As pharmacokinetics represents a bridge between pharmacological concentrations and clinical regimens, the pharmacokinetic exploration of the therapeutic dose range is a major outcome. This study was aimed at assessing pharmacokinetic linearity of i.v. vinorelbine through an open design with intra-patient dose escalation (3 doses/group). Three groups of six patients received either 20-25-30 mg/m2; or 25-30-35 mg/m2; or 30-35-40 mg/m2. The inclusion criteria were: histologically confirmed tumour with at least one assessable target lesion, age 25-75 years, WHO PS < or =2, normal haematology and biochemistry, life expectancy > or =3 months. The pharmacokinetics was evaluated in both whole blood and plasma over 120 h. Twenty-six patients were recruited and 18 were evaluable for pharmacokinetics. The toxicity consisted in grade < or =3 leucopenia and neutropenia (<20% of courses) and two grade 4 constipation with rapid recovery (2/54 courses). Compared to blood, plasma was demonstrated to underestimate the pharmacokinetic parameters. In blood, the drug total clearance was about 0.6 l/h/kg, with minor contribution of renal clearance, steady state volume of distribution close to 13 l/h/kg, and elimination half-life at about 40 h. A pharmacokinetic linearity was demonstrated up to 40 mg/m2, and even up to 45 mg/m2 when pooling data from another study. A pharmacokinetic-pharmacodynamic relationship was evidenced on leucopenia and neutropenia when pooling the data from the two studies.