Abstract
BACKGROUND & AIMS: Infliximab is an established rescue therapy for patients with steroid-refractory acute severe ulcerative colitis (ASUC), yet optimal dosing strategies minimizing colectomy risk remain unclear. We aimed to develop a model-informed risk stratification algorithm to identify patients at high risk of colectomy within 90 days of initiating infliximab to support personalized dosing.
METHODS: We conducted a multicenter, retrospective population pharmacokinetics (popPK) and exposure-response study using data from patients with ASUC. A parametric time-to-event model was developed to characterize the 90-day colectomy risk. Patient characteristics and pharmacokinetic projections were evaluated as predictors. These modelling results informed the development of an algorithm for risk stratification and personalized infliximab rescue dosing.
RESULTS: Seven medical centers contributed data from 72 patients with ASUC, yielding a total of 152 infliximab serum concentrations. Eleven patients underwent colectomy within 90 days. The strongest predictor of colectomy was the clearance-normalized exposure between weeks 2 and 4 (area under the concentration-time curve, AUCw2-4 to Bayesian-forecasted infliximab clearance, CL), with an area under the receiver operating characteristic curve of 0.79 (95% confidence interval [CI], 0.52-1.00). The AUCw2-4/CL ratio was calculated by individualizing the popPK model using the patient's body weight, baseline C-reactive protein, and infliximab concentrations. Patients with a log-transformed AUCw2-4/CL ratio < 5.79 were classified as high risk for colectomy (sensitivity 83%, specificity 85%). Overall classification accuracy was 85% (95% CI, 74-92).
CONCLUSIONS: We developed a model-based dose-exposure-response framework to predict colectomy risk in ASUC. We integrated the algorithm into an interactive tool to enable individualized infliximab rescue therapy.