteristic clinical features together with laboratory evidence of CFTR dysfunction (elevated sweat chloride or abnormal nasal potential difference) or identification of two pathologic CFTR mutations. Individuals for whom complete records were not available, including a detailed medical history, physical examination, sweat chloride (x2), and/or CFTR genotype, microbiological sputum analysis, and pulmonary physiology testing, were excluded from the analysis. A total of 59 patients older than 40 years were identified; 55 of these patients had sufficiently complete data to allow inclusion in the study.
Reported measurements of FEV1 (% predicted), pancreatic function, CF-related diabetes (CFRD), and nutritional status were obtained from the clinic visit closest to age 40. For patients lacking definitive testing, pancreatic enzymes were used as a marker of pancreatic insufficiency, and insulin was used as a marker of CFRD. Analysis of sputum microbiology reflected the summation of all available cultures, because identification of organisms by sputum culture lags behind the time of infection, and single-culture results are not necessarily representative of the overall infectious burden of an individual patient. All data analyses and determination of significance of detected differences were conducted using Statview software (SAS Institute, Cary, NC). Because a normal distribution was not present for age of diagnosis or greatest known age, analyses of variance were determined by Wilcoxon unpaired exact test for Figures 1 and 3AeC3B. Fisher's exact test was used to determine differences in prevalence of disease, CFTR mutation, or infection in Figures 3CeC3D, 4, and 5. For all tests, p < 0.05 was considered significant. This study was approved by the institutional review board of the National Jewish Medical and Research Center.
RESULTS
Age at Diagnosis
Records from all patients 40 years or older were reviewed, and 55 individuals were identified who fulfilled the study criteria. For the group as a whole, the median age at diagnosis was 15 years (Figure 1A). A scatterplot suggests that two distributions within the study population could be differentiated on the basis of age at diagnosis: an early diagnosis (ED) group (n = 28), with age at diagnosis ranging from 0.1 to 15 years, and a late diagnosis (LD) group (n = 27), with age at diagnosis ranging from 24 to 72.8 years. Figure 1B demonstrates the greatest known age of each patient depicted in Figure 1A. The median age of the ED group (45.4 years) was significantly lower than the LD group (52.7 years, p < 0.001). During the 12-year period of the survey, only four individuals had died from each group, although there was a tendency for more individuals in the ED group to have received a lung transplant (five compared with two).
To put this information into perspective, a similar analysis was performed using the 2003 CFF registry database of 23,105 individuals followed at U.S. CF centers (Figure 2) (3). The median age at diagnosis for the registry population as a whole was less than age 6 months. When the subset of 1,160 individuals who were older than age 40 was analyzed, the median age at diagnosis was 13 years. Figure 2B shows a frequency histogram for age at diagnosis in individuals 40 years or older. When a logarithmic function was applied to the group as a whole (dashed line), an R value of 0.72 was achieved, whereas restricting the fit to only those individuals diagnosed before age 20 (solid line) improved the R value to 0.78 (p < 0.001). In either case, similar to our study population, this simple model was unable to explain the excess frequency of individuals diagnosed between ages 25 and 50.
Demographics of the Study Population
Because the greatest known age of the LD group was significantly older than the ED group (Figure 1B), demographic data were obtained from the clinic visit closest to each individual's 40th birthday. Comparison between the ED and LD groups shows that the LD group had significantly milder lung disease as measured by FEV1 (Figure 3A) and a tendency toward lower sweat chloride values (Figure 3B), although values exceeded 60 mmol for all but two individuals in the LD group. A higher proportion of subjects in the ED group had pancreatic insufficiency (p < 0.05), and thus a greater proportion in the ED group had CFRD (p < 0.05; Figure 3C). Not surprisingly, nutritional status at 40 years, as measured by percentage of ideal body weight, was lower in the ED group (Figure 3A). Only 35.7% of the ED group was female; however, the LD group was 74% female (p = 0.0045; Figure 3D). When patients from National Jewish were compared with data from the CFF patient registry, patients with CF who were older than 40 years nationwide most closely resembled the ED group of our population, consistent with a median age at diagnosis of less then 15 years.
Genotype Distribution
The distribution of CFTR genotypes is shown in Table 1 and CFTR alleles in Figure 4. The most notable difference was in the frequency of patients homozygous for F508. In the ED group, 10 patients were homozygous (37%), and every individual with a known mutation had at least one F508 allele (85%). The CFF registry population is 50.6% F508 homozygote, and 88% of the population have at least one copy of F508 (3). In contrast, although F508 was the most common allele in the LD group, four individuals lacked a F508 allele, and only one individual was homozygous. Figure 4 shows the classes of mutations seen in each group. The proportion of individuals with milder class IIIeCV or unidentified mutations was higher in the LD group, whereas the proportion with severe class IeCII (including F508) mutations was higher in the ED group (p < 0.05).
Sputum Microbiology
Sputum microbiology was analyzed to determine if there were major differences in the types of pathogens isolated from lower airway secretions in the two subgroups. The mean number of bacterial cultures available per patient for the ED groups was 7.4 compared with 6.4 for the LD groups (p = 0.39). The ED group had a mean of 2.3 mycobacterial cultures per patient compared with 3.3 for the LD group (p = 0.06). Using the Dice similarity coefficient (Sorenson and Dice), an analysis of similarity to assess the global pattern of infectious agents (17), a significant difference was detected between the two groups (p < 0.02; data not shown). The difference could be ascribed to the pathogens shown in Figure 5. The prevalence of Pseudomonas aeruginosa, both mucoid and nonmucoid strains, was high in both groups but significantly greater in the ED group. An equal prevalence of Aspergillus species, Staphylococcus aureus, and Burkholderia cepacia were found in each group.
A total of 18 patients had at least one positive culture for NTM. In contrast to P. aeruginosa, NTM were detected three times more often in the LD group (p < 0.006). Mycobacterium avium complex was the most commonly detected species (3 of 4 patients in the ED group, 12 of 14 patients in the LD group). Although six of the LD patients (five with M. avium complex and one with M. abscessus) met American Thoracic Society criteria for NTM infection (18), none of the ED patients were classified as infected or received treatment for mycobacterial infection. Only the six individuals in the LD group who met American Thoracic Society criteria for NTM infection had more than two positive acid-fast bacilli (AFB) cultures or a positive AFB smear. The NTM isolated other than M. avium complex were as follows: M. abscessus (three LD patients), M. fortuitum (one LD patient), M. chelonae (one ED patient), M. simiae (one LD patient), and M. scrofulaceum (one ED patient). Two LD patients had positive cultures for both M. avium complex and M. abscessus, and one LD patient had positive culture for both M. avium complex and M. fortuitum.
When compared with patients nationwide who were older than 40 years and enrolled in the CFF registry, the patients in our study had a similar prevalence of P. aeruginosa and S. aureus. A lower percentage of CFF registry patients had positive cultures for B. cepacia, Aspergillus species, and NTM. However, the requirement for special culture techniques for NTM may have resulted in an underrepresentation of actual prevalence.
DISCUSSION
As the CF population ages and testing for rare CFTR mutations has become widely available, increased attention has been placed on the late diagnosis of the adult patient. Several series describing characteristics of younger adult patients, including those diagnosed after childhood, have been published in the last decade. McCloskey and coworkers (12) described 18 patients with CF (aged 14eC36 years) diagnosed after age 10 in Northern Ireland. Gan and coworkers (13) reported on 25 patients with CF (aged 21eC55 years) in the Netherlands diagnosed after age 16, with a median age at diagnosis of 27 years and a mean age of 35.7 years. An analysis of the CFF patient registry data from 1996 by Widerman and colleagues (14) identified a cohort diagnosed after age 18, with a mean age at diagnosis of 27 years. Although these reports describe populations of LD patients with CF who were diagnosed 20 to 30 years earlier than the LD group in this report, the investigators' findings support the conclusions of this study. Specifically, patients with CF diagnosed after childhood have less severe lung disease, are more likely to have pancreatic sufficiency, have a lower prevalence of P. aeruginosa, and have a lower prevalence of the F508 mutation (12eC14).
CF is a monogenic, autosomal-recessive, inherited disorder. Before identification of the disease locus, CFTR, it was widely assumed that much of the variability in disease phenotype would be explained by functional differences in specific disease-gene mutations (19). Although this assumption has proved to be generally true, in certain individuals profound differences have been observed between actual and predicted disease severity based on the class of CFTR mutation (6). In these patients with CF, the lack of a tight phenotypeeCgenotype correlation can be explained by two hypotheses: (1) differences in environmental cofactors and (2) expression of modifier genes at other loci. These hypotheses are not mutually exclusive.
The testing of these hypotheses requires identification of subpopulations of affected individuals who have significantly different clinical phenotypes. The long-term survivors of CF, defined for the purposes of this study as 40 years or older, represent the oldest 5% of individuals nationally and appear to be an important population to study factors that allow for increased survival. Two relatively distinct subpopulations were apparent within the long-term survivor group. One group, referred to as "ED," had a median age at diagnosis of 2 years and a second population, referred to as "LD," had a median age at diagnosis of 48.8 years. Thus, for the purposes of this study, the ED group was defined as age at diagnosis of less than 15 years, and the LD group was defined as age at diagnosis of more than 24 years.
Several limitations of using age at diagnosis as a discriminate variable should be acknowledged. In addition to differences in phenotype, delay in diagnosis may be influenced by a variety of nonclinical factors, such as access to, and quality of, health care, the ability to tolerate symptoms, and implementation of newborn screening. Although newborn screening is too recent a development to have affected our study population, in the future, newborn screening (and prenatal screening) may decrease the proportion of individuals diagnosed at later ages, thus limiting the ability to use this variable to discriminate discrete subpopulations of CF-affected individuals (20).
An analysis of diagnostic testing, including CFTR genotype, showed that the average sweat chloride value, loosely correlated to the presence of missense mutations, was slightly lower in the LD group (21). However, in part because of the entry criteria established for this study, all but two individuals had a diagnostic sweat chloride test. Other individuals in our center's database had a clinical syndrome consistent with CF: a borderline sweat chloride value and one CFTR mutation identified. Although these individuals may also represent the mild spectrum of CFTR-related disease (22), they were not included in this analysis. As expected, the F508 mutation was most common in both groups. All patients with known mutations in the ED group had at least one F508 allele, and 10 were F508 homozygous. Nineteen of the 27 patients in the LD group had at least one F508 allele, but only one patient was F508 homozygous. Patients diagnosed after age 30 were more likely to have a genotype that included one or more class IIIeCV mutations or an unidentified mutation, consistent with the milder clinical phenotype (23). It is also likely that the unidentified mutations in this cohort represent uncommon, mild, missense mutations not detected by the clinical screening for fewer than 100 mutations, which was used for genotyping most subjects.
Not surprisingly, patients in the LD group had a lower prevalence of gastrointestinal manifestations of CF as well as less CFRD and better overall nutrition. Prior analyses have associated exocrine pancreatic sufficiency with prolonged survival (24), and our analysis is in agreement with the findings of Durno and coworkers (25) who found that pancreatic sufficiency and pancreatitis are associated with LD. Class IV CFTR mutations, which were more prevalent in this group, have also been associated with preserved pancreatic function (25). The prevalence of CFRD, which is directly related to the severity of exocrine pancreatic insufficiency, is inversely related to survival, and our analysis suggests that pancreatic insufficiency is also inversely related to diagnosis after age 24 (26).
As the general population of patients with CF ages, it has been observed that women have a significantly increased risk of death, and a median lifespan 3 to 5 years shorter then men (27, 28). The factors that cause the CF "gender gap" are not completely understood but may relate to issues of nutrition and time of first P. aerugnosa acquisition (29). Our ED group and the CFF registry population older than 40 years appear to reflect this phenomenon, because only 35.7 and 44.5%, respectively, of these patients were women. Surprisingly, women in the LD group outnumbered men by 3 to 1. This pattern of increased female representation in patients with CF diagnosed in adulthood was detected by Widerman and coworkers (14) in their analysis of CFF registry data, but the extent of the difference was much smaller, possibly because their population was much younger than the one described in this article. Although a number of potential factors could contribute to this divergence in female survival, it appears that long-term survivors of CF represent a population that could be studied to test the contribution of sex in CF disease progression.
The global assessment of sputum microbiology revealed significant differences between the two groups, with greater heterogeneity in the LD group. Although our a priori prediction was that patients in the LD group would have a lower prevalence of P. aeruginosa (30), the majority of patients in both groups were infected. This finding is particularly interesting, because prior analyses of the national CFF registry database demonstrated that colonization with P. aeruginosa had an extremely strong negative prognostic impact (31eC33). This suggests that long-term survivors, regardless of age at diagnosis, may have a better ability to tolerate lower respiratory tract infection with P. aeruginosa. Although deficiency of mannose binding lectin is a polymorphism associated with more severe lung disease in CF (34), other mutations and polymorphisms of immune-related genes may be associated with the ability of long-term survivors of CF to tolerate chronic P. aeruginosa infection, and represents an important topic for future studies (35, 36).
Another interesting microbiological difference between the two groups was the prevalence of NTM in the sputum, which was three times more common in the LD group. The prevalence of NTM was previously studied in the North American CF population (37). Of patients with CF who were 10 years or older, 13% were found to be culture-positive for NTM. This finding is similar to the 10% prevalence we saw in the ED group, but considerably less than the 52% we found in the LD age group.
Ascertainment bias is particularly likely in our study: patients are often referred to our center because of unusual clinical symptoms or specifically for treatment of NTM infection. Of the 27 LD patients, six were referred to National Jewish for treatment of NTM infections and were subsequently diagnosed with CF. The remaining 21 patients were referred for evaluation of bronchiectasis, were self-referred for chronic lung disease, or were diagnosed by physicians in the region and referred for CF center care. Although systematic CF diagnostic testing in our infectious disease clinics together with efforts to routinely culture for NTM in the CF clinic certainly contributed to the high prevalence of mycobacterial infection in the LD group, conclusions from this analysis support the findings of other reports of NTM in adults with CF. In the national epidemiologic study of NTM in patients with CF, the age of mycobacteria-positive individuals was older (26 vs. 22 years), and they tended to have milder abnormalities in lung function (FEV1 [% predicted] 60 vs. 54%) (38). Although age at diagnosis was not analyzed in that study, those data support the hypothesis that older patients with milder disease are predisposed to infection with NTM (39eC42). Thus, although our finding that more than 50% of LD adults older than 40 years have at least one culture positive for NTM may be somewhat skewed, it is clear that a positive culture for NTM is a common presentation of undiagnosed CF in adults older than 40. It seems likely that in the general population a significant number of similar patients remain undiagnosed, and thus are underrepresented in the CFF registry database.
Although the significance of culturing NTM from the sputum of patients with CF is often uncertain with respect to pathogenesis or clinical course (43), the high prevalence of NTM in all our patients older than 40 highlights an important point for clinicians. All patients who present with bronchiectasis and are culture positive for NTM, regardless of age and associated symptoms, should be screened for CF with both a sweat chloride test and genotyping, and if CF is diagnosed, referral to a cystic fibrosis care and research center should be considered. Our analysis suggests that a significant population of individuals who meet classic diagnostic criteria for CF present late in life, often with respiratory symptoms that have been present for decades. Remarkably, 47 of the 55 patients described in this report are still alive, emphasizing that long-term survival of patients with CF may not be as uncommon as previously believed.
In summary, this study is the first to examine a large cohort of very long-term survivors with CF, which can be divided into two distinct populations on the basis of age at diagnosis. The subset of individuals who were diagnosed early in life matched the overall CFF registry database and the prior literature with respect to CFTR allelic frequency, prevalence of pancreatic insufficiency, CFRD, and microbiology. Yet, the average current age of this group was 45.4 years, placing these individuals in the oldest 5% of patients nationally. This ED subset could therefore be particularly useful for identifying modifier genes, because these individuals' genotype is typical of the CF population in the United States, but their survival is significantly greater. The population that was diagnosed after age 24 was more heterogeneous, with fewer typical manifestations of CF and less severe CFTR mutations, and thus may be less suitable for analysis of genetic modifiers. The high incidence of NTM as a presenting manifestation of CF suggests that acquisition of NTM may exacerbate the phenotype of some of these patients after a long period of compensated, undiagnosed disease, ultimately bringing them to clinical attention.
Acknowledgments
The authors thank Hebe B. Quinton at Dartmouth Hitchcock Medical Center for providing CFF registry data.
REFERENCES
Rosenstein BJ, Zeitlin PL. Cystic fibrosis. Lancet 1998;351:277eC282.
Morrissey BM, Schock BC, Marelich GP, Cross CE. Cystic fibrosis in adults: current and future management strategies. Clin Rev Allergy Immunol 2003;25:275eC287.
Cystic Fibrosis Foundation. Patient registry 2003 annual data report. Bethesda, MD: Cystic Fibrosis Foundation; 2004.
Boyle MP. Nonclassic cystic fibrosis and CFTR-related diseases. Curr Opin Pulm Med 2003;9:498eC503.
Highsmith WE, Overbeek SE, Hilvering C. A novel mutation in the CF gene in patients with pulmonary disease but normal sweat chloride concentrations. N Engl J Med 1994;331:974eC980.
Mickle JE, Cutting GR. Genotype-phenotype relationships in cystic fibrosis. Med Clin North Am 2000;84:597eC607.
Burke W, Aitken M, Chen SH, Scott CR. Variable severity of pulmonary disease in adults with identical cystic fibrosis mutations. Chest 1992;102:506eC509.
Sanders JS, Pryor TD, Wedel MK. Prolonged survival in an adult with cystic fibrosis. Chest 1980;77:226eC227.
Fiel S. Cystic Fibrosis in an adult. Emerg Med 1988;20:109eC121.
Hunt B, Geddes D. Newly diagnosed CF in middle and later life. Thorax 1984;40:23eC26.
Su C, Beanblossom B. Typical CF in an elderly woman. Am J Med 1989;86:701eC703.
McCloskey M, Redmond AO, Hill A, Elborn JS. Clinical features associated with a delayed diagnosis of cystic fibrosis. Respiration (Herrlisheim) 2000;67:402eC407.
Gan KH, Geus WP, Bakker W, Cornelis CBHW, Heijerman HGM. Genetic and clinical features of patients with cystic fibrosis diagnosed after the age of 16 years. Thorax 1995;50:1301eC1304.
Widerman E, Millner L, Sexauer W, Fiel S. Health status and sociodemographic characteristics of adults receiving a CF diagnosis after age 18 years. Chest 2000;118:427eC433.
Wallick K, Sontag M, Jones M, Brayshaw S, Accurso F, Rodman D, Nick J. Clinical characteristics of adult patients over age 40 with CF: significance of early and delayed diagnosis. Paediatr Pulm 2002;S24:319.
Rosenstein BJ, Cutting GR. The diagnosis of cystic fibrosis: a consensus statement. Cystic Fibrosis Foundation consensus panel. J Pediatr 1998;132:589eC595.
Smith EP, Pontasch KW, Cains J Jr. Community similarity and the analysis of multispecies environmental data: a unified statistical approach. Water Res 1990;24:507eC514.
Wallace RJ, Glassroth J, Griffith DE, Olivier KN, Cook JL, Gordin F. Diagnosis and treatment of disease caused by nontuberculous mycobacteria. Am J Respir Crit Care Med 1997;156:S1eCS25.
Bonforte RJ. Variability in cystic fibrosis. JAMA 1978;240:1855.
Lai HJ, Cheng Y, Cho H, Kosorok MR, Farrell PM. Association between initial disease presentation, lung disease outcomes, and survival in patients with cystic fibrosis. Am J Epidemiol 2004;159:537eC546.
Feldmann D, Couderc R, Audrezet MP, Ferec C, Bienvenu T, Desgeorges M, Claustres M, Mittre H, Blayau M, Bozon D, et al. CFTR genotypes in patients with normal or borderline sweat chloride levels. Hum Mutat 2003;22:340.
Groman JD, Meyer ME, Wilmott RW, Zeitlin PL, Cutting GR. Variant cystic fibrosis phenotypes in the absence of CFTR mutations. N Engl J Med 2002;347:401eC407.
Choo-Kang LR, Zeitlin PL. Type I, II, III, IV, and V cystic fibrosis transmembrane conductance regulator defects and opportunities for therapy. Curr Opin Pulm Med 2000;6:521eC529.
Corey M, Edwards L, Levison H, Knowles M. Longitudinal analysis of pulmonary function decline in patients with cystic fibrosis. J Pediatr 1997;131:809eC814.
Durno C, Corey M, Zielenski J, Tullis E, Tsui L, Durie P. Genotype and phenotype correlations in patients with CF and pancreatits. Gastroenterology 2002;123:1857eC1864.
Vizza CD, Yusen RD, Lynch JP, Fedele F, Alexander Patterson G, Trulock EP. Outcome of patients with cystic fibrosis awaiting lung transplantation. Am J Respir Crit Care Med 2000;162:819eC825.
The Cystic Fibrosis Genotype-Phenotype Consortium. Correlation between genotype and phenotype in patients with CF. N Engl J Med 1993;329:1308eC1313.
O'Connor GT, Quinton HB, Kahn R, Robichaud P, Maddock J, Lever T, Detzer M, Brooks JG. Case-mix adjustment for evaluation of mortality in CF. Pediatr Pulmonol 2002;33:99eC105.
Demko CA, Byard PJ, Davis PB. Gender differences in CF: Pseudomonas aeruginosa infection. J Clin Epidemiol 1994;43:125eC131.
Gibson RL, Burns JL, Ramsey BW. Pathophysiology and management of pulmonary infections in cystic fibrosis. Am J Respir Crit Care Med 2003;168:918eC951.
Emerson J, Rosenfeld M, McNamara S, Ramsey B, Gibson RL. Pseudomonas aeruginosa and other predictors of mortality and morbidity in young children with cystic fibrosis. Pediatr Pulmonol 2002;34:91eC100.
Huang NN, Schidlow DV, Szatrowski TH, Palmer J, Laraya-Cuasay LR, Yeung W, Hardy K, Quitell L, Fiel S. Clinical features, survival rate, and prognostic factors in young adults with cystic fibrosis. Am J Med 1987;82:871eC879.
Nixon GM, Armstrong DS, Carzino R, Carlin JB, Olinsky A, Robertson CF, Grimwood K. Clinical outcome after early Pseudomonas aeruginosa infection in cystic fibrosis. J Pediatr 2001;138:699eC704.
Gabolde M, Guilloud-Bataille M, Feingold J, Besmond C. Association of variant alleles of mannose binding lectin with severity of pulmonary disease in cystic fibrosis: cohort study. BMJ 1999;319:1166eC1167.
Merlo CA, Boyle MP. Modifier genes in cystic fibrosis lung disease. J Lab Clin Med 2003;141:237eC241.
Salvatore F, Scudiero O, Castaldo G. Genotype-phenotype correlation in cystic fibrosis: the role of modifier genes. Am J Med Genet 2002;111:88eC95.
Olivier KN, Weber DJ, Wallace RJ Jr, Faiz AR, Lee JH, Zhang Y, Brown-Elliot BA, Handler A, Wilson RW, Schechter MS, et al. Nontuberculous mycobacteria. I: multicenter prevalence study in cystic fibrosis. Am J Respir Crit Care Med 2003;167:828eC834.
Olivier KN, Weber DJ, Lee JH, Handler A, Tudor G, Molina PL, Tomashefski J, Knowles MR. Nontuberculous mycobacteria. II: nested-cohort study of impact on cystic fibrosis lung disease. Am J Respir Crit Care Med 2003;167:835eC840.
Olivier KN, Weber DJ, Wallace RJ, Faiz AR, Lee JH, Zhang Y, Brown-Elliot BA, Handler A, Wilson RW, Schechter MS, et al. Nontuberculous mycobacteria: I. Multicenter prevalence study in cytic fibrosis. Am J Respir Crit Care Med 2003;167:828eC834.
Kilby JM, Gilligan PH, Yankaskas JR, Highsmith WE, Edwards LJ, Knowles MR. Nontuberculous mycobacteria in adult patients with cystic fibrosis. Chest 1992;102:70eC75.
Aitken ML, Burke W, McDonald G, Wallis C, Ramsey B, Nolan C. Nontuberculous mycobacterial disease in adult cystic fibrosis patients. Chest 1993;103:1096eC1099.
Fauroux B, Delaisi B, Clement A, Saizou C, Moissenet D, Truffot-Pernot C, Tournier G, Vu Thien H. Mycobacterial lung disease in cystic fibrosis: a prospective study. Pediatr Infect Dis J 1997;16:354eC358.
Nick JA. Nontuberculous mycobacteria in cystic fibrosis. Semin Respir Crit Care Med 2003;24:693eC702.(David M. Rodman, Jillian )
National Jewish Medical and Research Center
The Children's Hospital, Denver, Colorado
ABSTRACT
Although the median survival for patients with cystic fibrosis (CF) is 32.9 years, a small group of patients live much longer. We analyzed the genotype and phenotype of CF patients 40 years and older seen between 1992 and 2004 at the National Jewish Medical and Research Center (n = 55). These patients were divided into two groups according to age at diagnosis: an early diagnosis (ED) group, median age at diagnosis 2.0 years (range 0.1eC15 years, n = 28), and a late diagnosis (LD) group, median age of diagnosis 48.8 years (range 24eC72.8 years, n = 27). Consistent with the hypothesis that the CFTR genotype affects the age at diagnosis, CFTR F508 homozygous individuals were more common in the ED group. Although patients in the ED group were predominantly male, the majority of LD patients were female. Patients with CF diagnosed late had a significantly lower prevalence of pancreatic insufficiency and CF-related diabetes, and better lung function. Fewer patients in the LD groups were infected with Pseudomonas aeruginosa, whereas a greater percentage had cultures positive for nontuberculous mycobacteria. This is the largest cohort of older patients with CF described to date, and our findings indicate that patients diagnosed as adults differ distinctly from survivors of long-term CF diagnosed as children.
Key Words: genotype lung mutation Pseudomonas aeruginosa nontuberculous mycobacteria sweat chloride
Cystic fibrosis (CF) is the most common, lethal, inherited disease in whites, affecting 1 in 2,500 individuals (1). When CF was originally described by D. Andersen in 1936, the affected population was identified as children with severe pancreatic insufficiency, recurrent pulmonary infections, and a life expectancy of less than 2 years. Since then, improvements in treatment have extended the median age of survival to 32.9 years (2, 3). In addition, identification of the CF gene (cystic fibrosis transmembrane conductance regulator [CFTR]) and improved diagnostic testing has expanded the spectrum of recognized disease phenotypes to include milder aspects of CFTR dysfunction (4, 5).
Although certain missense mutations are associated with milder disease, phenotypic variability in CF cannot be explained on the basis of the CFTR genotype alone (6, 7). In this regard, an important study population is the cohort of individuals who survive to older age despite having typical diagnostic features of CF. In the past, individuals who survived to old age, particularly those diagnosed after the third decade of life, were considered curiosities, often the subject of individual case reports in the medical literature (8eC11). More recently, several series have described characteristics of adult patients, including those diagnosed after childhood (12eC14). Our adult CF center is located at a national respiratory disease referral hospital, the National Jewish Medical and Research Center (Denver, CO), and testing for CF is done routinely in patients with persistent airway infections. We undertook a retrospective chart review to characterize older patients with CF identified or followed up at our center. Fifty-five individuals 40 years or older were identified, falling above the 95th percentile of patients with CF currently under care at 117 CF centers in the United States (3). As a group, extensive heterogeneity in pulmonary disease severity, spectrum of lung infection, pancreatic involvement, and severity of CFTR mutations were observed. However, two discrete populations, differentiated by their age at diagnosis, were present with respect to genotype and phenotype. The series of adult patients with CF described here is considerably older than previously reported cohorts, and the subgroup of patients diagnosed in adulthood were diagnosed 20 to 30 years later than previously reported series of adults diagnosed with CF (12eC14). Our analysis demonstrates that, although older patients with CF diagnosed in adulthood often have fairly classic features of CF, they are more frequently women, typically have normal pancreatic function, have less severe pulmonary disease, and have a different spectrum of pulmonary infection. In particular, infection with nontuberculous mycobacteria (NTM) may represent a common mode of presentation in these patients. Components of this report were previously reported as an abstract at the 16th annual North American CF conference (15).
METHODS
A retrospective chart review of the clinical records of the University of Colorado adult CF clinic between 1992 and 2004, encompassing 334 individuals, was performed with all data entered into an Excel spreadsheet (Microsoft, Redmond, WA). The subset of individuals 40 years or older at the time of last contact was analyzed. All patients included in the analysis met the criteria for diagnosis of CF as proposed by the statement of the CF Foundation (CFF) consensus panel (16). These diagnostic criteria require one or more characteristic clinical features together with laboratory evidence of CFTR dysfunction (elevated sweat chloride or abnormal nasal potential difference) or identification of two pathologic CFTR mutations. Individuals for whom complete records were not available, including a detailed medical history, physical examination, sweat chloride (x2), and/or CFTR genotype, microbiological sputum analysis, and pulmonary physiology testing, were excluded from the analysis. A total of 59 patients older than 40 years were identified; 55 of these patients had sufficiently complete data to allow inclusion in the study.
Reported measurements of FEV1 (% predicted), pancreatic function, CF-related diabetes (CFRD), and nutritional status were obtained from the clinic visit closest to age 40. For patients lacking definitive testing, pancreatic enzymes were used as a marker of pancreatic insufficiency, and insulin was used as a marker of CFRD. Analysis of sputum microbiology reflected the summation of all available cultures, because identification of organisms by sputum culture lags behind the time of infection, and single-culture results are not necessarily representative of the overall infectious burden of an individual patient. All data analyses and determination of significance of detected differences were conducted using Statview software (SAS Institute, Cary, NC). Because a normal distribution was not present for age of diagnosis or greatest known age, analyses of variance were determined by Wilcoxon unpaired exact test for Figures 1 and 3AeC3B. Fisher's exact test was used to determine differences in prevalence of disease, CFTR mutation, or infection in Figures 3CeC3D, 4, and 5. For all tests, p < 0.05 was considered significant. This study was approved by the institutional review board of the National Jewish Medical and Research Center.
RESULTS
Age at Diagnosis
Records from all patients 40 years or older were reviewed, and 55 individuals were identified who fulfilled the study criteria. For the group as a whole, the median age at diagnosis was 15 years (Figure 1A). A scatterplot suggests that two distributions within the study population could be differentiated on the basis of age at diagnosis: an early diagnosis (ED) group (n = 28), with age at diagnosis ranging from 0.1 to 15 years, and a late diagnosis (LD) group (n = 27), with age at diagnosis ranging from 24 to 72.8 years. Figure 1B demonstrates the greatest known age of each patient depicted in Figure 1A. The median age of the ED group (45.4 years) was significantly lower than the LD group (52.7 years, p < 0.001). During the 12-year period of the survey, only four individuals had died from each group, although there was a tendency for more individuals in the ED group to have received a lung transplant (five compared with two).
To put this information into perspective, a similar analysis was performed using the 2003 CFF registry database of 23,105 individuals followed at U.S. CF centers (Figure 2) (3). The median age at diagnosis for the registry population as a whole was less than age 6 months. When the subset of 1,160 individuals who were older than age 40 was analyzed, the median age at diagnosis was 13 years. Figure 2B shows a frequency histogram for age at diagnosis in individuals 40 years or older. When a logarithmic function was applied to the group as a whole (dashed line), an R value of 0.72 was achieved, whereas restricting the fit to only those individuals diagnosed before age 20 (solid line) improved the R value to 0.78 (p < 0.001). In either case, similar to our study population, this simple model was unable to explain the excess frequency of individuals diagnosed between ages 25 and 50.
Demographics of the Study Population
Because the greatest known age of the LD group was significantly older than the ED group (Figure 1B), demographic data were obtained from the clinic visit closest to each individual's 40th birthday. Comparison between the ED and LD groups shows that the LD group had significantly milder lung disease as measured by FEV1 (Figure 3A) and a tendency toward lower sweat chloride values (Figure 3B), although values exceeded 60 mmol for all but two individuals in the LD group. A higher proportion of subjects in the ED group had pancreatic insufficiency (p < 0.05), and thus a greater proportion in the ED group had CFRD (p < 0.05; Figure 3C). Not surprisingly, nutritional status at 40 years, as measured by percentage of ideal body weight, was lower in the ED group (Figure 3A). Only 35.7% of the ED group was female; however, the LD group was 74% female (p = 0.0045; Figure 3D). When patients from National Jewish were compared with data from the CFF patient registry, patients with CF who were older than 40 years nationwide most closely resembled the ED group of our population, consistent with a median age at diagnosis of less then 15 years.
Genotype Distribution
The distribution of CFTR genotypes is shown in Table 1 and CFTR alleles in Figure 4. The most notable difference was in the frequency of patients homozygous for F508. In the ED group, 10 patients were homozygous (37%), and every individual with a known mutation had at least one F508 allele (85%). The CFF registry population is 50.6% F508 homozygote, and 88% of the population have at least one copy of F508 (3). In contrast, although F508 was the most common allele in the LD group, four individuals lacked a F508 allele, and only one individual was homozygous. Figure 4 shows the classes of mutations seen in each group. The proportion of individuals with milder class IIIeCV or unidentified mutations was higher in the LD group, whereas the proportion with severe class IeCII (including F508) mutations was higher in the ED group (p < 0.05).
Sputum Microbiology
Sputum microbiology was analyzed to determine if there were major differences in the types of pathogens isolated from lower airway secretions in the two subgroups. The mean number of bacterial cultures available per patient for the ED groups was 7.4 compared with 6.4 for the LD groups (p = 0.39). The ED group had a mean of 2.3 mycobacterial cultures per patient compared with 3.3 for the LD group (p = 0.06). Using the Dice similarity coefficient (Sorenson and Dice), an analysis of similarity to assess the global pattern of infectious agents (17), a significant difference was detected between the two groups (p < 0.02; data not shown). The difference could be ascribed to the pathogens shown in Figure 5. The prevalence of Pseudomonas aeruginosa, both mucoid and nonmucoid strains, was high in both groups but significantly greater in the ED group. An equal prevalence of Aspergillus species, Staphylococcus aureus, and Burkholderia cepacia were found in each group.
A total of 18 patients had at least one positive culture for NTM. In contrast to P. aeruginosa, NTM were detected three times more often in the LD group (p < 0.006). Mycobacterium avium complex was the most commonly detected species (3 of 4 patients in the ED group, 12 of 14 patients in the LD group). Although six of the LD patients (five with M. avium complex and one with M. abscessus) met American Thoracic Society criteria for NTM infection (18), none of the ED patients were classified as infected or received treatment for mycobacterial infection. Only the six individuals in the LD group who met American Thoracic Society criteria for NTM infection had more than two positive acid-fast bacilli (AFB) cultures or a positive AFB smear. The NTM isolated other than M. avium complex were as follows: M. abscessus (three LD patients), M. fortuitum (one LD patient), M. chelonae (one ED patient), M. simiae (one LD patient), and M. scrofulaceum (one ED patient). Two LD patients had positive cultures for both M. avium complex and M. abscessus, and one LD patient had positive culture for both M. avium complex and M. fortuitum.
When compared with patients nationwide who were older than 40 years and enrolled in the CFF registry, the patients in our study had a similar prevalence of P. aeruginosa and S. aureus. A lower percentage of CFF registry patients had positive cultures for B. cepacia, Aspergillus species, and NTM. However, the requirement for special culture techniques for NTM may have resulted in an underrepresentation of actual prevalence.
DISCUSSION
As the CF population ages and testing for rare CFTR mutations has become widely available, increased attention has been placed on the late diagnosis of the adult patient. Several series describing characteristics of younger adult patients, including those diagnosed after childhood, have been published in the last decade. McCloskey and coworkers (12) described 18 patients with CF (aged 14eC36 years) diagnosed after age 10 in Northern Ireland. Gan and coworkers (13) reported on 25 patients with CF (aged 21eC55 years) in the Netherlands diagnosed after age 16, with a median age at diagnosis of 27 years and a mean age of 35.7 years. An analysis of the CFF patient registry data from 1996 by Widerman and colleagues (14) identified a cohort diagnosed after age 18, with a mean age at diagnosis of 27 years. Although these reports describe populations of LD patients with CF who were diagnosed 20 to 30 years earlier than the LD group in this report, the investigators' findings support the conclusions of this study. Specifically, patients with CF diagnosed after childhood have less severe lung disease, are more likely to have pancreatic sufficiency, have a lower prevalence of P. aeruginosa, and have a lower prevalence of the F508 mutation (12eC14).
CF is a monogenic, autosomal-recessive, inherited disorder. Before identification of the disease locus, CFTR, it was widely assumed that much of the variability in disease phenotype would be explained by functional differences in specific disease-gene mutations (19). Although this assumption has proved to be generally true, in certain individuals profound differences have been observed between actual and predicted disease severity based on the class of CFTR mutation (6). In these patients with CF, the lack of a tight phenotypeeCgenotype correlation can be explained by two hypotheses: (1) differences in environmental cofactors and (2) expression of modifier genes at other loci. These hypotheses are not mutually exclusive.
The testing of these hypotheses requires identification of subpopulations of affected individuals who have significantly different clinical phenotypes. The long-term survivors of CF, defined for the purposes of this study as 40 years or older, represent the oldest 5% of individuals nationally and appear to be an important population to study factors that allow for increased survival. Two relatively distinct subpopulations were apparent within the long-term survivor group. One group, referred to as "ED," had a median age at diagnosis of 2 years and a second population, referred to as "LD," had a median age at diagnosis of 48.8 years. Thus, for the purposes of this study, the ED group was defined as age at diagnosis of less than 15 years, and the LD group was defined as age at diagnosis of more than 24 years.
Several limitations of using age at diagnosis as a discriminate variable should be acknowledged. In addition to differences in phenotype, delay in diagnosis may be influenced by a variety of nonclinical factors, such as access to, and quality of, health care, the ability to tolerate symptoms, and implementation of newborn screening. Although newborn screening is too recent a development to have affected our study population, in the future, newborn screening (and prenatal screening) may decrease the proportion of individuals diagnosed at later ages, thus limiting the ability to use this variable to discriminate discrete subpopulations of CF-affected individuals (20).
An analysis of diagnostic testing, including CFTR genotype, showed that the average sweat chloride value, loosely correlated to the presence of missense mutations, was slightly lower in the LD group (21). However, in part because of the entry criteria established for this study, all but two individuals had a diagnostic sweat chloride test. Other individuals in our center's database had a clinical syndrome consistent with CF: a borderline sweat chloride value and one CFTR mutation identified. Although these individuals may also represent the mild spectrum of CFTR-related disease (22), they were not included in this analysis. As expected, the F508 mutation was most common in both groups. All patients with known mutations in the ED group had at least one F508 allele, and 10 were F508 homozygous. Nineteen of the 27 patients in the LD group had at least one F508 allele, but only one patient was F508 homozygous. Patients diagnosed after age 30 were more likely to have a genotype that included one or more class IIIeCV mutations or an unidentified mutation, consistent with the milder clinical phenotype (23). It is also likely that the unidentified mutations in this cohort represent uncommon, mild, missense mutations not detected by the clinical screening for fewer than 100 mutations, which was used for genotyping most subjects.
Not surprisingly, patients in the LD group had a lower prevalence of gastrointestinal manifestations of CF as well as less CFRD and better overall nutrition. Prior analyses have associated exocrine pancreatic sufficiency with prolonged survival (24), and our analysis is in agreement with the findings of Durno and coworkers (25) who found that pancreatic sufficiency and pancreatitis are associated with LD. Class IV CFTR mutations, which were more prevalent in this group, have also been associated with preserved pancreatic function (25). The prevalence of CFRD, which is directly related to the severity of exocrine pancreatic insufficiency, is inversely related to survival, and our analysis suggests that pancreatic insufficiency is also inversely related to diagnosis after age 24 (26).
As the general population of patients with CF ages, it has been observed that women have a significantly increased risk of death, and a median lifespan 3 to 5 years shorter then men (27, 28). The factors that cause the CF "gender gap" are not completely understood but may relate to issues of nutrition and time of first P. aerugnosa acquisition (29). Our ED group and the CFF registry population older than 40 years appear to reflect this phenomenon, because only 35.7 and 44.5%, respectively, of these patients were women. Surprisingly, women in the LD group outnumbered men by 3 to 1. This pattern of increased female representation in patients with CF diagnosed in adulthood was detected by Widerman and coworkers (14) in their analysis of CFF registry data, but the extent of the difference was much smaller, possibly because their population was much younger than the one described in this article. Although a number of potential factors could contribute to this divergence in female survival, it appears that long-term survivors of CF represent a population that could be studied to test the contribution of sex in CF disease progression.
The global assessment of sputum microbiology revealed significant differences between the two groups, with greater heterogeneity in the LD group. Although our a priori prediction was that patients in the LD group would have a lower prevalence of P. aeruginosa (30), the majority of patients in both groups were infected. This finding is particularly interesting, because prior analyses of the national CFF registry database demonstrated that colonization with P. aeruginosa had an extremely strong negative prognostic impact (31eC33). This suggests that long-term survivors, regardless of age at diagnosis, may have a better ability to tolerate lower respiratory tract infection with P. aeruginosa. Although deficiency of mannose binding lectin is a polymorphism associated with more severe lung disease in CF (34), other mutations and polymorphisms of immune-related genes may be associated with the ability of long-term survivors of CF to tolerate chronic P. aeruginosa infection, and represents an important topic for future studies (35, 36).
Another interesting microbiological difference between the two groups was the prevalence of NTM in the sputum, which was three times more common in the LD group. The prevalence of NTM was previously studied in the North American CF population (37). Of patients with CF who were 10 years or older, 13% were found to be culture-positive for NTM. This finding is similar to the 10% prevalence we saw in the ED group, but considerably less than the 52% we found in the LD age group.
Ascertainment bias is particularly likely in our study: patients are often referred to our center because of unusual clinical symptoms or specifically for treatment of NTM infection. Of the 27 LD patients, six were referred to National Jewish for treatment of NTM infections and were subsequently diagnosed with CF. The remaining 21 patients were referred for evaluation of bronchiectasis, were self-referred for chronic lung disease, or were diagnosed by physicians in the region and referred for CF center care. Although systematic CF diagnostic testing in our infectious disease clinics together with efforts to routinely culture for NTM in the CF clinic certainly contributed to the high prevalence of mycobacterial infection in the LD group, conclusions from this analysis support the findings of other reports of NTM in adults with CF. In the national epidemiologic study of NTM in patients with CF, the age of mycobacteria-positive individuals was older (26 vs. 22 years), and they tended to have milder abnormalities in lung function (FEV1 [% predicted] 60 vs. 54%) (38). Although age at diagnosis was not analyzed in that study, those data support the hypothesis that older patients with milder disease are predisposed to infection with NTM (39eC42). Thus, although our finding that more than 50% of LD adults older than 40 years have at least one culture positive for NTM may be somewhat skewed, it is clear that a positive culture for NTM is a common presentation of undiagnosed CF in adults older than 40. It seems likely that in the general population a significant number of similar patients remain undiagnosed, and thus are underrepresented in the CFF registry database.
Although the significance of culturing NTM from the sputum of patients with CF is often uncertain with respect to pathogenesis or clinical course (43), the high prevalence of NTM in all our patients older than 40 highlights an important point for clinicians. All patients who present with bronchiectasis and are culture positive for NTM, regardless of age and associated symptoms, should be screened for CF with both a sweat chloride test and genotyping, and if CF is diagnosed, referral to a cystic fibrosis care and research center should be considered. Our analysis suggests that a significant population of individuals who meet classic diagnostic criteria for CF present late in life, often with respiratory symptoms that have been present for decades. Remarkably, 47 of the 55 patients described in this report are still alive, emphasizing that long-term survival of patients with CF may not be as uncommon as previously believed.
In summary, this study is the first to examine a large cohort of very long-term survivors with CF, which can be divided into two distinct populations on the basis of age at diagnosis. The subset of individuals who were diagnosed early in life matched the overall CFF registry database and the prior literature with respect to CFTR allelic frequency, prevalence of pancreatic insufficiency, CFRD, and microbiology. Yet, the average current age of this group was 45.4 years, placing these individuals in the oldest 5% of patients nationally. This ED subset could therefore be particularly useful for identifying modifier genes, because these individuals' genotype is typical of the CF population in the United States, but their survival is significantly greater. The population that was diagnosed after age 24 was more heterogeneous, with fewer typical manifestations of CF and less severe CFTR mutations, and thus may be less suitable for analysis of genetic modifiers. The high incidence of NTM as a presenting manifestation of CF suggests that acquisition of NTM may exacerbate the phenotype of some of these patients after a long period of compensated, undiagnosed disease, ultimately bringing them to clinical attention.
Acknowledgments
The authors thank Hebe B. Quinton at Dartmouth Hitchcock Medical Center for providing CFF registry data.
REFERENCES
Rosenstein BJ, Zeitlin PL. Cystic fibrosis. Lancet 1998;351:277eC282.
Morrissey BM, Schock BC, Marelich GP, Cross CE. Cystic fibrosis in adults: current and future management strategies. Clin Rev Allergy Immunol 2003;25:275eC287.
Cystic Fibrosis Foundation. Patient registry 2003 annual data report. Bethesda, MD: Cystic Fibrosis Foundation; 2004.
Boyle MP. Nonclassic cystic fibrosis and CFTR-related diseases. Curr Opin Pulm Med 2003;9:498eC503.
Highsmith WE, Overbeek SE, Hilvering C. A novel mutation in the CF gene in patients with pulmonary disease but normal sweat chloride concentrations. N Engl J Med 1994;331:974eC980.
Mickle JE, Cutting GR. Genotype-phenotype relationships in cystic fibrosis. Med Clin North Am 2000;84:597eC607.
Burke W, Aitken M, Chen SH, Scott CR. Variable severity of pulmonary disease in adults with identical cystic fibrosis mutations. Chest 1992;102:506eC509.
Sanders JS, Pryor TD, Wedel MK. Prolonged survival in an adult with cystic fibrosis. Chest 1980;77:226eC227.
Fiel S. Cystic Fibrosis in an adult. Emerg Med 1988;20:109eC121.
Hunt B, Geddes D. Newly diagnosed CF in middle and later life. Thorax 1984;40:23eC26.
Su C, Beanblossom B. Typical CF in an elderly woman. Am J Med 1989;86:701eC703.
McCloskey M, Redmond AO, Hill A, Elborn JS. Clinical features associated with a delayed diagnosis of cystic fibrosis. Respiration (Herrlisheim) 2000;67:402eC407.
Gan KH, Geus WP, Bakker W, Cornelis CBHW, Heijerman HGM. Genetic and clinical features of patients with cystic fibrosis diagnosed after the age of 16 years. Thorax 1995;50:1301eC1304.
Widerman E, Millner L, Sexauer W, Fiel S. Health status and sociodemographic characteristics of adults receiving a CF diagnosis after age 18 years. Chest 2000;118:427eC433.
Wallick K, Sontag M, Jones M, Brayshaw S, Accurso F, Rodman D, Nick J. Clinical characteristics of adult patients over age 40 with CF: significance of early and delayed diagnosis. Paediatr Pulm 2002;S24:319.
Rosenstein BJ, Cutting GR. The diagnosis of cystic fibrosis: a consensus statement. Cystic Fibrosis Foundation consensus panel. J Pediatr 1998;132:589eC595.
Smith EP, Pontasch KW, Cains J Jr. Community similarity and the analysis of multispecies environmental data: a unified statistical approach. Water Res 1990;24:507eC514.
Wallace RJ, Glassroth J, Griffith DE, Olivier KN, Cook JL, Gordin F. Diagnosis and treatment of disease caused by nontuberculous mycobacteria. Am J Respir Crit Care Med 1997;156:S1eCS25.
Bonforte RJ. Variability in cystic fibrosis. JAMA 1978;240:1855.
Lai HJ, Cheng Y, Cho H, Kosorok MR, Farrell PM. Association between initial disease presentation, lung disease outcomes, and survival in patients with cystic fibrosis. Am J Epidemiol 2004;159:537eC546.
Feldmann D, Couderc R, Audrezet MP, Ferec C, Bienvenu T, Desgeorges M, Claustres M, Mittre H, Blayau M, Bozon D, et al. CFTR genotypes in patients with normal or borderline sweat chloride levels. Hum Mutat 2003;22:340.
Groman JD, Meyer ME, Wilmott RW, Zeitlin PL, Cutting GR. Variant cystic fibrosis phenotypes in the absence of CFTR mutations. N Engl J Med 2002;347:401eC407.
Choo-Kang LR, Zeitlin PL. Type I, II, III, IV, and V cystic fibrosis transmembrane conductance regulator defects and opportunities for therapy. Curr Opin Pulm Med 2000;6:521eC529.
Corey M, Edwards L, Levison H, Knowles M. Longitudinal analysis of pulmonary function decline in patients with cystic fibrosis. J Pediatr 1997;131:809eC814.
Durno C, Corey M, Zielenski J, Tullis E, Tsui L, Durie P. Genotype and phenotype correlations in patients with CF and pancreatits. Gastroenterology 2002;123:1857eC1864.
Vizza CD, Yusen RD, Lynch JP, Fedele F, Alexander Patterson G, Trulock EP. Outcome of patients with cystic fibrosis awaiting lung transplantation. Am J Respir Crit Care Med 2000;162:819eC825.
The Cystic Fibrosis Genotype-Phenotype Consortium. Correlation between genotype and phenotype in patients with CF. N Engl J Med 1993;329:1308eC1313.
O'Connor GT, Quinton HB, Kahn R, Robichaud P, Maddock J, Lever T, Detzer M, Brooks JG. Case-mix adjustment for evaluation of mortality in CF. Pediatr Pulmonol 2002;33:99eC105.
Demko CA, Byard PJ, Davis PB. Gender differences in CF: Pseudomonas aeruginosa infection. J Clin Epidemiol 1994;43:125eC131.
Gibson RL, Burns JL, Ramsey BW. Pathophysiology and management of pulmonary infections in cystic fibrosis. Am J Respir Crit Care Med 2003;168:918eC951.
Emerson J, Rosenfeld M, McNamara S, Ramsey B, Gibson RL. Pseudomonas aeruginosa and other predictors of mortality and morbidity in young children with cystic fibrosis. Pediatr Pulmonol 2002;34:91eC100.
Huang NN, Schidlow DV, Szatrowski TH, Palmer J, Laraya-Cuasay LR, Yeung W, Hardy K, Quitell L, Fiel S. Clinical features, survival rate, and prognostic factors in young adults with cystic fibrosis. Am J Med 1987;82:871eC879.
Nixon GM, Armstrong DS, Carzino R, Carlin JB, Olinsky A, Robertson CF, Grimwood K. Clinical outcome after early Pseudomonas aeruginosa infection in cystic fibrosis. J Pediatr 2001;138:699eC704.
Gabolde M, Guilloud-Bataille M, Feingold J, Besmond C. Association of variant alleles of mannose binding lectin with severity of pulmonary disease in cystic fibrosis: cohort study. BMJ 1999;319:1166eC1167.
Merlo CA, Boyle MP. Modifier genes in cystic fibrosis lung disease. J Lab Clin Med 2003;141:237eC241.
Salvatore F, Scudiero O, Castaldo G. Genotype-phenotype correlation in cystic fibrosis: the role of modifier genes. Am J Med Genet 2002;111:88eC95.
Olivier KN, Weber DJ, Wallace RJ Jr, Faiz AR, Lee JH, Zhang Y, Brown-Elliot BA, Handler A, Wilson RW, Schechter MS, et al. Nontuberculous mycobacteria. I: multicenter prevalence study in cystic fibrosis. Am J Respir Crit Care Med 2003;167:828eC834.
Olivier KN, Weber DJ, Lee JH, Handler A, Tudor G, Molina PL, Tomashefski J, Knowles MR. Nontuberculous mycobacteria. II: nested-cohort study of impact on cystic fibrosis lung disease. Am J Respir Crit Care Med 2003;167:835eC840.
Olivier KN, Weber DJ, Wallace RJ, Faiz AR, Lee JH, Zhang Y, Brown-Elliot BA, Handler A, Wilson RW, Schechter MS, et al. Nontuberculous mycobacteria: I. Multicenter prevalence study in cytic fibrosis. Am J Respir Crit Care Med 2003;167:828eC834.
Kilby JM, Gilligan PH, Yankaskas JR, Highsmith WE, Edwards LJ, Knowles MR. Nontuberculous mycobacteria in adult patients with cystic fibrosis. Chest 1992;102:70eC75.
Aitken ML, Burke W, McDonald G, Wallis C, Ramsey B, Nolan C. Nontuberculous mycobacterial disease in adult cystic fibrosis patients. Chest 1993;103:1096eC1099.
Fauroux B, Delaisi B, Clement A, Saizou C, Moissenet D, Truffot-Pernot C, Tournier G, Vu Thien H. Mycobacterial lung disease in cystic fibrosis: a prospective study. Pediatr Infect Dis J 1997;16:354eC358.
Nick JA. Nontuberculous mycobacteria in cystic fibrosis. Semin Respir Crit Care Med 2003;24:693eC702.(David M. Rodman, Jillian )