A prospective audit of acute pancreatitis involving nine hospitals in
the North-West Thames Region recruited 631 patients over 54 months. Th
ere were 57 deaths (9 per cent); a diagnosis had been reached in 50 pa
tients (88 per cent) before death and in seven (12 per cent) at autops
y. Eighteen patients (32 per cent) died within the first week, usually
as a result of multisystem organ failure (15 patients). Thirty-nine p
atients (68 per cent) died after the first week from complications rel
ated to infection (26 patients), co-morbid conditions (nine) or non-in
fective complications (four). Twenty-one patients (42 per cent) had be
en inadequately evaluated by Ranson's criteria, and only 22 (44 per ce
nt) of 50 with a premortem diagnosis of pancreatitis had undergone com
puted tomography (CT). Fifteen of 26 patients who died from infection-
related complications had CT and only nine underwent necrosectomy or s
urgical drainage. These data suggest that improved diagnosis, investig
ation and management of patients with acute pancreatitis is possible,
and may result in improved clinical outcome.