bacteriologic results of FNA and those of surgical specimens is greater than 95%.34 The peak occurrence

of infected necrosis is between 2 and 4 weeks after presentation but it may occur at any time during the

clinical course.

COMPLICATIONS

Table 53-4 Local Complications of Acute Pancreatitis

10 Invasive intervention is usually indicated in the presence of infected necrosis. In contrast, sterile

necrotic collections almost never require intervention in the acute phase of necrotizing pancreatitis. As

infected necrosis often consists of thick and tenacious materials, traditionally mechanical débridement

via an open surgical approach was the primary treatment. Recently, less invasive approaches have been

popularized. Minimally invasive approaches for the treatment of infected pancreatic necrosis are

associated with fewer complications, however the mortality rate and the length of hospital stay appear

to be predominantly determined by the disease process itself rather than the interventional approach. In

centers with the appropriate expertise, most cases of infected pancreatic necrosis can now be managed

using minimally invasive techniques.

11 Procedures for the treatment of infected necrosis are best performed when collections become

walled-off and demarcated from viable pancreatic tissue with at least partial liquefaction, which

typically requires a delay of 4 to 6 weeks after disease onset. This is especially true for débridement, as

when it is performed too early, bleeding, incomplete removal of infected necrosis, resection of vital

tissue, and loss of endocrine and exocrine function are more likely to ensue. Convincing evidence now

supports delaying intervention whenever possible. Antibiotics can often be used to temporize the

situation allowing necrotic collections to mature. A subset of patients may even be successfully treated

with antibiotics alone. Patients with clinical deterioration and signs of sepsis despite the use of

antibiotics who have clearly infected acute necrotic collections may require intervention within the first

few weeks of acute pancreatitis.9,10,35

12 Drainage alone is the initial recommended intervention for infected pancreatic necrosis. This is

most often accomplished through a percutaneous image-guided approach, which is technically feasible

in the vast majority of cases. When percutaneous drains are placed, preference should be given to a

retroperitoneal approach so that the drain tract can be used to perform video-assisted retroperitoneal

débridement (VARD, see below). Drainage can also be accomplished through an endoscopic transluminal

approach, particularly in the rare cases where a percutaneous approach is not feasible. Drainage alone

can successfully treat infected necrosis in approximately 1/3 to 1/2 of cases.9,10,35–38

13 When required, débridement can often be performed through minimally invasive techniques.

These techniques are reserved for patients without intra-abdominal catastrophes or other complications

of acute pancreatitis mandating surgical exploration. VARD and direct endoscopic necrosectomy (DEN)

are the two most widely utilized techniques for minimally invasive débridement.

VARD is usually accomplished through a retroperitoneal approach via flank incisions. The previously

placed drain tracts are dilated to insert an operative nephroscope over a wire or to place laparoscopic

ports that follow the drain tract into the retroperitoneum. Fluid that is encountered is suctioned and

submitted for culture. Hydrodissection is used liberally. Gentle débridement of solid debris can be

accomplished under direct vision through the entry site or under visualization with the videoscope. The

goal of VARD is to facilitate drainage and not necessarily to perform a complete necrosectomy.

Irrigation is usually continued postoperatively through surgically placed drains. In a multicenter

randomized controlled trial, patients with infected pancreatic necrosis were randomized to undergo

primary open necrosectomy or a step-up approach consisting of percutaneous or endoscopic drainage

followed, in most cases (65%), by VARD. Open necrosectomy was rarely used when VARD could not be

1376

accomplished. Although mortality between the two groups did not differ, primary open necrosectomy

was associated with a higher incidence of major complications and increased cost.39

DEN is performed via transmural puncture into a necrotic collection after insertion of an endoscopic.

It is required that the collection abuts or is in close proximity to the stomach or duodenum. The

collection can be visualized as a bulge in the wall of the viscera or using EUS. The tract into the

collection is then dilated and large bore stents are placed. Mechanical débridement can be performed

with endoscopic instruments and/or irrigation. Nasocystic or percutaneous drains can be used to provide

irrigation after the procedure. Typically, multiple sessions are necessary to completely débride the

cavity. A small randomized controlled trial showed a decreased inflammatory response as well as

decreased complications for patients undergoing endoscopic transgastric necrosectomy as opposed to

VARD after failure of drainage for infected pancreatic necrosis.

Open necrosectomy may still be performed in the subset of patients not amenable to a minimally

invasive approach or in centers without expertise in these techniques. For open necrosectomy, the

abdomen is usually entered through a vertical midline or bilateral subcostal incision. The anterior

surface of the pancreas can be exposed by dividing the gastrocolic ligament (greater omentum) and

entering the lesser sac. If inflammatory changes have obliterated the lesser sac, an alternative route to

the pancreas is achieved by dividing avascular portions of the transverse mesocolon. Clearly nonviable

tissue should be débrided bluntly without attempting to perform anatomic resections. The débridement

field is irrigated with several liters of sterile saline. There are several options for assuring that all

necrotic tissue is removed after surgery. Traditionally, when the abdomen is closed, large-bore drains

are placed, and postoperatively the drains are left in place at least 7 days and until the effluent becomes

clear. This procedure is known as necrosectomy with closed drainage. Several authors prefer variants of

this procedure in which either both gauze packing and drains are placed at the time of surgery and

gradually withdrawn postoperatively (necrosectomy with closed packing) or where high-volume lavage of

the lesser sac is performed through the drains placed at the time of surgery until the effluent becomes

clear and the patient’s clinical course improves (necrosectomy with continuous lavage). Other procedures

including necrosectomy with open packing, where the retroperitoneum is marsupialized and the abdomen

left open, and necrosectomy with planned, staged relaparotomy, in which the initial operation is followed

by repeat laparotomies to change gauze packing or to perform additional débridement have fallen out

of favor but may be necessary if necrosectomy is performed early in the course of disease before clear

demarcation between necrotic and viable tissues has occurred. Modern mortality rates associated with

necrosectomy performed for infected necrosis range from 10% to 20%.40,41 On long-term follow-up,

approximately 25% of survivors develop exocrine insufficiency, and 30% develop endocrine

insufficiency. It should be noted that open necrosectomy is associated with better results in more recent

series due to improvements in ICU care and implementation of delayed timing for surgery. The results

of open necrosectomy in relation to VARD or DEN have not been directly compared after initial

drainage. Although surgical débridement is clearly indicated for infected necrosis, its role in sterile

necrosis has undergone evolution. In the past, early necrosectomy was recommended for patients with

necrotizing pancreatitis, even in the absence of documented infection. The rationale for this approach

was to prevent infection from developing and to remove the source of toxins and inflammatory

mediators. Today, it is recommended that surgery should be avoided in patients without documentation

of infected necrosis, based on favorable outcomes reported using this conservative approach.40,41

However, there remains a subset of patients with sterile necrosis who, despite prolonged supportive

care, have persistent problems, including disabling pain, malaise, and gastric outlet obstruction who

may benefit from interventions for sterile necrosis late in the course of disease.

Because of the high morbidity and mortality rates associated with infected necrosis, there has been

much investigation into the use of antibiotics as prophylaxis against infection. Initial clinical trials failed

to demonstrate a benefit of prophylactic antibiotics; however, these studies were flawed by the

inclusion of patients with mild disease who were at low risk for developing infected necrosis and the

use of antibiotics with poor penetration into the pancreas. Trials were published in the 1990s showing a

significant reduction in the incidence of pancreatic infection among patients receiving antibiotic

prophylaxis, and based on this evidence, the use of antibiotic prophylaxis in patients documented to

have necrotizing pancreatitis has become a widespread practice. Several recent trials that failed to show

a benefit for prophylactic antibiotics have now been published.42,43 Similarly, recent meta-analyses of

the available trials have failed to demonstrate a benefit for patients receiving prophylactic

antibiotics.44,45 Disadvantages of using prophylactic antibiotics include the risks of fungal superinfection

and the selection of resistant organisms. Another strategy for prophylaxis against infection in patients

1377

with acute pancreatitis has been the administration of probiotic bacteria to reduce the load of

pathogenic bacteria in the bowel; however, prospective evaluation has yielded disappointing results.46

An algorithm for the general management of acute pancreatitis and pancreatic necrosis is shown in

Algorithm 53-1.

Algorithm 53-1. Algorithm for the management of acute pancreatitis.

Pseudocysts and Sterile Walled-Off Necrosis

Acute fluid collections develop in up to 30% to 50% of patients with acute pancreatitis. As most resolve

spontaneously, no specific treatment directed at acute fluid collections is necessary in the absence of

evidence that the fluid is infected. However, in up to 10% of patients with acute pancreatitis, these fluid

collections progress to develop a wall of fibrous granulation tissue. Until recently, predominantly liquid

collections with a well-formed wall appearing after an episode of acute pancreatitis were classified as

pseudocysts. It is now appreciated that many of the chronic collections seen after bouts of acute

pancreatitis contain variable amounts of solid materials from the necrosis of pancreatic and

peripancreatic tissues. In the new taxonomy, these collections are called walled-off necrosis. This

distinction may better indicate how these collections should be treated, however, there is no distinction

in the indication for intervention between pseudocysts and sterile walled-off necrosis.35 The walls of

these collections generally require at least 4 weeks from the onset of pancreatitis to mature. In contrast

to true cysts, pseudocysts (and walled-off necrosis) do not have epithelium-lined walls.

Most pseudocysts and sterile walled-off necrosis are asymptomatic; however, they can cause upper

abdominal pain, gastric outlet obstruction, and obstructive jaundice. Pseudocysts and walled-off necrosis

can be diagnosed on ultrasonography or CT scanning (Figs. 53-5 and 53-6). It is important to distinguish

1378