Purpose: The present study was performed to examine relationships between systemic exposure of capecitabine metabolites (5-FU, 5'-DFCR and 5'-DFUR) and toxicity or clinical response in patients with metastatic breast cancer.
Methods: A population pharmacokinetic model for capecitabine and its three metabolites was built. Typical parameter values, characteristics of random distributions, associated with parameters, and covariates impact were estimated. Area under the curve (AUC) were computed for 5-FU and compared with grades of toxicity. Pharmacokinetic modeling was based on data collected on the first treatment cycle. Toxicity was assessed on the two first treatment cycles.
Results: The study was conducted in 43 patients. The population pharmacokinetic model (a one-compartment model per compound) was able to capture the very complex absorption process of capecitabine. Statistically significant covariates were cytidine deaminase, alkaline phosphatase and dihydrouracilemia (UH2)/uracilemia (U) ratio. UH2/U ratio was the most significant covariate on 5-FU elimination and CDA on the transformation of 5'-DFCR in 5'-DFUR. A trend was observed between 5-FU AUC and thrombopenia toxicity grades, but not with other toxicities. Best clinical response was not linked to systemic exposure of capecitabine metabolites.
Conclusion: In our study, we propose a model able to describe, meanwhile, and its main metabolites, with a complex absorption process and inclusion of enzyme activity covariates such as CDA and UH2/U ratio. Trial registration Eudract 2008-004136-20, 2008/11/26.
Keywords: Breast cancer; Capecitabine; Modelling; PD; PK; Toxicity.