We investigated the antinociceptive efficacy of systemic and centrally injected oxycodone on thermal hyperalgesia in streptozotocin-induced diabetic mice. The antinociceptive response was assessed by recording the latency in the tail-flick test using the radiant heat from a 50-W projection bulb on the tail. The tail-flick latency in diabetic mice was significantly shorter than that in non-diabetic mice. Oral (p.o.) and i.t., but not i.c.v., administration of oxycodone prolonged the tail-flick latency in diabetic mice to a level that was considerably longer than the baseline latency in non-diabetic mice. However, morphine did not significantly inhibit the tail-flick response in diabetic mice. The antinociceptive effect of either p.o. or i.t. oxycodone in non-diabetic mice, but not in diabetic mice, was antagonized by pretreatment with a selective mu-opioid receptor antagonist, beta-funaltrexamine. In non-diabetic mice, pretreatment with a selective kappa-opioid receptor antagonist, nor-binaltorphimine, had no effect on the peak antinociceptive effect of either p.o. or i.t. oxycodone at 30 min after administration, however, it slightly but significantly reduced oxycodone-induced antinociception at 60 and 90 min after administration. On the other hand, pretreatment with nor-binaltorphimine practically abolished the antinociceptive effects of both p.o.- and i.t.-administered oxycodone in diabetic mice. Naltrindole, a selective delta-opioid receptor antagonist, had no effects on the antinociceptive effect of oxycodone in either non-diabetic or diabetic mice. These results suggest that the antinociceptive effects of oxycodone may be mediated by spinal kappa-opioid receptors in diabetic mice, whereas it may interact primarily with supraspinal and spinal mu-opioid receptors in non-diabetic mice.