Current therapies for arthritis are unsatisfactory and cause serious side effects and morbidity. It has been postulated that opioid drugs may block inflammatory mediators and attenuate the joint damage in adjuvant arthritis. However, the importance of opioid receptor subtypes involved in inflammation remains to be determined because data are conflicting in this regard. The present investigation was designed to test the effects of both a kappa-agonist, (+/-)U50488H and a kappa-antagonist, MR2266 on the progression of experimental arthritis. To produce adjuvant arthritis, male Lewis rats were inoculated subcutaneously (s.c.) with 0.05 ml of Freund's complete adjuvant (10 mg/ml) into the right hind paw. The kappa-opioid agonist, (+/-)U50488H (20 mg/kg/d s.c.) and the kappa-opioid antagonist, MR2266 (20 mg/kg/d s.c.) were administered for 3 days during the primary inflammatory phase of adjuvant arthritis. There were four treatment groups; group I were non-arthritic controls and received paraffin oil vehicle and opioid injections; group II were arthritic controls and received adjuvant and saline injections; group III received adjuvant and agonist and group IV received adjuvant and antagonist. The progression of adjuvant arthritis from day 0 to 24 was monitored by body weight change, hind limb size (ipsilateral and contralateral) and a total severity score for each clinical observation of gait, coat and limb condition. On day 24 histology and radiography of the contralateral limb was performed. There was less soft-tissue swelling, as judged by time-averaged % change in the volume of the contralateral limb, in both agonist (mean +/- se: 82 +/- 5) and antagonist (77 +/- 4) treated rats compared to untreated arthritic controls (99 +/- 5, p < 0.05). Other clinical measures of severity were not different between untreated and opioid-treated arthritic rats. However, the joint damage as judged by radiography was lower in kappa agonist treated rats (2.6 +/- 0.5, p < 0.05) compared to untreated controls (4.1 +/- 0.5) and antagonist treatment (4.4 +/- 0.5). Microscopic pathological scores were also significantly lower in agonist (2.8 +/- 0.3, p < 0.05) compared to both antagonist treated rats (4.2 +/- 0.1) and vehicle-treated controls (3.6 +/- 0.2). The results of this study show that kappa-opioid receptor agonists but not antagonists attenuate the progression of experimental arthritis. These observations have important implications for the evaluation and use of kappa-opioid agents in the management of arthritis.