J. Savader, MD #{149} Anthony C. Venbrux, M. Gittelsohn, PhD #{149} Floyd A. Osterman,

Scott Alan

Pancreatic Biliary

Index

terms: 76.1229

Bile

dures,

interventional

Pancreas.

stenosis

Pancreatitis,

1991;

proce-

palliative

the

proce-

Radiology A.C.V., statistics

and

H. Morgan

Radiological

K.V.R., F.A.O.) (A.M.G.), The

Sciences

revision

10. Address c RSNA,

received

(S.J.S.,

and Department Johns Hopkins

tal, 600 N Wolfe St. Baltimore, ceived June 15, 1990; revision 15;

Department

and

MD 21205. Rerequested August

accepted

reprint requests 1991

of BioHospi-

September

to S.J.S.

of

and

of percutaneous (PBD) in the

presurgical

PATIENTS Fifty

MD

#{149}

study

patients

Hospital 1990 were men (range,

at The

were

average years)

age was

34-80 Baseline

years) composed serum amylase

sured

in each

patient

years

(range,

48 hours

pri-

or to PBD. the drainage the ampulla,

After completion of PBD, with catheter positioned across serum amylase levels were

measured

for

amylase

7 consecutive levels

were

days. measured

All

struction

In

was

The

the

categorized

44

of

patients,

ob-

according

to

three levels: (a) the hepatic duct confluence to the common hepatic duct, (b) the proximal part of the common bile duct to the (c)

middle the

of the

distal

common

common

bile

bile

duct,

duct

and

to the

am-

pulla of Vater. Fifty-three PBD procedures were performed in 50 patients. The final diagnoses, proved with surgery or biopsy, were as follows: cholangiocarcinoma, 14 patients (28%); pancreatic carcinoma, eight patients (16%); periportal metastasis, three patients noma, two patients noma,

two

carcinoma, pancreatitis, the common benign four

(6%); ampullary (4%); duodenal

patients

(4%);

stricture

gallbladder

in (8%);

the

common

sclerosing

23 patients

(46%);

bile

duct,

cholangitis,

three patients (6%); choledochal patients (4%); and intrahepatic ma, one patient (2%). The indications for biliary were as follows: preoperative sion,

carcicarci-

one patient (2%); chronic five patients (10%); stones in bile duct, five patients (10%);

patients

sclerosing (18%).

the study group. levels were meawithin

information.

obstruction.

(iatrogenic

age was 60.7 and 20 women

60.2

biliary

benign

Johns

from July 1989 to included in our study.

whose 35-82

average

who

PBD

supplemental

cholangiograms in 44 patients reviewed to evaluate the level

primary

cyst, two hamartodrainage decomprespalliative

decompression, 13 patients (26%); percutaneous access and bile diversion for stones in the common bile duct, five patients (10%); and placement of stents for

METHODS

to undergo

for

initial were

decompres-

AND

consecutive

scheduled

whose

178:343-346

Russell

V. Robbins,

sion of malignant biliary obstruction has been well documented (1-3). More recently, PBD has proved valuable in the relief and treatment of benign biliary obstruction, including strictures, calculi, and sclerosing cholangitis (4-6). Because PBD is an invasive procedure, potential complications can occur, including cholangitis (13.0%-47.0% of patients), fever (1 1 .2%), hemobilia (7.0%-9.6%), sepsis (3.9%-8.0%), bile hypersecretion with electrolyte imbalance and hypotension (5.0%), and death (0.6%-5.6%) (7-11). Previous reports on complications of PBD rarely mention pancreatitis. In a review of 718 cases, one study noted pancreatitis as a complication of PBD in one patient (0.1%) only (6). Our concern oven postprocedural hypenamylasemia, frequently noted in follow-up laboratory tests, prompted us to conduct this pnospective study to evaluate the pancreatic response to PBD.

rum From

effectiveness biliany drainage

HE

Thirty years

77.291

#{149}

T

Hopkins March

or obstruc-

interventional

#{149}

77.1229

Radiology

ducts,

Bile ducts,

#{149}

tion, 76.289

#{149} Kenneth

MD

Response to Percutaneous Drainage: A Prospective Study’

To evaluate the effects of percutaneous biliary drainage (PBD) on the pancreas, serum amylase levels were measured for 7 consecutive days after PBD and compared with baseline values in 50 patients who underwent a total of 53 PBD procedures. Of the 45 patients with normal baseline serum amylase levels, 12 patients (24%) developed postprocedural hyperamylasemia without clinical symptoms and five patients (10%) developed postprocedural hyperamylasemia with clinical signs of pancreatitis. Five patients who presented with elevated baseline serum amylase levels demonstrated decreases into the normal range aften placement of stents without mitiation of bowel rest or liquid diet. The level of biliary obstruction proved insignificant, as did the nature of the obstructing disease, in determining which patients would experience hyperamylasemia or pancreatitis after PBD. It is concluded that the frequency of pancreatic insult from PBD may be more common than previously reported and that patient susceptibility is not dependent on the level of biliary obstruction or the nature of the disease.

dure,

MD

stricture

of the

injury,

common

chronic

cholangitis),

bile

duct

pancreatitis, nine

or

patients

PBD was performed with an 8.3-F biliary catheter (Ring; Cook, Bloomington, md) (47 procedures) and the 8-F straight drainage

catheter

(Kaufman;

Cook)

(five

patients). An 18-F soft Silastic biliary stent (Mentor; Goleta, Calif) was used one patient (the tube replaced a 16-F

in

tube).

seby

of the p-nitro-phenol reaction with spectrophotometnic analysis. Patients were also followed up during this period for

clinical signs of pancreatitis, including abdominal or back pain, epigastric tenderness, nausea, vomiting, and/or fever. The medical records of these patients were reviewed after completion of the

use

RESULTS The results of this study are summarized in Table 1. On the basis of the analysis of serum amylase levels

Abbreviation:

PBD

=

percutaneous

biliary

drainage.

343

obtained during the study, all patients were classified into one of four groups: patients with normal pneand postpnocedunal serum amylase levels (group 1), patients with normal pre- and abnormal postpnocedural serum amylase levels who had no clinical symptoms (group 2), patients with normal pre- and abnonmal postprocedunal serum amylase levels who had clinical symptoms (group 3), and patients with abnormal preprocedunal serum amylase levels and normal postpnocedunal serum amylase levels who had no clinical

symptoms (group Group 1, the largest

ed of 28 patients tients

in this

for relief

group

consist-

Eighteen

underwent

of malignant

level

lase

follow-up

during

the

procedure

biliary

of serum

ob-

In

has done

well since surgery. Group 2, the second largest subset, consisted of 12 patients (24%). Six patients in this group underwent PBD for relief of malignant biliary obstruction; six, for benign biliary obstruction. This group presented with

a mean

baseline

serum

amylase

level

of 142 UIL (Fig 1). The patients developed an abnormal elevated mean serum amylase level of 420 U/L the day after PBD. All patients had mdividual abnormal serum amylase 1evels 3 days after PBD. Despite laboratory confirmation of elevated amylase levels (range, 224-1,610 U/L), no patient in group 2 developed clinical symptoms of pancreatitis. The two patients in this group with a diagnosis of chronic pancreatitis at admission both underwent Whipple procedunes; one patient has done well, and the other patient continues to experi-

344

Radiology

#{149}

No.

Group

1 (without

of

Patients clinical

after PBD (U/L)

Mean No. of Days for Amylase Level to Return to Normal Range

90

(2)

NA

Mean

Mean Amylase Level before

Peak Level

Amylase

PBD (U/L)

28

82

12

142

420

(1)

1.4*

5

160

1,656

(2)

2.8t

5

373

422 (1)

2.4

symptoms; normal before and after PBD) Group

2 (without

clinical

symptoms; before

after Group

normal

PBD,

abnormal

PBD)

3 (with

clinical

symptoms; normal before PBD, abnormal Group

4 (without clinical symptoms; abnormal before PBD, normal after

PBD)

Note-Range applicable.

of normal

In four patients, t In one patient, *

level

amylase

the amylase

40-220

of amylase,

levels

did

level

did

not

return

not return

U/L.

Numbers

to normal

in parentheses

within

to normal

are

days.

NA

not

-

7 days.

within

7 days.

amy-

period.

and

in 50 Patients

Level

PBD

all 28 patients, the peak mean serum amylase level of 90 U/L occurred the day after PBD. None of the patients experienced clinical symptoms of pancreatitis related to catheter placement. One patient with chronic pancreatitis presented with moderate chronic abdominal pain, which was unchanged in nature or intensity aften catheter placement. A stent was placed across the common bile duct stricture; 2 months later it was removed. She continues to do well, with only intermittent episodes of mild abdominal pain. The second patient in this group, who had chronic pancreatitis at admission, underwent

a Whipple

of Data Serum Amylase

pa-

struction; 10 patients, for relief of benign biliary obstruction. Group 1 presented with a mean baseline Serum amylase of 82 U/L (Figure). No individual in this group demonstrat-

ed an abnormal

1

Summary

after PBD)

4). subset,

(56%).

Table

ence problems with malnutrition. Group 3 consisted of five patients (10%). Four patients in this group underwent PBD for relief of biliary obstruction caused by malignant disease; one, for biliary obstruction caused by benign disease. This group presented with a mean baseline serum amylase level of 160 U/L (Figure). This group developed an abnonmally elevated mean serum amylase level of 1,581 U/L and 1,656 U/L on postpnocedural days one and two, respectively. One day after PBD, patients developed clinical symptoms corresponding to their rising serum amylase level. Their symptoms were abdominal pain (three patients), feyen (three patients), nausea and vomiting (three patients), and abdominal tenderness (one patient). Patients in group 3 underwent conservative treatment with bowel nest, intravenous hydration, and control of pain. No patient developed signs or symptoms necessitating radiologic evaluation. Complete resolution of clinical symptoms were seen at 2 days (two patients), 3 days (one patient), 4 days (one patient), and 14 days (one patient, in whom mild persistent abdominal pain lasted 2 weeks). Group 4 also consisted of five patients (10%). Four patients in this group underwent PBD for relief of malignant biliary obstruction; one, for benign biliary obstruction. The mean baseline serum amylase level in this group was 373 U/L (Figure). None of the patients had symptoms of pancreatitis

PBD.

Two

of the five

clinical prior

patients

to

experi-

2000

1

900 w U)

4

800

1 >-

n 600

a:

5O0

:4\

:

400

a.,

Ui

n

__

-\

i

._..a.

/!

#{149}

200 #{149}

0

PRE

I

2

of mean

,..v

MA)

: 3

DAYS

Comparison

%

4

POST

5

6

7

PBD

#{149}

serum

amylase

level

to

(in days) after PBD. group 1, A group 2, 0 = group 3, group 4, MAX.NML = maximum normal level of serum amylase, PRE = before PBD. time

enced serum PBD.

a transient amylase In all five

further level the patients,

elevation in day after serum amy-

lase levels returned to the normal range within 5 days after PBD; in four patients they returned to the normal range 2 days after PBD. Of the two patients who experienced a further transient rise in the level of serum amylase, neither developed clinical symptoms. The single patient in this group who had been admitted with a diagnosis of chronic pancneatitis underwent a Whipple procedure and has since done relatively well. Cholangiognams were available for review in 44 patients. Forty-one dem-

February

1991

onstrated biliary obstruction (Table 2). Sixteen patients had obstruction at the level of the hepatic duct confluence on common hepatic duct, six at the level of the proximal to middle

common of the

ampulla

bile distal

duct,

and

common

of Vater.

19 at the bile

level to the

Hyperamylasemia

and/or pancreatitis was of patients with hepatic ence or common hepatic and 21% of patients with mon bile duct or ampullary No patient with lesions mal to middle common veloped hyperamylasemia atitis (P = .027). In our comparison of tients with normal serum

levels

duct

throughout

the

seen in 6.25% duct confluduct lesions distal cornlesions. of the pnoxibile duct deor pancregroup 1 (paamylase

study)

with

groups 2 and 3 combined (patients with abnormal serum amylase levels after PBD), nearly equal percentages of patients with malignant disease and benign disease were noted.

DISCUSSION Previous studies of PBD have noted prevalences of up to 25% for major complications (8). Cholangitis is often cited as the most common cornplication, but catheter-related problems, including migration, obstruction, dislocation, and pericatheter leakage, may occur cumulatively in 65%-100% of patients (3,6-8). Despite the relatively high frequency of hepatobiliary and catheter-related complications noted in the literature, reports of pancreatic complications are few (6). Many studies do not even mention pancreatitis as a complication (3,5,7,8). In light of the high complication rate associated with PBD and the difficulty that can be encountered in obtaining duodenal drainage in patients with significant obstruction, it seems curious that the pancreas is not traumatized more frequently, particularly in patients with

Volume

178

#{149} Number

2

disease involving the pancreaticobiliary junction. The cause of pancreatitis, regardless of the inciting event, is intrapancreatic activation of the enzymes trypsin, elastase, and phospholipase A. During PBD, at least three of the described mechanisms for intrapancreatic activation of enzymes may occur: bile reflux, hypersecretion and obstruction, and duodenal reflux (12). Bile reflux has long been considered to play an important role in the activation of pancreatic enzymes. Studies have shown that reflux of bile into the pancreatic duct at abovenormal pressures can produce significant inflammatory reaction. In addition, entenic bacteria can convert conjugated bile salts into deconjugated bile salts, a product highly toxic to pancreatic ductal epithelium (12). Patients with biliary obstruction by definition have abnormal biliary pressure. Infected bile is encountered in up to 68% of these patients (8). After PBD, all patients experience colonization of the biliary tract, and, as noted previously, up to 47% can experience subsequent cholangitis (7,8). These factors support bile reflux as an important pathway leading to pancreatitis after PBD. With regard to hypersecretion and obstruction, patients with pancreatic and distal lesions of the common bile duct commonly have pancreatic ductal narrowing. Placement of stents in the area of ductal narrowing may involve considerable trauma at or near the pancreaticobiliary junction. Edema may result, further contributing to obstruction of the pancreatic duct orifice. In addition, it also seems possible that placement of stents in patients with disease involving the pancreaticobiliary junction can at times either partially or completely occlude the pancreatic duct, because many stents used in the biliary tract have a significant amount of nonfen-

estrated surface area. Duodenal reflux may occur in patients with hepatobiliary disease both before and after stent placement. Neoplastic involvement or inflammation of the ampulla or distal common bile duct may result in ampullary dysfunction with subsequent duodenal reflux (12). Placement of a stent across the ampulla prevents sphincter of Oddi function because of the mechanics of the position of the stent. Reflux of duodenal contents into the biliary system may then occur either around the stent or via the lumen of the stent, resulting in intrapancreatic activation of enzymes with resultant pancreatitis. This study was designed to evaluate the pancreatic response to PBD through analysis of pre- and postprocedural serum amylase levels. Serum amylase levels were assessed for 7 days after PBD. At completion of the analysis, patients could be separated into four distinct groups. Of the 45 patients who entered the study with a normal baseline serum amylase 1evel, 28 patients (group 1) maintained normal serum amylase levels throughout the study. Eighteen patients in this group were treated for malignant biliary obstruction; 10, for benign biliary obstruction. Catheter size was 8.3 F in 27 procedures and 8.0 F in four procedures. Twelve patients (group 2) developed hyperamylasemia after PBD but remained asymptomatic. Serum amylase levels in this group quickly returned to the normal range (mean, 1.4 days), with a peak serum amylase level of 1,610 U/L noted in one patient (mean, 420 U/L). This suggests that some patients (12 patients [24%] in our study) can be expected to suffer minor pancreatic trauma as a direct result of the procedure but to a

degree

not

severe

enough

to cause

clinical symptoms. The transient hyperamylasemia experienced by this group is similar to that experienced by many patients after endoscopic retrograde cholangiopancreatography. This transient phenomenon has been shown to be harmless. These patients showed rapid development of abnormal peak serum amylase 1evels after PBD (mean, 1.4 days); in nine of 12 patients the serum amylase level returned to the normal range by the next morning. Five patients (group 3) developed both hyperamylasemia (mean peak serum amylase level, 1,656 U/L) and corresponding clinical symptoms of pancreatitis. Peak serum amylase levels in this group ranged from 1,452

Radiology

#{149} 345

U/L lase

to 4,010 U/L. levels indicate

level

of pancreatic

These serum amya more significant

trauma

overall

than that in group 2 patients, but note the overlap in peak serum amylase levels in the two groups. Thus it is important to emphasize that while a twofold or greater increase in serum amylase level is considered good laboratory evidence for pancreatitis, clinical correlation is necessary. It is also important to note that in all 12 patients in group 2 the serum amylase level returned to normal without medical treatment, yet all five patients in group 3 required conserva-

tive

treatment

nous hydration, antiemetics) ical picture.

(bowel on

rest,

intrave-

pain medication, the basis of their

and din-

Five patients (group 4) had abnormal baseline serum amylase levels (mean, 373 U/L) without symptoms and, after a minor mean postprocedunal rise in levels, developed normal serum amylase levels in a mean time of 2.4 days. None of these patients received medical therapy based on their baseline amylase level, suggesting that, in some patients, PBD may also relieve a component of pancreatic obstruction. Patients in groups 1-4 were subdivided by the level of their lesion (Table 2), and the test was applied to groups 1-3 to find out if the location of the obstructing lesion would help determine which patients were at

x2

346

Radiology

#{149}

risk for postpnocedural hyperamylasemia on clinical pancreatitis. A P value of .027 indicates pancreatic response to PBD is not dependent on lesion location, while nearly equal ratios of patients with benign and malignant disease (group 1 versus group 2 plus group 3) suggest that benignity or malignancy does not play a significant role either. In conclusion, the sensitivity of the pancreas to trauma as a result of PBD appears greater than previously noted in the literature (6). Ten percent of patients in our study developed clinical pancreatitis. Ten percent developed serum hyperamylasemia. Such sensitivity does not appear to be dependent on benignity versus malignancy of the obstructing disease, nor is it dependent on the location of the obstructing lesion. This study also suggests that as patients are followed up for potential complications such as cholangitis after PBD, one should be careful to watch for clinical and laboratory evidence of pancreatitis. U Acknowledgments:

sincere

gratitude

Offenbacker tion

of our

for

The

authors

to Valerie

Hux

their

expertise

express

and in the

3.

for

combined

trointest

Radiol

Cameron

JL,

1978;

stents.

1985;

RW,

biliary

Schild

PR,

ment

of complex

Surg

1983; 197:584-593.

biliary

problems.

Ann

et

ACN,

Ho

Clark

RA,

Carrasco lignant

C.

pro-

Oleaga

Grune O’Brien

problems 138:17-23.

in

SE,

200

of percu-

benign

1987;

JT

techni-

Complications

Mitchell

IT,

drainage:

drainage: AJR

biliary

1984;

vs malig-

148:1207-1209. Colley

DP,

Alexan-

catheter biliary 1981; 137:503-509.

de-

Zernoza J, Bechtel WJ. Maobstruction: complications biliary

drainage.

Radiolo-

152:343-346. JA,

Ring

EJ. In:

SL,

of disease.

Interventional

Felson B, ed. Vol 16. New

& Stratton, MJ, Gottlieb

Robbins

1979;

311

Radiol E, Ferrucci

related 1982;

of percutaneous

basis

Re-

with

biliary

biliary diseases.

In:

M.

transhepatic

vanSonnenberg

der E. Percutaneous compression. AJR

11.

Thelen

Intervent

Percutaneous

nant

gy

et al.

11:65-71.

taneous

10.

SL,

experience

cal and catheter procedures. AJR

9.

H,

Cardiovasc

Mueller

Yee

S. Kaufman

percutaneous

drainage:

1988;

8.

trans-

94:324-330.

156:625-629. article:

Jr.

silastic

1983;

management of benign biliary strictures. Radiology

cedures.

12. Sitzman JV, manage-

with

Kadir

HF,

cholangitis:

Surgery

DJ,

Gunther view

7.

Herlong

Sclerosing

Nonoperative postoperative 6.

3:23-31. BW,

reconstruction

Gallacher

Percuta-

and palliative jaundice. Gas-

Gayler

WC.

hepatic

I.

of bile

internal-external

in pre-operative of obstructive

biliary

WC,

Percutaneous-

intubation

drainage treatment

Mackdrey

5.

Maddrey

JL.

transhepatic

ducts

4.

KH,

stents in the managebiliary obstruction.

ary radiology. in roentgenology.

Cameron JL, transhepatic

Barth

Dig Dis Sci 1979; 24:849-857. Hoevels J, Lunderquist A, Ihse neous

their

prepara-

DP,

SL, Cameron

ly placed biliary ment of malignant

manuscript.

Zuidema GD, al. Percutaneous

Harrington

Kaufman

Debbie

References 1.

2.

biliSeminars York:

1981; 116-134. LS. The pancreas.

Cotran Philadelphia:

RS,

eds.

Pathologic Saunders,

1092-1114.

February

1991