936 A bayesian dosing method for carboplatin (CBDCA) in clinical practice

The experience we have acquired with cisplatinum (CDDP) has permitted us to use a similar Bayesian dosing method for the administration of carboplatin (CBDCA): with continuous infusion of the drug during 120 hours, prior estimation of pharmacokinetic parameters and adaptation of the daily dose according to total platinum plasma concentration. Our first step was to define a reference population and to validate our method for theoretical larger plasma concentrations of 1.0, 1.5 and 1.8 mg/l. 79 patients with a median age of 56 years were treated with CBDCA for different types of cancer (head and neck, gastrointestinal, genitourinary…). Treatment protocols differed according to the type of cancer and its grade (associations with radiotherapy and/or polychemotherapy). The infusion times were 120 hours according the use of a volumetric pomp and an implanted venous access port. The measured platinum concentrations at the end of infusions were respectively 1.0 mg/l ± 0.095, 1.49 mg/l ± 0.13 and 1.8 mg/l ± 0.17, compared with the theoretical end-point (1.0, 1.5 and 1.8 mg/l). The median dosages of CBDCA were 280.0 mg/m 2  ± 40.0 mg/m 2 for a maximal theoretical concentration of 1.0 mg/l, 416.5 mg/m 2  ± 90.0 for 1.5 mg/l and 523.5 mg/m 2  ± 101.0 for 1.8 mg/l. These doses lead to a total platinum plasmatic AUC of 218.0 mg/l × h ± 61.7 for theoretical end-point of 1.0 mg/l, 293.1 mg/l × h ± 81.4 for 1.5 mg/l and 375.0 mg/l × h = 129.3 for 1.8 mg/l. The residual concentration of total platinum varied from 0.01 mg/l to 0.57 mg/l for the alternated courses and the non-alternated courses. The Bayesian dosing method of CBDCA is perfectly applicable in clinical practice. It allows us to control the major side effects of the drug.