Hospital Length of Stay Independently Predicts Mortality in Patients Emergently Admitted for Esophageal Hemorrhage: Sex, Frailty, and Age as Additional Mortality Factors

Introduction: Upper gastrointestinal bleeding results in greater than $7.6 billion of in-hospital economic burden in the United States yearly. With a worldwide incidence between 40–100/100,000 individuals and a mortality rate of approximately 2–10%, upper gastrointestinal bleeding represents a major source of mortality and morbidity. The goal of this study was to describe mortality risk factors in patients emergently admitted with esophageal hemorrhage, the second most common etiology of upper gastrointestinal bleeding. Materials and Methods: Patients emergently admitted with esophageal hemorrhage between 2005–2014 were evaluated using the National Inpatient Sample database. Patient characteristics, clinical outcomes, and therapeutic trends were obtained. Relationships between morality and all other variables were determined via univariable and multivariable logistic regression analyses. Results: In total, 4,607 patients were included, of which 2,045 (44.4%) were adults, 2,562 (55.6%) were elderly, 2,761 (59.9%) were males, and 1,846 (40.1%) were females. The average age of adult and elderly patients were 50.1 and 78.7 years, respectively. The multivariable logistic regression analysis revealed, for every additional day of hospitalization, the odds of mortality for nonoperatively treated adult and elderly patients increased by 7.5% (p=<0.001) and 6.6% (p=<0.001), respectively. Every additional year of age was associated with a 5.4% (p=0.012) increase in mortality odds for nonoperatively managed adult patients. Frailty increased the odds of mortality by 31.1% (p=0.009) in nonoperatively treated elderly patients. Undergoing invasive diagnostic procedures in conservatively treated adults reduced mortality significantly (odds ratio=0.400, p=0.021). Frailty, age, and hospital length of stay demonstrated no significant association with mortality in surgically managed adult and elderly patients. Conclusion: Nonoperatively managed patients emergently admitted for esophageal hemorrhage with longer hospital length of stay and higher modified frailty index exhibited higher odds of mortality. Invasive diagnostic procedures were negatively correlated with mortality in nonoperatively treated adult patients. Age is only associated with higher mortality rates in adults, while elderly patients revealed no association between age and mortality.