Evaluation of 5-FU pharmacokinetics in cancer patients with DPD deficiency using a Bayesian limited sampling strategy

Aims: Dihydropyrimidine dehydrogenase (DPD) is the initial enzyme in the catabolism of 5-fluorouracil (5FU) and DPD deficiency is an important pharmacogenetic syndrome. The main purpose of this study was to develop a limited sampling strategy to evaluate the pharmacokinetics of 5FU and to detect decreased 5FU elimination in patients with c.1905+1G>A (IVS14+1G>A) mutation related DPD deficiency. Methods: Thirty patients, heterozygous for the c.1905+1G>A (IVS14+1G>A) mutation in DPYD, and 16 control patients received an intravenous dose administered over 2 min of 300 mg 5FU/m2 and/or 450 mg 5FU/m2, followed by blood sampling over 2 hrs and subsequently pharmacokinetic analysis of the 5FU plasma levels. A population pharmacokinetic analysis was performed in order to develop a quantitative compartmental pharmacokinetic model suitable for limited sampling strategy. Clinical aspects of treating DPD patients with 5FUbased chemotherapy was assessed from the retrospectively collected clinical data. Results: A two compartment model with Michaelis-Menten elimination from the central compartment was used in the population pharmacokinetic analysis. The mean (€ SD) AUC (mg.h/L), T1/2 s (h), Km (mg/L) and Vmax (mg/h) of 5FU in DPD deficient patients vs. controls were 9.1 € 4.0 vs. 6.0 € 2.1 (AUC), 0.268 € 0.116 vs. 0.128 € 0.043 (T1/2 s), 4.81 € 0.39 vs. 4.39 € 0.72 (Km) and 942 € 310 vs. 1749 € 380 (Vmax) at a dose of 300 mg 5FU/m2. For a dose of 450 mg 5FU/m2, the PK values in DPD deficient patients vs. controls were 17.7 € 5.4 vs. 13.4 € 3.9 (AUC), 0.306 € 0.103 vs. 0.181 € 0.041 (T1/2 s), 5.02 € 0.32 vs. 4.58 € 0.25 (Km) and 900 € 194 vs. 1370 € 267 (Vmax). The differences in AUC, T1/2 s and Vmax between DPD deficient patients and controls were all highly significant (p <0.05). Using a single sample t = 30 min limited sampling strategy, the positive predictive value and negative predictive value for Vmax were 96% and 88 %, respectively. Conclusion: Profound differences in pharmacokinetics of 5FU existed between DPD deficient patients and control patients. Pharmacokinetic 5FU profiling using a limited sampling model may be useful for identification of DPD deficient patients in order to reduce severe toxicity.