Endoscopy within six hours does not improve outcomes in non-variceal upper GI bleeding

Design

Systematic review and meta-analysis

Population

10,785 adults with non-variceal upper GI bleeding in 9 studies

Primary outcome

Mortality and rebleeding

Effect

Mortality OR 0.88 (0.58 to 1.34); ICU admission OR 1.41 (1.20 to 1.66)

This meta-analysis pooled nine studies of 10,785 adults with non-variceal upper gastrointestinal bleeding, comparing endoscopy within 6 hours (2,638) with later endoscopy (8,147), most within the 24-hour window guidelines recommend.

Ultra-early endoscopy did not reduce mortality (OR 0.88, 95% CI 0.58 to 1.34) or rebleeding (OR 1.23, 0.90 to 1.67), and made no difference to transfusion, surgery, repeat endoscopy or length of stay. ICU admission was more frequent in the ultra-early group (OR 1.41, 1.20 to 1.66). Results were consistent in high-risk subgroups, most often defined as Glasgow-Blatchford score above 12.

For surgeons covering acute admissions, this supports resuscitating first and scoping within 24 hours, rather than pushing for an out-of-hours list for every bleed. The pooled studies are mostly observational, and patients scoped very early may have been sicker, which could explain the ICU finding. Haemodynamically unstable patients who do not respond to resuscitation still need urgent endoscopy or intervention.

• Resuscitate and risk-score with Glasgow-Blatchford before arranging endoscopy for non-variceal bleeding.
• Aim for endoscopy within 24 hours in stable patients; this analysis found no gain from scoping within 6 hours.
• Escalate immediately for patients who stay unstable despite resuscitation; they were not the question here.
• Involve interventional radiology and surgery early when endoscopic control fails.

Why it matters

It weakens the case for routine out-of-hours endoscopy lists for stable upper GI bleeds.

Don't overread it

The pooled studies are largely observational, so sicker patients may have been scoped earlier.

The statistics, in plain English

An odds ratio of 0.88 for mortality with an interval from 0.58 to 1.34 means the data are compatible with anything from a 42% reduction to a 34% increase — no clear effect either way. The ICU interval (1.20 to 1.66) stays above 1, but that may reflect who was chosen for early scoping rather than harm from it.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.