Predicting Morbidity and Mortality After Transjugular Intrahepatic Portosystemic Shunt Placement: A Review of Existing Models and Future Directions.
Journal:
Techniques in vascular and interventional radiology
Published Date:
Jul 14, 2025
Abstract
Transjugular intrahepatic portosystemic shunt (TIPS) is a key therapeutic intervention in the management of portal hypertension and its complications, such as variceal bleeding, hepatic hydrothorax, and refractory ascites. TIPS has historically been used as a lifesaving measure or as a bridge to liver transplantation (LT). Despite its efficacy, creation of a TIPS can be associated with significant morbidity, particularly in patients with decompensated cirrhosis. Complications include hepatic encephalopathy (HE), progressive liver dysfunction, and cardiovascular compromise. As such, accurate patient selection and risk stratification are essential to optimize clinical outcomes. This review synthesizes current evidence on predictive models for post-TIPS mortality. Traditional scoring systems such as the Child-Turcotte-Pugh (CTP) score and the Model for End-Stage Liver Disease (MELD) remain widely used, with newer iterations such as the MELD-Na and MELD 3.0 demonstrating improved prognostic accuracy. Notably, MELD 3.0 offers enhanced prediction of long-term mortality. In contrast, the Freiburg Index of Post-TIPS Survival (FIPS) has become a valuable tool for short-term mortality prediction. Additional models, including the Bilirubin-Platelet (Bili-PLT) score, offer further refinement. At the same time, the role of sarcopenia has gained attention as an independent and synergistic predictor of poor outcomes, especially when combined with MELD-based scores. Beyond mortality, this review explores the multifactorial pathophysiology of post-TIPS complications such as hepatic encephalopathy, liver failure, and right heart dysfunction that can cause significant morbidity. These outcomes are influenced by a spectrum of patient-related and procedural factors. Novel predictive approaches-encompassing clinical, radiological, and machine learning-based models-are being developed to better anticipate these risks.
