Background: Infliximab (IFX) trough levels vary markedly between patients with inflammatory bowel disease (IBD), which is important for clinical response. The aim of this study was to evaluate the performance of previously developed population pharmacokinetic models in patients with IBD for dose individualization for Crohn disease (CD) and ulcerative colitis in our clinical setting.
Methods: The authors collected 370 trough levels prospectively from 100 adult patients with IBD who were undergoing IFX treatment between July 2013 and August 2016. The external evaluation included prediction- and simulation-based diagnostics [prediction-corrected visual predictive check, prediction- and variability-corrected visual predictive check, and normalized prediction distribution error tests].
Results: In prediction-based diagnostics, the authors observed a nonsignificant overall mean relative bias of -6.87% and an acceptable imprecision of 8.45%. Approximately 100% of the prediction error was within ±30%, indicating satisfactory predictability. Simulation-based diagnostics indicated model misspecification; thus, the model may not be appropriate for simulation-based applications.
Conclusions: While simulation-based diagnostics provided unsatisfactory results, the prediction-based diagnostics demonstrate that the population pharmacokinetic model developed by Fasanmade et al for CD can be used to predict and design individualized IFX dose regimens that meet the individual needs of patients with CD and ulcerative colitis.