Experience with the whipple procedure (pancreaticoduodenectomy) in a University-affiliated community hospital

Experience with the Whipple Procedure (Pancreaticoduodenectomy) in a University@ affiliated Community Hospital David K. W. Chew,

MD, Fadi F. Attiyeh,

BACKGROUND: The purpose of this report is to review the current standards of the Whipple pancreaticoduodenectomy and show that excellent results are achievable in a low-volume, univershy-affiliated community hospital. METHODS: A case series of consecutive patients operated on during the period November 1981 to June 1998 was evaluated retrospectively. Medical records were abstracted for demographic data, clinical presentation, comorbid factors, pathological diagnosis and staging, operative records, perioperative mortality, morbidity, and length of stay. Postoperative follow-up data were obtained from telephone interviews and from the primary referring physicians. RESULTS: A total of 29 patients underwent a pancreaticoduodenectomy procedure during this 15year period. Twenty-eight patients underwent the standard Whipple resection and 1 patient underwent an extended resection owing to the extent of the disease. The average age was 64 years (range 41 to 82). Comorbid diseases were present in 59% of cases. Jaundice was the main presenting complaint (62%), loss of weight and appetite was present in 34%. The most common indication for this procedure was malignant periampullary disease (83% of cases). Of patients with adenocarcinoma of the pancreas, 67% were stage I and 33% were stage Ill. The operation lasted an average of 5.5 hours (range 3.5 to 8 h). The mean operative blood loss was 1153 mL (range 250 to 4,000). The median length of stay was 11 days (range 7 to 81). There was 1 operative mortality (3%), and the overall major morbidity rate was 28%. Three patients required reoperation (IO%), 2 for intraabdominal hemorrhage and 1 for delayed gastric emptying. The major morbidity was hemorrhage at the gastrojejunostomy site (14%); 2 cases were intraabdominal and 2 were intraluminal. Pancreaticojejunostomy leak occurred in 1 patient, resulting in a localized intraabdominal abscess. Delayed gastric emptying, defined as

From the Department Of Surgery, Center, Columbia University College New York, New York. Requests for reprints should be MD, 1755 York Avenue, New York, Manuscript submitted October 2, form March 19, 1997.

the need for nasogastric suctioning for more than 10 days postoperatively, occurred in only 1 patient. Overall, an oral diet was tolerated after a median of 6 days (range 4 to 61). Seventy-two percent of patients had no major complications at all, 17% had one major complication, and 10% had two or more major complications. Pancreatic insufficiency was the major long-term complication, developing in about 50% of patients. There were no biliary strictures. The median survival for patients with carcinoma of the pancreas was 21 months and the 5-year survival was 15%. CONCLUSIONS: The above study demonstrates that a complicated procedure like the Whipple pancreaticoduodenectomy can be performed with excellent results in a community hospital. The most important prerequisite is thlat the surgeon be adequately trained in the procedure. In low-volume hospitals, the case load should be restricted to a minimal number of trained surgeons in order to concentrate the experience. Am J Surg. 1997;174:312-315. 0 1997 by Excerpta Medica, Inc. ancreatic cancer has been increasing in incidence over the last 4 decades, and an estimated 26,000 cases will be seen in 1996.’ Most of these patients will die P of the disease, and surgical resection is currently the only chance for cure. The first successful pancreaticoduodenectomy was performed by Kausch’ in 1912. It was not until 1935, when Whipple3 presented his series of 3 patients, that the procedure that now bears his name was popula.rized. Initial experience with this operation up to the 1970s was marred by a high mortality and morbidity rate, to the point that Crile4 and Shapiro’ advocated bypass over radical pancreaticoduodenectomy for pancreatic cancer. Fortunately, over the last 20 years, we have witnessed a dramatic decline in the perioperative mortality rate to less than 5%. An improved long-term survival for patients with periampullary tumors following the Whipple procedure has also been reported.h Two recent outcome studies have suggested a lower mortality rate as well as medical costs when this operation is performed at high-volume medical centers.“* We present our experience with this procedure over th.elast 15 years in a 700-bed university-affiliated community hospital.

St. Luke’sRoosevelt Hospital Of Physicians And Surgeons, addressed to Fadi NY 10128. 1996, and accepted

F. Attiyeh,

METHODS

in revised I

312

0 1997

by Excerpta

All rights

reserved.

Medica,

Inc.

MD, New York, New York

A retrospective who underwent

chart review was performed for all patients the Whipple procedure from November 9002-961 O/97/$1 7.00 PII SOOO2-961 0(97)00110-4

I

1981 to June 1996. A total of 29 cases were treated during this period. All the operations were performed by the second author and independently assessed by the first author. Medical records were abstracted for demographic data, clinical presentation, comorbid factors, pathological diagnosis and staging, operative records, perioperative mortality, morbidity, and length of stay. Long-term follow-up data were obtained from telephone interviews and from the primary referring physicians. This information was available for 26 out of the 29 cases. Only 1 patient with carcinoma of the pancreas was lost to follow-up at 12 months and was assumed dead of the disease. Tumor staging was according to the Union Intemationale Contre le Cancer (UICC) classification. Operative mortality was defined as death in the hospital following the operation. Morbidity was classified as major if an invasive intervention was required or it was potentially life-threatening; and minor if it was not. Delayed gastric emptying was defined as the need for nasogastric suctioning for more than 10 days postoperatively. A pancreatic fistula was defined as a persistently high output (>50 mL) of amylase-rich fluid from the peripancreatic drains after postoperative day 10 or a leak that was demonstrated roentgenographically. A biliary fistula was defined as the output of bile from the drains or a leak that was demonstrated on HIDA scan. The diagnosis of bile reflux gastritis was made endoscopically when symptomatic patients underwent an upper endoscopy. The standard Whipple resection was performed in 28 cases, and an extended resection (including a total pancreatectomy and splenectomy) was performed in 1 case owing to the extent of the disease. In 1 patient with Gardner’s syndrome, a left adrenalectomy for a coexistent adrenal tumor was performed at the same time. Truncal vagotomy was performed in 2 patients. Our standard resection comprised a 60% partial gastrectomy and pancreaticoduodenectomy, the pancreas being transected to the left of the portal vein. Reconstruction was performed via a pancreaticojejunostomy (end-to-end dunking type), a hepaticojejunostomy (end-to-side), and a gastrojejunostomy (either stapled or hand-sewn). The main pancreatic duct was ligated with a fine absorbable suture, eg, 00000 Dexon, to protect the anastomoses from the digestive effects of activated pancreatic enzymes in the early postoperative period. The jejunal limb was drained with a red Robinson catheter; an increased output of amylase-containing fluid around postoperative day 5 usually signified that the duct had reopened.

RESULTS During this 15year period, we treated a total of 29 patients. Of these, 15 were male and 14 were female. The average age was 64 years (range 41 to 82 years). Sixteen of the patients were white, 6 black, 6 Hispanic, and 1 Asian. Comorbid diseases were present in 59% of cases, with 2 patients having 2 or more such factors present (see Table I); and 17% of our patients also had prior upper abdominal or biliary tract surgery. A history of smoking was present in 38% of cases and 24% admitted ro regular alcohol consumption. The majority of the patients presented with jaundice (62%); loss of weight and appetite was present in 34%. Other symptoms included abdominal pain (24%) and gasTHE AMERICAN

TABLE

I Comorbid

Factors

Hypertension Diabetes mellitus Cardiac (coronary artery disease, aortic stenosis) Chronic obstructive pulmonary disease Peripheral vascular disease Obesity Prior history of cancer (chronic lymphocytic leukemia)

TABLE

5 5 2 1 1 1 1

II Pathological

Diagnosis

Adenocarcinoma pancreas Adenocarcinoma ampulla of Vater Adenocarcinoma duodenum Cholangiocarcinoma Cystadenoma pancreas Pancreatitis Villous adenoma duodenum Lymphoma

15 3 3 2 2 2 1 1

trointestinal hemorrhage (10%); 1 patient presented with acute pancreatitis. The distribution of the pathological diagnoses can be seen in Table II. As has been the experience of others, it is often difficult to be certain of the exact origin of the primary tumor based on gross examination of the lesion alone.Y Of patients with pancreatic cancer, 67% were stage I and 33% were stage III. The average number of lymph nodes sampled in each pathological specimen was 15. All of the operations were performed with curative intent, none for palliation only. However, of the 24 patients with periamFlullary malignancy, 3 (13%) had positive surgical margins at the remnant pancreas. The recent availability of dynamic, spiral computed tomography (CT) scans, and staging laparoscopy have improved the resectability rates by excluding patients with metastases and superior mesenteric arterial involvement. The operation lasted an average of 5.5 hours, with a median of 5 hours (range 3.5 to 8 h). The mean operative blood loss was 1153 mL, with a median of 1,000 .mL (range 250 to 4,000 mL). Postoperative length of stay averaged 19 days, with a median of 11 days (range 7 to 81). There was 1 operative mortality, resulting in an in-hospital mortality rate of 3%. This patient was elderly and debilitated before surgery, with a preoperative albumin level of 2.1 g/dL. He developed intraabdominal hemorrhage from the gastrojejunostomy site on postoperative day 2, requiring reoperation. He rebled from diffuse gastritis and finally succumbed to multiple organ dysfunction syndrome on posroperative day 18. The overall major morbidity rate was 28%. Three patients required reoperation (lo%), 2 for intraabdominal hemorrhage and 1 for delayed gastric emptying (Table III). The major morbidity has been hemorrhage at the gastrojejunostomy site (14%). In 2 cases, the bleeding was intraabdominal and in the other 2, intraluminal. There was no significant relationship to the technique of the anastomosis, ie, stapled versus hand-sewn. Both cases of gastrointestinal hemorrhage were controlled endoscopically, whereas both JOURNAL

OF

SURGERY@’

VOLUME

174

SEPTEMBER

1997

313

TABLE

III Morbidity Gastrointestinal lntraabdominal

After

Surgery

hemorrhage hemorrhage

lntraabdominal abscess Pancreaticojejunostomy

2 2 leak

2 1

Pancreatitis Delayed gastric emptying Segmental liver infarction

1 1 1

Reflux Colitis

32

esophagitis/gastritis

Wound infection Urinary tract infection Pneumonia Line sepsis

2= s lz 6 -2 :: ae

1 4 1 2

120

-

100

-

No. of Patients = 15 No. censored = 3

aoEO4020 o::, 0

20

r 40 Survival

cases of intraabdominal hemorrhage required reoperation. There was only 1 anastomotic leak at the pancreaticojejunostomy site resulting in an intraabdominal abscess. This was localized and was successfully drained via the peripancreatic drains placed at the time of the operation. There was no biliary fistula in this series. The single case of delayed gastric emptying occurred in a patient who had a truncal vagotomy. Upper endoscopy and barium studies had demonstrated stenosis of the anastomosis requiring revision. Despite revision, the patient continued to have gastroparesis, which did not respond to erythromycin and metoclopramide. Eventually, after prolonged nasogastric drainage, she was able to tolerate a soft diet. For the majority of the patients, however, an oral diet was tolerated after a median of 6 days (range 4 to 61). In our series, 21 (72%) patients had no major complications at all, 5 (17%) had one major complication, and 3 (10%) had two or more majorcomplications. As expected, such complications resulted in a prolonged length of stay. The major long-term complication was pancreatic insufficiency, developing in about 50% of our patients and responding to oral enzyme supplements. None of our patients had biliary strictures or significant dumping symptoms. The Kaplan-Meier survival curve for patients with carcinoma of the pancreas showed a median survival of 2 1 months and a 5-year survival of 15% (Figure). The small numbers in the other categories of periampullary malignancy negate any useful comments regarding their survivorshllp.

COMMENTS Pancreaticoduodenectomy is a formidable operation with many potential pitfalls for the inexperienced surgeon. Although several large series have reported operative mortalities between 0% and 9%, morbidity rates remain high at 35% to 50%.“-‘@ Fortunately, most of the complications can be managed nonoperatively. Our mortality rate of 3% and overall morbidity rate of 28% compare favorably with the current standards. The most common indication for this procedure remains malignant periampullary tumors, comprising 83% of cases. Delayed gastric emptying, a frequent complication in some series, has not been a problem in our patients. Miedema and associates” did not find truncal vagotomy to increase the incidence of this complication. We do not routinely include a truncal vagotomy in the Whipple 314

THE

AMERICAN

JOURNAL

OF SURGERY@

VOLUME

174

Figure.

Kaplan-Meier

.I:

60

80

After Surgery

survival

curve

100

120

(Months)

for Ca pancreas.

procedure. Several authors have also found a higher incidence of this problem with the pylorus-pres#erving pancreaticoduodenectomy.“,‘” Until the superiority of this operation has been proven by prospective, randomized trials, we believe that the standard Whipple should be the procedure of choice for malignant periampullary disease. A leak at the pancreaticojejunostomy s,ite is the most dreaded complication, with a reported incidence of between 5% and 20% and a leakage-related mortality rate of 40% to 50%. A recent meta-analysis of series on pancreatic resections revealed an overall fistula rate of 14% and an overall fistula-related mortality rate of 8%.20 Numerous methods of management of the pancreatic stump have been described. We have enjoyed success with the technique of ligating the pancreatic duct with a fine absorbable suture as described by Shiu” in 1982. This protects the anastomosis from the damaging effects of activated pancreatic enzymes in the early postoperative period. Our leak rate of 3% attests to its reliability. Ligation of the pancreatic duct did not seem to result in pancreatitis. The only case of postoperative pancreatitis occurred in the patient who also underwent a resection of a left adrenal tumor. l?erhaps the mobilization of the body and tail of the pancreas may have contributed to this complication. Finally, two recent outcome studies hav’? shown a lower mortality rate and reduced medical costs when the Whipple procedure was performed at high-volume medical centers.7.8 This study, however, shows that excellent results are achievable at a low-volume hospital. The important prerequisite is that the surgeon should be tmined in the operation. In hospitals that do not treat this condition frequently, it is best to restrict the case-load to a minimal number of surgeons in order to concentrate the experience.

REFERENCES 1. Parker SL, Tone: T, Bolden S, Wingo PA. Cancer statistics, 1996. Ca Cancer J Clin. 1996;46:5-27. _ 2. Kausch W. Das carcinom der papilla duodeni und seine radikale entfernung. Beirrage Zur Klinischen Chirurgie. 1912;78:439-486. 3. Whipple AO, Parsons WB, Mullins CR. Treatment of carcinoma of the ampulla of Vater. Ann Surg. 1935;102:76~3-779. 4. Crile G Jr. The advantages of bypass operations over radical pancreaticoduodenectomy in the treatment of pancreatic carcinoma. Surg Gynecol Obstet. 1970;130:1049-1053. SEPTEMBER

1997

1EXPERIENCE

WITH

WHIPPLE

PROCEDUREKHEW~

5. Shapiro TM. Adenocarcinoma of the pancreas: a statistical analysis of bypass vs Whipple resection in good risk patients. Ann Surg. 1975;182:715-721. 6. Crist DW, Cameron JL. The current status of the Whipple Operation for periampullary carcinoma. Adw~nces Surg. 1992;25:21-49. 7. Gordon TA, Burleyson GP, Tielsch JM, Cameron JL. The effects of regionalization on cost and outcome for one general high-risk surgical procedure. Ann Surg. 1995;221:43-49. 8. Lieberman MD, Kilburn H, Lindsey M, Brennan MF. Relation of perioperative deaths to hospital volume among patients underresection for malignancy. Ann Surg. going pancreatic

rus-preserving pancreaticoduodenectomy. ] Am Coil Surg. 1994;178:443-453. 14. Tsao ]I, Rossi RL, Lowell JA. Pyl orus-preserving pancreaticod, uodenectomy. Is it an adequate cancer operation? Arch Surg. 1994;129:405-411. 15. Hannoun L, Christophe M, Ribeiro J, et al. A report of fortyfour instances of pancreaticoduodenal resections in patients more than seventy years of age. Surg Gynecol Obster. 1993;177:556-560. 16. Zerbi A, Balzano G, Patuzzo R, et al. Comparison between py lorus-preserving and Whipple pancreaticoduodenectomy. Br J Surg.

1995;222:638-645. 9. Forrest JF, Longmire

W, Siewert JR. Pylclrus-preserving versus standard pancreaticoduodenectomy: an analysis of 110 pancreatic and periampullary carcinomas. BrJ Surg. 199.2;79:152-155. 18. Pate1 AC, Toyama MT, Kusske AM, et al. Pylorus-preserving Whipple resection for pancreatic cancer. Is it any better? Arch Surg. 1995;130:838-843. 19. Miedema BW, Sarr MG, van Heerden JA, et al. Complications following pancreaticoduodenectomy. Current management. Arch slug. 1992;127:945-950. 20. Sikora SS, Posner MC. Management of the pancreatic stump following pancreaticoduodenectomy. Br 1 Surg. 1995;82:1590-

WP. Cancer of the pancreas and periampullary region. A study of 279 patients. Ann Surg. 1979;189:129-138. 10. Cameron JL, Pitt HA, Yeo CJ, et al. One hundred and foftyfive consecutive pancreaticoduodenectomies without mortality. Ann Surg. 1993;217:430-438. 11. Castillo CF, Rattner DW, Warshaw AL. Standards for pancreatic resection in the 1990s. Arch Surg. 1995;130:295-300. 12. Klinkenbijl JHG, van der Schelling Cl’, Hop WCJ, et al. The advantages of pylorus-preserving pancreaticoduodenectomy in malignant disease of the pancreas and periampullary region. Ann Surg. 1992;216:142-145. 13. Kozuschek W, Reith HB, Waleczek H, et al. A comparison of long term results of the standard Whipple procedure and the pylo-

THE AMERICAN

1995;82:975-979. 17. Roder JD, Stein HJ, Huttl

1597. 2 1. Shiu MH. Resection tula. Surg Gpecol O&et.

JOURNAL

OF

SURGERY@

of the pancreas without 1982;154:497-500.

VOLUME

174

production

SEPTEMBEFI

of fis-

1997

315