BACKGROUND We developed and compared 2 different methods for quantifying uracil (U) and dihydrouracil (UH(2)) in BSA and human plasma. Special attention was paid to the selectivity/specificity and the absence of a matrix effect. The UH(2)/U ratio is intended as a biomarker to identify patients with deficiency in 5-fluorouracil metabolism. METHODS We quantified U and UH(2) with 2 liquid chromatography methods after solid-phase extraction, one with UV detection (LC-UV) and the other with mass spectrometric detection (LC-MS). We selected 2 internal standards to prevent the risk of interferences. Separation was achieved with a Waters Atlantis dC18 column (LC-MS) or a Waters SymmetryShield RP18 column connected with an Atlantis dC18 (LC-UV). Mass spectrometric data were acquired in single-ion monitoring mode. RESULTS Assay imprecision in BSA solution was <15% (LC-UV) and <12% (LC-MS); in plasma, assay imprecision was <9.5% and <9.0%, respectively. Recoveries were 88.2%-110% (LC-UV) and 94.8%-107% (LC-MS). Extraction efficiencies were >or=89.0%. In BSA, the lower limits of quantification for U and UH(2) were 2.5 microg/L and 6.25 microg/L, respectively, for the LC-UV method and 2.5 microg/L and 3.1 microg/L for LC-MS. The corresponding values in plasma were 11.6 microg/L and 21.5 microg/L, and 4.1 microg/L and 12.1 microg/L. CONCLUSIONS To estimate endogenous U and UH(2) concentrations and their ratio, we recommend the use of a drug-free human plasma pool in which baseline U and UH(2) concentrations have previously been measured with the standard-addition method. Our LC-MS method, which has the better test performance and is useful for measuring UH(2)/U ratios in cancer patients, is preferred when this equipment is available.