The bioequivalence of two megestrol acetate formulations, 160-mg "tablets" and 160-mg "sachets," was investigated in a single-dose, open-label, balanced-for-sequence cross-over study involving 12 advanced-cancer patients. The observed plasma megestrol-acetate time course obtained with both formulations was consistent with the literature data. The main source of variability in the pharmacokinetic parameters was intersubject variability; drug formulation played only a minor (and nonsignificant) role. The width of the 90% confidence interval of the area-under-the-curve (AUC) ratio (sachets: tablets) computed according to Schuirmann (0.9-1.4) was mainly due to the presence of a single outlier, showing an AUC ratio of 2.7. The trend to higher bioavailability of the new formulation was not significant, especially as compared with the dose-response data reported in the literature.