A method for simultaneous quantitation of both the acid and lactone forms of atorvastatin, a new synthetic inhibitor of HMG-CoA reductase that is being marketed for the treatment of high serum cholesterol, and both the acid and lactone forms of its two biotransformation products, 2-hydroxyatorvastatin and 4-hydroxyatorvastatin, in human serum (a total of six analytes) by high-performance liquid chromatography with electrospray tandem mass spectrometry was developed and validated. A deuterium labeled analog was used as internal standard for each of the six analytes. Each point of the calibration standard curve, which ranged from 0.5 to 200 ng/mL, contained the six analytes at equal concentrations. Three groups of quality control (QC) samples were used. In the first group, combination QC samples contained all six analytes at equal concentrations. In the second group, acid-only QC samples contained only the acid forms (i.e. three analytes) at equal concentrations. In the third group, lactone-only QC samples contained only the lactone forms (i.e. three analytes) at equal concentrations. After adding the internal standard to 0.5 mL of each standard and the QC sample kept at 4 degrees C, the samples were acidified with sodium acetate buffer (pH 5.0) and then extracted with methyl tert-butyl ether. Detection was by positive ion electrospray tandem mass spectrometry using eight selected reaction monitoring channels. The acid compounds were stable in human serum at room temperature but the lactone compounds were unstable as they hydrolyzed rapidly to their respective acid forms. The conversion of the lactone compounds in both QC and post-dose human serum samples was nearly complete after 24 h at room temperature. The lactone compounds in serum could be stabilized by lowering the working temperature to 4 degrees C or lowering the serum pH to 6.0. The acid-only and the lactone-only QC samples showed that, under the sample processing conditions used, the degree of the hydrolysis of the lactone compounds or the lactonization of the acid compounds during the assay procedure was minimal (< 5%). The intra-day C.V., inter-day C.V. and the deviations from the nominal concentrations for all six analytes were within 15%, demonstrating good precision and accuracy. The required lower limit of quantitation (LLQ) of 0.5 ng/mL was achieved for each analyte.