Summary of main findings
The 5-year cumulative incidence of moderate COPD (GOLD II) in a cohort of male smokers, either with normal spirometry or mild COPD (GOLD I) at baseline, was 8.3% (95% CI = 5.8 to 11.4%), with a mean annual incidence of 1.6%. The incidence of GOLD II was five times higher in participants with baseline GOLD I than in those with baseline normal spirometry. Patient characteristics that were independently associated with incident GOLD II were increasing age, childhood smoking (≤15 year of age), cough, and one or more GP contacts for lower respiratory tract problems at baseline.
Strengths and limitations of the study
Some limitations of the study should be considered. Only men were included. Sex differences in lung vulnerability to tobacco smoking are currently being investigated; therefore, the estimated incidence of moderate COPD in this study of male smokers can not be generalised to female smokers.14–17 There were 702 participants in the baseline survey and 436 in the second survey. In those only participating in the baseline survey the prevalence of GOLD II at baseline was slightly higher (4.0 versus 2.4%) and FEV1 percentage predicted was slightly lower (98 versus 102%) than in those participating in both surveys. This may indicate that a slightly higher proportion of participants with relatively poorer lung function tended to discontinue participation, and that the incidence rate of GOLD II is possibly slightly underestimated. As non-participation in the baseline survey could have been higher among smokers with poorer lung function, the estimated 8.3% cumulative incidence is likely to be conservative.
To assess potential bias associated with using different spirometers at baseline (Vitalograph) and follow-up (Jaeger), the change in FEV1 in a subset of 171 participants who performed an additional lung function test in 2002 was compared with results from the Vitalograph spirometer. The change in FEV1 between the baseline and additional tests, both performed with the Vitalograph spirometer, was similar to the change in FEV1 between the baseline and the follow-up tests performed with the Vitalograph and Jaeger spirometer, respectively (63.5 versus 60.6 ml/year, t-test for difference P = 0.6).
One of the strengths of the study is that the survey was performed in a population representative of the Dutch population; for example 35% of those who returned the questionnaire on smoking habits at the baseline survey were current smokers: a figure similar to the expected proportion of smokers (35–36%) in men aged 40–65 years in the Netherlands.18 Moreover, a high proportion (87.7%) of the eligible male smokers participated. Thus, selective response seems unlikely. In addition, all participants with a new diagnosis of moderate COPD performed postbronchodilator lung function measurements as recommend by the GOLD guidelines.
Comparison with existing literature
A limited number of studies have provided incidence rates of COPD.19–22 In a Finnish study the mean annual incidence of COPD (FEV1/FVC <0.6) among male smokers aged 40–64 years was about 0.5%.19 In a Polish study the mean annual incidence of COPD (FEV1 <65% predicted) in male smokers aged 41–60 years was 1.2–1.6%.20 Using different definitions of COPD and prebronchodilator lung function values in those studies, limits comparison of the annual incidence estimates with the current study. A more recent study showed a 10-year cumulative incidence of moderate COPD (GOLD II) in male smokers aged 51–52 years of about 25%,22 which is comparable to the current study's 5-year incidence in 50–54 year-old male smokers.
Univariate analysis in the current study showed that the incidence of COPD in smokers is associated with age and smoking history, that is, the number of pack years. In the multivariate analysis the number of pack years disappeared when age and childhood smoking were included.
Several determinants of COPD, such as childhood smoking, respiratory symptoms, a family history of obstructive lung disease, and parental smoking, have been implicated to varying degrees.1,23,24 In one study, childhood smoking was an independent risk factor of obstructive airways disease in women but not in men as diagnosed by a doctor.23 In contrast, the current study showed that childhood smoking was an independent risk factor of GOLD II, which can possibly be explained using different definitions of airflow obstruction. Support for the increased risk of COPD with childhood smoking in both sexes was given in a prognostic study, which found that cigarette smoking in males and females, 10–18 years of age, is associated with mild airflow obstruction and slowed growth of lung function.24 A similar incidence of GOLD II in symptomatic and asymptomatic smokers with normal baseline spirometry (7.4 versus 6.7%) has been previously reported,25 while the current study found that the incidence of GOLD II in symptomatic smokers was three times higher than in asymptomatic smokers (9.3 versus 2.9%).
An association between bronchitic symptoms, such as cough or sputum production, and the incidence of GOLD II is supported by the results of a Swedish study which found a significant association between each bronchitic symptom and the incidence of GOLD II after adjustment for possible confounders.22
In the current study, no association between a family history of obstructive lung disease and the incidence of GOLD II was found. An association between positive family history and the development of airflow obstruction may be undetected by poor recall in participants completing questionnaires, or by smaller family size in contemporary generations compared with earlier generations.
In a European study, parental smoking, especially smoking by fathers, was related to a poorer lung function among smoking and non-smoking adults aged 20–44 years.26,27 However, the current study found no significant association between parental smoking and the incidence of GOLD II. This may be due to lack of contrast, as parental smoking during childhood was reported by almost 90% of the participants in the current study.