P614 Reduction of bowel wall thickness and doppler signals as early as 36 hours predicts corticosteroid response in a multi-centre prospective cohort of Acute Severe Ulcerative Colitis: USUC Study including the GENIUS cohort

Abstract

Abstract Background Gastrointestinal ultrasound (GIUS) can accurately assess disease activity in atients with ulcerative colitis (UC). The aim of this study was to determine the accuracy of GIUS in predicting IV corticosteroid (CS) failure in patients with Acute Severe Ulcerative Colitis (ASUC) and the requirement for rescue therapy. Methods We conducted a multi-centre prospective observational cohort study of adult ASUC admitted to hospital between November 2019 to July 2022. GIUS was performed at six time points for each patient: at hospital admission (SV1), day 3 (SV2), at discharge (SV3), and thereafter during outpatient review in the 10 months of follow-up. Medical rescue therapy was given at the discretion of treating physician according to best clinical practice. Results A total of 32 consecutive patients with ASUC were recruited with a median follow-up duration of 8.9 months: median age 32 years, 38% male and median duration of disease 19 months. Seven patients had newly diagnosed UC and six patients (19%) were biologic experienced. Sixteen patients (50%) were CS responders and the remainder required medical rescue therapy; two patients required colectomy. The median time to rescue therapy and colectomy was 3.5 and 11 days from admission, respectively. At SV1, the median BWT of the worst affected segment was similar between CS responders and non-responders (5.50mm vs 5.85mm, p=0.66), however at SV2, responders had a significantly lower BWT (3.85mm vs 5.05mm, p=0.02). CS responders demonstrated a significant reduction in width of the muscularis propria (p=0.02) and submucosal (p=0.006) layer, but not the mucosal (p=0.61) layer (Figure 1). CS non-responders had a smaller absolute and relative reduction in median BWT at SV2. Receiver operating characteristic curve analysis showed that an absolute reduction of <1.40mm (Sn 63%, Sp 75%, AUROC 0.76) or a relative BWT reduction of <20% (Sn 81%, Sp 75%, AUROC 0.78) predicted CS non-response and need for rescue therapy well. A reduction in BWT and the absence of doppler activity at SV2 further enhanced the predictive capability of GIUS (Sn 81.3%, Sp 87.5%, AUROC 0.91) (Figure 2). Conclusion This study demonstrates the predictive utility of early GIUS in the management of ASUC. A reduction in absolute BWT of <1.40mm or relative BWT of <20% at 36-48 hours of admission identifies CS non-responders with good sensitivity and specificity. BWT reduction at SV2 in responders was driven mainly by decreased muscularis propria and submucosal thickness. Accuracy is further augmented when reduction in BWT and resolution of doppler signal are combined. This study suggests GIUS may allow clinicians to expedite initiation of rescue therapy for patients requiring salvage treatment.