Models for the prediction of mycophenolic acid area under the curve using a limited-sampling strategy and an enzyme multiplied immunoassay technique in chinese patients undergoing liver transplantation

Abstract

Background: An enzyme multiplied immunoassay technique (EMIT) provides convenient and accurate measurements of mycophenolic acid (MPA) concentrations for determination of immunosuppression during treatment with mycophenolate mofetil (MMF). No abbreviated model for estimating the full 12-hour MPA AUC using an EMIT assay in liver transplant recipients has been described previously. Objective: This study was conducted to determine the best model for predicting the MPA AUC using the EMIT method and a limited-sampling strategy in Chinese patients undergoing liver transplantation. Methods: The study enrolled consecutive liver transplant patients who were receiving MMF 1 g BID along with tacrolimus. A complete MPA pharmacokinetic profile was obtained for each patient on a single day, 7 to 14 days after transplantation. The EMIT method was used to determine MPA concentrations before dosing and at 0.5, 1, 1.5, 2, 4, 6, 8, 10, and 12 hours after dosing on the sampling day. Multiple linear regression analysis was used to evaluate potential models for estimating the full 12-hour MPA AUC. The accuracy and robustness of the models were evaluated using bootstrap analysis. Prediction error and prediction bias were calculated. Agreement between the estimated MPA AUC 0-12 and the full 12-hour MPA AUC was investigated using Bland-Altman analysis. Results: The study enrolled 48 Chinese liver transplant recipients (45 male, 3 female) with a mean (SD) age of 50 (12) years, mean weight of 64 (12) kg, and mean height of 169 (6) cm. Twenty-four models that included blood sampling at 1 through 4 time points were developed ( r 2 = 0.015–0.950). Four models with the highest r 2 values were selected; the lack of significant differences from the original dataset on bootstrap analysis indicated acceptable accuracy and robustness. The best model for predicting the MPA AUC 0-12 employed concentrations at 1, 2, 4, and 8 hours; 40 of 48 (83.3%) MPA AUC 0-12 values estimated using this model were within 15% of the full 12-hour MPA AUC. This model had a minimal mean prediction error (mean [SD], 0.27% [1.79%]) and mean absolute prediction error (8.83% [1.24%]). On Bland-Altman analysis, this model also had the best agreement between the estimated MAP AUC 0-12 and the full 12-hour MPA AUC, with a mean error of 9.02 mg · h/L. Conclusions: In this small group of Chinese liver transplant patients receiving MMF and concomitant tacrolimus, models for estimating the MPA AUC 0-12 were developed using the EMIT method and a limited-sampling strategy. The best model for prediction of the full 12-hour MPA AUC was 4.46 + 0.81 · C1 + 1.78 · C 2+2.51·C 4+4.94·C8.