Fourteen studies, including two hospital-based controls, eight population-based controls, three cohort studies, and one RCT, carried out between 1967 and 2005, were included in the meta-analysis. These included 5,640 case patients; sample sizes ranged from 20 to 11,023.
Overall, there was no statistically significant association between aspirin use and gastric cancer risk (OR 0.80, 95% CI 0.54 to 1.19; I2=97%; 14 studies).
There was no association between aspirin use and gastric cancer risk in the subgroup analysis of two hospital-based population control studies (OR 0.71, 95% CI 0.49 to 1.02; I2=0%) or eight population-based control studies (OR 0.88, 95% CI 0.51 to 1.54; I2 = 98%). However, there was a significant reduction in gastric cancer risk in people using aspirin in the three RCTs and one cohort study (OR 0.72, 95% CI 0.62 to 0.84; I2 = 2%).
There was a significant reduction in gastric cancer risk for the non-cardia gastric cancer (OR 0.62, 95% CI 0.55 to 0.69; seven studies) and in people with Helicobacter pylori infection (OR 0.62, 95% CI 0.42 to 0.90; two studies). Similar results were seen in separate analyses of Caucasian population subgroups.
There was evidence of publication bias in all analyses but, because the number of fail-safes was reported to be large, the authors concluded that publication bias had little effect on summary estimates.