Abstract Background Transjugular intrahepatic portasystemic stent shunt (TIPSS), instead of surgical shunt, has become the standard treatment for patients with complicated portal hypertension. This study compared outcomes in patients who underwent TIPSS or surgical shunting for complicated portal hypertension. Methods This was a retrospective study of all consecutive patients who received portasystemic shunts from 1994 to 2014 at a single institution. Patients who underwent surgical shunting were compared with those who had a TIPSS procedure following one-to-one propensity score matching. The primary study endpoints were overall survival and shunt failure, defined as major variceal rebleeding, relapse of refractory ascites, irreversible shunt occlusion, liver failure requiring liver transplantation, or death. Results A total of 471 patients received either a surgical shunt or TIPSS. Of these, 334 consecutive patients with cirrhosis who underwent elective surgical shunting (34) or TIPSS (300) for repeated variceal bleeding or refractory ascites were evaluated. Propensity score matching yielded 31 pairs of patients. There were no between-group differences in morbidity and 30-day mortality rates. However, shunt failure was less frequent after surgical shunting than TIPSS (6 of 31 versus 16 of 31; P = 0·016). The 5-year shunt failure-free survival (77 versus 15 per cent; P = 0·008) and overall survival (93 versus 42 per cent; P = 0·037) rates were higher for patients with surgical shunts. Multivariable analysis revealed that a Model for End-Stage Liver Disease (MELD) score exceeding14 and TIPSS were independently associated with shunt failure. In patients with MELD scores of 14 or less, the 5-year overall survival rate remained higher after surgical shunting than TIPSS (100 versus 40 per cent; P < 0·001). Conclusion Surgical shunting achieved better results than TIPSS in patients with complicated portal hypertension and low MELD scores.