The objective of the study was to develop a population pharmacokinetic model of pemetrexed and identify factors contributing to variability in exposure in Indian patients. Plasma samples were obtained from a cohort of 85 patients following 500 mg/m2 intravenous infusion and population pharmacokinetic analysis was performed using NONMEM (version 7.3.0). The stochastic approximation expectation maximization method was used to estimate parameters. The full covariate model approach was used by specifying clinically meaningful covariates a priori. Credible intervals obtained using Markov chain Monte Carlo Bayesian analysis were used to reduce the full covariate model by eliminating the covariates whose CI included the null. Model qualification was performed using visual predictive check and bootstrap. The final population parameter estimates and relative standard error for clearance (CL) was 3.3 L/h (10.8), central volume of distribution (V1) was 5.2 L (7.8), peripheral volume of distribution (V2) was 5.9 L (14.5) and intercompartmental clearance (Q) was 6.8 L/h (14.3). A large between-subject variability (50%-108% coefficient of variation) was observed in pharmacokinetic parameters. The percent coefficient of variation for the area under the plasma concentration-time curve from time zero to infinity was 72% and for maximum concentration was 68.25%. Diagnostic plots showed no major bias in the model. The final model included V1, V2, and Q scaled to body surface area raised to a fixed exponent of 1. Creatinine clearance and sex on clearance and albumin on V1 were statistically significant covariates based on Bayesian credible interval. However, traditional bootstrap resulted in a 95% confidence interval of the sex effect parameter including null. Given the size and nonsignificant sex effect in traditional bootstrap, it is considered clinically not significant.
Keywords: BAYES analysis; full covariate modeling; pemetrexed; population pharmacokinetics.
© 2019, The American College of Clinical Pharmacology.