Risk Stratification for Sarcopenic Obesity in Subjects With Nonalcoholic Fatty Liver Disease

Abstract

Background & aims: The impact of the severity of sarcopenic obesity (SO) in nonalcoholic fatty liver disease (NAFLD) on the risk of significant liver fibrosis or cardiovascular disease (CVD) remains unclear. We aimed to identify high-risk subjects with SO for significant liver fibrosis or CVD among subjects with SO and NAFLD.

Methods: This multicenter, retrospective study involved 23,889 subjects with NAFLD who underwent a health screening program (2014-2020). Sarcopenia was defined based on gender-specific sarcopenia index cutoff using multi-frequency bioelectric impedance analysis. High-risk subjects with SO were defined as those with significant liver fibrosis by fibrosis-4 index >2.67 or atherosclerotic CVD risk score >20%. Multivariable logistic regression analysis for identifying high-risk subjects with SO was performed in a cross-sectional cohort with SO, and further validation was performed in a longitudinal cohort.

Results: SO prevalence was 5.4% (n = 1297 of 23,889). Older age (unstandardized beta [β] = 3.23; P < .001), male (β = 1.66; P = .027), sarcopenia index (β = -6.25; P = .019), and metabolic syndrome (β = 1.75; P < .001) were significant risk factors for high-risk SO. Based on a high-risk SO screening model, high-risk subjects with SO had significantly higher odds of significant liver fibrosis (training: adjusted odds ratio [aOR], 3.72; validation: aOR, 2.38) or CVD (training: aOR, 5.20; validation: aOR, 3.71) than subjects without SO (all P < .001). In subgroup analyses, the cumulative incidence of significant liver fibrosis or CVD development was significantly higher in high-risk subjects with SO than in low-risk subjects with SO in a longitudinal cohort considering all-cause mortality and liver transplantation as competing risks (sub-distribution hazard ratio, 5.37; P < .001).

Conclusion: The high-risk screening model may enable the identification of high-risk subjects with SO with NAFLD.