U-shaped relationship between subcutaneous adipose tissue index and mortality in liver cirrhosis

Abstract

Background: Subcutaneous and visceral adipose tissues are important body components, but their effects on the mortality in patients with liver cirrhosis remain controversial based on the current evidence. Methods: We retrospectively identified 372 eligible patients in whom subcutaneous adipose tissue index (SATI) and visceral adipose tissue index (VATI) could be measured by computed tomography images at the third lumbar vertebra. The association of SATI and VATI with the risk of death was evaluated on a continuous scale with restricted cubic spline curves based on Cox proportional hazards models. Cumulative probability of mortality was estimated by Nelson–Aalen cumulative risk curve analyses. Independent predictors of death were evaluated by competing risk analyses after adjusting for age, sex, and model for end-stage liver disease score. Results: Majority of patients were male (69.4%) with a mean age of 55.40 ± 10.68 years. SATI had a U-shaped association with mortality (P for non-linearity <0.001). Cutoff values of SATI were 19.7 and 51.8 cm2/m2 at the points where hazard ratios were just <1.2. SATI was categorized as low (<19.7 cm2/m2), moderate (19.7–51.8 cm2/m2), and high (>51.8 cm2/m2) level. There was no significant difference in the cumulative probability of mortality between low versus moderate SATI groups (Gray's test, P = 0.052) and high versus moderate SATI groups (Gray's test, P = 0.054). Competing risk analyses demonstrated that low SATI could increase the mortality compared with moderate SATI (subdistribution hazard ratio [sHR] = 1.66, 95% confidence interval [CI]: 0.992–2.78, P = 0.054) and was an independent predictor of death (sHR = 1.86, 95% CI: 1.059–3.28, P = 0.031). Competing risk analyses also demonstrated that high SATI could significantly increase the mortality compared with moderate SATI (sHR = 1.6, 95% CI: 1–2.54, P = 0.049), and was an independent predictor of death (sHR = 2.007, 95% CI: 1.195–3.37, P = 0.0085). VATI had an irregularly shaped association with mortality (P for non-linearity <0.001). Cutoff values of VATI were 9.8 and 40.2 cm2/m2 at the points where hazard ratios were just <1.2. VATI was categorized as low (<9.8 cm2/m2), moderate (9.8–40.2 cm2/m2), and high (>40.2 cm2/m2) level. There was no significant difference in the cumulative probability of mortality between low versus moderate VATI groups (Gray's test, P = 0.381) and high versus moderate VATI groups (Gray's test, P = 0.787). Competing risk analyses demonstrated that neither low (sHR = 1.27, 95% CI: 0.599–2.7, P = 0.53) nor high VATI (sHR = 0.848, 95% CI: 0.539–1.34, P = 0.48) was an independent predictor of death compared with moderate VATI. Conclusions: Both excessive deficiency and accumulation of subcutaneous adipose tissues negatively influence the outcomes of cirrhotic patients.