Randomised clinical trial: gastrointestinal events in arthritis patients treated with celecoxib, ibuprofen or naproxen in the PRECISION trial

AIM: To evaluate GI safety of celecoxib compared with 2 nonselective (ns) NSAIDs, as a secondary objective of a large trial examining multiorgan safety. METHODS: This randomised, double-blind controlled trial analysed 24 081 patients. Osteoarthritis or rheumatoid arthritis patients, needing ongoing NSAID treatment, were randomised to receive celecoxib 100-200 mg b.d., ibuprofen 600-800 mg t.d.s. or naproxen 375-500 mg b.d. plus esomeprazole, and low-dose aspirin or corticosteroids if already prescribed. Clinically significant GI events (CSGIE-bleeding, obstruction, perforation events from stomach downwards or symptomatic ulcers) and iron deficiency anaemia (IDA) were adjudicated blindly. RESULTS: Mean treatment and follow-up durations were 20.3 and 34.1 months. While on treatment or 30 days after, CSGIE occurred in 0.34%, 0.74% and 0.66% taking celecoxib, ibuprofen and naproxen. Hazard ratios (HR) were 0.43 (95% CI 0.27-0.68, P = 0.0003) celecoxib vs ibuprofen and 0.51 (0.32-0.81, P = 0.004) vs naproxen. There was also less IDA on celecoxib: HR 0.43 (0.27-0.68, P = 0.0003) vs ibuprofen; 0.40 (0.25-0.62, P < 0.0001) vs naproxen. Even taken with low-dose aspirin, fewer CSGIE occurred on celecoxib than ibuprofen (HR 0.52 [0.29-0.94], P = 0.03), and less IDA vs naproxen (0.42 [0.23-0.77, P = 0.005]). Corticosteroid use increased total GI events and CSGIE. H. pylori serological status had no influence. CONCLUSIONS: Arthritis patients taking NSAIDs plus esomeprazole have infrequent clinically significant gastrointestinal events. Co-prescribed with esomeprazole, celecoxib has better overall GI safety than ibuprofen or naproxen at these doses, despite treatment with low-dose aspirin or corticosteroids.