Real-world success of de-escalating upadacitinib induction dosing in ulcerative colitis is predicted by steroid-free clinical and biochemical remission, not induction duration
Raj M Dalal, Sarah Rotondo-Trivette, Alec R Mason, Lynn Leveille, Jianyi Yin, Chelsea Missi, Moheb Boktor, David I FudmanAbstract
Background
Data guiding upadacitinib dose de-escalation in ulcerative colitis (UC) are limited. We evaluated factors associated with UC-related complications (UCRC) after upadacitinib was reduced from 45 mg daily.
Methods
This retrospective study included adults with UC at a tertiary center who underwent upadacitinib de-escalation (4/2022-11/2025). The primary outcome was a 12-month composite of UCRC (upadacitinib re-escalation or discontinuation, new systemic steroids, new advanced therapy, hospitalization, or colectomy). Predictors of UCRC after de-escalation were assessed using Cox proportional hazards modeling, including steroid-free clinical remission (SFCR), defined as a stool frequency subscore of 0 or 1 and a rectal bleeding subscore of 0 without systemic steroid use, and biochemical remission, defined as a fecal calprotectin (FC) < 150 μg/g or a C-reactive protein (CRP) < 5 mg/L when FC was unavailable. Optimal CRP and FC cut points were identified via ROC curves.
Results
Among 70 patients, 25 (35.7%) experienced UCRC (median follow-up 56 weeks). Induction duration, prior advanced therapy exposure, and endoscopic severity did not differ significantly between those with and without UCRC. Patients without UCRC were more likely to have achieved SFCR (51.1% vs 20.0%, P=.01) and biochemical remission (53.3% vs 28.0%, P=.04) prior to de-escalation, with lower median CRP (0.8 vs 1.5 mg/L, P=.04) and FC (79.5 vs 208.0 μg/g, P=.03). In multivariable analysis, achieving SFCR (aHR 0.33, 95% CI, 0.12-0.91) and biochemical remission (aHR 0.34, 95% CI, 0.13-0.85) independently predicted lower UCRC risk, and achieving both conferred additional UCAE-free survival benefit. Optimal cut points predicting UCRC were CRP > 12.7 mg/L (sensitivity 40%, specificity 93%) and FC > 58.7 μg/g (sensitivity 100%, specificity 47%).
Conclusions
SFCR and biochemical remission prior to upadacitinib de-escalation, but not induction duration, predict lower UCRC risk. Upadacitinib de-escalation should be guided by response assessment rather than a predetermined induction duration.