A case–control study of transjugular intrahepatic portosystemic stent shunts for patients admitted to intensive care following variceal bleeding

Abstract

Variceal bleeding has a 6-week mortality of 20%. Recent evidence suggests that early covered transjugular intrahepatic portosystemic stent shunts (TIPSS) can improve outcomes following a variceal bleed in selected patients. We aim to assess the outcomes following the insertion of covered TIPSS in a real-life intensive care setting. This is a retrospective matched cohort study of all patients referred for TIPSS with variceal bleeding admitted to intensive care (2007-2009). Patients were matched with others admitted to intensive therapy unit following a variceal bleed but did not proceed to TIPSS. All TIPSS procedures were carried out using polytetrafluoroethylene-covered stents. Thirty-eight patients [mean age 55.2 years; mean model for end-stage liver disease (MELD)=14.0; and median follow-up 458 days] were assessed. Nineteen underwent TIPSS and were well matched to the controls. All patients received terlipressin and antibiotics and 86% had active bleeding at endoscopy. Indication for TIPSS was salvage therapy (47%), rebleeding after day 5 (11%) and as secondary prophylaxis (42%). There was 34% all-cause inpatient mortality. The TIPSS group had lower mortality than the non-TIPSS group at 6 weeks (10.5 vs. 47.4%, P<0.05) that persisted at 1 year (21.1 vs. 52.6%, P<0.05). Multivariate analysis indicated MELD [HR 1.131, 95% confidence interval (CI) 1.018-1.257] and TIPSS (HR 0.301, 95% CI 0.091-0.995) as significant predictors of mortality (P<0.05). TIPSS was found to significantly reduce the incidence of failure to control bleeding and rebleeding (HR 0.120, 95% CI 0.015-0.978, P<0.05). Patients with recent severe variceal bleeding admitted to intensive care have significantly better outcomes following covered TIPSS insertion. These findings should be validated in randomized-controlled trials.