Characterization of gastrointestinal bleeding in severely ill hospitalized patients

OBJECTIVE: To characterize the source of bleeding and the prognosis in critically ill patients with upper gastrointestinal hemorrhage that developed while in the hospital. SETTING: Intensive care units of a large academic tertiary-care center. DESIGN: Retrospective cohort study. SUBJECTS: Patients undergoing endoscopy in intensive care units for gastrointestinal bleeding that developed while in the hospital. MEASUREMENTS AND MAIN RESULTS: Medical records were available for 142 patients. Of these, 66 met the criteria for in-hospital bleeding. Peptic ulcer disease, present in 56% of patients, was the most common bleeding source identified. Of patients with peptic ulcer disease, nine of 37 (24%) had stigmata of recent hemorrhage. Ten patients (15%) received endoscopic hemostasis interventions (eight receiving therapy for bleeding ulcers, two receiving therapy for esophageal varices). The in-hospital mortality rate was 42%. The cause of death was sepsis and/or multiple system organ failure in 21 patients (75%); the gastrointestinal bleeding may have contributed to the onset of sepsis in one of these patients. No patients died directly of gastrointestinal bleeding. CONCLUSIONS: Critically ill patients who bleed while in the hospital have similar sources of bleeding and rates of endoscopically directed therapy as patients admitted to hospital with bleeding. The mortality rate is very high in patients with bleeding that develops in the hospital, and this is usually a result of systemic disease. These data may help clinicians and patients to estimate the potential benefit of urgent endoscopy in critically ill patients.