Introduction

Acute upper gastrointestinal bleeding(AUGIB) is a common emergency, with 10% linked to varices¹, requiring swift evaluation.

Aims & Methods

We conducted a prospective multi-center study from 3 May to 2 July 2022, involving adults (>16 years old) presenting with AUGIB in 152 UK hospitals. We evaluated the patient characteristics of those identified to have varices on index endoscopy as the cause of bleeding and used regression analysis to identify predictors of rebleeding and mortality.

Results

Of 4278 AUGIB cases undergoing endoscopy, 10% (418/4278) were noted to have varices(91% oesophageal, 16% gastric, and 2% duodenal). Median age was 58 yrs and 80% (338) were new admissions. At presentation, 76%(330) had known chronic liver disease (Child-Pugh categories: A:3%, B: 85%, C: 12%.), 13%(54) were on antiplatelets and 15%(61) on anticoagulants. At presentation, 35%(145) had a history of previous AUGIB, 45%(189) were on non-selective beta-blocker prophylaxis and 23%(98) were on a variceal band ligation programme. Glasgow-Blatchford score (GBS) stratified patients into 13% medium-risk (GBS 2-6), 44% higher-risk (7-11), and 29% very high-risk (≥12). For pre-endoscopy management: 66%(278) received PPI, 65%(269) vasopressors (98% terlipressin, 2% octreotide) and 56%(235) antibiotics. Endotracheal intubation for endoscopy was required for 33%(137) and admission under critical care for 20%(82). Time to endoscopy from presentation: 0-12 hrs:36%, 12-24 hrs:24%, and >24 hrs:32%. Stigmata of recent bleeding observed in 72%(300) – 41% blood in the upper GI tract and 64% high risk markings on varices. Variceal therapy applied in 67%(281 – 91% banding, 10% injection therapy), Danis stent 2%(7) and Sengstaken tube 3%(14). 17% (69) had evidence of further in-patient bleeding after index endoscopy. 26%(109) needed >1 endoscopy during inpatient stay, and 2%(8) underwent transjugular intrahepatic portosystemic shunt. 67%(281) were transfused ≥1 packed red blood cells, 15%(66) platelets, and 20%(86) fresh frozen plasma. Median length of stay was 7days(IQR 4-11) and in-hospital mortality was 13%(54). In a multi-variable logistic model assessing predictors of rebleeding and mortality, significant factors for rebleeding included use of pre-endoscopic vasopressors (OR=2.49,95%CI:1.59-3.88), pre-endoscopic antibiotics (OR=0.4,95%CI:0.28-0.59) and presence of endoscopic stigmata (OR=6.89,95%CI:5.11-9.30). For mortality, the significant factors were very high-risk GBS (OR=4.83,95%CI:2.71-24.09), presence of endoscopic stigmata (OR=3.8,95%CI:2.82-138.72) and the use of endotherapy (OR=0.04,95%CI:0.001-0.53).

Conclusion

Pre-endoscopic vasopressors were associated with increased risk of rebleeding, while pre-endoscopic antibiotics with decreased risk. Endoscopic stigmata was significantly associated with both rebleeding and mortality, with endotherapy a crucial influencer for reducing mortality. These findings offer valuable insights for risk stratification and guide appropriate therapeutic interventions in managing AUGIB with varices as per international guidelines.

References

1. Hearnshaw et al Gut 2011