The use of NSAIDs is strongly associated with peptic ulceration. The inhibition of prostaglandin synthesis with the consequent increase of gastric acidity is considered a possible mechanism. Therefore we decided to assess the effect of one-month treatment with NSAIDs on the circadian gastric pH of rheumatoid arthritis (RA) patients. We studied 11 consecutive patients (one man and 10 women, median age 55, range 26-72 years) with confirmed RA. None was H. pylori positive. A 24-hr gastric pH recording was performed both in basal conditions and after one-month treatment with either indomethacin 150 mg/day (eight cases) or ketoprofen 300 mg/day (three cases). Only the 10 female patients were eligible for final analysis, and six matched healthy subjects not taking NSAIDs were used as control group. The number of 24-hr pH readings for various pH thresholds was calculated for both populations. The highest acid levels (pH < 3.0) did not differ between the two pH profiles of the control group (7440 vs 7391, P = NS), while they predominated after the one-month NSAID treatment (10,339 vs 11,440, P < 0.001) in RA patients. These findings show that there is an increased gastric acidity after one-month of treatment with NSAIDs in female patients with RA of recent onset. This may sustain the rationale of using antisecretory agents to prevent gastroduodenal ulcerations in these patients.