238 - Age-Stratified Risk of Hepatic Encephalopathy After TIPS: Defining a Clinical Inflection Point in a Multi-System EHR Cohort

Tuesday, April 14, 2026

3:18 PM - 3:27 PM ET

Presenting Author(s)

Rohit Reddy, MD

IR Resident
Department of Interventional Radiology, University of Miami, United States

Author/Co-author(s)

AM

Adrielly Martins, MD

Research Fellow
Schiff Center for Liver Diseases, University of Miami Miller School of Medicine, United States
CL

Cynthia Levy, MD

Professor of Clinical Hepatology
Division of Digestive Health and Liver Diseases, Department of Medicine, University of Miami Miller School of Medicine, United States
US

Usman Shahid, MD

Attending
University of Miami, United States

Purpose: Hepatic encephalopathy (HE) is a frequent complication after TIPS, but contemporary, age-stratified incidence remains poorly defined. We evaluated age-stratified outcomes across a multi-system EHR to identify a potential clinical inflection point for de novo HE risk.

Materials and Methods: We analyzed adults undergoing first-ever TIPS (2010–2025) across a multi-system national EHR, and excluded patients with prior HE, prior liver transplants, and prior TIPS. Outcomes included incident HE, 30d readmission, mortality, transplant, and transplant-free survival (TFS) at 30, 90, and 180d. MELD-Na, MELD 3.0, and FIPS were calculated. Age was stratified into prespecified categories (< 50, 50–55, 55–60, 60–65, 65–70, 70–75, 75–80, >80). Logistic regression models evaluated predictors of incident HE and 30-day readmission using both complete case and multiple imputation (MI) analyses.

Results: Among 18,804 TIPS patients (median age 59 [IQR 51–66], 64% male), incident HE occurred in 10,369 patients (55.1%) and 30-day readmission in 9,478 (50.4%). TFS was 69.5%, 56.6%, and 46.7% at 30, 90, and 180 days, with a median of 146 days. Odds of incident HE rose with age and peaked at 65 – 70 years (OR 1.90 vs < 50, p< 0.001), declining modestly thereafter. Readmission odds increased with age, peaking around 70–80 years (OR ≈2.0 vs < 50). In regression models, age was the dominant predictor of incident HE and readmission, while MELD 3.0 showed inconsistent or opposite associations. Findings were consistent across multiple imputation analyses.

Conclusion: In this multi-system cohort, age demonstrated a clinical inflection point for post-TIPS HE risk, peaking at 65–70 years. This is the first study to evaluate fine-grained age strata in a national multi-system TIPS cohort restricted to incident HE, directly compare age with MELD-Na, MELD 3.0, and FIPS, and quantify post-TIPS TFS alongside HE and readmission. These results support incorporating age-stratified thresholds into TIPS candidacy decisions, as conventional severity scores may underestimate HE risk relative to patient age.