Comparing Risk Factors of Atopic and Non-Atopic Asthma May Lead to Misleading Conclusions

ISEE-0294 Background and Objective: It has become commonplace to analyze separately atopic and non-atopic asthma in studies on genetic, environmental, and immunological characteristics of asthma. Atopic asthma is in most studies defined simply based on presence of both asthma and positive skin prick test or specific IgE antibodies. We here illustrate the problems introduced by using atopic and non-atopic asthma as endpoint in analyses exploring characteristics associated with asthma/atopy. Methods: Analyses of the 31 year follow-up of the Northern Finland 1966 birth cohort (n = 5417). Asthma was defined as doctor-diagnosed asthma ever and atopy as any skin prick test reaction >3 mm to common allergens at age 31. Atopic and non-atopic asthma were defined as asthma with and without atopy, respectively. Results: The prevalence of asthma was 7.8% and of atopy 30%. Having a cat in childhood was associated with decreased risk of atopy at age 31 years (OR 0.59, 95% CI 0.52-0.66). This created a spurious protective association with atopic asthma (OR 0.69, 95% CI 0.53-0.88) and an opposite association with non-atopic asthma (OR 1.27, 95% CI 0.93-1.75). There was no association between cat ownership and asthma, when analysis was stratified by atopy or in a multivariate model adjusted for atopy (OR 1.05, 95% CI 0.85-1.29). Similar spurious associations were seen for gender. Findings were confirmed with simulated datasets. Conclusion: If a characteristic is associated with atopy, this will produce a similar association with atopic asthma and a reverse association with non-atopic asthma, even if the characteristic has no association with asthma. Therefore, before better methods become available to define phenotypes of asthma, asthma and atopy should be analyzed as separate end points.