The modified computed tomography severity index combined with low skeletal muscle mass can better predict the severity of hypertriglyceridemia-induced pancreatitis.

Abstract

Body composition parameters are associated with hypertriglyceridemia-induced pancreatitis (HTGP). This study investigated the association between the quantity of muscle assessed using computed tomography (CT) and the severity of HTGP. The modified CT severity index (MCTSI) was calculated from admission examination data. Patients' characteristics and body composition parameters were collected. Univariate and multivariate logistic regression analyses were also performed. The receiver operating characteristic curves and corresponding area under the curves (AUC) were calculated to test the efficiency of the model. A nomogram was then constructed. Of the 175 included patients, 138 were male, of which 85 had moderately severe to severe HTGP. Patients with low skeletal muscle mass (LSMM) and high MCTSI were significantly more likely to have moderately severe to severe HTGP. Patients with LSMM had lower body mass index, lower HDL-C level, higher amylase level, prevalence of surgery, shorter umbilical waist circumference, and longer length of hospital stay. Univariate and multivariate logistic regression analyses confirmed that female sex, lipase, total cholesterol, LSMM-MCTSI (P= .004, odds ratio=23.105), and albumin were risk factors. The TOTAL model that combined LSMM-MCTSI and clinical risk parameters performed best (AUCs=0.875), followed by other models (LSMM-MCTSI: AUCs=0.762, MCTSI: AUCs=0.728). The Delong test revealed significant difference. Finally, a nomogram was developed to predict the severity of HTGP. The performance of MCTSI in predicting severity can be improved by considering LSMM, which is a promising strategy for the treatment of HTGP.