Gastrointestinal Hemorrhage among LVAD Recipients. A Nationwide Inpatient Sample Study

Abstract

Purpose Patients with left ventricular assist device (LVAD) have high incidence of gastrointestinal bleeding (GIB). We sought to investigate in-hospital outcomes of patients with LVAD who were admitted with GIB. Methods We utilized data from the National Inpatient Sample and the US Census Bureau to calculate annual national rates of GI bleeding admissions, in-hospital mortality, length of stay and hospitalization cost among LVAD patients with GIB for years 2011 to 2014. Results We identified 695 (weighted n=3440) LVAD patients admitted with GIB with LVAD and 5047 (weighted n=24,903) admitted without GIB. Patients with GIB were older (GIB: 62.5 years vs. non-GIB 57.99 years, P<0.0001) and higher comorbidity burden as assessed by Elixhauser score (GIB: 4.63 vs. non-GIB 3.96, P<0.0001). We found a small but statistically significant decline of in-hospital mortality among GIB patients (4.96% in 2011 vs. 3.85% in 2014, p=0.03). In-hospital mortality did not differ significantly between patients admitted with and without GIB (5% among GIB patients vs. 4.8 among non-GIB patients, p=0.8). Although length of stay was longer among GIB patients (10.7 days vs. 9.2 days, p=0.006), the total hospital cost did not differ significantly ($31768 for GIB vs $29345 for non-GIB, p=0.3). Conclusion In-hospital mortality from GIB among LVAD recipients decreased during the study period. GIB admissions were associated with similar in-hospital mortality and cost but slightly more prolonged length of stay compared with non-GIB admissions among LVAD recipients. Patients with left ventricular assist device (LVAD) have high incidence of gastrointestinal bleeding (GIB). We sought to investigate in-hospital outcomes of patients with LVAD who were admitted with GIB. We utilized data from the National Inpatient Sample and the US Census Bureau to calculate annual national rates of GI bleeding admissions, in-hospital mortality, length of stay and hospitalization cost among LVAD patients with GIB for years 2011 to 2014. We identified 695 (weighted n=3440) LVAD patients admitted with GIB with LVAD and 5047 (weighted n=24,903) admitted without GIB. Patients with GIB were older (GIB: 62.5 years vs. non-GIB 57.99 years, P<0.0001) and higher comorbidity burden as assessed by Elixhauser score (GIB: 4.63 vs. non-GIB 3.96, P<0.0001). We found a small but statistically significant decline of in-hospital mortality among GIB patients (4.96% in 2011 vs. 3.85% in 2014, p=0.03). In-hospital mortality did not differ significantly between patients admitted with and without GIB (5% among GIB patients vs. 4.8 among non-GIB patients, p=0.8). Although length of stay was longer among GIB patients (10.7 days vs. 9.2 days, p=0.006), the total hospital cost did not differ significantly ($31768 for GIB vs $29345 for non-GIB, p=0.3). In-hospital mortality from GIB among LVAD recipients decreased during the study period. GIB admissions were associated with similar in-hospital mortality and cost but slightly more prolonged length of stay compared with non-GIB admissions among LVAD recipients.