Selective cyclooxygenase (COX)-2 inhibitors produce less gastric damage than conventional nonsteroidal anti-inflammatory drugs (NSAIDs), suggesting that NSAIDs cause damage by inhibiting COX-1. We tested this hypothesis in rats by using a selective COX-1 inhibitor (SC-560).

The effects of SC-560, celecoxib (selective COX-2 inhibitor), or a combination of both inhibitors on gastric damage and prostaglandin synthesis were determined. Selectivity of the drugs for COX-1 vs. COX-2 was assessed in the carrageenan-airpouch model. A COX-1-preferential inhibitor, ketorolac, was also evaluated. The effects of these inhibitors on leukocyte adherence to vascular endothelium and on gastric blood flow were assessed.

SC-560 markedly reduced gastric prostaglandin synthesis and platelet COX-1 activity, but spared COX-2 and did not cause gastric damage. Celecoxib did not affect gastric prostaglandin E₂ synthesis and did not cause gastric damage. However, the combination of SC-560 and celecoxib invariably caused hemorrhagic erosion formation, comparable to that seen with indomethacin. Ketorolac caused damage only at doses that inhibited both COX isoforms, or when given with a COX-2 inhibitor. Celecoxib, but not SC-560, significantly increased leukocyte adherence, whereas SC-560, but not celecoxib, reduced gastric blood flow.

Inhibition of both COX-1 and COX-2 is required for NSAID-induced gastric injury in the rat.