BACKGROUND The main elimination pathway of vinorelbine is hepatic metabolism, and the clearance of vinorelbine could be reduced in patients with liver metastases. OBJECTIVE To study the pharmacokinetics of vinorelbine in patients who have advanced breast cancer with or without liver metastases and to study the relationship between hepatic function and vinorelbine clearance. METHODS We studied 29 patients with advanced breast cancer: 19 with liver metastases and 10 control patients with extrahepatic metastases (mean age, 61 years; age range, 38 to 81 years). The vinorelbine dose was 30 mg/m2 as a short intravenous infusion; the dose was reduced by 50% in patients with bilirubin > 2 mg/dl. Patients were classified by ultrasonographic estimation of the liver volume replaced by tumor (%LVRT). Standard liver function tests and a monoethylglycinexylidide test (a quantitative liver function test based on lidocaine metabolite formation) were performed. Vinorelbine was assayed in plasma by HPLC with fluorescence detection. Vinorelbine determination was impossible in two patients with more than 75% LVRT because of interferences. Pharmacokinetic parameters were calculated with a noncompartimental method and compared by means of the Kruskal-Wallis test. RESULTS A lower vinorelbine clearance rate was observed in the five patients with more than 75% LVRT (22.9 L/hr/m2) compared with the 10 patients with no liver metastases (48.0 L/hr/m2) and the 12 patients with 25% to 75% LVRT (45.3 L/hr/m2). Terminal elimination half-life and apparent volume of distribution were not significantly different among groups. The monoethylglycinexylidide test had a significant correlation with vinorelbine clearance. (r2 = 0.70; p = 10(-4). CONCLUSIONS These results support vinorelbine dose reduction in patients with severe liver failure but not in patients with moderate secondary liver involvement. The monoethylglycinexylidide test may prove to be useful for vinorelbine dose individualization.