Eight RCTS were included in the review (938 patients, range 40 to 194). Study quality scores ranged from 4 to 14 (median 8.5). Most patients were monitored until discharge from hospital or up to a 30-day follow-up, where reported.
Routine second-look endoscopy significantly decreased re-bleeding compared with the control group (OR 0.55, 95% CI 0.37 to 0.81; Ι²=9%; eight RCTs). Similar results were found when performing sensitivity analysis by removing one trial at a time. However, when removing two studies with high-risks patients, no significant benefit was found for routine second-look endoscopy.
The meta-analysis showed that second-look endoscopy significantly reduced the need for surgery in patients with peptic ulcer bleeding (OR 0.43, 95% CI 0.19 to 0.96; Ι²=3%; five trials) but not mortality (five trials), length of hospital stay (one trial) and blood transfusion (five trials).
Subgroup analyses showed the benefit of second-look endoscopy with varying definitions of re-bleeding. However, no significant difference in re-bleeding was found between the two groups with varying endoscopic haemostasis method and proton-pump inhibitor therapies. Re-bleeding was significantly reduced by second-look endoscopy in studies with a lower quality score but not with a higher quality score (>8). There was no evidence of publication bias.