Background: Gastrointestinal bleeding (GIB) is a common and serious complication in patients with left ventricular assist devices (LVADs), contributing to significant morbidity, prolonged hospitalization, and increased healthcare costs. We evaluated national trends, demographic disparities, and outcomes of GIB in hospitalized LVAD patients. Methods: We analyzed adult (≥18 years) LVAD hospitalizations in the National Inpatient Sample (2016–2021), identifying internal LVADs using ICD-10-PCS code 02HA0QZ. GIB was defined using ICD-10-CM codes and classified into upper (UGIB) and lower (LGIB) sources. Survey-weighted logistic and linear regression models assessed associations with mortality, length of stay (LOS), and total charges. Subgroup analyses explored sex and racial disparities. Results: Among 20, 785 weighted adult LVAD admissions, 9. 8% had GIB. Of these, 72. 3% had LGIB and 31. 0% had UGIB. Patients with GIB were older (59. 2 vs. 54. 8 years) and more likely to be female (43% vs. 40%) and Black (9. 2% vs. 7. 8%). GIB was associated with longer LOS (+15. 3 days, 95% CI: 12. 0–18. 5), higher charges (+316, 031, 95% CI: 212, 435–419, 627), and greater in-hospital mortality (OR 1. 69, 95% CI: 1. 25–2. 29; p < 0. 001). Female patients with GIB had higher odds of mortality (OR 1. 37) and increased LOS (+5. 6 days), though this was not statistically significant. Racial disparities were evident: Black patients with GIB had longer LOS (+8. 9 days), while Asian/Pacific Islander patients had shorter LOS (–23. 3 days, p < 0. 001). From 2016 to 2021, GIB prevalence rose modestly (from 9. 4% to 10. 7%, p = 0. 33), with no significant change in mortality trends (p = 0. 13). Conclusions: GIB complicates nearly 1 in 10 LVAD hospitalizations, with lower GI bleeds being most common. GIB is independently associated with higher mortality, LOS, and costs. Persistent gender and racial disparities highlight the need for targeted strategies to improve outcomes in this high-risk population.
Basetty et al. (2026) studied this question.