Implanted Left Ventricular Assist Devices Impacts Hospital Resource Utilization Without Affecting Patient Mortality: A Population-Based Analysis

Abstract

Introduction: Left Ventricular Assist Devices(LVADs) are increasingly being utilized for management of end stage heart failure. Gastrointestinal (GI) bleeding (GIB) is one of the most common adverse events following LVAD implantation. The aim of this study was to investigate the impact of LVAD implants on in-hospital outcomes of patients admitted with GIB. Methods: This was a cross-sectional study using the Nationwide Inpatient Sample (NIS) 2010-2012. The NIS was queried to identify all adult (≥18years) admissions with a primary diagnosis of GIB. All GIB admissions were sub-divided into patients with (cases) and without LVADs (controls). The main outcome measurements included in-hospital mortality, length of stay, and hospitalization costs. Multivariate analysis controlled for demographic, hospital and etiological differences, site of GIB, and patient comorbidities. Results: There were a total of 1,002,299 total hospitalizations for GIB from 2010-2012 in the United States. Patients with LVAD implants accounted for 1,112 (0.11%) of all admissions for GIB. Univariate analysis revealed that patients with LVADs admitted with GI bleeding were younger, more often male, with higher comorbidities, and treated at large volume, urban teaching hospitals compared to patients without LVADs. Early endoscopy (< 1 day) was less frequently (37.39% vs. 66.01%, p < 0.001) performed in patients with LVADs than without LVADs. Angiodysplastic lesions were the most common etiology of bleeding in patients with LVADs, occurring in 35.4% of patients. Whereas for patients without LVADs, peptic ulcer disease (27.10%) and diverticular disease (14.77%) were the two most common etiologies of GIB. (Table 1) On multivariate analysis, presence of LVADs was not associated with increased mortality in GIB (OR 0.53, 95% CI 0.07-4.15); however, LVADs independently accounted for prolonged hospitalization (by 3.5 days, p < 0.001) and higher hospital costs ($37,032, p = 0.01). Conclusion: In patients admitted with GIB, presence of LVAD implants accounts for higher healthcare utilization, but is not adversely associated with mortality despite therapeutic anticoagulation, increased comorbidities, and delayed endoscopy. These observations are pertinent for current clinical practice and future research.Figure 1