Limited sampling model for the analysis of 5-fluorouracil pharmacokinetics in adjuvant chemotherapy for colorectal cancer

Background: Administration of 5-fluorouracil may be associated with life-threatening toxicities, resulting from a reduced drug biotransformation to the inactive metabolite 5-fluoro-5,6-dihydrouracil. Patients with severe toxicities display significant alterations of 5-fluorouracil pharmacokinetics, the monitoring of which may be made easier by the availability of a limited sampling model (LSM). Methods: LSMs for 5-fluorouracil and 5-fluoro-5,6-dihydrouracil therapeutic monitoring have been developed in 80 patients with colorectal cancer (training set) given 5-fluorouracil, 370 mg/m(2) per day as an intravenous bolus, plus leucovorin, 100 mg/m(2) per day, for 5 days every 4 weeks. Pharmacokinetic analysis was performed on plasma levels of 5-fluorouracil and 5-fluoro-5,6-dihydrouracil obtained on day 1 of the first cycle of chemotherapy, and backward stepwise regression analysis was used to determine the optimal LSM on the basis of bias (percentage mean prediction error [MPE]) and precision (percentage root mean square prediction error [RMSE]). Results: An optimal model based on 2 time points was obtained (percentage MPE=1.99%+/-1.41%; percentage RMSE=12.70%+/-1.27%), and the predicted area under the time versus plasma concentration curve (AUC) was calculated as follows: predicted AUC (h.mug/mL)=0.119.C-5+1.436.C-45+2.066, in which C-5 and C-45 are plasma concentrations of 5-fluorouracil at 5 and 45 minutes after drug administration, respectively. The application of this algorithm to pharmacokinetic analysis of plasma levels of 5-fluorouracil in 80 patients (test set) allowed a precise estimation of AUC (percentage MPE=-0.09%+/-1.37%; percentage RMSE=12.17%+/-1.23%). The best LSM for 5-fluoro-5,6-dihydrouracil was characterized by a percentage MPE of -0.64%+/-0.86% and a percentage RMSE of 7.64%+/-0.81%, and the optimal sampling time points were 45 and 180 minutes. Conclusion: The current LSM allows a reliable assessment of drug exposure and improves the use of therapeutic drug monitoring for treatment optimization of 5-fluorouracil in patients with cancer.