We studied the anti-inflammatory activity of meloxicam on rat carrageenin-induced pleurisy and its toxicity for rat gastric mucosa, relative to its in vitro inhibitory potency against partially purified cyclooxygenase (COX)-1 and COX-2 preparations in order to clarify the pharmacological profile of the compound as an anti-inflammatory agent. In rat carrageenin-induced pleurisy, the plasma exudation rate peaked at 5 h, at which time COX-2 was detectable in cells from the pleural exudate. Meloxicam and piroxicam (1 and 3 mg/kg) and NS-398 (3 mg/kg) showed almost equal anti-inflammatory potency against 5-hour pleurisy. A single oral administration of the compounds caused a dose-dependent increase in the number of rats with gastric mucosal erosion. The ED50 value for meloxicam (5.92 mg/kg) was significantly higher than that for piroxicam (1.76 mg/kg), indicating that meloxicam is safer. Indometacin showed intermediate safety (2.59 mg/kg). In in vitro experiments, indometacin inhibited COX-1 about 1.7 times more potently than COX-2. NS-398 inhibited COX-2 with an IC50 of 0.32 microM, but never affected COX-1 activity, even at 100 microM. In the same assay system, meloxicam inhibited COX-2 about 12 times more selectively than COX-1. Piroxicam, however, inhibited both isoforms almost equally. These results indicate that meloxicam is a potent anti-inflammatory agent with low gastric toxicity. One reason for its in vivo pharmacological profile may be related to its relative selectivity for COX-2 over COX-1. Thus, meloxicam may belong to a group of COX-2 selective anti-inflammatory agents with a better safety profile than conventional COX-1 and COX-2 nonselective anti-inflammatory agents.