The synergistic interaction of etoposide with cisplatin in certain tumors prompted an evaluation of its potential role in the i.p. treatment of ovarian cancer and other intraabdominal malignancies. We conducted a Phase I evaluation of etoposide as a single agent to determine the maximum tolerated dose i.p., to describe dose-limiting and other toxic effects, and to examine the pharmacokinetics of etoposide in this setting. Etoposide was diluted in 2 liters of normal saline, and instilled i.p. over 10 to 25 min following maximal drainage of ascites. The dwelling time was 4 h, followed by peritoneal drainage. Twenty-two patients received 56 courses at doses which ranged from 100 to 800 mg/m2. The median age was 49, the median performance status was 1, and 18 patients had received prior chemotherapy, with or without radiation. The principal acute toxicity was abdominal pain in 10 patients; this was usually accompanied by signs of peritoneal irritation and was always responsive to nonsteroidal antiinflammatory medications. The major toxicity was dose-related neutropenia; Grade 3 or 4 toxicity affected five of six patients at 800 mg/m2. Thrombocytopenia, nausea and vomiting, and alopecia were also observed. The recommended dose for further study is 700 mg/m2. The pharmacokinetics of etoposide in plasma and peritoneal fluid was measured in 19 patients. Peritoneal levels over the 4-h dwelling time declined monoexponentially with a harmonic mean half-life of 3.5 h (range, 1.9 to 7.8). Plasma levels rose to a peak at 2.9 +/- 1.7 (SD) h and then declined exponentially with a harmonic mean terminal half-life of 7.7 h (range, 4.2 to 15.6). The plasma area under the concentration-time curve increased linearly with respect to dose. The relative pharmacological advantage (ratio of peritoneal to plasma area under concentration-time curve) for i.p. administration was measured as 2.8 and was independent of dose. Based on the high plasma protein binding of etoposide (94%) and the minimal protein binding in the fluid instilled i.p., the ratio of the areas under the concentration-time curves of free drug is estimated to be 4%. These results illustrate that tumor confined to the peritoneal cavity would be exposed to substantially higher free (diffusible) drug concentrations following i.p. than following i.v. administration and support the further evaluation of etoposide by this route.