Population pharmacokinetics of raltitrexed in patients with advanced solid tumours

AIMS: To investigate the population pharmacokinetics of raltitrexed in patients with advanced solid tumours and to identify patient covariates contributing to the interpatient variability in the pharmacokinetics of raltitrexed. METHODS: Patient covariate and concentration-time data were collected from patients receiving 0.1-4.5 mg m(-2) raltitrexed during the early clinical trials of raltitrexed. Data were fitted using nonlinear mixed effects modelling to generate population mean estimates for clearance (CL) and central volume of distribution (V). The relationship between individual estimates of the pharmacokinetic parameters and patient covariates was examined and the influence of significant covariates on the population parameter estimates and their variance was investigated using stepwise multiple linear regression. The performance of the developed model was tested using an independent validation dataset. All patient data were pooled in the total cohort to refine the population pharmacokinetic model for raltitrexed. RESULTS: three-compartment pharmacokinetic model was used to fit the concentration-time data of raltitrexed. Estimated creatinine clearance (CL(CR)) was found to influence significantly the CL of raltitrexed and explained 35% of variability in this parameter, whilst body weight (WT) and serum albumin concentrations (ALB) accounted for 56% of the variability in V. Satisfactory prediction (mean prediction error 0.17 micro g l(-1) and root mean square prediction error 4.99 micro g l(-1)) of the observed raltitrexed concentrations was obtained in the model validation step. The final population mean estimates were 2.17 l h(-1)95% confidence interval (CI) 2.06, 2.28] and 6.36 l (95% CI 6.02, 6.70) for CL and V, respectively. Interpatient variability in the pharmacokinetic parameters was reduced (CL 28%, V 25%) when influential covariates were included in the final model. The following covariate relationships with raltitrexed parameters were described by the final population model: CL (l h(-1)) = 0.54 + 0.02 CL(CR) (ml min(-1)) and V (l) = 6.64 + 0.08 WT (kg) - 0.16 ALB (g l(-1)). CONCLUSIONS: A population pharmacokinetic model has been developed for raltitrexed in patients with advanced cancer. Pharmacokinetic parameters of raltitrexed are markedly influenced by the patient's renal function, body weight and serum albumin levels, which may be taken into account in dose individualization. The use of influential covariates to guide anticancer dosage selection may result in less variability in drug exposure and potentially a better clinical outcome.