Summary Dihydropyrimidine dehydrogenase (DPD) is the rate-limiting enzyme for degradation of 5-fluorouracil (5-FU). DPD activity is highly variable in liver and peripheral mononuclear cells (PMNCs) and it has not been well studied in human tumours. Characterization of DPD in colorectal cancer is of clinical interest through its role in the regulation of 5-FU, the main chemotherapeutic agent used in this disease. Therefore, DPD activity was analysed in colorectal tumour and adjacent normal tissue from 63 patients, including three liver metastasis. DPD activity was highly variable in all tissues studied (coefficient of variation 43-61 %) and was higher in normal tissue than in tumour. The tumournormal activity ratio ranged from 0.19 to 3.32 (median 0.76). PMNC DPD activity was available for 57 patients and was correlated with tumour activity (r5 = 0.29, P < 0.001). A higher correlation was observed between PMNCs and tumour samples that were both obtained in the morning (r, = 0.49), consistent with circadian variation in DPD activity. Normal tissue DPD activity was not correlated with either tumour (r5 = 0.11) or PMNC activity (rs = -0.06). This study provides the first analysis of DPD activity in colorectal cancer and illustrates the large degree of variation in tumour activity. The tumour-normal activity ratio results suggest that elevated tumour DPD can play a role in 5-FU resistance through increased inactivation in tumour cells, but is an uncommon event in colorectal tumours. The results support the use of PMNCs for monitoring tumour DPD activity, particularly when circadian variation is taken into account. As a large degree of the variation in tumour DPD activity is not explained by PMNC activity, more accurate alternatives are needed before DPD activity can be used for targeting 5-FU therapy.Keywords: dihydropyrimidine dehydrogenase; enzyme activity; colorectal cancer; 5-fluorouracil; pharmacogenetics 5-Fluorouracil (5-FU) is commonly used in the treatment of gastrointestinal, head and neck, and breast tumours. 5-FU is itself inactive and requires intracellular conversion to form cytotoxic nucleotides. Several cellular targets for fluoropyrimidines have been well characterized, including inhibition of thymidylate synthase (TS) by FdUMP and false-base incorporation into RNA or DNA. Most investigations into cellular resistance factors regulating 5-FU activity have focused on alterations in TS levels and reduced folate pools, the required cofactor for binding dUMP to TS (Wang et al, 1993;Johnston et al, 1995). However, most of an administered 5-FU dose undergoes metabolism to an inactive species through a threeenzyme process, which is initiated and rate limited by dihydropyrimidine dehydrogenase (DPD; EC 1.3.1.2). After a bolus injection of 5-FU, 80% is degraded via DPD 24 h after administration (Heggie et al, 1987). Studies of 19F-NMR spectroscopy in mice bearing colon tumours found that catabolites made up 51% of labelled drug in the tumour, compared with 26% for the anabolic products (Kamm et al, 1994). Ther...