Sarcopenia assessed by computed tomography or magnetic resonance imaging is associated with the loss of response to biologic therapies in adult patients with Crohn's disease

Abstract

Sarcopenia occurs in patients with Crohn's disease (CD). However, the association between sarcopenia and loss of response (LOR) to biologic agents remains unclear. This study explored such an association in CD patients. This retrospective study included 94 CD patients who received biologic therapy. The skeletal muscle cross-sectional area at the third lumbar was assessed by computed tomography or magnetic resonance imaging for sarcopenia evaluation. A LOR was defined by fecal calprotectin (FC) < 250 μg/g or >50% reduction from baseline levels or other factors, such as the used agent being replaced by other biologic agents. The association between sarcopenia and LOR was assessed by logistic regression analysis. LOR was observed in 54 patients (57.4%). The prevalence of sarcopenia in the LOR group was higher than that in response group (70.4% vs. 40.0%, p = 0.003). Sarcopenia (odds ratio [OR] = 3.89, 95% confidence interval [CI]: 1.31-11.54), Montreal L1 type (OR = 0.20, 95% CI: 0.06-0.60), perianal lesions (OR = 4.08, 95% CI: 1.31-12.70), and monocytes percentage (OR = 1.27, 95% CI: 1.02-1.57) at baseline were independent associated factors for LOR. Sarcopenia was also associated with LOR in patients who received infliximab (OR = 3.31, 95% CI: 1.11-9.87). Montreal L1 type, perianal lesions, and monocytes percentage (Model 1), and with additional consideration of sarcopenia (Model 2), were developed to predict LOR. Model 2 showed better performance than Model 1 (area under the curve [AUC] 0.82 vs. 0.75). Sarcopenia was associated with the LOR to biological agents or infliximab in adult patients with CD.

FIGURE 1

Flowchart of study patients: 284 patients with Crohn's disease (CD) were found in the medical database; 94 patients with biologic therapy were finally included. CTE, computed tomography enterography; MRE, magnetic resonance enterography.

FIGURE 2

Performance of models in predicting loss of response (LOR). Model 1 was based on Montreal L1 type, perianal lesions, and monocytes percentage (a). Sarcopeania was additionally added in Model 2 (b).

FIGURE 3

Nomogram to predict loss of response (LOR) based on Montreal L1 type (1, yes), perianal lesions (1, yes), monocytes percentage, and sarcopeania (1, yes) (a). Calibration curve further showed a good agreement between the predicted risk and observed outcomes (b).