Sarcopenia is a common finding in patients with decompensated cirrhosis without effective therapy. We aimed to examine whether a transjugular portosystemic shunt (TIPS) could improve the abdominal muscle mass assessed by cross-sectional images in patients with decompensated cirrhosis and to investigate the association of imaging-defined sarcopenia with the prognosis of such patients. In this retrospective observational study, we enrolled 25 Decompensated cirrhosis patients aged >20 who received TIPS for the control of variceal bleeding or refractory ascites between April 2008 and April 2021. All of them underwent preoperative computed tomography or magnetic resonance imaging, which was used to determine psoas muscle (PM) and paraspinal muscle (PS) indices at the third lumbar vertebra. First, we compared baseline muscle mass with muscle mass at 6 and 12 months after TIPS placement and analyzed PM- and PS-defined sarcopenia to predict mortality. Among 25 patients, 20 (80.0%) and 12 (48.0%) had PM- and PS-defined sarcopenia, respectively, at baseline. In total, 16 and 8 patients were followed up for 6 and 12 months, respectively. All imaging-based muscle measurements performed 12 months after TIPS placement were significantly greater than the baseline measurements (all p<0.05). Unlike patients with PS-defined sarcopenia (p=0.529), patients with PM-defined sarcopenia had poorer survival than did patients without (p=0.036). PM mass in patients with decompensated cirrhosis may increase by 6 or 12 months after TIPS placement and imply a better prognosis. Patients with preoperative PM-defined sarcopenia may suggest poorer survival.