P65 Assessing predictors of mortality following transjugular intrahepatic portosystemic shunts for variceal bleeding

<h3>Background and Aims</h3> Transjugular intra-hepatic portosystemic shunts (TIPSS) are used as salvage therapy in variceal bleeding and are also increasingly being utilised pre-emptively. Our aim was to evaluate the outcomes of salvage TIPSS performed for variceal bleeding at a tertiary liver transplant centre and identify predictors of mortality. <h3>Method</h3> Data was collected for patients undergoing TIPSS procedures between April 2010 and September 2022. Statistical analyses included t-tests and Mann-Whitney tests for comparison of continuous variables between groups and the Cox proportional hazards model to identify predictive factors for mortality. Patients were censored at the point of death, their last known follow-up or the date of transplantation. <h3>Results</h3> 193 patients had an average age of 52.7 years and were 68.4% male. Aetiology was predominantly alcohol (64.8%), followed by metabolic dysfunction-associated steatotic liver disease (10.9%) and hepatitis C infection (7.8%). 9 patients progressed to transplant and in those who did not, survival was 74.6% at 90 days and 64.8% at 12 months. Documented causes of death revealed 79.4% as liver-related. The median MELD was 13 (IQR 10–17). Baseline MELD, MELDNa, GEMA and GEMANa were all associated with 90-day, 12 month and overall survival. In predicting 90-day mortality, AUROC analysis of the MELD score was 0.819 (0.753–0.884), MELD Na 0.813 (0.747–0.880), GEMA 0.809 (0.743–0.875) and GEMA Na 0.801 (0.732–0.870). Both 90-day and overall survival analysis showed significance at univariate analysis for bilirubin, creatinine, INR, albumin, WBC, MELD, MELDNa, GEMA, GEMANa, Bureau’s criteria, previous hepatic encephalopathy and the RFH eGFR. All prognostic scores (tested in separate multivariate models) except for Bureau’s were independently predictive of mortality at multivariable analysis at both 90 days (MELD aHR 1.138, 95% C.I. 1.090–1.188, MELDNa aHR 1.123, 95% C.I.1.076–1.172, GEMA aHR 1.136, 95% C.I. 1.078–1.185, GEMANa aHR 1.130, 95% C.I.1.078–1.185) and overall (MELD aHR 1.083, 95% C.I 1.045–1.122, MELDNa aHR 1.074, 95% C.I.1.037–1.111, GEMA aHR 1.077, 95% C.I. 1.040–1.115, GEMANa aHR 1.079, 95% C.I 1.040–1.119). Previous encephalopathy, smoking and age at TIPSS insertion remained independently significant in all models as did WBC (aHR 1.040, 95% C.I.1.008–1.072), RFH e-GFR (aHR 0.990, 95% C.I. 0.983–0.998) and MetALD (HR 1.912, 95% C.I 1.144–3.197) when modelled with the Bureau criteria for overall survival. <h3>Conclusion</h3> Prognostic scores assessed on this large cohort of TIPSS procedures performed for variceal bleeding were good predictors of mortality. As increasing numbers of pre-emptive procedures are being implemented, these will need to be re-assessed in this subset of patients.