Selective, accurate, and reproducible liquid chromatography/tandem mass spectrometry (LC/MS/MS) methods were developed and validated for the determination of mevalonic acid, an intermediate in the biosynthesis of cholesterol and therefore a useful biomarker in the development of cholesterol lowering drugs, in human plasma and urine. A hepta-deuterated analog of mevalonic acid was used as the internal standard. For both methods, calibration standards were prepared in water, instead of human plasma and urine, due to unacceptably high levels of endogenous mevalonic acid. The lower quality control (QC) samples were prepared in water while the higher QC samples were prepared in the biological matrices. For the isolation/purification of mevalonic acid from the plasma and urine matrices, the samples were first acidified to convert the acid analyte into its lactone form. For the plasma samples, the lactone analyte was retained on and then eluted off a polymeric solid-phase extraction (SPE) sorbent. For the urine method, the sample containing the lactone analyte was passed through a C-18 SPE column, which did not retain the analyte, with the subsequent analyte retention on and then elution off a polymeric SPE sorbent. Chromatographic separation was achieved isocratically on a polar-endcapped C-18 analytical column with a water/methanol mobile phase containing 0.5 mM formic acid. Detection was by negative-ion electrospray tandem mass spectrometry. The standard curve range was 0.500-20.0 ng/mL for the plasma method and 25.0-1,000 ng/mL for the urine method. Excellent accuracy and precision were obtained for both methods at all concentration levels tested. It was interesting to note that for certain batches of urine, when a larger sample volume was used for analysis, a high degree of matrix effect was observed which resulted not only in the attenuation of the absolute response, but also in a change of analyte/internal standard response ratio. This demonstrated that, under certain conditions, the use of a stable isotope analog internal standard does not, contrary to conventional thinking, guarantee the constancy of the analyte/internal response ratio, which is a prerequisite for a rugged bioanalytical method. On the other hand, under conditions where the sample matrix does not have such a deleterious effect, we have found that a stable isotope analog could serve as a surrogate (substitute) analyte. Thus, we have shown that using calibration standards prepared by spiking plasma with tri-deuterated or tetra-deuterated mevalonic acid, instead of mevalonic acid itself (the analyte), plasma QC samples that contain mevalonic acid can be successfully analyzed for the accurate and precise quantitation of mevalonic acid. The use of a surrogate analyte provides the opportunity to gauge the daily performance of the method for the low concentration levels prepared in the biological matrix, which otherwise is not achievable because of the endogenous concentrations of the analyte in the biological matrices. Copyright 2003 John Wiley & Sons, Ltd.