The objectives of this study were to establish dose-response and blood concentration-response relationships for robenacoxib, a novel nonsteroidal anti-inflammatory drug with selectivity for inhibition of the cyclooxygenase (COX)-2 isoenzyme, in a canine model of synovitis. Acute synovitis of the stifle joint was induced by intra-articular injection of sodium urate crystals. Robenacoxib (0.25, 0.5, 1.0, 2.0 and 4.0 mg/kg), placebo and meloxicam (0.2 mg/kg) were administered subcutaneously (s.c.) 3 h after the urate crystals. Pharmacodynamic endpoints included data from forceplate analyses, clinical orthopaedic examinations and time course of inhibition of COX-1 and COX-2 in ex vivo whole blood assays. Blood was collected for pharmacokinetics. Robenacoxib produced dose-related improvement in weight-bearing, pain and swelling as assessed objectively by forceplate analysis (estimated ED(50) was 1.23 mg/kg for z peak force) and subjectively by clinical orthopaedic assessments. The analgesic and anti-inflammatory effects of robenacoxib were significantly superior to placebo (0.25-4 mg/kg robenacoxib) and were non-inferior to meloxicam (0.5-4 mg/kg robenacoxib). All dosages of robenacoxib produced significant dose-related inhibition of COX-2 (estimated ED(50) was 0.52 mg/kg) but no inhibition of COX-1. At a dosage of 1-2 mg/kg administered s.c., robenacoxib should be at least as effective as 0.2 mg/kg of meloxicam in suppressing acute joint pain and inflammation in dogs.