To evaluate the rationale of using antibiotics in acute pancreatitis a
nd to determine whether the indication for their use depends upon the
etiology of the pancreatitis, the records of 202 patients with acute p
ancreatitis were retrospectively reviewed. The incidence of abnormal b
ody temperature, leukocytosis, bacteremia and the results of biochemis
try tests in different etiologies of the disease were investigated. Pa
ncreatitis was found to be alcohol-related (47 patients), gallstone-re
lated (105 patients), idiopathic (26 patients) and miscellaneous (24 p
atients). On admission, 83 patients had abnormal body temperature and
146 patients showed leukocytosis. Bacteremia occurred in 20 patients.
Of these, 15 had gallstone-related pancreatitis, two had pancreatic ca
ncers and one developed bacteremia after endoscopic retrograde cholang
io-pancreatography (ERCP). These 18 patients had abnormal biochemistry
results (including high serum levels of direct bilirubin, alkaline ph
osphatase and gamma-glutamyltransferase) and dilated bile ducts on ima
ging studies, indicating biliary infections. The remaining two patient
s with bacteremia included one alcoholic patient and one patient with
idiopathic pancreatitis. The most commonly involved pathogens were Esc
herichia coli and Klebsiella pneumoniae. In addition, eight patients (
4%) developed secondary pancreatic infections during hospitalization;
the blood cultures of seven of these patients were negative on admissi
on. Although fever and leukocytosis are not good predictors of infecti
on in acute pancreatitis our results showed that bacteremia is common
in patients whose pancreatitis is related to gallstones, ERCP or pancr
eatic malignancy with obstructive jaundice. We recommend that antibiot
ics be used only in this subset of acute pancreatitis patients.