Pharmacological evaluation of experimental isolated liver perfusion and hepatic artery infusion with 5-fluorouracil.

The intention of this study was to estimate the pharmacological advantage of a clinically applicable method of isolated liver perfusion (ILP) over hepatic artery infusion (HAI) administering various doses of 5-fluorouracil (FUra). FUra concentrations were measured using high-performance liquid chromatography in liver tissue (pigs and rats), hepatic tumor tissue (rats), and in the systemic circulation (pigs) following ILP and HAI. Forty-two pigs and 36 rats were subjected to either ILP or HAI with 20, 40 or 80 mg of FUra/kg of body weight. ILP resulted in significantly increased FUra concentrations in the liver as compared with the results with HAI in rats and pigs. Median areas under the concentration-time curve in liver tissue were 122.7 mumol.g-1.min and 59.9 mumol.g-1.min (40-mg/kg dose-group) and 236.3 mumol.g-1.min and 45.1 mumol.g-1.min (80 mg/kg) for ILP and HAI, respectively in pigs (both P less than 0.05). Systemic plasma areas under the curve were significantly lower for ILP as compared with HAI in 40- and 80-mg/kg dose-groups with 2.2 mumol.ml-1.min and 9.2 mumol.ml-1.min (40 mg/kg; P less than 0.01) and 6.8 mumol.ml-1.min and 43.2 mumol.ml-1.min (80 mg/kg; P less than 0.01) for ILP- and HAI-treated pigs, respectively. In hepatic tumor tissue a dose-dependent increase of mean FUra concentration was found for ILP-treated rats (P less than 0.05). No significant differences were observed in median FUra concentrations in tumor tissue between ILP- and HAI-treated rats (0.66 mumol.g-1 and 0.63 mumol.g-1 for ILP- and HAI-treated groups with 80 mg/kg; P greater than 0.05). The mean FUra concentration tumor/liver ratio was 0.26. In order to clarify the metabolic fate of high-dose FUra, five rats were subjected to HAI with 150 mg of FUra/kg, and hepatic tumor extracts excised at t = 0 min, t = 5 min, and t = 15 min after infusion were analyzed using 19F nuclear magnetic resonance. Catabolite alpha-fluoro-beta-alanine appeared rapidly at t = 5 min and t = 15 min in liver tissue. Significant amounts of the presumed active nucleotides were not detected in tumor tissue. We conclude that ILP is a means to improve selectivity of administration of antitumor agents to the liver, as compared with HAI. The pharmacological advantage of ILP over HAI administering equivalent doses of FUra was not demonstrated in tumor tissue, because of a large differential between liver tissue extraction and tumor tissue extraction of FUra, which was influenced by the mode of administration.