Katherine A Barraclough1, Nicole M Isbel, Christine E Staatz. 1. Department of Renal Medicine, University of Queensland at the Princess Alexandra Hospital, Woolloongabba, Brisbane, Australia. katherine_barraclough@health.qld.gov.au
Abstract
BACKGROUND: Three randomized-controlled trials have examined the utility of mycophenolic acid (MPA) therapeutic drug monitoring in kidney transplant recipients. Each used a different methodology to estimate MPA exposure (predose concentrations [C(0)], multiple regression-derived limited sampling strategies [LSSs], and Bayesian estimation in Opticept, FDCC [fixed dose versus concentration controlled] and APOMYGERE [adaption de Posologie du MMF en Greffe Renale] trials, respectively). Results were conflicting. This study aimed to compare the ability of these methodologies to predict AUC(0-12) in an independent cohort of kidney transplant recipients. METHODS: Sixty-nine full AUC (AUCf) profiles were collected from 45 subjects. C(0) and AUC(0-12) predicted by the LSSs and Bayesian estimation (AUCp) were compared with AUCf. RESULTS: Correlation between C(0) and AUCf was poor. There was better correlation between LSSs and AUCf, and correlation was higher again between Bayesian estimates and AUCf. Bias and precision associated with the LSS and Bayesian-forecasting methods fell close to the level considered acceptable in clinical studies (<15%). However, neither method quite achieved this, and regardless of whether LSSs or Bayesian forecasting were applied, only 45% to 47% of AUCp values fell within 15% of AUCf. Discordance between C(0) and AUCp and AUCf occurred in 15% to 27% of cases. CONCLUSIONS: This study highlights the difficulties of MPA therapeutic drug monitoring. Bayesian estimations showed slight superiority over C(0) and LSS predictions, suggesting that this may be the preferable methodology.
BACKGROUND: Three randomized-controlled trials have examined the utility of mycophenolic acid (MPA) therapeutic drug monitoring in kidney transplant recipients. Each used a different methodology to estimate MPA exposure (predose concentrations [C(0)], multiple regression-derived limited sampling strategies [LSSs], and Bayesian estimation in Opticept, FDCC [fixed dose versus concentration controlled] and APOMYGERE [adaption de Posologie du MMF en Greffe Renale] trials, respectively). Results were conflicting. This study aimed to compare the ability of these methodologies to predict AUC(0-12) in an independent cohort of kidney transplant recipients. METHODS: Sixty-nine full AUC (AUCf) profiles were collected from 45 subjects. C(0) and AUC(0-12) predicted by the LSSs and Bayesian estimation (AUCp) were compared with AUCf. RESULTS: Correlation between C(0) and AUCf was poor. There was better correlation between LSSs and AUCf, and correlation was higher again between Bayesian estimates and AUCf. Bias and precision associated with the LSS and Bayesian-forecasting methods fell close to the level considered acceptable in clinical studies (<15%). However, neither method quite achieved this, and regardless of whether LSSs or Bayesian forecasting were applied, only 45% to 47% of AUCp values fell within 15% of AUCf. Discordance between C(0) and AUCp and AUCf occurred in 15% to 27% of cases. CONCLUSIONS: This study highlights the difficulties of MPA therapeutic drug monitoring. Bayesian estimations showed slight superiority over C(0) and LSS predictions, suggesting that this may be the preferable methodology.