Aspirin causes peptic ulcers predominately by reducing gastric mucosal cyclooxygenase (COX) activity and prostaglandin synthesis. Because aspirin circulates for only a few hours, we hypothesized that aspirin's inhibitory effect on gastric COX activity must be prolonged. We performed a placebo-controlled experiment in healthy humans to determine the duration of inhibition of aspirin on gastric mucosal COX activity (PGE₂ and PGF₂ synthesis rates). Recovery of gastric COX activity after stopping aspirin was slow and linear. Seventy-two hours after 325-mg aspirin, gastric COX activity was still reduced by 57% (P < 0.001). Duration of inhibition of gastric COX activity was estimated to be 7-8 days after 325-mg aspirin and 5 days after 81-mg aspirin. Recovery of gastric prostaglandin synthesis after 325-mg but not after 81-mg aspirin occurred at slower rates in subjects with Helicobacter pylori-associated gastritis than in those with normal histology. In conclusion, aspirin inhibits gastric COX activity for much longer than predicted from its pharmacokinetic profile, explaining why aspirin at widely spaced intervals is ulcerogenic.