We utilized a canine renal transplant model to estimate the first-pass extraction of mizoribine (MZB) during renal artery infusion and to compare the efficacy and toxicity of continuous intraarterial (i.a.) versus intravenous (i.v.) MZB delivery, with and without a background of oral cyclosporine. Five autotransplanted mongrel dogs with programmable, implantable pump/catheter systems were given MZB by both i.v. bolus (5 mg/ kg) and i.a. infusion (5.0 mg/kg/day). Mean ± SD elimination half-life was 3.02±0.81 hr, and the transplanted kidney removed as much as 47–59% (mean 56%) of locally infused MZB. With increasing local and systemic MZB delivery in a single autografted dog undergoing both i.a. and i.v. pump/catheter placement, renal extraction decreased from at least 47% (5.0 mg/kg/day) to 33% (7.5 mg/kg/day), and finally to 18% (10.0 mg/kg/day). A dose of 3.0 mg/kg/day MZB did not significantly prolong survival of renal allograft recipients over that of untreated controls (median survival time [MST]=8 days) when administered either locally (MST=9 days) or systemically (MST=12 days). All dogs receiving 4.0 mg/kg/ day MZB i.a. died from rejection, and a survival advantage was still not realized (MST=7 days). In contrast, 4.0 mg/kg/day i.v. prolonged survival over controls (MST=14 days; P=0.03) but not when directly compared with the i.a. group (P=0.30), and produced death from severe debility in five of seven animals with significantly higher mean systemic MZB levels (P=0.02). Four of six dogs receiving 5.0 mg/kg/day MZB i.a. (MST=14 days) and two of four dogs receiving 5.0 mg/kg/day i.v. (MST=14 days) died from severe debility, though survival in both groups was prolonged over control values (P=0.01 and P=0.05, respectively). Coad-ministration of a subtherapeutic dose of oral CsA (5 mg/ kg/day) significantly prolonged the overall survival of dogs receiving MZB 4.0 mg/kg/day i.a. (MST=23; P=0.01) but not i.v. (MST=11; P=1.00), so that a significant difference in overall survival between the combined MZB i.a. + CSA and MZB i.v. + CSA groups was now realized in favor of the former (P=0.04). We conclude that at local doses required to achieve immunosuppression, the transplanted kidney was not able to extract enough MZB to prevent death from systemic toxicity, presumably as a result of saturation of renal elimination mechanisms, so that an overall survival benefit was not realized. However, combination of oral CsA with i.a., but not i.v., MZB infusion conferred a survival advantage with lower systemic MZB concentrations, suggesting mediation via a local immunosuppressive effect.