Pancreatic Cancer Obstructive jaundice

Pancreatic cancer usually requires palliative rather than curative therapy. Palliative procedures should have low morbidity and mortality, provide a good quality of life, and necessitate minimal hospitalization. The surgeon, endoscopist, and radiologist all play a role in delivering effective palliation to properly selected patients.

Pancreatic Cancer is diagnosed in 28,000 people annually in the United States. Surgical resection offers the best hope for long-term survival, but unfortunately only 20% of cases at presentation are resectable for cure. Thus, most treatment efforts are directed at palliation of jaundice, gastric outlet obstruction, and pain.

Effective palliation should relieve symptoms, while causing little morbidity or mortality and limited hospitalization. Staging the disease at the time of diagnosis is useful in deciding the most appropriate strategy for palliation. Patients with locally advanced but unresectable pancreatic cancer have a median survival of 8.3 months, versus 2.5 months for patients with documented distant metastases.[ 1] Techniques for palliation vary. With an understanding of these techniques, the physician can choose the best method of palliation while maintaining the patient's quality of life.

JAUNDICE RELIEF

Obstructive jaundice occurs in 70% of all-patients with pancreatic cancer and up to 90% when the tumor involves the pancreatic head. Jaundice results from periampullary obstruction of the bile duct via direct invasion or compression, from portal adenopathy, or from extensive hepatic metastases. Palliation is recommended to avoid the sequelae of hepatic dysfunction, pruritus, cholangitis, and the psychologic impact of jaundice on the patient and family. Data from retrospective studies imply that survival is prolonged with biliary decompression. Thus, relief of biliary obstruction is recommended to avoid complications, increase performance status, and possibly prolong survival.

In the past, a biliary enteric anastomosis was routinely used, since all patients had an exploratory laparotomy for diagnosis of obstructive jaundice. Currently, nonoperative biliary decompression with biliary endoprostheses (stents) is being used increasingly because improved imaging studies allow both diagnosis and staging to be accomplished non-operatively, and patients with advanced disease and limited survival thus avoid laparotomy. Newer stents with large internal diameters provide effective nonoperative palliation for the lifetime of the patient. Stents can be inserted percutaneously or endoscopically.

The optimal strategy for decompression of obstructive jaundice depends on the projected survival of the patient. For example, patients with periampullary cancer can often have resection for cure. Patients with relatively early but noncurable disease can be treated either by palliative surgical bypass of the biliary system or with a biliary endoprosthesis that minimizes short-term morbidity and has few long-term complications and low mortality. Patients with extensive unresectable disease or with underlying medical illness contraindicating surgery should receive endoscopic or radiologic biliary decompression. Finally, patients whose advanced disease indicates short-term survival may be best managed by medical therapy for pruritus.

Although choledochoduodenostomy and cholecystoduodenostomy are effective operations for benign conditions obstructing the distal common bile duct, most surgeons avoid using the duodenum when palliating malignant periampullary biliary obstruction because of the potential for progressive luminal obstruction. Although the total combined experience with duodenal drainage is small when compared with series using jejunum, results nonetheless illustrate an overall operative mortality of 14% to 18% and a survival of 6 to 7 months, which compares favorably with choledochojejunostomy and cholecystojejunostomy."

The decision to perform a cholecystojejunostomy or choledochojejunostomy (Fig 1) is influenced as much by surgeon preference as by objective data. In theory, a cholecystojejunostomy is technically less demanding and requires less operative time than a choledochojejunostomy. However, initial success and durability of jaundice relief depend on cystic duct patency. Sarr and Cameron[ 4] reviewed the experience of 8,000 palliative procedures for pancreatic cancer and found little difference in operative mortality, survival, or jaundice relief between cholecystojejunostomy and choledochojejunostomy. Watanapa and Williamson" confirmed the similarities of operative mortality, survival, and initial jaundice relief afforded by these two procedures. However, they noted a higher rate of recurrent jaundice after cholecystojejunostomy than with choledochojejunostomy (20% vs 8%).

Sarfeh et al conducted a prospective randomized trial comparing cholecystojejunostomy and choledochojejunostomy in patients with malignant distal common bile duct obstruction. Choiecystojejunostomy required less operative time (2.6 vs 3.7 hours) and resulted in less operative blood loss (256 mL vs 600 mL) than choledochojejunostomy; however, both postoperative morbidity (53% vs 19%) and recurrent jaundice (46% vs 0%) were greater when cholecystojejunostomy was done. Lillemoe et al[ 6] recently reviewed the surgical experience at Johns Hopkins Medical Center for palliation of pancreatic cancer. Choledochojejunostomy was done in 82 of 118 patients (69%) having surgical biliary bypass. There was no difference in survival or recurrent jaundice between choledochojejunostomy and cholecystojejunostomy, though two patients with choledochojejunostomy had obstructive jaundice before death. These authors recommend choledochojejunostomy over cholecystojejunostomy unless the patient has portal hypertension or extensive collateralization after superior roesenteric vein or portal vein occlusion?

Although a Roux-en-Y reconstruction is the procedure of choice for benign common bile duct obstruction, it has limited usefulness in the palliation of malignant processes. A Roux-en-Y loop has been shown to reduce reflux of enteric contents into the biliary tree and subsequent cholangitis after biliary-enteric anastomosis. Singh et al suggested that Roux-en-Y reconstruction should be considered in patients with a more favorable diagnosis. Lillemoe et al[ 6] used a Roux-en-Y limb in 63 of 92 patients (68%), and cholangitis developed in only 3%. Thus, a Roux-en-Y choledochojejunostomy may be useful in a patient with an expected median survival of at least 6 months. Those who are expected to have a limited survival are adequately palliated with a loop cholecystojejunostomy or choledochojejunostomy.

In patients with incurable disease, bile duct endoprostheses (Fig 2) provide effective relief of jaundice without surgery. These devices are indicated in patients with unresectable disease or those who cannot undergo operation. Both surgical decompression of the biliary tree and the plastic biliary endoprosthesis have similar success rates for the relief of jaundice. However, insertion of an endoprosthesis requires a significantly shorter initial hospitalization and produces less mortality. Morbidity is similar for the two approaches, though a plastic stent is more likely to become occluded than a biliary enteric anastomosis, leading to repeat hospitalization and reduction in quality of life. For this reason most surgeons avoid the use of endoprostheses.

Several randomized studies have compared surgical biliary bypass and endoscopically placed stents. These data are summarized in the Table. Both methods equally relieve jaundice, but perioperative mortality is lower with the endoprosthesis. However, the reported surgical mortality rates in these series are high compared with today's reports of 2.5% to 10% postoperative mortality. Late complications are primarily related to stent occlusion, which requires repeat endoscopy for replacement of the stent. Some studies" note a lower overall hospitalization for endoscopically placed stents despite the higher readmission rate, while others[ 12] show that multiple hospitalizations negate the advantage.

Effective long-term stent drainage requires a stent caliber of at least 10F to 12F and a straight configuration for improved flow. Stents can be placed endoscopically or percutaneously. One randomized trial compared endoscopic and percutaneous stent insertion, using traditional plastic stents, for the palliation of malignant obstructive jaundice in non-operative candidates. Analysis of these 75 cases showed the endoscopic method had a significantly higher success rate for relief of jaundice (81% versus 61%) and a significantly lower 30-day mortality (15 % versus 33%). The higher mortality in the percutaneous group was due to complications of liver puncture (bleeding and bile leaks). Additionally, endoscopic procedures can often be accomplished in one session, while percutaneous procedures may take two or more sessions before they are completely internalized. Most authorities recommend an initial attempt at endoscopic palliation of malignant obstructive jaundice. If this is unsuccessful, then a percutaneous approach can be used. Local expertise may influence the choice of therapy.

Expandable metal mesh stents (ie, Wall-stent, Schneider) were initially used in the biliary system by interventional radiologists. The advantage of these stents is the much larger internal diameter (up to 10 mm when fully expanded). Thus, stent occlusion by bacterial biofilm (sludge) is uncommon. These stents can also be placed endoscopically. In a recent randomized trial using endoscopically placed stents, Davids et al[ 16] prospectively compared metal stents and plastic stents (at least 10F size) in patients with unresectable malignant distal bile duct obstruction. The median stent patency was significantly longer with metal stents (273 days versus 126 days). Stent occlusion in the metal stent group occurs mainly by tumor ingrowth through the mesh and can easily be managed by placing a plastic stent through the metal stent. Studies are already under way using plastic-coated metal stents that prevent tumor ingrowth. It is imperative that metal mesh stents be used only in patients who are not candidates for operation, since metal stents cannot be removed.

An additional factor to consider when deciding between surgical bypass and endoscopic stent is financial cost. One study[ 12] investigated the financial impact of the two methods. Despite higher rehospitalization rates, stents saved an estimated $8,000 per patient in 1987.

In summary, surgical biliary bypass and endoprostheses can equally reduce jaundice. Proper selection is necessary to avoid lapatotomy in patients with limited survival. Both surgical biliary enteric anastomosis and expandable metal stent placement provide excellent palliation in patients with obstructive jaundice and reasonable prognosis (expected survival of at least 6 months).
DUODENAL OBSTRUCTION

Nausea and vomiting are common symptoms in patients with pancreatic cancer, especially when the tumor is located in the pancreatic head. Although 30% to 40% of patients have these symptoms, only 5% have true mechanical obstruction. This suggests the presence of a functional delay in gastric emptying) In 10% to 20% of patients duodenal (gastric outlet) obstruction will develop if steps are not taken to prevent this complication.

The need for a prophylactic gastrojejunostomy to prevent gastric outlet obstruction is controversial. The addition of a gastrojejunostomy during surgical biliary bypass was initially believed to increase operative morbidity and mortality. In their review, however, Sarr and Cameron[ 4] found an operative mortality of 17% with biliary bypass alone and 18% when gastrojejunostomy was done at the time of biliary bypass. Watanapa and Williamson confirmed that the addition of a gastrojejunostomy to a biliary bypass did not increase the operative mortality. However, operative mortality is higher when gastrojejunostomy is done for gastric outlet obstruction after a biliary bypass has already been done.

The decision to do a prophylactic gastrojejunostomy is complicated because despite a widely patent anastomosis, delayed gastric emptying has been reported in 14% to 26% of patients postoperatively. Doberneck and Berndt noted that delayed gastric emptying was significantly more common in patients who had preoperative duodenal obstruction (57% versus 16%), suggesting that preexisting gastric atony contributes to delayed emptying during the postoperative period. Gastroparesis may also be due to tumor infiltration of retroperitoneal nerves, which results in denervation of the stomach. However, the cause is not clean Lillemoe et al[ 6] reported an 8% incidence of postoperative delayed gastric emptying, which may be explained by their use of retrocolic gastrojejunostomy as well as metoclopramide.

Finally, the overall usefulness of gastric bypass for symptom relief has been questioned. Although 10% to 20% of patients with cancer of the head of the pancreas will have gastric outlet obstruction, the percentage of patients who die with symptoms of nausea and vomiting after undergoing duodenal bypass is not well documented in surgical series. Weaver et al attempted to analyze the palliative effects of gastrojejunostomy in 81 patients with pancreatic carcinoma. The outcome of the procedure was classified as poor if the patient died within 30 days and good if the patient was alive and eating a regular diet. The authors concluded that only 14% of patients derived a benefit, but 81% of this group (9/11) had no preoperative symptoms of nausea and vomiting. This emphasizes the difficulty in determining the amount of gastric dysfunction as a result of the gastrojejunostomy itself and implies that delayed gastric emptying is seen in the natural history of pancreatic cancer.

Practical considerations for gastrojejunostomy are ( 1) its position related to the transverse colon, ( 2) use of a loop or Roux-en-Y limb, and ( 3) the need for vagotomy. Traditionally, the jejunum has been positioned anterior to the transverse colon to avoid the potential of late obstruction by tumor infiltration of the transverse mesocolon. Most studies do not report the position of the anastomosis, but recommend an antecolic position. However, Lillemoe et al[ 6] reported the cases of 118 patients having 107 gastrojejunostomies, 79% of which were retro-colic. The incidence of delayed gastric emptying was 8% overall--17% (4/23) for antecolic and 6% (5/84) for retrocolic procedures. Late readmission for gastric outlet obstruction occurred in four patients with gastrojejunostomy, equally distributed between antecolic and retrocolic reconstructions. Although the numbers did not reach statistical significance, this study showed a trend toward better results with a retrocolic location.

Most surgeons favor a loop gastrojejunostomy over a Roux-en-Y for its ease of formation and because it allows alkaline juice to bathe the gastrojejunal anastomosis. The Roux-en-Y limb is advocated to avoid bile reflux gastritis. However, Potts et al found that clinical bile reflux did not occur. Additionally, no difference in delayed gastric emptying or late gastric outlet obstruction was seen between loop and Roux-en-Y gastrojejunostomies. Thus, there is no reason to do a more involved reconstruction.

Vagotomy is infrequently used when a palliative procedure is being done for gastric outlet obstruction, since the incidence of stomal ulceration is less than 20% and highly effective acid-reducing agents are available. In addition, a vagotomy adds operative time and may add to delayed gastric emptying.

In summary, most surgeons favor a palliative gastrojejunostomy because it probably reduces the risk of late gastric outlet obstruction without increasing morbidity and mortality. Delayed gastric emptying does occur in 10% to 20% of patients after gastrojejunostomy, but usually resolves. Retrocolic location may have little long-term effect, but a recent single-institutional report has shown a trend toward better results with the retrocolic position. Finally, since there is no advantage for the Roux-en-Y limb in the prevention of bile reflux gastritis or delayed gastric emptying, a loop gastrojejunostomy is the preferred method.
PAIN

Pain may be the most incapacitating and dehumanizing symptom of pancreatic cancer, especially those tumors involving the body or tail of the pancreas. The pain usually begins as a dull ache in the upper abdomen, which is frequently episodic and aggravated by eating. As the disease progresses, a severe constant pain develops which penetrates into the patient's back and is extremely difficult to control. Some patients may experience pain relief after relief of obstructive jaundice, but eventually 50% to 90% of patients with pancreatic cancer will have incapacitating pain.

Pain relief can be provided with biliary or pancreatic decompression, narcotics, chemical or surgical splanchnicectomy, or radiation therapy. Decompression of an obstructed biliary or pancreatic duct seldom provides long lasting relief. Narcotics may be helpful for a time, but escalating the dose as well as the potency of the narcotic is usually necessary.

Chemical neurolysis is an attractive alternative to a large dose of narcotics. This is usually accomplished at the time of laparotomy by injecting 30 to 50 mL of 50% ethanol or 6% phenol on either side of the aorta at the celiac axis. Percutaneous fluoroscopic injections can also be carried out around the celiac axis when a laparotomy has not been done.[ Retrospective data on the effectiveness of chemical splanchnicectomy suggest that up to 81% of patients may have initial pain control and 71%persistent pain control.[ 21] Recent randomized prospective evaluation of intraoperatire chemical neurolysis has shown significant retraction and prevention of pain in patients with unresectable pancreatic carcinoma.

Another encouraging method of pain control is transthoracic splanchnicectomy. Reports for both benign and malignant pancreatic disease are favorable, especially since the denervation can be accomplished with a minimally invasive thoscopic technique.

External beam radiotherapy is another potential method of pain control. A beneficial effect is reported in one third to three fourths of patients. However, the onset of pain relief may be delayed for several weeks, and few patients receive total pain relief.
CONCLUSION

Palliation is the most common treatment for pancreatic cancer. It should cause little morbidity and mortality while providing a high quality of life. The method of palliation can be tailored to the patient's estimated survival. Those with distant metastases or locally advanced primary tumors are best managed nonoperatively. Patients found to have unresectable disease while undergoing exploration benefit from biliary bypass, as well as possible gastric bypass. Pain control can be provided by a number of methods. Surgical or chemical neurolysis are the most durable, and radiation may also have a role.

Last updated Jan 2/07