The mechanism by which nonsteroidal anti-inflammatory drugs (NSAIDs) suppress gastric mucosal blood flow is not fully understood, although the depletion of mucosal prostaglandin E 2 has been proposed as one possible explanation. We investigated the role of gastric acid on gastric mucosal blood flow in NSAID-treated rats. A rat stomach was mounted in an ex vivo chamber, and gastric mucosal blood flow was measured sequentially in a 5-mm 2 area of the gastric corpus using a scanning laser Doppler perfusion image system. Results showed that diclofenac (5 mg/kg s.c.) and indomethacin (10 mg/kg s.c.) did not affect gastric mucosal blood flow, although both strongly decreased mucosal prostaglandin E 2 when saline was instilled into the gastric chamber. On replacement of the saline in the chamber with 100 mM hydrochloric acid, these drugs caused a decrease in gastric mucosal blood flow levels within 30 min. The specific cyclooxygenase (COX)-2 inhibitors celecoxib (50 mg/kg s.c.) and rofecoxib (25 mg/kg s.c.) did not affect mucosal prostaglandin E 2 level, nor did they decrease gastric mucosal blood flow, even when hydrochloric acid was added to the chamber. Furthermore, measurement of vasoconstrictive factors present in the mucosa showed that endothelin-1 levels increased after administration of diclofenac s.c. in the presence of intragastric hydrochloric acid. This indicates that the presence of mucosal hydrochloric acid plays an important role in the NSAID-induced decrease in gastric mucosal blood flow, while the COX-1-derived basal prostaglandin E 2, which is unlikely to control gastric mucosal blood flow itself, protects microcirculatory systems from mucosal hydrochloric acid.