Although continuous infusion of 5-fluorouracil (5-FU) has been clinically demonstrated to be superior to bolus administration, the mechanistic difference between the treatments still remains unclear. Here, we investigated in vitro whether there were any differences in the metabolism and activity of 5-FU between these schedules. To simulate bolus and infusional treatments of 5-FU, HST-1 human squamous carcinoma cells were treated with short-term, high-doses and long-term, low-doses so that the area under the curve (AUC) of 5-FU became equivalent between both schedules, and compared the cytotoxicity, fluorinated RNA (F-RNA) levels, 5-fluorodeoxyuridine monophosphate (FdUMP) content and thymidylate synthase (TS) activity. F-RNA and FdUMP were measured by capillary gas chromatography-mass spectrometry and competitive ligand-binding assay, respectively. The [H]FdUMP binding site in TS was determined as an index of the amount of TS using the radio-binding assay. Long-term, low-dose treatment of 5-FU was found to be 1.3-1.7 times more cytotoxic than the short-term, high-dose treatment. F-RNA content increased as the AUC of 5-FU was increased and was 2-4 times significantly higher in the cells treated with the long-term, low-dose than those with the short-term, high-dose schedule, indicating that the levels of F-RNA are AUC and schedule dependent. In contrast, there were no significant differences in FdUMP levels, TS activity and TS inhibition rate between the schedules. These data suggest that the superior activity of 5-FU administered long-term, low-dose over short-term, high-dose could be explained by more 5-FU incorporated into RNA during a long-term, low-dose exposure, thus providing a strategic rationale for the clinical advantage of continuous infusion over bolus administration.