S0622 Upper Endoscopic Outcomes in High MELD Patients Undergoing Inpatient Transplant Evaluation

Abstract

INTRODUCTION: Patients with cirrhosis have increased risks of bleeding related to portal hypertension (including esophageal varices, and portal hypertensive gastropathy), as well coagulopathy from hepatic dysfunction, thrombocytopenia, and impaired INR. There is limited data evaluating the upper endoscopy outcomes of interventions in patient with increased MELD scores, or those undergoing transplant evaluation. Our aims were to determine upper endoscopy outcomes of intra-procedure bleeding, post-procedure bleeding, and IR intervention in patients undergoing inpatient liver transplant evaluations at a single high-volume liver transplant center. METHODS: Retrospective review of patients who underwent inpatient liver transplant evaluation at a single high-volume center were evaluated between January 2017 and July 2019, with 235 patients identified. 62 of those required an inpatient upper endoscopy as a part of their transplant evaluation. Univariate analysis with Pearson Chi-squared or ANOVA was performed. Multivariate logistic model selection was performed using AIC. The final logistic model utilized MELD, INR, platelets, and banding to predict the probability of intra or post-procedure bleeding after endoscopies during urgent liver transplant evaluation. The model was tested using the Hosmer-Lemeshow goodness of fit test. A P-value of 0.05 was considered statistically significant. RESULTS: Of sixty-two patients who underwent upper endoscopy, 9 were noted to have intra-procedure or post-procedure bleeding. Of those 9 patients that bled, 6 were related to varices, and 1 patient required an interventional radiology procedure for hemostasis. While not statistically significant, 7 of the patients with procedure-related bleeding had underlying alcoholic cirrhosis (P = 0.089). Univariate MELD differences between the two groups were not significant (P = 0.40) with group means of 23.39 for the outcome group and 20.33 for the non-outcome group. On multivariate analysis (Table 1), a platelet count > 100 demonstrated a statistically significant 0.10 times decreased odds of bleeding (P = 0.035). MELD, INR, and banding were not significant on multivariate analysis. CONCLUSION: Based on our single center results, we have demonstrated that performing upper endoscopies in patients with a high MELD can be done safely. Those with a platelet count below 100 may have a slightly increased risk of bleeding based on multivariate analysis. A larger sample size is needed to evaluate these findings along with external validation.Table 1