The role of the intestine in the pathophysiology and management of severe acute pancreatitis

Abstract

Background

The outcome of severe acute pancreatitis has scarcely improved in 10 years. Further impact will require new paradigms in pathophysiology and treatment. There is accumulating evidence to support the concept that the intestine has a key role in the pathophysiology of severe acute pancreatitis which goes beyond the notion of secondary pancreatic infection. Intestinal ischaemia and reperfusion and barrier failure are implicated in the development of multiple organ failure.

Discussion

Conventional management of severe acute pancreatitis has tended to ignore the intestine. More recent attempts to rectify this problem have included 1) resuscitation aimed at restoring intestinal blood flow through the use of appropriate fluids and splanchnic-sparing vasoconstrictors or inotropes; 2) enteral nutrition to help maintain the integrity of the intestinal barrier; 3) selective gut decontamination and prophylactic antibiotics to reduce bacterial translocation and secondary infection. Novel therapies are being developed to limit intestinal injury, and these include antioxidants and anti-cytokine agents. This paper focuses on the role of the intestine in the pathogenesis of severe acute pancreatitis and reviews the implications for management.

Introduction

Severe acute pancreatitis (SAP) remains a significant clinical challenge. It is associated with a mortality rate of 10–40% depending on whether there is sterile or infected pancreatic necrosis ¹. These patients have a profound systemic inflammatory response and usually die with multiple organ dysfunction. A third of deaths occur within days of hospital admission, without an obvious focus of infection ². The remaining deaths occur later and are usually the result of secondarily infected pancreatic necrosis (the so-called ‘late septic deaths’) ³. Mortality rates have not altered over the last decade ⁴, and a reduction will only occur when there is a better understanding of the underlying pathophysiology of this disease.

Recent reviews on intestinal barrier dysfunction in acute pancreatitis have focused on bacterial translocation and the intestinal origin of sepsis ⁵,⁶. However, this paper will explore how the intestine, as both culprit and victim, influences the pathophysiology of SAP (Figure 1). The implications of these interactions with regard to treatment will also be discussed.

Figure 1. .

The interactions between the pancreas and intestine during severe acute pancreatitis with monitoring and interventions used in its treatment.

The role of the intestine in the development of SIRS/MODS in critical illness

Once considered a quiescent organ, the intestine is now seen to have a pivotal role in the development of the systemic inflammatory response syndrome (SIRS) and multiple organ dysfunction syndrome (MODS) in critical illness ⁷. In the 1960s Jacob Fine and his colleagues postulated that severe illness enabled gut organisms and toxins to gain access to the systemic circulation ⁸. Wolochow and associates ⁹ labelled this phenomenon ‘bacterial translocation’ but it received little attention for 20 years. Renewed interest was generated when it was reported that the intestine might be an important source of infecting microbes in immunocompromised patients ¹⁰. There is now little doubt that bacterial translocation occurs but uncertainty remains over the mechanism and how this relates to MODS ¹¹. The same bacteria responsible for postoperative septicaemia have been cultured from mesenteric lymph nodes but this does not always lead to SIRS and MODS ¹²,¹³.

Studies of haemorrhagic shock, gram-negative sepsis and trauma have provided an insight into the way that the intestine influences the course of critical illness. The similar systemic manifestations in these diseases suggest common mechanisms that are relevant to severe acute pancreatitis, especially those with late sepsis.

Splanchnic circulation

Acute pancreatitis is characterised by retroperitoneal oedema, ‘third space’ fluid loss, hypovolaemia and circulatory shock. The physiological response is to prioritise blood flow to vital organs at the expense of the splanchnic circulation. The increased sympathetic outflow that accompanies circulatory shock causes vasoconstriction of the mesenteric post-capillary veins and venules via alpha-adrenergic receptors. This causes an ‘auto-transfusion’ of up to 30% of the total circulating blood volume, so improving the cardiac output and perfusion of other vital organs ¹⁴. Vasoconstriction of the mesenteric afferent arterioles also occurs but this effect is mediated by the renin-angiotensin axis rather than the sympathetic nervous system ¹⁵. The subsequent increase in systemic vascular resistance also helps to sustain the systemic blood pressure but results in low perfusion through the splanchnic bed.

Splanchnic tissues can adapt to low perfusion states by extracting up to 90% of the oxygen from the blood. This protective effect is limited because prolonged extraction rates of >70% lead to regional ischaemia ¹⁶. Resuscitation does not provide immediate relief, as the splanchnic region is the last to be reperfused ¹⁷. Indeed, patients who appear to be adequately resuscitated, without clinical evidence of hypovolaemia, may continue to suffer from intestinal ischaemia ¹⁸. Intestinal ischaemia can also occur in inflammatory states when the splanchnic blood flow is normal or increased through increased metabolic demand for oxygen ¹⁹.

The relationship between intestinal ischaemia and SIRS/MODS has not been fully elucidated, but can be discussed within two broad hypotheses: the ‘gut starter’ and the ‘gut motor’ hypotheses.

The ‘gut starter’ hypothesis is also called the ‘two hit model of MODS’ and focuses on the role of the neutrophil. Neutrophils are primed as they pass through the mesenteric circulation during reperfusion of the ischaemic intestine. They continue to circulate around the body in this primed yet inactive state until they are provoked by a second insult such as exposure to endotoxin ²⁰. Priming of the neutrophil is initiated by mesenteric lymph rather than portal blood and it is the lipid fraction generated by phospholipase A₂ that appears to be most important in priming neutrophils ²¹. Neutrophils primed in this way exhibit an increased oxidative burst, augmented release of proteases and cytokines ²², and reduced apoptosis ²³ so becoming potent mediators of distant organ dysfunction when activated.

The ‘gut motor’ hypothesis focuses on the role of the intestinal barrier in the development of SIRS and MODS. When this barrier is disrupted the luminal content invades the portal venous and lymphatic systems. This translocation activates immune cells downstream from the intestinal mucosa, i.e. Peyers patches, macrophages in the lamina propria of the gut, mesenteric lymph nodes and Kuppfer cells in the liver. These activated immune cells release inflammatory mediators that drive the onset of SIRS and MODS even though a focus of infection may not be evident ²⁴.

Components of the intestinal barrier

The intestinal barrier comprises several functional and structural components. These include the intestinal microflora, gut motility, digestive enzymes, unstirred water and mucous layer, epithelial layer, endothelial layer, mucosal associated lymphoid tissue and the gut-liver axis. The most critical is the epithelial layer. It comprises a single layer of columnar cells arranged into villi and crypts. Microvilli line the apex of the cell and provide a barrier to microbes by virtue of the small space between microvilli and a negative charge ²⁵. Tight junctions encircle the apical pole (the zonula occludens) connecting adjacent cells by strands that are linked to the cell cytoskeleton ²⁶. These prevent passage of molecules >11.5 Å in radius ²⁷, that include lipopolysaccharide and many other bacterial proinflammatory mediators.

Disruption of the intestinal barrier

Ischaemia increases the permeability of the intestinal barrier and there is mucosal acidosis (possibly by lipid peroxidation from free radicals) ²⁸ and ATP depletion ²⁹. As the ischaemia worsens, epithelial cells separate from the tips and then the sides of the villus ³⁰. Surviving cells develop an increased rate of apoptosis ³¹ and decreased nutrient transport ³². Eventually the lamina propria disrupts resulting in mucosal ulceration, haemorrhage and transmural necrosis.

The intestinal barrier is particularly susceptible to ischaemia because of the anatomy of the villus micro-circulation. The countercurrent arrangement of veins around a central arteriole allows for arteriovenous shunting of oxygen and anoxia of the villus tip ³³. This effect is exacerbated by haemodilution due to plasma skimming of central arterioles that branch off at right angles from their parent vessel ³⁴.

Ischaemia causes hypoxanthine to accumulate and ATP is depleted through inadequate oxidative phosphorylation. In addition, xanthine dehydrogenase is converted to xanthine oxidase. When blood flow is restored xanthine oxidase converts hypoxanthine to xanthine with the liberation of the superoxide ion. The superoxide ion leads to the formation of further oxygen free radicals that injure cell membranes by lipid peroxidation ³⁵. Thus, intestinal reperfusion causes further damage through the production of oxygen free radicals and inflammatory mediators.

Oxygen free radicals are also chemoattractant for neutrophils, and along with other stimuli (i.e. TNF-α IL-1, IL-8, GM-CSF, IFN-γ, leukotriene B4, platelet activating factor (PAF), ICAM-1, ELAM-1, and complement C3a and C5a) initiate migration of activated neutrophils into the reperfused tissue ³⁶.

The accumulation of activated neutrophils may be the main cause of intestinal injury, and they mediate damage in three ways: enhancing the ischaemic effect, releasing of oxygen free radicals and proteases, and amplifying the inflammatory response.

When the endothelial cell is stimulated by cytokines, oxygen free radicals, complement or thrombin, it expresses adhesion molecules that bind activated neutrophils. These aggregate at the pre-capillary sphincter and impair blood flow. The accumulation of platelets, products of the coagulation pathway and cellular debris amplifies this effect and the result increases the ischaemia ³⁷.

Oxygen free radicals derived from activated neutrophils are produced by cell surface nicotinamide adenine dinucleotide phosphate oxidase and intracytosolic myelo-peroxidase. Direct intestinal injury is caused by the release of lyzozymes containing neutral serine proteases that degrade elastin, collagen I, II, III and IV, fibrinogen, fibronectin and proteoglycans, and metallo-proteases (MMP-1, MMP-2, MMP-3) that digest collagen, gelatin and proteoglycans ³⁸.

Neutrophils amplify the inflammatory response by secreting a wide range of mediators. When activated they release proinflammatory (IL-1, TNF-α, IL-6, IL-8, IFN-β) and anti-inflammatory (TGF-β) cytokines and their antagonists (IL-lra) ³⁹. In addition they activate platelets and coagulation cascades, and inhibit fibrinolysis ⁴⁰.

Intestinal injury is also mediated through the complement system by the membrane attack complex ⁴¹. Complement can attract neutrophils ⁴², stimulate their migration across endothelial cells ⁴³, increase myelo-peroxidase activity and the production of cytokines ⁴⁴, and can mediate damage in distant organs ⁴⁵.

Cytokines that are liberated from the intestine during ischaemia-reperfusion can directly affect the intestinal barrier. TNF-α induces the permeability of intestinal epithelial ⁴⁶ and endothelial cells ⁴⁷ by altering the actin cytoskeleton ⁴⁸. Blockade of TNF-α significantly reduces the extent of bacterial translocation ⁴⁹. IFN-γ increases nitric oxide production ⁵⁰ that inhibits the Na+, K + -ATPase pump ⁵¹ so increasing the permeability of model epithelial cells ⁵². The effects of IFN-γ are further enhanced by TNF-α ⁵³ but limited by IL-10 ⁵⁴. IL-4 also increases epithelial permeability ⁵⁵.

Phospholipase A₂, found in high concentration in the intestinal mucosa, plays an important role in mucosal injury and translocation ⁵⁶. It is excessively activated in intestinal ischaemia ⁵⁷ and in sepsis is released under the partial control of TNF-α ⁵⁸. It generates proinflammatory lipid mediators such as platelet activating factor that stimulates endothelial cells and neutrophils, enhancing neutrophil-mediated tissue injury. It also generates arachidonic acid so liberating eicosanoids that cause direct tissue injury ⁵⁹.

Nitric oxide (NO) has been shown to have both a protective ⁶⁰ and injurious effect ⁶¹ on the intestinal barrier. Under normal conditions it is an antioxidant and a signalling molecule that maintains intestinal micro-circulation. But during ischaemia and reperfusion, NO is converted to cytotoxic nitrites by superoxide radicals. One such nitrite, peroxynitrite, activates poly(ADP-ribose) synthetase by breaking single strands of DNA. This nuclear enzyme then depletes enterocyte NAD+ and ATP and increases intestinal permeability ⁶². The large concentrations of NO liberated from the intestine during ischaemia and reperfusion are probably due to an upregulation of inducible NO synthase (iNOS) ⁶³. Blockade of iNOS prevents the increased intestinal permeability ⁶⁴ and mesenteric vascular hyporeactivity ⁶⁵ caused by endotoxin. It also limits bacterial translocation during intestinal ischaemia and reperfusion ⁶⁶.

Evidence of impaired intestinal function in acute pancreatitis

There is increasing evidence that the pathophysiologic mechanisms outlined above are relevant to acute pancreatitis. This suggests that the intestine is not only adversely impacted by acute pancreatitis, but plays a role in its progression.

Intestinal atrophy with nutritional depletion

Traditional management of acute pancreatitis has aimed to rest the pancreas, by keeping the patients ‘nil by mouth’. It is now recognised that the intestines require intraluminal nutrition for the maintenance of their structure and function. Fasting leads to intestinal mucosal atrophy ⁶⁷, increased rate of enterocyte apoptosis ⁶⁸, decreased glutamine and arginine transport ⁶⁹, and altered mucin composition of goblet and deep crypt cells ⁷⁰. These changes result in a breakdown of the intestinal barrier as evidenced by increased intestinal permeability ⁷¹.

Intestinal ischaemia

Acute pancreatitis reduces intestinal blood flow ⁷², resulting in mucosal acidosis. Gastric tonometry enables calculation of the intramucosal pH (pHi) by measuring the partial pressure of carbon dioxide in the lumen and the bicarbonate in the blood. This indirect measurement of the adequacy of intestinal perfusion has become a valuable tool in a number of clinical contexts and it has been validated in low flow states ⁷³. In a series of patients with acute pancreatitis the pHi was more closely related to MODS and death than to APACHE II, arterial pH, mixed venous pH, lactic acid and oxygen delivery and consumption. The pHi within 48 hours of intensive care admission was lower in those who died than in those who survived. A pHi of 7.25 or lower predicted death with an overall accuracy of 82% ⁷⁴

Intestinal barrier failure and translocation

The alterations in the intestinal barrier function observed during circulatory shock and sepsis models have been identified in acute pancreatitis. An increase in intestinal permeability has been reported in animal pancreatitis ⁷⁵ and the degree of permeability correlated with the severity of pancreatitis ⁷⁶. This has been confirmed in human studies where increased intestinal permeability occurred early in pancreatitis ⁷⁷, was more likely to be seen in those with MODS ⁷⁸, and normalised during recovery ⁷⁹. It is unclear whether the changes in intestinal permeability contribute to the development of more severe pancreatitis or are just a consequence of it.

One way that the changes in the intestinal barrier may influence the course of pancreatitis is in secondary infection of pancreatic necrosis. These infections are usually due to bacteria, both aerobic and anaerobic, of a similar spectrum to intestinal microflora ⁸⁰. Bacterial translocation from the intestine is well documented in experimental acute pancreatitis but the mechanism is still unclear ⁸¹. Human studies demonstrating increased permeability have not demonstrated an increased incidence of septic complications. Furthermore, the increased permeability occurred early but the peak incidence of infection is known to be 3 weeks into the course of the disease.

Acute pancreatitis may enhance bacterial translocation by altering intestinal motility. Ileus promotes a change in the intestinal microflora and bacterial overgrowth leading to translocation ⁸². Pancreatitis delays intestinal transit time ⁸³ and this has been identified as a contributing factor in translocation during experimental pancreatitis ⁸⁴.

Endotoxaemia

Animal studies have shown that pre-treatment with low dose lipopolysaccharide (50 µg kg) mitigates the formation of pancreatic oedema, lowers the serum amylase ⁸⁵, upregulates acinar cell apoptosis and reduces the severity of caerulein pancreatitis ⁸⁶. Acinar cell apoptosis appears to have a protective role, as it is inversely related to the severity of pancreatitis in several different models ⁸⁷. Pancreatic PAF and phospholipase A₂ may be the triggers of this apoptosis ⁸⁸,⁸⁹. While these studies suggest that the pathological features of acute pancreatitis might be modified by endotoxaemia through the induction of acinar cell apoptosis, there are other studies that suggest that endotoxaemia exacerbates experimental pancreatitis ⁹⁰. For instance, intraperitoneal injection of endotoxin (0.5 and 10 mg/100 g) transformed experimental acute pancreatitis (duct ligation and arginine models) to a more severe form ⁹¹.

Endotoxaemia appears to correlate with the severity, incidence of systemic complications and mortality rates of acute pancreatitis in patients ⁹². The usefulness of measuring endotoxaemia is limited by its intermittent appearance in the peripheral circulation. Another approach is to measure the endogenous antibodies to endotoxin, both IgM and IgG. It has been shown that endotoxaemia occurs in almost all patients with severe acute pancreatitis by the significant decrease and/or depletion of IgM anti-endotoxin antibodies. Similar changes in the serum levels of IgG anti-endotoxin antibody accurately predict death in patients with severe acute pancreatitis ⁹³.

Implications for management: towards an intestinal focus

The conventional management of acute pancreatitis has given scant regard to the intestine, and as a result may increase disease severity. As discussed, initial resuscitation may not overcome splanchnic hypoperfusion, and aggressive resuscitation may enhance reperfusion injury. Inotropic support may further compromise the intestine ⁹⁴ and keeping the patient nil by mouth for even short periods leads to sloughing of enterocytes, villus atrophy and increased intestinal permeability.

In the light of the emerging evidence regarding the role of the intestine in acute pancreatitis there is a need to re-evaluate the approach to management so that there is an emphasis on the preservation of intestinal integrity and function. In particular there should be recognition of the need for sufficient resuscitation to overcome intestinal ischaemia and for adjunctive measures to reduce reperfusion injury and loss of the intestinal barrier. Whether cytokine-based therapies will have a clinical role in preserving intestinal integrity and function has yet to be determined ⁹⁵.

Fluids and endpoints for resuscitation

The rapid restoration and maintenance of intravascular fluid volume remains one of the few therapeutic measures generally accepted as effective in the management of acute pancreatitis ⁹⁶. Improvements in fluid resuscitation, oxygen delivery and intensive care mean that patients are more likely to avoid acute renal failure and survive the initial hypovolaemic insult ⁹⁷. A significant proportion will still go on to develop early multiple organ failure and death ⁹⁸.

The standard clinical parameters used to monitor the resuscitation of hypovolaemia, i.e. heart rate, blood pressure, urine output and haemoglobin, do not detect the residual mesenteric hypoxia ⁹⁹. The extent of this hypoxia is difficult to estimate clinically, even for experienced intensive care physicians ¹⁰⁰. The search for an appropriate endpoint in resuscitation has led to the gastric tonometer as an indirect marker of the adequacy of intestinal perfusion. Although it appears to be a sensitive predictor of outcome in the critically ill, its role in guiding resuscitation is not well established. Further studies are required to determine whether tonometric guided resuscitation will improve management of patients with acute pancreatitis ¹⁰¹.

There is some evidence that cardiac preload (measured as right ventricular end diastolic volume (RVEDV) index by means of a pulmonary artery catheter) may be a better guide to resuscitation. It has been shown that splanchnic perfusion and outcome improved in trauma patients resuscitated to a RVEDV index >100 ml/m². This approach appears to be more effective than inotropic support and is less likely to impact pulmonary function ¹⁰².

Less invasive methods of monitoring the effects of resuscitation include skeletal muscle monitoring. Multi-parameter fiberoptic sensors placed in the deltoid detect PCO₂ and evidence suggests that is a better guide than the restoration of PAO₂¹⁰³.

Bioimpedence is a recently developed technique that records an alternating electrical current of high frequency and low amplitude through limb electrodes. Resistance is due to water and electrolytes and vector analysis is used to determine changes in resistance and hence fluid balance. It has been validated as an accurate measurement of central venous pressure ¹⁰⁴ and cardiac output ¹⁰⁵ in critically ill patients.

The most appropriate resuscitation fluid in acute pancreatitis has not been determined. Impairment of the pancreatic microcirculation and the development of pancreatic necrosis are not prevented by conventional resuscitation using crystalloid ¹⁰⁶. Furthermore, lactated Ringer's solution has been shown to be proinflammatory, inducing cell activation and expression of genes involved in inflammation, cell migration and tissue remodelling ¹⁰⁷. Ringer's ethyl pyruvate solution (REPS) is a novel resuscitation fluid that replaces the lactate in lactated Ringer's with the stable form of pyruvate. Pyruvate is a potent antioxidant and it has been shown to ameliorate the structural and functional damage to rodent intestinal mucosa following mesenteric ischaemia-reperfusion ¹⁰⁸. Its effect in pancreatitis has yet to be determined.

Hypertonic saline may be a more immediate alternative to lactated Ringer's due to its immunosuppressive properties. It abrogates neutrophil priming by inhibiting the intracellular signalling pathways by cell shrinkage ¹⁰⁹, and enhances the protective function of T cells ¹¹⁰. It has been shown to attenuate end-organ damage during rodent pancreatitis ¹¹¹ and limits bacterial translocation during haemorrhagic shock ¹¹². However, in a clinical trial of burn patients hypertonic saline led to increased renal failure and mortality, casting doubt on its safety in restoring volume deficits ¹¹³.

Dextran is a colloid that improves pancreatic micro-circulation, reduces trypsinogen activation, prevents acinar necrosis and improves survival of necrotising pancreatitis in rats ¹¹⁴. A phase-I study has suggested that these effects are relevant to patients ¹¹⁵. A major disadvantage, however, is its tendency to be coagulopathic by reducing platelet adhesiveness, depressing factor VII and inhibiting fibrinolysis ¹¹⁶. This has led to calls for its abandonment as a resuscitative fluid ¹¹⁷.

The role of artificial blood substitutes is still to be defined. Diaspirin cross-linked haemoglobin restores the pancreatic microcirculation during hypovolaemic shock ¹¹⁸, improves mucosal perfusion in rodent ileum during sepsis ¹¹⁹ and abrogates pathologic post-injury neutrophil cytotoxic function ¹²⁰. However, its clinical application has been halted by two trials where mortality was increased following its use in stroke ¹²¹ and trauma victims ¹²². Trauma trials with other groups of blood substitutes (that use polymerised haemoglobin rather than cross-linked) have been encouraging ¹²³, and further trials with bovine and recombinant haemoglobin are still in progress.

Low molecular weight heparin used during resuscitation has been shown to prevent mesenteric neutrophil rolling and adhesion that may modulate the damaging effects of the early inflammatory response ¹²⁴. Other resuscitation therapies include those aimed at correcting the splanchnic vasoconstriction. The use of enalapril, an angiotensin-converting enzyme inhibitor, improves intestinal perfusion in the critically ill ¹²⁵. The use of oestradiol in rodent models of haemorrhagic shock and sepsis also improved intestinal perfusion if given at the beginning of resuscitation, but had no effect on systemic oxygen consumption or cardiac output ¹²⁶. Further studies are required to determine whether these therapies have a role in the treatment of pancreatitis.

Inotropes and selective splanchnic vasoconstriction

The use of inotropes to maintain systemic arterial blood pressure has come under scrutiny because of their effect on intestinal blood flow. Adrenaline is not recommended in septic shock as it impairs intestinal perfusion and leads to gut mucosal damage ¹²⁷. Dobutamine has a variable effect, and fails to improve intestinal blood flow when measured by regional catheterisation and dilution dye test ¹²⁸. In contrast noradrenaline enhances intestinal oxygen utilisation as measured by pHi ¹²⁹. Low dose dopamine increases the intestinal blood flow and oxygen delivery providing the fractional intestinal flow is not already high before treatment ¹³⁰.

Dopexamine, a dopaminergic and beta-adrenergic agonist, improves intestinal perfusion and oxygenation ¹³¹, and maintains intestinal villus microcirculation during normotensive endotoxaemia ¹³². Like dopamine, however, it does not improve intestinal oxygenation in states of hyperdynamic septic shock where the intestinal flow is already increased ¹³³.

The appropriate inotrope for patients with severe acute pancreatitis has yet to be determined. The use of alpha-adrenergic agonists exacerbates rodent pancreatitis by reducing microcirculatory flow in the pancreas ¹³⁴. However, beta-adrenergic and dopaminergic stimulation improve pancreatitis through reduction in the micro-vascular permeability ¹³⁵. The consequences of indiscriminate inotrope use on intestinal and pancreatic circulation need to be recognised. There is a need for selectivity to preserve intestinal circulation and enhance pancreatic perfusion.

Enteral nutrition and preventing intestinal barrier failure

Acute pancreatitis is characterised by hypermetabolism ¹³⁶ and there has been strong advocacy for total parenteral nutrition ¹³⁷. This has been based on the premise that ‘gut rest’ and the ‘avoidance of pancreatic stimulation’ are important to the management of acute pancreatitis. However, human studies show that jejunal feeding only results in minimal increases in pancreatic enzyme, bicarbonate and volume output ¹³⁸ and may even suppress enzyme production ¹³⁹. In experimental models enteral feeding attenuates caecal bacterial overgrowth ¹⁴⁰, maintains the immune responsiveness of the host ¹⁴¹, preserves intestinal mucosal integrity ¹⁴² and permeability ¹⁴³, and prevents bacterial translocation ¹⁴⁴.

Clinical trials comparing parenteral and enteral nutrition have shown the latter to reduce the inflammatory response, disease severity ¹⁴⁵, and the rate of septic complications in acute pancreatitis ¹⁴⁶. Furthermore parenteral nutrition does not hasten the resolution of acute pancreatitis and it is significantly more expensive than enteral feeding ¹⁴⁷. As a result enteral feeding has become the preferred mode of nutritional support for patients with acute pancreatitis. The most beneficial enteral nutritional formula has yet to be determined, but enteral feeds fortified with immune-enhancing substances (e.g. glutamine, arginine and omega-3 fatty acids) may decrease the length of hospital stay and septic complications in patients with pancreatitis ¹⁴⁸.

Parenteral nutrition does have a role, however, in some patients who cannot tolerate the full requirement of protein and calories by enteral means ¹⁴⁹. In these cases, only a small volume of low residue enteral diet is required to protect the intestinal mucosa ¹⁵⁰.

Prevention of secondary infections

Infection of necrotic pancreas by enteric bacteria is a key determinant of clinical outcome ¹⁵¹. Reducing micro-flora overgrowth by restoring gut motility reduces bacterial translocation and improves outcome in acute pancreatitis ¹⁵². Animal studies have shown a decrease in pancreatic infection and mortality with the use of either oral antibiotics to selectively decontaminate the gut ¹⁵³,¹⁵⁴, or intravenous antibiotics with high pancreatic tissue penetration ¹⁵⁵,¹⁵⁶. Similar studies in humans have shown benefits from both systemic antibiotics and selective gut decontamination ¹⁵⁷.

Early prospective randomised trials that used antibiotics in necrotising pancreatitis showed a reduction in pancreatic infection although mortality was not altered ¹⁵⁸,¹⁵⁹,¹⁶⁰, except for one study that used selective gut decontamination with norfloxacin, colistin and amphotericin ¹⁶¹. An early meta-analysis that included these four studies did show a reduced overall mortality ¹⁶² and this has been supported by a more recent meta-analysis ¹⁶³.

A broad-spectrum antibiotic that achieves therapeutic pancreatic tissue levels is important. Imipenem is most widely used because it has been shown to reduce pancreatic infection and mortality ¹⁶⁴. In addition, it appears to reduce the need for surgery and the overall number of major organ complications in acute necrotising pancreatitis ¹⁶⁵. The current recommendation is to administer imipenem-cilastin for 2–4 weeks once the diagnosis of necrotising pancreatitis has been made ¹⁶⁶.

Concerns over multi-resistance and fungal infections have arisen with the more widespread use of antibiotic prophylaxis. A recent prospective study revealed that 24% of pancreatic infections are due to fungi, but these are easily controlled. Infection with multi-resistant organisms was rare but significant as it led to a negative outcome ¹⁶⁷.

On the basis of available evidence many centres consider that early antibiotic prophylaxis is mandatory in the management of severe acute pancreatitis to reduce secondary infections by enteric organisms. Questions still remain, however, over its role in less severe forms of pancreatitis.

Novel therapies

The complex cytokine interactions and inflammatory pathways which mediate SIRS/MODS in acute pancreatitis are being unravelled and may provide potential sites for intervention ¹⁶⁸. Blocking a single cytokine or inflammatory mediator is unlikely to be successful owing to the large redundancy in these pathways ¹⁶⁹. There is a current trend to target multiple sites, and it will be necessary to give consideration to the balance between the pro- and anti-inflammatory responses in any intervention strategy ¹⁷⁰.

Antioxidant therapy

Antioxidant therapy in acute pancreatitis needs to be formally evaluated in a randomised controlled clinical trial. A series of prospective studies have demonstrated that significant oxidative stress occurs within 48 hours of the onset of symptoms and that it is correlated with the severity of acute pancreatitis ¹⁷¹,¹⁷²,¹⁷³,¹⁷⁴. While the source of the reactive oxygen species will include activated neutrophils, it is likely that the intestinal ischaemia and reperfusion also contribute. Bolstering the antioxidant defences should be of benefit, especially during resuscitation when there is the risk of reperfusion injury. As demonstrated in a study of lower limb ischaemia reperfusion, patients with a low antioxidant reserve are at an increased risk of developing SIRS in the postoperative period ¹⁷⁵.

Antioxidants that have improved the outcome of animal models of pancreatitis include N-acetylcysteine ¹⁷⁶, melatonin ¹⁷⁷, lazaroid ¹⁷⁸ and taurine ¹⁷⁹. Human trials with mannitol (a hydroxyl scavenger) and folate (xanthine oxidase inhibitor) along with splanchnic perfusion support decreased the incidence of MODS in critically ill patients ¹⁸⁰. The addition of intravenous lidocaine, vitamin C, selenium, polymyxin B (reduces endotoxin) and hydrocortisone, and enteral administration of glutamine, acetylcysteine, and vitamins A and E reduced the mortality and length of ICU stay ¹⁸¹. Such a ‘shotgun approach’ to therapy fails to isolate the relative benefit of each component, but does highlight the need for further clinical trials of antioxidant therapy in acute pancreatitis.

Anti-cytokine therapy

Almost all the experimental studies examining the effects of anti-cytokine therapies in acute pancreatitis have neglected the effect on the intestine and the intestine-derived cytokine response ¹⁸².

TNF-α and IL-1 are important mediators in SIRS and probably play a role in the early stages of acute pancreatitis. Strategies aimed at blocking these cytokines in human sepsis trials have been disappointing ¹⁸³,¹⁸⁴ but the recent discovery of a single allele in the TNF-α promotor region that is associated with increased risk for sepsis may help explain the broad range of observed responses ¹⁸⁵. Improved survival and lung function follow pretreatment with anti-TNF polyclonal antibody in rodent pancreatitis ¹⁸⁶. Delayed antagonism of TNF-α may be more effective than early treatment, because of the beneficial effect TNF-α has on regulating the quiescent immune system ¹⁸⁷. IL-1 blockade with the naturally occurring receptor antagonist, IL-1ra, attenuates the release of TNF-α, IL-6, amylase, lipase, neutrophil accumulation in the lung, and pancreatic damage in rodent necrotising pancreatitis. Mortality was decreased if IL-1ra was given prophylactically ¹⁸⁸.

IL-10 is a natural anti-inflammatory cytokine that reduces the release of TNF-α, IL-1B and IL-6 when given to caerulein-induced rodent pancreatitis ¹⁸⁹. It also reduces the mortality in diet-induced mouse pancreatitis ¹⁹⁰. Successful transcription of the IL-10 gene into murine pancreas by means of a plasmid/liposome vector also decreased the severity of pancreatitis ¹⁹¹. Two randomised studies have been performed to assess the efficacy of IL-10 in preventing post-ERCP pancreatitis ¹⁹²,¹⁹³. Results were conflicting and the study showing benefit had a particularly high rate of post-ERCP pancreatitis (30%). Further studies are warranted.

Platelet activating factor is a biologically active phospholipid that is activated by phospholipase A₂. It is released from neutrophils, macrophages, platelets and endothelial cells and is an important mediator in the inflammatory processes. It causes intestinal barrier dysfunction by upregulating intestinal enzyme activity and gene expression of phospholipase A₂¹⁹⁴. Antagonism of PAF attenuates intestinal barrier dysfunction in rodent pancreatitis ¹⁹⁵, thereby reducing bacterial translocation ¹⁹⁶. It also reduced lung changes ¹⁹⁷ and mortality ¹⁹⁸. Randomised trials on humans with septic shock have shown reduced organ dysfunction and morbidity when treated with a PAF antagonist ¹⁹⁹. These improvements have not been confirmed in a human trial of severe acute pancreatitis, where PAF antagonism failed to prevent new onset of organ failure or reduce mortality ²⁰⁰.

Other novel therapies aim at the cell adhesion molecules. Treatment with monoclonal antibodies to ICAM-1 and PECAM-1 reduces IL-1 release and leukocyte recruitment so maintaining endothelial barrier function ²⁰¹. Lung injury is ameliorated ²⁰² as is local pancreatic damage ²⁰³. Lung injury can also be prevented in animal pancreatitis by blocking the signals to prime neutrophils that are released by macrophages. Antibodies and carboxamide derivatives have been used to block cytokine-induced neutrophil chemoattraction and result in less lung damage. Unfortunately they do not reduce pancreatic damage ²⁰⁴,²⁰⁵.

It can be expected that therapies based on a better understanding of the cytokine events underpinning severe acute pancreatitis and its attendant MODS will be developed to limit intestinal dysfunction and improve clinical outcome.

Monitoring intestinal function in acute pancreatitis

The wider recognition of the role of the intestine in the pathophysiology of acute pancreatitis provides the impetus to identify markers of intestinal integrity and function that will have clinical and research utility. Intestinal mucosal acidosis, for instance, indicates ongoing splanchnic hypoperfusion, and has been used to predict organ failure and death in critically ill patients ²⁰⁶.

The use of tonometry has already been discussed. The pHi has been used as an index of adequacy of resuscitation after trauma, where it provided an early warning of systemic complications in the post-resuscitation period ²⁰⁷. There are still doubts over its accuracy in assessing splanchnic mucosal blood flow but the importance of monitoring the gastric pHi has been highlighted. A recent modification to the tonometric method has been a recognition that pHi may be replaced by intramucosal pCO₂ or the gradient of intraluminal to end-expired paCO₂²⁰⁸.

Markers of intestinal barrier failure that may prove useful are D-lactate, intestinal fatty acid-binding protein (IFABP) and calcitonin precursors. D-Lactate is a stereo-isomer of mammalian L-lactate and it is strictly the product of bacterial fermentation. The fact that the liver does not metabolise it means that peripheral levels reflect portal venous levels. It has been shown to have discriminatory value in the diagnosis of mesenteric ischaemia in a prospective clinical study ²⁰⁹. IFABP is uniquely located in the cytoplasm of mature enterocytes of the small intestine and has been shown to be a sensitive marker of intestinal ischaemia and SIRS ²¹⁰. Studies have shown that there is a significant increase in IFABP in the serum of patients with severe acute pancreatitis ²¹¹. CTpr, a serum calcitonin precursor, is also raised in severe acute pancreatitis, the concentration of which correlates with the degree of intestinal permeability ²¹².

The use of the polymerase chain reaction (PCR) to identify bacterial DNA fragments in the circulation may yet prove useful as an historical marker of bacterial translocation ²¹³. At present this approach is too sensitive, as the DNA detected represents live invading organisms, dead pre-absorbed organisms and the aftermath of bacterial phagocytosis.

Conclusion

Severe acute pancreatitis is associated with impaired intestinal function. Intestinal dysmotility, ischaemia, hyperpermeability, bacterial translocation and endotoxaemia are all features of the disease. Better understanding of the pathophysiological role of the intestine in acute pancreatitis will result in new treatment paradigms warranting careful evaluation.

It is likely that a management approach will emerge to include resuscitation aimed at restoring intestinal blood flow and oxygenation through the use of appropriate fluids, inotropic agents and monitoring. The addition of antioxidant therapy will help to reduce free radical-mediated reperfusion injury. The integrity of the gut barrier will be further enhanced by therapies aimed at dampening the local inflammatory response, and early enteral nutrition of an appropriate formula. Prophylactic antibiotics and selective gut decontamination will help prevent the added insult of secondary infections.

More is known about the impact of the intestine on the course of acute pancreatitis, than on the impact of acute pancreatitis on the intestine. It is reasonable to assume that there is a reciprocal impact. The management of acute pancreatitis needs to better reflect the important role of the intestine in its pathophysiology.

English or American?

For the first four volumes of HPB the Editor has ‘translated’ all papers into British English in order to achieve uniformity of style. However, as this is an international journal the Editorial Board has decided that from now on manuscripts may be written in either British English or American English as long as usage is consistent throughout the paper.

It is a requirement that laboratory values should be given in SI (Standard International) units. However, if authors so wish American units can be given as well in brackets, e.g. bilirubin 35.9 µmol/L (2.1 mg/dl).

Robin Williamson

Editor