Population PK/PD model for diflomotecan, a new E-ring modified camptothecin analogue, in patients of a phase I study

Abstract

2046 Background: To develop a semimechanistic population PK/PD model of neutropenic effects of diflomotecan, to establish the concentration-effect relationship and to quantify the degree of interpatient variability. Methods: PK data were obtained from 24 patients with advanced cancer receiving diflomotecan as 20 min i.v. infusion every 3 weeks at planned starting doses of 2, 4, 5 and 6 mg/m2. The number of cycles ranged from 1 to 7. Diflomotecan was measured in 258 plasma samples after the 1st administration. Neutrophil cell counts in peripheral blood were followed every 3–7 days for the entire course of the trial with a total of 505 counts. Population PK/PD analysis was done with NONMEM. Dose, demographic and biochemical variables were tested as covariates. Results: Plasma levels-time data were described with a 3-compartment model and first order elimination. Typical estimates of distribution volumes were 41.5, 39.2 and 12.8 L for central (Vc), shallow (VP1) and deep (VP2) compartments, respectively. Body surface area (BSA) and gender were the only covariates included in the model. Vc increased linearly with BSA: Vc = 41.5 x (BSA/1.85), and VP1 was lower in females (29.5 vs 48.8 L). Simulation analysis showed that neither covariate exerts a clinically relevant influence. The typical value of CL was 21.6 L/h. Interpatient variability was 52% for Vc and 51% for CL. The time-course of neutrophil cell counts was described by means of a semimechanistic model including self-renewal, maturation, homeostatic regulation processes and drug effects (Friberg et al, 2002, J. Clin. Oncol., 20:4713–4721). Values of 4.58 cells x 109/L, 5.38 days and 637.4 μM-1 (based on unbound concentrations) were estimated for Circ0 (baseline neutrophil counts), MTT (mean transit time) and slope (parameter related to the cytotoxic potential of the drug), respectively. Conclusions: The semimechanistic population PK/PD model described the time course of diflomotecan plasma concentrations and its haematological toxicity profiles appropriately. Consequently, this model may be used to predict the time-course of neutrophil cell count in future clinical studies. No significant financial relationships to disclose.