Background/Aims: The transjugular intrahepatic portosystemic shunt (TIPS) is technically divided into TIPS through the left branch of the portal vein (TIPS-LBPV) and TIPS through the right branch of the portal vein (TIPS-RBPV). In order to compare their advantages and disadvantages, this randomized, controlled trial was designed to investigate their outcomes in advanced cirrhotic patients. Methods: Seventy-two patients were randomly placed into TIPS-LBPV (36 patients) and TIPS-RBPV (36 patients, with four failures) groups, and they were prospectively followed for 2 years after TIPS implantation. Results: Patients who underwent the two different kinds of TIPS were balanced during recruitment for this study. The incidences of overall encephalopathy and de novo encephalopathy in the TIPS-LBPV group were significantly lower than that of the TIPS-RBPV group during follow-up (P = 0.036 and 0.012 respectively). The incidences of rebleeding or re-intervention and improvement of ascites were similar between groups (P > 0.05). Patients undergoing TIPS-RBPV required more rehospitalization and incurred more costs than those who underwent TIPS-LBPV (P = 0.030 and 0.039 respectively). There was no significant difference between the two groups in survival based on a survival curve constructed according to the Kaplan-Meier method (P > 0.05). Conclusion: Patients undergoing TIPS-LBPV had a lower incidence of encephalopathy, less rehospitalization and lower costs after TIPS implantation compared with patients undergoing TIPS-RBPV. © 2009 John Wiley & Sons A/S.