For decades, timing of endoscopy has been a controversy in the management of patients who present with upper gastrointestinal bleeding (GIB). The advent of endoscopic hemostatic therapy led to reduced further bleeding, surgery and mortality. Observational studies suggest that in patients at low risk of further bleeding, early endoscopy establishes diagnosis and allows their prompt hospital discharge. In the high-risk patients, early endoscopy with hemostatic treatment can stop bleeding and improve outcomes. Sample size in early randomized controlled trials (RCTs) was small. They included low-risk patients or patients with poorly defined risks. We designed a RCT to test the hypothesis that in high-risk patients (defined by those with an admission Glasgow Blatchford Score of 12 or greater), endoscopy within 6 h of gastrointestinal consultation, when compared to the standard of care i.e. endoscopy within 24 h, would improve outcomes. The primary outcomes, all-cause mortality at 30 days did not differ between groups; 23 of 258 (8.9%) in the urgent-endoscopy group and 17 of 258 (6.6%) in the early-endoscopy group died (difference 2.3%, 95% confidence interval -2.3 to 6.9%). Further bleeding was similar (10.9% vs. 7.8%) between groups. A higher rate in endoscopic hemostatic treatment was observed in the urgent-endoscopy group (60.1% vs. 48.4%). In patients with peptic ulcers, active bleeding or visible vessels were found on initial endoscopy in 105 of the 158 patients (66.4%) and in 76 of 159 (47.8%) in the respective group. In the majority of patients with GIB, endoscopy earlier than 24 h is not indicated.