The objective of this study was to explore correlations between a variety of covariates and oxaliplatin ultrafilterable and blood pharmacokinetic parameters. Data from 40 patients receiving oxaliplatin combined with 5-fluorouracil and levofolinic acid as standard treatment for advanced colorectal cancer were analysed. Plasma ultrafilterable, blood, and urine platinum concentrations were determined by flameless atomic absorption spectrophotometry. Data were analysed according to a population pharmacokinetic method using the NONMEM program. The best fit for oxaliplatin plasma ultrafilterable clearance (CL) was given by the following equation, which considers four covariates: body surface area (BSA, in metres squared), age (in years), sex (0 if male, 1 if female), and serum creatinine (Scr, in micromoles per liter): CL (l/h)=5.49xBSA+4.55xBSAx(140-AGE)x(1-0.15xSEX)/Scr. By taking into account these covariates, the interindividual variability in CL decreased from 43% to 33%. Renal clearance represented 34% of the overall elimination. This value was obtained by recovering urine over only 5 h from the beginning of the infusion and modelling the data using NONMEM. We would recommend the use of this methodology for pharmacokinetic studies in oncology in which renal clearances of the drug are presently rarely explored. The oxaliplatin blood concentrations versus time observed during the three-cycle period were well-described by a three-compartment model with first-order elimination from the central compartment. No significant intrapatient pharmacokinetic variability was observed between cycles. The relationship we obtained using the population approach between oxaliplatin CL and covariates may allow rational reduction of oxaliplatin dose in cases of elevated serum creatinine levels.