INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: chronic obstructive pulmonary disease; epidemiology; airflow obstruction; family studies; genetics

Chronic obstructive pulmonary disease (COPD) is defined as airflow obstruction that does not change appreciably over a period of several months (1). It is a syndrome that includes chronic bronchitis, emphysema, and chronic irreversible asthma. COPD is a major cause of morbidity and mortality that affected 44 million people worldwide in 1990. It is estimated that 2.88 million people died from this condition in the year 2000 (2). Indeed, 14 million people suffer from COPD in the United States alone, and this resulted in nearly 92,000 deaths in 1995 (3).

The major environmental risk factor for the development of COPD is cigarette smoking (4). In nonsmokers the FEV₁ declines at a rate of approximately 20 to 30 ml/yr during adult life. In most smokers this is increased to 30 to 45 ml/yr, but in the 10 to 15% of smokers who are particularly susceptible to cigarette smoke, the rate of decline is 80 to 100 ml/yr (5). There is evidence of a dose-response between the severity of lung disease and the pack-years of cigarettes smoked (4, 6) but the correlation coefficient is low, indicating that only 15% of the variability in FEV₁ is accounted for by this factor. The low percentage of variance suggests that additional factors contribute to the impact of smoking on the development of airflow obstruction (see references 9 and 10 for review). The most important genetic factor in the development of emphysema is the Z allele of ₁-antitrypsin, which results in plasma levels that are 10 to 15% of the normal M allele (11). Homozygotes for the Z allele are greatly predisposed to developing emphysema if they smoke (12, 13). However Z ₁-antitrypsin deficiency makes up only 1 to 2% of all cases of emphysema and heterozygotes do not have a clearly increased risk of lung damage (14). Three previous studies have suggested that genetic factors other than ₁-antitrypsin deficiency may be involved in the susceptibility of smokers to chronic airflow obstruction (17). In all of these studies there was a significantly higher prevalence of COPD among relatives of index patients than among the control group. These findings have been confirmed by Silverman and colleagues (20) in a study of patients with early onset severe COPD.

We report here a significantly increased risk of airflow obstruction in the siblings of probands with severe COPD, which we have defined as irreversible airflow obstruction with a reduced gas transfer factor (21).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Identification of Probands

Index cases (probands) with severe COPD were identified from chest clinics and lung function laboratories in the East Anglian region of the United Kingdom and from referrals to a regional transplant unit for assessment for lung volume reduction surgery and lung transplantation. The probands were enrolled into the study if they had: FEV₁ < 60% predicted, and were 55 yr of age or younger, FEV₁ < 40% predicted and were 56 to 60 yr of age, or FEV₁ < 20% predicted and were 61 to 65 yr of age; FEV₁/FVC ratio < 0.6, gas transfer factor (Kco) more than 10% below the lower limit of normal, ₁-antitrypsin level in the normal range, or an MM, MS, or MZ phenotype and a chest radiograph that was compatible with COPD. The probands were contacted by one of us by phone or in person and asked to complete the American Thoracic Society respiratory questionnaire (24) to record smoking history, respiratory symptoms, medical history, occupational history, and demographic details. Lung function data, ₁-antitrypsin level and/or phenotype, and CT scan appearances (when available) were then recorded from the patients' hospital notes.

Screening of Parent and Siblings

Living parents and siblings were contacted by telephone or in person and asked to complete the same questionnaire as the probands. They were then invited to a clinic run by their local general practitioner or a research nurse for simple spirometry (FEV₁ and FVC). If the screening spirometry showed airflow obstruction (FEV₁/FVC < 0.7) then they were invited to attend their local hospital for full lung function tests (RV, TLC, and gas transfer factor) and reversibility to ₂-agonists.

Estimation of the Association between Family History and COPD

Control subjects were randomly selected from the population-based EPIC-Norfolk cohort study. This cohort forms part of a multicenter international cohort designed to investigate the relationship between diet, cancer, and chronic disease. The detailed design and operation of this study have been described previously (25). Briefly, at baseline survey between 1993 and 1998, 77,630 men and women 45 to 74 yr of age were identified from general practices in Norfolk, UK, and invited to participate. In the United Kingdom virtually all citizens are registered with a general practice, and thus the cohort represents a population-based sampling frame. From these, 30,447 agreed to participate and provided informed consent and 25,633 volunteers attended for a health check that included the completion of a detailed health and lifestyle questionnaire. This contained questions on lifetime smoking, including the daily number of cigarettes smoked at ages 20, 30, 40, and 50 years, and the number smoked at the time of the study. Cumulative cigarette consumption in pack-years was calculated from these data.

FEV₁ and FVC were measured in all participants with an electronic turbine spirometer (Micro Medical Instruments, Rochester, UK). Two measurements were made with the subjects standing and facing forward by nurses who had been trained in spirometry. The nurses made a subjective judgement of the subject's spirometry technique and recorded it as good or poor. The higher of the two values for FEV₁ and FVC were used for analysis. Student's t test was used to compare FEV₁ and FVC between the study group and the population- based control subjects, and Pearson's chi-square test was used to compare the proportion of participants with airflow obstruction within the two groups. The odds ratio for the development of COPD in the study group relative to control group was assessed using matched (conditional logistic regression) analysis. All analyses were performed using the statistical package Stata (Stata Corp, College Station, TX, USA).

Ethical Committee Approval

The project was approved by the Anglia and Oxford Multicenter research ethics committee and by local hospital research ethics committees. All patients gave informed consent.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Probands

One hundred fifty-two probands were identified who fulfilled the inclusion criteria for severe COPD. Two declined to join the study, and 150 (98.7%) were enrolled; 149 were Caucasian and one was from Pakistan. The mean age of the group was 52.3 SD ± 6.9 yr, the male:female ratio was 1:0.86, the mean FEV₁ was 0.9 L (SD ± 0.4) and mean Kco was 47% predicted (Table 1). All the probands smoked cigarettes, with a mean of 46.8 pack-years for men and 37.4 pack-years for women (Table 2). Only four of the probands were current or ex-pipe smokers and eight were current or ex-cigar smokers. These subjects also smoked cigarettes (mean pack-years smoked, 72.0 [SD ± 19.9] and 27.0 [SD ± 12.2] for pipe and cigar smokers, respectively). The minimum exposure to cigarette smoking by any proband was 4 pack-years (three of 150 probands, two male and one female). In two of these patients CT scans confirmed extensive emphysema, and the third underwent lung transplantion for emphysema. One of the male patients worked as a coal fire engineer, but the other two had no occupational dust exposure.

Six of the probands reported a pneumothorax, and two had undergone pleurectomy. Two had undergone a bullectomy, and one required surgery to the chest for a stab wound. Nineteen of the probands required heart-lung or lung transplantation for emphysema, and two had had lung volume reduction surgery, but in all cases their preoperative lung function tests fulfilled the inclusion criteria. The mean age of the patients at transplantation was 52.2 (SD ± 6.1) and this was an average of 3.8 (SD ± 1.9) yr before they were recruited into the study. Twenty-seven of the probands gave a history of pleurisy, and two had a previous history of tuberculosis. Breathlessness was scored according to the MRC classification; Grade I, have to walk more slowly than someone of your own age because of breathlessness; Grade II have to stop for breath when walking on the flat; Grade III short of breath after walking 100 yards and Grade IV too breathless to leave the house. After exclusion of 24 patients who had undergone surgical intervention, 11 (9%) of the probands had MRC Grade I breathlessness, 23 (18%) Grade II, 12 (10%) Grade III, and 76 (60%) Grade IV; four patients were not limited by breathlessness. Thirty-eight (47%) of men and 29 (42%) of women fulfilled the criteria for chronic bronchitis, and 28% of men and 35% of women reported respiratory symptoms before the age of 16. Fifty-eight of the 150 probands (39%) had had CT scans, and all of these were consistent with the diagnosis of emphysema. In three patients (2%) the diagnosis of emphysema was confirmed by histologic examination of the lung removed at transplantation.

Parents and Siblings of the Probands

Thirty-three of the 150 probands had one living parent and five had two living parents (total, 38). The male:female ratio of living parents was 0.3:1 and their mean age was 75.9 (SD ± 9.8) yr. The mean FEV₁ was 1.52 L (SD ± 0.62), which was 80.2% predicted. Twenty seven (71%) of the parents were current or ex-smokers (mean pack-years 38.4 ± 24.6) and 12 of these (44%) had airflow obstruction as defined by FEV₁/ FVC < 0.7. The mean FEV₁ in the parents with airflow obstruction was 1.1 L (SD ± 0.47; 50% predicted).

Two of the probands were adopted, 14 had no siblings, and 21 had no relatives that were contactable. The remaining 113 probands had a total of 221 siblings who could be contacted, of whom 192 (87%) completed the questionnaire and 173 (78%) completed the study. Of the 19 patients who completed the questionnaire but did not undergo spirometry, one developed carcinoma of the breast, one moved to the United States, one moved without a forwarding address, and 16 did not attend at least two follow-up appointments for spirometry. Thus, information on 173 siblings was available for analysis (Table 3). The ratio of men:women was 1:1.2, and 126 of 173 (73%) of the siblings (62 men, 64 women) had smoked more than one-pack year of cigarettes. Those siblings who were never smokers had normal spirometry. Twenty-four of the 62 men who were current or ex-smokers (39%) had airflow obstruction with an FEV₁/FVC < 0.7. In 15 participants this was irreversible after the administration of 200 µg salbutamol. In five, there was an increase in FEV₁ that met the British Thoracic Society definition of reversibility (15% and at least 200 ml) (1), but in all five the postbronchodilator dynamic lung volumes still showed airflow obstruction (FEV₁/FVC < 0.7). Four of the 24 siblings with airflow obstruction did not have formal reversibility studies; two refused to use the inhaler, one failed to attend for further spirometry, and one had a reduced gas transfer on full lung function testing suggestive of emphysema but did not have an assessment of reversibility. Thus, 20 of the 62 men (32%) who were current or ex-smokers had proven irreversible airflow obstruction. If a FEV₁/FVC < 0.7 and a FEV₁ < 80% predicted was used to define cases (1), then 22 of the 62 men who were current or ex-smokers (35%) would be affected.

Twenty of the 64 women who smoked (31%) had airflow obstruction with an FEV₁/FVC < 0.7. In 17, this was irreversible after the administration of 200 µg salbutamol. In two, there was an increase in FEV₁ that met the British Thoracic Society definition of reversibility (15% and at least 200 ml) (1), but in both siblings the postbronchodilator dynamic lung volumes still showed airflow obstruction (FEV₁/FVC < 0.7). One subject failed to attend for further spirometry. Thus, 19 of the 64 women (30%) who were current or ex-smokers had proven irreversible airflow obstruction. A total of 14 of the 64 women (22%) fulfilled the affected criteria (FEV₁/FVC < 0.7 and a FEV₁ < 80% predicted) (1). Taken together 44 of 126 current or ex-smoking siblings had airflow obstruction (35%), and in all 39 who were tested this was irreversible to bronchodilator therapy. A total of 36 of 126 (29%) siblings fulfilled the criteria for affected status.

Of the 44 current or ex-smokers with airflow obstruction 22 agreed to undergo formal lung function tests, and in these 12 (55%) had a Kco < 70% predicted in keeping with emphysema. If the criteria for emphysema was raised to a Kco < 80% predicted then 13 of the 22 who underwent formal lung function tests would be affected. High-resolution CT scans were available on five others with airflow obstruction who refused to have full lung function tests. In all cases the CT scans showed evidence of emphysema, which ranged from mild to severe disease. Seven of the 113 families had two or more siblings with airflow obstruction, and in six of these families a sibling fulfilled the criteria for emphysema (five had emphysema confirmed by a Kco < 70% predicted and one had emphysema confirmed by HRCT scan). Attempts to assess the prevalence of lung disease in extended families (first cousins, uncles, and aunts) were unsuccessful.

Association of Family History of Emphysema with Risk of COPD

Current and ex-smoking siblings were matched with control subjects from the European Prospective Investigation of Cancer (EPIC) database (25). An attempt was made to match each current and ex-smoking sibling 39 to 79 yr of age with four control subjects from the database. Each sibling was matched by age (± 5 yr), sex, and smoking history (± 2 pack-years). Siblings who smoked "roll up" tobacco were matched by converting each ounce of tobacco smoked to 30 cigarettes. Pipe and cigar tobacco consumption was not included in the matching. One hundred fifteen current or ex-smoking siblings were within the age range, four of these were not matched because of incomplete FEV₁ or FVC data. In each of these patients only the percent predicted values were available, and all were within the normal range. Therefore, 111 siblings were matched to a total of 439 control subjects, 100 were matched with four control subjects, three were matched with three control subjects, two were matched with two control subjects and six were matched with one control subject. The characteristics of the siblings and control subjects are summarized in Table 4. The mean FEV₁ was not significantly different between the two groups (mean FEV₁, 2.34 L in the siblings compared with 2.50 L in the matched control subjects, p = 0.065). The control group had a lower mean FVC than the study group (3.05 L in control subjects and 3.30 L in the siblings, p = 0.018) and a significantly lower mean FEV₁/FVC ratio (0.72 compared with 0.83 in the matched control subjects, p < 0.00005). The proportion with airflow obstruction (FEV₁/FVC < 0.7) was much lower in the control group (52 of 439) when compared with matched siblings (45 of 111, p < 0.0005). Moreover, 9.3% of the EPIC cohort fulfilled the criteria for COPD (FEV₁ < 80% predicted and FEV₁/FVC < 0.7) compared with 31.5% of the siblings (p < 0.0005). Matched analysis, using conditional logistic regression, gave an odds ratio for COPD of 4.70 in siblings relative to control subjects (p < 0.0001, 95% confidence interval, 2.63 to 8.41).

The control subjects were matched for total pack-years of smoking and not by current smoking status. The inclusion of smoking status (current or ex-smoker) as an additional variate did not significantly affect the result of the matched analysis (odds ratio for an association between family history of emphysema and COPD, 4.64; p < 0.0001; 95% confidence interval, 2.47 to 8.09). When the analysis was repeated for siblings with less than a 30 pack-year smoking history, the odds ratio for COPD was 5.39 for siblings compared with matched control subjects (p < 0.0001; 95% confidence interval, 2.49 to 11.67). The odds ratio for siblings with a greater then 30 pack-year history was 3.67 (p < 0.0017; 90% confidence interval, 1.53 to 8.80).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our results demonstrate conclusively the familial clustering of airflow obstruction in families ascertained through a proband with airflow obstruction and a low Kco. The proband phenotype was chosen to represent emphysema (26). However, recent studies by Gelb and colleagues (21) have demonstrated that Kco does not correlate well with emphysema in patients with a FEV₁ < 1 L. Thus, the probands have been defined as having severe COPD. Our study is different from previous investigations as (1) COPD has been defined as airflow obstruction with a graded fall in FEV₁ with increasing age, (2) the low Kco defines a group of patients with severe COPD and helps to exclude asthma, (3) a large proportion of the probands and siblings who were approached agreed to join the study, and (4) it is the only study that has used a large population-based control group to determine a value for relative risk.

Although not an inclusion criteria, all the probands were current or ex-cigarette smokers. None of the probands smoked only cigars or a pipe, which is consistent with previous reports that these tobacco products are less likely to be inhaled and to cause airflow obstruction (27). There was no correlation between the values of FEV₁ and Kco and the number of pack-years of cigarettes smoked (Figure 1). Indeed three of the probands had only a 4 pack-year smoking history. These findings are in keeping with population studies that have also shown a poor correlation between the level of FEV₁ and smoking history (28, 29).

View larger version (13K): [in this window] [in a new window]   View larger version (12K): [in this window] [in a new window]   Figure 1.   Correlation between FEV1 (upper panel  ) and gas transfer factor (Kco) (bottom panel  ) with pack-years of cigarettes smoked in probands with emphysema. There was no significant correlation between either parameter and smoking history.

A previous study by Silverman and colleagues (20) assessed probands with early onset severe COPD. They identified probands 52 yr of age or younger with a FEV₁ < 40% predicted. They found a similar aggregation of disease within families, but their data differed from ours in having a large excess of female probands (30). If our current study is analyzed using the same inclusion criteria (FEV₁ < 40% predicted, 52 yr of age or younger) then 55 probands would meet the criteria of whom 23 were female (42%) and 32 were male (58%). Thus, our data do not support the hypothesis that women develop more severe emphysema at a younger age.

Approximately one quarter of the siblings who completed the questionnaire and spirometry were nonsmokers. These siblings had essentially normal lung function. This reinforces the clear message that irreversible airflow obstruction is linked to tobacco smoking. Moreover, the normal FEV₁ in nonsmoking siblings argues against other environmental factors preventing normal lung development in adolescence and so explaining the familial clustering of COPD. Thirty-two percent of the current or ex-smoking siblings met the British Thoracic Society criteria for COPD (1). This is much higher than in the EPIC-Norfolk matched control subjects, and the risk of airflow obstruction was even greater among siblings who had smoked less than 30 pack-years. Our results show that the risk of COPD in siblings of an affected individual was significantly greater (odds ratio, 4.7) than that determined previously by other groups who focused on milder COPD as the inclusion criteria for the proband (17, 18). However, the result was similar to the odds ratio of 4.5 reported by Silverman and colleagues (20) who recruited families via a proband with severe early onset COPD and compared current and ex-smoking siblings with unmatched control subjects.

The findings reported here are unlikely to be due to chance in view of the highly significant difference between the two populations. The possibility of confounding has been diminished by matching for age, sex, and smoking history. Moreover, the use of a population-based cohort for the control group will minimize selection bias. Although we cannot exclude residual confounding, for example, by industrial exposure, it would be unlikely to explain an association of this magnitude. The most important issue is of potential bias in the measurement of pulmonary function in the two populations. The recordings of FEV₁ and FVC in subjects recruited to the EPIC cohort are without inhaled bronchodilator medication. A small proportion of the control population may have unrecognized airflow obstruction, which would underestimate the strength of the association between a positive family history and the development of airflow obstruction. Moreover, some of the control group may have unreported first-degree relatives with COPD, which will also result in a conservative estimate of the association between family history and risk of COPD. Although care was taken in all recording of spirometry, and the nurses in the study and control group used the same make of portable spirometer, neither the siblings nor the control subjects had their spirometry performed in the same lung function laboratories or by the same nurses. This may introduce error into the data. Overall, however, it is likely that the results reported here are a conservative estimate of the true familial risk of airflow obstruction in smoking siblings of a patient with severe COPD.

Taken together this study demonstrates marked familial clustering of airflow obstruction in persons who smoke cigarettes. These data also provide an assessment of the ability to recruit nuclear families for genetic studies after the identification of a proband with severe COPD. It is possible that factors such as childhood passive smoking, infections (31), or diet (32) could interact with smoking to account for the familial aggregation of disease. However, the most likely explanation is a common genetic susceptibility to COPD. Our conclusions are in keeping with those of previous studies (17) and provide strong support for strategies designed to identify genes that predispose smokers to airflow obstruction (10, 33).