This study determines objectively the extent of nonsteroidal anti-inflammatory drug (NSAID) use in upper gastrointestinal (GI) mucosal acid-peptic diseases by supplementing the conventional interview with two tests of current aspirin (ASA) use--high-performance liquid chromatography (HPLC) for the presence of salicylates in serum and platelet cyclooxygenase activity, which detects ASA use within 5 days of testing. Of 186 consecutive patients undergoing upper endoscopy, 62% of 55 patients with esophagitis had evidence of current NSAID use, vs. 26% of 42 control patients with normal endoscopy (p less than 0.001), 12% of 17 patients with recently healed peptic ulcer (p less than 0.001), and 36% of 25 patients who had an active peptic ulcer (p less than 0.05), five of whom had concomitant esophagitis. Another 52 patients were ineligible for this analysis. Testing for platelet cyclooxygenase activity uncovered 26% more ASA users than history alone. In considering age, sex, smoking and drinking habits, arthritis, and ASA use by logistic regression, ASA use was the only factor contributing to esophagitis; ASA could not be further associated with severity, stricture or symptoms, however. In these patients, 95% of NSAID use was chronic, and 84% of that was ASA. These data show a previously unreported, strong association of ASA use with esophagitis, which suggests that ASA may be a significant factor in the resistance of esophagitis to current therapies as well as the frequently rapid relapse after therapy is withdrawn.