Abstract
PURPOSE: Carboplatin doses are often determined by using modified Calvert formulas. It has been observed that the area under the concentration versus time curve (AUC) for free carboplatin is lower than expected when modified formulas are used for carboplatin/paclitaxel chemotherapy combination regimens. By using limited-sampling models, the carboplatin AUC actually reached can easily be verified, and the dose adjusted accordingly. METHODS: In this report, we describe the development and validation of a limited-sampling model for carboplatin from 77 pharmacokinetic curves, when carboplatin is used in combination with paclitaxel. RESULTS: The following single-point model was selected as optimal: AUC carboplatin (min mg(-1) ml(-1)) = 418. c(2.5 h)(mg/ml) + 0.43 (min mg(-1) ml(-1)), where c(2.5 h) is the concentration (mg/ml) of carboplatin 2.5 h after the start of a 30-min infusion. This model proved to be unbiased (mean prediction error = 3.4 +/- 1.6%) and precise (root mean square error = 10.1 +/- 1.5%). CONCLUSIONS: The proposed model can be very useful for ongoing and future carboplatin/paclitaxel studies aimed to optimise and individualize treatment.