Clearing the air. Smoking and incident asthma in adults.

Abstract

Cigarette smoke exposure is associated with numerous mechanisms thought to play an important role in the pathogenesis of asthma, including increased Th2 pathway activation and allergic sensitization (1–4), as well as airway hyperreactivity (5, 6). Many studies have identified an association between cigarette smoke and asthma exacerbations in adults (7, 8); however, whether cigarette smoke is linked to the development of new adult-onset asthma is less clear. Several cross-sectional studies have demonstrated that patients with adult-onset asthma were more likely to be smokers (9–11), but interpreting causality from these studies is difficult, as a temporal relationship between cigarette smoke and asthma development cannot be established. Several cohort studies have suggested an association between cigarette smoking and incident asthma (12, 13), although other cohort studies have found no association (14, 15) or even a lower incidence of asthma in active smokers (16). This latter finding may be a result of the “healthy smoker effect,” in which subjects who are more prone to the negative effects of smoking quit at a higher rate. On the basis of these prior studies, the 2014 US Surgeon General’s Report on the Health Effects of Smoking concluded that the evidence was “suggestive but not sufficient to infer a causal relationship between active smoking and the incidence of asthma in adults” (17). Less is known about the effects of passive smoking on the incidence of asthma. Although several studies have suggested an association between passive smoking and new adult-onset asthma (18–21), many were limited by recall bias, leaving the question of the association between passive smoking and new adult-onset asthma not fully answered. In this issue of the Journal, Coogan and colleagues (pp. 168–176) report their analysis of the relationship between active and passive cigarette smoking and the incidence of adult-onset asthma in the Black Women’s Health Study, a prospective cohort of African American women that was established in 1995 (22). From 1995 to 2011, 46,182 women without baseline asthma were followed for an average of 14.7 years. Subjects enrolled in the study were mailed biennial health questionnaires assessing a variety of exposures, as well as adult-onset asthma, which was defined as a self-reported first diagnosis of asthma with concurrent use of asthma medications. Information on smoking status was obtained by self-report and included a one-time assessment of secondhand smoke exposure in 1997. The authors found that both active and passive smoking were associated with an increased incidence of adult-onset asthma in a multivariable model that adjusted for numerous potential confounders including age, body mass index (BMI), education, parental history of asthma, female hormone usage, household income, health insurance, and access to a physician. Among both former and active smokers, there appeared to be a dose–response relationship between asthma and smoking, with an increased incidence of adult-onset asthma with a higher number of pack-years smoked. Interestingly, when the authors stratified subjects by age and BMI, the associations between active and passive smoking and incident adult-onset asthma were more pronounced in those with a BMI lower than 30 kg/m2 and age 50 years or older (populations that are typically thought to be at lower risk for asthma development in adulthood). This study has several strengths that make it a significant addition to the existing body of literature on cigarette smoking and incident adult-onset asthma. First, this is one of the largest longitudinal studies examining this relationship and boasts both an impressive duration of follow-up as well as a strong retention of subjects compared with prior studies. In addition, subjects were assessed frequently throughout the duration of the study, allowing for a more accurate assessment of changes in smoking habits. Furthermore, this analysis addressed several limitations of prior studies by excluding subjects with asthma at baseline, assessing exclusively for a diagnosis of asthma rather than just symptoms of wheezing, and adjusting for a number of potential confounders such as socioeconomic status and family history of asthma in multivariable models. Although this study represents an important contribution to the literature on the association between smoking and asthma, some areas of uncertainty remain. First, distinguishing between chronic obstructive pulmonary disease and asthma in large epidemiologic studies using self-report without accompanying pulmonary function tests or physician diagnosis remains challenging. In this study, the diagnosis of asthma was captured by self-report. The authors performed several sensitivity analyses that used stricter definitions of asthma, for instance, including only subjects who reported the use of a preventative asthma medication, which they were also required to name. Despite a smaller sample size, the association between both passive and active smoking and incident adult-onset asthma was maintained with this stricter case definition. Likewise, in a subset of patients, the study team contacted subjects’ physicians, who confirmed a diagnosis of asthma 91% of the time. Although these measures lend some confidence to the diagnosis of asthma, it remains possible that some of the patients who self-identified as having asthma may actually have had chronic obstructive pulmonary disease. This concern is strengthened by the finding that the hazard ratio between smoking and incident asthma was highest in subjects aged 50 years or older, who are at higher risk for the development of chronic obstructive pulmonary disease. However, given the size of the cohort, the strength of the association, and the findings of the sensitivity analyses, it seems reasonable to conclude that active smoking is associated with increased adult-onset asthma, even in the presence of some misclassification. A second important area for future study involves the relationship between passive smoking and incident asthma. Passive smoking, which was defined in the present study as being in the same room as a smoker for an hour per day for 12 consecutive months, was assessed at only one point. As the authors acknowledge, this assessment is relatively crude, may miss biologically significant lower levels of exposure, and is prone to the same recall biases as prior studies exploring this relationship. More significantly, clean indoor air laws and public smoking bans that went into effect after 1997, when passive smoking was assessed in this study, likely altered the intensity of passive smoking exposure subjects experienced throughout the 15 years of follow-up. As the authors acknowledge, this misclassification would likely have biased the relationship of passive smoking and incident asthma toward the null, underestimating the true effect size. Thus, although an association between passive smoking and incident adult-onset asthma was found in this study, further studies with more frequent and precise assessments of passive smoking are needed. Understanding and quantifying the association between cigarette smoke exposure and incident asthma has significant implications for public health and beyond. First, defining the strength of this association has a clear and significant effect on tobacco product regulation and on assessment of the global burden of disease and health costs related to tobacco products. Furthermore, these analyses have the potential to provide insight into asthma endotypes. Prior studies have reported that smoking negatively affects the response to inhaled corticosteroids in asthmatics (23) and that smoking-related asthma may be a distinct asthma phenotype (24). In this study, Coogan and colleagues provide evidence that the association between smoking and asthma varies by age and BMI, emphasizing the intriguing possibility that smoking may be linked to specific asthma endotypes. These findings should spur additional mechanistic and epidemiologic studies of smoking-related asthma to ultimately develop effective therapeutic and regulatory strategies for this subset of patients.