Of 279 patients admitted to a specialist unit with acute pancreatitis,
210 were admitted directly and 69 were transferred for treatment of l
ocal or systemic complications. Outcome was assessed in terms of morta
lity and morbidity and in relation to aetiology, predicted severity of
disease (modified Glasgow score), organ failure (modified Goris multi
ple organ failure score), and need for surgical intervention. The deat
h rate was 1.9% in patients admitted directly but was 18.8% in those t
ransferred hom other units. Mortality in gall stone related pancreatit
is was 3% compared with 15% (p=0.03) in pancreatitis of unknown aetiol
ogy and 27% (p=0.01) in post-endoscopic retrograde cholangiopancreatog
raphy pancreatitis. Mortality was related to age (mortality >55 years
old 11% v 2%; p=0.003) and Goris score (score 0, mortality 0% v score
5-9, mortality 67%; p=0.001). In patients transferred fi om other unit
s, mortality was 11% in those transferred within a week of diagnosis a
nd 35% when transfer was delayed p=0.04). Thirty six patients had panc
reatic necrosis on dynamic computed tomography of whom 29 underwent pa
ncreatic necrosectomy with a 34% mortality. Mortality was related to t
he modified Goris score (median score 2 in survivors v 6 in non-surviv
ors; p=0.005) and was higher when necrosectomy was performed within th
e first two weeks of admission (100% v 21%; p=0.004). In conclusion, m
ortality in acute pancreatitis is influenced by age, aetiology of the
disease, and presence of organ failure. Patients transferred for speci
alist care have a 10-fold greater mortality than those admitted direct
ly and mortality is greatest when transfer is delayed. Early necrosect
omy carries a prohibitively high mortality.