Vedolizumab (VDZ) is an effective therapy for the management of patients with moderately to severely active ulcerative colitis (UC) or Crohn’s disease (CD) who have failed conventional therapy with aminosalicylates, corticosteroids, and thiopurines, as well as biologic therapy with tumor necrosis factor (TNF) antagonists. Several studies have identified potential predictors of treatment outcomes; however, the optimal approach to integrating predictors into routine practice is uncertain.No prior decision support tools exist to predict VDZ drug exposure in UC and CD and link this back to differences in effectiveness or response to VDZ dose escalation. By having a tool that can predict at baseline prior to start of therapy whether VDZ will be effective and what a patients drug exposure profile will be with VDZ, the provider can 1) determine if VDZ is an appropriate therapy to begin, 2) proactively monitor those patients deemed high risk for treatment failure with VDZ, and 3) proactively measure drug concentrations for VDZ to then increase the dose or the interval at which VDZ is administered to improve outcomes.
Researchers at UC San Diego have addressed this gap by deriving multivariable logistic regression prediction models and clinical design support tools (CDST) using the GEMINI clinical trial data sets for the outcomes of corticosteroid-free clinical and endoscopic remission. The correlation between variability in VDZ exposure and differences in efficacy across predicted probability groups was explored in the derivation set, and the CDST was subsequently validated in an external cohort of UC and CD patients treated with VDZ in routine practice. The intent was to create a CDST that will help clinicians optimize VDZ therapy for individual patients.
The models give a score and thus predicts how an UC or CD patient will respond to VDZ, how quickly they will respond, and it predicts what the individual patient’s drug exposure profile will be during therapy and whether they would benefit from VDZ dose or interval changes.
The scores put the patients into a high, intermediate, or low probability of response to VDZ category. Patients in the high probability group are predicted to have higher drug exposure and can be monitored less frequently with no drug concentration testing. Patients in the intermediate or low probability group will have incrementally lower drug exposure and efficacy. In these patients providers can have the patient return for follow-up sooner, check drug concentrations, and increase the dose of VDZ or the interval of VDZ administration until they achieve a concentration consistent with the concentrations achieved in the high probability of response group.
The models performance in clinical practice was confirmed through an external validation in a large multi-center cohort of VDZ treated UC and CD patients.
This technology is patent pending and available for licensing and/or research sponsorship.
Biomarker, prediction model, therapeutic drug monitoring, ulcerative colitis, vedolizumab