Aspirin Use and Gastric Adenocarcinoma—Reply

Abstract

We appreciate Lai and colleagues’ interest in our study in which we reported associations between aspirin use and gastric adenocarcinoma in two large prospective cohort studies of U.S. men and women (1). Lai and colleagues calculated that comparing aspirin users with nonusers, the absolute rate difference was 2/100,000 person-years with a number-needed-to-treat (NNT) of 50,000 in women. Perhaps a more clinically interpretable metric of absolute risk would be based on persons at risk rather than person-years using an estimate calculated according to the mean follow-up time (30.6 years) of cohort participants. We would then estimate the risk of gastric cancer was 0.0018 [125 cases/(2,097,414 person-years/30.6 years)] among nonregular aspirin users, and 0.0012 [51 cases/(1,258,158 person-years/30.6 years)] among regular users. This translates to an NNT of 1,714, suggesting one case of gastric cancer could potentially be prevented by treating 1,714 women with aspirin regularly. Nonetheless, we concur that public health recommendations should weigh potential complications associated with aspirin use such as upper gastrointestinal bleeding against the absolute benefits.In the case that exposure status is determined after the start of follow-up of a cohort, immortal time bias can be introduced if the person-time in the interval between the start of follow-up and when exposure data are collected is misattributed or ignored completely (2). In our study, person-time was accrued only from the return of a baseline questionnaire when the participants were first queried about aspirin use, to the date of gastric adenocarcinoma diagnosis, death, or the end of follow-up, whichever came first. We then updated aspirin use with each biennial questionnaire cycle as a time-varying variable. Thus, our results were not influenced by immortal time bias.Aspirin has been found to inhibit the growth of Helicobacter pylori (H. pylori) in a dose–response manner and enhance its susceptibility to antimicrobial agents (3). Consistently, we observed that aspirin use was associated with a lower risk of H. pylori infection among women. However, the hypothesis of inhibitory effects of aspirin on H. pylori and gastric cancer is speculative, and more evidence is needed to understand potential mechanisms. In addition, aspirin and salicylic acid, its primary bioactive metabolite, can influence other bacteria through both antimicrobial and regulatory effects, and manipulation of microbiota has been proposed as a mechanism for aspirin chemoprevention (4).A.T. Chan reports personal fees from Bayer Pharma AG and Boehringer Ingelheim, grants and personal fees from Pfizer Inc., and grants from Zoe, Ltd. outside the submitted work. No disclosures were reported by the other author.See the original Letter to the Editor, p. 553