The present phase I trial was planned to assess the maximum tolerated dose, the dose-limiting toxicity and the pharmacokinetics of bendamustine hydrochloride in a once every 3 weeks schedule, and to recommend a safe dose for future phase II studies. Included were patients with refractory solid tumors. Bendamustine hydrochloride was administered as a short intravenous infusion over 30 min. The starting dose was defined at 160 mg/m2 and dose escalation used increments of 20 mg/m2. Plasma and urine samples were analyzed using validated high-pressure liquid chromatography/fluorescence assays. Twenty-six patients (14 men, 12 women) were enrolled for the study. At 280 mg/m2, one out of four patients developed a thrombocytopenia grade 4, two experienced grade 3 fatigue and three experienced cardiac toxicity (grade 2). The latter toxicity was considered dose limiting also and further dose escalation was stopped. Plasma pharmacokinetics parameters of bendamustine hydrochloride and its metabolites were assessed in 15 patients. Mean pharmacokinetic parameters of bendamustine hydrochloride were a tmax of 32.3 min, a t1/2 of 37.8 min, a volume of distribution of 14.2 l/m and a clearance of 287.8 ml/min/m2. No dose dependency of bendamustine hydrochloride was observed within the used dose range. The metabolites comprised only 23% of the overall area under the concentration-time curve. The maximum tolerated dose of bendamustine hydrochloride on day 1 q 3 weeks is 280 mg/m2. Fatigue and cardiac toxicity were dose limiting. The plasma pharmacokinetics data of bendamustine and its metabolites were in accordance with previous reports. The recommended dose for future trials is 260 mg/m2 every 3 weeks.