Exposure-response analysis of Raltitrexed assessing liver toxicity

Raltitrexed (RTX) is a thymidylate synthase inhibitor with large pharmacokinetics (PK) variability that can be administered in case of 5-fluorouracil (5FU) intolerance or dihydropyrimidine dehydrogenase deficiency. While it is a more potent thymidylate synthase inhibitor than 5FU, RTX failed to replace this drug for colorectal cancer patients, mainly due to its toxicity at the recommended dose of 3 mg/m² every 3 weeks. However, every 2 weeks administration at 2 mg/m² demonstrated a favourable toxicity profile.

We performed a randomized crossover comparative population PK study between every 2 weeks TOMOX (RTX 2 mg/m² ) and every 3 weeks TOMOX (RTX 3 mg/m² ).

A three-compartment model and a proportional error model best describe the data. Creatinine clearance and sex, but not body surface area (BSA), were covariates of RTX clearance leading to decrease of its interindividual variability of 28%. Weight and body surface area were covariates of central and peripheral volumes of distribution, respectively, leading to decreases of interindividual variability of 34.6% and 100%, respectively. In contrast to the dose, AUC was a good predictor of liver toxicity (P = 0.006, OR = 3.91, 95%CI = [1.48-10.34]). Using covariates to compute individual clearance and a threshold AUC (1.639, determined in this study), a covariates-based dose was calculated, leading to less variability in AUC than observed with the actual BSA-based or fixed doses.

These results advocate for the use of creatinine clearance and sex to determine the RTX dose instead of BSA.