Low testosterone as a better predictor of mortality than sarcopenia in men with advanced liver disease.

Abstract

Both sarcopenia and low serum testosterone have been associated with increased mortality in men with cirrhosis. It is not known how these variables interact. We conducted a retrospective longitudinal cohort study of 145 men referred for liver transplant evaluation between 2005 and 2012. Baseline demographics included hormone profile and model of end-stage liver disease (MELD) score. Baseline computerized tomography was reformatted to calculate skeletal muscle area at L4 using validated, Tomovision software-based methodology. The primary outcome was time to death or liver transplantation. Median testosterone was low at 6.2 nmol/L (ref. 10-27.6 nmol/L) as was muscle mass at 48.0 cm(2)/m(2) (ref. > 52.4 cm(2)/m(2)). Muscle mass correlated with both serum testosterone (tau = 0.132, P = 0.019) and MELD score (tau = -0.155, P = 0.007). In separate multivariable models, both sarcopenia (hazard ratio [HR] 1.05, P = 0.04) and low testosterone (HR 1.08, P = 0.01) were significantly associated with mortality independent of MELD score. When the variables MELD score, muscle area, and testosterone were entered into a single model, low testosterone but not sarcopenia remained significantly predictive of mortality (HR 1.07, P = 0.02, and HR 1.04, P = 0.09, respectively). Low testosterone and sarcopenia are both associated with increased mortality in men with advanced liver disease and may identify patients at high risk of mortality that would be missed by the MELD score alone. Low testosterone appears to be a better predictor of mortality than sarcopenia and is a simpler test to improve the prognostic value of the MELD score. Interventional trials are required to determine whether low testosterone and sarcopenia are markers or mediators of mortality in this population.