The subcutaneous sponge implant model of acute inflammation in the rat has been evaluated as a suitable test system for evaluating the potential anti-inflammatory efficacy of 5-lipoxygenase inhibitors. The inflammatory parameters measured were exudate volume and leukocyte recruitment. Specific radioimmunoassays were used to measure (1) 5-lipoxygenase (LPO) and cyclo-oxygenase (CO) activity in exudate leukocytes stimulated ex vivo with A23187, and (2) the LTB4 and PGE2 content of inflammatory exudate. The NSAIDs flurbiprofen and indomethacin inhibited cell recruitment, exudate volume and CO activity with ED50S of approximately 1 mg per kg p.o. but failed to inhibit LPO activity at 10 mg per kg p.o. Nafazatrom (Bayer 6575), quercetin and NDGA, which inhibit LPO activity in vitro, were inactive against all parameters when dosed at 100 mg per kg p.o. The "mixed inhibitors" BW755C and phenidone were approximately equipotent inhibitors of LPO activity but BW755C was 10 times more potent than phenidone against CO activity. BW755C was also greater than 10 times more potent at inhibiting cell recruitment and exudate volume than phenidone suggesting that the anti-inflammatory efficacy of the mixed inhibitors reflect their potency against CO rather than LPO activity. Time course studies demonstrated that the inhibitor effects of BW755C and phenidone on leukocyte recruitment reflected a reduction in the PGE2 but not the LTB4 content of the inflammatory exudate. Polyester sponges soaked in high concentrations of LTB4 caused only a modest (2-fold) increase in leukocyte recruitment whilst physiological levels were inactive. The results taken together suggest that CO products make a major contribution to leukocyte recruitment in this model whilst the LPO product LTB4 has little role. This model therefore is of little value for evaluating the anti-inflammatory efficacy of 5-lipoxygenase inhibitors. Moreover, the rat would appear to be unsuitable for evaluating the role of LTB4 in acute inflammation.