BACKGROUND Ustekinumab was recently approved for the treatment of moderate to severe Crohn's disease (CD). Although the ustekinumab Clinical Decision Support Tool (UST-CDST) was able to predict ustekinumab responsiveness in a clinical trial, it is not clear whether UST-CDST can also predict a future clinical relapse following ustekinumab therapy in the real-life setting.

METHODS We enrolled patients with moderate to severe CD who were refractory to conventional therapies and who showed a clinical response after induction therapy with ustekinumab and monitored them until the relapse. We performed a Cox proportional hazard analysis to investigate the predictive capability of UST-CDST for a clinical disease relapse.

RESULTS Clinical remission rates at week 20 were 25.0% for low-probability responders, 66.7% for intermediate-probability responders, and 75.0% for high-probability responders. The high-probability responders were more likely to achieve clinical remission at week 20 compared with the low-probability responders. Among 99 patients with moderate to severe CD, 37 (37.4%) experienced a clinical relapse during the median follow-up period of 18.0 months of ustekinumab treatment. The cumulative relapse rates were 70.0% in the low-probability responders, 35.9% in the intermediate-probability responders, and 22.5% in the high-probability responders (P = .001). In a multivariable Cox proportional hazard analysis, the high-probability responders and intermediate-probability responders had a lower risk of clinical relapse than the low-probability responders. Receiver operating characteristic analysis using UST-CDST to predict relapse revealed an area under the curve of 0.698.

CONCLUSIONS The UST-CDST can predict clinical relapse in patients with moderate to severe CD subjected to ustekinumab therapy.