Endoscopic Screening and the Risk of Post Left Ventricular Assist Device Gastrointestinal Bleeding: A Multicenter Retrospective Cohort Study

Introduction: LVADs are increasingly used as life prolonging therapy in advanced heart failure. Unfortunately, post LVAD GI hemorrhage is a common complication, contributing to morbidity and increased healthcare costs. Some centers attempt to mitigate bleeding by utilizing preoperative endoscopic screening; however, the evidence to support this practice is scant. Our study aimed to evaluate whether preoperative screening endoscopy was associated with reduced risk of post-LVAD GI bleeding. Methods: A multicenter retrospective cohort study was performed at three academic sites of patients who underwent LVAD insertion from 2010 to 2019. Study participants were categorized based on whether they underwent preoperative screening endoscopy or not. Patients were followed for a year after insertion and various outcomes were collected. Primary outcome was bleeding at 1 year. Secondary outcomes were severe bleeding and intraprocedural complications. 398 patients met inclusion criteria. Baseline patient characteristics were compared by using the chi-square test for categorical variables and Wilcoxon rank sum test for continuous variables. Univariable and multivariable logistic regression analyses were performed for primary and secondary endpoints. All P values were two-tailed, with statistical significance thresholds set at p < 0.05. All statistical analyses were performed using SAS Software (version 9.4; SAS Institute Inc. Cary, NC, USA). Results: Screening cohort had increased bleeding at 1-year (36.2% v 25%, p = 0.019). After adjusting for covariables (age, race, history of GI bleed, CAD, CKD, RV dysfunction, center, PPI use & anticoagulant use) screening cohort remained at increased risk of GI bleed (aOR 1.719 [1.018-2.901], p = 0.0425). Severe bleeding was common (47.4%). Severe bleeding cohort was associated with a borderline-significant trend towards increased LV distention (P=0.083) and supratherapeutic INR (P=0.097). AVMs (55.5%) and peptic ulcer disease (37.5%) were the most common identified source of bleeding. Only 1 complication was reported with screening endoscopy (Figure 1, Table 1). Conclusion: Our study suggests that pre-LVAD endoscopic screening is safe, but increases rather than decreases risk of GI bleed post LVAD, despite controlling for known confounders. While this was an observational study and thus may not have captured all confounders, it appears that endoscopic screening may not be warranted.Figure 1.: Alluvial diagram of screening endoscopic findings and post endoscopic findings in LVAD patients presenting with GI bleed. Table 1. - Baseline Variables of Patients Who Underwent LVAD Insertion Screening P value No Yes N/Median %/IQR N/Median %/IQR Age 56 43-65 61 55-68 < 0.001 Gender 0.073 Female 70 26.92 26 18.84 Male 190 73.08 112 81.16 Race 0.017 White 99 38.08 63 45.65 Black 127 48.85 46 33.33 Hispanic 21 8.08 20 14.49 Others 13 5 9 6.52 Hx of GIB < 0.001 No 258 99.23 125 90.58 Yes 2 0.77 13 9.42 CAD 0.008 0 147 56.54 59 42.75 1 110 42.31 78 56.52 CKD 0.836 0 121 46.54 63 45.65 1 136 52.31 74 53.62 RV dysfunction 0.736 None 93 35.77 54 39.13 Mild 67 25.77 35 25.36 Moderate 65 25 29 21.01 Severe 28 10.77 18 13.04 LV distention 0.148 None 53 20.38 20 14.49 Mild 30 11.54 25 18.12 Moderate 21 8.08 15 10.87 Severe 147 56.54 75 54.35 Outpatient_PPI 0.004 No 134 51.54 53 38.41 Yes 110 42.31 82 59.42 Bleeding_within_1_year 0.019 No 195 75 88 63.77 Yes 65 25 50 36.23 Severe bleeding 0.076 No 219 84.23 106 76.81 Yes 40 15.38 31 22.46 Bleed_from Mucosal 0.596 No 225 86.54 122 88.41 Yes 35 13.46 16 11.59 Bleed_from Polyps 0.236 No 247 95 127 92.03 Yes 13 5 11 7.97 Bleed_from AVM < 0.001 No 221 85 91 65.94 Yes 39 15 47 34.06 Bleed_from Hemorrhoids 0.208 No 257 98.85 134 97.1 Yes 3 1.15 4 2.9 Bleed_from Diverticulosis 0.77 No 252 96.92 133 96.38 Yes 8 3.08 5 3.62 Anticoagulation usage < 0.001 No 7 2.69 17 12.32 Yes 252 96.92 121 87.68 Antiplatelets usage 0.097 No 16 6.15 15 10.87 Yes 243 93.46 123 89.13