Population pharmacokinetics of 5-fluorouracil in colorectal cancer patients

Aims. The pharmacokinetics of 5-fluorouracil (5-FU) after intravenous administration in color- ectal cancer patients were examined using population analysis. The relevant covariates and the extent of inter- and intraindividual variability were evaluated.

Methods. Data from 27 patients with diagnosis of nonmetastatic colorectal adenocarcinoma receiving weekly 5-FU (450 mg/m²), plus levamisol 50 mg/8 hours by oral route for 3 days every 15 days, were pooled with data from 17 patients with diagnosis of metastatic colorectal adenocarcinoma, receiving daily 5-FU (425 mg/m²) and intravenous folinic acid (20 mg/m²) over five consecutive days (daily times five), every four weeks. In both groups 5-FU was administered as a 60-minute infusion and blood samples were collected at 10, 30 and 60 minutes from the end of the infusion, and analysed using a validated high-performance liquid chromatography assay. An open two-compartmental pharmacokinetic model with first-order elimination from central compartment was fitted to the plasma concentration data using nonlinear mixed effect modelling (NONMEM). The potential effect of patient covariates was evaluated using a stepwise method. Model evaluation was performed using bootstrap method.

Results. The pharmacokinetic model was successfully fitted to the data. None of the covariates tested were significantly correlated to the pharmacokinetic parameters. The mean parameters’ estimates (%CV) and the percent coefficient of variation of the central tendency parameters, interindividual (IIV), interoccasion (IOV) and residual variability (s) for the final model were: CL (L/h), 65.3 (13.2); Vc (L), 14.7 (11.8); Vp (L), 334.0 (31.4); Q (L/h), 19.6 (25.5); IIV_CL (%), 76.5 (34.6); IIV_Vc (%), 82.3 (31.0); IIV_Vp (%), 137.5 (35.1); IIV_Q (%), 117.5 (38.5); IOV_CL (%), 66.1 (45.3); IOV_Vc (%), 70.8 (39.5); IOV_Q (%), 81.1 (27.8) and s (%), 3.0 (25.4). The bootstrap resampling method confirmed the stability of the final model. The estimates of the central tendency parameters, IIV, IOV and residual variability were essentially equal to those generated with the original dataset (0% to 18% deviation) and the 95% confidence intervals included the mean parameters’ estimates obtained from the former set.

Conclusions. The two-compartmental model accurately described the pharmacokinetics of 5-FU administered by short-term infusion. A population pharmacokinetic approach is a useful tool to integrate the knowledge gathered in the clinical setting. The model developed may help in dose adaptation and will further be used in PK/PD modelling of therapeutic outcomes and adverse events.