Abstract
BACKGROUND: Acute upper gastrointestinal bleeding (AUGIB) is a common medical emergency with evolving demographics and management strategies, particularly in medical/endoscopic therapy and transfusion strategies. OBJECTIVE: To provide key data of the most recent 2022 UK audit and compare it with the preceding audit in 2007. DESIGN: Prospective multicentre audit conducted from 3 May to 2 July 2022, including adults (≥16 years) with AUGIB across 147 UK hospitals (response rate 86% vs 84% in 2007). AUGIB was defined by clinical symptoms (haematemesis, haematochezia, coffee ground vomiting or melaena confirmed by medical personnel). Patients were followed until discharge, death or 28 days, with re-admissions during the study period counted as new episodes. RESULTS: Among 5141 patients (59% male; median age 69), 15% had cirrhosis, 19% reported excess alcohol use, 7% used non-steroidal anti-inflammatory drugs (NSAIDs) and 46% were on antithrombotics. Most (77%) were new admissions, who were younger with fewer comorbidities, while the remainder bled during hospitalisation. Peptic ulcer disease accounted for 32% of cases, varices for 10% and no abnormality was found in 33%. Pre-endoscopic risk stratification was not performed in 42%.Compared with 2007, patients in 2022 had higher comorbidity (67% vs 50%), more cirrhosis (15% vs 9%), greater anticoagulant use (31% vs 13%) and higher transfusion rates (50% vs 43%). In 2022, among early transfusions (pre-endoscopy or within first 24 hrs; 38%), 43% were given at haemoglobin (Hb)>70 g/L, with 24% classified as inappropriate due to haemodynamic stability. A signal of harm was observed: while inappropriate transfusion was not associated with rebleeding at either 70 or 80 g/L, at 80 g/L it was linked to higher adjusted mortality (adjusted OR (aOR) 1.60, 95% CI 1.00 to 2.56).Inpatient endoscopy was more common (83% vs 74%), though endotherapy use remained modest (27% vs 23%). Salvage therapy rates were unchanged (3.3% vs 3.1%) but shifted from surgery to interventional radiology. Outcomes improved, with lower rebleeding (9.7% vs 13.3%), reduced in-hospital mortality (8.8% vs 10.0%) and shorter median stay (5 vs 6 days). In multivariate analysis, mortality was independently predicted by older age (≥80 years: aOR 2.32, 95% CI 1.64 to 3.30), shock (aOR 2.22, 95% CI 1.53 to 3.17) and comorbidity, while lower Hb at presentation increased risk (≤70 g/L: aOR 1.56, 95% CI 1.15 to 2.11). Anticoagulant use was associated with increased mortality (aOR 1.43, 95% CI 1.11 to 1.85), whereas NSAID use (aOR 0.49, 95% CI 0.25 to 0.96) and antiplatelet use (aOR 0.68, 95% CI 0.54 to 0.87) were associated with lower mortality. CONCLUSIONS: Despite a higher-risk case mix and incomplete adherence to guidelines (notably in transfusion thresholds and risk stratification), outcomes in AUGIB have improved. The observation of increased mortality with liberal transfusion above 80 g/L in stable patients reinforces the importance of restrictive transfusion practice. Quality improvement initiatives focused on risk stratification, endoscopic training and multidisciplinary care could further enhance outcomes in the UK and internationally.