Pharmacokinetic studies of cyclosporine in Oriental kidney transplantation patients

Abstract

SUMMARY: Pharmacokinetic studies were performed in 10 stable kidney transplantation patients who received microemulsion formulation (Neoral®) of cyclosporine A (CsA) twice daily. No agents having pharmacokinetic effects on CsA had been used in these patients. The values of various basic pharmacokinetic parameters were similar to those reported in Western literature. The complete area under the blood concentration–time curve (AUC) of CsA for the duration of 12 h (12‐h AUC) was determined using the linear trapezoidal rule from seven concentrations at 0, 1, 2, 4, 6, 8, and 12 h after CsA administration. The mean values of 12‐h AUC were 4603.63 ± 344.61 ng h/mL. CsA concentrations at 2 h after dosing (not the trough levels) showed the best correlation with the complete AUC (r2 = 0.9322). The abbreviated AUC of CsA was calculated either by stepwise multiple linear regression analysis or by the linear trapezoidal rule from a few sampling time points. Using stepwise multiple linear regression analysis, which was used in calculating abbreviated AUC in all previous studies, the model equation that had the highest correlation and the lowest prediction error with the complete AUC was derived by using CsA concentrations at 2 and 8 h after dosing (12‐h AUC = 4.262C2 + 8.390C8− 669.417; r2 = 0.9808, absolute prediction error = 3.97 ± 0.96). Two model equations derived using the linear trapezoidal rule provided the best correlation with the complete AUC: (1) The two time points selected model equation 12‐h AUC = 4C2 + 5C8; r2 = 0.9780, absolute prediction error = 6.41 ± 1.22). (2) The three time points selected model equation 12‐h AUC = 4C0 + 3C2 + 5C6; r2 = 0.9475, absolute prediction error = 5.00 ± 1.41). When different pharmacokinetic data sets were applied to the model equations derived using regression analysis, the values of coefficients and the constant of the regression equation had changed from the initial equation. Thus, new model equations will emerge every time the new data are applied. In contrast, the values of coefficients in the model equation calculated using trapezoidal rule were unaltered when tested by the new pharmacokinetic data set. Thus, the abbreviated AUC derived using the linear trapezoidal rule would be simpler than and superior to that obtained using stepwise multiple linear regression analysis in prediction of the complete AUC.