BACKGROUND We aimed to elucidate the relationship between serum myostatin levels and other markers including skeletal muscle mass and to investigate the influence of serum myostatin levels on survival for patients with liver cirrhosis (LC). METHODS A total of 198 LC subjects were analysed in this study. Myostatin levels were measured using stored sera. We retrospectively investigated the relationship between myostatin level and other markers, and the influence of myostatin level on overall survival (OS). Assessment of skeletal muscle mass was performed using the psoas muscle index (PMI) on computed tomography images at baseline. PMI indicates the sum of bilateral psoas muscle mass calculated by hand tracing at the lumber three level on computed tomography images divided by height squared (cm2 /m2 ). The study cohort was divided into two groups based on the median myostatin value in each gender. RESULTS Our study cohort included 108 male and 90 female patients with a median age of 67.5 years. The median (range) myostatin level for male patients was 3419.6 pg/mL (578.4-12897.7 pg/mL), whereas that for female patients was 2662.4 pg/mL (710.4-8782.0 pg/mL) (P = 0.0024). Median (range) serum myostatin level for Child-Pugh A patients (n = 123) was 2726.0 pg/mL (578.4-12667.2 pg/mL), whereas that for Child-Pugh B or C patients (n = 75) was 3615.2 pg/mL (663.3-12897.7 pg/mL) (P = 0.0011). For the entire cohort, the 1-, 3-, 5-, and 7-year cumulative OS rates were 93.94%, 72.71%, 50.37%, and 38.47%, respectively, in the high-myostatin group and 96.97%, 83.27%, 73.60%, and 69.95%, respectively, in the low-myostatin group (P = 0.0001). After excluding hepatocellular carcinoma patients (at baseline) from our analysis (n = 158), the 1-, 3-, 5-, and 7-year cumulative OS rates were 96.0%, 77.93%, 52.97%, and 39.08%, respectively, in the high-myostatin group and 96.39%, 87.58%, 77.63%, and 73.24%, respectively, in the low-myostatin group (P = 0.0005). Higher age (P = 0.0111) and lower PMI (P < 0.0001) were identified as significant predictors of poorer OS in our multivariate analysis, while higher serum myostatin (P = 0.0855) tended to be a significant adverse predictor. In both genders, PMI, serum albumin, prothrombin time, and branched-chain amino acid to tyrosine ratio showed a significantly inverse correlation with myostatin levels, and serum ammonia levels showed a significantly positive correlation with myostatin levels. CONCLUSIONS Higher serum myostatin levels correlated with muscle mass loss, hyperammonemia, and impaired protein synthesis, as reflected by lower serum albumin levels and lower branched-chain amino acid to tyrosine ratio levels. High serum myostatin levels were also associated with a reduced OS rate in LC patients.