Journal of Surgical Oncology 26:282-284 (1984)

Biliary and Duodenal Bypass for Carcinoma of the Head of the Pancreas

IAN R. COUGH, MBBS, MRCP, FRcs, FRACS, AND GEOFFREY MUMME, MBBS, FRACS

From the Department of Surgery, University of Queensland, Clinical Sciences Building, Royal Brisbane Hospital, Brisbane, Queensland, Australia

A review was undertaken of 66 patients having palliative bypass surgery for carcinoma of the head of the pancreas. A second laparotomy for relief of duodenal obstruction was necessary in six of 29 (20%) of the patients who did not have gastroenterostomy performed initially, but late gastric outlet obstruction occurred only once in 37 patients having an initial gastroenterostomy (P < 0.05). Cholecystojejunostomy failed to provide permanent biliary bypass in 14 of 53 (26%) patients, compared to success- ful drainage in all 13 patients having anastomosis of the common hepatic duct to the jejunum (P < 0.05). In seven patients cholecystojejunostomy was performed when the cystic duct was already obstructed by tumour. Therefore prophylactic gastrojejunostomy is recommended as a routine. The gallbladder should only be used for biliary bypass when appropriate contrast x-rays have demonstrated the patency of the cystic duct.

KEY WORDS: biliary, duodenal, bypass, surgery, carcinoma, pancreas

INTRODUCTION Carcinoma of the pancreas is a common condition and

the incidence is increasing in many countries [ 1,2]. Only a small proportion of patients are suitable for excisional surgery and the extent and role of resectional surgery is controversial. Shapiro [3] retrospectively compared the reported results of Whipple pancreaticoduodenectomy and palliative bypass surgery in patients who had a simi- lar extent of disease and concluded that there was no advantage to those treated with a Whipple resection. Therefore for reasons of surgical philosophy or technical unresectability, most patients having a laparotomy for carcinoma of the head of the pancreas will have a pallia- tive procedure.

The main decisions to be made at laparotomy concern: 1 ) establishment of a tissue diagnosis, 2) coeliac plexus block, 3) biliary bypass, and 4) duodenal bypass. In general the authors support the concepts of biopsy to establish the diagnosis with certainty, and of prophylactic coeliac plexus blockade; these issues have been reviewed elsewhere [4-61.

The aim of the present study was to define the strategy of biliary and duodenal bypass surgery in the palliative management of carcinoma of the head of the pancreas.

MATERIALS AND METHODS A retrospective review was conducted of 66 patients

having palliative bypass surgery for carcinoma of the head of the pancreas performed by surgeons at the Royal Brisbane Hospital between 1970 and mid-1983. The di- agnosis was established by operative biopsy (using var- ious techniques) in 32 patients, autopsy in three patients, and by typical operative findings and clinical course resulting in death in 31 patients. Only two patients, both of whom have biopsy-proven carcinoma, are still alive at 6 and 9 months.

These 66 patients represent approximately half the pa- tients with pancreatic carcinoma treated during the time period. Other patients who were managed without lapa- rotomy, or in whom the diagnosis was uncertain, or who had surgery other than palliative bypass surgery were excluded from analysis.

Accepted for publication December 13, 1983. Address reprint requests to I.R. Cough, MD, Department of Surgery, University of Queensland, Clinical Sciences Building, Royal Brisbane Hospital, Brisbane, Queensland 4029, Australia.

0 1984 Alan R. Liss, Inc.

Bypass for Carcinoma of the Pancreas 283

RESULTS TABLE 1. Survival According to Extent of Disease

There were 42 males and 24 females with a median Survival (months) age of 65 years (range 42 to 85). The extent of disease Extent Number Median Range found at laparotomy was the main determinant of survival Confined to head of pancreas

local spread (P < 0.05) and those with extensive local spread survived longer than those with distant metastases (P < 0.05) (Mann-Whitney statistic). The data for groups incidences. Only one case occurred in the present series treated with and without initial gastroenterostomy are and it was due to progression of the cancer. Therefore presented in Table II. The only differences in outcome vagotomy is not necessary, and most patients with gas- involved postoperative gastric outlet obstruction. Func- tric, duodenal, or anastomotic ulceration could probably tional delayed gastric emptying was observed in one be managed with histamine H2 receptor antagonists. patient who had an initial gastroenterostomy . Six patients Several studies [6,8- 101 have reported substantial in the nongastroenterostomy group required a second numbers of patients in whom cholecystojejunostomy was laparotomy for subsequent gastric outlet obstruction and unsatisfactory as a biliary bypass procedure. Failure to this difference was significant (Fisher’s exact test, P = relieve the jaundice initially occurs in approximately 10% 0.02). Vagotomy was not performed in any of the pa- of patients and recurrent jaundice and/or cholangitis in tients. Stomal ulceration was not observed. Only one up to 40%. In contrast, anastomosis of the jejunum to the patient had upper gastrointestinal haemorrhage in the late common hepatic duct has a failure rate approaching zero. postoperative period and this was shown endoscopically It is evident that although it may be technically easier to be due to tumour eroding into the duodenal lumen. to perform a cholecystojejunostomy than to use the com-

Biliary-enteric anastomosis using the gallbladder was mon hepatic duct, the surgeon must ensure that the cystic performed in 53 patients. In seven of these the obstruc- duct is patent and will remain patent for a reasonable tive jaundice was not relieved; four patients underwent length of time. A tense distended gallbladder may be due successful reoperation with anastomosis between the to cystic duct obstruction and does not necessarily indi- common hepatic duct and jejunum, one was managed cate communication with an obstructed bile duct. Even with percutaneous biliary drainage, and the other two though the cystic duct may be patent at the time of had tumours obstructing the cystic duct at autopsy. In surgery it may be occluded by advancing tumour in the another seven patients recurrent biliary obstruction and/ early postoperative period if the cystic duct entrance is or cholangitis developed before the patient died. close to the point of common bile duct obstruction. It is

In the 13 patients who had an initial anastomosis be- well-known that the cystic duct may have a segment tween the common hepatic duct and the jejunum and also within the wall of the bile duct and therefore even careful in the four patients converted to this procedure, biliary dissection may be misleading. Although advances have drainage was successful and permanent. When the failure been made with radioisotope biliary scanning in jaun- rate of initial biliary bypass operations is considered, 14 diced patients, the gallbladder can rarely be demonstrated of 53 cholecy stojejunostomies were significantly inferior in the presence of complete extra hepatic biliary obstruc- to 0 of 13 anastomoses between the common hepatic duct tion because of the high intraductal pressures [ 121. It is and jejunum (Fisher’s exact test, P = 0.03). Median therefore strongly recommended that the patency of the survival time was 5 months for both types of biliary cystic duct be demonstrated by preoperative or intraoper- bypass.

33 7 0- 17 25 4 0-1 1 time (Table I). Patients with disease confined to the head Extensive local spread

of the pancreas survived longer than those with extensive Distant metastases 8 1 0-9

DISCUSSION In the present study, 30% of all patients had duodenal

obstruction. Of the patients who did not have an initial gastroenterostomy, 20 % subsequently developed gastric outlet obstruction and required a second laparotomy . There was no additional morbidity or mortality associ- ated with the performance of prophylactic gastrojejunos- tomy. These data accord closely with several other recent series [6-lo].

Gastrointestinal bleeding after palliative surgery for carcinoma of the pancreas may occur in up to 20% of patients [ll] although most series have recorded lower

TABLE 11. Outcome of Gastroenterostomy vs. Nongastroenterostomy

Gastroenterostomy Nongastroenterostomy

Number 37 29 Preoperative gastric outlet

obstruction 12 1 Postoperative gastric outlet

obstruction 1 6 Wound dehiscence 4 3 Gut obstruction 4 4 Postoperative death (< 1 month) 2 4 Median survival (months) 5 5 Survival range (months) 0-17 0-17

284 Gough and Mumme

ative contrast x-rays. If the patency of the cystic duct cannot be established with certainty it is preferable to use the dilated common hepatic duct for the biliary enteric anastomosis.

Ten of 66 patients (15%) in the present series required reoperation (four biliary, six gastrojejunostomy) because of failure of the initial procedures. In another ten patients the choice of cholecystojejunostomy increased their mor- bidity. The patients with and without initial gastrojeju- nostomy and cholecystojejunostomy constitute groups whose treatment was selected by the operating surgeon; the statistically significant differences in outcome which have been demonstrated indicate that the selection of the procedures was frequently in error.

In these patients with limited life-expectancy, palliative surgery that is sufficient to permanently control symp- toms should be performed at the first laparotomy. The present study has demonstrated that this can be safely achieved by employing prophylactic gastroenterostomy, and that the gallbladder should not be used for biliary enteric anastomosis unless the surgeon is certain that the procedure will adequately and permanently drain the bil- iary system.

REFERENCES 1. Forrest JF, Longmire WP: Carcinoma of the pancreas and peri-

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

ampullary region: A study of 279 patients. Ann Surg 189:129- 138, 1979. Stewart RJ, Stewart AW, Stewart JM, Isbister WH: Cancer of the pancreas in New Zealand 1970-1974. Aust NZ J Surg 52:379- 384, 1982. Shapiro TM: Adenocarcinorna of the pancreas: A statistical anal- ysis of biliary bypass vs. Whipple resection in good risk patients. Ann Surg 182:715-721, 1975. Carter DC: Surgery for pancreatic cancer. Br Med J 1:744 746, 1980. Weiss SM, Skibber J, Dobelbower RR, Whittington R, Rosato FE: Operative pancreatic biopsy: Ten year review of accuracy and complications. Am Surg 48:214-216, 1982. Sarr MG, Cameron JL: Surgical management of unresectable carcinoma of the pancreas. Surgery 91:123-133, 1982. Sarr MG, Gladen HE, Beart RW, van Heerden JA: Role of gastroenterostomy in patients with unresectable carcinoma of the pancreas. Surg Gynecol Obstet 152:597-600, 1981. Blievernicht SW, Neifeld JP, Terz JJ, Lawrence W: The role of prophylactic gastrojejunostomy for unresectable periampullary carcinoma. Surg Gynecol Obstet 151 :794-796, 1980. Eastman MC, Kune GA: The objectives of palliative surgery in pancreas cancer: A retrospective study of 73 cases. Aust NZ J Surg 50:462-464, 1980. Thompson JF, Walker CJ: The management of pancreatic carci- noma: A review of 173 cases. Aust NZ J Surg 53:25-30, 1983. Meinke WB, Twomey PL, Guernsey JM, Frey CF, Farias LR, Higgins G, Keehn R: Gastrointestinal bleeding after operation for pancreatic cancer. Am J Surg 146:57-60, 1983. Rosenthall L: Cholescintigraphy in the presence of jaundice uti- lizing Tc-IDA. Semin Nucl Med 12:53-63, 1982.