RESUMO
PURPOSE: This study investigated correlations of the clinical outcomes of oral metronomic vinorelbine (VNR) with VNR pharmacokinetics and MDR1 polymorphisms. METHODS: Eighty-two patients with metastatic non-small cell lung cancer (NSCLC) unfit for standard chemotherapy were treated with VNR at the oral doses of 20-30 mg every other day or 50 mg three times a week. They had a performance status (PS) ≤ 3, were > 70-year-old and drug-naïve or cisplatin-pretreated. MDR1 2677G > T and 3435C > T polymorphisms were analysed and blood concentrations of VNR and desacetyl-VNR (dVNR: active metabolite) assayed. Overall survival (OS), treatment duration and drug-related toxicity were the main endpoints. RESULTS: Median OS and treatment duration were 27 weeks (range 1.3-183) and 15 weeks (range 1.3-144), respectively. OS was directly correlated with the duration of VNR treatment and number of therapy lines after VNR treatment (multiple linear regression: adjusted r2 = 0.71; p < 0.00001). Neither MDR1 genotypes nor VNR/dVNR concentrations predicted OS. VNR blood levels were positively correlated with platelet counts (r2 = 0.12; p = 0.0036). Patients who had long-term benefit (treated for ≥ 6 month without toxicity) showed lower VNR concentrations than those who had not. Twelve patients stopped therapy due to grade 3-4 toxicity. Toxicity was associated with blood concentrations of VNR ≥ 1.57 ng/mL and dVNR ≥ 3.04 ng/mL, but not with MDR1 polymorphisms. CONCLUSIONS: Neither pharmacokinetic nor pharmacogenetic monitoring seem useful to predict OS. On the other hand, high VNR and dVNR blood levels were associated with severe toxicity.