Efficacy of variceal embolisation in patients with TIPS implantation: a propensity score analysis

Background:

Implantation of transjugular intrahepatic portosystemic shunt (TIPS) is an effective treatment for complications of portal hypertension such as variceal bleeding in patients with liver cirrhosis. However, some patients experience rebleeding from varices after successful TIPS implantation. During TIPS implantation varices can be detected oder: with angiography and embolized. As clinical data for variceal embolization are rare, we set out to analyze the safety and efficacy of this procedure.

Methods:

Patients receiving TIPS implantation with (n = 104) and without variceal embolization (n = 311) between 2004 and 2017 were retrospectively enrolled and analyzed by reviewing medical records, laboratory parameters and technical data of TIPS implantation and variceal embolization. Propensity score matching was performed to adjust for differences in baseline characteristics of patients with and without embolization. The primary endpoint was the re-bleeding rate and overall survival.

Results:

Rebleeding-rates were 16.5% in the embolization group and 8.2% in patients without embolization (p = 0.060) showing a trend to higher re-bleeding rates in patients with variceal embolization. Patients with additional embolization had significantly less varices during follow-up compared to patients who were only treated with TIPS implantation (38.5% vs. 62.2%; p = 0.033). Additional embolization had no effect on overall survival (HR 0.78 [0.59 – 1.06], p = 0.108). After propensity score matching 76 patients in each group entered the analyses and re-bleeding rates were similiar in both groups (14.5% vs. 15.7%, p = 0.999).

13 of 104 patients (12.5%) with embolization experienced complications (portal vein thrombosis: n = 2, outflow of the embolization agent in the lung: n = 9 and in the portal vein: n = 2). These complications were not clinically relevant.

Conclusions:

Additional embolization during TIPS implantation showed similiar re-bleeding rates compared to patients without embolization. Further, prospective and well-powered studies are warranted to analyze if there may be a benefit concerning re-bleeding which might be underestimated in our cohort due to the small sample size.