BACKGROUND: The prevalence rates of smoking in subjects with asthma have frequently been reported as similar to those in the general population; however, available data are not up-to-date. There is only limited and somewhat conflicting information on the long-term effects of smoking on health outcomes among population-based cohorts of subjects with asthma. We aimed to investigate changes in smoking habits and their effects on forced expiratory volume in 1 s (FEV(1)) in subjects with asthma in comparison with the rest of the population, focusing on the healthy smoker effect. METHODS: We studied 9,092 subjects without asthma and 1,045 with asthma at baseline who participated in both the European Community Respiratory Health Survey I (20-44 years old in 1991-1993) and II (1999-2002). RESULTS: At follow-up, smoking was significantly less frequent among subjects with asthma than in the rest of the population (26 vs. 31%; p < 0.001). Subjects with asthma who were already ex-smokers at the beginning of the follow-up in the 1990 s had the highest mean asthma score (number of reported asthma-like symptoms, range 0-5), probably as a result of the healthy smoker effect (2.80 vs. 2.44 in never smokers, 2.19 in quitters and 2.24 in smokers; p < 0.001). The influence of smoking on FEV(1) decline did not depend on asthma status. Smokers had the highest proportion of subjects with chronic cough/phlegm (p < 0.01). CONCLUSION: One out of 4 subjects with asthma continues smoking and reports significantly more chronic cough and phlegm than never smokers and ex-smokers. This stresses the importance of smoking cessation in all patients with asthma, even in those with less severe asthma.