BACKGROUND: The pharmacology of oxycodone is poorly understood despite its growing clinical use. The discrepancy between its good clinical effectiveness after systemic administration and the loss of potency after spinal administration led the authors to study the pharmacodynamic effects of oxycodone and its metabolites using in vivo and in vitro models in rats. METHODS: Male Sprague-Dawley rats were used in hot-plate, tail-flick, and paw-pressure tests to study the antinociceptive properties of morphine, oxycodone, and its metabolites oxymorphone and noroxycodone. Mu-opioid receptor agonist-stimulated GTPgamma[S] autoradiography was used to study G-protein activation induced by morphine, oxycodone, and oxymorphone in the rat brain and spinal cord. Spontaneous locomotor activity was measured to assess possible sedation or motor dysfunction. Naloxone and the selective kappa-opioid receptor antagonist nor-binaltorphimine were used to study the opioid receptor selectivity of the drugs. RESULTS: Oxycodone showed lower efficacy and potency to stimulate GTPgamma[S] binding in the spinal cord and periaqueductal gray compared with morphine and oxymorphone. This could relate to the fact that oxycodone produced only weak naloxone-reversible antinociception after intrathecal administration. It also suggests that the metabolites may have a role in oxycodone-induced analgesia in rats. Intrathecal oxymorphone produced strong long-lasting antinociception, whereas noroxycodone produced antinociception with very high doses only. Subcutaneous administration of oxycodone and oxymorphone produced thermal and mechanical antinociception that was reversed by naloxone but not by nor-binaltorphimine. Oxymorphone was more potent than oxycodone, particularly in the hot-plate and paw-pressure tests. CONCLUSIONS: The low intrathecal potency of oxycodone in rats seems be related to its low efficacy and potency to stimulate mu-opioid receptor activation in the spinal cord.