Introduction Patients admitted out of hours or at weekends may have an excess mortality due to the fact that they undergo emergency endoscopy by junior staff. We retrospectively looked at the predictors of mortality in patients undergoing emergency endoscopy for severe bleeding in Leeds.

Methods The survival of patients with the most significant upper GI bleeding lesions who underwent emergency endoscopy in Leeds between end of April 2008 and middle of December 2011 were selected for retrospective analysis using data from our endoscopy reporting system and hospital records.

Results A total of 1663 emergency endoscopies were carried out during this period. From this number, 390 examinations (23% of total) were selected for more careful scrutiny as the following significant bleeding lesions had been recorded; 136 duodenal ulcers, 84 gastric ulcers, 134 oesophageal varices, 25 gastric varices, 11 Dieulafoy’s and 9 other bleeding lesions.

After 23% of procedures (91/390), the patient died within 30 days. As expected, patients who died had a significantly higher Rockall score (7.5 vs. 5.9 p < 0.0001), a higher ASA level (3.5 vs. 2.8 p < 0.001) and a lower systolic BP at the time of the examination (95 vs. 102 p = 0.025). Patients who died following endoscopy for bleeding ulcers were significantly older than those who survived (76.6 vs. 67.2 yrs, p = 0.006). There was no significant difference in mortality with the type of bleeding lesion, Hb (7.9 vs. 8.0) or heart rate (100 vs. 102 bpm) at the time of the endoscopy between those who survived and those who died.

Furthermore, undergoing an emergency gastroscopy at night or during the weekend or bank holiday was not associated with an increased risk of death (P = 0.24 and p = 0.53 respectively).

Whether the examination was carried out by an SpR or a Consultant made no difference to the survival of the patient. The only link between endoscopic intervention and patient outcome was the finding that patients with varices requiring balloon tamponade where significantly more likely to die (15/36 vs. 16/98 p = 0.01).

Conclusion Our study had the statistical power to detect all the recognised risk factors for death following admission with an acute upper GI bleed including advancing age, increasing comorbidity and hypotension. We found no evidence that undergoing an emergency endoscopy at night or during the weekend or a bank holiday had any adverse effect on outcomes. Similarly, the level of seniority of the endoscopist did not affect outcomes.

Disclosure of Interest None Declared.