A phase I and pharmacokinetic study of gemcitabine given by 24-h hepatic arterial infusion

Abstract Purpose This study was performed to assess the toxicities, the maximum-tolerated dose (MTD), the pharmacokinetics and the anti-tumour activity of gemcitabine given by 24-h hepatic arterial infusion (HAI). Patients and methods Patients with liver malignancies received gemcitabine by 24-h HAI, weekly × 3, every 4 weeks. On day 1 or day 8 of the first cycle, patients received one administration by 24-h intravenous infusion for pharmacokinetic comparison and to determine hepatic extraction. Results Thirteen patients received gemcitabine at the dose levels of 75, 135 and 180 mg/m2 . The MTD was 180 mg/m2 with thrombocytopaenia as the dose-limiting toxicity. Pharmacokinetic analysis showed a significantly lower maximum gemcitabine plasma concentration ( Cmax : HAI, 26, 80 and 128 nM, respectively; IV, 229, 264 and 293 nM, respectively) and area under the plasma-concentration-versus-time curve (AUC0–24h : HAI, 386, 1247 and 2033 nmol × h/L, respectively; IV, 3526, 4818 and 5363 nmol × h/L, respectively) during HAI, compared with intravenous infusion (both P < 0.001). Additionally, the mean hepatic extraction ratios of gemcitabine at the 75, 135 and 180 mg/m2 dose level were 0.89, 0.75 and 0.55, respectively. Hepatic extraction decreased linearly with increasing dose. The Cmax and AUC0–24h of 2′,2′-difluoro-2′-deoxyuridine, the deaminated product of gemcitabine, were similar for HAI and intravenous infusion. Seven patients had stable disease for a median duration of 9 months (range: 2–11 months). Conclusions Gemcitabine given by 24-h HAI was well tolerated and resulted in significantly lower systemic gemcitabine plasma concentrations than intravenous infusion due to a relatively high hepatic extraction.