Acute pancreatitis

Published on 11/04/2015 by admin

Filed under Surgery

Last modified 22/04/2025

Print this page

This article have been viewed 1454 times

Acute pancreatitis

Colin J. McKay, Euan J. Dickson and C. Ross Carter

General description

Acute pancreatitis is a common cause for emergency hospital admission, with approximately 40 cases per year for each 100 000 population in Scotland,¹ Norway² and Sweden.³ There has been a steady increase in the incidence and a slight reduction in case mortality, although not population mortality, over the past 45 years.⁴ In approximately 80% of patients, acute pancreatitis is a rapidly-resolving condition requiring little more than analgesia and a short period of intravenous fluid resuscitation, with the remainder developing a multisystem illness characterised by a systemic inflammatory response with a variable degree of organ dysfunction.

Pathophysiology

The mechanism by which gallstones trigger an attack of acute pancreatitis has not been clearly defined. The bile reflux theory, proposed by Opie in 1901, suggested that obstruction of the common bile duct/pancreatic duct common channel by a gallstone caused reflux of bile into the pancreatic duct resulting in acute pancreatitis. While there is no doubt that passage of and at least transient obstruction by a gallstone is the initial step in biliary acute pancreatitis, there is little evidence that bile reflux is involved.

Experimental models have shed some light on the mechanism by which pancreatic duct obstruction induces acute pancreatitis. Changes in the pattern of enzyme secretion within pancreatic acinar cells, coupled with intracellular zymogen activation, are considered the important early events in the development of acute pancreatitis. Disruption of the paracellular barrier allows release of pancreatic enzymes into the paracellular space. Inflammatory cells and inflammatory mediators may further exacerbate the acinar cell injury.⁴ Research has shown that many of these early events can be triggered by an increase in intracellular calcium.⁵

The mechanism of alcohol-induced acute pancreatitis is less clear, but alcohol has been shown to increase the sensitivity of acinar cells to cholecystokinin hyperstimulation, resulting in enhanced intracellular protease activation.⁶ Alcohol also influences acinar cell calcium homeostasis, but several alternative theories have been proposed.

Natural history

Acute pancreatitis varies from a mild, self-limiting attack to a severe life-threatening illness and, for this reason, patients are often classified as having either mild or severe acute pancreatitis (see Box 13.1). This rather simplistic description ignores the wide variety of clinical behaviour that can be observed in these patients but helps to focus attention on the subgroup of patients who develop complications. Currently, the internationally accepted classification of acute pancreatitis and its complications is set out in the paper arising from the Atlanta Conference.⁷ Improved understanding of treatment concepts and the dynamic nature of the pathophysiology has rendered a number of the concepts outlined in the Atlanta Conference outdated and a revision has recently been published.⁸

Box 13.1 Definitions

Acute pancreatitis

Acute pancreatitis is an acute inflammatory process of the pancreas, with variable involvement of other regional tissues or remote organ systems.

Mild acute pancreatitis

Mild acute pancreatitis is associated with minimal organ dysfunction and an uneventful recovery. The predominant feature is interstitial oedema of the gland.

Severe acute pancreatitis

Severe acute pancreatitis is associated with organ failure and/or local complication such as necrosis (with infection), pseudocyst or abscess. Most often this is an expression of the development of pancreatic necrosis, although patients with oedematous pancreatitis may manifest clinical features of a severe attack.

Systemic inflammatory response syndrome

Response to a variety of severe clinical insults, manifested by two or more of the following conditions:

• Temperature > 38 or < 360 °C

• Heart rate > 90 beats/min

• Respiratory rate > 20/min or PaCO₂ < 32 mmHg (< 4.3 kPa)

• White blood cell count (WBC) > 12 000 cells/mm³, < 4000 cells/mm³ or > 10% immature (band) forms

Multiple organ dysfunction syndrome

Presence of altered organ function in an acutely ill patient such that homeostasis cannot be maintained without intervention.

An acute pseudocyst is a collection of pancreatic juice enclosed in a wall of fibrous or granulation tissue that arises following an attack of acute pancreatitis. Formation of a pseudocyst requires 4 or more weeks from the onset of acute pancreatitis.

A pancreatic abscess is a circumscribed intra-abdominal collection of pus, usually in proximity to the pancreas, containing little or no pancreatic necrosis, which arises as a consequence of acute pancreatitis.

Within this framework different patterns of disease have emerged. Multicentre trials in acute pancreatitis have enabled prospective study of severe acute pancreatitis and several important points have emerged. Firstly, the majority of patients who develop severe acute pancreatitis have evidence of early systemic organ dysfunction.⁹ It is exceptional for a patient to have no evidence of organ failure in the first week of illness and to subsequently develop a significant late local complication. Secondly, most patients who develop organ failure have evidence of this at the time of admission or very shortly thereafter.¹⁰ Thirdly, while the tendency is for early organ dysfunction to recover without further problems, worsening organ failure is associated with a high mortality.⁹^,¹¹^,¹²

These observations have important implications for patient management. The presence of early organ dysfunction identifies a high-risk group of patients who merit close observation for both early and late clinical complications. In particular, deteriorating organ failure carries a high mortality and should prompt early involvement of the intensive care team and consideration of transfer to a specialist unit if possible. The fact that organ dysfunction is present at, or shortly after, admission in the majority of patients in whom it develops means that efforts should be directed at early recognition of this rather than employing prediction systems of disease severity.

The majority of patients with severe early organ dysfunction will have pancreatic necrosis on computed tomography (CT) scan. A significant proportion (30–40%) of patients with pancreatic necrosis will develop secondary pancreatic infection, usually in the second to third week after admission,¹³ which may be associated with a deterioration in organ failure. Patients who have infected pancreatic necrosis complicated by multiple organ failure represent a formidable management challenge.

Diagnosis

In the majority of patients the diagnosis of acute pancreatitis is relatively easy, characterised by a clinical presentation of sudden severe epigastric pain radiating through to the back. Vomiting within the first 24 hours is very frequently severe and contributes to dehydration. Other signs and symptoms such as tachycardia, tachypnoea and circulatory collapse are dependent on the severity of the attack. A raised serum amylase (at least three times the upper limit of normal) supports the diagnosis in >95% of cases. Serum amylase estimation may be inaccurate in association with hyperlipidaemia, where a raised urinary amylase can be diagnostic. Serum lipase may be marginally more accurate but is not commonly available in routine clinical practice. CT can confirm the diagnosis where doubt exists, or in patients with delayed presentation, and it should therefore be very uncommon for the diagnosis to be made at laparotomy (Fig. 13.1).

Figure 13.1 Extensive fat necrosis in the infracolic compartment in a patient with severe acute pancreatitis.

Aetiology

Obstructive factors

Genetic defects

Genetic familial defects of the cationic trypsinogen gene ¹⁴ (N29I, RII7H) and the cystic fibrosis gene (CTFR) may be associated with recurrent pancreatitis, but severe acute inflammatory changes are uncommon.

Trauma

Hyperamylasaemia may occur after blunt abdominal trauma, usually from a crush injury to the body of the pancreas against the vertebral column, and is suggestive of pancreatic injury. Investigation is by contrast-enhanced CT to determine the extent of pancreatic and associated visceral injury.

Iatrogenic causes

Hyperamylasaemia may follow surgical or endoscopic procedures on the pancreas, and is usually self-limiting. The risk increases following a therapeutic endoscopic retrograde cholangiopancreatography (ERCP; 3%), especially when a sphincterotomy has been performed. Where a patient has significant symptoms or clinical signs, the possibility of iatrogenic duodenal perforation should be considered and where necessary excluded by CT.

Drug-induced acute pancreatitis may occur following ingestion of a number of drugs;¹⁵ those most commonly implicated are valproic acid, azathioprine, L-asparaginase and corticosteroids. However, unless gallstone disease has been excluded with confidence it is unwise to ascribe acute pancreatitis to a particular drug. Repeat exposure to the same drug again causing acute pancreatitis is the strongest evidence of a direct association.

Inflammatory

Autoimmune pancreatitis is a rare condition, considered part of the IgG4-related autoimmune disease spectrum.¹⁶ This presents as abdominal pain associated with homogeneous gland enlargement with a well-defined edge on CT, an increased IgG4/IgG ratio and a periductal lymphoplasmocytic infiltrate on biopsy. This may also be associated with abnormalities in the extrahepatic biliary tree resembling sclerosis cholangitis and a response to steroids is diagnostic. Focal autoimmune pancreatitis may prove difficult to differentiate from carcinoma. There are established associations with other autoimmune diseases (polyarteritis nodosa, systemic lupus erythematosus, vasculitis) and inflammatory bowel disease (Crohn’s and ulcerative colitis), and many are now considered part of the autoimmune spectrum, although only a small proportion appear to have an association with IgG4 serum or tissue abnormalities.

Physiological

Sphincter manometric abnormalities

Type 1 pancreatic sphincter dysfunction ¹⁷ may be associated with hyperamylasaemia and abnormalities on sphincter manometry, as part of a global gut dysmotility spectrum. Managament of sphincter spasm may only partly resolve the patient’s symptoms. Conventional treatment involves endoscopic sphincterotomy, but the risk of post-ERCP pancreatitis in these patients is high (30%).

Assessment of severity

The dynamic nature of organ dysfunction in patients presenting with acute pancreatitis has been well described,⁹ and for over 30 years authors have explored ways of ‘predicting’ those patients with more severe disease. Overall mortality, whether early or late, is also associated with the development and persistence of organ failure.¹⁸ This was indirectly shown, if not recognised, 25 years previously with the development of the predictive multifactorial scoring systems – Ranson,¹⁹ Glasgow²⁰ and APACHE II²¹ – which, rather than predicting the subsequent development of organ failure, more accurately identified established multisystem organ dysfunction. Their principal use is to remind the inexperienced of the multisystem nature of the disease process, or as a method of stratifying patients within a study protocol. Of the multiple factor scoring systems, APACHE II provides the best prediction of mortality but the mainstay of assessment remains repeated, careful clinical observation.²²

Single biochemical measures

Many studies have attempted to find a single biochemical or clinical marker that would allow adequate prediction of severity without the need for cumbersome scoring systems. In addition, the need for an objective marker of severity at the point of hospital admission is well recognised and would greatly facilitate the entry of patients with severe acute pancreatitis into clinical studies.

C-reactive protein (CRP)

The most widely studied single predictive marker is serum CRP. The major advantage of CRP is its routine availability in clinical practice. Patients with clinically severe pancreatitis usually have a CRP > 200 mg/L, with a practical cut-off being 150 mg/L, but its serum peak is not reached for 36–48 h. The positive predictive value of CRP is similar to that of APACHE II ²² but its major use is in monitoring the clinical course during the acute and recovery phase.

Other single predictive markers

Peak levels of interleukin-6 (IL-6) occur within 24 h and this has aroused interest in its use as a predictor of outcome. IL-6 is a pro-inflammatory cytokine induced by stimuli such as tumour necrosis factor (TNF) and interleukin-1 (IL-1). Other single predictive markers that have been studied include trypsinogen activation peptide (TAP), leucocyte polymorph neutrophil (PMN) elastase, TNF and serum procalcitonin. Procalcitonin is the precursor of the hormone calcitonin and is raised in the presence of an inflammatory response, particularly where this is bacterial in origin. Serum procalcitonin is a promising marker of both severe acute pancreatitis and of infected pancreatic necrosis in the later phase.²²^,²³

Intra-abdominal hypertension (IAH)

IAH is recognised as a contributing factor to organ dysfunction in the context of a variety of acute abdominal processes. Most of the literature to date focuses on trauma patients, but there is increasing interest in its role in patients with severe acute pancreatitis (SAP). There are data to suggest that raised intra-abdominal pressure (IAP) may be associated with disease severity,²⁴ organ failure and mortality in SAP.²⁵ There are, however, no data to suggest improved outcome following surgical decompression for raised IAP in acute pancreatitis, and indeed this may be harmful. At present the monitoring of IAP cannot be recommended outside of a clinical trial.

Repeated clinical assessment

In the absence of clinically useful predictive systems, our own practice is to monitor patients for the development of systemic organ dysfunction by repeated clinical and biochemical assessment. The presence of a systemic inflammatory response syndrome (SIRS; defined as two or more of the following: fever, tachycardia, tachypnoea or leucocytosis) identifies patients at risk of multiple organ dysfunction syndrome (MODS), particularly when three or four SIRS criteria are present or when SIRS persists for 48 h or more after admission. Patients without SIRS at admission are at very low risk. The development of systemic organ dysfunction (usually clinically manifest as hypoxaemia) mandates careful monitoring in a high-dependency or intensive care unit environment.

Imaging

Role of ultrasound (US)

All patients with acute pancreatitis should have an ultrasonic assessment of the biliary tree within 24 h of admission.²⁶

All patients with acute pancreatitis should have an ultrasonic assessment of the biliary tree within 24 h of admission.²⁶ In those with gallstones, the majority will have mild disease, and this will facilitate definitive treatment of cholelithiasis prior to discharge. In the emergency situation, ultrasonography can be difficult due to a number of factors, including the presence of intraluminal bowel gas or lack of patient cooperation. Therefore, in patients with a negative initial US, and no other obvious aetiological factor, the US should be repeated prior to discharge before excluding gallstones as a potential aetiological factor.

Role of CT

The main role of CT in the early phase of acute pancreatitis is to clarify the diagnosis in cases where there is diagnostic uncertainty. In patients with severe acute pancreatitis, particularly when complicated by MODS, early CT helps to exclude other pathology such as intestinal perforation, gut ischaemia or dissecting aortic aneurysm. Dynamic contrast-enhanced CT can also be used to assess disease severity (Fig. 13.2) and predict the potential for complications.²⁷ Although not widely used for this purpose clinically in the UK, it is routinely used in many other countries, and can be useful for stratification of patients in clinical trials. The CT Severity Index (CTSI)²⁷ combines a score for the radiological pancreatic and peripancreatic abnormalities with a weighting for the extent of necrosis. More recently there are reports of perfusion CT, which may be a useful modality for detecting early, subclinical ischaemic changes in the pancreas that then lead to pancreatic necrosis.

Figure 13.2 CT scan showing acute pancreatitis with necrosis (a) and retroperitoneal gas (b).

Role of magnetic resonance (MR)/magnetic resonance cholangiopancreatography (MRCP)

Magnetic resonance imaging (MRI) offers a realistic alternative to contrast-enhanced CT ²⁸ in the assessment of patients with acute pancreatitis. The avoidance of cumulative radiation exposure, potentially nephrotoxic iodinated contrast media, and the excellent contrast sensitivity and spatial resolution would make it an attractive alternative. Axial T1- and T2-weighted scans produce images analogous to those of CT. Gadolinium contrast enhancement can infer viability and improve anatomical definition. Heavily T2-weighted image acquisition, using a single breath hold and long repetition (TR) and echo (TE) times, results in little signal being produced by solid tissue, and a high signal from static fluid in the biliary and pancreatic ducts, enabling images anatomically comparable to those of ERCP to be produced.

Whilst technically feasible in most centres, the MR environment is unsuitable for patients requiring significant circulatory or respiratory support, and at present few centres have the capability to perform MR-guided intervention. Contrast-enhanced CT therefore remains the imaging modality of choice for assessment and intervention in severe acute pancreatitis. However, MR has a role in the follow-up of acute inflammatory collections, where it is superior to CT in determining the extent of solid material within a collection (Fig. 13.3), and in excluding choledocholithiasis in selected patients.

Figure 13.3 MRI scan showing extensive necrosis (arrowed) within a post-acute fluid collection (FC).

Endoscopic ultrasound (EUS)

EUS has rapidly emerged as an important tool in diagnostic and therapeutic algorithms for patients with acute pancreatitis. Its main areas of use are in the diagnosis of microlithiasis in patients with idiopathic pancreatitis and for linear EUS-guided drainage of peripancreatic collections, as discussed in the relevant sections below.

Management

Initial management

Management of acute pancreatitis in the UK is still based on two key guideline documents, now 7 ²⁶ and 10²⁹ years old, respectively. The initial management of patients presenting with acute pancreatitis should be directed at the early identification and management of organ failure, most frequently renal and respiratory dysfunction.

The initial management of patients presenting with acute pancreatitis should be directed at the early identification and management of organ failure, most frequently renal and respiratory dysfunction.

At present there are no established end-points of resuscitation to confirm that tissue perfusion and oxygen delivery have been restored adequately in patients with acute pancreatitis. Aggressive fluid resuscitation is often required, and monitoring the response to this relies upon traditional markers, in particular urine output, blood pressure and pulse oximetry. Physiological markers of resuscitation, for example acid–base balance from an arterial blood gas, may be helpful in detecting clinically occult hypoperfusion. Patients who do not respond to initial resuscitation, or who have evidence of organ dysfunction, should be transferred to a critical care environment for more invasive and intensive monitoring with central venous and arterial catheters. Respiratory failure should be treated with humidified oxygen, and this will be guided by continuous pulse oximetry and arterial blood gas analysis. Deteriorating respiratory function is a sign of disease progression. As many as half of all deaths from acute pancreatitis occur in less than 7 days, and the majority of these occur within 72 h of admission.¹ There is evidence that patients managed in specialist institutions have a reduced risk of early death and this may be an indication that management of early MODS could be improved.

Supportive management

All patients with severe pancreatitis should be managed within a high-dependency/intensive care environment with the potential for organ support.²⁶

Where possible, patients with severe acute pancreatitis should be managed by a designated multidisciplinary team who have an interest in pancreaticobiliary pathology. Facilities should be available for patients to undergo ERCP/sphincterotomy when indicated (see below). Management of these patients is complex and should be discussed with a specialist unit at an early stage. Specific measures will be determined by the clinical situation. However, therapeutic interventions are aimed at restoring tissue perfusion as rapidly as possible. This is achieved initially with volume resuscitation titrated to haemodynamic and physiological response, rather than the early use of vasopressors or inotropes in patients who are still intravascularly deplete. Acute dialysis for acidosis has not been shown to improve outcome but is required for established renal failure. Currently there is no specific therapy to reverse respiratory failure other than ventilatory support.

Specific medical management

There have been many attempts to introduce specific medical treatments for acute pancreatitis and these broadly fall into the following categories.

Prevention of infection

In patients who survive the early, systemic complications of acute pancreatitis, secondary infection of pancreatic necrosis is the most important late complication. Infection occurs in 30–40% of patients with a minimum 30% pancreatic necrosis ¹³ and is responsible for the majority of late deaths from acute pancreatitis. Secondary infection manifests as escalating sepsis or a deterioration in organ failure scores, usually in the second (36%) and third (71%) weeks of the illness.³⁰

The role of prophylactic antibiotics to prevent secondary infection has been widely studied. The most recent Cochrane review ³¹ found no evidence of a significant reduction in mortality with antibiotic prophylaxis and no difference in the incidence of infected pancreatic necrosis. Even non-pancreatic infections showed no significant difference with antibiotic therapy. It was noted, however, that all of the studies were underpowered and a definitive answer to this question will require better quality clinical trials. The most recent meta-analysis of 14 trials including 841 patients also found no evidence of benefit with antibiotic prophylaxis.³²

Prophylactic antibiotics should not be used in the management of acute pancreatitis.³¹^,³²

Nutritional support

There are two main, and very separate, considerations in determining the mode of nutritional support. Severe acute pancreatitis is often a prolonged and profoundly catabolic illness, and there is no doubt that, throughout the illness, nutritional integrity should be maintained – the question in these patients is not whether nutritional support is necessary, but rather how it can be best administered.

The second consideration relates to the potential of modulating the disease process by the mode of delivery, either through maintenance of host defences or through the use of immunomodulating feeds. The first issue is a practical problem faced by clinicians on a daily basis, the second remains somewhat speculative, with interesting but inconclusive evidence so far. These will be dealt with separately.

Nutritional delivery in the patient with acute pancreatitis: The key study in this regard was the randomised study of Kalfarentzos et al. in 1996, who randomised 38 patients with severe acute pancreatitis to total parenteral nutrition (TPN) or nasojejunal feeding.³³ The most recent Cochrane review, including eight randomised trials, demonstrated a reduction in mortality, systemic complications and surgical intervention in patients given enteral nutrition.³⁴

Most experience to date has been with enteral feeding distal to the duodenojejunal flexure. More recently, four randomised studies have shown nasogastric feeding to be a practical alternative to jejunal feeding.³⁵ All studies in this area are underpowered and it is therefore difficult to recommend this feeding route in routine clinical practice.

It is important to recognise that there are situations where parenteral nutrition must be considered, such as where complex fistulas develop, and sometimes a combination of enteral and parenteral routes is required. Combined feeding is most commonly required when enteral feed is not adequately absorbed, leading to intractable diarrhoea and fluid losses.

Nutritional support should be by the enteral route where possible.³⁴

Disease modulation through content or mode of delivery: There has been interest in the role of the intestine in the pathophysiology of multiple organ failure in critical illness, with loss of gut barrier function potentially leading to endotoxaemia and the systemic inflammatory response syndrome (SIRS). In a small study from Leeds,³⁶ the authors reported a reduction in the inflammatory response and organ failure in those receiving enteral support, but unfortunately there were only 13 patients with severe disease, limiting the validity of the conclusions. There have been several trials comparing so-called ‘immunonutrition’ with standard enteral feeding in critically ill patients, but so far no evidence of benefit has been demonstrated in acute pancreatitis.³⁷ Similarly, there has been interest in the role of ‘probiotics’, but a randomised trial from the Netherlands found an increase in fatal complications in the probiotic group, with an unexpectedly high incidence of intestinal necrosis.³⁸

Other medical therapies

Inhibition of pancreatic secretion: Pharmacological attempts to suppress pancreatic function have included intravenous glucagon, somatostatin and, more recently, the somatostatin analogue octreotide. On the basis of the available literature there is no justification for the use of octreotide or any other pharmacological inhibitor of pancreatic secretion in acute pancreatitis.

Inhibition of pancreatic enzymes: Many studies have evaluated the concept of supporting the endogenous antiprotease defence mechanisms. Randomised trials of intravenous aprotinin (Trasylol), gabexatemesilate, intraperitoneal aprotinin, and both low- and high-dose fresh frozen plasma (FFP) have shown no therapeutic benefit.

Inhibition of the inflammatory response: Following initial promising results with the platelet-activating factor antagonist, lexipafant, a multicentre, randomised, placebo-controlled study of anticytokine therapy recruited 1518 patients. This study recruited only those patients with symptoms of less than 48 h duration and was restricted to those with predicted severe attacks. Not only was there no difference in mortality between groups, the incidence of local complications, length of ICU stay, hospital stay and change in organ failure scores were all similar in the three study groups.

The potential for other agents that modify the inflammatory response to influence outcome in acute pancreatitis has only been assessed in experimental models.

Role of ERCP: There have now been three published randomised trials addressing this issue, and four smaller studies. Contradicting an earlier Cochrane review,³⁹ the most recent meta-analysis⁴⁰ has shown early ERCP in patients with either predicted mild or severe acute biliary pancreatitis without acute cholangitis did not lead to a significant reduction in the risk of overall complications and mortality. There is no role for urgent ERCP in patients with mild disease. All patients with jaundice who exhibit signs of sepsis should undergo urgent ERCP and sphincterotomy as cholangitis may coexist with acute pancreatitis and hyperamylasaemia, but there appears to be little role for early ERCP outside this scenario.

In the non-jaundiced patient there is no role for urgent ERCP and sphincterotomy.⁴⁰

Definitive management issues

The definitive management issues may be considered as, firstly, those designed to prevent further attacks once a mild attack has subsided, and secondly those specifically related to the management of early and late complications.

Prevention of recurrent acute pancreatitis

Management of gallstones

The timing of cholecystectomy will obviously depend on the clinical situation. In patients recovering from mild acute biliary pancreatitis, definitive management of the gallstones to prevent a further attack should ideally be achieved during the same admission, and certainly no later than 4 weeks following discharge from hospital.²⁶ This will normally involve a cholecystectomy (laparoscopic or open) with intraoperative cholangiography, or alternatively duct imaging by MRCP (or EUS) followed by cholecystectomy. Elderly patients or those with significant medical comorbidity may be managed by an endoscopic sphincterotomy, although this may not be as effective as definitive surgery in preventing further attacks.

In severe acute pancreatitis, interval cholecystectomy should be performed when the inflammatory process has subsided and the procedure is potentially easier.²⁹

Definitive management of cholelithiasis in uncomplicated acute pancreatitis should be achieved within 4 weeks of discharge from hospital.

Investigation of non-gallstone-associated pancreatitis

Following resolution of an attack of acute pancreatitis, an assessment of potential aetiological factors is an important aspect of care, and a diagnosis of idiopathic pancreatitis should be made in less than 20% of patients.²⁶ Evaluation of the initial acute attack should include an adequate history (alcohol/drugs/familial), biochemical tests (liver function tests/lipids (hypertriglyceridaemia)/calcium) and biliary ultrasound. If these investigations are normal, axial imaging (CT or MR/MRCP) may be appropriate to exclude a mechanical cause. Patients in whom no cause is identified following these investigations should be considered for EUS, as a cause will be identified in the majority of cases.⁴¹ A cholecystectomy or biliary sphincterotomy is justified in patients with recurrent idiopathic pancreatitis in whom microlithiasis or biliary sludge is identified. With the increasing use of EUS in these patients, it is increasingly recognised that many of these patients will have changes consistent with early chronic pancreatitis,⁴¹ rather than biliary microlithiasis as suggested by earlier studies. The prevalence of microlithiasis appears to be higher in regions where gallstone disease is the predominant aetiology.

Peripancreatic fluid collections

Management of an early fluid collection

In the early stages of acute pancreatitis, up to 25% of patients with acute pancreatitis will develop a fluid collection in the peripancreatic area identifiable on CT. In themselves, these collections are of little significance and require no intervention. There are significant risks associated with aspiration and especially external drainage, in particular the development of secondary infection, fistula and recurrence, and they therefore cannot be recommended.

Aspiration or drainage of sterile acute fluid collections should be discouraged.

Management of a pseudocyst

Management of pseudocysts is determined by an understanding of the anatomy (based on CT), the degree of necrosis (MR/EUS) and the clinical condition of the patient. As a general rule, definitive management should be delayed until all organ dysfunction has resolved and can often be performed simultaneously with management of cholelithiasis (see above).

In our experience, it is helpful to characterise pseudocysts associated with acute pancreatitis as either fluid predominant or necrosis predominant.

In each case the collection may be sterile or infected and associated with varying degrees of systemic disturbance. The size of the collection and its relationship to adjacent structures, in particular the stomach, are also important factors when considering treatment options.

Asymptomatic pseudocysts do not require treatment, and many will eventually resolve spontaneously. Acute pseudocysts are most commonly retrogastric, and may or may not link to a disrupted pancreatic duct. Three-quarters will be associated with a mild to moderate hyperamylasaemia. In symptomatic cysts, a conservative policy may be warranted for up to 12 weeks from the onset of acute pancreatitis. This policy is not, however, without risk as pseudocyst rupture, bleeding or abscess formation may occur. The likelihood of resolution is related, at least in part, to pseudocyst size. Should conservative treatment fail, the options are percutaneous, endoscopic or surgical drainage.

Percutaneous drainage: Results of percutaneous drainage suggest wide variation in success (40–96%)⁴² and the introduction of infection is a risk that must be considered. In practical terms, the risk of pancreatic fistula limits this approach and there is evidence that it can make subsequent surgical intervention more hazardous if this becomes necessary.⁴³ In our practice we restrict the use of percutaneous drainage to occasional patients with infected, fluid-predominant collections, particularly where there is a degree of systemic organ dysfunction. Increasingly, however, endoscopic or laparoscopic drainage is employed.

Endoscopic drainage: The technique of endoscopic cystgastrostomy as first described by Baron et al.,⁴⁴ initially by blind puncture of a cyst bulging into the stomach wall using a side-viewing endoscope, has been subsequently refined using endoscopic ultrasound guidance. EUS guidance enables drainage of cysts where no intraluminal bulge is present and also helps avoid intervening vessels. Further procedures may be required to facilitate drainage, particularly where cysts are large or where there is much necrotic debris, but we have not found it necessary to pursue active endoscopic necrosectomy in this group of patients. Where disruption of the pancreatic duct has occurred, transpapillary duct stenting can aid resolution. Baron and colleagues have recently updated their experience in 104 patients with walled-off pancreatic necrosis with successful resolution in 95 patients (91%). The mean time to intervention was 63 days and mean duration of treatment was 4.1 months. The complication rate, mainly haemorrhage and perforation, was 14%.⁴⁵

Surgical drainage of an acute post-inflammatory collection: Surgery may be considered as a primary mode of intervention in selected patients, particularly for patients with large, necrosis-predominant collections without infection or systemic organ dysfunction. This can be readily achieved by a laparoscopic transgastric procedure, or a direct cyst-enterostomy, and allows simultaneous laparoscopic cholecystectomy where appropriate. In those patients in whom endoscopic drainage fails to achieve complete resolution, or a cyst recurs, simple surgical drainage is rarely an option as this often results from separation of the head/body and tail of the pancreas due to prior necrosis of the central pancreas – termed a ‘disconnected tail’. These patients often require a challenging distal pancreatico-splenectomy.

Symptomatic or persistent pseudocysts following acute pancreatitis should be managed in a specialist unit where the full range of interventional procedures is available.

Failure of percutaneous or endoscopic management is associated with the need for complex surgery and should therefore only be undertaken by or following consultation with a pancreatic specialist.⁴³

Management of a pancreatic duct fistula

This complication most commonly follows prior intervention for an acute post-inflammatory collection or infected necrosis, and manifests as persistent drainage of amylase-rich opalescent fluid, in the absence of significant sepsis. Management is similar to that of a communicating pseudocyst, initially by transpapillary stenting where possible. Intraperitoneal rupture of a pseudocyst can result in pancreatic ascites or pleural effusion. More invasive management of a persistent fistula, either inaccessible or failing to respond to ductal stenting, should be delayed by percutaneous or endoscopic control, until the patient has made a full recovery, and again often requires surgical resection (distal pancreatico-splenectomy).

Management of necrosis

The management algorithm surrounding necrotising pancreatitis has altered radically in the last 15 years in response to evolving concepts, improved understanding and the development of minimally invasive techniques, including percutaneous necrosectomy and laparoscopic or EUS-guided cystgastrostomy, as an alternative to conventional open debridement. A multidisciplinary approach has evolved, and it is now common for several techniques to be utilised in an individual patient, as the indications and clinical condition of the patient alter during the course of the disease process.

The development of secondary septic complications is the usual initiator demanding invasive treatment. The choice of intervention technique is underpinned by an understanding of the dynamic evolution of post-acute, necrosis-associated collections in pancreatitis. The previously held concept that recovery would not occur until almost complete removal of necrosis had been achieved has been progressively challenged and the focus of intervention is now on the ‘adequate and maintained control of sepsis’. The success of various approaches will be dependent on the anatomical position and particularly the ratio of solid to fluid components within the collection.

The process of maturation or ‘organisation’, with separation and partial liquefaction of the solid components within a collection, takes in excess of 12 weeks to complete, during which four stages can be recognised:

1. True pancreatic necrosis – minimal separation of devitalised tissue with a high solid/fluid ratio.

2. Transitional pancreatic necrosis with partial but incomplete separation.

3. Organised pancreatic necrosis (OPN) – good separation of devitalised tissue within a fluid-filled cavity and formation of a fibrous wall lined with granulation tissue.

4. Pseudocyst – almost complete resolution of any solid component and a well-formed fibrous wall lined with granulation tissue.

The necrotic process associated with pancreatitis tends to involve both the pancreatic parenchyma and surrounding adipose tissue. Indeed, significant quantities of necrotic peripancreatic tissue can be present with an essentially viable gland. Complications relate to the extent of the necrotic process, and in particular the extent of parenchymal necrosis. Early aggressive debridement in the absence of infection has been advocated. However, mortality in this series was 25% overall, and the only randomised study of early versus late (> 12 days) necrosectomy was discontinued as a result of the mortality rate in the early treatment group (56% vs. 27%).⁴⁶ The general principle is now to withhold surgery in the early phase of disease, operating for complications ideally once the acute inflammatory insult has subsided.

There is no role for early surgical intervention other than for the management of complications.

Management of sterile necrosis

The development of retroperitoneal necrosis secondary to acute pancreatitis is not in itself an indication for intervention. Pancreatic necrosis, where sterile, can usually be adequately treated by conservative means.⁴⁷ There remains debate regarding the role of debridement in patients failing to respond to conservative treatment. Early debridement does not improve outcome; however, some specialists advocate debridement in patients with continuing organ dysfunction after several weeks, but this may also be detrimental.⁴⁸ The majority of sterile post-inflammatory fluid collections progress into organised pancreatic necrosis, which can be managed with low morbidity and mortality, and our policy is to delay intervention where possible. The management of these patients is discussed in the section dealing with fluid collections associated with necrosis.

Sterile necrosis should initially be treated conservatively where possible, allowing delayed definitive treatment.

Management of infected necrosis (early phase, 2–6 weeks)

Infected pancreatic necrosis has previously been described as the most feared surgical complication of acute pancreatitis. This led to the development of protocol-driven management in the 1990s, aimed at the early identification of secondary infection within necrosis. The presence of a persistent SIRS response made clinical differentiation between SIRS and sepsis difficult and this led to CT- or US-guided fine-needle aspiration (FNA) of the pancreatic or peripancreatic collections as a diagnostic test. FNA was considered the cornerstone of management as it was thought infected necrosis mandated radical surgical intervention.⁴⁹^,⁵⁰

Our own approach has progressed from one based on the presence or absence of infection to one based on organ dysfunction. Infected collections, even those containing gas, may be observed when the patient is clinically well and recovering with conservative treatment. Patients with profound organ dysfunction in whom we suspect secondary infection, with a drainable peripancreatic collection, will undergo percutaneous drainage aimed at sepsis control, with later staged percutaneous or open management of the necrosis as clinically appropriate, and we no longer perform diagnostic FNA. Decision-making in these patients is extremely difficult and is best carried out within an experienced multidisciplinary team.

Methods of necrosectomy: The traditional approach to infected necrosis was open laparotomy/debridement. These approaches are falling from favour with increasing evidence that minimally invasive intervention may reduce morbidity/mortality;51–53 however, they remain the method of choice in some countries.

Open laparotomy/debridement: The technique of pancreatic debridement involves a wide exposure of the abdomen, usually through a bilateral subcostal/rooftop incision. Both colonic flexures are mobilised to expose the retroperitoneum and the lesser sac entered via the gastrocolic omentum, or occasionally the transverse mesocolon. Pus is aspirated from the abscess cavity, leaving the solid component behind, which is then removed by ‘blunt finger’ dissection (Fig. 13.4). Tissue that will not come away by finger teasing should be left in situ to demarcate and subsequent removal at a later procedure. The procedure may also include a cholecystectomy, operative cholangiogram and feeding jejunostomy.

Figure 13.4 Solid necrotic material removed at open necrosectomy by blunt finger dissection.

Methods for postoperative management of the debridement cavity after laparotomy are as follows:

• With drainage/‘closed packing’. Simple drainage, often with multiple retroperitoneal tube drains, was the conventional approach to the postoperative management of the debrided pancreatic and peripancreatic bed. Whilst mortality was less than with resective procedures, multiple second-look laparotomies were often required for residual sepsis. The initial results of Warshaw and colleagues reported respectable mortality figures of 24% using this technique. Their technique has been modified using multiple soft Penrose drains containing cotton gauze to pack the cavity following completion of the necrosectomy.⁵⁴ These are subsequently removed at intervals, allowing the cavity to collapse around the drains. Their reported mortality rate using this technique is the lowest in the literature (6.2%), although the series included patients with sterile necrosis (11%) and pancreatic abscess (39%), and only 14% had both sepsis syndrome and a positive culture requiring early intervention, which is indicative of the difficulties in interpretation of the available literature.

• With open packing. Bradley and colleagues from Atlanta have been the principal proponents of the open laparostomy technique.⁵⁵ In this, at the conclusion of the debridement, the lesser sac is packed with lubricated cotton gauze and the abdomen left open, allowing planned re-explorations every few days until granulation tissue forms. Enteric fistula and secondary haemorrhage are not uncommon, and the technique is rarely performed as a first option. Surgical packing and planned re-operation is, however, sometimes required to control blood loss from the retroperitoneum following the development of an intraoperative coagulopathy, a lavage system being created, following correction of the coagulopathy, at the time of subsequent pack removal.

• With closed lavage. Postoperative closed lavage as described by Beger et al.⁵⁶ is the most popular method for postoperative sepsis control following open debridement, the aim of the lavage being the continuous removal of devitalised necrotic material and bacteria. Several (four to six) large-diameter tube drains are inserted in the lesser sac and throughout the abdomen, and the abdomen closed. Continuous lavage is then commenced, our own preference being for CAPD dialysis fluid (Dianil 7, Baxter Healthcare, potassium free, Iso-osmolar) warmed through a blood warmer and delivered at 500 mL/h. The lavage is continued, for around 3–4 weeks on average, until the return fluid is clear and the patient has no residual signs of systemic sepsis. This technique has been adopted with minor variations by centres on both sides of the Atlantic.

Minimally invasive approaches to infected necrosis: Minimally invasive surgery has been shown consistently to be associated with a lesser activation of the inflammatory response than equivalent open surgery, and there is experimental evidence suggesting that local sepsis and the inflammatory response may be lessened by a minimally invasive rather than an open technique. The widespread belief that formal necrosectomy is required has recently been challenged and there is evidence that patients can resolve following simple percutaneous drainage or following limited necrosectomy. The PANTER trial and several prospective cohort series 51–53 have suggested that by minimising the massive inflammatory ‘hit’ of open pancreatic necrosectomy, a minimally invasive approach to the management of infected pancreatic necrosis may lessen the risk of post-procedural organ failure, respiratory and wound morbidity in these patients. However, there is as yet no evidence that one minimal approach is superior to another.

• Percutaneous drainage. Freeny et al.,⁵⁷ combining aggressive CT-guided percutaneous drainage with continuous post-drainage lavage, showed that nearly half the pancreatic abscesses may resolve. However, more than half of these patients required subsequent surgical intervention for residual sepsis. Drain occlusion is common due to necrotic debris and repeated drains may be necessary. Simple drainage, even with small-diameter drains, may be associated with complete resolution and within the PANTER trial ‘step-up’ arm, 35% of patients were successfully managed by small-bore (4-mm) percutaneous drainage alone.

• Minimally invasive surgery. Simple percutaneous or endoscopic drainage alone may result in complete resolution; however, they also have a useful role providing initial sepsis control, associated with an improvement in organ dysfunction. Careful drain management is required to recognise drain blockage early and to prevent recurrent sepsis. As a result of the difficulties in maintaining drain patency, we have developed a technique ⁵⁸ to allow minimally invasive drainage, and in addition removal of the necrotic component. This involves the intraoperative dilatation of a previously placed CT-guided percutaneous drain tract and subsequent necrosectomy using a urological rigid-rod lens system (Fig. 13.5). Complete resolution of sepsis and necrosis can occur without recourse to further surgery, with a reduced need for postoperative organ support compared to the open procedures. In other centres as well as within our own patient cohort, this technique has significantly reduced mortality.⁵² The Dutch Pancreatitis study group have popularised a video-assisted retroperitonael debridement technique (VARD variation of the Fagniez technique) through a small 5-cm incision in the left flank. Their management approach has evolved from being initially performed on all patients with infected necrosis, to now being used as a ‘step-up’ approach should initial percutaneous drainage fail to control sepsis. The Dutch group have completed a randomised trial⁵¹ comparing this minimally invasive two-stage ‘step-up’ approach with open necrosectomy and have shown a reduction in early organ failure (respiratory) and late morbidity, but the study was underpowered to address mortality.

Figure 13.5 Percutaneous necrosectomy showing the rod lens scope and a guiding catheter to ensure accurate drain placement.

• Endoscopic necrosectomy. The principle of tract dilatation and minimally invasive necrosectomy has also been used with the endoscopic approach. Seifert et al.⁵⁹ have reported the dilatation of an endocyst-gastrotomy tract, allowing insertion of the endoscope into the retroperitoneum and subsequent piecemeal debridement. More recently the use of multiple transgastric cyst gastrostomy puncture sites has been reported, allowing nasocystic lavage and stent-assisted drainage through alternative drainage sites with good sepsis control.⁶⁰

There is currently no evidence that one minimally invasive technique has any advantage over another and choice is often determined by local expertise and resources.