Limited performance of subjective global assessment compared to computed tomography-determined sarcopenia in predicting adverse clinical outcomes in patients with cirrhosis

Abstract

The subjective global assessment (SGA) is commonly used to assess nutritional status in patients with cirrhosis. Sarcopenia, a major component of malnutrition, is associated with survival in cirrhosis, and can be objectively diagnosed by computing the skeletal muscle index (SMI) using cross-sectional imaging. The aim of this study was to assess the prevalence of sarcopenia between SGA categories in patients with cirrhosis, and to determine their association with decompensation/mortality. We included 315 patients (66% males) who were assessed for liver transplantation. All patients had SGA and SMI, and were evaluated for the presence of hepatic encephalopathy (HE) and ascites. Mean age was 54±8 years. SGA categories were 126 SGA A (40%), 155 SGA B (49%), 34 SGA C (11%). Sarcopenia was present in 121 (38%) patients; of these, 82% were SGA A/B. Of SGA A patients, 25 (20%) had sarcopenia. There was a significant but only weak concordance between sarcopenia and SGA B/C (κ=0.28, p<0.001), and SGA C (κ=0.13, p<0.001). The latter was lost in overweight/obese patients. SGA B/C was associated with HE (OR 2.8, p=0.01) and ascites (OR 2.3, p=0.002). Median survival was shorter in patients with sarcopenia (20 [IQR 15.9-24.5] vs. 42 [IQR: 25.8-58.9] months, p<0.001) and in SGA C patients (9.4 [IQR: 0-26.2] vs. 33 [IQR 20.2-45.7] months, p=0.01). In univariate analysis both sarcopenia and SGA C were associated with mortality, but sarcopenia was the only factor that remained significant on multivariate analysis. There was only a weak concordance between SGA and sarcopenia. This concordance was non-significant in patients who were overweight/obese. Sarcopenia was associated with mortality, whereas SGA was not. Sarcopenia by the SMI is a more efficient method to predict adverse outcomes in a timely fashion and has prognostic implications.