Data from various phase I/II studies of carboplatin in combination with paclitaxel (Taxol; Bristol-Myers Squibb Company, Princeton, NJ) have suggested that the degree of thrombocytopenia seen is less than that expected when carboplatin is given alone. However, some studies also have suggested that the area under the plasma concentration-time curve (AUC) of carboplatin is lower than that expected, raising the possibility of a pharmacokinetic interaction. Patients with advanced epithelial ovarian cancer were treated with first-line carboplatin (AUC = 7, using the 51Cr EDTA [edetic acid] clearance method) and escalating doses of paclitaxel. Thrombocytopenia was mild and was significantly less when the paclitaxel dose was 175 mg/m2 versus 150 mg/m2. Paclitaxel kinetics were nonlinear, as previously reported. The achieved carboplatin AUC was 7 +/- 1 mg/mL.min, indicating that the pharmacokinetics of carboplatin are not affected by paclitaxel. Glomerular filtration rates measured in 184 patients using the 51Cr EDTA clearance method were compared with rates estimated from the plasma creatinine level using the Cockcroft-Gault or Jeliffe formulas and showed a significant bias of these two formulas. Clearances above 50 mL/min were underestimated by an amount that became greater as the clearance increased and was approximately 25% to 35% for patients with clearances in the normal range. Since creatinine-based methods have been used in many previous studies, care is needed in interpreting the predicted AUC values from these studies. Carboplatin and paclitaxel may be given safely in combination at full doses, and the thrombocytopenia seen is significantly less than that observed with single-agent carboplatin. No evidence exists of a pharmacokinetic interaction, and the observation in some studies that the carboplatin AUC was lower than expected was probably due to the methodology used to estimate the glomerular filtration rate.