P481 Visceral fat to skeletal muscle area ratio predicts for Ustekinumab loss of response.

Tan, Z.(1)*;Chin, A.(2);Welman, C.J.(3,4);Thin, L.W.Y.(1,5);

(1)Fiona Stanley Hospital, Gastroenterology, Murdoch, Australia;(2)Royal Perth Hospital, Gastroenterology, Perth, Australia;(3)Fiona Stanley Hospital, Medical Imaging, Murdoch, Australia;(4)Royal Perth Hospital, Radiology, Perth, Australia;(5)University of Western Australia, Internal Medicine, Perth, Australia;


Loss of response (LOR) to anti-TNFα therapy in inflammatory bowel disease (IBD) is common and biologic therapies such as anti-IL12/23 agents have shown greater durability in efficacy. Anti-TNFα LOR is increased in higher body mass index (BMI) patients however the evidence for the association between body composition or BMI and Ustekinumab’s (Ust) therapeutic response is limited, with the latter not impacting clinical efficacy in the ‘IM-MUNITI’ post-hoc analyses. This study’s primary aim was to evaluate the impact of body composition on time to LOR to Ustekinumab in Crohn’s disease (CD) patients.


99 adult CD patients aged ≥16 yrs. from two tertiary IBD centres were studied retrospectively. Included patients had received Ust ≥12 weeks, abdominal imaging (CT or MRI) within 6 months of Ust induction, and documented follow up until 30th April 2022. A single experienced abdominal radiologist blinded to the clinical information measured the area of visceral fat (VFA), skeletal muscle (SMA), and subcutaneous fat at the mid L3 vertebral level.  Each compartment was corrected for height2 to derive visceral fat (VFI), skeletal muscle (SMI) and subcutaneous fat (SFI) indices. VFA:SMA ratio was also calculated. Clinico-demographic details including therapeutic and surgical history were recorded. Kaplan Meir survival analyses, univariate and multivariate cox regression analyses were performed to evaluate factors predictive of Ust LOR.


99 patients met the inclusion criteria with mean age (±SE) 46.6yrs (±1.6) and 49 (49.5%) male. 63 (63.6%) patients were bio-experienced. Univariate Cox regression analyses of age, sex, BMI, peak C-reactive protein (CRP) and nadir Albumin level within 3 months of the CT/MRI scan, disease duration, prior biologic failure and concurrent immunomodulator used within 6 months of Ust induction showed that none of these were predictive of Ust LOR (P>0.05). There was a numerical trend for the middle VFI tertile (40.01-79.99cm2/m2) to having a lower risk for LOR compared to the lowest and highest VFI tertile (Fig. 1a), (HR=0.37, 95%CI= 0.13, 1.11, P=0.08), as well as a higher SMI for having a lower risk of LOR (Fig. 1b), (HR= 0.6, 95%CI= 0.24,1.31, P= 0.18), but these were not statistically significant.  A higher VFA: SMA ratio however, was a significant predictor of LOR (HR=2.98, 95%CI 1.2,7.3, P= 0.02) (Fig. 1c), which persisted when it was adjusted on multivariate analysis for age, sex, albumin, CRP, BMI, disease duration, prior biologic failure and concurrent immunomodulator use (HR= 10.6, 95%CI 2.8-39.8, P=<0.0001).


The ratio of  VFA:SMA, but not BMI, is a superior predictor of losing response to Ustekinumab.