OBJECTIVE The purpose of this study was to evaluate the bioequivalence of two formulations of atorvastatin using the reference-scaled average bioequivalence (RSABE) method and to study the pharmacokinetics of atorvastatin in healthy Chinese subjects under fed conditions. MATERIALS AND METHODS A single-dose, randomized, open-label, four-way crossover study was conducted in healthy Chinese subjects after informed consent was obtained. Healthy subjects were randomly assigned to receive 20 mg of either the test or reference formulation, following a 7-day washout period. The formulations were considered bioequivalent if 90% confidence intervals (CIs) for the ln-transformed ratios and ratio of geometric means (GMR) of AUC and Cmax of atorvastatin were within the bioequivalence range (80 - 125%). Plasma atorvastatin, ortho-hydroxy atorvastatin and para-hydroxy atorvastatin concentrations were analyzed by liquid chromatography-tandem mass spectrometry. Tolerability was assessed during the entire study period. RESULTS ANOVA indicated that the period, sequence, and formulation had no significant effect on the pharmacokinetic parameters (p < 0.05). The test formulation was bioequivalent to the marketed formulation as the 90% CIs for natural log-transformed ratios of atorvastatin of Cmax (88.45 - 103.57%), AUC0-t (98.08 - 104.89%) and AUC0-∞ (98.15 - 104.87%) were within equivalence limits (80 - 125%). No serious adverse events were found among the subjects. CONCLUSION The RSABE approach was successful in evaluating the bioequivalence of these two formulations. This study confirmed that test and reference atorvastatin calcium tablets were bioequivalent under fed condition.