Oral uoropyrimidines represent a new treatment option in colorectal cancer. So far, two drugs have drawn major attention. UFT is a combination of tegafur and uracil in a 1 : 4 molar ratio. Tegafur is a pro-drug to 5- uorouracil (5-FU) and uracil inhibits dihydropyrimidindehydrogenase, which is a key enzyme in the catabolism of 5-FU. Combined with leucovorin, the drug appears to have an effect equal to that of 5-FU and leucovorin, as shown in two large randomized trials (1, 2), but with a lower risk of toxicity compared with that in the Mayo schedule. This schedule is probably one of the more toxic bolus 5-FU regimens (3). Capecitabine is activated to 5-FU through a number of steps. Thymidinfosforylase catalyses the nal conversion and this enzyme is present in higher concentrations in tumour tissue than in normal tissue, resulting in a tumour selective generation of 5-FU (4). In randomized trials, capecitabine has an activity in line with 5-FU and leucovorin (5, 6). Both of these oral treatments are attractive for a number of reasons, but they have not been directly compared. We conducted two consecutive phase II trials to elucidate the advantages and disadvantages of the two drugs. The two trials had identical criteria for inclusion, evaluation of response, toxicity and follow-up. The inclusion criteria included histopathologic veri cation of the diagnosis; age 575 years; measurable lesions; performance status (WHO) 52; and informed consent. The patients were required to have a normal bone marrow function as estimated by white blood cell count ]31⁄210:L and thrombocytes ]1001⁄2 10:L. Previous chemotherapy was not allowed, with the exception of adjuvant chemotherapy if terminated at least 12 months before protocol entrance. Patients with a previous malignant disease or serum bilirubin ]40 mmol:L were also excluded. Pretreatment evaluation included full history, physical examination, CT or abdominal ultrasound (UL), chest x-ray and measurement of the lesion(s) chosen for evaluation. The rst phase II trial included 77 patients. The treatment, which lasted 28 days, included UFT in a daily dose of 300 mg:m and l-leucovorin 22.5 mg:m. Both drugs were divided into three daily doses (q 8 h) and hydroxyurea (HU) in a xed daily dose of 0.5 g. The treatment period was followed by a 7-day rest period for a total cycle time of 35 days. Details of this trial have been presented previously (7). The second phase II trial included 63 patients treated with capecitabine in a dose of 2500 mg:m divided into two daily doses for a treatment period of 14 days followed by a 7-day rest period. The total cycle time was thus 21 days. The results are presented in Table 1. In the two trials 63 (82%) and 52 (83%) patients, respectively, were evaluable for response, with an objective response rate of 23% and 22%. The table also indicates a similar progression-free interval of 6.8 and 4.7 months, and an almost identical median survival (11 and 10.6 months). Important side effects are listed in Table 2. We observed no signi cant myelotoxicity. Severe (grade 3) nausea and vomiting were seen in 4% and 2%, respectively, and severe diarrhoea (grade 3) occurred in 9% of patients in the UFT trial and in 3% of patients in the capecitabine trial. A major difference was seen with respect to hand–foot syndrome, which occurred in 54% of the patients in the capecitabine trial. This side effect was not observed