Our objective was to determine the maximum tolerated dose of cyclosporin A (CsA) delivered as a loading dose (LD) and continuous i.v. infusion (CI) in combination with carboplatin in patients with refractory gynecologic cancers. Twenty-nine heavily pretreated patients (25 ovarian epithelial, 2 cervical, and 2 endometrial carcinomas) received 113 cycles of CsA and carboplatin from September 1989 to September 1991. Twenty-four of these 29 carcinomas were strictly defined to be platinum resistant. CsA was administered as a LD escalated from 6 to 10 mg/kg followed by a 24-h CI from 2.5 to 14.5 mg/kg/day. Carboplatin was targeted to an area under the time versus concentration curve (AUC) of 6 mg/ml x min and was not dose escalated. Whole-blood CsA concentrations (fluorescence polarization immunoassay) at the maximum tolerated dose (10 mg/kg LD, 14.5 mg/kg/day CI) ranged from 2.4 to 3.0 microgram/ml over 12 h. Estimated median carboplatin AUC, based on calculated carboplatin clearance, was 7.9 mg/ml x min. The dose-limiting toxicity of the combination of CsA and carboplatin was grade 4 thrombocytopenia. Grade 3 or 4 thrombocytopenia occurred in 35% of the patients, which could be explained by the effects of carboplatin (AUC of 6 mg/ml x min) alone. Overall, neutropenia occurred in 24% of the patients and anemia in 17% of the patients. Grade 3 or 4 nausea or vomiting was noted in 10 and 14% of the patients, respectively. Grade 3 hypertension during CsA administration occurred in 14% of the patients. No grade 3 or 4 nephrotoxicity was seen in this trial. Three objective responses were noted: one complete response (11 months) and one partial response (5 months), both in potentially platinum-sensitive patients with platinum-free intervals of only 9 months each. One platinum-resistant patient had a partial response for 21 months. Five additional patients experienced >75% reduction of CA-125 or a return to a normal CA-125 titer. We concluded that whole-blood CsA concentrations of >3.0 microgram/ml (as seen when CsA is used as a modulator of multidrug resistance) were not achievable in this combination with carboplatin in this population of heavily pretreated gynecologic cancer patients. However, because CsA is used in this trial as a chemosensitizer in platinum-sensitive tumors and as a chemomodulator of platinum resistance, we targeted a CsA concentration of >1.0 microgram/ml, which was achieved. The CsA dose recommended for a Phase II trial of this combination is 10 mg/kg LD and 11.6 mg/kg/day CI, which results in blood CsA concentrations ranging from 1.2 to 1.3 microgram/ml over 12 h. Responses in this population of refractory gynecologic cancer patients are unusual, and these encouraging results form the basis for a Phase II trial of this combination.