Prediction of treatment failure during infliximab induction therapy in inflammatory bowel disease patients based on pharmacokinetic and pharmacodynamic modeling

Background: In infliximab (IFX) treatment for Crohn's disease (CD) and ulcerative colitis (UC), it is difficult to predict treatment failure during the induction phase. In the present study for optimal IFX treatment, we attempted to estimate serum IFX concentration and clinical response in individual patients during the induction phase to predict the indication of therapeutic effect and the possibility of treatment failure in the maintenance phase. Methods: We estimated pharmacokinetic and pharmacodynamic (PK/PD) parameters and predicted the serum IFX concentration and clinical response using a PK/PD model and Markov chain Monte Carlo Bayesian analysis method during the induction phase. Then, we determined whether the indication of therapeutic effect between predicted and observed clinical response were matched during the maintenance phase. Results: Data obtained from 15 patients were analyzed. The correlation between predicted and observed values of serum IFX concentration (Pearson product-moment correlation coefficient, 0.700; P < 0.0001, n = 68) and clinical response of CD patients (0.790; P < 0.0001, n = 25) and UC patients (0.702; P = 0.0004, n = 21) were significantly high. The indication of therapeutic effect at the final time point of each patient (from day 115 to day 203) were successfully predicted in 14 of 15 patients (93.3%). Conclusions: This study presents prediction of serum IFX concentration and clinical response in individual patients during induction therapy, with presumption of the indication of therapeutic effect and the treatment failure in the maintenance phase. Our results show the possibility of optimizing IFX therapy during the induction phase. © 2020 Elsevier B.V.