Asthma and chronic obstructive pulmonary disease (COPD) are caused by complex interactions between environmental and genetic factors. A putative genetic risk factor for asthma and COPD is the cystic fibrosis transmembrane conductance regulator (CFTR) gene 123. This gene encodes a cAMP-regulated channel with chloride activity in pulmonary epithelia. When channel activities are absent, cystic fibrosis with life-threatening airways obstruction due to thickened secretions and secondary pulmonary infection develop 4. The most common cause of cystic fibrosis is homozygosity for the phenylalanine-508 deletion (F508del), explaining about 70% of cystic fibrosis worldwide 45.

We previously showed that persons heterozygous for a F508del deletion are overrepresented among people with asthma 16. Another more common variant, the 5T allele, could likewise be involved in asthma 7 or COPD. This variation is in the polythymidine tract of the CFTR gene and has mainly been associated with congenital bilateral absence of the vas deferens, a monosymptomatic form of cystic fibrosis 8910. However, it may also be associated with increased risk of obstructive lung disease, particularly bronchiectasis 91011121314. Because most previous studies on lung disease in 5T carriers were based on case patients 29101112131415161718192021222324, currently we know little about the risk for obstructive lung disease in 5T carriers in the general population.

Three common alleles are known in the polythymidine tract, 5T, 7T, and 9T. The polythymidine tract is situated in intron-8 near the acceptor splice site for exon-9 2526. The shorter this polythymidine tract is, the more often exon-9 is skipped from CFTR mRNA. Transcripts missing exon-9 increases from 1%–13% in 9T homozygotes 272829 to 12%–25% in 7T homozygotes 1327282930 to 66%–90% in 5T homozygotes 13273132. CFTR mRNA without exon-9 leads to a protein with no chloride channel activity 3334. Thus, carriers of 5T with high exon-9 skipping have reduced channel activities and could have increased susceptibility for obstructive lung disease. This could be particularly relevant for 5T carriers exposed to additional risk factors for lung disease such as tobacco smoke or familial predisposition to lung disease. Variations in the genes for mannose-binding lectin and α₁-antitrypsin have been studied as modifiers of cystic fibrosis lung disease 353637 and could also potentially influence risk of lung disease in 5T heterozygotes. Allele frequencies in whites are approximately 5% for the 5T allele, 84% for 7T, and 11% for 9T 2526.

We hypothesised that carriers of the 5T allele have increased annual lung function decline and increased risk for asthma or COPD. To test this hypothesis, we genotyped 9131 individuals from the adult Danish population for the 5T, 7T, and 9T alleles in the CFTR gene. We combined polythymidine and F508del genotypes 1, and examined associations between 11 different genotype combinations, and annual FEV₁ decline and risk of asthma or COPD. We also examined whether other common risk factors for lung disease or variations in the genes for mannose-binding lectin and α₁-antitrypsin significantly add to risk of lung disease in 5T carriers.

Subjects participated in the 1976–78, 1981–83, and/or 1991–94 examination of the Copenhagen City Heart Study, a prospective epidemiological study initiated in 1976–78 38. Participants aged 20 years and above were selected randomly after age stratification into 5-year age groups from among residents of Copenhagen. Of the 17180 individuals invited, 10135 participated, 9259 gave blood, and 9131 were genotyped for the polythymidine tract variants of the cystic fibrosis conductance membrane regulator (CFTR) gene. Details of study procedures and some characteristics of non-responders are described elsewhere 3839. More than 99% were Whites of Danish descent. All participants gave written informed consent, and Herlev University Hospital and the ethics committee for Copenhagen and Frederiksberg approved the study (# 100.2039/91).

Participants filled out a self-administered questionnaire, which was validated by the participant and an investigator on the day of attendance. Participants reported on long-term occupational exposure to dust or welding fumes, pulmonary symptoms (dyspnea, wheezing, bringing up phlegm), familial predisposition to asthma (having at least one sibling with asthma), smoking habits (current smoker, ex-smoker, never-smoker), type of smoking and daily tobacco consumption. An estimate of life-time tobacco exposure (in packyears) was calculated as: daily tobacco consumption (g) times duration of smoking (years) divided by 20 (g/pack). If at least once during the study period participants aswered "Yes" to the question "Do you suffer from asthma?", we recorded they had self-reported asthma. Medication for asthma / bronchitis was "Yes" to the question "Do you daily take medication for asthma / bronchitis?" Additional information on hospitalizations due to asthma (ICD8: 493; ICD10: J45–46) and COPD (ICD8: 491–492; ICD10: J41–44) was drawn from the Danish National Hospital Discharge Register from May 1st 1976 through December 31st 2000. We confirmed in the Danish National Hospital Discharge Register covering all hospital discharges in Denmark, that no participants in the sample were ever hospitalized for cystic fibrosis.

Forced expiratory volume in one second (FEV₁) and forced vital capacity (FVC) were measured with an electronic spirometer (model N403, Monaghan, Littleton, Colo.) at the 1976–78 and 1981–83 examinations and with a dry wedge spirometer (Vitalograph, Maidenhead, UK) at the 1991–94 examination. At each examination, three sets of values were obtained, and as a criterion for correct performance of the procedure, at least two measurements of FEV₁ and FVC differing by less than 5% had to be produced. The highest set of FEV₁ and FVC were used in the analyses as percentage of predicted value using internally derived reference values based on a subsample of healthy never smokers 40. Annual decline in FEV₁ (ml/year) was calculated as FEV₁ (ml) obtained at the latest measurement minus the FEV₁ value obtained at the first measurement, times 365.25 divided by the number of days between the two measurements (in years^-1). Spirometry defined COPD was FEV₁<80% predicted and FEV₁/FVC<0.7, excluding self-reported asthma 41.

We amplified the polythymidine tract variants in intron-8 by nested polymerase chain reaction using the primerpairs: 5'-TAATGGATCATGGGCCATGT-3'and 5'-ACAGTGTTGAATGTGGTGCA-3' (first step reaction), and 5'-

CCGCCGC

Linkage disequilibrium between the 9T and F508del alleles was tested by the linkage utility program "EH" http://linkage.rockefeller.edu, which estimates allele and haplotype frequencies with and without allelic association. The linkage disequilibrium coefficient D was calculated as D = P₂₂ - p₂q₂, where P₂₂ is the observed frequency of the 9T/F508del haplotype, p₂ is the frequency of the F508del allele in the general population and q₂ is the population frequency of the 9T allele. The degree of linkage disequilibrium was expressed as D' = D/D_max × 100%.

Statistical analysis was performed with SPSS; for power calculations, NCSS-PASS and StatMate were used. P < 0.05 on a two-sided test was considered significant. Pearson's χ²-test or analysis of variance (ANOVA) was used for overall comparisons between several genotypes; Pearson's or Fisher's Exact χ²-test were used for post-hoc two-genotype comparisons. The most common genotype combination in the population, 7T homozygosity without F508del, was used as reference group for statistical comparisons. We evaluated asthma and COPD prevalences between genotypes using unadjusted and adjusted logistic regression with Wald's test as a measure of significance; the adjusted model included gender, age at study entry (deciles), and packyears at study entry (never smokers and deciles). We evaluated asthma and COPD incidences between genotypes using the log-rank test 424344. Unadjusted and adjusted Cox regression with forced entry examined time to disease by using hazard ratios (relative risks) and 95% confidence intervals; the adjusted model included gender, age at study entry (deciles), tobacco use during follow-up (never smokers and deciles), and FEV₁ % predicted at study entry (deciles). We tested possible interactions between the 5T/7T genotype and smoking habits, long-term occupational exposure to dust or welding fumes, familial predisposition to asthma, α₁-antitrypsin MS genotype, α₁-antitrypsin MZ genotype, or mannose-binding lectin deficiency in predicting FEV₁ at study entry in ANCOVA models.

Characteristics of participants are given in Table 1; genotypes are ordered according to predicted increased skipping of exon-9 of the cystic fibrosis transmembrane conductance regulator gene, stratified for presence or absence of F508del heterozygosity. Among the 9,131 participants selected randomly from the Danish general population, 352 (3.9%) were 5T heterozygotes and 249 (2.7%) were F508del heterozygotes. Expected numbers of 5T and F508del heterozygotes according to the Hardy Weinberg equilibrium were 349 and 246, respectively. Allele frequencies did not differ from those predicted by the Hardy Weinberg equilibrium (χ²-test for 7T allele: P = 0.84; 9T allele: P = 0.60; 6T allele: P = 0.98; 5T allele: P = 0.42; F508del allele: P = 0.19). The novel intron-8 polythymidine tract variant, the 6T allele 45, was identified in four individuals. The 9T and F508del alleles were in linkage disequilibrium with a degree of linkage of 98% (χ²-test: P < 0.001).

This study shows that polythymidine 5T heterozygosity is not associated with increased annual decline in FEV₁ or risk of asthma or COPD in the adult Caucasian population; these results are independent of age, gender, tobacco smoking, and other potential confounders. Interestingly, however, both 5T homozygotes showed evidence of asthma. Furthermore, our results support that F508del heterozygosity is associated with increased asthma risk independently of the 5T allele.

Because 1 in 26 carries a 5T allele in this population, it is indeed important that 5T heterozygosity does not increase risk of obstructive lung disease in the population at-large. It appears that the 5T allele causes lung disease only in very rare circumstances 91011121314, leaving the average heterozygous individual unaffected by obstructive lung disease. Previous results suggest that penetrance of pulmonary manifestations in 5T carriers might depend on the length of an adjacent TG repeat 4647. This could be particularly relevant for 5T homozygotes and compound heterozygotes. In 5T heterozygotes, however, longer TG repeats seem less likely to affect risk of pulmonary disease. This is because 5T heterozygosity was not associated with risk of lung disease in this study although predicted TG12 and TG13 allele frequency in 5T carriers in our population was 31% 47. Other additional genetic variations have also been shown to influence exon-9 skipping in 5T carriers, but to a lesser degree than the TG repeat.

Because all 5T/F508del compound heterozygotes were free from severe pulmonary disease, the 5T allele did not appear to explain our previous results 16 suggesting that F508del heterozygosity may be overrepresented among asthmatics. A few recent studies also support this observation 21948, while others have found no 202149 or negative associations 50. In the present analyses, 7T/9T and 7T/7T individuals with F508del heterozygosity had higher prevalences of self-reported asthma, and 7T/7T individuals with F508del heterozygosity also had higher incidence of hospitalization from asthma. F508del heterozygosity was only associated with increased asthma risk in individuals without the 5T allele, indicating that our previous observations are independent of influence from this allele. In addition, both 5T homozygotes showed evidence of asthma supporting the hypothesis that CFTR variations may be associated with asthma 219.

To identify factors in the population that significantly add to risk of lung disease in 5T heterozygotes, we tested for interactions between 5T/7T genotype and potential risk factors for lung disease, but found no significant interactions. Garred 35 and coworkers found a worse prognosis in cystic fibrosis patients with MBL deficiency. We were not able to extend this finding, since lung function in 5T or F508del heterozygotes was not reduced by MBL deficiency. Previous studies by Mahadeva 36 and Frangolias 37 showed that pulmonary disease severity in cystic fibrosis patients were unaffected by α₁-antitrypsin S and Z alleles. In line with this, we also observed no increased risk for pulmonary dysfunction in 5T carriers with α₁-antitrypsin MS or MZ genotypes.

In the present study, bias caused by investigators' knowledge of disease or risk-factor status seems unlikely, because we selected from a general population and genotyped our sample without knowledge of disease status or lung function test results. Selection bias is possible if severe lung disease in some individuals with 5T genotypes prevented them from participating in our study; however, expected and observed numbers of these genotypes according to the Hardy-Weinberg equilibrium were similar. The 2.7% frequency of F508del heterozygosity found in this study is in accordance with the 2.9% frequency of F508del heterozygosity observed in another previous study of the Danish population 51. Annual decline in FEV₁ was reduced in 7T/9T individuals and incidence of COPD hospitalization was reduced in 5T/7T individuals. If correction for multiple comparisons was performed, these significant findings become nonsignificant. Therefore, and because reduced COPD risk in 5T/7T individuals is less biologically plausible, the findings are likely due to chance alone rather than representing real phenomena. Misclassification of genotypes is unlikely, because diagnoses were confirmed by sequencing a subsample of different poly-T variants.

Polythymidine 5T heterozygosity was not associated with increased annual decline in FEV₁ or risk of asthma or COPD in adults in this population-based study; however, both 5T homozygotes showed evidence of asthma. Furthermore, our results also support that F508del heterozygosity may be associated with increased asthma risk independently of the 5T allele.

The author(s) declare that they have no competing interests.

Morten Dahl, Anne Tybjærg-Hansen, and Børge G. Nordestgaard carried out the genotyping and statistical analysis. Peter Lange helped collect the data and was involved in the statistical analysis. All investigators participated in designing the study and in writing the paper, and all authors read and approved the final version of the manuscript.