A Prediction Model for Successful Increase of Adalimumab Dose Intervals in Patients with Crohn's Disease: Secondary Analysis of the Pragmatic Open-Label Randomised Controlled Non-inferiority LADI Trial

Background In the pragmatic open-label randomised controlled non-inferiority LADI trial we showed that increasing adalimumab (ADA) dose intervals was non-inferior to conventional dosing for persistent flares in patients with Crohn's disease (CD) in clinical and biochemical remission. Aims To develop a prediction model to identify patients who can successfully increase their ADA dose interval based on secondary analysis of trial data. Methods Patients in the intervention group of the LADI trial increased ADA intervals to 3 and then to 4 weeks. The dose interval increase was defined as successful when patients had no persistent flare (>8 weeks), no intervention-related severe adverse events, no rescue medication use during the study, and were on an increased dose interval while in clinical and biochemical remission at week 48. Prediction models were based on logistic regression with relaxed LASSO. Models were internally validated using bootstrap optimism correction. Results We included 109 patients, of which 60.6% successfully increased their dose interval. Patients that were active smokers (odds ratio [OR] 0.90), had previous CD-related intra-abdominal surgeries (OR 0.85), proximal small bowel disease (OR 0.92), an increased Harvey-Bradshaw Index (OR 0.99) or increased faecal calprotectin (OR 0.997) were less likely to successfully increase their dose interval. The model had fair discriminative ability (AUC=0.63) and net benefit analysis showed that the model could be used to select patients who could increase their dose interval. Conclusion The final prediction model seems promising to select patients who could successfully increase their ADA dose interval. The model should be validated externally before it may be applied in clinical practice.