Artigo recomendado: Coffee Drinking and Mortality

Scott K. Aberegg, M.D., M.P.H.

n engl j med 367;6 nejm.org august 9, 2012

To the Editor: Freedman et al. (May 17 issue) report the results of a widely publicized study of the association between coffee consumption and mortality. Several limitations and alternative explanations of their findings deserve consideration. The biologic plausibility of a reduction in mortality associated with coffee consumption is inadequately explicated. This is especially noteworthy because coffee consumption appeared to reduce the rate of death from accidents and injuries. In essence, this is a positive result on a negative control embedded within the study, and it undermines the study’s main findings. People frequently drink coffee at work (the proverbial “coffee break”), and the findings of this study could be due to a “healthy worker” effect not captured by other covariates, such as educational attainment and health status. The inclusion of relevant covariates is fraught with difficulty when a plausible mechanism supporting the main study findings is not apparent. The simplest explanation for the aggregate findings seems to be the most likely — in this case, residual confounding from unmeasured behaviors associated with coffee consumption that protect against everything, including accidents and injuries.

Caso queira, deixe seu e-mail nos comentários abaixo desta postagem ou solicite pelo e-mail anestesiasegura@sma.com.br e lhe enviaremos o artigo completo.