Optimum sampling time for cyclosporin therapeutic drug monitoring

Abstract

Since dosage of Cyclosporin (CsA) is complicated by its pharmacokinetics variability, and by the narrow therapeutic window, attention to blood level is crucial for optimization of therapy. Trough level monitoring is generally used but previous studies suggested that AUC monitoring is superior. Although AUC could provide more precise information, it is expensive and time consuming. Thus a careful choice of only few sampling times is desired. The purpose of the study were to (1) determine the optimum sampling time point for predicting AUC by multiple linear regression and by trapezoidal rule. (2) evaluate the correlation between CsA level at different time points including AUC and dose. Pharmacokinetic studies were performed in 25 stable kidney transplant patients who received cyclosporin (CsA) microemulsion formulation twice daily. At least 3 days after constant dosage regimen were administered, whole-blood FPIA levels were obtained before morning dose of CsA and at 1, 2, 3, 4, 6, 8, and 12 hours after the dose. No single CsA concentrations, including trough levels, could provide a value of correlation coefficient (r2) of more than 0.9 with the complete AUC. The concentration at 2 hours after CsA dosing was the best single-level predictors of AUC (r2 = 0.8845). Stepwise multiple linear regression analysis revealed that the best predictive model for CsA AUC incorporated 2 levels at 2 and 6 hours post dost is 3.085*C2+6.019*C6+376.893 (r2 = 0.9638, mean absolute prediction error+_SE = 5.40+_0.88) while the best predictive model incorporated 3 levels at 1, 2, and 6 hours post dose is 0.738*C1+2.112*C2+7.02*C6+263.108 (r2 = 0.9823, mean absolute prediction error+_SE = 3.01+_0.81). Additional time points increased the accuracy only slightly. Our results also suggested that the measure AUC could be accurately estimated from using 3 sampling time points at 0, 2, and 6 hours post dose to calculate AUC by trapezoidal rule (r2 = 0.9695, mean absolute prediction error+_SE = 4.94+_0.81). Cyclosporin-sparing agents did not effect the optimum sampling time but did prolong the half-life, decrease clearance, increase C0/mg dose and increase Cmax/mg dose resulted in higher AUC/mg dose. Trough level, which is generally used for guide CsA dosing, showed poor correlation with CsA dosage. Our result suggested that either level at 2 hours post dose or AUC is more appropriate than trough level for CsA therapeutic drug monitoring because they displayed the best correlation with dose. Since the number of patients incorporated in this study was too small and the time of observation was not long enough, no significant relationship between blood level and clinical effects could be determined.