P0225An Early Integrated Predictive Score for Steroid Refractoriness in Hospitalized Adult Patients with Acute Severe Ulcerative Colitis

Abstract Background One-third of patients with acute severe ulcerative colitis (ASUC) are refractory to intravenous glucocorticoids therapy (ivGC) and require salvage therapy or colectomy. Most available predictive scores rely on third-day data, while admission-based models have shown limited performance. We aimed to develop an early and simple predictive score based on routine clinical and laboratory data and compare it to existing models. Methods This retrospective, single-center study, included hospitalized patients with ASUC. The predictive model was developed using an early cohort (2017-2022) by examining clinical and laboratory variables presumed to be associated with ivGC refractoriness at admission and on the third day of hospitalization. Continuous variables were dichotomized according to receiver operating characteristic (ROC) analysis, and combinations of up to three components were evaluated. The weight of each component was determined based on its regression coefficient (B value) in a logistic regression model. The combination with the best performance metrics, focusing on area under the curve (AUC), sensitivity, and specificity was suggested as the study score (Figure 1). The proposed score was validated in an independent cohort (2022-2024). For both cohorts, the study score was compared for significance and performance metrics with existing early and third-day models1,2. Results Ninety-six patients were included in the early cohort; 56 (58%) were ivGC responders and 40 (42%) were refractory. On admission, patients with ivGC refractoriness presented with higher number of stools per day (median 15 (interquartile range (IQR) 10-20) vs. 10 (IQR (7-15), p = 0.005), higher CRP to albumin (CAR), and CRP to lymphocytes ratios (CLR) CAR: median 2 (IQR: 0.6-4) vs. 0.8 (IQR: 0.45-2.3) p = 0.012; CLR: median 40 (IQR: 11.5-122) vs. 17 (IQR: 7.5-43) p = 0.023. UCEIS endoscopic score was higher among ivGC-refractory patients (median 5 points (IQR: 4-6.7) vs. 4 (IQR: 2-5), p = 0.001), (Figure 1). IvGC refractoriness was associated with the following criteria: ≥ 9 stools/day on admission and either CLR ≥ 20 or UCEIS ≥ 5 (odds ratio (OR): 5.26, 95% confidence interval (CI): 1.9–14.2, p = 0.001). The study score outperformed existing early scores (AUC: 0.68). In the validation cohort (n = 53), this model was the only one to retain significance (OR: 6.37, 95% CI: 1.7–23.6, p = 0.006), with the highest AUC (study score: 0.72, Gibson: 0.63, Adams: 0.57, Oxford: 0.64, Table 1). Conclusion We propose a simple and early score to predict ivGC refractoriness in patients with ASUC. This model outperformed existing early scores and demonstrated performance comparable to third-day models. It may support timely clinical decision-making in this high-risk setting. References: 1. Calméjane L, Laharie D, Kirchgesner J, Uzzan M. Review article: Updated management of acute severe ulcerative colitis: From steroids to novel medical strategies. United European Gastroenterology Journal. 2023;11(8):722. doi:10.1002/ueg2.12442 2. Kayal M, Meringer H, Martin L, Colombel JF. Systematic review: Scores used to predict outcomes in acute severe ulcerative colitis. Alimentary Pharmacology 58(10):974-983. doi:10.1111/apt.17731 Conflict of interest: Dr. Sror, Neta: I confirm that I have no conflict of interest, neither financial nor personal, regarding any of the subjects presented at the ECCO 2026 Conference. Leibovitzh, Haim: I declare that they I have no conflict of interest Shaham, Daniel: No conflicts of interests Hirsch, Ayal: No conflicts of interests Thurm, Tamar: Lecture fees from Takeda, Johnson & Johnson, AbbVie Ron, Yulia: No conflicts of interests Maharshak, Nitsan: I confirm that I have no conflict of interest, neither financial nor personal, regarding any of the subjects presented at the ECCO 2026 Conference. Cohen, Nathaniel Aviv: Served as a consultant for or received speaker’s fees from Abbvie, Abivax, Bristol-Myers Squibb, Eli-Lilly & Co., Iterative Health, Janssen Pharmaceuticals, Padagis, Rafa Pharmaceuticals, Seres Pharmaceuticals and Takeda and received travel support from Pfizer and Takeda.