Development and Validation of a TIPS Prognostic Model

This multicenter, prospective observational study aims to learn whether a prediction model using routine clinical data, imaging findings (including spleen stiffness measurement), and circulating biomarkers can predict stent dysfunction and longer-term outcomes after transjugular intrahepatic portosystemic shunt (TIPS) in adults with portal hypertension. The main question it aims to answer is: Does this model predict TIPS stent dysfunction within 1 year after the procedure, and does it perform better than Doppler ultrasound for assessing stent function? Secondary questions include whether the model can predict further decompensation, liver failure, recompensation, and liver-related death over 3 years, and whether periodontal status and circulating biomarkers are associated with adverse outcomes after TIPS. About 1,490 participants aged 18-80 years who are scheduled to undergo TIPS as part of their routine medical care will be enrolled at multiple centers and followed for up to 3 years. This is an observational study; participants receive usual medical care, and no treatment is assigned by the study. No additional study-specific visits are required. Data will be collected from routine clinical records at baseline, during TIPS, within 1 week after TIPS, and at follow-up visits (1, 3, and 6 months, then every 6 months). Optional blood, stool, urine, gingival crevicular fluid, and subgingival plaque samples may be collected for biomarker and mechanism research. The primary endpoint is TIPS stent dysfunction within 1 year after the procedure.

Inclusion Criteria: * Voluntary written informed consent provided by the participant. * Age ≥18 years and ≤80 years; male or female. * Scheduled to undergo transjugular intrahepatic portosystemic shunt (TIPS) per physician order as part of routine clinical care. Exclusion Criteria: * Concomitant intrahepatic or extrahepatic malignancy, including solid tumors and hematologic malignancies, or any other disease with expected survival \