The steady-state pharmacokinetics and tolerability of a microemulsion formulation of cyclosporine (Sandimmune Neoral) were compared with Sandimmune in 18 clinically stable renal allograft recipients. In study period I (2 weeks duration), patients entered the study on a stable, individualized twice-daily dosage regimen of Sandimmune. Two approaches were assessed for changing patients over from Sandimmune to Sandimmune Neoral. In period II (2 weeks), doses were converted based on the area under the curve ratio derived from a relative bioavailability study comparing the two formulations in healthy volunteers. In period III (2 weeks), doses were titrated to provide comparable steady-state trough concentrations as at study entry. Sandimmune was reinstituted during period IV (2 weeks). Safety and tolerability were assessed at weekly clinic visits and the steady-state pharmacokinetics of cyclosporine in whole blood were characterized at the end of each study period. Dose conversion in period II based on the AUC ratio derived from healthy volunteers was inadequate for achieving comparable cyclosporine exposure as assessed by steady-state AUC and troughs. The concentration-controlled approach (period III) indicated that maintaining the same cyclosporine dose when changing between formulations yields comparable steady-state trough concentrations. Concomitant with this conversion, steady-state peak concentration and AUC increased on average by 39% and 15%, respectively, due to absorption-related differences between the formulations. These increases were not associated with adverse events or changes in blood pressure or clinical laboratory parameters. Furthermore, they were not detrimental to the transplanted kidney as monitored by ultrasound examination. The pharmacokinetic profiles from Sandimmune Neoral exhibited less variability and yielded a stronger correlation between trough concentration and systemic exposure (AUC) compared with Sandimmune.