Continuous portal venous infusion of 5-fluorouracil (5-FU) and/or heparin immediately following injection of tumor cells into the portal system in unrestrained rats has been assessed as a pilot study to examine whether the incidence of liver metastases could be reduced and survival therapy improved. Thirty-six rats were divided into four groups according to the treatment administered for 7 days by portal infusion: heparin, 5-FU, 5-FU + heparin and physiological saline solution as a control. After a 90-day follow-up, metastases were observed in every rat injected with tumor cells with the exception of one rat in the heparin group. The 90-day actuarial survival rates for the treatment groups of heparin, 5-FU, 5-FU + heparin and control were 44% (4/9), 11% (1/9), 33% (3/9) and 22% (2/9) respectively, with the survival curves not differing significantly. The apparently improved survival rate in the heparin group may be due to the insufficient period of observation. The conclusions of this study may be only preliminary owing to both the insufficient period of observation and the small number of rats, but the inhibitory effect of 5-FU on the development of liver metastases was disappointing as a result of the large number of tumor cells injected into the portal system. The technique used in this study, which enabled continuous portal venous infusion of drugs to be performed in unrestrained rats, may provide a useful model for the study of continuous infusion in the rat.