A sensitive, specific, accurate and reproducible analytical method was developed and validated to quantify pravastatin (Prav), pravastatin-d5 (Prav-d5), SQ-31906, SQ-31906-d5, and pravastatin lactone (Prav-Lac) in human serum samples. Serum samples (0.5 ml) were acidified and extracted by a solid-phase extraction procedure to isolate all five analytes from human serum. Sample extracts were reconstituted and analyzed by turbo ion spray liquid chromatography/tandem mass spectrometry (LC/MS/MS) in the positive ion mode. The total run time was 9 min between injections. The assay demonstrated a lower limit of quantitation (LLQ) of 0.5 ng/ml for all five analytes. The calibration curves were linear from 0.5 ng ml to 100 ng/ml for all five analytes. The coefficients of determination of all calibration curves were > or = 0.999. Precision and accuracy quality control (QC) samples were prepared at concentrations of 2, 30, 80, and 500 ng/ml for all analytes. The intra-assay and inter-assay precision calculated from QC samples were within 8%, for all analytes. The inter-assay accuracy calculated from QC samples was within 8% for all analytes. The extraction recoveries were > or = 90% for all analytes. Benchtop stability experiments in an ice-water bath ( < or = 10 degrees C) demonstrated that over time, Prav-Lac hydrolyzes to Prav in serum. Prav, Prav-d5, SQ-31906, and SQ-31906-d5 were stable under these conditions for up to 24 h. Hydrolysis was minimized by buffering the serum to pH 4.5 and maintaining the serum sample in an ice-water bath. All analytes were stable after three freeze/thaw cycles and in reconstitution solution after 1 week at 4 degrees C. Stability of all analytes in human serum was demonstrated after storage at -70 degrees C for 77 days. The benchtop (< or = 10 degrees C) stability of pooled study samples was also investigated and the results were comparable to those obtained from serum QC samples.