Surgical Management of Necrotizing Pancreatitis

Pancreatic necrosis is a highly morbid condition. It is most commonly associated with severe, acute pancreatitis, but can also be caused by trauma or chronic pancreatitis. Once diagnosed, management of pancreatic necrosis begins with supportive care, with an emphasis on early, and preferably, enteral nutrition. Intervention for necrosis, sterile or infected, is dictated by patient symptoms and response to conservative management. When possible, intervention should be delayed to allow the necrotic collection to form a capsule. First-line treatment for necrosis is with percutaneous drainage or endoscopic, transmural drainage. These strategies can be effective as monotherapy, but the need for repeated interventions, or for progression to more invasive interventions, is not uncommon. Necrosectomy may be performed using a previously established drainage tract, as in percutaneous endoscopic necrosectomy (PEN), video-assisted retroperitoneal debridement (VARD), and direct endoscopic necrosectomy (DEN). Although outcomes for these minimally-invasive techniques are better than for traditional necrosectomy, both laparoscopic and open techniques remain important for patients with extensive disease that cannot otherwise be adequately treated. This is especially true when pancreatic necrosis is complicated by disconnected pancreatic duct syndrome (DPDS), where necrosectomy remains standard of care.


Introduction
Pancreatic necrosis is the presence of nonviable pancreatic parenchyma or peripancreatic fat which may be localized or diffuse. It is classified radiologically according to the Atlanta Criteria as an acute necrotic collection (ANC), which is defined as a non-encapsulated area of necrosis, or as walled-off necrosis (WON), which is encapsulated. [1] Although pancreatic necrosis may result from trauma, malignancy, or chronic pancreatitis, the most common cause is acute pancreatitis; 20% of patients with acute pancreatitis develop necrosis. For patients who develop necrosis, the mortality rate is 15-30%. [2] Surgery has historically been the primary treatment for pancreatic necrosis. However, the superior outcomes associated with new, less invasive techniques have narrowed the scope for surgical intervention. Despite these shifts in practice, surgical care remains an important tool for the treatment of pancreatic necrosis. lumen, these stents also allow access to the area for debridement, via irrigation or DEN. [21] (Figure 2) In DEN, an endoscope with one or two working ports is advanced through the previously placed, transluminal stent. Upon entering the WON, a number of tools, including forceps and snares are used to remove debris that would otherwise not be susceptible to removal with irrigation. [21] On average, 3-6 endoscopic interventions are necessary prior to resolution of necrosis. [22] DEN was first compared to surgical necrosectomy in the Pancreatitis, Endoscopic Transgastric vs. Primary Necrosectomy in Patients with Infected Necrosis (PENQUIN) Trial. In this trial, patients in the surgery group underwent a number of different operations, including 6 video-assisted retroperitoneal debridement (VARD) surgeries, 4 open necrosectomies, and 2 percutaneous drainage placements without need for more invasive therapy. The two patient who did not have a necrosectomy were excluded from final statistical analysis. All 10 patients in the endoscopic group had ultrasound guided stent placement, irrigation, and DEN. Following intervention, the rates of new-onset organ failure and pancreatic fistula were lower in the endoscopic group. The trial also compared the groups with regard to a composite clinical outcome, which included major post-operative complications and mortality, and found a lower rate in the endoscopic group. [23,24] These findings were later replicated in the Minimallyinvasive Surgery Versus Endoscopy Randomized (MISER) Trial. Additionally, MISER showed lower rates of pancreatic fistula formation and a higher quality of life at 3 months after surgery in the endoscopic group.
[25] In the Transluminal Endoscopic Step-up Approach Versus Minimally-invasive Surgical Step-up Approach in Patients with Infected Necrotizing Pancreatitis (TENSION) Trial, a larger randomized trial, no difference in mortality was observed. However, the rates of pancreatic fistula and length of stay favored the endoscopic group. [26] Percutaneous drainage is preferable in patients that are deemed too unstable to tolerate endoscopic drainage or if the area of necrosis extends into a dependent  space, such as the paracolic gutters or pelvis. It is also an acceptable alternative when endoscopic drainage is unavailable or not technically feasible, specifically in the setting of ANC, when there is no capsule that could support an endoscopic stent. [27] Percutaneous drainage is usually CT-guided, although ultrasound-guided drainage can also be performed. These drains may be transperitoneal, with the external portion of the drain fixed in the anterior abdominal wall. These drains may also be placed through the flank, directly into the retroperitoneum, without traversing the peritoneum. In addition to draining ANC and WON, percutaneous drains can also be used for irrigation. [28] Although percutaneous drainage is successful as monotherapy in some patients, patients with larger areas of necrosis, multifocal necrosis, incomplete liquefaction, and pre-procedural organ failure are less likely to be adequately treated. While some of these factors can be overcome with larger drainage catheters, for these reasons, percutaneous drainage remains a bridge to therapy, allowing patients to survive the acute period of disease, and undergo definitive management later, with improved outcomes. [26,29,30]

General considerations for surgical management
Surgical management may be minimally-invasive or open, but has the same two primary goals: obtaining source control by removing as much necrotic tissue as possible and providing access for irrigation and drainage. As a general principle, minimally-invasive approaches are preferred to open necrosectomy as first-line treatment. The improved outcomes of minimally-invasive technique lead to development of the "step-up" approach to management, which begins with percutaneous or endoscopic intervention, followed by a progression to surgical intervention as indicated by unresolved disease. However, the final treatment decision is dictated by the patient, surgeon, and available resources. A second principle is that surgical intervention should be delayed as long as possible in order to improve outcomes. Operating during the early, acute phase of pancreatitis, especially in the presence of ANC, rather than WON, is associated with higher morbidity and mortality regardless of surgical approach. A third principle is that long-term nutritional access, through a gastrostomy or gastrojejunostomy tube, should be obtained prior to concluding the procedure if no other method for enteral feeding has been established. Fourth, a cholecystectomy may also be performed if gallstones were implicated in the etiology of pancreatitis, provided the patient is adequately stable to undergo an additional procedure. (Figure 1).

Minimally-invasive Necrosectomy
VARD is a technique, used as the final phase of the step-up approach, where the retroperitoneum is accessed through a previously established, left flank, percutaneous drainage tract. (Figure 2) The tract is then serially dilated, in order to accommodate progressively larger drainage catheters. At the time of surgery, in order to facilitate introduction of laparoscopic instruments, a small, 4-6 centimeter incision is made where the tract exits the skin. After confirming entry into the WON with a probe, tissue and fluid are removed with suction. The laparoscope is then inserted, with or without CO 2 insufflation, for continued debridement under direct visualization, using blunt laparoscopic forceps. Following debridement, again under direct visualization, large drainage catheters or chest tubes, 28-French or greater, are placed. After surgery, these catheters are used for repeated lavage as well as for drainage. [31] The superiority of VARD, and the step-up approach, compared to surgery for the treatment of necrotizing pancreatitis was first published in the Minimallyinvasive Step-up Approach Versus Maximal Necrosectomy in Patients with Acute Necrotizing Pancreatitis (PANTER) Trial. In this study, 35% of the patients assigned to the step-up arm were successfully treated with percutaneous drainage alone. When comparing the step-up and surgical groups, the step-up group was less likely to have new-onset organ failure, less likely to develop an incisional hernia, and had an overall lower rate of endocrine insufficiency. However, the mortality rate was not significantly different, 19% in the step-up group versus 16% in surgery group. [31] A similar procedure, percutaneous endoscopic necrosectomy (PEN), can be performed utilizing a previously established percutaneous drainage tract. Unlike VARD, PEN utilizes a flexible endoscope, as compared to a rigid laparoscope. Because the endoscope has working ports, in addition to irrigation and suction, an additional incision around the tract is not needed. Also unlike VARD, PEN can be performed at bedside, with conscious sedation. [32] PEN was shown to be effective in a large, prospective study of 171 patients with infected pancreatic necrosis. The primary outcome investigated was control of sepsis and resolution of the infected collection. In this study, 18 of 26 (69%) patients with infected ANC and 23 of 27 (85%) with infected WON who underwent PEN were successfully treated, while the remainder required surgical necrosectomy. Predictors of failure included >50% parenchymal necrosis and early organ failure. ANC was not predictive. The overall mortality rate for this study was 38%. [32,33] Although this technique has not been directly compared to surgery, VARD, or transmural endoscopy, this study demonstrated the safety and utility of PEN in patients with infected pancreatic necrosis.
Regardless of the type of minimally-invasive drainage, VARD or PEN, it has been shown that the "step-up approach," beginning with drainage and progressing to debridement, is superior to upfront surgical approaches in terms of mortality, rates of pancreaticocutaneous fistula formation, and long-term morbidity. [25, 30,34]

Transgastric necrosectomy
In addition to utilizing a percutaneous drainage tract for necrosectomy, access can also be gained through the stomach. By entering the abdomen and opening the anterior wall of the stomach and then opening the posterior aspect of the stomach, access to the lesser sac and underlying pancreas is achieved. (Figure 2) An aperture between the WON and posterior wall of the stomach is then created, either with sutures or by stapling, providing a definitive drainage tract. This tract is then used for necrosectomy following the same principles as DEN.
This approach is most well suited for WON limited to the lesser sac. When there is extensive necrosis extending to the retroperitoneum or paracolic gutters, VARD or traditional necrosectomy are more appropriate, due to the limited exposure with this method. These limitations are counterbalanced by the minimal amount of mobilization required to enter the lesser sac by the transgastric method. [35] When performed laparoscopically, five ports are typically placed; in addition to an umbilical port, two ports are placed in the right upper quadrant, one port is placed in the left upper quadrant, and one port in the epigastrium. After entering the abdomen and creating the anterior gastrotomy, an ultrasound is used to identify the necrosis and plan the locations of the posterior gastrotomy. Ultrasound is adjunctive to preoperative imaging, which is also essential to surgical planning. Both anterior and posterior gastrotomies should be made after placement of stay sutures. Upon entering the lesser sac, necrosectomy should be performed with blunt instruments, such as a ring forceps, taking great care to remove only loose material and avoid avulsing adherent tissue or vessels that may be bridging the area of necrosis. Following necrosectomy, a cystogastrostomy is created with an endoscopic stapler, or suture. The anterior gastrotomy is then closed with sutures or with a stapler. [36] When performed open, an upper midline incision is made, and the procedure proceeds in the same fashion as in the laparoscopic procedure. One difference in the open procedure is that many surgeons elect to use digital dissection for the necrosectomy, as opposed to instruments. [37] Open and laparoscopic approaches to transgastric drainage have been shown to have similar outcomes. In a recent retrospective review of patients from three tertiary referral centers, rates of morbidity, including rates of reoperation and hemorrhage, and mortality were not significantly different. However, the patients who underwent laparoscopic drainage had a higher rate of readmission. It should be noted that the overall mortality in this study was 2% at an average follow-up of 21 months, significantly less than reported elsewhere in the literature. The overall morbidity rate of 38% is in alignment with commonly reported rates elsewhere in literature. [38] Although surgical transgastric necrosectomy is relatively well tolerated, outcomes favor endoscopic transgastric drainage. Meta-analysis comparing the two show lower rates of overall major complications, pancreatic fistula formation, post-procedural organ failure, and hernia with an endoscopic approach. However, the overall rate of clinical resolution, post-operative bleeding, endocrine dysfunction, exocrine insufficiency, and mortality were not significantly different. [39] Thus, surgical transgastric necrosectomy is a valid alternative to other approaches of necrosectomy in the absence of an experienced endoscopist or at a center without access to advanced endoscopic tools.

Laparoscopic and open necrosectomy
Although utilization of a drainage tract and the transgastric approach are important for management of pancreatic necrosis, traditional laparoscopic and open necrosectomy methods also continue to be utilized.
For laparoscopic necrosectomy, patients are typically placed in lithotomy position, with the operating surgeon standing between the legs. An umbilical port is placed first. Upon entering the abdomen, a diagnostic laparoscopy should be performed. Subsequently, two left lateral ports and an epigastric port are placed. In some cases, a hand-assist port is placed to augment dissection and removal of tissue. Following lysis of adhesions, a transgastrocolic, for pancreatic head or body necrosis, or transmesocolic, for pancreatic tail necrosis, approach to retrogastric necrosectomy can be taken. (Figure 2) Upon entering the area of necrosis, blunt instruments are used to remove loose, necrotic tissue. This tissue is then placed into an endocatch bag for removal from the abdomen. Dissection is alternated with irrigation and suction to remove as much necrotic tissue as possible. [40] Once the necrosectomy is complete, large drainage catheters are placed in the cavity, which also allow for post-operative irrigation. At this time, consideration should also be given to cholecystectomy, if gallstones were implicated in the development of pancreatitis, and to nutritional access. Depending on the specific study, mortality for patients who require laparoscopic necrosectomy ranges from 10 to 18%. Rates of reoperation also vary widely, ranging from 11 to 38%. [41,42] The most invasive procedure used for the treatment of pancreatic necrosis is the open debridement. This technique is reserved for patients that fail other less invasive techniques, or patients who require concurrent intervention for another intraabdominal process, such as bowel ischemia or abdominal compartment syndrome. Unless midline laparotomy is required for another indication, the abdomen can be opened with bilateral, subcostal incisions. The gastrocolic ligament is then opened, and the stomach is reflected superiorly, exposing the lesser sac. (Figure 2) The transverse mesocolon is then opened, exposing the retroperitoneum. The hepatic and splenic flexures of the transverse colon are often taken down at this point. A Kocher maneuver may also be necessary if the area of necrosis involves the head of the pancreas. Once the pancreas is adequately exposed, blunt debridement can begin. This is usually accomplished with digital dissection or with lavage in order to minimize the risk of bleeding or bile duct injury. These risks must be balances with adequate removal of loose, nonviable tissue. Wide drainage of the area with a sumping tube (i.e. Abramson drain) can facilitate continue lavage and debridement. The quality of the initial necrosectomy predicts the need for subsequent operations.
After necrosectomy, the abdomen may be kept open, with packing in place, to allow for repeated removal of necrotic tissue. Alternatively, the closed packing technique can also be used. This technique consists of filling the cavity created by the necrosectomy with gauze-filled Penrose drains. The drains are removed one at a time, until the cavity closes. [43] A third option is continuous irrigation, where large catheters are placed into the lesser sac under direct visualization. Additional drainage catheters are left in the peritoneal cavity. The abdomen is then closed and the large catheters are used for continuous installation of hypertonic fluid. [44] As in patients who undergo laparoscopic necrosectomy, the rates of morbidity and mortality following open necrosectomy are high. Rates of post-operative morbidity range from 34 to 95% and mortality ranges from 6 to 47%, depending on the pre-operative severity of illness. Rates of reoperation vary depending on the packing technique. When the abdomen is left open, reoperation is planned rather than required because of deterioration or other complications, such as hemorrhage. Depending on the study, when the abdomen is left open, patients may return to the operating room from 1 to 17 time. Comparatively, relaparotomy is required in 17% of patients treated with closed packing require and 17-27% of patients treated with continuous irrigation. Rates of pancreatic fistula also differ depending on packing technique with a 25-46% rate in open abdomens, 53% rate in closed packing, and 13-19% rate with continuous irrigation. [45] The outcomes for both of these techniques are improved when intervention can be delayed at least 3 weeks. Delayed necrosectomy is associated with lower rates of exocrine and endocrine insufficiency, adverse post-operative events, including bleeding, and mortality. [17,46] Early surgical intervention only provides a survival benefit in the case of decompression of abdominal compartment syndrome. [47,48] When compared directly, in a retrospective case series, the rates of pancreatic fistula, post-operative pulmonary infections, and surgical site infections were all significantly lower with laparoscopic necrosectomy. Additionally, patients who underwent laparoscopic necrosectomy also had a shorter length of stay, but a longer initial operation. There was no difference in need for reoperation, overall morbidity, or mortality. It should be noted that mortality was very low compared to other literature in this study, 5.9% in the open group and 4% in the laparoscopic group. [49]

Disconnected pancreatic duct syndrome
While parenchymal destruction in pancreatic necrosis confers significant morbidity and mortality, the seriousness of this condition can be further compounded by concurrent disruption of the pancreatic duct. Disconnected pancreatic duct syndrome (DPDS) occurs when the remnant of pancreas distal to the necrosis, and duct disruption, remains viable and continues to release digestive enzymes into the retroperitoneum. DPDS most commonly occurs in the setting of severe acute pancreatitis, and can be found in up to 46% of patients with pancreatic necrosis. [50] DPDS can also occur as the result of trauma and chronic pancreatitis. The clinical presentation of DPDS is heterogenous. Some patients are asymptomatic and the injury is incidentally diagnosed on radiology. While others may have early satiety due to the size of the resulting fluid collection or symptomatic ascites. [51,52] DPDS is an often overlooked complication due to the low accuracy of imaging in differentiating between full-thickness pancreatic necrosis, affecting the pancreatic duct, and partial thickness or peripancreatic necrosis. Often multiple imaging modalities are required for accurate diagnosis, which in turn leads to delays in diagnosis, increased morbidity, and increased costs. [53][54][55] Diagnostic criteria for DPDS include: necrosis of ≥2 cm of pancreatic parenchyma, viable pancreatic tissue distal to the area of necrosis, and extravasation of contrast when injected into the main pancreatic duct during ERCP. [56] Once DPDS is diagnosed, choice of intervention is dependent on the patient's clinical condition and the phase of disease. As in pancreatic necrosis without DPDS, intervention during the acute phase, when inflammation is high, is not only challenging, but also hazardous. Although the historical standard of care for these patients was surgery, if a patient deteriorates during the acute phase, initial therapy should be percutaneous or endoscopic. Percutaneous drainage, although useful as a temporizing measure, especially in unstable patients, is unlikely to succeed as monotherapy. [57,58] Although success rates are dependent on the extent of necrosis, transpapillary and transmural endoscopic interventions have better shortterm outcomes, with up to an 87% success rate of fistula resolution. [50,59,60] However, in order for endoscopic treatment to be successful, multiple interventions are often required, including hybrid approaches with percutaneous drains. Further, long-term data regarding patency and migration of indwelling stents is not available. [60,61] Thus, percutaneous and endoscopic treatments remain temporizing measures, rather than definitive treatment, for DPDS, except for in patients who are poor surgical candidates. [62] Once a patient reaches the late stage of disease, or if a patient deteriorates despite optimal percutaneous and endoscopic intervention during the acute phase, surgery becomes the primary treatment for DPDS. Because of the technical difficulty of operating in the retroperitoneum after tissue planes have been obscured by inflammation, and because of the frequency of splenic vein thrombosis, and resulting sinistral portal hypertension, this operation is usually performed with a midline laparotomy and not laparoscopically. [63] Surgery for DPDS consists of resection of the distal, disconnected pancreas, and creation of internal drainage tracts. These techniques may be used independently or in combination. When the entirety of the disconnected pancreas is resected, splenectomy is also performed in almost all cases. However, when a pancreatojejunostomy, pancreaticogastrostomy, or fistuloenterostomy is made with the viable distal pancreas, the spleen may be preserved, in addition to preserving the pancreatic remnant. In this way, internal drainage not only provide a conduit for pancreatic secretions, but also decreases the risk of exocrine pancreatic insufficiency and diabetes. Importantly, patients who undergo internal drainage, compared to other surgical modalities, also have lower incidence of organ failure, development of pancreatic fistula, and need for long-term percutaneous drainage. [50,64]

Conclusions
Pancreatic necrosis is a significant and challenging disease process with mortality reaching beyond 30% in most studies. Intervention begins with supportive care and nutritional support. However, invasive therapy is often needed, especially when necrosis becomes infected.
First-line interventions for pancreatic necrosis may be percutaneous or transmural endoscopic drainage depending on if the necrosis is encapsulated, the accessibility of the necrosis, the patient's clinical condition, and the capabilities of the hospital. These minimally-invasive interventions are often successful as monotherapy, without the need for further intervention. They are also preferable to open or laparoscopic necrosectomy when performed as part of a step-up approach.
Despite all of the improvement in minimally-invasive management of pancreatic necrosis, some percentage of patients continue to require surgical intervention.
Both laparoscopic and open approaches have been shown to be effective via transgastric, transgastrocolic, and transmesocolic routes.
When pancreatic necrosis is further complicated by a disconnected pancreatic duct, although minimally-invasive management has been described and shown to be effective, surgical management remains standard of care.
Despite the advances in care driven by clinical trials and new technology, management of pancreatic necrosis remains difficult. Further study is needed to reduce the morbidity and mortality of this devastating disease.