It has been suggested that the placement of endoscopic or percutaneous bili
ary stents prior to pancreaticoduodenectomy increases postoperative morbidi
ty. A retrospective review of a prospectively collected database was perfor
med. Patients undergoing preoperative biliary stenting were compared with,
patients who did not undergo stenting. In addition, outcomes after endoscop
ic and percutaneous stenting were compared. Patients who had undergone oper
ative biliary bypass prior to pancreaticoduodenectomy were excluded from th
e analysis. Between January 1994 and December 1997, 567 patients underwent
pancreaticoduodenectomy without prior operative biliary bypass. Preoperativ
e biliary stenting was performed in 408 patients (72%), whereas the remaini
ng 159 patients (28%) did not undergo biliary stenting. In the stented grou
p, 64% had stents placed via a percutaneous approach and 36% had stents pla
ced endoscopically. The stented patients were older (mean 63.1 years vs. 61
.4 years; P = 0.05) and were more likely to be white (92% vs. 82%; P = 0.00
5). Those who had stents placed were more likely to have jaundice (67% vs.
38%; P <0.0001) and fever (5% vs. 1%; P = 0.03) as presenting symptoms. The
re were no differences in multiple intraoperative parameters when the two g
roups were compared. Patients who had stents placed had a perioperative mor
tality rate of 1.7% compared to 2.5% in those who did not (P = 0.3). Althou
gh the overall complication rates were 35% in those who had stents placed a
nd 30% in those who did not (P = NS), patients with stents experienced a si
gnificantly increased incidence of pancreatic fistula (10% vs. 4%; P = 0.02
) and wound infection (10% vs. 4%; P = 0.02). The incidences of other posto
perative complications were similar between the stented and unstented group
s. Eight patients (3%) in the percutaneously stented group developed signif
icant hemobilia after stent placement, whereas none of the patients undergo
ing endoscopic stent placement developed hemobilia (P = 0.03). There were n
o statistical differences in other complications between the percutaneously
and endoscopically stented groups. Preoperative biliary stenting did not i
ncrease the overall complication rate or mortality rate in patients undergo
ing pancreaticoduodenectomy. Stenting does appear to increase the rate of p
ancreatic fistula formation, possibly as a result of pancreatic inflammatio
n related to the stenting procedure. Stenting also increases the rate of wo
und infection, likely secondary to contaminated bile (bactibilia) after ins
trumentation of the biliary tree. Preoperative biliary stenting is safe but
should be used selectively because of the above-mentioned risks. The metho
d of stenting should be based on local expertise.