Risk of portal hypertensive complications preventable by TIPS in patients with ascites

Background & Aims: Transjugular intrahepatic portosystemic shunt (TIPS) is an effective treatment of recurrent/refractory ascites in patients with cirrhosis. The aim of this study is to identify patients with ascites as index decompensation who are at risk of developing portal hypertension (PH)-related complications within 12 months that seem preventable by TIPS. Methods: We included 451 patients from two tertiary care centres (Vienna and Padua, derivation cohort) with clinically significant ascites (grade 2/3) as a single first decompensating event and without contraindications for TIPS placement. Multivariable logistic regression analysis was used to identify variables independently associated with a composite endpoint of PH-related complications (encephalopathy excluded), liver transplantation, or liver-related death. A classification tree was used to identify patients at highest risk for these PH-related complications. Risk estimates were validated in a temporal validation cohort from Vienna (n = 84). Results: In the derivation cohort (mean age 56 ± 11 years; 69% male; 51% alcohol-related cirrhosis; 44% ascites grade 3; median model for end-stage liver disease [MELD] 12 points), 152 (34%) patients developed the composite endpoint within 12 months. A model including ascites grade, sodium, and MELD accurately predicted the occurrence of this composite endpoint (area under the receiver operator characteristics curve: 0.79 [95% CI: 0.75–0.84]). Two high-risk clusters were identified: patients with grade 3 ascites and either (i) sodium ≤135 mmol/L, or (ii) MELD ≥12 points, with a pooled absolute risk of 64.3% (derivation cohort) and 68.9% (validation cohort) to develop the composite endpoint. Conclusions: Patients with first decompensation caused by ascites grade 3 and either sodium ≤135 mmol/L or MELD ≥12 are at high risk for PH-related complications that are likely preventable by early TIPS placement. A trial investigating ‘early’ TIPS in this at-risk population is warranted. Impact and implications: We identified ascites grade, sodium, and model for end-stage liver disease (MELD) as key predictors of portal hypertension-related complications that may be preventable by TIPS in patients with ascites. Specifically, patients with ascites grade 3 and either sodium ≤135 mmol/L or MELD ≥12 are at risk to experience early clinical deterioration and may benefit from TIPS. A trial investigating ‘early’ TIPS in this at-risk population is warranted.