595

M yer H. Roszler, M .D.

W illiam L. Cam pbell, M .D.
Post-ERCP Pancreatitis: Association
with Urographic Visualization
during ERCP1
To investigate the possible association of
urographic visualization and acinariza-
tion of contrast m aterial with postproce-
dure pancreatitis, 140 consecutive endo-
scopic retrograde cholangiopancreatogram s
(ERCP) with pancreatic duct filling were
reviewed. Urographic visualization was
identified in 29 patients (21% ); pancreati-
tis developed in 13 of these patients
(45% ). Pancreatitis occurred in five of 111
patients (4% ) without urographic visual-
ization. Of 19 patients who dem onstrated
both acinarization and urographic visual-
ization, ten (53% ) had postprocedure pan-
creatitis. Twenty-six patients exhibited
acinarization without urographic visual-
ization; one (4% ) had pancreatitis. Uro-
graphic visualization during ERCP is
probably m ore com m on than generally
recognized and indicates patients who are
at high risk for postprocedure pancreati-
tis. Although acinarization accom panied
by urographic visualization is associated
with a high risk of pancreatitis, acinariza-
tion alone was not associated with this
com plication in this study. Detection of
renal opacification during ERCP requires
close scrutiny of film s and is best accom -
plished on overhead radiographs.
Index terms: Contrast media, complications, 77.44.
Endoscopic retrograde cholangiopancreatography
(ERCP), 77.1229 #{149} Pancreatitis, 77.291
Radiology 1985; 157:595-598
From the Department of Radiology, Presbyterian-
University Hospital, Pittsburgh. Presented at the 70th
Scientific Assembly and Annual M eeting of the Radio-
logical Society of North America, W ashington, D.C.,
November 5-30, 1984. Received December 4, 1984;
revision requested February 15, 1985, and received
June 4, 1985; accepted June 17, 1985.
RSNA, 1985
CUTE pancreatitis is the most frequent, clinically significant
complication following endoscopic retrograde cholangiopan-
creatography (ERCP) (1). Attacks of postprocedure pancreatitis
have been observed with frequencies varying between .5% for
ERCP and 12% for endoscopic retrograde pancreatography (ER?) (2,
3). Although the exact causes of post-ERCP pancreatitis are uncer-
tam, overdistention of pancreatic ducts, pancreatic duct manipula-
tion, and acinar rupture have been implicated (1, 4). Radiographic
signs suggestive of ductal overdistention include acinarization (pa-
renchymal opacification) and renal excretion of contrast material.
Previous investigations have suggested that postprocedure pancre-
atitis may be more likely to develop in patients exhibiting acinari-
zation (3, 5-7). W hether urographic visualization is predictive of an
increased risk of post-ERCP pancreatitis has been unclear (1, 8, 9).
At our institution, we have occasionally observed urographic
visualization at the time of ERCP. Postprocedure pancreatitis devel-
oped in some of these patients during a period when the incidence
of this complication at our hospital seemed to be rising. If urogra-
phic visualization during or after ERCP were demonstrated to sig-
nify an increased risk of pancreatitis, early therapeutic precautions
could be initiated in selected patients. W e therefore undertook a
retrospective review of our experience with ERCP to investigate the
possible association of urographic visualization and acinarization
of contrast material with postprocedure pancreatitis.
M ATERIALS AND M ETHODS
W e retrospectively studied 140 consecutive patients who had undergone
ERCP in which pancreatic duct opacification was achieved. Patients in
whom the pancreatic duct was not visualized were excluded since previous
experience suggested that postprocedure pancreatitis is rare without pan-
creatic duct injection (1).
Endoscopic cannulation of the papilla of Vater was performed by or
under the direction of one of several staff gastroenterologists. The endo-
scope used was the Olympus JF-1T model (New Hyde Park, N.Y.). M edica-
tions given included intravenous diazepam (Valium; Roche Products), me-
peridine hydrochloride, (Demerol Hydrochloride; W inthrop-Breon) and
glucagon (Lilly). Prophylactic antibiotics were not administered. The con-
trast medium employed in the pancreatic duct was Renografin-60 (diatri-
zoate meglumine and diatrizoate sodium solution; Squibb). For biliary in-
jection, the contrast material was diluted 1:1 with sodium chloride injection
(USP). All examinations were performed under fluoroscopic guidance with
one of several staff radiologists present.
Radiographs on each patient included a preliminary anteroposterior su-
pine abdominal film, multiple spot films, and an anteroposterior supine
abdominal film obtained at the conclusion of the procedure. In 20 patients,
an immediate postexamination lateral radiograph was also obtained. All
films were evaluated for the amount of pancreatic duct filling, acinariza-
tion, and urographic visualization.
The patients medical records were reviewed for past history of pancreati-
tis, development of postprocedure pancreatitis, treatment of pancreatitis,
596. Radiology Decem ber 1985
and final diagnosis. M inimal criteria for
the diagnosis of post-ERCP pancreatitis
were characteristic abdominal pain and
tenderness accompanied by hyperamylas-
emia within 48 hours of the procedure.
Only patients with symptoms and signs
severe enough to warrant the diagnosis of
post-ERCP pancreatitis by the referring
gastroenterologists were considered to
have pancreatitis in this study.
Groups were compared for statistical
significance using the x2 test.
RESULTS
Twenty-nine of 140 patients (21% )
exhibited urographic visualization
(Fig. 1). In 27 patients, a pyelogram
was seen; in two patients, only the
bladder was opacified on delayed ra-
diographs taken within 12 hours. Py-
elograms were usually seen on post-
procedure anteroposterior or on
lateral overhead radiographs and
were infrequently identified on spot
films. In two individuals, a pyelo-
gram was identified only on lateral
overhead radiographs taken immedi-
ately after ERCP (Fig. 2).
Postprocedure pancreatitis devel-
oped in 13 of the 29 (45% ) patients
who exhibited urographic visualiza-
tion. Pancreatitis developed in only
five of 111 patients (4% ) who did not
have urographic visualization (P <
.005). Ten of the 13 patients with uro-
graphic visualization and postproce-
dure pancreatitis exhibited acinariza-
tion. The overall occurrence of
pancreatitis in our series was 13%
(18/140), with 72% (13/18) of the pa-
tients with postprocedure pancreati-
tis also exhibiting urographic visual-
ization. Of 122 patients without
pancreatitis, 16 (13% ) exhibited uro-
graphic opacification (P < .005).
Of the 18 patients with post-ERCP
pancreatitis, six were considered to
have relatively mild pancreatitis, con-
sisting of moderate abdominal pain
and elevated serum amylase. The oth-
er 12 patients had more severe pan-
creatitis characterized by marked
abdominal pain, tenderness, and hy-
peramylasemia. Some had fever and
rebound abdominal tenderness.
Treatment included nasogastric suc-
tion and intravenous fluids.
Of 32 patients with a past medical
history of pancreatitis, seven (22% )
had post-ERCP pancreatitis. Diag-
noses in these seven individuals were
posttraumatic noncommunicating
pseudocysts in two, prior acute pan-
creatitis in two, and chronic pancre-
atitis in three. Of 108 patients without
a history of pancreatitis, postproce-
dure pancreatitis occurred in 11(10% )
Acinarization, with or without uro-
graphic visualization, was observed
in 45 (32% ) patients. The degree of
acinarization varied from focal areas
of parenchymal opacification to in-
volvement of virtually the entire pan-
creas (Fig. 3). Post-ERCP pancreatitis
developed in 11 (24% ) individuals
with acinar filling. Pancreatitis oc-
curred in seven of 95 patients (7% )
without acinarization. Of 19 patients
who exhibited both acinarization and
urographic visualization following
ERCP, ten (53% ) had pancreatitis.
Acinarization without urographic vi-
sualization was seen in 26 patients.
Only one of these 26 (4% ) had post-
procedure pancreatitis. Table 1 sum-
marizes the data on urographic visu-
alization, acinarization, and post-
procedure pancreatitis.
The degree of pancreatic ductal fill-
ing ranged from partial opacification,
to filling of the main duct, to opacifi-
cation of the main duct and side
branches. Patients with urographic
visualization tended to have a greater
degree of ductal filling than did those
without. However, the exact volume
of contrast material injected into the
pancreatic duct could not be deter-
mined because of variable leakage of
contrast material into the duodenum
and filling of the bile duct. In patients
in whom the total volume of contrast
material used during ERCP was
known, there was no correlation be-
tween the total amount of contrast
material used and postprocedure pan-
creatitis. Thus, in 25 patients without
postprocedure pancreatitis, the mean
volume of contrast material was 54
ml; in three patients with postproce-
dure pancreatitis, the mean amount
was 53 ml.
DISCUSSION
The reported prevalence of urogra-
phic visualization during ERCP has
varied from 0% (0-25 patients) (10) to
37% (10/27 patients) (9). The largest
published series was that of Sahel and
Sarles (8), in which pyelograms were
seen in 27 of 500 (5% ) ERCPs. The as-
sociation of postprocedure pancreati-
tis with a pyelogram has previously
received scant attention, having been
reported in zero of ten (9), one of 27
(8), and five of seven patients (1). All
of these reported cases of pancreatitis
and urographic visualization also ex-
hibited acinarization.
Acinarization during ERCP has
been well described, with occur-
rences during ERP ranging from 7%
(4) to 34% (3). In one large series, aci-
narization occurred in 27% (156/569) of
ERPs (5). The association of acinariza-
tion with higher than expected levels
of serum amylase and postprocedure
pancreatitis has been noted by a num-
ber of investigators (1, 6, 9-11). The
Figure 1. Urographic visualization dur-
ing ERCP (arrows). Contrast material in the
intrahepatic bile ducts demonstrates changes
of primary sclerosing cholangitis. The left
kidney is displaced inferiorly by an enlarged
spleen.
Figure 2. Pyelogram (arrow) during
ERCP seen only on a lateral overhead radio-
graph.
prevalence of post-ERP pancreatitis
has ranged from 13% (4) to 26% (3).
Our own experience suggests a
strong association between urogra-
phic visualization and post-ERCP
pancreatitis. Urographic visualiza-
tion, irrespective of the presence or
absence of acinarization, indicated a
45% risk of pancreatitis; if acinariza-
tion was present, the risk was slightly
but not significantly higher (53% ). W e
noted, as have others, that in some
patients, acinarization is associated
with post-ERCP pancreatitis. Howev-
er, in our series, all but one of these
individuals also exhibited renal ex-
cretion of contrast material. The prey-
Volum e 157 Num ber 3 Radiology. 597
Figure 3. Diffuse pancreatic acinariza-
tion (parenchymal opacification) during
ERCP.
Figure 4. Pyelogram during ERCP was
mistaken at fluoroscopy and on initial spot
films for an obstructed common bile duct.
alence of pancreatitis among our pa-
tients who demonstrated only
acinarization without urographic vi-
sualization was much lower (4% ) (Ta-
ble 1).
A history of pancreatic disease has
been reported to be a risk factor for
the development of postprocedure
pancreatitis (12). We also found this to
be true, although post-ERCP pancre-
atitis occurred both in patients with
(22%) and without (10%) a prior histo-
ry of disease.
Various possible pathways have
been suggested to explain the system-
ic absorption of contrast media dur-
ing ERCP. Bognel et al. (13) injected
dye into the pancreatic ducts of dogs
and were unable to recover it in
lymph obtained from the thoracic
duct; absorption via lymphatics thus
appeared unlikely. Bockman (14) in-
jected India ink and ferritin into the
common bile duct of mice; the tracer
was subsequently found in the liver
sinusoids. It was concluded that high-
pressure biliary injection of contrast
material may disrupt hepatic cells and
allow escape into the sinusoids
through the space of Disse. However,
Sable et al. (15) showed that serum
diatrizoate levels after injection of
contrast material into the biliary tree
were the same as levels obtained after
only simple administration into the
duodenum. W hen the pancreatic duct
alone was cannulated, a marked in-
crease in serum diatrizoate levels was
obtained. They concluded that ab-
sorption of contrast media occurs
mainly via the pancreatic duct.
A pancreatic ducto-interstitial-ve-
nous pathway has been postulated
(16, 17) and documented by W aldron
et al. (18). They injected Hypaque (so-
dium diatrizoate) and thorotrast into
the pancreatic ducts of dogs. W ith
electron microscopy, particles were
demonstrated in the pancreatic duct,
interstitial spaces between the pan-
creatic cells, perivascular spaces, and
in the proximal renal tubules. Rare fo-
cal injury to pancreatic cells was also
noted.
Sable et al. (15) demonstrated that
significant amounts of contrast mate-
rial appeared in the urine of patients
undergoing ERCP with pancreatic
duct cannulation, even when a pyelo-
gram was not seen. They concluded
that absorption was probably via the
ducto-interstitial-venous pathway.
Urographic visualization would occur
if a sufficient volume of contrast ma-
terial was absorbed.
Acinarization is thought to result
from high intraductal pressure and
can be minimized by monitoring in-
jection pressures (19). Injected con-
trast material that reaches the acinus
has been demonstrated to irritate pan-
creatic tissue (7). In our study, howev-
er, acinarization without urographic
visualization was associated with a
relatively low occurrence (4%) of
postprocedure pancreatitis. Appar-
ently, in such patients, the injection
pressure is sufficiently high to cause
some acinarization but not enough
contrast material contacts the acini to
produce clinical pancreatitis. Howev-
er, if volumes and pressures were
high enough to obtain both urogra-
phic visualization and acinarization,
we postulate that greater contact with
contrast material would lead to acinar
damage and an increased occurrence
of pancreatitis. A pyelogram seen
without subsequent pancreatitis pre-
sumably may result from a relatively
high volume of contrast material be-
ing absorbed via the ducto-intersti-
tial-venous pathway without suffi-
cient injection pressure to cause
significant acinar damage.
W e found that pyelogram s were
frequently not visible on spot films
taken during ERCP. This is probably
because the spot films imaged a small
area and because they were taken rel-
atively early in the examination be-
fore a dense pyelogram could devel-
op. Occasionally, a renal collecting
system was mistaken . for another
structure (Fig. 4). M ost pyelograms
were seen only on postprocedure an-
teroposterior and lateral overhead ra-
diographs and easily could have been
overlooked because of overlying
bowel gas as well as contrast material
in the bowel and biliary tree. Indeed,
in few cases was the presence of renal
contrast material prospectively men-
tioned in the radiographic reports.
This is in keeping with the lack of
mention of urographic visualization
in many published studies of ERCP.
The ability to identify patients at
high risk for post-ERCP pancreatitis
has potential practical clinical value.
Awareness of the significant possibil-
ity of this complication can allow the
clinician to institute close observation
and early therapeutic measures in se-
lected cases. Currently, a minority of
ERCPs are performed on outpatients,
in part because of the fear of postpro-
cedure pancreatitis occurring when
the patient is away from the hospital.
Economic incentives now encourage
the performance of outpatient radio-
logic procedures. Increased confi-
dence in predicting the risk of post-
ERCP pancreatitis can potentially
help distinguish individuals able to
remain outpatients from those requir-
ing extended outpatient or overnight
observation.
598 #{149} Radiology
Decem ber 1985
The prevalence of post-ERCP pan-
creatitis at our institution is disturb-
ing when compared with the mostly
lower incidences reported elsewhere
(1, 2, 8). Technical factors must be pre-
sumed to be at least partially responsi-
ble. This has prompted our radiolo-
gists and endoscopists to reemphasize
precautions against overinjection of
the pancreas. Other variables also
probably influenced our incidence of
post-ERCP pancreatitis. Thus, we
studied only patients who had under-
gone ERCP and exhibited pancreatic
duct filling. Patients with unsuccess-
ful cannulation or only biliary filling,
in whom pancreatitis would be un-
likely to develop, were excluded.
Also, many of the patients examined
at our institution were undergoing
investigation for serious hepatobil-
iary problems; opacification of the
biliary tree during ERCP was of prime
importance in these individuals. Not
infrequently, multiple attempts at
bile duct cannulation were made,
with incidental filling of the pancre-
atic duct. This resulting constant ma-
nipulation has been suggested as a
likely mechanism leading to postpro-
cedure pancreatitis (4). Finally, the
exact definition of significant post-
ERCP pancreatitis may vary from in-
stitution to institution, giving rise to
differences in the reported preva-
lence of this complication. I
Acknowledgment: W e wish to thank Donna
Scahill for valued assistance in manuscript
preparation.
Send correspondence and reprint requests to:
M yer H. Roszler, M .D., Department of Radiolo-
gy, Presbyterian-University Hospital, DeSoto at
OHara Streets, Pittsburgh, Pennsylvania 15213.
References
1. Bilbao M K, Dotter CT, Lee TG, Katon RM .
Complications of endoscopic retrograde
cholangiopancreatography: a study of
10,000 cases. Gastroenterology 1976;
70:314-320.
2. Seifert E. Endoscopic retrograde cholan-
giopancreatography. Am J Gastroenterol
1977; 68:542-549.
3. Classen M , Koch H, Ruskin H, Pesch HJ,
Demling L. Pancreatitis after endoscopic
retrograde pancreatography (abstr.). Gut
1973; 14:431.
4. Hamilton I, Lintott DJ, Rotwell J, Axon
ATR. Acute pancreatitis following endo-
scopic retrograde cholangiopancreatog-
raphy. Clin Radio! 1983; 34:543-546.
5. Ruppin H, Amon R, EttI W , Classen M ,
Demling L. Acute pancreatitis after en-
doscopic/radiological pancreaticography.
Endoscopy 1974; 6:94-98.
6. Varley PF, Rohrmann CA, Silvis SE,
Vennes JA. The normal endoscopic pan-
creatogram. Radiology 1976; 118:295-300.
7. Kivisaari L. Contrast absorption and pan-
creatic inflammation following experi-
mental ERCP. Invest Radiol 1979;
14:493-497.
8. Sahel J, Sarles H. Endoscopic pancreatog-
raphy and urograms. Gastroenterology
1976; 71:1109.
9. Goldberg HI, Bilbao M K, Stewart ET, Rohr-
mann CA, M oss AA. Endoscopic retro-
grade cholangiopancreatography: radio-
graphic technique. Am J Dig Dis 1976;
21:270-278.
10. Kasugai T, Kuno N, Kobayashi 5, Hattori
K. Endoscopic pancreatocholangio-
graphy. I. The normal endoscopic pancrea-
tocholangiogram. Gastroenterology 1972;
63:217-226.
11. Skude G, W ehlin L, M aruyama T, Ariyama
J. Hyperamylasaemia after duodenoscopy
and retrograde cholangiopancreatog-
raphy. Gut 1976; 17:127-132.
12. Zimmon DS, Falkenstein DB, Riccobono C,
Aaron B. Complications of endoscopic
retrograde cholangiopancreatography:
analysis of 300 consecutive cases. Castro-
enterology 1975; 69:303-309.
13. Bognel JC, Levy E, M onnier A. Experi-
mental pancreatography (abstr.). 2d Inter-
national Symposium of Digestive Endosco-
py, Brussels, 1975. Cited in: Sable RA,
Rosenthal W S, Siege J, Ho R, Jankowski
RH. Absorption of contrast medium dur-
ing ERCP. Dig Dis Sci 1983; 28:801-806.
14. Bockman DE. Route of flow and micro-
pathology resulting from retrograde intra-
biliary injection of India ink and ferritin in
experimental animals: a combined light
and electron microscopic study. Gastroen-
terology 1974; 67:324-332.
15. Sable RA, Rosenthal W S, Siegel J, Ho R,
Jankowski RH. Absorption of contrast
medium during ERCP. Dig Dis Sci 1983;
28:801-806.
16. Stein AA, Powers SR. Browne HH. Ex-
perimental hemorrhagic pancreatitis: new
concepts of pathogenesis. Ann Surg 1956;
143:508-516.
17. Edlund Y, Ekholm R. M icrostructure and
ultrastructure of human pancreas. Acta
ChirScand 1957; 113:469-471.
18. W aldron RL, Luse SA, W ollowick HE, Sea-
man W B. Demonstration of a retrograde
pancreatic pathway: correlation of roent-
genographic and electron microscopic
studies. AJR 1971; 111:695-699.
19. Kasugai T, Kuno N, Kizu M . M anometric
endoscopic retrograde pancreatocholan-
giotography: technique, significance, and
evaluation. Am J Dig Dis 1974; 19:485-502.