OBJECTIVE To describe the pharmacokinetics of intra-articularly (IA) administered morphine. METHODS Experimental randomized, cross-over study. METHODS Eight adult healthy mixed breed horses aged 6.5 +/- 2.3 (mean +/- SD) years and weighing 535 +/- 86 kg. METHODS Unilateral radiocarpal synovitis was induced by IA injection of 3 microg lipopolysaccharide (LPS) on two occasions (right and left radiocarpal joint, respectively) separated by a 3-week wash-out period. Treatments were administered 4 hours post-LPS-injection: Treatment IA; preservative free morphine IA (0.05 mg kg(-1)) plus saline intravenous (IV) and treatment IV; saline IA plus preservative free morphine IV (0.05 mg kg(-1)). Concentrations of morphine, morphine-3-glucuronide and morphine-6-glucuronide (M6G) were determined repeatedly in serum and synovial fluid (SF) by high-performance liquid chromatography mass spectrometry, at 2 and 4 hours and then at 4 hours intervals until 28 hours post-treatment. RESULTS Injection of LPS elicited a marked and comparable synovitis in all LPS-injected radiocarpal joints. IA administered morphine was detectable in SF of all eight joints 24 hours post-treatment and in 6/8 joints 28 hours post-treatment. The terminal half-life of morphine in SF was estimated to be 2.6 hours. IA administration of morphine resulted in mean serum concentrations of morphine below 5 ng mL(-1) from 2 to 28 hours after treatment. CONCLUSIONS Intra-articularly administered morphine remained within the joint for at least 24 hours. At the same time only very low serum concentrations of morphine and M6G were detected. The present results suggest that IA morphine at 0.05 mg kg(-1) may be used for IA analgesia lasting at least 24 hours and give strong support to the theory that previously observed analgesic and anti-inflammatory effects of IA morphine in horses are most likely to be mediated peripherally.