© 2002 American Thoracic Society
Smoking and adult asthmaA healthy smoker effect?To the Editor :I read with interest the article by Basagaña and colleagues that examined the incidence of asthma among adults in Spain (1). This article is important because there are few studies of adult-onset asthma that ascertain incident, rather than prevalent, asthma cases (2). As the authors correctly point out, studies focusing on risk factors for prevalent asthma cannot distinguish whether an exposure causes new cases of asthma or adversely affects persons with pre-existing disease, resulting in a longer duration of symptoms. As such, the present study makes a significant contribution to the literature. Unfortunately, the authors missed an opportunity to advance our understanding of cigarette smoking as a possible cause of adult asthma. This is because they examined the impact of current smoking at baseline interview on the risk of subsequent asthma, which may have biased their results. In Table 3 of their article, there is no statistical relationship between current smoking and asthma risk. In fact, the odds ratio is less than 1.0, which could erroneously suggest a protective effect of smoking. These observations may be explained by a selection bias termed the "healthy smoker effect" (3). Because most people start smoking during adolescence, a substantial proportion will have stopped by young or middle adulthood. Basagaña and colleagues observed this phenomenon: at the baseline interview, 276 of the 1,264 ever smokers had quit. Many of those who quit smoking probably did so because they experienced respiratory symptoms, such as cough or wheezing. As a consequence, the pool of current smokers at baseline interview is probably enriched for persons who have experienced fewer smoking-related respiratory symptoms. In other words, the current smokers are "healthier" from the respiratory standpoint and are less likely to have, or develop, asthma. This selection process can bias the relationship between smoking and asthma, masking a true causal effect of smoking. In fact, this bias can even make smoking appear to reduce the risk of asthma, as recognized by previous investigators (4). To properly examine the effect of cigarette smoking on the incidence of adult asthma, the authors needed to study lifetime smoking rather than only current smoking. By reconstructing previous smoking history, they could have provided a more accurate understanding of smoking and asthma.
University of California, San Francisco San Francisco, California REFERENCES
Our thanks to Dr. Mark D. Eisner for his thoughtful letter regarding our article on asthma incidence and smoking (1). We agree that this is an unexplored field. The role of smoking (or smoking cessation) on the incidence rather than exacerbation or attack of asthma is difficult to assess in an observational study where smoking history is assessed retrospectively because of the difficulty of disentangling the time-pattern between onset/cessation of smoking from the onset of asthma. In addition, in our study, we did not collect changes in smoking during the period 1991–1999. In 1999, we only collected cross-sectional information. We are currently completing the follow-up of the 15-country study European Community Respiratory Health Survey (ECRHS-II), and are collecting a complete smoking history. In this study, we will be able to better evaluate the complex association between smoking and asthma, including potential interactions with atopy and other concomitant exposures. Finally, can smokers be healthy? Perhaps we should stop referring to the "healthy smoker effect" and find a different name for this type of bias. Maybe, "smoking cessation effect"?
Respiratory and Environmental Health Research Unit, IMIM Barcelona, Spain REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
