The bioavailability of two oral formulations of cyclosporin A (Sandimmun and Neoral) was assessed in 10 children with end-stage renal disease (ESRD), while at a steady state on a dialytic procedure. The study was performed according to a randomized, double blind, cross-over design, allowing a 1-month washout period between studies. Each patient received 2.5 mg/microg of oral cyclosporin A every 12 h, either Sandimmun (SAN) or Neoral (NEO). Serum concentrations of cyclosporin A were determined serially during a 24 h period, after the 5th dose of cylosporin. Serum concentrations against time curves were constructed and bioavailability of both medications, expressed as AUC and Cmax, were compared. A statistically significant increase was observed in the AUC and Cmax of NEO, which were 90% and 130% higher, respectively, than those of SAN. Considering that the internationally accepted criteria for bioequivalence allows a 20% variation in AUC and Cmax, it appears that Neoral and Sandimmun do not bear bioequivalence in children with ESRD. Notwithstanding, there were no significant differences in trough levels between both formulations. We conclude that, if trough levels are the only source of information for dosing design, Neoral could be substituted for Sandimmun on a 1:1 basis. However, a 1:1 drug substitution is not suitable when AUC is used in children with ESRD.