UBC Theses and Dissertations
Developing a limited sampling strategy for cyclosporine area under the curve monitering in lung transplant patients Dumont, Randall John
This study developed a limited sampling strategy (LSS) to provide an estimate of cyclosporine (Neoral®) area-under-the-curve (AUC) in lung transplant recipients, a population for which a cyclosporine LSS has not yet been delineated. The predictive performance of the LSS, and other published ..LSS in other transplant types, was evaluated. Finally, the pharmacokinetic parameters of the lung transplant patients were calculated. Fourteen stable lung transplant patients (n = 7 male; n = 7 female) were entered into the study. Upon administration of a steady-state morning cyclosporine dose, blood samples were collected at 0, 1, 2, 3, 4, 5, 6, 8, 9, 10, and 12 hours post-dose in 12 patients (and up to 8 hours post-dose in 2 patients on a q8h regimen). Blood samples were analyzed by monoclonal fluorescence polarization immunoassay. AUC was calculated by the linear trapezoidal method, and the LSS was calculated using multiple regression analysis. Predictive performance was evaluated using methods proposed by Sheiner and Beal. Pharmacokinetic analysis was performed using WinNonlin® computer software. Patient characteristics (mean + SD) are as follows: age: 48 ± 12 years; weight: 69 ± 17 kg; transplant type: 6 double lung, 8 single lung; total daily cyclosporine dose: 4.3 ± 1.7 mg/kg; time post-transplant: 5.1 + 3.4 years. Eight patients were used to determine the LSS. Analysis of all available concentration-time data revealed the following equation: AUC = 17.24xC6 - 58.96xC8 + 23.39xC9 + 52.29xC12 - 796.07, r² = 0.999. In order to provide a clinically feasible LSS, the remainder of the analysis was restricted to the data collected in the first 3 hours post-dose. One 4-point, four 3-point, six 2-point, and four 1-point equations were determined. On the basis of the number of samples required, the coefficient of determination, comparison of predictive performance, and the percent prediction error in AUC estimation (%pe), we selected the following equation for analysis of predictive performance: AUC = 1.46xCl + 5.36xC3 + 274.49; r² = 0.975; %pe (range) = -4.47 - 8.47%. For this LSS, mean prediction error (ME, bias) was 195 ngxhr/mL, and mean absolute error (MAE, precision) was 299 ngxhr/mL. There was no significant difference in predictive performance between the LSS for lung transplant patients and other published LSS in other transplant types, with 5 exceptions. This 2-concentration LSS for lung transplant patients was significantly more biased than a 3-concentration LSS developed for renal transplant patients, and was significantly less biased and significantly more precise than 2 other LSS that were developed for renal transplant patients. The best clinically feasible LSS for cyclosporine AUC estimation requires 2 concentrations drawn at 1 and 3 hours post-dose.
Item Citations and Data