COX-2 inhibition, H. pylori infection and the risk of gastrointestinal complications.

Current data on the gastric safety of cyclooxygenase-2 (COX-2) inhibitors in the presence of H. pylori infection are largely derived from animal experiments and indirect clinical evidence. In animal models of H. pylori gastritis, COX-2 inhibitors suppressed prostaglandin synthesis and aggravated mucosal damage. In the human stomach, COX-1 appears to be the predominant source of prostaglandins despite the fact that COX-2 is upregulated in H. pylori gastritis. There are conflicting data on whether H. pylori alters the risk of ulcer in patients receiving COX-2 inhibitors. Among patients with H. pylori infection, rofecoxib reduced the risk of complicated gastric but not duodenal ulcers as compared to naproxen. The advantage of rofecoxib over naproxen also disappeared in patients with H. pylori infection and prior upper gastrointestinal events. In contrast, pooled data suggested that H. pylori increases the risk of ulcer in patients receiving nonselective nonsteroidal anti-inflammatory drugs but not in patients receiving celecoxib. In rodent gastric ulcers, COX-2 was upregulated in the granulation tissue and ulcer margin. Inhibition of COX-2 delayed healing of experimental gastric ulcer. Limited data showed that COX-2 expression was also increased in human gastric ulcer regardless of the H. pylori status. The functional significance of COX-2 in human gastric ulcer is unknown.