To determine the association between delayed (>24 h) endoscopy and hospital mortality in patients with upper gastrointestinal hemorrhage (UGIH).

We retrospectively analyzed all adult patients with UGIH who underwent endoscopy in a single emergency room for 2 years. The primary exposure was defined as >24 h from the ED visit to the first endoscopy. The primary outcome was defined as all cause hospital mortality. Secondary outcomes were intensive care unit admission rate, ED length of stay, and hospital length of stay.

Among 1101 patients enrolled, 898 received endoscopy within 24 h (early group) and 203 received endoscopy after 24 h (delayed group). The hospital mortality of early and delayed group was 2.8% and 6.4%, respectively (unadjusted relative risk [RR] 2.30: 95% CI, 1.20–4.42, p = 0.012). This was significant after adjusting covariates including AIMS65 and Glasgow-Blatchford score (adjusted RR 2.23: 95% CI, 1.18–4.20, p = 0.013). Intensive care unit admission rate was not different between two groups. ED and hospital length of stay were significantly longer in delayed group.

Endoscopy performed after 24 h was associated with increased hospital mortality in UGIH. Patients in the delayed group stayed longer in the ED and in the hospital.