Meta-analysis of aspirin use and risk of lung cancer shows notable results

Abstract

Aspirin is a promising agent for chemoprevention of lung cancer. We assessed the association of aspirin use and the development of lung cancer, with a focus on heterogeneity between studies. Databases were searched for relevant studies until September 2014. Studies evaluating the relationship of aspirin use and incidence of lung cancer were considered. Relative risks (RR) were extracted and a pooled estimate was calculated. Heterogeneity was assessed by the I measure, random-effects models, and finite-mixture models. Sources of heterogeneity were investigated using a meta-regression. A decreased risk of lung cancer was found including 20 studies [RR=0.87, 95% confidence interval (CI): 0.79-0.95] on the basis of a random-effects model. Strong heterogeneity was observed (τ=0.0258, I=74.4%). As a result, two subpopulations of studies were identified on the basis of a mixture model. The first subpopulation (42%) has an average RR of 0.64. The remaining subpopulation (58%) shows an RR of 1.04. Different results were found for case-control (RR=0.74, 95% CI: 0.60-0.90) and cohort studies (RR=0.99, 95% CI: 0.93-1.06) in a stratified analysis. In a subgroup analysis, use of aspirin was associated with a decreased risk of non-small-cell lung cancer in case-control studies (RR=0.74; 95% CI: 0.58-0.94). At first glance, our meta-analysis shows an average protective effect. A second glance indicates that there is strong heterogeneity. This leads to a subpopulation with considerable benefit and another subpopulation with no benefit. For further investigations, it is important to identify populations that benefit from aspirin use.