586 Investigation of Gastrointestinal Bleeding Among Left Ventricular Assist Device Recipients

Abstract

INTRODUCTION: Left ventricular assist device (LVAD) therapy is a form of mechanical support for patients with advanced heart failure as either a bridge to heart transplant or destination therapy. The association between gastrointestinal bleeding (GIB) and LVAD recipients is well established. There has been reported literature comparing demographics and clinical associations between patients with and without GIB. The aim of this study was to further investigate these characteristics, identify associations with GIB, and further understand endoscopic management of GIB among LVAD recipients. METHODS: This was a single-center retrospective study analyzing randomly selected patients from a population of 425 Heartmate (HMII) and 193 Heartware recipients from 2007 to 2018. Chart review was performed and data pertaining to demographics, LVAD therapy intention and settings, endoscopic reports, home medications, and echocardiogram reports was collected. Univariate analysis was performed. RESULTS: A total of 188 HMII and 134 Heartware patients were included in this study. Patients that developed GIB tended to be older (56.4 vs. 61.7 years; P = 0.0002). No differences were found in race and gender. Patients that developed a GIB had a higher baseline Cr (1.2 vs. 1.4; P = 0.0284) and lower baseline Hg (10 vs. 11; <0.001). A supratherapeutic INR was associated with GIB (14.7% vs. 26.7%; P = 0.0075). Patients prescribed angiotensin II receptor blockers had less occurrence of GIB (Table 1). Melena was the most common presenting sign. Bleeding lesions were most frequent in the small bowel (42.6%). Of all bleeding lesions identified, 73% were found with EGD and 80.4% were found with push enteroscopy. Arteriovenous malformations were the predominant etiology (Table 2). CONCLUSION: This is one of the largest single-center studies of GIB among patients with LVADs. Small bowel gastric arteriovenous malformations were the most common etiology. The majority of upper GI bleeds were within reach of a push enteroscope (proximal small bowel), which had an incremental diagnostic yield over EGD. There was a protective association with reduced occurrence of GIB among patients taking angiotensin II receptor blockers. The use of proton pump inhibitors showed no association in GIB occurrence and may be overused in this population. Further development of the association between angiotensin II receptor blockers and GIB effect in LVADs are warranted. Multivariate analyses to validate these associations are in progress.