Despite being one of the oldest anti-cancer drugs, fluorouracil (FU) is still being increasingly used in cancer chemotherapy. The source of variability for FU sensitivity in patients may be complex, although an overproduction of thymidylate synthase (TS) was the only mechanism of resistance identified in tumours from FU-resistant patients. Dihydropyrimidine dehydrogenase (DPD) is the first and rate-limiting enzyme of FU catabolism. Thus, DPD activity may be a potential factor for controlling FU responsiveness. A panel of 19 human tumour cell lines, including digestive tract, breast and head and neck cancer cells, were investigated. Both TS and DPD activities were measured in parallel to FU responsiveness. None of the cell lines had been previously exposed to FU, and thus expressed a spontaneous sensitivity to FU. Sensitivity between cell lines showed marked differences, with IC50 values ranging from 45 ng/ml (colon cell line) to 5063 ng/ml (head and neck cell line). TS activity was measurable in all cell lines and varied within a 46-fold range. DPD activity was detected in all but four cell lines, showing a 100-fold range of variation. Cell lines most sensitive to FU exhibited the lowest DPD and TS activities and vice versa. Simple linear regression analysis showed that both TS (r2 = 0.22, P = 0.042) and DPD (r2 = 0.27, P = 0.022) activities were significantly correlated to FU effectiveness (log 10 IC50): the greater the enzyme activities, the higher the FU IC50. TS and DPD were demonstrated to be independent variables. A multiple regression analysis showed that the combination of TS and DPD activities explained 36% of the variability in FU IC50 (r2 = 0.36, P = 0.01). Two groups of cell lines could be identified, one group with both low TS and low DPD activities (G1), and the other with either high TS and/or high DPD activities (G2). Mean FU IC50 values were 193 and 930 ng/ml in G1 and G2, respectively, and this difference in FU sensitivity was highly significant (P = 0.009). The present study shows, for the first time, that DPD activity in tumour cells is an independent factor significantly related to FU sensitivity. These results should encourage DPD and TS coupled measurements in tumours of patients before FU treatment in order to establish their prognostic relevance. DPD and TS measurements could also be used during the treatment course to determine the implication of these enzymes in the development of tumour resistance to FU.