To examine the antitumor effectiveness of continuous venous infusion of 5-fluorouracil (5-FU), such a regimen was mimicked in vitro using cultured ovarian cancer cell lines. Two ovarian carcinoma cell lines, HRA and KK with cell doubling times (DT) of 16 h and 45 h, respectively, were grown under three incubation conditions with respect to 5-FU exposure, and the extent of growth inhibition was compared among the three conditions and between the two cell lines. Protocol I: 6 h-intermittent exposure to 5-FU, protocol II: 24 h-intermittent exposure to 5-FU, protocol III: uninterrupted exposure to 5-FU. The 50% growth inhibitory concentrations (IC50) of 5-FU obtained for KK cells grown under protocols I and II were both two times higher than that obtained under protocol III. Regarding HRA cells, IC50 obtained under protocol I was two times higher than that obtained under protocol III, while that obtained under protocol II was 3.7 times higher than that obtained under protocol III. Results obtained from the present in vitro examination suggested that the drug-free interval in the intermittent infusion of 5-FU should be shorter than the DT of the targeted tumor cells to obtain an efficacy corresponding to that obtained by uninterrupted exposure to the drug. In establishing an effective intermittent continuous infusion therapy with 5-FU with fewer side effects, the drug-free interval should be shorter than the DT of tumor cells, but longer than the DT of proliferating cells in normal tissues, including gastrointestinal mucosa cells assuming that these normal cells possess shorter DT than tumor cells.