Incidence of gastroduodenal ulcers during treatment with celecoxib or diclofenac: pooled results from three 12‐week trials in Chinese patients with osteoarthritis or rheumatoid arthritis

AIM: To test whether treatment with celecoxib reduces the incidence of gastroduodenal ulcers compared to diclofenac in Asian patients with osteoarthritis (OA) or rheumatoid arthritis (RA) with minimal significant risk factors. METHODS: Patients with a clinical diagnosis of OA or RA of at least 3 months were randomized to 12 weeks of double-blind treatment with celecoxib 100 mg twice daily (n = 440) or diclofenac 50 mg twice daily (n = 440). The primary outcome was the gastric and/or duodenal ulcer rate at endpoint as determined by upper gastrointestinal endoscopy performed during the screening week, and at endpoint. RESULTS: There was no significant difference in the overall incidence of gastroduodenal ulcers at 12-week endpoint for celecoxib compared to diclofenac (2.8% vs. 5.1%; Cochran-Mantel-Haenszel [CMH] chi(2) P = 0.083). However, there was a significantly lower incidence of gastric ulcers on celecoxib versus diclofenac (0.5% vs. 3.6%; CMH chi(2) P = 0.002). Approximately 59% of patients in both treatment groups had no visible gastric lesions at endpoint; and a similar proportion were found to have one or more erosions on celecoxib (n = 85; 21.4%) and diclofenac (N = 91; 23.3%). A survival analysis of time to ulcer was significant for gastric ulcers (log-rank P = 0.004), but not for duodenal ulcers, or for gastroduodenal ulcers combined. Fewer patients reported at least one adverse event on celecoxib compared to diclofenac (42.4% vs. 50.3%; chi(2), 5.52; P = 0.019). CONCLUSIONS: In Asian patients with minimal significant risk factors, treatment with celecoxib was associated with a modest but significantly reduced incidence of gastric ulcers at the end of 12 weeks.