Traumatic duodenal perforations in children pose a diagnostic and ther
apeutic challenge. To identify specific diagnostic criteria and define
an optimal therapeutic protocol, we reviewed all duodenal injuries tr
eated at our institution in the past 10 years. There were 14 hematomas
and 13 perforations, The diagnosis was confirmed by computed tomograp
hy (CT), ultrasound scan (US), upper gastrointestinal contrast studies
(UGI), or at laparotomy. The clinical findings and CT findings of the
two groups were compared, Children with suspected duodenal hematomas
were treated expectantly, and children with duodenal perforations were
treated surgically, Twenty-five associated injuries(10 pancreatic) oc
curred in 19 children, Children with perforations had higher injury se
verity scores (ISS) (25 v 9), but the two groups could not be differen
tiated based on presenting signs, symptoms, or laboratory findings. CT
findings of retroperitoneal air or contrast were seen in 9 of 9 perfo
rations and in 0 of 10 hematomas. CT findings of intraabdominal or ret
roperitoneal fluid, mesenteric enhancement, and thickened duodenal wal
l did not differentiate the two groups. Duodenojejunostomy was perform
ed in one patient, and primary repair was performed in 11 children who
had perforation, In five children, duodenostomy tube drainage with fe
eding jejunostomy or gastrojejunostomy were added. Complications occur
red in three of four children in the first 5 years of the study and in
two of nine children in the last 5 years. The decreased morbidity rat
e correlated with reduced time to definitive therapy (28 v 7.8 hours),
Duodenal fistulae resulted in three of seven children treated without
duodenostomy tube drainage and zero of five treated with drainage. En
teral feeds resumed faster (average, 12 v 27 days) if repair of perfor
ation was combined with feeding jejunostomy or pyloric exclusion and g
astrojejunostomy. Children with duodenal hematoma resumed eating an av
erage of 16 days after injury. Only one child required surgery for per
sistent obstruction. The findings of retroperitoneal air and contrast
extravasation on CT accurately distinguish duodenal perforation from h
ematoma. Conservative management of hematoma is safe and effective. Pr
imary repair of perforation with duodenal drainage results in fewer po
stoperative complications, and gastrojejunostomy or feeding jejunostom
y shorten the time to resumption of feeds. Copyright (C) 1997 by W. B.
Saunders Company.