The purpose of this study is to compare the effect of Famotidine plus a COX-2 inhibitor (celecoxib) with Famotidine plus dologesics in ulcer healing in arthritis patients.

For many years the integrity of the stomach mucosal barrier is thought to be maintained by mucosal prostaglandins (PG) synthesized by COX-1. However, the notion that COX-1 protects the stomach and COX-2 induces inflammation may be over-simplistic. In animal studies, COX-2, but not COX-1, is expressed in experimental gastric ulcer. Inhibition of COX-2 delays ulcer healing, indicating that PG derived from COX-2 contributes to restoring the mucosal barrier [1]. Whether this animal observation can be generalized to the human stomach is unknown. To date the biological functions of COX-1 and COX-2 in the healing of human gastric ulcer healing is unclear. Unlike experimental ulcers that only express COX-2, recently we have shown that both COX-1 and COX-2 are up-regulated in human gastric ulcers [2]. Furthermore, our preliminary results suggest that inhibition of COX-2 alone may not lead to a clinically significant delay in ulcer healing (refer to progress report). These observations suggest that peptic ulcer healing is more complex in the human stomach - both COX isoforms may be involved in the healing process. Inhibition of COX-2 alone may have less adverse effect than non-selective inhibition of both COX isoforms in ulcer healing. The current study aims to resolve the functional significance of COX-2 in human gastric ulcer from a biological and clinical perspective.

Inclusion Criteria:

• Gastric ulcers confirmed by endoscopy
• Stop taking NSAIDs for 1 week prior to endoscopy
• Age 18
• H. pylori negative
• Informed written consent

Exclusion Criteria:

• Actively bleeding ulcers
• Ulcers showing dysplasia or malignancy
• Renal failure (serum creatinine >200umol/l)
• Previous gastric surgery
• Moribund or terminal malignancy
• Concomitant use of proton pump inhibitor, misoprostol, aspirin, steroid or anticoagulant