Chapter 47 Acute Pancreatitis and Peripancreatic Fluid Collections
Video related to this chapter’s topics: Transgastric Drainage of Acute Pancreatic Fluid Collection
Introduction
Acute pancreatitis may be clinically mild or severe. Clinically severe acute pancreatitis is usually a result of pancreatic glandular necrosis. The morbidity and mortality of acute pancreatitis are significantly higher when pancreatic necrosis is present, especially when infection of the necrosis occurs.1 Patients with pancreatic necrosis must be identified so that appropriate management can be undertaken. The management of patients with necrotizing pancreatitis has shifted from early surgical débridement (necrosectomy) to aggressive intensive medical care. Specific criteria for operative or nonoperative intervention have been developed.2,3 Advances in radiologic imaging and aggressive medical management with emphasis on prevention of infection have allowed for prompt identification of complications and improvement in outcome for these patients.4 Several types of pancreatic and peripancreatic fluid collections may arise as a result of acute pancreatitis,5 including acute fluid collections, acute pancreatic pseudocysts, pancreatic abscesses, and organized pancreatic necrosis. This chapter reviews more recent advances in the diagnosis and treatment of acute pancreatitis and peripancreatic fluid collections.
Presentation and Classification of Acute Pancreatitis
Acute pancreatitis usually has a rapid onset manifested by upper abdominal pain, vomiting, fever, tachycardia, leukocytosis, and elevated serum levels of pancreatic enzymes. Gallstone and alcohol-induced pancreatitis are the most common causes in the United States. Box 47.1 lists causes of pancreatitis. Several classifications of severity of illness for acute pancreatitis are used to identify patients at risk for developing complications (Table 47.1).6,7
RANSON’s CRITERIA OF SEVERITY | |
---|---|
At Admission | During Initial 48 Hours |
Age >55 yr | Hematocrit decrease >10% |
WBC >16,000/mm3 | BUN increase >5 mg/dL |
Blood glucose >200 mg/dL | Serum calcium <8 mg/dL |
Serum LDH >350 IU/L | Pao2 <60 mm Hg |
Serum AST >250 IU/L | Base deficit >4 mEq/L |
Fluid sequestration >6 L | |
Score ≥3 is considered severe | |
APACHE II SCORE* | |
Score ≥8 is considered severe | |
GLASGOW CRITERIA | |
Within 48 hr of hospitalization | |
Age >55 yr | |
WBC >15,000/mm3 | |
Glucose >180 mg/dL | |
BUN >45 mg/dL | |
LDH >600 U/L | |
Albumin <3.3 g/dL | |
Calcium <8 mg/dL | |
Pao2 <60 mm Hg | |
Score ≥3 is considered severe | |
BISAP SCORE | |
BUN >25 mg/dL | |
Impaired mental status (Glasgow Coma Scale score <15) | |
SIRS—defined as ≥2 of the following | |
Temperature of <36° C or >38° C | |
Respiratory rate >20 breaths/min or Paco2 <32 mm Hg | |
Pulse >90 beats/min | |
WBC <4000/mm3 or >12,000/mm3 or >10% immature bands | |
Age >60 yr | |
Pleural effusion detected on imaging |
AST, aspartate aminotransferase; BISAP, bedside index for severity in acute pancreatitis; BUN, blood urea nitrogen; LDH, lactate dehydrogenase; Paco2, arterial carbon dioxide tension; Pao, arterial oxygen tension; SIRS, systemic inflammatory response syndrome; WBC, white blood cell count.
* Acute physiology score + age points + chronic health points.5
Ranson’s score consists of 11 clinical signs with prognostic significance; 5 signs are measured at the time of admission, and 6 signs are measured between admission and 48 hours later. There is good correlation between the number of Ranson signs and the incidence of systemic complications and the presence of pancreatic necrosis.6 The Acute Physiology, Age, and Chronic Health Evaluation (APACHE) II score is a grading system based on 12 physiologic variables, patient age, and prior history of severe organ system insufficiency or immunocompromised state.6 This score allows stratification of illness severity on admission and may be recalculated daily. Severe acute pancreatitis is present if there are three or more Ranson’s criteria, if the APACHE II score is 8 or greater, or if clinical findings of one or more of the following are present: shock, renal insufficiency, or pulmonary insufficiency.6 The Glasgow scoring system is another classification system.7 In contrast to Ranson’s criteria, the variables apply if they occur at any time within 48 hours. More recently, a bedside index for severity in acute pancreatitis (BISAP) score was developed.8 The BISAP score has five variables: blood urea nitrogen greater than 25 mg/dL, impaired mental status, systemic inflammatory response syndrome, age older than 60 years, and pleural effusion detected on imaging. One point is assigned for each variable within 24 hours of presentation to yield a composite score of 0 to 5 (see Table 47.1).
Acute pancreatitis may be classified histologically as interstitial-edematous or necrotizing based on the inflammatory changes of the pancreatic parenchyma.9 According to the International Symposium on Acute Pancreatitis in 1992, pancreatic necrosis is defined as one or more diffuse or focal areas of nonviable pancreatic parenchyma (Fig. 47.1).9 Pancreatic glandular necrosis is usually associated with peripancreatic fat necrosis.9–11 By definition, the presence of pancreatic necrosis represents a severe form of acute pancreatitis.9 Approximately 20% to 30% of the 185,000 new cases of acute pancreatitis per year in the United States are necrotizing.12,13
Management of Mild Acute Pancreatitis (Interstitial-Edematous)
The management of patients with clinically mild or interstitial pancreatitis is almost entirely supportive. Patients with this form of pancreatitis generally have a self-limited course. There is no role for antibiotic therapy or nutritional support. The goal of management is to identify the etiology of the pancreatitis and to treat it appropriately. Identification of gallstones in the absence of other factors should be managed with cholecystectomy. Potentially offending drugs should be discontinued, and hyperlipidemia should be addressed. If there is alcohol abuse, a rehabilitation program should be offered to the patient. In some patients, an etiology for acute pancreatitis is not identified, and recurrent attacks occur. The management of these patients is controversial, specifically as it relates to sphincter of Oddi dysfunction and pancreas divisum. Data support pursuing endoscopic evaluation of these entities,14 although there are no randomized trials proving effectiveness of endoscopic therapy.15
Identification and Clinical Importance of Pancreatic Necrosis
Pancreatic necrosis may be pathologically identified at surgery or autopsy. The radiographic diagnosis of pancreatic necrosis is determined by dynamic intravenous contrast–enhanced abdominal computed tomography (CT).10 Because the normal pancreatic microcirculation is disrupted during acute necrotizing pancreatitis, contrast-enhanced abdominal CT shows a lack of normal contrast enhancement of affected portions of the pancreas (see Fig. 47.1)16; this may be better detected several days after initial clinical presentation. Contrast-enhanced abdominal CT is the “gold standard” for noninvasive diagnosis of pancreatic necrosis, with an accuracy of greater than 90% when more than 30% glandular necrosis is present.10 The presence of radiographically detected pancreatic necrosis markedly increases the morbidity and mortality associated with acute pancreatitis.17 As the percentage of glandular necrosis increases, the morbidity increases.
Theoretically, contrast medium may cause significant additional reductions of capillary flow, which has been shown to aggravate acute pancreatitis in experimental studies. However, a more recent study in men with severe acute pancreatitis compared patients who did not receive contrast medium with patients who did. Patients in whom contrast medium was administered did not show deterioration of acute pancreatitis.18 The overall mortality in severe acute pancreatitis is approximately 30%.13 The mortality occurs in two phases. Early deaths (1 to 2 weeks after onset of pancreatitis) are due to multisystem organ failure from release of inflammatory mediators and cytokines.1 Late deaths result from local or systemic infections.19 As long as acute necrotizing pancreatitis remains sterile, the overall mortality is approximately 10%. The mortality rate at least triples if infected necrosis occurs.12
Patients with sterile necrosis and high severity of illness scores (Ranson’s scores, APACHE II scores) accompanied by multisystem organ failure, shock, or renal insufficiency have a significantly higher mortality.20 Myriad systemic and local complications of acute necrotizing pancreatitis may occur. Systemic complications have been detailed elsewhere2 and include adult respiratory distress syndrome, acute renal failure, shock, coagulopathy, hyperglycemia, and hypocalcemia. Local complications include gastrointestinal bleeding, infected necrosis, and adjacent bowel necrosis. Late local complications that may require therapy include development of pancreatic abscess or pancreatic pseudocyst. Early therapy of acute necrotizing pancreatitis consists of the combination of aggressive supportive intensive medical care and prevention of infection using prophylactic antibiotics. Late management requires recognition of local infectious complications (pancreatic infection) and the initiation of aggressive débridement strategies. Infected necrosis develops in 30% to 70% of patients with acute necrotizing pancreatitis and accounts for more than 80% of deaths from acute pancreatitis.1,3 The risk of infected necrosis increases with increasing amounts of pancreatic glandular necrosis and length of time from onset of acute pancreatitis, peaking at 3 weeks.1,3
Infection in Acute Necrotizing Pancreatitis
Because the development of infected necrosis significantly increases the mortality of acute necrotizing pancreatitis, prevention of infection is critical. In experimental acute necrotizing pancreatitis, pancreatic infection occurs primarily as a result of bacterial translocation from the colon.21 Several animal studies have shown a decrease in pancreatic infection and mortality using orally administered antibiotics for “selective decontamination” of the gut or intravenous antibiotics with high pancreatic tissue penetration.21–23 Similarly, human studies have shown benefits from orally administered antibiotics with or without rectally administered antibiotics for “selective decontamination” of the gut.24,25 This regimen has not gained acceptance.
The use of systemic antibiotics for the prevention of pancreatic infection has been studied extensively. Numerous studies have been published in animals and humans regarding the use of prophylactic antibiotics. In addition, several meta-analyses have been performed showing conflicting results. The evidence does not show convincingly that the use of prophylactic antibiotics decreases mortality and need for surgery, although these do seem to decrease in nonpancreatic infections.26–30 At the present time, if administration of prophylactic antibiotics is chosen, intravenous antibiotics with excellent pancreatic tissue penetration (extended penicillins, carbapenems, or fluoroquinolones) is recommended. Duration of therapy of 2 weeks is advocated by some experts.
Sterile and infected acute necrotizing pancreatitis can be difficult to distinguish clinically because both may produce fever, leukocytosis, and severe abdominal pain. The distinction is important because the mortality in patients with infected acute necrotizing pancreatitis without intervention is nearly 100%.12 The bacteriologic status of the pancreas may be determined by CT-guided fine needle aspiration of pancreatic and peripancreatic tissue or fluid (Fig. 47.2).31,32 This aspiration method is safe and accurate with a sensitivity of 96% and specificity of 99%, and it is recommended in patients with acute necrotizing pancreatitis who experience clinical deterioration or who do not experience clinical improvement despite aggressive supportive care.2 Ultrasound-guided aspiration may have a lower sensitivity and specificity33 but can be performed at the bedside. Surveillance aspiration may be repeated on a weekly basis as clinically indicated.
Role of Endoscopic Retrograde Cholangiopancreatography in Severe Acute Gallstone Pancreatitis
Gallstone pancreatitis is caused by impaction of a stone within the common channel of the ampulla of Vater. In most cases, the stone passes without therapy. It is assumed that endoscopic retrograde cholangiopancreatography (ERCP) and biliary sphincterotomy would improve the outcome of gallstone pancreatitis because removal of an impacted stone would relieve obstruction to the flow of pancreatic secretions. Initial studies performing urgent (within 72 hours of admission) ERCP and biliary sphincterotomy (if a stone is identified) in patients with acute gallstone pancreatitis and choledocholithiasis showed an improved outcome in only the group of patients presenting with clinically severe acute pancreatitis.34 The improvement was attributed to relief of pancreatic ductal obstruction produced by an impacted gallstone in the common biliary-pancreatic channel of the ampulla of Vater. More recent studies suggest the improved outcome after ERCP and sphincterotomy in gallstone pancreatitis results from reduced biliary sepsis, rather than a true improvement in pancreatitis.35,36
In the presence of pancreatic ductal disruption, a frequent occurrence in acute necrotizing pancreatitis,37 introduction of infection by incidental pancreatography during ERCP may theoretically occur, transforming acute necrotizing pancreatitis from sterile to infected. ERCP in patients with severe gallstone acute pancreatitis must be employed judiciously and reserved for patients with suspected biliary obstruction based on hyperbilirubinemia and clinical cholangitis because it is unlikely that the ampulla is obstructed in the presence of a normal serum bilirubin.38,39 Other imaging modalities, such as endoscopic ultrasound (EUS) and magnetic resonance cholangiopancreatography, may be used in patients with severe biliary pancreatitis with the goal of selectively performing ERCP in patients with documented bile duct stones.7,40 Whether empiric biliary sphincterotomy should be performed if bile duct stones are not identified is unknown, but this seems to be a reasonable approach.41
Nutritional Support for Acute Necrotizing Pancreatitis
To meet increased metabolic demands and to “rest” the pancreas, total parenteral nutrition (TPN) administered through a central venous catheter was used for nutritional support in patients with acute necrotizing pancreatitis. TPN does not hasten resolution of acute pancreatitis or preserve gut integrity, an important factor in preventing bacterial translocation. Randomized prospective studies of TPN and enteral feeding (through a nasoenteric feeding tube placed radiographically beyond the ligament of Treitz) instituted within 48 hours of onset of severe acute pancreatitis42,43 showed that enteral feeding was well tolerated without adverse clinical effects and resulted in significantly fewer total and infectious complications. In a meta-analysis, enteral nutrition significantly reduced mortality, infectious complications, and pancreatic infections in patients with predicted severe pancreatitis compared with TPN.44 The cost of nutritional support was threefold higher in the TPN group.42 Acute phase response and disease severity scores were significantly improved after enteral nutrition.43
This form of enteral feeding seems to be preferable in patients with acute necrotizing pancreatitis in the absence of a significant ileus.45 This approach has been extended to nasogastric feeding. In a review of four studies comprising 92 patients in which nasogastric feeding was compared with nasojejunal feeding, mortality and intolerance to feeding were similar.46
Enteral feeding is the preferred strategy with TPN reserved for patients when the gut has failed or administration of enteral nutrition is impossible for other reasons (e.g., prolonged ileus, complex pancreatic fistulas, abdominal compartment syndrome).47 Various endoscopic techniques are available for placing nasojejunal feeding tubes in the setting of acute pancreatitis.47 One method avoids the need to transfer the guidewire from the mouth to the nose. A small-caliber endoscope is passed transnasally into the duodenum.48 A guidewire is advanced through the endoscope beyond the ligament of Treitz. The endoscope is withdrawn leaving the guidewire in place. The tube is passed over the guidewire with or without fluoroscopic guidance.
Interventions for Pancreatic Necrosis
The timing and type of pancreatic intervention for patients with acute necrotizing pancreatitis are controversial. Because the mortality from sterile acute necrotizing pancreatitis is approximately 10%, and surgical intervention has not been shown to reduce this figure, most investigators recommend supportive medical therapy in this group.12 Conversely, infected acute necrotizing pancreatitis is traditionally considered uniformly fatal without intervention,12 although small, retrospective studies found that antibiotic therapy alone was effective in a select group of patients.49,50 Surgical pancreatic débridement (necrosectomy) remains the standard with which other drainage strategies are compared and may require multiple abdominal reexplorations.45
Necrosectomy should be undertaken soon after confirmation of infected necrosis. The role of surgery in patients with multisystem organ failure and sterile necrosis remains unproved, although this scenario is frequently cited as an indication for surgical débridement.51 Additionally, the longer that surgical intervention can be delayed from the onset of acute necrotizing pancreatitis, the better the patient survival52; this is probably related to improved demarcation between viable and necrotic tissue at the time of operation. The role of delayed necrosectomy (after resolution of multisystem organ failure) in sterile acute necrotizing pancreatitis likewise is controversial. Some investigators advocate débridement in patients who remain systemically ill 4 to 6 weeks after onset of acute pancreatitis with fever, weight loss, intractable abdominal pain, inability to eat, and “failure to thrive.”2,53,54 Others believe that delayed necrosectomy is unnecessary as long as the process remains sterile.54
Surgical Débridement for Pancreatic Necrosis
Surgical methods for treatment of necrosis vary. There are three main types of surgical débridement: conventional drainage, open or semiopen procedures, or closed procedures.45 Conventional drainage involves necrosectomy with placement of standard surgical drains and reoperation on demand (fever, leukocytosis, lack of improvement by imaging studies). Open or semiopen management employs necrosectomy and either scheduled repeat laparotomies or open packing that leaves the abdominal wound exposed for frequent dressing changes. Closed management involves necrosectomy with extensive intraoperative lavage of the pancreatic bed. The abdomen is closed over large-bore drains for continuous high-volume postoperative lavage of the lesser sac.
Most surgeons have abandoned the conventional surgical approach of débridement because inadequately removed necrotic tissue becomes or remains infected and results in a mortality of approximately 40%.3 In all procedures except the closed technique, multiple operations are frequently required to remove the necrotic pancreatic and peripancreatic material.3 Leaving the abdomen open avoids the need for formal laparotomies; packing may be changed in the intensive care unit. Repeated débridement and manipulation of the abdominal viscera using the open and semiopen techniques result in a high rate of postoperative local complications such as pancreatic fistulas, small and large bowel complications, and bleeding from the pancreatic bed. Pancreatic and gastrointestinal tract fistulas occur in 41% of patients after surgical necrosectomy and often require additional surgery for closure.55,56 The mortality using open or closed techniques is approximately 20%.3
Percutaneous (Interventional Radiology) Therapy
Successful percutaneous therapy for infected acute necrotizing pancreatitis has been described using large-bore percutaneous catheters up to 28-Fr diameter in conjunction with aggressive irrigation.57 At a mean of 9 days after hospital admission for necrotizing pancreatitis with medically uncontrolled sepsis, 34 patients had percutaneous drainage and irrigation catheters inserted into the pancreatic collection. An average of three separate catheter sites per patient and four catheter exchanges per patient was necessary for removal of necrotic material. Pancreatic surgery was completely avoided in 16 patients (47%). Control of sepsis with delayed elective surgery for repair of external pancreatic fistulas related to catheter placement was achieved in nine patients. Nine patients required immediate surgery for failure of percutaneous therapy. The mortality was 12% in these ill patients, many of whom had multisystem organ failure. In a similar fashion, Echenique and associates58 described successful percutaneous drainage of necrosis in 20 patients with documented necrosis. Solid debris was removed percutaneously using basket extraction techniques. Similar results have been obtained by other authors, although percutaneous therapy is most often used to improve immediate sepsis and delay necrosectomy.59,60 Percutaneous access has also been used to allow subsequent passage of rigid endoscopes for débridement of necrosis.61
Flexible Endoscopic Therapy
Successful endoscopic drainage of symptomatic sterile or infected pancreatic necrosis several weeks after the onset of severe necrotizing pancreatitis has been described.62 The initial descriptions used transmural placement (transgastric or transduodenal) of internal 10-Fr diameter drainage catheters plus a 7-Fr nasopancreatic irrigation tube into the retroperitoneum. The catheters are placed through a tract dilated up to 20 mm (Fig. 47.3). With this method, solid debris flows around the catheters through the transenteric tract. Complete nonsurgical resolution has been achieved in 84% of patients with this form of late, or “organized” pancreatic necrosis.63,64
Complications of endoscopic therapy (described in more detail in the section on peripancreatic fluid collections) include perforation, bleeding, and infection. Adjuvant percutaneous drains using this technique are often required to drain peripheral collections away from the body of the pancreas. Beginning in 2005, the flexible endoscopic approach evolved to incorporate direct endoscopic débridement.65 Rather than relying on irrigation to débride necrotic material, the transmural entry site is dilated to a much larger diameter to allow passage of an upper endoscope directly inside the necrotic cavity to allow “direct” débridement (Fig. 47.4