Objective. To clarify the pharmacological advantage of carboplatin-based intraperitoneal chemotherapy using the three-compartment mathematical model. Methods. Eleven consecutive patients in one institution underwent intraperitoneal administration of carboplatin, and 11 consecutive patients in another institution received intravenous administration. Carboplatin (AUC = 6 mg x min/ml) was diluted in 500 ml 5% glucose and administered either as an intraperitoneal bolus infusion or intravenous drip infusion during I h. Patients undergoing intravenous injection also received an infusion of 500 ml 5% glucose to obtain intraperitoneal samples. Intraperitoneal fluid and blood samples were obtained, immediately and 1, 2, 4, 8 12, and 24 h after administration. The mathematical model consisting of a three-compartment model was applied to analyze the pharmacokinetics. The model was created with simultaneous differential equations and was solved by the Runge-Kutta method. Results. The rate constants of platinum diffusion from the peritoneal cavity to serum, serum to peritoneal cavity, serum to peripheral space, peripheral space to serum, and elimination were 0.94 +/- 0.79 (mean +/- SD), 1.28 +/- 2.50, 16.50 +/- 9.26, 0.99 +/- 0.62, and 4.14 +/- 1.45 (h(-1)), respectively. When the theoretical pharmacological concentration of platinum was calculated using this mathematical model, 24-h platinum AUC in the serum was exactly the same regardless of intraperitoneal or intravenous administration of carboplatin. However, the 24-h platinum AUC in the peritoneal cavity was approximately 17 times higher when carboplatin was administered by the intraperitoneal route. Conclusion. The present pharmacological analysis suggests that intraperitoneal infusion of carboplatin is feasible not only as an intraperitoneal regional therapy but also as a more reasonable route for systemic chemotherapy. (c) 2005 Elsevier Inc. All rights reserved.