To develop limited-sampling strategy (LSS) models for estimating prednisolone's area under plasma concentration versus time curve (AUC(0-infinity)), its maximum concentration in plasma (C(max)), and total clearance (CL/F). Healthy subjects (n = 24), enrolled in a bioequivalence study, received 20 mg PO of the prodrug prednisone as reference and test tablets, and plasma prednisolone concentrations (n = 576) were measured by a validated HPLC assay. A linear regression analysis of AUC(0-infinity), C(max), CL/F, and log(CL/F) against the plasma prednisolone concentrations for the reference formulation was carried out to develop LSS models to estimate these parameters. The LSS models were validated on the test formulation data sets and on simulated sets generated by the software ADAPT II. LSS models based on a single [1.5 hours for C(max) and 7 hours for AUC(0-infinity), CL/F, and log(CL/F)] plasma sample, accurately estimated (R2 = 0.84-0.97, mean bias < 1%; mean precision < 10%) these pharmacokinetic parameters. Validation tests indicated that the most informative single-point LSS models developed for the reference formulation provide precise estimates (R(2) > 0.83; mean bias < 3%; mean precision < 10%) of the corresponding pharmacokinetic parameters for the test formulation. LSS models based on the two most informative sampling points (1.5 and 7 hours) were required for accurate estimates (R(2) > 0.87; mean bias < 6%; mean precision < 8%) of prednisolone's C(max), AUC(0-infinity), CL/F, and log(CL/F) for the simulated data sets. Finally, bioequivalence assessment of the prednisone formulations, based on LSS-derived AUC(0-infinity) and C(max) values provided results identical to those obtained using the original values for these parameters. One- and 2-point LSS models provided accurate estimates of prednisolone's C(max), AUC(0-infinity), and CL/F, following single oral doses of prednisone, and allowed correct assessment of bioequivalence between two prednisone formulations.