Background Acute upper gastrointestinal haemorrhage is a common medical emergency and hospital admission has usually been regarded as obligatory until the risk of further haemorrhage has receded. This policy means that some patients at low risk stay in hospital for longer than is necessary especially when delayed. We attempted to negligible risk of further bleeding or death and for whom early discharge or even outpatient management would be possible with no adverse effect on standards of care. Methods We used a validated risk scoring system based on age (score 0-2), presence of shock (0-2), comorbidity (0-3), diagnosis (0-2), and endoscopic stigmata of recent haemorrhage (0-2); the maximum possible score was 11. We studied patients identified through the UK National Audit of acute upper gastrointestinal haemorrhage; they had been admitted with upper gastrointestinal haemorrhage to hospitals in the UK during 4 months of 1993. This analysis was based on the 2531 patients from the national audit who underwent endoscopy after an acute admission. Findings 744 (29.4%) of the 2531 patients scored 2 or less on the risk score. Of these patients only 32 (4.3% [95% CI 3.0-6.0] rebled and only one (0.1% [0.006-0.75] died). Thus, the risk score identifies patients at low risk of rebleeding or death. The median hospital stay for these patients was 4 days; duration of hospital stay increased with risk score. Within risk score categories of 5 or less, there was a trend of increasing hospital stay as the time between admission and endoscopy increased. Interpretation Our risk score identifies a large proportion of patients with acute upper gastrointestinal haemorrhage who are at low risk of further bleeding or death. Early endoscopy and discharge of such patients could allow substantial resource savings.
