- Email: [email protected]
Urgent biliary decompression after endoscopic retrograde cholangiopancreatography
Urgent Biliary Decompression After Endoscopic Retrograde Cholangiopancreatography Edward C.S. Lai, Fxcs(Ed), FRACS, Chung-Mao Lo, MB, BS, Tat-Kuen Choi, MD, Wing-Keung Cheng, MB, BS, Sheung-Tat Fan, FRCS(Glasg), John WOng, PhD, FRACS, FRCS(Ed), FACS, Hong Kong
Acute cholangitis complicating diagnostic endoscopic retrograde cholangiopancreatography (ERCP) is potentially fatal. Among 323 consecutive patients with proved hiliary obstruction, 21 (7 percent) developed acute cholangitis after examination. Four patients underwent emergency surgery for the control of sepsis with two deaths. Of the 21 parameters chosen for evaluation, malignant obstruction, fever (higher than 37.5”C) within 72 hours before the procedure or when afebrile, and an increased aspartate transaminase level of 70 IU or more were the independent predictive factors identified by multivariate analysis. An increased temperature should be regarded as an absolute contraindication to examination unless followed by immediate ductal drainage. Since the risk of septic complications is minimal when none of the risk factors are present, routine urgent biliary decompression after ERCP is probably unnecessary for these selected patients. For patients with malignant obstruction or other risk factors, early elective surgical drainage is advisable. When surgery is not feasible, nonoperative drainage of the obstructed biliary system as a preventive measure might be considered.
ince its introduction over two decades ago, endoscopic retrograde cholangiopancreatography (ERCP) has S been established as a useful direct imaging technique for both biliary and pancreatic ductal systems. Among various common complications reported, acute pancreatitis and duodenal perforation probably represent technical mishaps which would decrease with increasing skill. Acute cholangitis, on the other hand, occurs in the presence of biliary obstruction in most instances and relates less to the experience of the endoscopist [ 2-41. The incidence of biliary sepsis after ERCP as a diagnostic examination has ranged from 0.12 percent to 6.5 percent in various series and is the most common cause of death ]1,2,5-71. Since most cholangitis develops within 72 From the Division of Hepatobiliary Surgery, Department of Surgery, University of Hong Kong, Queen Mary Hospital, Hong Kong. Requests for reprints should he addressed to Edward C.S. Lai, Division of Hepatobiliary Surgery, Department of Surgery, University of Hong Kong, Queen Mary Hospital, Hong Kong. presented at the 29th Annual Meeting of the Society for Surgery of the Alimentary Tract, New Orleans, Louisiana, May 17- 18, 1988.
hours after ERCP, routine urgent biliary decompression within 24 hours after the procedure has been advocated by many investigators [4,6,7]. We undertook a retrospective study to examine the validity of this doctrine in patients with definite biliary obstruction as delineated successfully by endoscopic cholangiography. PATIENTS
AND METHODS
The clinical records and cholangiograms of all patients who underwent diagnostic ERCP at the endoscopic unit of the Department of Surgery, University of Hong Kong, Queen Mary Hospital, between January 1, 1984 and July 3 1, 1987 were retrospectively reviewed. All patients had demonstrable biliary obstruction, such as calculi, residing in the extrahepatic or intrahepatic ductal system and stricture of a benign or malignant nature on endoscopic cholangiograms. In the presence of complete obstruction at the ampullary level preventing opacification, patients were also evaluated if cannulation had been attempted. Ail patients who underwent the endoscopic procedure fasted for a minimum of 6 hours. Examinations were conducted under local anesthesia with oropharyngeal anesthetic spray and intravenous diazepam (Diazemulb; Kabivitrum, Stockholm, Sweden). A single 40 mg dose of gentamicin was routinely given immediately before examination as a prophylactic antibiotic but omitted when the patient was already on appropriate therapeutic antimicrobials. No additional antibiotics were mixed with the contrast medium for opacification of the biliary system. Duodenal relaxation was achieved by intravenous scopolamine n-butylbromide (Buscopana, German Remedies Ltd., Bombay, India) and occasionally supplemented by intravenous glucagon. The commonly used side-view endoscopes were the Olympus lT, lT-10, and JF lT-10. Complete opacification of the entire biliary system was attempted whenever possible. During the 43-month study period, 323 patients were found to be eligible. There were 17 1 men and 152 women ranging in age from 15 to 94 years (mean age 63.84 f 0.77 years, mean f SE). After exclusion of other probable causes of fever, the patient was considered to have acute cholangitis related to ERCP if a temperature greater than or equal to 383°C developed within 72 hours after the procedure in conjunction with either abdominal pain or obstructive jaundice. The diagnosis was made on the aforementioned clinical features irrespective of the result of blood cultures. Death was considered related to the endoscopic examination when the patient died within the same hospitalization period after an operation for the endoscopic complication or within 7 days after the procedure without any further therapeutic intervention. The following clinical and laboratory data were evalu-
THE AMERICAN JOURNAL OF SURGERY
VOLUME 157 JANUARY 1989
121
TABLE I
TABLE II Final Pathologic Dlagnosls Based on Cllnlcal and Cholanglographlc Flndlngs
Parameters for Risk Factors of Acute Cholangitls After Endoscoplc Retrograde Cholanglopancreatography (ERCP) Clinical (n = 3)
Laboratory (n = 14)
Cholangiographic (n = 4)
Pathologic nature of obstruction, Maximum temperature within 72 hours before ERCP, Nature of antibiotic coverage WBC count; total bilirubin, ALP, GGTP, AST, ALT, & amylase at admission & within 1 week before ERCP Number of obstructed areas, Level of obstruction, Extent of opacification, Cause of obstruction
ALP = alkaline phosphatase; ALT = alanine transaminase; AST = aspartate transaminase; GGTP = gamma glutamyl transpeptidase; WBC = white blood cell.
Benign
Malignant
Total
Pathologic Diagnosis
Patients
Common duct stone with or without gallstone Recurrent pyogenic cholangltis latrogenic stricture Bile duct cancer Carcinoma of pancreas Carcinoma of liver Metastatic disease
217
..
323
(n)
60 1 27 11 2 5
RESULTS Among the 323 patients, the majority (54 percent) had previous attacks of acute cholangitis. The remaining patients had an endoscopic examination as part of the investigation to determine the underlying cause of acute pancreatitis (48 patients) and obstructive jaundice (102 patients). Jaundice with serum total bilirubin levels over
26 mmol/liter was present in 80 percent and 61 percent of the patients at the time of presentation and before endoscopy, respectively. Prophylactic antibiotic alone was given to 267 patients (83 percent) and the rest had different durations of antibiotic therapy. Although most patients were afebrile within the last 72 hours before the procedure, 43 patients had temperatures greater than 37.5’C. Benign lesions (86 percent) were detected more often than malignant lesions. The final diagnoses are shown in Table II. Ninety-five patients had obstruction confined to only one anatomic region, and the rest (71 percent) had more than one region involved. Isolated intrahepatic calculus or stricture without distal abnormalities occurred in 7 percent of the patients. Most patients had choledocholithiasis with or without associated biliary stricture (11 percent and 72 percent, respectively). Opacification of the biliary tract was complete in 256 patients but incomplete due to the presence of high-grade obstruction in 38 patients. Complete blockage preventing filling of the proximal ductal system was found in the remaining 29 patients. Acute cholangitis developed in 21 patients (7 percent) after the diagnostic endoscopic procedure. Of the 29 patients who had complete ductal obstruction, 2 with neoplastic lesions located proximal to the ampulla had septic complications. Culture was positive in 4 of the 11 patients who had blood sampled during febrile attacks. Sepsis resolved with conservative measures in 17 patients, but emergency surgery was necessary in 4 patients. Two patients who had bile duct cancer died from persistent septicemia despite urgent surgical decompression. Both were afebrile before the procedure and had complete opacification of the ductal system. Thus, when only the 45 patients with malignant biliary obstruction were considered, the mortality rate among the 10 patients (22 percent) who developed cholangitis was 20 percent. The remaining two patients with benign obstruction recovered uneventfully after surgery. The other significant complication was acute pancreatitis, which occurred in two patients. The third death in the present series was an elderly lady with carcinoma of the gallbladder who died from myocardial infarction 48 hours after an uneventful examination.
122
157
ated to determine the level of risk of acute cholangitis complicating ERCP: pathologic nature of the biliary obstruction (benign or malignant); maximum temperature recorded within the last 72 hours before ERCP; nature of the antibiotic coverage (prophylactic or therapeutic); and white cell count and routine liver function tests sampled at the time of admission as well as the last available set collected within 1 week prior to the examination. Four cholangiographic features were analyzed according to the following definitions: (1) If each part of the ductal system (right and left intrahepatic ducts, common hepatic duct, and common bile duct) was taken as an individual anatomic region, the number of obstructed areas was considered multiple when more than one region were obstructed by either stricture or calculous disease. (2) The level of obstruction was labeled as proximal when biliary abnormalities were confined to regions proximal to the common hepatic duct with no distal involvement. In the presence of a distal obstructive element, it was defined as distal. (3) The extent of opacification was judged according to the difference between the areas proximal and distal to the site of obstruction. (4) The obstructive lesions were categorized according to the underlying cause (stones, strictures, or a combination of both). The variables analyzed are shown in Table I. Statistical calculations were performed with standard computer programs. The chi-square test, with or without Yates correction, and the Mann-Whitney U test were used when appropriate for categoric and continuous variables. Multivariate analysis of the identification of risk factors was performed by stepwise logistic regression (statistical significance p <0.05).
THE AMERICAN
JOURNAL
OF SURGERY
VOLUME
JANUARY
1989
BILIABY DECOMPRESSION AFTER ENDOSCOPIC RETROGRADE CHOLANGIOPANCREATOGRAPHY
TABLE IV PredIcted Risk of Acute Cholangltis After Endoscoplc Retrograde Cholanglopancreatography in 323 Patients*
TABLE III Factors Associated With Acute Cholangiils After Endoscopic Retrograde Cholangiopancreatography (ERCP) In 323 Patients* Patients+ Total Afebrile Obstruction Benign Malignant Lesions Stone Stricture Both Admission Total billrubln _<120 mmollllter >I20 mmobliter Alkaline phosphatase 1400 ~mol/min/llter >400 jrmol/min/liter Pm-ERCP Billrubin I120 mmol/liter > 120 mmol/llter Glutamyl trenspeptkiase 1300 ~mol/mln/liter >300 pmol/mln/llter Aspartate transmlnase 570 IU >70 IU White blood cell count (cellslmm3) 110 x 103 >I0 x 103
p value Total Afebrile
4 22
1 14
0.0006
0.005
4 17 9
2 13 4
0.002
0.002
0.012
0.042
4 13
0.004
4 15
0.05
...
0.0025
0.011
0.026
* Afebrile patients (n = 260) had maximum temperature of 37.5OC or lower after ERCP. + Values expressed as percentages.
Hence, the overall morbidity and mortality rates in the present series were 7 percent and 1 percent, respectively. Seven factors were found to be significantly related to increased risk of acute cholangitis after ERCP on univariate analysis (Table III). Stepwise logistic regression, however, showed that only the pathologic nature of the obstructive lesion and fever (higher than 37S°C) within the 72 hours prior to examination were independent predictive factors (discriminant coefficients 1.75 and 2.73; p <0.002 and p
THE AMERICAN
17 2
54 11
4 3
3 21
COMMENTS
0.029
4 10
Febrile (n = 43) Afebrile (n = 260) Aspartate transaminase level 570 IU >70 IU
* Febrile patients had maximum temperature above 37.5’C and afebrile patients ha.d maximum temperature of 37.5’C or below before endoscopic retrograde cholangiopancreatography.
0.035
4 15
Nature of Obstruction (% ) Benign Malignant
With increasing use of ERCP, acute cholangitis complicating the procedure is a growing concern since it is a definite risk that occasionally leads to death. However, few attempts have been made to identify the factors that lead to biliary sepsis after ERCP was carried out to aid better patient selection for appropriate management. A transient temperature increase that subsides rapidly is not infrequent after ERCP [8,9]. In a prospective study by Parker et al [9], of 50 patients, 20 percent had temperature over 37.8’C within the subsequent 72 hours after examination. Since the present study included patients with proved ductal obstruction, the presence of high fever (38.3’C or higher), together with abdominal pain or jaundice would strongly suggest the diagnosis of acute cholangitis. The incidence of 7 percent is comparable to other reports when patients with documented ductal obstruction are considered [6,10,11]. Different clinical, laboratory, and radiographic features were chosen on the basis of easy accessibility at the time of ERCP. Assessment of the risk of subsequent cholangitis could hence be made readily, allowing a prompt decision for appropriate management afterwards. When the data of the 323 patients included were studied, the presence of malignant biliary obstruction and fever within 72 hours before ERCP were the only significant variables that stood the test of multivariate analysis. Similarly, Takada [II] noted that acute cholangitis occurred significantly more frequently in patients with neoplastic stricture among a group of 364 patients with proved biliary obstruction. Early elective biliary decompression would be warranted for all patients with malignant lesions, since both procedure-related deaths in the present series occurred in this group despite the absence of fever before examination. From our analysis, it is clear that the presence of even a low-grade fever prior to ERCP, indicating residual biliary sepsis, would increase the subsequent risk of acute cholangitis. Similarly, Takada [I I] cautioned that ERCP should not be performed in persons with symptoms of cholangitis. Recently there has been a surge of interest in the use of endoscopic sphincterotomy and nasobiliary drainage, both in combination and alone, immediately after diagnostic ERCP in the management of
JOURNAL
OF SURGERY
VOLUME
157
JANUARY
1989
123
patients during an attack of acute suppurative cholangitis. Septic episodes have usually resolved after the endoscopic intervention without being aggravated as predicted by our analysis [ 12-141. The present study might serve as indirect evidence of the efficacy of these therapeutic endoscopic procedures. Nonetheless, when the examination is conducted solely for diagnostic purposes, an increased temperature within the last 72 hours should be regarded as a contraindication. The extent of opacification of the ductal system proximal to an obstructive lesion is another probable determinant [Ill. A ductal system with high-grade obstruction, as indicated by incomplete opacification, allows stasis of bile and introduction of bacterial flora from endoscopic instruments [ 15,16]. With injection of hyperosmolar contrast media, increased intraductal pressure secondary to the rapid fluid influx would generate biliary sepsis [4]. Of the 132 patients reported by Takada [ll] who had complete neoplastic occlusion of the biliary tract, none developed acute cholangitis after ERCP. In contrast to his experience, 2 of our 29 patients with complete ductal obstruction had septic complication. It is possible that bacteria might have been introduced into the proximal obstructed system which remained unopacified by the amount of contrast medium injected. Judging from the incidence of cholangitis we reported, a single dose of gentimicin before examination offered no definite benefit in the reduction of biliary sepsis. Other measures described, including proper disinfection of the instruments and prophylactic antibiotics either mixed with the contrast medium for injection or given parentally after the procedure, have been used with apparent success [2,15-181. Surgical decompression within 24 hours after ERCP for all patients is the most secure and widely accepted preventive measure for biliary sepsis [4,6,7]. Recently, biliary drainage, by either endoscopic sphincterotomy or the percutaneous route, has also been suggested [ 11,I 91. The present study established two pairs of independent risk factors: malignancy and fever within 3 days prior to examination and increased serum aspartate transaminase levels before the endoscopic study in afebrile patients with proved biliary obstruction. The danger of acute cholangitis complicating ERCP among those low-risk patients who have none of the aforementioned risk factors is minimal. Unless used as a definitive measure, routine prophylactic biliary decompression by nonoperative means for this selected group of patients is not justified. Complication rates of 7 to 20 percent, either from endoscopic sphincterotomy or percutaneous transhepatic drainage, are higher than those of biliary sepsis [20-231. When these patients are encountered at the time of ERCP, precautions such as limiting the amount of medium injected to a minimum for diagnosis and commencing with antibiotics for coverage after the examination would probably suffice [24]. For patients with underlying malignancies or other risk factors who are at high risk for subsequent cholangitis, the application of nonsurgical drainage procedure might be considered when urgent surgical decompression within 24 hours is not feasible. 124
THE AMERICAN JOURNAL OF SURGERY
However, the potential benefits derived from such an approach would require further prospective evaluation. REFERENCES 1. Bilbao MK, Dotter CT, Lee TG, Katon RM. Complications of endoscopic retrograde cholangiopancreatography (ERCP). Gastroenterology 1976; 70: 3 14-20. 2. Zimmon DS, Falkenstein DB, Riccobono C, Aaron B. Complications of endoscopic retrograde cholangiopancreatography. Analysis of 300 consecutive cases. Gastroenterology 1975; 69: 303-9. 3. Kessler RE, Falkenstein DB, Clemett AR, Zimmon DS. Indications, clinical value and complications of endoscopic retrograde cholangiopancreatography. Surg Gynecol Obstet 1976; 142: 865 70. 4. Davis JL, Milligan FD, Cameron JL. Septic complications following endoscopic retrograde cholangiopancreatography. Surg Gynecol Obstet 1975; 140: 365-7. 5. Seifert E. Endoscopic retrograde cholangiopancreatography. Am J Gastroenterol 1977; 68: 542-9. 6. Vennes JA, Jacobson JR, Silvis SE. Endoscopic cholangiography for biliary system diagnosis. Ann Intern Med 1974; 80: 61-4. 7. Brandes JW, Scheffer B, Lorenz-Meyer H, Korst HA, Littmann KP. ERCP: Complications and prophylaxis. A controlled study. Endoscopy 1981; 13: 27-30. 8. Hershey SD, Sugawa C, Cushing R, Ledgerwood AM, Lucas CE. The value of prophylactic antibiotic therapy during endoscopic retrograde cholangiopancreatography. Surg Gynecol Obstet 1982; 155: 801-3. 9. Parker HW, Geenan JE, Bjork JT, Stewart ET. A prospective analysis of fever and bacteraemia following ERCP. Gastrointest Endosc 1979; 25: 102-3. 10. Thurnherr N, Bruhlman KF, Kreys GI, Bianchi L, Faust H, Blum AL. Fulminant cholangitis and septicemia after endoscopic retrograde cholangiography (ERC) in two patients with obstructive jaundice. Dig Dis 1976; 21: 477-81. 11. Takada T. Cholangitis. In: Takemoto T, Kasugai T, eds. Endoscopic retrograde cholangiopancreatography. New York: Igakushoin, 1978: 277-93. 12. Leese T, Neoptolemos JP, Baker AR, Carr-Locke DL. Management of acute cholangitis and the impact of endoscopic sphincterotomy. Br J Surg 1986; 73: 988-92. 13. Gogel HK, Runyon BA, Volpicelli NA, Palmer RC. Acute suppurative obstructive cholangitis due to stones: treatment by urgent endoscopic sphincterotomy. Gastrointest Endosc 1987; 33: 210-3. 14. Cotton PB, Burney PGJ, Mason RR. Transnasal bile duct catheterization after endoscopic sphincterotomy. Method for biliary drainage, perfusion and sequential cholangiography. Gut 1979; 20: 285-70. 15. Elson CO, Hattori K, Blackstone MO. Polymicrobial sepsis following endoscopic retrograde cholangiopancreatography. Gastroenterology 1975; 69: 507-10. 16. Low DE, Mickflikier AB, Kennedy JK, Stiver HG. Infectious complications of endoscopic retrograde cholangiopancreatography. Arch Intern Med 1980; 140: 1076-7. 17. Kasugai T, Kuno N, Kobayashi S, Hattori K. Endoscopic pancreatocholangiography I: the normal pancreatocholangiogram. Gastroenterology 1972; 63: 227-34. 18. Liguory C, Gouerou H, Chavy A, Coffin JC, Huguier M. Endoscopic retrograde cholangiopancreatography. Br J Surg 1974; 61: 359-62. 19. O’Connor HJ, Axon ATR. Gastrointestinal endoscopy: infection and disinfection. Gut 1983; 24: 1067-77. 20. Cotton PB, Vallon AG. British experience with duodenoscopic sphincterotomy for removal of bile duct stones. Br J Surg 1981; 68: 373-5. 21. Leese T, Neoptolemos JP, Carr-Locke DL. Successes, failures, early complications and their management following endoscopic
VOLUME 157 JANUARY 1989
BILIARY DECOMPRESSION AFTER ENDOSCOPIC RETROGRADE CHOLANGIOPANCREATOGRAPHY
sphincterotomy: results in 394 consecutive patients from a single centre. Br J Surg 1985; 72: 215-9. 22. Mueller PR, Van Sonnenberg E, Ferrucci JT Jr. Percutaneous biliary drainage: technical and catheter-related problems in 200 procedures. AJR 1982; 138: 17-23. 23. Pessa ME, Hawkins IF, Vogel SB. The treatment of acute cholangitis percutaneous transhepatic biliary drainage before definitive therapy. Ann Surg 1987; 205: 389-92. 24. Cotton PB. ERCP. Gut 1977; 18: 316-41.
DISCUSSION Melody 0. Allen (Minneapolis, MN): Dr. Lai, do
you attribute any significance to the rate at which the contrast material drained from the obstructed duct at the end of the procedure? Do you feel that it is a predictor of later development of sepsis? We usually think it better if it does drain rapidly. Is that consistent with your experience? In the 2 1 patients who did develop clinical evidence of cholangitis, what were the organisms eventually cultured from the bile? Do you believe that they were introduced into the ductal system at the time of ERCP? I am particularly interested in the presence of Pseudomonas or Candida, which are indicators of contamination during ERCP. Frank Toups (Metairie, LA): I think gentamicin is a dangerous drug. Aminoglycoside can cause kidney damage, and several of the patients did die with sepsis. Dr. Lai, did you administer just one dose or several doses of gentamicin? Would you consider ticarcillin (Timentina), which is a safer, yet as effective, drug? It has essentially the same bacterial effect as gentamicin but is not as dangerous. John W. Braasch (Burlington, MA): Dr. Lai, should you be prepared for a papillotomy in a seriously ill patient
at the time of ERCP, which could obviate cholangitis and septic death? Ali Ghazi (New York, NY): Did you perform endoscopic papillotomy in those patients with biliary obstruction? Did you use a stent to drain the malignant obstruction and prevent surgery and cholangitis? Edward C.S. Lai (closing): Dr. Allen, with regard to the delayed ductal drainage of contrast medium after examination, we believe it is an important determinant in the prediction of subsequent sepsis as it relates to the degree of obstruction. Due to the retrospective nature of this study, detailed documentation of the drainage time was not available. As far as the results of bacteriologic studies in those patients who developed acute cholangitis, unfortunately, only 11 patients had such studies. Escherichia coli was the most common organism isolated among the four patients with positive cultures. The other major organism was Klebsiella. Dr. Toups, the use of gentamicin as a prophylactic antibiotic was obviously not effective. We have since changed to other third-generation cephalosporins. The use of antibiotics was reserved for only those patients with documented obstruction. Three doses of antibiotics were given intravenously as prophylaxis. Drs. Braasch and Ghazi, we are opposed to the use of papillotomy as a prophylactic measure, particularly in patients without any of the mentioned risk factors. The use of nonoperative drainage for malignant obstruction is also probably not advisable, since there is no clear evidence to suggest its definite role in the prevention of sepsis. In fact, according to different series on either papillotomy or other nonoperative measures, the complication rates reported were higher than our predicted risk of sepsis.
THE AMERICAN JOURNAL OF SURGERY
VOLUME 157 JANUARY
1989
125
Acute cholangitis complicating diagnostic endoscopic retrograde cholangiopancreatography (ERCP) is potentially fatal. Among 323 consecutive patients with proved hiliary obstruction, 21 (7 percent) developed acute cholangitis after examination. Four patients underwent emergency surgery for the control of sepsis with two deaths. Of the 21 parameters chosen for evaluation, malignant obstruction, fever (higher than 37.5”C) within 72 hours before the procedure or when afebrile, and an increased aspartate transaminase level of 70 IU or more were the independent predictive factors identified by multivariate analysis. An increased temperature should be regarded as an absolute contraindication to examination unless followed by immediate ductal drainage. Since the risk of septic complications is minimal when none of the risk factors are present, routine urgent biliary decompression after ERCP is probably unnecessary for these selected patients. For patients with malignant obstruction or other risk factors, early elective surgical drainage is advisable. When surgery is not feasible, nonoperative drainage of the obstructed biliary system as a preventive measure might be considered.
ince its introduction over two decades ago, endoscopic retrograde cholangiopancreatography (ERCP) has S been established as a useful direct imaging technique for both biliary and pancreatic ductal systems. Among various common complications reported, acute pancreatitis and duodenal perforation probably represent technical mishaps which would decrease with increasing skill. Acute cholangitis, on the other hand, occurs in the presence of biliary obstruction in most instances and relates less to the experience of the endoscopist [ 2-41. The incidence of biliary sepsis after ERCP as a diagnostic examination has ranged from 0.12 percent to 6.5 percent in various series and is the most common cause of death ]1,2,5-71. Since most cholangitis develops within 72 From the Division of Hepatobiliary Surgery, Department of Surgery, University of Hong Kong, Queen Mary Hospital, Hong Kong. Requests for reprints should he addressed to Edward C.S. Lai, Division of Hepatobiliary Surgery, Department of Surgery, University of Hong Kong, Queen Mary Hospital, Hong Kong. presented at the 29th Annual Meeting of the Society for Surgery of the Alimentary Tract, New Orleans, Louisiana, May 17- 18, 1988.
hours after ERCP, routine urgent biliary decompression within 24 hours after the procedure has been advocated by many investigators [4,6,7]. We undertook a retrospective study to examine the validity of this doctrine in patients with definite biliary obstruction as delineated successfully by endoscopic cholangiography. PATIENTS
AND METHODS
The clinical records and cholangiograms of all patients who underwent diagnostic ERCP at the endoscopic unit of the Department of Surgery, University of Hong Kong, Queen Mary Hospital, between January 1, 1984 and July 3 1, 1987 were retrospectively reviewed. All patients had demonstrable biliary obstruction, such as calculi, residing in the extrahepatic or intrahepatic ductal system and stricture of a benign or malignant nature on endoscopic cholangiograms. In the presence of complete obstruction at the ampullary level preventing opacification, patients were also evaluated if cannulation had been attempted. Ail patients who underwent the endoscopic procedure fasted for a minimum of 6 hours. Examinations were conducted under local anesthesia with oropharyngeal anesthetic spray and intravenous diazepam (Diazemulb; Kabivitrum, Stockholm, Sweden). A single 40 mg dose of gentamicin was routinely given immediately before examination as a prophylactic antibiotic but omitted when the patient was already on appropriate therapeutic antimicrobials. No additional antibiotics were mixed with the contrast medium for opacification of the biliary system. Duodenal relaxation was achieved by intravenous scopolamine n-butylbromide (Buscopana, German Remedies Ltd., Bombay, India) and occasionally supplemented by intravenous glucagon. The commonly used side-view endoscopes were the Olympus lT, lT-10, and JF lT-10. Complete opacification of the entire biliary system was attempted whenever possible. During the 43-month study period, 323 patients were found to be eligible. There were 17 1 men and 152 women ranging in age from 15 to 94 years (mean age 63.84 f 0.77 years, mean f SE). After exclusion of other probable causes of fever, the patient was considered to have acute cholangitis related to ERCP if a temperature greater than or equal to 383°C developed within 72 hours after the procedure in conjunction with either abdominal pain or obstructive jaundice. The diagnosis was made on the aforementioned clinical features irrespective of the result of blood cultures. Death was considered related to the endoscopic examination when the patient died within the same hospitalization period after an operation for the endoscopic complication or within 7 days after the procedure without any further therapeutic intervention. The following clinical and laboratory data were evalu-
THE AMERICAN JOURNAL OF SURGERY
VOLUME 157 JANUARY 1989
121
TABLE I
TABLE II Final Pathologic Dlagnosls Based on Cllnlcal and Cholanglographlc Flndlngs
Parameters for Risk Factors of Acute Cholangitls After Endoscoplc Retrograde Cholanglopancreatography (ERCP) Clinical (n = 3)
Laboratory (n = 14)
Cholangiographic (n = 4)
Pathologic nature of obstruction, Maximum temperature within 72 hours before ERCP, Nature of antibiotic coverage WBC count; total bilirubin, ALP, GGTP, AST, ALT, & amylase at admission & within 1 week before ERCP Number of obstructed areas, Level of obstruction, Extent of opacification, Cause of obstruction
ALP = alkaline phosphatase; ALT = alanine transaminase; AST = aspartate transaminase; GGTP = gamma glutamyl transpeptidase; WBC = white blood cell.
Benign
Malignant
Total
Pathologic Diagnosis
Patients
Common duct stone with or without gallstone Recurrent pyogenic cholangltis latrogenic stricture Bile duct cancer Carcinoma of pancreas Carcinoma of liver Metastatic disease
217
..
323
(n)
60 1 27 11 2 5
RESULTS Among the 323 patients, the majority (54 percent) had previous attacks of acute cholangitis. The remaining patients had an endoscopic examination as part of the investigation to determine the underlying cause of acute pancreatitis (48 patients) and obstructive jaundice (102 patients). Jaundice with serum total bilirubin levels over
26 mmol/liter was present in 80 percent and 61 percent of the patients at the time of presentation and before endoscopy, respectively. Prophylactic antibiotic alone was given to 267 patients (83 percent) and the rest had different durations of antibiotic therapy. Although most patients were afebrile within the last 72 hours before the procedure, 43 patients had temperatures greater than 37.5’C. Benign lesions (86 percent) were detected more often than malignant lesions. The final diagnoses are shown in Table II. Ninety-five patients had obstruction confined to only one anatomic region, and the rest (71 percent) had more than one region involved. Isolated intrahepatic calculus or stricture without distal abnormalities occurred in 7 percent of the patients. Most patients had choledocholithiasis with or without associated biliary stricture (11 percent and 72 percent, respectively). Opacification of the biliary tract was complete in 256 patients but incomplete due to the presence of high-grade obstruction in 38 patients. Complete blockage preventing filling of the proximal ductal system was found in the remaining 29 patients. Acute cholangitis developed in 21 patients (7 percent) after the diagnostic endoscopic procedure. Of the 29 patients who had complete ductal obstruction, 2 with neoplastic lesions located proximal to the ampulla had septic complications. Culture was positive in 4 of the 11 patients who had blood sampled during febrile attacks. Sepsis resolved with conservative measures in 17 patients, but emergency surgery was necessary in 4 patients. Two patients who had bile duct cancer died from persistent septicemia despite urgent surgical decompression. Both were afebrile before the procedure and had complete opacification of the ductal system. Thus, when only the 45 patients with malignant biliary obstruction were considered, the mortality rate among the 10 patients (22 percent) who developed cholangitis was 20 percent. The remaining two patients with benign obstruction recovered uneventfully after surgery. The other significant complication was acute pancreatitis, which occurred in two patients. The third death in the present series was an elderly lady with carcinoma of the gallbladder who died from myocardial infarction 48 hours after an uneventful examination.
122
157
ated to determine the level of risk of acute cholangitis complicating ERCP: pathologic nature of the biliary obstruction (benign or malignant); maximum temperature recorded within the last 72 hours before ERCP; nature of the antibiotic coverage (prophylactic or therapeutic); and white cell count and routine liver function tests sampled at the time of admission as well as the last available set collected within 1 week prior to the examination. Four cholangiographic features were analyzed according to the following definitions: (1) If each part of the ductal system (right and left intrahepatic ducts, common hepatic duct, and common bile duct) was taken as an individual anatomic region, the number of obstructed areas was considered multiple when more than one region were obstructed by either stricture or calculous disease. (2) The level of obstruction was labeled as proximal when biliary abnormalities were confined to regions proximal to the common hepatic duct with no distal involvement. In the presence of a distal obstructive element, it was defined as distal. (3) The extent of opacification was judged according to the difference between the areas proximal and distal to the site of obstruction. (4) The obstructive lesions were categorized according to the underlying cause (stones, strictures, or a combination of both). The variables analyzed are shown in Table I. Statistical calculations were performed with standard computer programs. The chi-square test, with or without Yates correction, and the Mann-Whitney U test were used when appropriate for categoric and continuous variables. Multivariate analysis of the identification of risk factors was performed by stepwise logistic regression (statistical significance p <0.05).
THE AMERICAN
JOURNAL
OF SURGERY
VOLUME
JANUARY
1989
BILIABY DECOMPRESSION AFTER ENDOSCOPIC RETROGRADE CHOLANGIOPANCREATOGRAPHY
TABLE IV PredIcted Risk of Acute Cholangltis After Endoscoplc Retrograde Cholanglopancreatography in 323 Patients*
TABLE III Factors Associated With Acute Cholangiils After Endoscopic Retrograde Cholangiopancreatography (ERCP) In 323 Patients* Patients+ Total Afebrile Obstruction Benign Malignant Lesions Stone Stricture Both Admission Total billrubln _<120 mmollllter >I20 mmobliter Alkaline phosphatase 1400 ~mol/min/llter >400 jrmol/min/liter Pm-ERCP Billrubin I120 mmol/liter > 120 mmol/llter Glutamyl trenspeptkiase 1300 ~mol/mln/liter >300 pmol/mln/llter Aspartate transmlnase 570 IU >70 IU White blood cell count (cellslmm3) 110 x 103 >I0 x 103
p value Total Afebrile
4 22
1 14
0.0006
0.005
4 17 9
2 13 4
0.002
0.002
0.012
0.042
4 13
0.004
4 15
0.05
...
0.0025
0.011
0.026
* Afebrile patients (n = 260) had maximum temperature of 37.5OC or lower after ERCP. + Values expressed as percentages.
Hence, the overall morbidity and mortality rates in the present series were 7 percent and 1 percent, respectively. Seven factors were found to be significantly related to increased risk of acute cholangitis after ERCP on univariate analysis (Table III). Stepwise logistic regression, however, showed that only the pathologic nature of the obstructive lesion and fever (higher than 37S°C) within the 72 hours prior to examination were independent predictive factors (discriminant coefficients 1.75 and 2.73; p <0.002 and p
THE AMERICAN
17 2
54 11
4 3
3 21
COMMENTS
0.029
4 10
Febrile (n = 43) Afebrile (n = 260) Aspartate transaminase level 570 IU >70 IU
* Febrile patients had maximum temperature above 37.5’C and afebrile patients ha.d maximum temperature of 37.5’C or below before endoscopic retrograde cholangiopancreatography.
0.035
4 15
Nature of Obstruction (% ) Benign Malignant
With increasing use of ERCP, acute cholangitis complicating the procedure is a growing concern since it is a definite risk that occasionally leads to death. However, few attempts have been made to identify the factors that lead to biliary sepsis after ERCP was carried out to aid better patient selection for appropriate management. A transient temperature increase that subsides rapidly is not infrequent after ERCP [8,9]. In a prospective study by Parker et al [9], of 50 patients, 20 percent had temperature over 37.8’C within the subsequent 72 hours after examination. Since the present study included patients with proved ductal obstruction, the presence of high fever (38.3’C or higher), together with abdominal pain or jaundice would strongly suggest the diagnosis of acute cholangitis. The incidence of 7 percent is comparable to other reports when patients with documented ductal obstruction are considered [6,10,11]. Different clinical, laboratory, and radiographic features were chosen on the basis of easy accessibility at the time of ERCP. Assessment of the risk of subsequent cholangitis could hence be made readily, allowing a prompt decision for appropriate management afterwards. When the data of the 323 patients included were studied, the presence of malignant biliary obstruction and fever within 72 hours before ERCP were the only significant variables that stood the test of multivariate analysis. Similarly, Takada [II] noted that acute cholangitis occurred significantly more frequently in patients with neoplastic stricture among a group of 364 patients with proved biliary obstruction. Early elective biliary decompression would be warranted for all patients with malignant lesions, since both procedure-related deaths in the present series occurred in this group despite the absence of fever before examination. From our analysis, it is clear that the presence of even a low-grade fever prior to ERCP, indicating residual biliary sepsis, would increase the subsequent risk of acute cholangitis. Similarly, Takada [I I] cautioned that ERCP should not be performed in persons with symptoms of cholangitis. Recently there has been a surge of interest in the use of endoscopic sphincterotomy and nasobiliary drainage, both in combination and alone, immediately after diagnostic ERCP in the management of
JOURNAL
OF SURGERY
VOLUME
157
JANUARY
1989
123
patients during an attack of acute suppurative cholangitis. Septic episodes have usually resolved after the endoscopic intervention without being aggravated as predicted by our analysis [ 12-141. The present study might serve as indirect evidence of the efficacy of these therapeutic endoscopic procedures. Nonetheless, when the examination is conducted solely for diagnostic purposes, an increased temperature within the last 72 hours should be regarded as a contraindication. The extent of opacification of the ductal system proximal to an obstructive lesion is another probable determinant [Ill. A ductal system with high-grade obstruction, as indicated by incomplete opacification, allows stasis of bile and introduction of bacterial flora from endoscopic instruments [ 15,16]. With injection of hyperosmolar contrast media, increased intraductal pressure secondary to the rapid fluid influx would generate biliary sepsis [4]. Of the 132 patients reported by Takada [ll] who had complete neoplastic occlusion of the biliary tract, none developed acute cholangitis after ERCP. In contrast to his experience, 2 of our 29 patients with complete ductal obstruction had septic complication. It is possible that bacteria might have been introduced into the proximal obstructed system which remained unopacified by the amount of contrast medium injected. Judging from the incidence of cholangitis we reported, a single dose of gentimicin before examination offered no definite benefit in the reduction of biliary sepsis. Other measures described, including proper disinfection of the instruments and prophylactic antibiotics either mixed with the contrast medium for injection or given parentally after the procedure, have been used with apparent success [2,15-181. Surgical decompression within 24 hours after ERCP for all patients is the most secure and widely accepted preventive measure for biliary sepsis [4,6,7]. Recently, biliary drainage, by either endoscopic sphincterotomy or the percutaneous route, has also been suggested [ 11,I 91. The present study established two pairs of independent risk factors: malignancy and fever within 3 days prior to examination and increased serum aspartate transaminase levels before the endoscopic study in afebrile patients with proved biliary obstruction. The danger of acute cholangitis complicating ERCP among those low-risk patients who have none of the aforementioned risk factors is minimal. Unless used as a definitive measure, routine prophylactic biliary decompression by nonoperative means for this selected group of patients is not justified. Complication rates of 7 to 20 percent, either from endoscopic sphincterotomy or percutaneous transhepatic drainage, are higher than those of biliary sepsis [20-231. When these patients are encountered at the time of ERCP, precautions such as limiting the amount of medium injected to a minimum for diagnosis and commencing with antibiotics for coverage after the examination would probably suffice [24]. For patients with underlying malignancies or other risk factors who are at high risk for subsequent cholangitis, the application of nonsurgical drainage procedure might be considered when urgent surgical decompression within 24 hours is not feasible. 124
THE AMERICAN JOURNAL OF SURGERY
However, the potential benefits derived from such an approach would require further prospective evaluation. REFERENCES 1. Bilbao MK, Dotter CT, Lee TG, Katon RM. Complications of endoscopic retrograde cholangiopancreatography (ERCP). Gastroenterology 1976; 70: 3 14-20. 2. Zimmon DS, Falkenstein DB, Riccobono C, Aaron B. Complications of endoscopic retrograde cholangiopancreatography. Analysis of 300 consecutive cases. Gastroenterology 1975; 69: 303-9. 3. Kessler RE, Falkenstein DB, Clemett AR, Zimmon DS. Indications, clinical value and complications of endoscopic retrograde cholangiopancreatography. Surg Gynecol Obstet 1976; 142: 865 70. 4. Davis JL, Milligan FD, Cameron JL. Septic complications following endoscopic retrograde cholangiopancreatography. Surg Gynecol Obstet 1975; 140: 365-7. 5. Seifert E. Endoscopic retrograde cholangiopancreatography. Am J Gastroenterol 1977; 68: 542-9. 6. Vennes JA, Jacobson JR, Silvis SE. Endoscopic cholangiography for biliary system diagnosis. Ann Intern Med 1974; 80: 61-4. 7. Brandes JW, Scheffer B, Lorenz-Meyer H, Korst HA, Littmann KP. ERCP: Complications and prophylaxis. A controlled study. Endoscopy 1981; 13: 27-30. 8. Hershey SD, Sugawa C, Cushing R, Ledgerwood AM, Lucas CE. The value of prophylactic antibiotic therapy during endoscopic retrograde cholangiopancreatography. Surg Gynecol Obstet 1982; 155: 801-3. 9. Parker HW, Geenan JE, Bjork JT, Stewart ET. A prospective analysis of fever and bacteraemia following ERCP. Gastrointest Endosc 1979; 25: 102-3. 10. Thurnherr N, Bruhlman KF, Kreys GI, Bianchi L, Faust H, Blum AL. Fulminant cholangitis and septicemia after endoscopic retrograde cholangiography (ERC) in two patients with obstructive jaundice. Dig Dis 1976; 21: 477-81. 11. Takada T. Cholangitis. In: Takemoto T, Kasugai T, eds. Endoscopic retrograde cholangiopancreatography. New York: Igakushoin, 1978: 277-93. 12. Leese T, Neoptolemos JP, Baker AR, Carr-Locke DL. Management of acute cholangitis and the impact of endoscopic sphincterotomy. Br J Surg 1986; 73: 988-92. 13. Gogel HK, Runyon BA, Volpicelli NA, Palmer RC. Acute suppurative obstructive cholangitis due to stones: treatment by urgent endoscopic sphincterotomy. Gastrointest Endosc 1987; 33: 210-3. 14. Cotton PB, Burney PGJ, Mason RR. Transnasal bile duct catheterization after endoscopic sphincterotomy. Method for biliary drainage, perfusion and sequential cholangiography. Gut 1979; 20: 285-70. 15. Elson CO, Hattori K, Blackstone MO. Polymicrobial sepsis following endoscopic retrograde cholangiopancreatography. Gastroenterology 1975; 69: 507-10. 16. Low DE, Mickflikier AB, Kennedy JK, Stiver HG. Infectious complications of endoscopic retrograde cholangiopancreatography. Arch Intern Med 1980; 140: 1076-7. 17. Kasugai T, Kuno N, Kobayashi S, Hattori K. Endoscopic pancreatocholangiography I: the normal pancreatocholangiogram. Gastroenterology 1972; 63: 227-34. 18. Liguory C, Gouerou H, Chavy A, Coffin JC, Huguier M. Endoscopic retrograde cholangiopancreatography. Br J Surg 1974; 61: 359-62. 19. O’Connor HJ, Axon ATR. Gastrointestinal endoscopy: infection and disinfection. Gut 1983; 24: 1067-77. 20. Cotton PB, Vallon AG. British experience with duodenoscopic sphincterotomy for removal of bile duct stones. Br J Surg 1981; 68: 373-5. 21. Leese T, Neoptolemos JP, Carr-Locke DL. Successes, failures, early complications and their management following endoscopic
VOLUME 157 JANUARY 1989
BILIARY DECOMPRESSION AFTER ENDOSCOPIC RETROGRADE CHOLANGIOPANCREATOGRAPHY
sphincterotomy: results in 394 consecutive patients from a single centre. Br J Surg 1985; 72: 215-9. 22. Mueller PR, Van Sonnenberg E, Ferrucci JT Jr. Percutaneous biliary drainage: technical and catheter-related problems in 200 procedures. AJR 1982; 138: 17-23. 23. Pessa ME, Hawkins IF, Vogel SB. The treatment of acute cholangitis percutaneous transhepatic biliary drainage before definitive therapy. Ann Surg 1987; 205: 389-92. 24. Cotton PB. ERCP. Gut 1977; 18: 316-41.
DISCUSSION Melody 0. Allen (Minneapolis, MN): Dr. Lai, do
you attribute any significance to the rate at which the contrast material drained from the obstructed duct at the end of the procedure? Do you feel that it is a predictor of later development of sepsis? We usually think it better if it does drain rapidly. Is that consistent with your experience? In the 2 1 patients who did develop clinical evidence of cholangitis, what were the organisms eventually cultured from the bile? Do you believe that they were introduced into the ductal system at the time of ERCP? I am particularly interested in the presence of Pseudomonas or Candida, which are indicators of contamination during ERCP. Frank Toups (Metairie, LA): I think gentamicin is a dangerous drug. Aminoglycoside can cause kidney damage, and several of the patients did die with sepsis. Dr. Lai, did you administer just one dose or several doses of gentamicin? Would you consider ticarcillin (Timentina), which is a safer, yet as effective, drug? It has essentially the same bacterial effect as gentamicin but is not as dangerous. John W. Braasch (Burlington, MA): Dr. Lai, should you be prepared for a papillotomy in a seriously ill patient
at the time of ERCP, which could obviate cholangitis and septic death? Ali Ghazi (New York, NY): Did you perform endoscopic papillotomy in those patients with biliary obstruction? Did you use a stent to drain the malignant obstruction and prevent surgery and cholangitis? Edward C.S. Lai (closing): Dr. Allen, with regard to the delayed ductal drainage of contrast medium after examination, we believe it is an important determinant in the prediction of subsequent sepsis as it relates to the degree of obstruction. Due to the retrospective nature of this study, detailed documentation of the drainage time was not available. As far as the results of bacteriologic studies in those patients who developed acute cholangitis, unfortunately, only 11 patients had such studies. Escherichia coli was the most common organism isolated among the four patients with positive cultures. The other major organism was Klebsiella. Dr. Toups, the use of gentamicin as a prophylactic antibiotic was obviously not effective. We have since changed to other third-generation cephalosporins. The use of antibiotics was reserved for only those patients with documented obstruction. Three doses of antibiotics were given intravenously as prophylaxis. Drs. Braasch and Ghazi, we are opposed to the use of papillotomy as a prophylactic measure, particularly in patients without any of the mentioned risk factors. The use of nonoperative drainage for malignant obstruction is also probably not advisable, since there is no clear evidence to suggest its definite role in the prevention of sepsis. In fact, according to different series on either papillotomy or other nonoperative measures, the complication rates reported were higher than our predicted risk of sepsis.
THE AMERICAN JOURNAL OF SURGERY
VOLUME 157 JANUARY
1989
125