Background Ageing populations are increasing in many countries and ble
eding peptic ulcers in patients older than 60 years carry a greater ri
sk of rebleeding and death. This study aimed to identify the risk fact
ors for rebleeding and death in very elderly patients with peptic ulce
r bleeding. The efficacy of treatment in preventing recurrent bleeding
and death in this group of patients was also studied by means of pros
pective data collection and analysis. Methods Data relating to 1744 pa
tients treated between September 1985 and January 1994 for peptic ulce
r bleeding were collected prospectively and analysed. Patients were st
ratified by age to one of three groups: group 1 (less than 60 years, n
= 833), group 2 (60-79 years, n = 706) and group 3 (80 or more years,
n = 205). Results Univariate and multivariate analyses of 21 factors
possibly affecting either rebleeding or death identified age greater t
han 80 years as one of the factors significantly affecting rebleeding
and death. In a comparison of groups 1, 2 and 3, the likelihood of reb
leeding and death was significantly greater in group 3. Univariate and
multivariate analyses for rebleeding and death were performed for eac
h group. The severity of initial bleeding had a marked bearing on subs
equent rebleeding rates for all three groups. In group 3, however, lar
ge ulcer size and impaired liver function were additional factors whic
h correlated significantly with final outcome. No rebleeding or morbid
ity occurred when endoscopic treatment was performed early for patient
s in group 3 but there was a significantly greater risk of further rec
urrent haemorrhage and treatment-related morbidity when treatment was
performed after the onset of rebleeding. Conclusion Patients aged 80 y
ears or greater had the highest risk of rebleeding and death. For pati
ents below 80 years of age, significant factors related to a fatal out
come included co-morbid illness, complications and the need for mechan
ical ventilation. For patients aged 80 years or older, the significant
factors were ulcer size greater than 2 cm and admission with serum bi
lirubin level above 20 mmol/l. Endoscopic treatment for the very elder
ly was effective if carried out early.