Gallstone Pancreatitis: When to do ERCP?

[Guest guest]

Gallstone Pancreatitis: When to do ERCP?

by Ira son, MD

The concept of utilizing ERCP to treat patients with acute gallstone

pancreatitis has intrigued biliary endoscopists for two decades. Consensus was

reached long ago that the pathogenesis of the disease is related to impaction of

a gallstone in the distal bile duct, and that pancreatic ductal hypertension

related to obstruction unleashes the chain of events that result in acute

pancreatitis. Logic seemed to dictate that extraction of the culpable stone

should abort the episode. The challenges that had to be addressed included:

1.. correct identification of patients with biliary pancreatitis, as opposed

to other causes;

2.. the fact that many patients clear the culpable stone by the time they

present clinically;

3.. the uncertainty as to whether, once the cascade of events involved in

pancreatitis was initiated, extraction of a causative stone would alter the

evolution of the process.

An underlying concern at an early point was that, given the capacity for ERCP to

cause pancreatitis as a complication of the procedure, it seemed

counterintuitive to perform ERCP during an acute episode. However, a number of

early series demonstrated no apparent aggravating effect of ERCP on the course

of the pancreatitis. In contrast, the impression of early investigators was of

sometimes dramatic improvement beyond what would have been expected with

conservative medical management alone. These observations clearly warranted the

performance of randomized, controlled trials.

The first of these, and still most oft- quoted, appeared to vindicate the

growing suspicion of those who argued that early ERCP would favorably ameliorate

the course of some patients with acute biliary pancreatitis. Neoptolemos, Carr-

Locke and coinvestigators studied 121 patients with biliary pancreatitis out of

a total of 223 consecutive patients presenting with acute pancreatitis to their

hospital. Patients were randomized to urgent ERCP or conventional management

within 72 hours of admission. Sphincterotomy was performed only if stones were

found in the bile duct, and was followed immediately by stone extraction.

For purposes of data analysis, patients were stratified according to modified

Glasgow criteria into those with mild versus severe predicted attacks. This

stratification proved to be an important feature of the study, because in

patients with mild attacks of pancreatitis urgent ERCP did not reduce local or

systemic complications, nor did it decrease hospital stay.

In contrast, patients with severe attacks fared substantially better with urgent

ERCP with regard to local complications(e.g. pseudocysts, ascites, duodenal

obstruction) or systemic complications (e.g. cardiopulmonary failure, renal

failure, death), and duration of hospital stay:

Complications Hospital stay Death

Conservative 61% 17 days 5/28

Urgent ERCP 24% 9.5 days 1/25

Note that these data pertain only to patients with predicted severe

pancreatitis. When all patients were analyzed together, the mortality rates were

2% and 8% for the patients treated by urgent ERCP and for those treated

conservatively, respectively; this difference did not attain statistical

significance. There was no evidence of exacerbation of pancreatitis in this

study.

Publication of this paper ushered in a new era in this field, with many experts

now advocating, at long last on the basis of solid scientific data, urgent ERCP

for patients with severe disease on admission or failure to improve after a

relatively brief observation period (e.g. 12-24 hours). A second paper,

published from Hong Kong 5 years later, further supported the role of urgent

ERCP, but the benefit that emerged from early ERCP was a reduction in biliary

sepsis rather than in complications from the pancreatitis itself.

In the Hong Kong study (Fan et al), 195 patients with acute pancreatitis were

randomized within 24 hours of admission (not 72 hours, as in the British study)

to urgent ERCP versus conservative therapy. There was no apparent attempt to

establish whether other causes of pancreatitis might have been present, a

feature of the study that may have been justifiable in light of the high

prevalence of primary bile duct stones in Asia (at the end, biliary stones were

diagnosed in 127 patients). Several clinical developments permitted patients

randomized initially to medical therapy to be crossed over to ERCP at any time:

rising fever, leukocytosis, tachycardia, increasing serum bilirubin or jaundice,

and shock not responding rapidly to IV fluids. Of 98 patients randomized to

conservative treatment, 9 of 58 with predicted mild pancreatitis (by Ranson's

criteria) and 18 of 40 patients with predicted severe pancreatitis were crossed

over to ERCP, usually because of cholangitis and/or sepsis.

In the group as a whole, complications occurred in 18% of the patients receving

emergency ERCP and 29% of those initially receiving conservative treatment (p =

0.07). Mortality rates in the two groups were 5% and 9%, respectively. If only

patients who proved to have stones anywhere in the biliary tract were analyzed

separately, morbidity rates diverged sufficiently to attain statistical

significance: 16% with emergency ERCP versus 33% without it (p = 0.03), with

mortality rates of 2% and 8% (p=0.09). When the analysis was stratified to

compare only patients with severe pancreatitis and biliary stones (whether in

the gallbladder, bile duct, or both), the difference in incidence of

complications between the ERCP and conservatively treated groups was

accentuated.

However, the major benefit conferred by ERCP in this study was in the reduction

of biliary sepsis in the ERCP-treated group. The differences were as follows:

Biliary Sepsis

All patients

ERCP 0/97

Conservative 12/98 p=0.001

Mild pancreatitis

ERCP 0/56

Conservative 4/58 p=0.14

Severe pancreatitis

ERCP 0/41

Conservative 8/40 p=0.008

All patients with biliary sepsis had persistent ampullary or common bile duct

stones at ERCP.

From their data, the authors derived the sweeping conclusion that all patients

with acute pancreatitis should undergo ERCP and, if stones are found in the bile

duct, sphincterotomy. Needless to say, this conclusion did not receive

widespread acceptance. Grounds for criticism included the unwarranted nihilism

about the capacity to distinguish biliary pancreatitis from other causes by

history and noninvasive means; the high prevalence of primary common duct stones

in Asia and, perhaps with that, the high prevalence of antecedent bactobilia

that exists in these patients; the high incidence of biliary sepsis in the

patients with mild pancreatitis (4 of 58 seems atypical); and the fact that the

authors apparently did not give antibiotics, which have come into widespread -

though not universal - use in patients with acute gallstone pancreatitis.

Despite differences in the nature of the chief advantage conferred by urgent

ERCP and the scope of patients who appear to benefit, the British and Hong Kong

studies reinforced the existence of a role for ERCP, at least in some patients,

if not all, with acute biliary pancreatitis. A " wrench " was thrown into the

picture with the picture of another major randomized study, this time from

Germany, in 1997 (Folsch et al). Acknowledging the role of urgent ERCP in

patients with obstructive jaundice or cholangitis, these authors challenged the

applicability of urgent ERCP to patients with acute biliary pancreatitis but

lacking overt obstruction or cholangitis. Accordingly, their study design

excluded patients wtih bilirubin levels > 5 mg/dl. Like the British authors, but

unlike those in Hong Kong, Folsch et al made an attempt to identify patients

with gallstones as the etiology on radiologic or laboratory grounds, excluding

others from randomization. They required patients to be available for ERCP

within 72 hours of symptom onset, not admission. If anything, this should have

increased the number of patients with persistent CBD stones and enhanced the

advantage of ERCP, if any existed. As in previous trials, patients initially

selected for conservative treatment could be crossed over to ERCP, in this study

for temperature of over 39 degrees Centigrade, an increase in bilirubin of 3

mg/dl or more within 5 days, or " persistent biliary cramps. "

A total of 238 patients were randomized. Of 126 randomized to urgent ERCP, the

procedure was successful in 121 and sphincterotomy was performed to remove bile

duct stones detected on cholangiography in 58 patients - a relatively high

yield. In the group randomized to conservative treatment, ERCP was performed in

22 patients because of increasing jaundice (8 patients), high fever (8), and

biliary cramps (6). (The specificity of temperature over 39 degrees and " biliary

cramps " for biliary obstruction, rather than pancreatitis, is arguable).

In the urgent ERCP group, 14 patients died within 3 months compared with 7 in

the conservative treatment group, while deaths from acute biliary pancreatitis

occurred in 10 and 4 patients, respectively. These differences, though

disturbing, did not achieve statistical significance. Overall, complications

occurred in 46% and 51% of patients treated by ERCP and medically, respectively.

However, respiratory failure was more frequent in the ERCP group (p=0.03), while

jaundice was more common in the other group. There was no plausible explanation

for the excess incidence of respiratory failure in the ERCP group.

In an editorial accompanying this paper, Baillie accepted the conclusion

that ERCP in patients with acute pancreatitis should be limited to those with

cholangitis or progressive jaundice, and that " patients with severe or worsening

pancreatitis do not benefit from ERCP. " This paper was criticized on the grounds

that its multicenter nature may have resulted in the inclusion of relatively

inexperienced endoscopists (Tarnasky and Cotton), but the authors countered by

citing the identification of the participants on the basis of their reputations

as experts and, equally important, the high success rates of

cholangiography,sphincterotomy and stone extraction in their study. Also, if

anything a multicenter trial might reflect more accurately the level of

expertise available in many centers today.

The only other randomized trial, appearing superficially to counterbalance the

German study, was published in abstract form from Poland in 1995, and again as

an abstract in 1996 (Nowak et al). In this study, 280 consecutive patients

underwent ERCP. Of the 280, 75 had stones impacted in the ampulla and underwent

immediate sphincterotomy (group 1). While the remainder were randomized to ES

(group 2) or conservative treatment (group 3).

When groups 1 and 2 were combined, mortality was 2% compared to 13% in group 3,

with complication rates of 17% versus 36%, respectively. In their second

abstract, representing an expanded population of 307 patients, outcomes were

analyzed by time of procedure with the following results:

Timing of ES Complications Death

Within 24 hours 13% 0%

24-72 hours 34% 6%

4-7 days 22% 48%

The advantages of ERCP extended to patients with mild pancreatitis in this

study.

Unfortunately, the study has not been published in full, making analysis of the

data difficult. In addition, the frequency of stones impacted in the ampulla was

quite high, and no data were presented on the clinical and laboratory features

of these patients. It may be, for example, that most of them had high bilirubin

levels and/or cholangitis and would not be a focus of dispute anywhere about

whether they needed ERCP. In the 1995 abstract, the data analysis lumped these

patients together with those randomized to sphincterotomy despite the absence of

impacted stones. Finally, the death rate of 48% in the 71 patients undergoing

delayed sphincterotomy seems extraordinarily high.

In my experience, the results of the German study have been substantiated in

that the most dramatic improvement after ERCP and sphincterotomy has clearly

been achieved in patients with concomitant cholangitis and/or frank biliary

obstruction. I have been less impressed with the results obtained after ERCP in

patients with normal or minimally elevated bilirubin levels (e.g. less than 2-3

mg/dl) and no evidence of sepsis even when the patient has severe pancreatitis.

The frequency of ampullary or CBD stones in the latter patients has been low in

my experience.

These patients generally appear to be ill as a result of the cascade of events

that have already been set in motion within the pancreas by the time a decision

about performing ERCP must be made.

If the data from randomized controlled trials leaves one ambivalent about

performing ERCP in patients who are not jaundiced or septic, a noninvasive test

to resolve the issue of persistent choledocholithiasis would be most useful.

Several recent publications have addressed the utility of endoscopic ultrasound,

particularly with radial instruments, in patients with acute biliary

pancreatitis. The yield of EUS has rivalled that of ERCP in these studies. From

a logistical viewpoint, the ideal situation if EUS is to be used would involve

an endoscopist with expertise in both EUS and ERCP, or the simultaneous

availability of an endosonographer and biliary endoscopist so that the latter

can proceed with sphincterotomy and stone extraction if a CBD is found on the

EUS.

Even more attractive, because of its noninvasiveness, is MRCP, which has a yield

of 80-90% in the detection of CBD stones. However, several factors limit the

potential utility of MRCP in this setting. First, the technique is not

universally available, and even in institutions where it exists it may not be

obtainable on the short notice that would be required in acutely ill patients

with biliary pancreatitis. Second, MRCP may have a lower yield for small CBD

stones. Finally, the ability of MRCP to detect stones in the very distal CBD or

intra-ampullary segment has not been demonstrated rigorously in clinical trials.

Despite these potential limitations of MRCP, we have, in the last 1-2 years,

used it whenever possible in patients with moderate or severe biliary

pancreatitis who do not have jaundice or cholangitis. We personally review the

films with a radiologist to be confident of the technical quality of the study.

Thus far, the yield has been low and we have seldom subsequently felt compelled

to proceed to ERCP in these patients. This is clearly an area where further

clinical trials would be of interest.

When ERCP is undertaken in these patients, two practical points arise:

1.. The duodenum may be markedly edematous, compounding the difficulty of

cannulating the papilla.

2.. Once cannulation of the bile duct is achieved, should sphincterotomy be

done regardless of whether a stone is seen on the cholangiogram? I believe that

a compelling case can be made for doing so for several reasons. First, a very

distal small stone, or sludge, may be missed on flouroscopy or plain films.

Second, sphincterotomy will protect against recurrent stone impaction at the

papilla, with the potential for exacerbating the present episode or causing a

recurrent episode before the patient has had a cholecystectomy. In older

patients or those with significant comorbid disease, sphincterotomy actually

leaves open the option of deferring cholecystectomy indefinitely. Finally, there

is a theoretical, though unproven, possibility that stone passage may have left

the sphincter of Oddi in a state of spasm or stenosis that is further driving

the ongoing pancreatitis.

In light of these considerations, my approach to the patient with biliary

pancreatitis undergoing ERCP is to attempt initial cannulation with a

sphincterotome, obtain a cholangiogram, and then proceed with biliary

sphincterotomy even if the cholangiogram is negative.

REFERENCES

Baillie J: Treatment of acute biliary pancreatitis (editorial). N Engl J Med

1997;336:286-7.

Folsch U, Nitsche R, Ludtke R et al: Early ERCP and papillotomy compared with

conservative treatment for acute biliary pancreatitis. N Engl J Med

1997;336:237-42.

Frakes JT: Biliary pancreatitis: A review. J Clin Gastroenterol 1999;28:97-109.

Hochwalk SN, Dobryansky M, Rofsky NM et al: Magnetic resonance

cholangiopancreatography accurately predicts the presence or absence of

choledocholithiasis. J Gastrointest Surg 1998;2:573-9.

Gupta R, Toh SKC, CD: Early ERCP is an essential part of the management

of all cases of acute pancreatitis. Ann R Coll Surg Engl 1999;81:46-50.

Neoptolemos JP, Carr-Locke DL, London NJ et al: Controlled trial of urgent

endoscopic retrograde cholangiopancreatography and endoscopic sphincterotomy

versus conservative treatment for acute pancreatitis due to gallstones. Lancet

1988;vol 1, 979-83.

Nowak A, Nowakowska-Dulawa E, Marek TA, Rybicka J: Final results of the

prospective, randomized controlled study on endoscopic sphiuncterotomy or

conventional management in acute biliary pancreatitis. Gastroenterolgy

1995;108:A380.

Pezzilli R, Billi P, Barakat B et al: Ultrasonographic evaluation of teh common

bile duct in biliary acute pancreatitis patients: comparison with endoscopic

retrograde cholangiopancreatography,

J Ultrasound Med 1999;18:391-4.

Sugiyama M, Atomi Y: Acute biliary pancreatitis: the roles of endoscopic

ultrasonography and endoscopic retrograde cholangiopancreatography. Surgery

1998;124:14-21.