Preadmission EGD significantly reduces the hospitalization rate in acute upper gastrointestinal bleeding (UGIB) without effecting outcome: A prospective randomized controlled study

Purpose: to determine whether preadmission EGD performed in the emergency department (ED) reduces the hospitalization rate in UGIB. Methods: consecutive stable patients with UGIB hospitalized from the ED were randomized to EGD or control. Controls were admitted to the ICU, stepdown, or ward using available clinical criteria and underwent EGD on an elective basis- usually within 24 hrs. EGD group was admitted using EGD and clinical data; stable patients with clean based lesions were discharged, those with stigmata were admitted to the ICU or stepdown, and the rest to the ward. Patients with varices, coagulopathy, hemodynamic instability, severe comorbidity requiring ICU care, or unable to consent were excluded. All control and EGD patients were prospectively followed for 30 days to obtain clinical outcome data. Results: From 9/96-12/96, 72 of 3226 consecutive patients admitted from the ED had a primary diagnosis of UGIB; 32 met one or more of the exclusion criteria and 40 were randomized. The mean age (46 vs 56y) and admission hemoglobins (11.4 vs 12.5 g/dl) were comparable for the HGD and control patients and ulcer, Mallory Weiss tear, and esophagitis accounted for 82-85% of the bleeding sources. Controls underwent endoscopy later compared to the EGD group (median 1 vs 6.5 hrs). EGD data changed the admission decision in 75% of patients in the EGD group: 10 patients were discharged (8 were to have been admitted to the ward, and one each to ICU and stepdown), 3 were "downgraded" (ICU-stepdown, ICU-ward. stepdown-ward), and 2 were "upgraded" (ward-ICU, stepdown-ICU). Randomization Location of initial admission ICU Stepdown Ward Outpatient Pre-EGD admission 15% 30% 55% 0 Post-EGD admission 10% 20% 20% 50% Control 25% 30% 45% 0 *p<0.05. The prevalence of endoscopie stigmata (30 vs 20%) and transfusion requirements (2 vs 2u) for hospitalized patients were no different between the EGD and control groups. There were no significant differences in rebleeding (0 vs 1), re-EGD (0 vs 1), surgery (0 vs 1). mortality (0 vs 1), unplanned repeat hospitalization (1 vs 0) and doctor visits (2 vs 2) between EGD and control groups on 30 day prospective followup. All unplanned hospitalization and doctor visits were for non GI causes. Conclusions: data from the preadmission EGD can change the admission decision in 75% of patients with UGIB and reduce the hospitalization rate 50%, without effecting the clinical outcomes. Preadmission EGD may streamline the care of patients with UGIB and improve utilization of resources.