Abstract
Acute pancreatitis (AP) is an important cause of morbidity and mortality worldwide and the annual incidence appears to be increasing. It presents as a mild self-limiting illness in 80% of patients. However, one-fifth of these develop a severe complicated life-threatening disease requiring intensive and prolonged therapeutic intervention. Alcohol and gallstone disease remain the commonest causes of AP but metabolic abnormalities, obesity and genetic susceptibility are thought be increasingly important aetiological factors. The prompt diagnosis of AP and stratification of disease severity is essential in directing rapid delivery of appropriate therapeutic measures. In this review, the range of diagnostic and prognostic assays, severity scoring systems and radiological investigations used in current clinical practice are described, highlighting their strengths and weaknesses. Increased understanding of the complex pathophysiology of AP has generated an array of new potential diagnostic assays and these are discussed. The multidisciplinary approach to management of severe pancreatitis is outlined, including areas of controversy and novel treatments.
Introduction
Acute pancreatitis (AP) affects around 40/100,000 of the Western general population annually and its incidence appears to be gradually increasing, although overall mortality has remained unchanged. 1 The disease follows a mild self-limiting course in 80% of patients but around 20% of cases are severe and associated with significant mortality and morbidity. 2 The Atlanta symposium defined severe acute pancreatitis (SAP) as attacks associated with organ failure and/or local complications such as necrosis and pseudocyst formation. 3 Death secondary to SAP tends to be bimodal with approximately 50% of deaths occurring within a week of diagnosis as a result of rapidly progressive systemic inflammatory response syndrome (SIRS). Late deaths predominantly result from septic complications of pancreatic necrosis (PN) and subsequent multi-organ failure (MOF). 4,5
Clinical presentation
Presentation is classically with right upper quadrant or epigastric pain radiating through to the back, associated with nausea and vomiting. Local signs include abdominal tenderness with or without peritonism and discolouration in the periumbilical region (Cullen's sign) and flanks (Grey Turner's sign) due to tracking of inflammatory exudates from the peripancreatic region. Patients variably present with, or develop, signs of organ dysfunction such as respiratory distress, hypotension, oliguria and intestinal ileus. In addition, they are often profoundly dehydrated as a result of extensive third space fluid loss from the inflamed pancreas. Early clinical investigations may demonstrate hypoxia, deranged liver function tests (LFTs), elevated serum urea and creatinine concentrations, hyperglycaemia, hypoalbuminaemia, leukocytosis and hypocalcaemia. Hypocalcaemia in AP relates to saponification with free fatty acids, hypoalbuminaemia, hypomagnesaemia, increased calcitonin release and abrogated parathyroid hormone release.
In mild pancreatitis prompt resolution of local symptoms is observed and patients can usually be discharged from hospital within a few days. The clinical picture of severe pancreatitis results from two key underlying pathologies: MOF and PN. In a study examining the incidence of organ failure in SAP, 61.7% of patients developed MOF. 6 Single organ failure occurred in 26.6% patients including respiratory (35.1%), cardiovascular (22.3%), gastrointestinal (19.1%), hepatic (15.9%) and renal (14.9%). This finding is crucial as persistence of organ failure in early AP is an independent risk factor for mortality. 7 Pulmonary complications are particularly common throughout the course of SAP affecting up to half of patients and associated with significant mortality. The underlying pathology includes atelectasis, pulmonary oedema, pneumonia and adult respiratory distress syndrome. 8 Lung injury is mediated by an array of systemic factors including trypsin, cytokines, free fatty acids, platelet activating factor (PAF), nitric oxide and substance P. 9
PN occurs in 20% of patients with AP and directly correlates with disease severity, increased risk of organ failure and increased mortality. 10,11 In patients with PN compared with those without, the reported risk of organ failure is 50% versus 5% and mortality is 17% versus 3%, respectively. 12 The action of proteolytic and lysosomal enzymes, reactive oxygen species (ROS), vasoconstrictors and inflammatory-mediated tissue injury contribute to the development of necrosis. 13 Sterile PN is associated initially with SIRS and subsequently with complications such as pseudocyst formation, pancreatic fistulation and pancreatic insufficiency. However, the most important complication of PN is infected pancreatic necrosis (IPN) where mortality increases from 10% to 25%. 14 The pathogenesis of IPN remains unclear but bacterial translocation from the gut is thought to be the most important source. 15 SIRS, impaired gut motility, mucosal hypoxia and immune dysfunction may all contribute to bacterial translocation in AP. Overall the development of sepsis is a key prognostic factor with 60–80% of all AP mortality related to septic complications. 5
Pathophysiology
The central process in AP is the intracellular activation of trypsinogen leading to pancreatic autolysis. This results from a dysequilibrium between pancreatic over-stimulation and several intrinsic protective mechanisms against this, which include compartmentalization of lysosomal enzymes and digestive zymogens, trypsin autolysis, serine protease inhibitors and intracellular calcium homeostasis. 16 Pancreatic over-stimulation or injury results in the pathological co-localization of lysosomal enzymes, such as cathepsin B, with exocrine enzymes leading to the accumulation of activated intracellular trypsin. 17 Protection against prematurely activated trypsinogen is provided by serine protease inhibitors, such as pancreatic secretory trypsin inhibitor and protease-activated receptor-2, which is activated by trypsin and inhibits trypsinogen cleavage. 18 A key early event in AP is thought to be when these defences are overwhelmed by the extent of trypsin activation. Activated trypsin within pancreatic parenchyma leads to tissue destruction and release of further enzymes from the damaged cells generating a vicious cycle. Trypsin also activates other proteases such as elastase and components of the complement, kinin and coagulation cascades. Additionally, trypsin stimulates the local generation of ROS and upregulates cytokine production by both macrophages and acinar cells. 19,20
In addition to the direct action of pancreatic proteases, injured pancreatic and peripancreatic tissue stimulate a profound acute inflammatory response. Furthermore, pancreatic parenchymal cells, particularly acinar and stellate cells, play a direct role in leukocyte chemoattraction and migration through secretion of chemokines. 21 Early and marked infiltration of the pancreas with monocytes, macrophages and polymorphonucleated (PMN) cells leads to further tissue injury and generation of cytokines, ROS and arachidonic acid metabolites such as PAF. 22 Interleukin (IL) 1, 6, 8, 18 and tumour necrosis factor (TNFα) appear to be the key cytokines mediating the inflammatory response in AP. 23,24 TNFα in particular is the central early cytokine released by both immune and acinar cells stimulating further cytokine production, chemotaxis, cell death, endothelial activation and pro-inflammatory gene expression. TNFα also contributes to the oxidative response in the systemic component of AP by upregulating nitric oxide synthase activity and modification of xanthine dehydrogenase into xanthine oxidase. 25
Microcirculatory dysfunction rapidly develops within the inflamed pancreas, mediated by pancreatic proteases, lysosomal enzymes, leukocytes and vasoactive factors such as endothelin-1 and phospholipase A2. 26,27 The resultant ischaemia causes local pancreatic oedema and necrosis, exacerbates the systemic inflammatory response and leads to ischaemia–reperfusion injury and generation of ROS. 28 ROS cause direct acinar cell injury, upregulate apoptosis, inactivate protease inhibitors and activate nuclear factor-κB (NF-κB), which regulates gene expression for numerous proinflammatory mediators including cytokines, selectins and adhesion molecules. 29
The accelerated systemic response observed in AP is mediated by several factors generated by the inflamed pancreas. Cytokines are central to systemic disease as those released from the pancreas are powerful stimulants of Kupffer cells, the resident hepatic macrophages, amplifying pro-inflammatory signalling. 30 In particular, TNFα overexpression is directly associated with secondary lung injury by upregulating pulmonary adhesion molecules, neutrophil sequestration and liver injury by inducing hepatocyte apoptosis. 31,32 Adhesion molecules such as intercellular adhesion molecule-1 (ICAM-1) and matrix metallopeptidases, such as matrix metallopeptidase-9 (MMP-9), appear to be particularly important in secondary lung injury by increasing permeability of vascular basement membranes and facilitating neutrophil endothelial adhesion and translocation. 33,34 ROS clearly mediate distant organ dysfunction and the extent of oxidative stress correlates well with the severity of AP. Furthermore, a glutathione S-transferase genetic polymorphism that results in diminished antioxidant activity, is associated with high rates of progression of mild pancreatitis to severe disease. 35 Mast cells are thought to play a particularly important role in multi-organ dysfunction syndrome (MODS) associated with AP by mediating significant systemic endothelial barrier dysfunction. 36
Aetiology
The causes of AP are numerous and varied (Table 1). The majority of adult cases in developed countries are due to gallstones and alcohol, gallstones being more common in women and alcohol in men. 1 Less common causes include hyperlipidaemia, hypercalcaemia and post-endoscopic retrograde cholangiopancreatography (ERCP). The leading causes of AP in children are trauma, systemic diseases, infection and drugs. 37 In addition to traditional causes the impact of co-factors such as obesity and genetic susceptibility is thought to be increasingly important. 38 These co-factors can increase the risk of developing AP and have also been shown to increase the severity of the disease and the incidence of secondary complications. 39
Causes of acute pancreatitis
ERCP, endoscopic retrograde cholangiopancreatogram; PRSS1, protease serine 1; SPINK1, serine protease inhibitor kazal type 1 (SPINK1); CFTR, cystic fibrosis transmembrane conductance regulator; SLE, systemic lupus erythematosus
Gallstones
The prevalence of asymptomatic gallstones in the western population is estimated to be 10–15% and a significant proportion of patients presenting with biliary pancreatitis arise within this cohort. 40 Patients with a known history of symptomatic gallstone disease have a higher risk of developing AP, although their lifetime risk remains less than 10%. 41 Gallstones causing AP are usually less than 5 mm diameter due to the much higher probability of migration from the gallbladder into the common bile duct. 42 At the Ampulla of Vater gallstones cause obstruction of the pancreatic duct and the resultant back pressure leads to intrapancreatic enzyme activation through impedence of acinar exocytosis. 43
Elevated serum concentrations of enzymes arising from the liver, in particular alanine aminotransferase (ALT), may be suggestive of gallstone aetiology, although a significant proportion of patients will have normal LFTs. 44 Indeed, in a meta-analysis of studies examining LFTs in AP, the sensitivity of a serum ALT activity >150 U/L as a predictor for gallstone-induced pancreatitis was only 50%. 45 Patients with gallstone-induced pancreatitis are reported to be more likely to present with markedly elevated serum amylase activity compared with other aetiologies. In one study, 89% of patients with gallstone pancreatitis presented with an amylase activity >1000 IU/L compared with only 6% of those with alcoholic pancreatitis. 46 The lipase/amylase ratio has also been advocated as a predictor of aetiology with a high ratio (≥3) reported to be indicative of alcoholic aetiology and a low ratio (<2) indicative of gallstones. 47 However, the value of measuring this ratio has not been universally confirmed and it is rarely used in clinical practice. 48 Stimac et al. 49 devised a seven-variable scoring system including serum amylase, urinary amylase, aspartate aminotransferase (AST), ALT, alkaline phosphatase, lipase/amylase ratio and mean corpuscular volume. In 145 patients, biliary (score ≥4) and alcoholic (score <4) pancreatitis could be distinguished with a sensitivity and specificity of 92% and 94%, respectively.
Alcohol
The risk of developing alcohol-related pancreatitis varies significantly between individuals and appears to relate to the presence of co-factors, in particular genetic susceptibility. 50 However, AP is much more frequent following prolonged alcohol abuse than after isolated binges. 51
Chronic alcohol exposure leads to impaired exocytosis through acetaldehyde-induced microtubular dysfunction and apical cytoskeleton reorganization in acinar cells, with subsequent accumulation of intracellular enzymes. 21 In addition, alcohol decreases the stability of zymogen and lysosomal membranes and enhances acinar cell sensitivity to cholecystokinin (CCK) further increasing susceptibility to pathological enzyme activation. 52,53
The pancreas metabolizes ethanol via the oxidative and non-oxidative pathway, both of which can contribute to pancreatic injury. Oxidative metabolism by alcohol dehydrogenase and cytochrome P450 2E1 produces acetylaldehyde, which increases lysosomal instability and depletes antioxidant glutathione. 54 ROS formed as a by-product of oxidative ethanol metabolism further contribute to acinar cell oxidative stress and resultant damage. Non-oxidative metabolism of ethanol (and fatty acids) in the pancreas generates fatty acid ethyl esters (FAEEs) catalysed by FAEE synthases. 55 These contribute to the pathogenesis of AP through a number of mechanisms including lysosomal membrane destabilization, induction of pro-inflammatory transcription factors NF-κB and activator protein-1, 56 generation of fatty acids by hydrolysis 55 and disruption of Ca2+ homeostasis. 57 The importance of FAEEs in AP has been further supported by the association between the risk of developing alcoholic pancreatitis and polymorphism of the gene encoding carboxyl ester lipase, an enzyme responsible for the synthesis of FAEE. 58
Hyperlipidaemia
Hyperlipidaemic pancreatitis (HLP) results from significant hypertriglyceridaemia (>10 mmol/L) but is unrelated to hypercholesterolaemia. 59 FAEEs have been shown to accumulate within the pancreas in hyperlipidaemic pancreatitis producing a similar pathological process to that observed in alcoholic AP. In addition, hyperlipidaemic-associated mitochondrial accumulation of fatty acids leads to uncoupling of oxidative phosphorylation, depletion of the mitochondrial membrane potential and impaired ATP production. 60
HLP most commonly arises from an acute precipitating factor such as alcohol abuse or poor diabetic control on a background of a familial hyperlipidaemia. However, AP secondary to isolated hyperlipidaemia can occur rarely. 61 In a similar way to familial hypertriglyceridaemia, isolated acquired causes for HLP are uncommon and are mostly observed in diabetic ketoacidosis. 62 Alcohol abuse, pregnancy, oral contraceptive pill use and obesity frequently lead to an increase in concentration of serum triglycerides but only to concentrations needed to precipitate AP in the presence of an underlying lipid abnormality. 63
HLP should be suspected in patients with a firm diagnosis of AP combined with serum triglycerides >10 mmol/L or grossly lipaemic serum. 61 However, serum lipid concentrations may be elevated above normal in up to 50% of patients during the acute phase of pancreatitis and therefore need to be evaluated with caution and the sample usually repeated. 64 Similarly, in patients with true HLP, triglyceride concentrations often fall rapidly after 24–48 h due to therapeutic fasting and non-calorific fluid infusion.
Hypercalcaemia
Several potential underlying mechanisms for hypercalcaemic pancreatitis exist. Firstly, calcium deposition in the pancreatic duct can cause outflow obstruction in a similar way to gallstones. Secondly, elevated serum calcium concentration protects trypsin from autolysis and may therefore increase propensity to pancreatitis. A third mechanism relates to the central role of calcium in the control of pancreatic acinar cell secretory function. The secretory pole of the acinar cell contains a mixture of inactive digestive enzymes held within a condensed matrix tightly bound by high concentrations of calcium. 65 Stimulation of G-protein linked acetylcholine and CCK receptors generate second messengers, which bind to calcium channels, ryanodine receptors and inositol phoshate 3 receptors on the ER, leading to an increase in cytosolic calcium concentration. 66 In addition, calcium itself elicits further calcium release from these receptors via calcium-induced Ca2+ release, amplifying the signal. Spikes in ionized calcium concentration stimulate release and activation of trypsinogen in postexocytic endocytic vacuoles. 67 Calcium spikes are then terminated by the sarcoendoplasmic reticulum calcium – ATPase pump, the plasma membrane calcium – ATPase pump and the mitochondrial calcium uniporter restoring normal basal ionized calcium concentration. Prolonged or uncontrolled elevation of serum ionized calcium concentration, secondary to hypercalcaemia, but also induced by pancreatic hyperstimulation, bile salts, fatty acids and FAEEs, leads to excessive intracellular trypsinogen activation and pancreatitis. 68
The final mechanism by which calcium dysregulation mediates pancreatitis is by influencing cell fate signalling, particularly in the context of mitochondrial function. In experimental pancreatitis, the prolonged elevation of cytosolic calcium concentration appears to be a crucial factor leading to acinar cell necrosis and the abrogation of elevated serum ionized calcium concentration with calcium chelators is cytoprotective. 69 Pathological elevation of serum ionised calcium concentration increases the ionic load on mitochondria, depleting the mitochondrial membrane potential and exacerbating ATP deficiency, eventually leading to necrosis. 70 In addition, the activity of the mitochondrial permeability transition pore and release of mitochondrial cytochrome c appear to be central to acinar cell apoptotic signalling during pancreatitis. 70,71
Mechanisms of cell fate may have particular clinical importance in AP as promotion of apoptosis may improve outcome, whereas caspase inhibition appears to be detrimental. 72 It has been proposed that stressed acinar cells potentiate pro-inflammatory signalling and apoptotic deletion of these cells attenuate overall injury. Conversely, others suggest that apoptotic cell loss during AP may significantly contribute to renal, hepatic and pulmonary dysfunction, bacterial translocation in the gut and impaired resistance to infection. 73
Drug-related
Drug-related pancreatitis is uncommon (1.4–2%) and reactions are usually idiosyncratic and of low severity. Non-idiosyncratic reactions arise from a range of drug effects including direct toxicity (e.g. diuretics), pancreatic angioedema (e.g. ACE inhibitors) and hyperlipidaemia (e.g. beta-blockers). 74 The underlying mechanism of drug-induced AP is poorly understood but certain patient groups such as children, the elderly and those who are HIV may be more susceptible. 75 Drugs causing pancreatitis have been classified into classes I–III depending on the number of case reports (>20, 10–20 and <10 reports, respectively). 74
Iatrogenic
There are a number of iatrogenic causes of pancreatitis including ERCP, pancreaticobiliary surgery and cardiopulmonary bypass. Post-ERCP pancreatitis is by far the most common with a risk of around 5%, although the majority of cases are mild and self-limiting. 76
Idiopathic
The underlying cause for AP cannot be identified in a significant proportion of patients, although much attention has been focused on minimizing this group in recent years. The British Society of Gastroenterology guidelines on the management of AP suggest that a diagnosis of idiopathic pancreatitis should be made in less than 20% of cases. 77 In this context, a number of contentious aetiological factors have been suggested that would otherwise be labelled idiopathic; these include biliary microlithiasis, sphincter of Oddi dysfunction and pancreatic divisum.
Biliary microlithiasis, the presence of tiny stones or more commonly sludge, has been reported as the cause of idiopathic AP in up to 80% of cases. 78 Interestingly, Grau et al. 79 suggested that measuring LFTs may aid in the diagnosis of microlithiasis. They found that elevation of serum ALT and AST >1.2 times the upper reference limit (URL) within 24 h of admission, in patients with idiopathic pancreatitis, was diagnostic of microlithiasis with a positive predictive value (PPV) of 92% and negative predictive value (NPV) of 89%. Suspected cases of microlithiasis should be investigated with endoscopic ultrasound scanning (USS) and biliary crystal analysis, although empirical cholecystectomy is an accepted treatment approach.
Sphincter of Oddi dysfunction as a primary cause of AP remains unconfirmed but is identified in up to a third of patients with recurrent idiopathic AP and is usually managed by endoscopic sphincter ablation. 80 Pancreas divisum is a congenital variant present in 7% of the population in which the embryological dorsal and ventral ductal systems fail to fuse. 81 It is controversially proposed that a propensity to AP relates to impaired exocrine outflow from the main part of the gland via a small or stenosed accessory duct.
Hereditary
The role of genetic predisposition in AP remains much less clear than in chronic pancreatitis but may include a cohort of ‘idiopathic’ cases, particularly in the context of recurrent acute pancreatitis. There are currently three main genes thought to be important in this context, namely trypsinogen gene protease serine 1 (PRSS1), serine protease inhibitor kazal type 1 (SPINK1) and the cystic fibrosis transmembrane conductance regulator gene (CFTR). Mutations in PRSS1 can lead to the premature activation of trypsinogen and are found in 80% of patients with hereditary pancreatitis. 82 SPINK1 is a specific trypsin inhibitor and a crucial regulator of inappropriate intracellular activation of trypsin. SPINK 1 mutations, such as the N34S mutation, are very common in the general population and are likely to increase susceptibility to known risk factors such as alcohol rather than being an isolated aetiological factor. 83 CFTR control fluxes of chloride and bicarbonate in ductal cells and therefore the ‘washing out’ of trypsin into pancreatic ducts. Mutations in CFTR are similarly common and also thought to represent a modifier in a multifactorial process. 84
Diagnosis
Clinical history and examination findings may be highly suggestive of AP but are frequently non-specific. The prompt and accurate diagnosis of AP is crucial in order to allow expedient delivery of appropriate resuscitative management, thereby reducing the risk of early organ dysfunction. The differential diagnosis primarily includes other causes of an acute abdomen presentation, with systemic inflammatory response, many of which, in marked contrast to pancreatitis, warrant urgent laparotomy. The standard diagnostic test most commonly used is serum amylase, variably supplemented with urinary amylase and serum lipase. Urinary trypsinogen-2 has demonstrated promising results as an alternative to amylase but is not in widespread clinical use. Computed tomography (CT) now plays an increasingly important role in the management of acute surgical pathology and AP may on occasion be first diagnosed by emergency radiological imaging.
Amylase
A serum amylase activity greater than three times the URL remains the primary diagnostic test for AP in most centres and the evidence generally supports this. Gumaste et al. 85 using this cut-off for diagnosis of AP recorded a sensitivity of 72%, specificity of 99%, PPV of 98% and NPV of 82%. However, amylase has some well-recognized limitations as a diagnostic test and normal activities have been observed in 19–32% of cases of AP. 86 Serum activity starts to rise within hours of onset and falls variably after 3–5 days making it less accurate in late presentation. Urinary amylase is significantly elevated for several days after serum activities return to within reference limits in patients with normal renal function, and can therefore be a useful adjunct. The reported sensitivity and specificity (using a cut-off of 550 U/L) for urinary amylase activity is 62% and 97%, respectively. 87 Serum activities within the reference interval in patients with AP may also be found in the context of chronic pancreatic exocrine insufficiency secondary to parenchymal loss, particularly in the context of alcoholic pancreatitis. 88 Hypertriglyceridaemia is known to interfere with amylase assays, with early reports of up to 50% of patients with acute pancreatitis and elevated lipids having a normal serum amylase. 89 Explanations for this effect include optical interference of lipaemic serum during endpoint colorimetric methods, interaction between triglycerides and assay reagents and the presence of an amylase inhibitor in lipaemic serum. 90,91 Serial dilution or ultracentrifugation of lipaemic serum may be used to improve accuracy, although newer assays are significantly less susceptible to interference. 92–95 Rarely, very rapid normalization of serum amylase activity may arise from prompt disease resolution, profound early necrosis or accelerated renal clearance of amylase.
The specificity of serum amylase is also challenged by the frequent elevation of activity observed in numerous non-pancreatic intra-abdominal presentations, including peptic ulcer disease, cholecystitis, intestinal obstruction, mesenteric ischaemia and gynaecological disorders. Abnormal serum activities can also be observed in extra-abdominal pathology including salivary gland disease, pneumonia and head injury. A number of metabolic disorders have been implicated including renal failure, liver failure, diabetic ketoacidosis and anorexia nervosa. Asymptomatic hyperamylasaemia has been observed as a benign familial variant and in high proportions of intoxicated patients with a background of chronic alcoholic abuse, HIV-positive persons and patients following ERCP. 96 Macroamylasaemia gives rise to elevated activities of serum amylase and is found in up to 2.7% of hospitalized patients. It arises from the formation of large molecular mass amylase–immunoglobulin complexes. Diagnostic confusion can be clarified by measuring urinary amylase (in the absence of renal impairment) which will be normal as the size of the complexes prevent renal filtration. 97 Similarly, the amylase–creatinine clearance ratio can be calculated, with a ratio <1% reported to be consistent with macroamylasaemia, compared with 2–5% observed in healthy individuals and >5% in individuals with acute pancreatitis. 98
The lack of a universally agreed URL remains a hindrance to the effective use of amylase as a diagnostic tool. Studies have used activities ranging from as low as 114 IU/L to over 1000 IU/L. 99,100 A cut-off value of 1000 IU/L generates a specificity near to 100% but at the cost of sensitivity, which can fall to 61% compared with over 91% using a cut-off of 300 IU/L. 96 It can be argued that in AP, in which early diagnosis and intervention is crucial, sensitivity should be increased at the expense of specificity.
Lipase
The reported sensitivity of serum lipase ranges from 85% to 100% and is generally found to be greater than that of amylase, 85,96,100,101 although not exclusively. 102 In particular, the half-life of lipase is longer than amylase at 8–14 days and has four times the activity, making loss of sensitivity due to late presentation or chronic insufficiency less likely. 103 Furthermore, analysis of pancreatic tissue in chronic pancreatitis demonstrated that the decline in amylase activity is significantly greater than lipase activity (91% versus 26%). 104
Pancreatic lipase activity is 100-fold greater than that found in the liver and small bowel and 20,000-fold greater than serum activity. Consequently, measurement of serum lipase activity may therefore be expected to have greater specificity than measuring amylase. However, most studies indicate similar specificities for amylase and lipase in the range of 92–99% 85,96,99,101 Certainly, non-pancreatic hyperlipasaemia has been observed in a range of pathologies including peptic ulcer disease, mesenteric ischaemia, acute renal failure, bone fractures, crush injury and fat embolism. 105 Inaccuracies in serum lipase estimation have also been shown to arise from the presence of multiple fractions of lipase in the serum of patients with pancreatitis and the existence of macroforms of lipase. 106 In contrast to the measurement of serum amylase, the lipase assay is unaffected by triglycerides, although drugs such as furosemide can cause inaccuracy.
As in the case of amylase, there is no consensus on the cut-off value of serum lipase activity in the diagnosis of AP. The URL 2 × URL and 3 × URL have all been advocated. 85,102,107 Pezzilli et al. 108 found that increasing the serum lipase diagnostic cut-off value from 270 to 483 IU/L improved overall PPV/NPV from 62%/100% to 87%/99%, respectively. More recently, Smith et al. 99 found the diagnostic accuracy of serum amylase and lipase activities in 1880 patients measured on admission to the emergency department was greatest using much lower cut-off values. The cut-off point for lipase was 208 IU/L and for amylase was 114 IU/L with a sensitivity of 90.3% versus 76.8% and a specificity of 93% versus 92.6%, respectively. The area under the receiver operator curve (AUC ROC) for lipase was 0.948 and was significantly better than for amylase at 0.906.
Urinary trypsinogen-2
Trypsinogen is a pro-enzyme cleaved by enterokinase and trypsin itself to form active trypsin and trypsinogen activation protein (TAP). Two major isoenzymes exist, trypsinogen-1 and trypsinogen-2, the second being significantly elevated in AP. 109 Serum and urinary concentrations of trypsinogen-2 increase significantly within a few hours and fall to within reference limits after a few days. A large prospective study involving 500 patients presenting with acute abdominal pain compared a urinary dipstick test for trypsinogen-2 (detection level 50 μg/L) with serum and urinary amylase (cut-off 300 and 2000 IU/L, respectively). 110 The trypsinogen test had a sensitivity and specificity of 94% and 95%, respectively, significantly higher than those of serum amylase (85% and 91%) and urinary amylase (83% and 88%). In a similar study, urinary trypsinogen-2 also compared favourably with serum lipase demonstrating a sensitivity and specificity of 93% and 92% versus 79% and 88%, respectively. 111 In both these studies the NPV for urinary trypsinogen-2 was 99%.
Conclusions
In current practice, in most centres, the prompt diagnosis of AP relies on the evaluation of serum amylase activity in the context of clinical suspicion. In this context most clinicians avoid using rigid diagnostic amylase cut-off values, in order to make allowance for common confounding factors, in particular, time of presentation, underlying chronic pancreatitis and presence or absence of hyperamylasaemia in non-pancreatic intra-abdominal pathology. The measurement of urinary amylase and serum lipase are generally reserved as adjuncts when the diagnosis of AP remains unclear in the presence of such confounding factors. Combining amylase and lipase has not been found to increase diagnostic accuracy. 112 However, the literature might suggest that lipase or indeed trypsinogen-2 could offer more accurate first-line investigations for AP, although this has not translated widely into clinical practice. In acutely unwell patients with marked peritonitis and equivocal evidence of AP, the gold standard investigation is an abdominal CT scan. This is primarily to identify non-pancreatitic pathology, particularly that requiring urgent laparotomy and secondarily, by exclusion, to avoid laparotomy in AP, which is associated with increased mortality. Otherwise, early cross-sectional imaging (before 3–5 days of onset) is frequently unhelpful as the radiological features of AP are often absent or non-specific.
Severity stratification
The rapid and accurate stratification of severity in AP is critical to early management as delay in the appropriate transfer to a critical care setting increases mortality and morbidity. Simple clinical assessment on admission has been shown to identify less than half of patients with severe pancreatitis. 113 Numerous scoring systems incorporating physiological parameters, laboratory investigations and radiological studies have been developed to improve accuracy of severity prediction. Those in current clinical use include Ranson's criteria, the modified Glasgow score, APACHE II and the computed tomography severity index (CTSI). The main disadvantages of these systems are the time delay to score completion (generally 48 h) and/or the complexity in score calculation. The ideal test of a simple, rapid and accurate assay remains elusive. C-reactive protein (CRP) represents the only simple marker widely used in current clinical practice, although numerous other candidates have been identified, including acute-phase reactants, inflammatory mediators and pancreatic enzyme products. Serum amylase activity at diagnosis does not correlate well with severity of AP. 114 In addition, ongoing monitoring of amylase in early disease does not aid in the evaluation of clinical progress or prognosis. However, persistent late elevation of serum amylase may be associated with certain complications, in particular pancreatic pseudocyst formation.
Multifactorial scoring systems
Multifactorial scoring systems have been used in the assessment of patients with pancreatitis for several decades as a method of quantifying disease severity (Table 2). Most systems aim to grade global organ dysfunction as the extent of this in the first week of admission is closely associated with mortality. 7 In Ranson's criteria, the first scoring system to be described, 11 physiological parameters are recorded on admission and at 48 h. The risk of mortality relates to the number of positive components: 1% with less than three components, 16% with 3–4, 40% with 5–6 and 100% with more than six components. 115 In a meta-analysis including 1300 patients Ranson's criteria (score ≥3) was found to be moderately accurate with a sensitivity of 75%, specificity 77%, PPV 49% and NPV of 91%. 116 The modified Glasgow criteria is a more widely used adaption of Ranson's criteria with the advantages of requiring fewer measurements (8 instead of 11) and being applicable to all cases of AP regardless of aetiology (Ranson's was originally described for evaluating alcoholic pancreatitis with a separate modified version designed for gallstone pancreatitis.) 117
Multiple factor scoring systems for severity stratification in pancreatitis
BISAP, bedside index for severity in acute pancreatitis; AST, aspartate transferase; LDH, lactate dehydrogenase; SIRS, systemic inflammatory response syndrome
The APACHE II score was originally designed to predict patient survival on intensive care admission for all causes, but has been found to be more accurate than Ranson's criteria in predicting SAP and can be calculated at any time point. 118 However, generating the score can be cumbersome in that it requires collection of a large number of parameters, several of which are unlikely to be relevant to AP. The sensitivity of the APACHE II system in predicting SAP (score >9) is 63–82% increasing to 75–89% if combined with the Ranson criteria (APACHE II >9, Ranson's >2). 113
The bedside index for severity in acute pancreatitis (BISAP) has been recently proposed as an alternative scoring system with the dual advantages of being simple to perform and accurate at 24 h. 119 Five parameters are recorded: (i) serum urea >8.9 mmol/L, (ii) disturbed mental status, (iii) presence of SIRS (≥2 of: pulse >90 beats/min, respiratory rate >20 breaths/min or PaCO2 <4.3 mmol/L, temperature >38 or <36°C, white blood cell count >12,000 cells/mm3 or <4000 cells/mm3 or >10% immature neutrophils), (iv) age >60 years, and (v) the presence of a pleural effusion. The BISAP was found to have AUC ROC of 0.81(CI 0.74–0.87) for prediction of SAP in a prospective study of 185 patients. 120 The accuracy of BISAP was very similar to APACHE II and CTSI but slightly less (although not significantly) accurate than Ranson's with a AUC ROC of 0.94 (0.89–0.97). Another alternative, the Simple Prognostic Score (SPS), has recently been shown to be as accurate as standard scoring systems allowing prognostic evaluation on admission. The system includes just three factors: serum urea concentration >8.9 mmol/L, serum lactate dehydrogenase activity >900 U/L and the presence of necrosis on CT. 121
The modified Glasgow score is currently the most widely used in the UK as the calculation can be performed easily using parameters recorded during standard clinical practice. However, the score is limited by a 48 h delay to completion and as such may be superseded by newer models such as BISAP and SPS, or ideally a single prognostic marker, in the future.
Laboratory prognostic markers
Acute-phase reactants
C-reactive protein
CRP is an acute-phase protein produced by hepatocytes in response to cytokines including IL-1 and IL-6. A CRP concentration >150 mg/L at 48 h is widely used in clinical practice to predict SAP, having a reported sensitivity and specificity of around 80%. 122 Al-Bahrani and Ammori 123 performed a review of studies investigating the accuracy of CRP (cut-off range 110–150 mg/L) in predicting SAP at different time points following admission. The sensitivity and specificity of CRP were 36–61% and 89–90%, respectively, on admission, 44–83% and 70–96% at 24 h, 57–89% and 55–82% at 48 h, and 83–90% and 60–84% at 72 h. The relatively late increase in sensitivity and specificity of CRP remains the main limitation of its use in the early identification of patients with severe disease. However, elevated serum CRP at 72 h (cut-off range 200 mg/L) has been reported to be predictive of PN with an NPV of 97% and may be used clinically as an indication for CT scanning. 124
Procalcitonin
The 116 amino acid propeptide of calcitonin, procalcitonin, is a well-recognized acute-phase reactant and various methods are available for its measurement. In a study of 162 patients 24 h post-admission with AP, procalcitonin concentration >0.5 ng/mL on a semi-quantitative strip test had a sensitivity of 92% and NPV of 97% for predicting subsequent organ failure. 125 Procalcitonin measured within 12 h of presentation had a similar specificity as it had when measured at 24 h (84%) but slightly lower sensitivity (71%). Elevated procalcitonin concentrations beyond 72 h have been shown by some groups to be an accurate predictor of infected PN, although this is not universal. 126–128 In the study by Rau et al. 127 in which 11 out of 50 patients died, a late procalcitonin concentration >5.7 ng/mL predicted mortality with a sensitivity and specificity of 100% and 92%, respectively.
Serum amyloid A
Serum amyloid A (SAA) is synthesized in the liver in the context of acute inflammation acting as a chemoattractant for neutrophil, monocytes and T-cells. Reports of the efficacy of SAA in predicting severity of AP are conflicting. The strongest data derive from a multicentre randomized trial involving 172 patients used a plasma enzyme-linked immunosorbent assay with a cut-off concentration of 418 mg/L measured on admission. Severe AP was predicted with a sensitivity of 67% and NPV of 89% demonstrating moderate efficacy. 129 Similar results were obtained in a smaller study involving 31 patients and a lower cut-off value of 250 mg/L. 130
Inflammatory mediators
Interleukin -6
IL-6 acts as an inducer of hepatic CRP synthesis and is the most widely investigated cytokine marker. Pezzilli et al. 131 investigated the reliability of measuring IL-6 concentration at 24 h postadmission for AP in predicting severe disease. In their first study involving 38 patients, an IL-6 concentration >2.7 ng/L had a sensitivity of 100% and specificity of 86% for detecting SAP. Similar results were obtained in a further study involving 40 patients (15 with SAP) and a cut-off of 3.7 μg/L (sensitivity 100%, specificity 83%). When compared directly with CRP, TNFα, IL-1B and IL-8, IL-6 (using a cut-off >400 pg/mL at 24 h) was superior with an overall accuracy of 88%. 132 Stoelben et al. 133 found that late elevation of IL-6 (>600 pg/L at 72 h) was also a useful predictor of PN with a PPV and NPV of 89% and 85%, respectively. The limitations of IL-6 are the complexity of the assay and that accuracy diminishes if measured on admission rather than at 24 h with a reported sensitivity and specificity of 70% and 79%, respectively. 134
Interleukin-8
IL-8 is a key secondary mediator of TNFα-mediated neutrophil activation. Like IL-6, measuring IL-8 at 24 h post-admission, with a cut-off concentration of 30 pg/mL, shows the highest accuracy for diagnosing SAP with a PPV of 77% and NPV of 100%. 131 In a study of 50 patients with severe pancreatitis IL-8 concentrations >140 pg/mL were predictive for organ failure with a sensitivity and specificity of 79% and 81%, respectively, comparing favourably with CRP (sensitivity 71%, specificity 78%). 127
Interleukin-10
Increased concentrations of the anti-inflammatory cytokine IL-10 has been reported as a marker of mild pancreatitis rather than severe disease. 131 In contrast, Chen et al. found that post hoc measurement of IL-10 revealed significantly elevated concentrations (>30 pg/mL) in severe pancreatitis with an NPV of 83% at 24 h. 135 IL-10 has also been shown to be accurate in predicting MOF in AP when measured in combination with both serum calcium and IL-6 with an NPV of 93% and 83%, respectively. 136,137
Tumour necrosis factor α
TNFα is released by monocytes, macrophages and acinar cells, and plays a key role in the inflammatory response of AP by regulating other cytokines and adhesion molecules. Furthermore, the TNFα-related apoptosis-induced ligand receptor pathway is a central component of tissue injury and is associated with severe AP. 25 Serum TNFα concentrations increase early in pancreatitis and are cleared rapidly thus limiting its measurement as a clinical marker. 132 Inhibitory soluble receptors for TNFα (sTNFRα-1 and sTNFRα-2), released by target cells as a regulatory mechanism, have longer half-lives in the circulation and may therefore be more reliable markers of TNFα activity. 138 In a study involving 58 patients, an elevation in serum concentration of soluble TNF receptors from day 1 to 2 was identified as a marker for severe disease. 139 Overall, studies evaluating TNFα and sTNFRα as prognostic markers have yielded equivocal results and this probably arises from variations in production and clearance, activity of neutralizing serum enzymes, variability of TNFα/sTNFRα binding and poor uniformity of assay methodology. 25
Monocyte chemotactic protein-1
Monocyte chemotactic protein-1 (MCP-1) is a potent chemokine and elevated concentrations have been correlated with increased incidence of local complications and organ failure in AP. Additionally, the presence of a single nucleotide polymorphism, 2518 A/G in the MCP-1 gene, is associated with a increased risk of developing severe disease and mortality in AP. 140
Intercellular adhesion molecule-1
Elevation in serum concentrations of the central adhesion molecule ICAM-I at 48 h post onset of symptoms have been shown to discriminate between necrotizing and non-necrotizing pancreatitis with a PPV of 82% and NPV of 78%. 141 Perejaslov et al. found that a significantly elevated concentration of ICAM-1 on admission, was predictive of an increased risk of severe disease, development of MODS and PN. The investigators also measured serum IL-18 concentrations, which in contrast peaked much later with persistent elevation beyond 1 week being a marker for ongoing severe disease.
Matrix metallopeptidase-9
The value of measuring serum concentrations of MMP-9 as a very early marker of severe disease in AP has been reported in a small case-control study comparing healthy controls, mild AP and SAP (10 patients each arm). MMP-9 concentration was significantly elevated in SAP at one hour and correlated strongly with APACHE II score, CRP and TNFα concentrations. 142
Phospholipase A2
Phospholipase A2 is produced by the inflamed pancreas and activated neutrophils to generate lipid inflammatory mediators from phospholipids and lipoproteins. 143 In one study a secretory synovial type phospholipase A2 concentration of >300 ng/mL on two successive days in the first four days was highly predictive of infected necrosis. 144
Polymorphonuclear elastase
Polymorphonuclear elastase released by PMN cells to degrade extracellular matrix during the early phases of the acute inflammatory response has been proposed as an early marker of SAP. The findings of the largest study to date, based on a cohort of 182 patients, reported that a polymorphonuclear elastase concentration >300 μg/L at 24 h was associated with a severe disease course with a sensitivity of 93% and NPV of 98%. 145 Importantly, severe disease was also accurately predicted on admission using a serum cut-off of >250 μg/L (sensitivity 93%, specificity 94%). Similar findings were also reported by Viedma et al. in a study involving 80 patients (PPV 95%, NPV 100%), although much lower accuracy has also been reported. 146,147
Pancreatic enzyme products
TAP and trypsinogen-2
TAPs are tetra-1-aspartyl-1-lysyl amino terminal peptides released on cleavage of trypsinogen to form trypsin. In normal physiology they are produced in the gut and undergo minimal absorption as a result of degradation by enteric oligopeptidases. In pancreatitis, premature activation of trypsinogen leads to release of TAP into the circulation with peak serum concentrations observed at 12–24 h post onset. A urinary concentration of TAP of >10 ng/mL, at 48 h post onset of symptoms, has been shown to have a sensitivity and specificity of 100% and 85% for predicting SAP. 148 Johnson et al. 149 compared the prognostic performance of measuring urinary TAP and serum CRP at 24 h in 190 patients with AP. TAP was significantly more accurate than CRP with areas under the ROC curves of 0.81 and 0.66, respectively. In this study and others, TAP measured at 24 h was also found to have comparable accuracy to the APACHE II score. 150
Urinary trypsinogen-2 concentrations have been reported to be significantly less accurate than urinary TAP in predicting SAP, both on admission and at 24 h. 151 In a study using a diagnostic cut-off of 2 mg/L, a third of patients with SAP were missed. 152 In contrast, Kemppainen et al. found that serum trypsinogen-2 and trypsin-2-α 1-antitrypsin were potentially useful early markers of SAP. Trypsinogen concentrations were 32-fold the URL in patients with SAP compared with 10-fold in those with mild disease. This compared with concentrations of serum trypsin-2-α 1-antitrypsin, in severe and mild disease that were 100- and 52-times the URL, respectively. 153
Carboxypeptidase B activation peptide (CAPAP)
Carboxypeptidase B activation peptide is a stable 95-amino acid peptide released from activated pancreatic proenzymes. It has been investigated in several small clinical studies as a marker of SAP in both serum and urine, between admission and at 72 h, demonstrating a sensitivity and specificity of 85–100% and 59–89%, respectively. 130,154,155
Others
Transferrin sialylation
A small study identified increased transferrin sialylation between day 1 and 2, measured using enzyme-linked lectin assay, as an accurate marker of SAP with sensitivity and specificity close to 90%. 156
Poly-C avid ribonuclease
In a study involving 56 patients with pancreatitis, poly-C avid ribonuclease (P-RNase), a marker of pancreatic destruction, at a cut-off activity of 65.3 U/L, had a high specificity for diagnosing severe disease at day 1 (94.4%) but only reached a reasonable sensitivity at day 3 (72.2%). 157 In the same cohort of patients, serum P-RNase activities also showed a close correlation with concentrations of IL-6, IL-8 and TNF-soluble receptors over days 1–4. 158
Conclusions
The ideal prognostic test based on a simple, reliable and accurate assay remains to be found in the context of AP. The only marker used widely in clinical practice is CRP. In similarity to the commonly used scoring systems, CRP has the significant limitation of only being acceptably predictive after 48 h from onset of disease, thus delaying severity stratification. In this regard, IL-6, polymorphonuclear elastase and TAP represent the most thoroughly investigated markers with good accuracy at 24 h and as such may be used increasingly in the future. Among the array of other markers under early investigation, most interest lies with those displaying accuracy within hours of onset such as MMP-9.
Radiological severity scoring
The CTSI described by Balthazar et al. for acute pancreatitis is widely used to aid identification of severe, complicated AP and monitor disease progression. A score of 0–10 is generated from the grading of topological severity of pancreatitis and the extent of necrosis (Table 3). 159 In a comparative study, a CTSI >3 within the first five days had a PPV of 92% and NPV of 95% for predicting SAP. This compared favourably with CRP (concentration ≥150 mg/L at 48 h; PPV 50%, NPV 94%) and APACHE score (≥7 at 48 h; PPV 57%, NPV 88%). An elevated early CTSI has also been shown to be predictive of complications, sepsis, need for intensive care unit (ITU) admission and mortality. 160 CT has the additional advantage of demonstrating the precise extent of pancreatic inflammation or necrosis and identifying specific complications such as peri-pancreatic fluid collections. Typical radiological features of SAP on cross-sectional imaging are shown in Figure 1. The major limitation of the CTSI in predicting severity is that no evaluation of the systemic inflammatory response in AP is obtained.
Typical features of severe acute pancreatitis. (a) Normal pancreas with homogenous perfusion and distinct parenchymal outline (A) lying anterior to the splenic artery (B) and directly posterior to the stomach (C). (b) Necrotic attenuated pancreas with indistinct border representative of local inflammatory changes (A). The pancreas is almost entirely replaced by a large infected anterior collection containing pockets of gas (B) into which a radiological drain (C) has been inserted. The presence of a nasojejunal feeding tube, which is traversing the duodenum compressed by the collection, is also demonstrated (D)
CT severity index
Management
The initial management of acute pancreatitis entails cessation of oral intake (so-called ‘pancreatic rest’), supportive measures (in particular fluid resuscitation), close monitoring, and support of organ function and effective analgesia. Early identification of severe disease is crucial and should prompt immediate transfer to a high dependency or intensive care setting. Rapid confirmation of aetiology should be sought and specific interventions undertaken as appropriate. Patients should have an abdominal USS performed within 24 h of admission to exclude gallstones and those with signs of biliary sepsis or obstruction (with or without sepsis) undergo urgent ERCP and sphincterotomy. 77 Laparoscopic cholecystectomy should be performed during the same admission and at least within four weeks of discharge (or when inflammation has resolved in SAP) to avoid early recurrent attacks. In hyperlipidaemic pancreatitis, the hypertriglyceridaemia tends to improve spontaneously. However, effective acute and long-term control of serum triglycerides (at least below 5 mmol/L) is essential to reduce severity of AP and prevent recurrent attacks. 161,162 Acute treatment modalities remain controversial and include heparinization, intravenous insulin and plasma exchange, with maintenance therapy centring on lipid- lowering agents (particularly fibrates) and dietary control. 163 Genetic screening for familial hyperlipidaemia should be undertaken where appropriate, as suggested by family history and clinical presentation, in addition to exclusion and management of secondary factors such as alcohol and drug exposure, weight and dietary issues, hypothyroidism and poor diabetic control. Hypercalcaemia causing pancreatitis tends to fall spontaneously but specific measures to normalize serum calcium should be undertaken if this does not occur. A careful drug history is essential with cessation of any potential causative agents for AP. 75
The key ongoing issues in severe pancreatitis are nutritional support and management of PN and associated complications. Adequate nutritional support in AP is crucial as most patients develop a hyperdynamic and hypercatabolic state associated with significant energy expenditure and negative nitrogen balance. 164 Early enteral nutrition has been shown to improve outcomes in high-risk surgical patients by maintaining gut-related immune function, preventing bacterial translocation and reducing intestinal ischaemia–reperfusion. 165 This appears to extend to AP in which a meta-analysis has shown enteral feeding to be superior to TPN in terms of cost, length of hospital stay, septic complications and need for surgical intervention. 166 Nasojejunal feeding has the theoretical advantage of minimizing pancreatic stimulation and even increasing the release of inhibitory factors such as inhibitory polypeptide, polypeptide YY and somatostatin. However, clinical trials have demonstrated no significant difference between NG and NJ feeding in terms of morbidity and mortality with NG being simpler and cheaper to use. 167 Despite this, a significant proportion of patients with severe AP will be unable to tolerate NG feeding due to gastric outlet obstruction and occasionally even NJ feeding due to significant ileus. The benefits of additives to feeding regimens such as selenium, glutamine and probiotics remain unclear. 168
PN is a common feature of SAP and if sterile should be treated conservatively, avoiding invasive intervention, which might lead to iatrogenic superinfection. In contrast, infected PN is associated with a high mortality and warrants rapid diagnosis and treatment. Clinical suspicion of infected necrosis should prompt urgent USS/CT-guided fine needle aspiration for microscopy and culture and instigation of broad-spectrum intravenous antibiotics until microbiological results are available. 169 Infected necrosis invariably requires drainage or removal of infected tissue and numerous approaches are described including percutaneous radiological drainage, endoscopic drainage, minimally invasive surgery and extensive open debridement. 170 Prevention of infected necrosis using prophylactic antibiotics in AP remains controversial and the results of relevant studies equivocal. 170,171
Extensive peripancreatic fluid collections, including early pseudocysts (fluid in the lesser sac) are common in SAP and require percutaneous drainage if large, expanding or complicated by infection, bleeding, bowel obstruction or biliary obstruction. After several weeks, mature pseudocysts may form and most are effectively managed with endoscopic or surgical cystoenteric drainage procedures. 172 Pancreatic leaks arise secondary to necrotic duct disruption or in relation to pseudocyst formation and can present with ascites, pleural effusion or percutaneous fistula. Treatment options include somatostatin analogues, endoscopic and radiological drainage, pancreatic stenting and pancreatic resection. 173 The uncommon but high-risk complication of peripancreatic vascular erosion or pseudoaneurysm formation is most effectively managed with angiographic embolisation. 174
The complex pathophysiology of AP has elicited interest in numerous specific targets for intervention but clinical studies to date have been generally disappointing. The use of PAF receptor inhibitor Lexipafant was recently investigated in clinical trials but was not associated with any difference in severity of organ failure or mortality in AP despite promising preclinical data. 175 The protease inhibitor gabexate mesilate is used in clinical practice in some centres, although there is little evidence to support this. A meta-analysis of all trials investigating the use of protease inhibitors in AP demonstrated a very small reduction in mortality in the subgroup of patients with moderate to severe AP. 176 Numerous agents have been investigated for prophylaxis against post-ERCP pancreatitis (PEP). Non-steroidal anti-inflammatory drugs showed some benefit in preventing PEP in randomized trials in contrast to steroids, which were ineffective. Trials investigating nitroglycerin, a sphincter of Oddi relaxant and anti-inflammatory cytokine IL-10 in PEP prevention were equivocal. 177
Conclusions
Acute pancreatitis represents an important cause of worldwide mortality and morbidity. Improving the outcome for patients remains the focus of extensive basic science and clinical research and many of the current issues relating to AP are particularly pertinent to the clinical biochemist. These include: understanding and manipulation of the complex pathophysiology, prompt and accurate diagnostic testing, effective disease severity stratification, precise aetiologically based management and prevention of recurrent disease.
DECLARATIONS
This article was prepared at the invitation of the Clinical Sciences Reviews Committee of the Association for Clinical Biochemistry.