Introduction In acute severe gastrointestinal bleeding, endoscopy has a failure rate of 10–30% in detecting active bleeders and achieving haemostasis. Further aggressive treatment may be necessary and mesenteric angiography has the advantage of being both a therapeutic and diagnostic modality. Previous outcome data on patients undergoing angiography report a normal angiogram in 52% (due to the intermittent nature of bleeds), with lower rebleed rates than in those with positive extravasation, but a higher mortality. Where there was positive extravasation and subsequent embolization, a rebleed rate of 9–47% is reported, with a surgery rate of 0–35% and 30 day mortality of 3–27%. We would like to report our department’s experience of the outcomes of acute severe gastrointestinal bleeds requiring mesenteric angiography.

Methods Over a 5 year study period between 2006–2011 in the general intensive care unit of St George’s Hospital, patients were included if they had had a non-variceal acute severe gastrointestinal bleed that required mesenteric angiography after failed endoscopy. Retrospective data was collected on patient demographics, endoscopic findings, angiographic findings, embolization, rates of rebleeding, number of packed red blood cells (PRBC) transfused, hospital stay, and mortality at 30 days.

Results 26 patients were included, 18 male and 8 female. Overall, each patient needed an average of 22.2 units of PRBCs; rebleeding occurred in 28.9%; surgery was performed in 6 patients (including 4 colectomies); 30 day mortality was 35% (n = 9) and 90-day motality was 50% (n = 13). Subgroup analysis showed that for the 18 patients with positive angiographic findings and subsequent embolization, 10 had positive endoscopic findings (4 gastric ulcers, 5 duodenal ulcers, 1 Dieulefoy lesion) whilst 8 had inconclusive findings at endoscopy. For this group, an average of 18 units PRBCs were transfused, the rebleed rate was 35%, subsequent surgery was performed in 24%, 30 day mortality was 44% (n = 8) and 90 day mortality 54% (n = 10). For the 8 patients (27%) with a negative angiography, all had a negative endoscopy. Only 13% required surgery (n = 1), an average of 11 units was transfused, the rebleed rate was 13% (n = 1), 30 day mortality 13% (n = 1) and 90 day mortality 38% (n = 3).

Conclusion Overall, this group of patients needing angiography have a very high risk of mortality. Only 31% of our patients had a normal angiogram, suggesting a more astute selection of cases. Interestingly, for those with a negative angiography, the rates of rebleeding and surgery were lower than in the positive extravasation group, but the mortality is also lower, a finding in contrast to previous studies.

Disclosure of Interest None Declared.