Chapter 143: Principles of Palliative Care for Patients with Pancreatic Adenocarcinoma

Palliative care is an approach to care that improves the quality of life of patients and their families facing the problems associated with life-threatening illness, through the prevention and relief of suffering by means of early identification and assessment and treatment of pain and other problems—physical, psychosocial, and spiritual.¹ As with many health care disciplines, there are both medical and surgical components to palliative care. Palliative medical care employs expertise on the assessment and management of problems including pain, cachexia, delirium, fatigue, dyspnea, and end-of-life issues. Although palliative care has long been demonstrated to improve symptom control and quality of life, recent evidence suggests it may also be associated with longer survival.^2,3 Therefore, the American Society of Clinical Oncology has recently issued a statement suggesting that palliative care should be considered early in patients with metastatic disease and/or high symptom burden.⁴ Palliative surgery is defined as surgery for which the major intent is alleviation of symptoms and improving patient's quality of life.⁵

The need for palliative surgery in cancer patients is frequent. Palliative surgery comprises approximately 20% of surgical practice and over 1000 procedures per year at major cancer centers.^6,7 In addition, almost half of all inpatient consultations at cancer centers meet the criteria for palliative care.⁸ Although formal training in palliative surgery is infrequently provided during surgical residency, palliative surgeries are among the most high-risk surgical procedures performed and require some of the most time-intensive preoperative discussions. Surgical decision-making and risk assessment are critical as morbidity and mortality rates for palliative surgical procedures are commonly estimated at 30% and 10%, respectively. The risk–benefit ratio is often narrow and the considerable risks of surgery must be balanced against the prognosis and estimated remaining length of life.

Pancreatic cancer represents the 12th most common cancer type but the 4th most common cause of cancer death in the United States.⁹ An individual's lifetime risk of developing pancreatic cancer is approximately 1.5%. The majority of patients with pancreatic cancer are diagnosed with unresectable or metastatic disease at presentation, and the 5-year survival rate of pancreatic cancer is only 6%.¹⁰ In addition, curative-intent pancreatic resection only offers an expected long-term survival rate of approximately 20%.⁹ In light of the dismal rate of cure for pancreatic cancer surgery, a nihilistic approach would be to consider all pancreatic cancer surgery primarily palliative. However, as defined above, palliative surgery is defined as surgery in patients with advanced or incurable malignancy with the intent of improving quality of life or symptoms. Although improved quality of life may lead to improved length of life, all surgeries performed with an intent to cure, including for patients with resectable pancreatic cancer, are by definition excluded from classification as palliative surgery.

The more complex concept of preemptive palliation must also be addressed in discussions of palliative surgery for pancreatic cancer, as interventions to prevent subsequent potential symptoms have also been labeled as palliative surgery.¹¹ This issue has particular relevance to pancreatic cancer due to considerations for the performance of, for example, prophylactic gastrojejunostomy or celiac-plexus block during noncurative pancreatic cancer surgery, sometimes even in the absence of current symptoms of gastric outlet obstruction or pain, respectively. Although surgical interventions performed for possible symptoms that may develop later in the disease course can be controversial, we believe they are relevant to palliative care of patients with pancreatic adenocarcinoma and therefore include discussion of the relevant surgical techniques and research surrounding this issue.

The most common indications for surgical palliation of pancreatic cancer are obstructive conditions (gastric outlet obstruction and obstructive jaundice), clearly predicted on the location and adjacent anatomy of the pancreas. The pancreatic head, lodged within the curve of the duodenum, is the most common site of adenocarcinoma. The bile duct extends through the head of the pancreas, joins with the pancreatic duct, and terminates in the descending portion of the duodenum. Another common indication for palliative consultation, pain, can also be seen as an anatomic consequence of the anatomically adjacent celiac ganglion in combination with the propensity of pancreatic cancer for perineural invasion. The pancreas is well vascularized through the superior pancreaticoduodenal branches of the gastroduodenal artery and the inferior pancreaticoduodenal branches of the superior mesenteric artery. Therefore, gastrointestinal hemorrhage secondary to pancreatic tumors, including pancreatic cancer, can cause upper gastrointestinal hemorrhage which can be difficult to control. Large bowel obstruction, a rare palliative scenario with pancreatic cancer, can develop either through the association of the anterior surface of the pancreas with the transverse mesocolon or rarely from peritoneal carcinomatosis. Remaining palliative situations for which surgical consultation is requested include malignant ascites and malignant small-bowel obstruction, both a function of the aggressive metastatic potential of pancreatic cancer and its propensity to spread as carcinomatosis.

There is no clear data on the frequency of surgical palliation in pancreatic data, although there are older reports in the literature that up to 50% of patients may require surgical palliation.¹² Approximately 15% to 20% of patients with pancreatic adenocarcinoma develop gastric outlet obstruction. The reported rates of the need for palliative intervention for obstructive jaundice in patients with pancreatic cancer vary, but 30% to 60% of patients with unresectable or metastatic disease present with jaundice, depending on the anatomic site of the tumor.¹³ Pain is the most common symptom at presentation and reported in 80% of patients with locally advanced unresectable disease, and in 85% of patients with metastatic disease. Establishing an accurate estimate of prognosis is an essential part of the risk–benefit analysis in determining the advisability of palliative surgical intervention. The median survival rate for patients with metastatic disease is approximately 3 months, and it is as little as 7 months for patients with locally advanced unresectable disease.¹⁴ These estimates of median survival provide important background in determining the potential benefit of palliative surgery; with these estimates in mind, we can proceed with analysis of data supporting various palliative interventions. Also of note in reviewing the indications for palliation of the clinical scenarios of obstructive jaundice, duodenal obstruction, and celiac plexus–related pain is that palliative care decision making falls primarily into two categories: palliation in patients who are clearly not candidates for resection and palliation in patients who are deemed unresectable at exploration.

Medical Palliation

As there are effective surgical and endoscopic methods to alleviate obstructive jaundice, medical palliation is typically only required as a temporizing measure prior to definitive palliation. Topical agents are generally ineffective for pruritus secondary to obstructive jaundice. Antihistamines are often the initial treatment modality and although there is little supportive evidence, the sedating effects of H1 antihistamines are often beneficial for nocturnal pruritus. Mirtazapine, a selective norepinephrine reuptake inhibitor with antihistaminic properties, as well as opioid receptor antagonists have been associated with improvement in pruritus secondary to malignant cholestasis. Cholestyramine, a bile acid sequestrant, is also occasionally administered, but is of minimal benefit as the duct obstruction results in little bile entering the gastrointestinal tract for binding and subsequent excretion.

Procedural Palliation

Nonoperative palliation of obstructive jaundice can involve either endoscopic biliary stent placement, typically by gastroenterology specialists or surgeons with advanced endoscopic skills, or percutaneous transhepatic biliary drainage catheter placement by interventional radiology. As percutaneous transhepatic biliary drainage is more invasive and painful, and has a higher complication rate, this procedure is reserved for clinical situations in which endoscopic stent placement is not an option. Malignant strictures of the mid to lower bile ducts are usually amenable to endoscopic drainage, while strictures involving the hilum can be more challenging and may require a percutaneous approach.

ENDOSCOPIC BILIARY STENT PLACEMENT

There are several randomized controlled trials comparing plastic endobiliary stents to surgical biliary bypass in patients with pancreatic cancer requiring palliation of obstructive jaundice.¹⁵ The risk of complications was higher in the surgical group but this was offset by a higher rate of recurrent biliary obstruction in the patients undergoing endoscopic plastic stent placement. These trials were included in a recent meta-analysis demonstrating a 40% decrease in complications but 19-fold increased risk of recurrent biliary obstruction with stent placement.¹⁵ Table 143-1 contains the prospective randomized trials comparing the two approaches with the corresponding outcomes. In addition, there are also several randomized controlled trials comparing self-expanding metallic stents (SEMS) to plastic stents, and a recent meta-analysis of these trials demonstrated that the risk of recurrent obstruction with SEMS was approximately half that of plastic stents.^15,16 The obvious implication of these studies is that SEMS would likely improve upon the high risk of recurrent obstruction when compared to surgical bypass. A single randomized controlled trial of surgery compared to SEMS has been performed for patients with metastatic pancreatic cancer. Although limited by size (N = 30) and not inclusive of patients with locally advanced disease, the study demonstrated improved cost and quality-of-life scores at 60 days for patients undergoing endoscopic drainage.¹⁷

Various recommendations for stent type have been suggested based on prognostication. As an example, the European Society of Gastrointestinal Endoscopy recommends that if patient life expectancy is shorter than 4 months, initial insertion of a plastic stent is most cost-effective, but if expected survival is longer than 4 months then initial insertion of a SEMS is more cost-effective.¹⁸ Importantly, self-expanding metallic stents do not necessarily preclude subsequent pancreaticoduodenectomy in patients ultimately considered candidates for resection. Although it is always prudent to involve the surgeon in discussions prior to endobiliary stent placement, a short-segment stent can typically be removed at the time of surgery and is often within the field of resection. There is also some debate as to the application of covered versus uncovered SEMS. The current standard of care for endoscopic stenting favors placement of uncovered SEMS, as randomized trials indicate that covered stents are associated with higher rates of migration and other adverse events without an associated increase in stent patency.¹⁹

PERCUTANEOUS TRANSHEPATIC BILIARY CATHETER PLACEMENT

Percutaneous transhepatic biliary drainage is most often performed by an interventional radiologist under conscious sedation with local anesthetic administration. Unilateral drainage is often sufficient to provide symptomatic palliation of obstructive jaundice. Technical decision making by the radiologist can include approaching the side with less disease or away from an atrophic hemi-liver. The right-sided approach may also benefit from an easier angle of catheter placement and avoidance of the fluoroscopic field, while the left-sided approach may be less painful and easier for the patient to manage than a right intercostal catheter. Technical details involve first confirming access to the bile ducts through aspiration of bile or through injection of contrast and performance of a percutaneous transhepatic cholangiogram. External drainage is performed when a catheter is placed above the site of obstruction and connected to gravity drainage. External drainage alone is not ideal as patients are prone to infections, dehydration, and electrolyte abnormalities. External–internal drainage is performed when a catheter is placed through the obstruction, and allows for at least some drainage of bile into the duodenum. Although external loss of bile can be improved through this approach, patients are still prone to developing cholangitis. Strictures that do not allow immediate external–internal drainage may require a period of external drainage and a second procedure to allow passage through the stricture. Finally, conversion to internal drainage is completed when an internal stent is placed through the stricture, allowing bile drainage into the duodenum, without the need for an external catheter. Another option when an internal–external catheter is in place is to perform a combined transhepatic and endoscopic procedure, also known as a rendezvous procedure. The rendezvous procedure allows for the endoscopist to place an endoscopic stent via a wire passed through the transhepatic catheter, which may allow for decreased risk of bleeding and bile leakage compared to transhepatic dilation for stent placement. Postprocedural care for patients undergoing transhepatic drainage procedures is important for surgeons, as patients undergoing interventional procedures are often placed on surgical inpatient services. Patients should be hospitalized for at least 24 hours following transhepatic drainage to monitor for complications, and catheters should be flushed with sterile saline every 12 to 24 hours.²⁰ Patients should be well educated on the signs and symptoms of cholangitis including pain, fever, nausea/vomiting, and malaise, and should be informed that pericatheter leakage can indicate catheter occlusion.²⁰ Complications of transhepatic catheter placement can include hemobilia, cholangitis, pancreatitis, pericatheter leakage, pneumothorax, and cholecystitis due to blockage of the cystic duct. Technical and drainage success rates are routinely reported as greater than 80% and mortality rates vary widely, depending on the indication for placement and patient functional status and comorbidities.

Surgical Palliation

Only a minority of patients with obstructive jaundice require surgical palliation. Potential surgical candidates include patients for whom an endoscopic approach is not possible, patients who live remote from endoscopic and interventional expertise, and patients with concomitant duodenal obstruction. Consideration should also incorporate prognosis and performance status. Local expertise can factor heavily into the risk–benefit analysis, as patients who fail an endoscopic attempt at palliation must be considered for an interventional percutaneous procedure that may require multiple procedures to drain both sides of the liver without a guarantee that the drain can be internalized. The debate regarding the optimal approach to palliation remains active as demonstrated by a recent meta-analysis of randomized trials revealing no significant difference in major complications or death between biliary stent placement and surgical bypass, albeit with significantly lower rates of recurrent biliary obstruction in the surgical bypass group.²¹ Regardless, the most frequent current indication for surgical palliation of biliary obstruction is the patient identified with unresectable disease at laparotomy. Roux-en-Y hepaticojejunostomy/choledochojejunostomy is the most widely accepted surgical technique, while cholecystojejunostomy and choledochoduodenostomy are infrequently performed due to concerns over ineffective palliation and recurrent jaundice with tumor progression. After cholecystectomy has been performed, the common bile/hepatic duct is dissected and transected. A 60-cm Roux-en-Y limb is created and an end to side hepaticojejunostomy is performed in an interrupted fashion with 4-0 absorbable suture. Other suitable options include a loop hepaticojejunostomy rather than a roux limb. Also, some authors elect to perform a side-to-side hepaticojejunostomy without dividing the bile duct. A loop gastrojejunostomy can be added to either approach if there are coexisting concerns regarding the presence or development of duodenal obstruction.

Cholecystojejunostomy is frequently described in laparoscopic approaches to biliary bypass but is infrequently performed as an open operation due to the requirement for cystic duct patency. Although endoscopic studies of the hepatocystic junction suggest only half of patients with obstructive jaundice will be candidates based on cystic duct patency, recent case reports on laparoscopic cholecystojejunostomy have reported low rates of recurrent biliary obstruction.^22,23 Perhaps the best evidence regarding this issue can be found from a study that utilized Surveillance, Epidemiology, and End Results (SEER) Medicare claims data and demonstrated that patients treated with a gallbladder bypass had a subsequent biliary intervention rate at 1 year of 7.5%, compared to 2.9% for patients treated with a bile duct bypass.²⁴ The discrepancy in the rates of recurrent jaundice between the reports of open gallbladder bypass and newer reports of laparoscopic gallbladder bypass will need to be addressed prior to considering the laparoscopic approach a valid option. Similarly, choledochoduodenostomy is not routinely utilized for palliation due to concerns regarding recurrent jaundice with tumor progression. There are few studies comparing this approach to other forms of biliary bypass; however, small series suggest a low rate of recurrence and cholangitis.^25,26

Medical

Nausea and emesis in advanced cancer patients are due to many factors beyond mechanical obstruction, including opioid and other medication side effects, metabolic/electrolyte abnormalities, autonomic failure, peptic ulcer disease, constipation, and increased intracranial pressure. Gastroparesis can be differentiated from mechanical gastric outlet obstruction relatively easily with endoscopy or radiologic tests, but it is important to note that gastroparesis and mechanical obstruction can coexist.

Medical management of gastric outlet obstruction includes bowel rest with parenteral hydration and often decompression with a nasogastric tube. Nasogastric tube use beyond a few days is not ideal due to nasopharyngeal discomfort, patient immobilization, ulceration of the nares, and gastric and esophageal erosions. Pharmacological measures often include opioids administered intravenously, subcutaneously, or transdermally as 90% of patients also experience continuous abdominal pain. Anticholinergic agents such as scopolamine (transdermal) may help symptoms. Antiemetics include antihistamines (promethazine), antidopamine antagonists (metoclopramide—contraindicated in complete obstruction), and serotonin antagonists such as ondansetron. Corticosteroids may decrease gut wall edema and also have a central antiemetic effect. Somatostatin analogues such as octreotide inhibit the release of several GI secretions and hormones and are therefore useful in reducing emesis and abdominal pain.

As there are effective procedural and surgical options, patients chosen for solely medical management often have prohibitive functional status or advanced-stage disease. In such situations, it is often helpful to attempt to classify patients into clinical scenarios of complete versus partial obstruction. Patients with obstruction that is deemed partial can be treated with metoclopramide to improve gastric emptying and obtain a central mechanism effect to block dopamine. Adding dexamethasone will also increase the central effect on nausea, but there is debate on the efficacy of dexamethasone for decreasing peritumoral edema and a lack of support from randomized clinical trials. For patients with complete obstruction, a good medical treatment option is to utilize a combination of an anticholinergic medication that does not cross the blood–brain barrier with resultant central nervous system side effects, such as glycopyrrolate, and an antisecretory agent such as octreotide 100 μg subcutaneous three times per day.

Procedural

Procedural interventions can include stent placement or, in cases of advanced disease, venting gastrostomy tube placement. Gastric and duodenal stents are similar to biliary SEMS not only in technical aspects of placement but also in terms of having lower immediate complications with significantly less durability. As with biliary stents, uncovered stents also have lower rates of migration. Technical success rates, simply defined as the ability to deploy the stent, are consistently greater than 90% while clinical success rates, defined as the ability to tolerate oral intake or improvement in nausea/vomiting, are reported from 84% to 93%.^27,28 During follow-up approximately 30% to 40% of patients developed a recurrent inability to tolerate oral intake. As randomized controlled trials are challenging in palliative clinical scenarios, there is a paucity of high-quality evidence to guide decision making. A small, multicenter randomized trial utilizing a gastric outlet obstruction scoring system reported an improvement in food intake of 3 days for patients undergoing stent placement compared to gastrojejunostomy.²⁹ However, more major complications, recurrent obstructive symptoms, and need for reinterventions were observed after stent placement. The authors concluded that gastrojejunostomy should be considered the preferred treatment in patients with a life expectancy of 2 months or longer. Systematic reviews of randomized trials, comparative studies, and case series propose similar recommendations in favor of SEMS for patients with a relatively short life expectancy and surgery for patients with anticipated longer survival.³⁰ The prospective randomized trials comparing gastroduodenal stent placement and surgical bypass are outlined in Table 143-2.

Venting gastrostomy tubes may be placed via endoscopic or interventional radiologic methods. Gastrostomy tubes are typically utilized late in the course of disease for symptoms refractory to other measures. Median survival rates after gastrostomy tube placement are 13 to 17 days.^31,32 Although gastrostomy tubes fail to palliate through restoration of a functioning gastrointestinal tract, they have proven useful in improving nausea, allowing for removal of the nasogastric tube, allowing for liquid intake for comfort, and discharge home.^31-33 Controversy exists as to whether endoscopic or percutaneous radiologically guided placement is safer; review studies suggest endoscopic placement may have fewer early complications.³⁴ Complications for both approaches include leakage, peristomal infection, tube obstruction, and migration. Although placement in patients with ascites is technically feasible, successful placement of a gastrostomy tube in a patient with ascites may require simultaneous placement of an intraperitoneal catheter or paracentesis, and major complication rates of approximately 10% have been reported in patients with malignant bowel obstruction and ascites.³⁵

Surgical

The most common surgical approach to gastric outlet obstruction is gastrojejunostomy. Appropriate surgical candidates without prohibitive comorbidities or diminished performance status, patients with concomitant obstructive jaundice, and patients with an anticipated life expectancy beyond a few months should be considered for gastrojejunostomy. Similar to obstructive jaundice, a common indication for palliative surgical bypass is the finding of unresectable disease at laparotomy. The concept of preemptive palliation for gastric outlet obstruction has been addressed in randomized clinical trials of patients without symptoms of gastric outlet obstruction undergoing exploration for attempted curative resection with the finding of unresectable disease. In a study from Johns Hopkins Medical Center, 87 patients were identified by the operating surgeon not to be at significant risk for duodenal obstruction and were randomized to prophylactic gastrojejunostomy or no gastrojejunostomy.³⁶ Postoperative complications were similar between both groups. None of the patients in the prophylactic gastrojejunostomy group developed gastric outlet obstruction, while 19% of patients without gastrojejunostomy developed obstruction and required therapeutic intervention. In a similar study from the Netherlands, randomization for double bypass (biliary bypass and gastrojejunostomy) and single bypass (biliary bypass) was performed for 65 patients found to be unresectable during exploration for periampullary carcinoma.³⁷ Although quality-of-life scores did not differ between groups, likely due to similar rates of disease progression, clinical symptoms of gastric outlet obstruction developed in 6% of patients with a double bypass and 41% of patients with only a biliary bypass.

Many surgeon-dependent preferences exist for the technique of gastrojejunal anastomosis. Options include Roux-en-Y versus loop construction, stapled versus handsewn, antecolic versus retrocolic, isoperistaltic versus nonisoperistaltic, and multiple different versions of handsewn anastomoses. Our own intrainstitutional preference for double bypass is to perform a loop gastrojejunostomy proximal to the jejunojejunostomy of a Roux-en-Y hepaticojejunostomy as illustrated in Figure 143-1.

Medical

Pain in cancer patients is a multifactorial syndrome including tumor-related causes (70% to 75%), treatment-related causes (20% to 25%), and causes unrelated to cancer or its treatment such as chronic preexisting pain (5%).³⁸ Pancreatic cancer frequently displays perineural invasion with resulting painful neuropathic spread of tumor, often to the celiac axis. In addition, pancreatic cancer often obstructs the pancreatic duct with resulting pancreatic fibrosis and pancreatitis. The majority of cancer pain can be controlled with commonly accepted medication regimens, although pancreatic cancer–related pain can be particularly challenging. Opioids are often indicated in initial treatment and accepted initial regimens include (1) morphine 5 mg orally every 4 hours and 2.5 to 5 mg orally every 2 hours as needed, (2) hydromorphone 1 mg oral every 4 hours and 0.5 to 1 mg oral every 2 hours as needed, and (3) oxycodone 5 mg oral every 4 hours and 2.5 to 5 mg every 2 hours as needed.³⁸ Antiemetics and laxatives should also be added, such as metoclopramide 10 mg every 4 to 6 hours as needed for nausea and sennoside 1 to 2 tablets twice a day with other laxatives such as polyethylene glycol or lactulose as needed. Constipation can be a considerable cause of increased pain due to dilated bowel and gas retention. Because of large interpersonal variation in pain medication and laxative requirements, frequent follow-up and personal titration should be performed every few days during initial treatment. In addition, general principles for cancer-related pain management are provided in the National Comprehensive Cancer Network Guidelines for Adult Cancer Pain (available with registration at www.NCCN.org).

Although most studies of systemic chemotherapy and radiotherapy utilize survival as the primary endpoint, many studies of combination therapy have also demonstrated improvements in quality of life. Palliative systemic chemotherapy or chemoradiation has been associated with improvement in pain scores and reduction in analgesic use. Gemcitabine-based chemoradiotherapy regimens have demonstrated improved pain control over 5-fluorouracil regimens in patients with pancreatic adenocarcinoma, and hypofractionated-accelerated radiotherapy can provide pain relief as well.^39,40

Procedural

Perineural and celiac invasions from pancreatic cancer suggest that celiac plexus neurolysis can be an effective strategy to help manage pancreatic cancer–associated abdominal pain. Randomized clinical trials have demonstrated improvement in pain, with a decrease in both opioid consumption and side effects, following percutaneous celiac plexus neurolysis; evidence in support of endoscopic injection is based primarily on case series that demonstrate an improvement in pain.⁴¹ There are no high-quality comparative studies of the two approaches.

Although a previous Cochrane Review demonstrated minimal evidence for the superiority of celiac plexus block pain relief over analgesic therapy in a review of studies that included surgical approaches, a more recent systematic review of percutaneous and endoscopic plexus block procedures documented greater improvements in pain with plexus block.^41,42 Percutaneous celiac plexus block is performed by either an anterior or posterior approach with assistance from computed tomography, ultrasound, or fluoroscopy. Endoscopic injection is performed with endoscopic ultrasound guidance via a transgastric approach. Possible side effects with either approach include diarrhea (usually limited to 1 to 3 days), hypotension, and a transient increase in pain.

Surgical

Due to improvements in oral analgesia as well as percutaneous and endoscopic celiac plexus block injections, surgical chemical neurolysis is usually reserved for the finding of unresectable disease at laparotomy. A prospective randomized clinical trial from Johns Hopkins Medical Center has provided evidence supporting the indication for celiac block in such situations.⁴³ Mean pain scores were significantly lower in patients with unresectable pancreatic cancer undergoing chemical neurolysis with alcohol for patients both with and without preoperative pain. Injections can be performed with 20 mL of 50% ETOH utilizing a 20- or 22-gauge needle and injection on both sides of the celiac trunk and optionally at the base of the left gastric artery. Laparoscopic celiac plexus block has also been demonstrated efficacious in reducing pain scores and offers a minimally invasive approach to palliation for pancreatic cancer patients found to have unresectable disease at laparoscopy or for pain control during staging.⁴⁴ Thoracoscopic splanchnicectomy is also an option for pain control in patients with advanced cancer and has been shown to reduce pain scores by approximately half.⁴⁵

Palliative Pancreaticoduodenectomy

As the morbidity and mortality rates associated with pancreaticoduodenectomy continue to improve, the concept of palliative resection has been raised. As the approaches to management of obstruction and jaundice are well established, the most common recent indication for consultation for consideration of palliative pancreaticoduodenectomy is upper gastrointestinal bleeding. Surgery is rarely a viable option for such patients, and the majority are managed with endoscopy or angiography and embolization.

Several authors have sought to compare patients with positive margins after resection to patients not undergoing resection to investigate the potential benefit to palliative resection.⁴⁶ These studies should be recognized for their limitations as the patients included in these reports do not meet the criteria for palliative surgery or even the concept of preemptive palliation as the patients underwent resection with curative intent. The survival benefit demonstrated in nonrandomized studies comparing resection with positive margins and patients undergoing bypass may therefore be solely due to patient selection.⁴⁷ Although such studies are hypothesis generating, it should be emphasized that patients selected intraoperatively for a bypass procedure would be expected to have more advanced disease than those selected for resection. A recent study that can provide insight into this dilemma compared 42 patients undergoing palliative resection (again defined as resection with positive margins) with 154 patients undergoing double bypass.⁴⁸ Postoperative follow-up included administration of a validated quality-of-life instrument; better quality of life was found in the bypass group. Another study compared patients undergoing resection with grossly positive (R2) margins to patients undergoing surgical bypass procedures, with the finding of increased morbidity and mortality without a clear survival advantage for patients treated with R2 resection.⁴⁹

A more accepted definition of palliative pancreaticoduodenectomy would be surgery performed for symptoms attributable to an incurable malignancy and exclusive of attempt to cure. There are few proponents of routine resection in this situation and fewer studies of this clinical scenario, likely representing either the extremely poor outcomes that dissuade publication of the results or the recognition by surgeons that palliative pancreaticoduodenectomies are generally a poor surgical option, especially for patients with pancreatic adenocarcinoma.

Malignant Ascites

Ascites in pancreatic cancer is often a harbinger of limited survival, but occasional patients with preserved functional status and viable treatment options may be candidates for interventions to address discomfort and mass effects related to the fluid. Simple measures such as diuretics and percutaneous aspiration should constitute the initial approach. Frequent need for aspiration may lead to consideration for placement of indwelling intraperitoneal drainage catheters which can be placed by surgeons or interventional radiologists. Hyperthermic intraperitoneal chemotherapy (HIPEC) without cytoreduction has recently been described for treatment of malignant ascites, although this approach has rarely been applied to patients with pancreatic cancer. A recent multi-institutional analysis of 52 patients undergoing laparoscopic HIPEC for refractory malignant ascites reported a complication rate of 6%, no perioperative mortality, and prevention of ascites reaccumulation in all but one patient.⁵⁰ Of note, however, is that the majority of patients had colorectal and gastric malignancies and the series did not include any patients with pancreatic cancer.⁵⁰

Pancreatic Exocrine Insufficiency

Digestive enzyme insufficiency can be caused by pancreatic duct obstruction and tumor replacement of normal pancreatic parenchyma; patients treated with previous surgery or radiation therapy are at increased risk. Commonly reported symptoms include abdominal pain and cramps, steatorrhea and oily stools, flatulence, and weight loss. Pancreatic enzyme replacement is usually easily performed with pancrelipase capsules taken with meals and snacks. Capsules contain lipase, protease, and amylase in varying doses. A common initial regimen based on recommendations for patients with chronic pancreatitis would begin with 25,000 units of lipase with each meal and approximately 10,000 units of lipase with each snack.^51,52

Depression and Distress Management

Due to the association of pancreatic cancer diagnosis with anxiety, depression, and even increased suicide rate, National Comprehensive Cancer Network Guidelines recommend depression screening and consideration for evaluation by a Palliative Medicine Service for patients with this diagnosis.⁵² In addition, all patients with locally advanced or metastatic pancreatic cancer would meet the criteria for consideration for palliative care referral based on the American Society of Clinical Oncology provisional clinical opinion on the integration of palliative care into standard clinical care.⁴

In summary, the principles of palliative care for patients with pancreatic cancer include the three modalities of medical, procedural, and surgical palliations. Knowledge of all three modalities is critical to providing high-quality multidisciplinary care. As palliative care has been associated with improvements in quality of life and possibly length of life, early referral to a palliative care specialist should be considered for patients with pancreatic adenocarcinoma. Surgical interventions, although decreasing in frequency, are still indicated in selected situations and should be discussed preoperatively in all patients undergoing potentially curative resection.