5-fluororuracil, (5-FU), like the majority of anti-cancer agents, has a narrow therapeutic index. Many toxic side effects, sometimes severe, are reported in patients treated in conventional, adjuvant or metastatic situations. These toxic side effects are due to an overdosage, consequence of the wide interpatient metabolism variability. 5-FU metabolism mainly depends on dihydropyrimidine dehydrogenase (DPD) activity that follows a gaussian curve in the general population and is submitted to a genetic polymorphism. Thus, patients presenting a dihydropyrimidine dehydrogenase deficiency have an increased risk with 5-FU and oral fluoropyrimidines, such as capecitabine and UFT, which may induce early severe toxicity. More than 30 SNPs (Single Nucleotide Polymorphisms) have been reported on the DPD gene in the literature. Some of them are silent, and about 15 those that are located at very important sites for enzyme activity, such as those coding for enzyme binding sites for the substrate, or cofactors such as NADH and FAD interfere with enzyme activity. The frequent use of fluoropyrimidines in anticancer treatments, their high dosage, the new indications and the severity of acute toxic side effects due to enzyme deficiencies, make their detection a medical and public health priority. Different approaches have been developed: pharmacogenetic, detecting SNPs on the DPD gene itself or its promoter; pharmacogenomic, measuring DPD mRNA expression in white blood cells; enzymatic or radioenzymatic, directly quantifying enzyme activity; pharmacologic, measuring uracil and/or thymine in plasma or urine; pharmacologic, simultaneously measuring plasma uracil and dihydrouracil, respectively substrate and metabolite of DPD, then determining UH2/U ratio; and pharmacologic, again, by individual 5-FU dose adjustment with pharmacokinetic monitoring. These approaches must respect certain criteria and requirements, such as sensitivity and specificity. According to our own experience with 1,500 patients, none of those approaches alone can make it. Moreover, they have to provide results, quickly enough not to delay the start of treatment. Thus, our pretheraputic DPD deficiency detection activity combines in practice quick and accurate techniques so as to give results within 8 to 10 days. Lastly, DPD deficiency detection is not limited to a simple result since most often it is a partial deficiency that does not contra-indicate 5-FU based treatment, provided that some added precautions are taken. Therapeutic advice given by pharmacologists can help the clinician to find the adequate 5-FU dosage. An individual dose adjustment by pharmacokinetic monitoring can be proposed to obtain the efficient dosage, manage the treatment, and avoid a toxic effect due to an overdosage or, on the contrary, a therapeutic failure by underdosage. [ABSTRACT FROM AUTHOR]