Intraperitoneal administration of docetaxel has been used to treat peritoneal dissemination of cancer, but its safety has not yet been confirmed. We have compared the pharmacokinetic behaviour of docetaxel after intravenous and intraperitoneal administration in CD-1-nu/nu mice bearing MKN45P, a gastric cancer variant line producing peritoneal dissemination. Docetaxel (8 mg kg(-1)) was intravenously or intraperitoneally injected into the mice and at designated times the drug concentration was measured in plasma, ascites fluid, and abdominal tissues (liver, kidney, intestine and spleen, solid cancer, and suspended free cancer). The pharmacokinetic behaviour of docetaxel was similar in control mice and cancer-bearing mice after administration via either route, except that the transfer of docetaxel from the abdominal cavity to systemic blood (plasma) was slower in cancerbearing mice than in control mice. As expected, the intraperitoneal drug concentration was much higher (approximately 100-fold) and was maintained for a longer time in the intraperitoneal injection group than in the intravenous injection group. The drug concentrations in peritoneal solid cancer tissue and suspended free cancer cells were also significantly higher for a longer time in the intraperitoneal injection group than in the intravenous injection group. The values of the plasma area under concentration-time curves (AUC) were similar for both administration routes. The ratio of AUC ascite/AUC plasma after intraperitoneal administration was higher than after intravenous administration. The drug concentration in abdominal organs after intraperitoneal injection was lower during the first 2 h, then became similar to those after intravenous injection. These results indicated that the intraperitoneal administration of docetaxel for peritoneal dissemination was likely to be an effective treatment method, without causing any increase in systemic toxicity.