Pancreaticoduodenectomy

Perhaps one of the most technically challenging abdominal surgeries, pancreatectomy has evolved from a bold innovative intervention to a well-refined lifesaving procedure over the past decades. Pancreatectomy is, however, associated with a long history of high mortality and morbidity. In 1899, William S. Halsted performed the first successful resection of ampullary carcinoma through a transduodenal approach at Johns Hopkins Hospital. In this surgery, he reimplanted the common bile duct and pancreatic duct onto the duodenum, but did not resect the head of pancreas. The first true pancreatectomy did not occur until 1912, when the German surgeon Walther Carl Eduard Kausch performed the first two-stage pancreaticoduodenectomy with an en bloc resection of the head of pancreas for ampullary carcinoma. Mortality was as high as 25%, and mostly resulted from postoperative hemorrhage, peritonitis, and pancreatic fistula.

The transduodenal approach to pancreatectomy continued to be the approach of choice until 1935. In that year, Allen O. Whipple from Columbia University published his first three cases of the two-stage pancreaticoduodenectomy for ampullary carcinoma (Whipple et al., 1935). Whipple described oversewing the pancreatic stump in order to avoid disruption of the pancreaticojejunostomy. Of the three patients in this report, one died during the immediate postoperative period, one died of anastomotic leak a few months later, and the other suffered from pancreatic fistula but survived. Whipple later modified the pancreaticoduodenectomy into a one-stage procedure. Alexander Brunschwig from the University of Chicago is credited, though, as the first to perform a one-stage procedure to resect a pancreatic head cancer. In 1941, Whipple reported his experience with 41 cases of pancreaticoduodenectomy, emphasizing the importance of one-stage procedure to avoid inflammatory adhesions from a two-stage resection along with an end-to-side choledochojejunostomy and jejunojejunostomy for prevention of reflux cholangitis (Whipple, 1941).

From the 1940s to the early 1970s, the mortality from pancreaticoduodenectomy remained at least 25% in most series. The one exception was a report by Dr. John M. Howard in 1968, in which 41 cases of pancreaticoduodenectomy were performed without any mortality at the Hahnemann Hospital (Howard, 1968). With the advent of preoperative risk stratification improved imaging, high-volume pancreatic surgery centers, and interventional radiology, the mortality has decreased significantly, to below <2%. In 1990, Dr. Michael Trede from Heidelberg published the experience of 118 consecutive pancreaticoduodenectomies without mortality (Trede et al., 1990). In 1997, Charles Yeo et al. published the experience at Johns Hopkins Hospital with 650 pancreaticoduodenectomies and an overall mortality of 1.4%. Despite these improvements in mortality and the expansion of the indications for a variety of pathologies, the morbidity remains high to this day, ranging from 25% to 45% at most centers. These complications include pancreatic fistula (PF), delayed gastric emptying (DGE), post-pancreatectomy hemorrhage (PPH), biliary leaks, and pancreatic insufficiency. General complications nonspecific to pancreatectomy such as wound infections, sepsis, cardiac and pulmonary events, and renal failure have significantly decreased.

Distal Pancreatectomy

Distal pancreatectomy for resection of tumors in the body or tail of the pancreas developed at a slower pace than pancreaticoduodenectomy, largely due to its associated pathology. Carcinoma of the pancreas has been very difficult to diagnose historically—especially the lesions located in the body or tail of the pancreas, since these patients usually present later, manifesting metastatic disease. With improved imaging and the recognition of other entities like pancreatic neuroendocrine tumors, cystic neoplasms, and intraductal papillary mucinous neoplasm (IPMN), distal pancreatectomy has become a more frequent procedure.

The German surgeon Werner Körte reported that Trendelenburg performed perhaps the first reported distal pancreatectomy for sarcoma in the tail of the pancreas in 1882. The patient died after a few months. Theodor Billroth performed the first successful distal pancreatectomy with resection of the most of the body and tail of the pancreas in 1884. Since then there were not many reported cases of this procedure until the 1910s. John M.T. Finney from John Hopkins Hospital reviewed 16 cases of pancreatectomy from the literature and also reported one case of his own. In the 16 reviewed cases, four cases appeared to be distal pancreatectomy for a variety of pathologies and included the 1882 Trendelenburg case (Finney, 1910). His own case was actually a central pancreatectomy for pancreatic cystadenoma. This case was complicated postoperatively by pancreatic fistula that eventually resolved, and the patient survived. In 1913, William J. Mayo from the Mayo Clinic reported one case of distal pancreatectomy with the resection of the body and tail of the pancreas for a benign pancreatic cyst (Mayo, 1913). Keith D. Lillemoe published in 1999 the experience of 235 cases of distal pancreatectomies at Johns Hopkins Hospital (Lillemoe et al., 1999). The mortality was less than 1% but the morbidity was 30%. The most common complications were new-onset diabetes, pancreatic fistula, intraabdominal abscess, small bowel obstruction, and postoperative hemorrhage. Other studies have reported pancreatic fistula as being the most common complication in as many as 25% of patients.

Total Pancreatectomy

The most radical pancreatic resection is the total pancreatectomy, and this was first performed in 1944 by James T. Priestly from Mayo Clinic to remove an insulinoma (Priestley et al., 1944). Alexander Brunschwig also reported the first case of total pancreatectomy for carcinoma in 1944 (Brunschwig, 1944). The concept of total pancreatectomy became popularized in the 1960s and 1970s in order to prevent the high recurrence of pancreatic cancer after partial resection, which was thought to be from a multicentric disease. In addition, surgeons believed that the rate of pancreatic fistula was unacceptably high, and removing the entire gland would obviate this concern. However, studies in the 1980s demonstrated that total pancreatectomy carried the same oncological outcome as partial resection (Karpoff et al., 2001; Schnelldorfer et al., 2008; Sperti et al., 1997; Westerdahl et al., 1993). One reason is that total pancreatectomy does not increase the rate of a negative resection margin compared to partial pancreatectomy. In addition, most pancreatic fistulae can be managed medically without significant long-term complications, negating one potential benefit of total pancreatectomy.

The mortality of total pancreatectomy has decreased significantly over the years. In reports published since 1980s, the mortality ranges from 0% to 17% (Balcom et al., 2001; Karpoff et al., 2001; Swope et al., 1994) in high-volume centers. Total pancreatectomy carries significant endocrine and exocrine sequela that require preoperative evaluation and patient education along with lifelong management of this inherent pancreatic insufficiency. The concerns and risks of diabetes can be potentially mitigated with concurrent islet cell autotransplantation for patients undergoing resection for benign conditions such as chronic pancreatitis. However, a long-term follow-up study from the University of Minnesota has shown that 46% of patients became insulin-dependent again after 5 years despite initial insulin independence immediately after surgery (Sutherland et al., 2012).

In this chapter, we discuss the major complications from pancreatectomy: PF, DGE, PPH, biliary leak, and pancreatic insufficiency. These complications can significantly delay patient recovery, the initiation of adjuvant chemoradiotherapy, and impact patients for years following resection. Progress in the understanding and management of these complications has been made. The International Study Group of Pancreatic Surgery (ISGPS) has significantly contributed to these areas and has formulated consensus definitions and grading of PF, DGE, and PPH.

Pancreatic Fistula

Pancreatic fistula is one of the most common and feared complications from partial pancreatectomy and occurs generally in 10% to 25% of cases. The most common clinical presentation is the appearance of murky fluid in a drain left next to the pancreaticoenteric anastomosis or cut edge of the pancreas. Alternatively, the patient manifests an intraabdominal fluid collection of pancreatic juice or abscess following the operation. Historically there have been many different names for this complication, such as pancreatic leak and pancreatic anastomotic insufficiency. The ISGPS has determined these terms as interchangeable and prefers the use of the term “pancreatic fistula” (Shukla et al., 2010).

ETIOLOGY

Pancreatic fistula often results from the disruption of the pancreaticoenteric anastomosis either from ischemia, erosion by pancreatic enzymes, or other technical issues. Pancreatic fistula can also develop from the cut edge of the pancreas if the pancreatic duct has not been sufficiently ligated with suture or controlled with staples.

RISK FACTORS

The texture of the pancreas appears to be the most important risk factor of pancreatic fistula after pancreatectomy. A soft pancreas does not hold suture or staples well, and postoperatively the local inflammatory process can disrupt these. Soft glands are often found in patients with neuroendocrine tumors, ampullary tumors, and cystic lesions. In patients with chronic pancreatitis, the nature of this disease process leads to inflammation and fibrosis of the pancreas, making the gland firm, and the occurrence of pancreatic fistula is relatively low (Donahueand Reber, 2015). Pancreatic cancer is associated with a fibrotic pancreas and this is responsible for a lower PF rate in patients with this disease. Other risk factors include the size of the pancreatic duct diameter and intraoperative blood loss. The ISGPS has posted an on-line risk calculator to help inform the risk of developing a fistula following resection (https://www.pancreasclub.com/calculators/isgps-calculator/).

CLINICAL PRESENTATION

Most commonly, patients with a pancreatic fistula simply demonstrate a change in the character of the effluent found in a surgically placed drain. The fluid no longer appears serosanguinous and instead becomes cloudy or completely clear. Some patients will manifest systemic signs ranging from a prolonged ileus or an elevated white blood cell count with a fever to more ominous signs of a significant inflammatory response with tachycardia, renal dysfunction, and mental confusion. The surgeon should be suspicious of the occurrence of a PF, checking for these changes each postoperative day.

DIAGNOSIS

The criteria to diagnose a pancreatic fistula have varied considerably in the literature. A widely used criteria is fluid output via an operatively placed drain (or a subsequently placed percutaneous drain) of any measurable volume on or after postoperative day 3 with an amylase content greater than three times the upper normal serum value. Abnormal-appearing drainage other than clear pancreatic juice can also be a sign of pancreatic fistula. In the literature, there are at least four widely used criteria, summarized in Table 76-1 (Bassi et al., 2005).

In 2005, the ISGPS released a report to categorize pancreatic fistula into three grades: grade A (mild), grade B (moderate), and grade C (severe). This system was modified recently and the details are outline in Table 76-2.

Imaging is not required to diagnose pancreatic fistula; however, it may be helpful to assess the size and location of the potential intraabdominal abscess, placement of the surgical drain, and the existence of complications that lead to gastric outlet obstruction from anatomical abnormalities.

TREATMENT

The treatment of pancreatic fistula is mostly conservative, and fortunately, most pancreatic fistulae (70%-82%) will resolve within weeks with conservative management. This is true for both pancreaticoduodenectomy and distal pancreatectomy.

With biochemical leak pancreatic fistula, which is the most common form of pancreatic fistula, the patients can still feed orally. Total parenteral nutrition (TPN) or somatostatin analog such as octreotide is not required and this class of fistula rarely delays hospital discharge. In contrast, grade B pancreatic fistula requires significant adjustment from the standard clinical pathway. The patient may require strict NPO and TPN. Octreotide may be indicated if the volume is significant. If the patient has fever or leukocytosis, antibiotics are also needed. Hospital discharged is likely to be delayed as these patients may need interventional drainage of fluid collections or angiographic embolization for hemorrhage, and readmission is more likely to occur. However, the patient can often be discharged home with surgical drain in place and followed up in an outpatient setting.

Grade C pancreatic fistula requires major changes of the standard clinical pathway. The patient often requires NPO, TPN, intravenous antibiotics, and somatostatin analog and care in an intensive care unit. CT scan may show peripancreatic fluid collection. Hospital stays are often lengthened. If the patient continues to deteriorate clinically, reoperation may be required to repair or revise the pancreaticoenteric anastomosis. In extreme conditions, completion pancreatectomy may be necessary.

PREVENTION

Over the years, there have been many studies on potential methods to prevent pancreatic fistula. Fibrin glue and other hemostatic agents have been tested. One of the early trials by Kram et al. (1991) has shown promising results. In their report, no pancreatic fistula occurred in 15 patients. However, late reports consistently failed to show the advantage of fibrin glue (Lillemoe et al., 2004; Orci et al., 2014). For example, Lillemoe et al. (2004) reported that out of 125 patients, pancreatic fistula occurred in 26% of the fibrin-glue group, compared to 30% of the control group. A variation of this method, by internal occlusion of the pancreaticojejunostomy anastomosis, also failed to find a significant difference in the incidence of pancreatic fistula (Lorenz et al., 1988).

A double-blinded randomized clinical trial by Allen et al. (2014) from Memorial Sloan-Kettering Cancer Center has shown promising results with pasireotide for prevention of pancreatic fistula. Pasireotide is another somatostatin analog with higher half-life and better binding capacity than octreotide. Pasireotide or placebo is administered subcutaneously twice daily for 7 days after pancreatectomy. The authors found significant decrease in pancreatic fistula in the pasireotide group compared to the placebo group (9% vs 21%).

Delayed Gastric Emptying

DGE is characterized by oral intolerance, inability to remove the nasogastric tube, and/or the necessity of reinserting the nasogastric tube several days after the operation. It can significantly delay the patient recovery, nutritional improvement, and the initiation of adjuvant therapy. In most reports, the rate of DGE ranges from 19% to 57%. (Martignoni et al., 2000; Miedema et al., 1992; Richter et al., 2003; Wente et al., 2007a; Yamaguchi et al., 1999; Yeo et al., 1997).

ETIOLOGY/RISK FACTORS

The mechanism of DGE is largely unknown. It has been postulated that the resection of duodenum can trigger DGE, and this is supported by the fact that there is less DGE with duodenum-preserving pancreatic head resection. In addition, distal pancreatectomy that does not involve duodenal resection rarely causes DGE. Decreased motilin level has also been suggested to trigger DGE, given that the prokinetic drug erythromycin, which is a motilin agonist, can reduce the incidence of DGE.

Pylorus-preserving pancreaticoduodenectomy is one of the most common variations of the classic pancreaticoduodenectomy, and some reports have claimed that it is associated with higher incidence of DGE, while others have shown the opposite. The etiology of this may be that pylorus-preserving pancreaticoduodenectomy can cause devascularization or denervation of the pylorus with subsequent pylorospasm.