The cytotoxic effect of 5-fluorouracil (5-FU) is mediated by the inhibition of thymidylate synthase (TS), however, at the same time 5-FU is catabolized by dihydropyrimidine dehydrogenase (DPD). Efficacy of 5-FU may therefore depend on the TS and DPD activity and on pharmacogenetic factors influencing these enzymes. Our aims were (1) to determine the distribution of DPD activity, the frequency of DPD deficiency and the DPD (IVS14+1G>A) mutation in the peripheral blood mononuclear cells of colorectal cancer (CRC) patients, and study the relationship between DPD deficiency and toxicity of 5-FU; (2) to investigate the influence of TS polymorphisms and DPD activity on the survival of CRC patients receiving 5-FU-based adjuvant therapy. The frequency of DPD deficiency was determined by radiochemical methods in the peripheral blood mononuclear cells (PBMCs) of 764 CRC patients treated with 5-FU. The relationship between the TS polymorphisms, DPD activity and the disease-free and overall survival was studied in 166 CRC patients receiving 5-FU-based adjuvant therapy. TS polymorphisms were determined in the DNA samples separated from the PBMCs, by PCR-PAGE and PCR-RFLP-PAGE (restriction fragment length polymorphism) methods. Low DPD values (<10 pmol/min/106 PBMCs) were demonstrated in 160/764 patients (20.9%), and of those DPD deficiency (<5 pmol/min/106 PBMCs) was verified in 38 patients (4.9%). In the latter group severe (>Gr 3) toxicity was found in 87%. The prevalence of the DPD IVS14+1G>A mutation among the 38 DPD-deficient patients was 7.8% (3/38) and was accompanied by severe Gr 4 toxic symptoms (neutropenia, mucositis, diarrhea). TS polymorphisms showed a relationship with the survival of CRC patients. It is important to mention that by combining the 3-3 genotypes of 5'-TSER and 3'-TSUTR polymorphisms the obtained 8 genotype combinations showed significantly different Kaplan-Meier survival curves. The evaluation of these curves with Cox regression analysis resulted in two prognostically different groups: "A" good prognosis (RR<1) and "B" bad prognosis (RR>1). The disease-free- and overall survival of these two groups were significantly different. DPD activity also showed correlation with the survival; patients with DPD activity <10 pmol/min/106 PBMCs showed significantly longer disease-free and overall survival. The determination of DPD activity proved to be a more valuable parameter in the evaluation of serious 5-FU-related toxicity compared to the IVS14+1G>A mutation analysis. According to the Cox multivariate analysis the combination of germline TS polymorphisms and DPD activity is/an independent prognostic marker of survival in CRC patients treated with adjuvant 5-FU therapy.