Current status of bacterial translocation as a cause of surgical sepsis

In recent years it has been increasingly recognized that the gastrointestinal tract has functions other than simply the digestion and excretion of foodstuffs. The gut is also a metabolic and immunological organ that serves as a barrier against living organisms and antigens within its lumen. This role is termed ‘gut barrier function’. The fact that luminal contents in the caecum have a bacterial concentration of the order of 1012 organisms per millilitre of faeces,1 whilst portal blood and mesenteric lymph nodes are usually sterile, dramatically illustrates the efficacy of this barrier function. The idea that the alimentary tract, teeming with its own bacterial flora, could represent a source of sepsis under certain conditions has interested clinicians for many years. This theory, usually referred to as the ‘gut origin of sepsis’ hypothesis, is not new. In the late nineteenth century, the idea developed that peritonitis could result from the passage of bacteria from organs adjacent to the peritoneal cavity. In Germany this was referred to as durchwanderungs-peritonitis, literally translated as ‘wandering through peritonitis’. In 1891 and 1895, two separate investigators hypothesized that viable bacteria could pass through the intact gut wall in vivo.2,3 These findings were confirmed after the Second World War by Fine and colleagues4 who were able to detect viable bacteria in the peritoneal cavity of dogs in a haemorrhagic shock model. Schatten et al.5 demonstrated that bacteria could migrate from the gastrointestinal tract into the portal circulation in the absence of an infective process in humans. Subsequently it has become clear that, in addition to Gram-negative bacteria, endotoxin, Gram-positive bacteria and fungi can pass through the mucosal barrier.6 In 1979, Berg and Garlington7 defined this phenomenon as ‘bacterial translocation’. The term ‘bacterial translocation’ is used to describe the passage of viable resident bacteria from the gastrointestinal tract to normally sterile tissues such as the mesenteric lymph nodes and other internal organs.7 The term also applies to the passage of inert particles and other macromolecules, such as lipopolysaccharide endotoxin, across the intestinal mucosal barrier. There is no doubt that bacterial translocation occurs in humans.8 It has a prevalence of about 15% in elective surgical patients and occurs more frequently in patients with intestinal obstruction and those who are immunocompromised. There is good evidence to show that translocation is associated with an increased incidence of septic complications, but not with mortality.9 Many studies have established an association between gastrointestinal microflora and nosocomial infection, supporting the concept of the gut as a reservoir of bacteria and endotoxin.10,11 However, the evidence that the mechanism of this association between enteric organisms and subsequent sepsis is bacterial translocation remains, at least in humans, largely circumstantial. This may reflect the methodological and ethical difficulties involved in obtaining samples of portal venous blood or mesenteric lymph nodes, which are necessary for unequivocal confirmation of the occurrence of translocation. To date there has been only one clinical study providing compelling evidence for a mechanistic link between translocation and the development of late sepsis.12 This study demonstrated that septic complications in 448 surgical patients were significantly more prevalent in those who showed bacterial translocation when compared with patients with no organisms in their mesenteric lymph nodes collected at laparotomy (P < 0.001)12. Furthermore, the spectrum of organisms responsible for septic morbidity was similar to those observed in the mesenteric lymph nodes, many being from enteric strains. These data strongly support the gut origin of sepsis hypothesis. Whilst it is tempting to assume that any bacteria or endotoxin passing through the intestinal barrier might cause septic complications in the host, there is growing evidence to suggest that translocation may in fact be a normal phenomenon. It is possible that translocation occurs so that the alimentary tract can be exposed to and sample antigens within the lumen such that the gut can mount a controlled local immune response helping to keep these antigens away from the internal milieu; this process is known as ‘oral tolerance’. It is only when the host’s immune defences are overwhelmed that septic complications arise. Other authors go further and depict translocation not as the initiator of septic complications, but as a result of other insults which have triggered off a systemic inflammatory response, one of the effects of which is manifested as deterioration or breakdown of gut barrier function.13 Bacterial translocation has been shown to occur in various patient populations.6 As already stated, it occurs in patients undergoing elective abdominal surgery, organ donors and those with intestinal obstruction, colorectal cancer, ischaemia–reperfusion injury shock and pancreatitis. Many authors suggest an increased prevalence in patients with obstructive jaundice, those receiving parenteral nutrition and the malnourished, but the evidence for this is limited. Interestingly, translocation, assessed by endotoxin or bacterial culture of portal or systemic blood, has only rarely been demonstrated after trauma.14,15 Host defence mechanisms directed against invasion by microbes comprise many factors, such as gastric acid, pancreatic enzymes, bile, mucus, bowel motility, the antigen-specific local immune system known as gut-associated lymphoid tissue (GALT) and, arguably most important, the epithelial cell barrier.16 The intestinal epithelium is a polarized monolayer of enterocytes covered with mucus, a glycocalyx brush border and secretory immunoglobulin A (sIgA). It represents a specialized anatomical barrier separating the luminal contents of the bowel from the internal milieu. Physical breaches in this barrier, which can occur following ulceration, may predispose to translocation. A rapid rate of cellular turnover, specialization and migration helps the intestinal mucosa to maintain physical integrity. Translocation may occur via transcellular or paracellular routes, or a combination of the two. Transcellular translocation is under the control of specific membrane pumps and channels, whereas paracellular translocation is permitted (at least theoretically) by breakdown of tight junctions. Transcellular migration has been shown to occur in rats where various organisms, including Escherichia coli and Proteus mirabilis, were visualized within intact enterocytes.17 There is in vitro evidence that opening up the gaps between enterocytes by loosening the intercellular tight junctions may increase bacterial translocation.18 Tight junctions can adjust their degree of ‘leakiness’ to meet physiological needs, and are also susceptible to numerous pathological stimuli that may change the intestinal permeability. However, it is important to emphasize that there are, as yet, no data from animal or human studies that directly confirm a causative relationship between changes in permeability and bacterial translocation.19 Similarly, there is no evidence to confirm that gross changes in villus morphology are causally related to increased rates of translocation. The human gastrointestinal tract contains a wide variety of aerobic and anaerobic bacteria. Absolute bacterial counts vary along the length of the bowel, increasing from a concentration of 108 organisms/ml in the region of the distal ileum to ∼1012 organisms/ml beyond the ileocaecal valve. The upper gut and stomach are usually sterile or are sparsely populated with relatively avirulent bacteria. It is a remarkable testimony to the efficiency of the intestinal barrier that large populations of indigenous bacteria reside in the healthy lower gut without causing harm. This indigenous microflora exerts an important influence in preventing colonization with exogenous pathogens, so-called colonization resistance. Many factors have been implicated in the regulation of the different populations of microflora. These include the creation of ‘microclimates’ whereby facultative bacteria, for example, utilize oxygen, thereby ensuring a suitable environment for obligate anaerobes. This has prompted suggestions that obligate anaerobic bacteria may be the principal inhibitors of translocation of E.coli and other potentially pathogenic bacteria. Disruption of the normal flora may predispose to a breakdown in colonization resistance. This may occur, for example, with antibiotic therapy and there is evidence that substrate utilization by different organisms will also influence other bacterial concentrations.16 Critical illness is often associated with significant proximal gut overgrowth of enteric organisms, which may contribute to nosocomial infection.12 The similarity in the spectrum of organisms identified in septic foci and those cultured from gastric aspirates suggests that the infecting organisms are of gut origin. This infers, but does not prove, a role for bacterial translocation. Marshall et al.11 showed that >90% of their intensive care unit (ICU) patients with infection had at least one episode of infection with an organism that was simultaneously present in the upper gastrointestinal tract. In our own studies we have consistently found organisms in mesenteric lymph nodes that are similar to those identified in septic foci as overgrowth in the upper gut.9,12,19 All these data point to the importance of the gastrointestinal microflora in maintaining health and that disruption is associated with illness. However, it should be emphasized that the influence of enteric bacteria is not simply related to population density of specific bacteria. Gram-negative facultative anaerobes, particularly E.coli, are consistently shown to be the most common organisms identified in mesenteric lymph nodes. Obligate anaerobes are known to translocate but rarely do so despite their luminal concentrations. Thus it appears that absolute intraluminal population density is not a major determinant of translocation and therefore it is unlikely that rates of translocation will be influenced by alterations in gut permeability. Under normal circumstances, translocating bacteria should be phagocytosed before reaching mesenteric lymph nodes or lymphatic vessels. The presumption is that if the host is immunocompromised the normal defence mechanisms fail, permitting egress and survival of these bacteria at distant extra-intestinal sites. The inference is that with increasing severity of illness, bacterial translocation occurs because of the inability of the host to deal adequately with the numbers of bacteria present. Much experimental evidence exists to support this concept. Bacterial translocation is more common in athymic mice than in control animals, suggesting a specific role for T-cell-mediated immunity in inhibiting translocation.20 Immunosuppressant agents and leukaemia have been associated with increased translocation to mesenteric lymph nodes in rodents, and leukaemia has been associated with increased translocation to blood in humans. The possibility that bacterial translocation may not be an all-or-none phenomenon has been previously suggested. There is experimental evidence in human studies attesting to the invariable presence of E.coli β-galactosidase in the cytoplasm of mesenteric lymph nodes but positive cultures in only 5% of the same patients.21 This suggests translocation and subsequent control in the mesenteric nodes. The association between multi-organism colonies and an increasing incidence of postoperative sepsis, as shown in our own studies,12 is compatible with this speculative mechanism of translocation and is probably a manifestation of immunosuppression. Furthermore, it is of some interest to note that our results show that the incidence of bacterial translocation was greatest in patients aged >70 years, those undergoing urgent surgery and those with intestinal obstruction, all of whom might be expected to have a degree of impaired immunity.8,12 Bacterial translocation may thus serve as a promoter of septic morbidity, but not actually be the initiator. Many other factors have been shown to influence translocation, particularly in animals, and these have been extensively reviewed elsewhere.6,16,22 These include stress, jaundice, radiation, variations in intestinal motility and various drugs. Prokinetic agents such as cisapride and propranolol have been shown to decrease translocation, as have sucralfate, bile salts and prostaglandin analogues. Other drugs have been shown to increase translocation. These include chemotherapeutic and immunosuppressive agents as well as certain antibiotics. There is now a consensus that perioperative nutritional support is of benefit, particularly to patients with severe malnutrition. Most reviewers of nutritional support therapy urge the use of enteral nutrition (EN) as opposed to parenteral nutrition (TPN), on the assumption that the latter is associated with mucosal atrophy and increased intestinal permeability which reflect damage to the intestinal barrier. This, it is often stated, predisposes to bacterial translocation and may be one explanation for increased rates of septic complications observed in some studies investigating TPN. A number of assumptions are implicit in these commonly held views about TPN. The first assumption is that bacterial translocation does occur in humans, occurs more readily if intestinal barrier function is impaired and is associated with increased incidences of sepsis. The second is that septic morbidity has been proved to be significantly higher in patients receiving TPN. The third is that the absence of luminal nutrients, for example during starvation, malnutrition or TPN, is associated with deleterious consequences to the gut barrier which predispose to translocation. Bacterial translocation does occur on a frequent basis in humans. It is associated with increased septic morbidity supporting the concept of the gut origin of sepsis hypothesis. However, it probably also occurs in healthy individuals but is not significant in the non-immunocompromised host. There is no evidence to suggest that bacterial translocation is reduced by the use of EN or increased in patients receiving TPN.23 There is no evidence to confirm that short-term TPN is associated with villus atrophy or significant changes in intestinal permeability19 or to support the view that alterations in intestinal barrier function, as assessed from changes in mucosal architecture or from alterations in intestinal permeability, will predispose to an increased prevalence of bacterial translocation in humans. Starvation or malnutrition by themselves do not induce bacterial translocation. Alterations in mucosal architecture or intestinal permeability may indicate changes in intestinal barrier function, but do not necessarily equate to alterations in the prevalence of bacterial translocation. With the exception of trauma patients, there is no firm evidence that septic morbidity is increased in patients receiving TPN as opposed to EN.24,25 The nature of the nutritional support that should be provided to patients should be determined by their tolerance to EN and not by unproven assumptions concerning the role of gut barrier function. There is an increasing body of knowledge that suggests that certain nutrients have specific effects on gut function. These are sometimes referred to as ‘functional foods for the gastrointestinal tract’. For example, glutamine is known to be the preferred fuel source of the enterocyte and has been shown to impact on intestinal permeability, and may reduce rates of translocation in animal studies. Other ‘gut-specific nutrients’ include arginine, pre- and probiotics and certain antioxidants such as vitamin A. An excellent review is given by Duggan.26 There is now much circumstantial evidence to support the view that translocation is associated with an increase in septic morbidity. To date, however, there has been only one study attempting to prove a mechanistic link between translocation and the development of late sepsis. In this prospective controlled study of 279 surgical patients, cultures of nasogastric aspirates were compared with those from mesenteric lymph nodes taken at laparotomy and with organisms cultured from subsequent septic complications.12 Only 31% of patients had a sterile nasogastric aspirate; the most frequently identified organisms were Candida spp. (54%) and E.coli (20%). Multiple organisms were isolated in 39% of patients but were cultured more frequently in patients aged >70 years (P < 0.01), those undergoing non-elective surgery (P < 0.001) and those requiring proximal gastrointestinal surgery (P < 0.01). Postoperative sepsis was more common in these patients (P < 0.01). Bacterial translocation, assessed by culture of mesenteric lymph nodes, occurred in 21% of patients and was significantly more frequent in those with multiple organisms in their nasogastric aspirates (P < 0.01). Escherichia coli was the most common organism to be isolated from the lymph node specimens (48%) and septic foci (53%). Fungal translocation did not occur. An identical genus was identified in the nasogastric and septic in of patients, in the nasogastric and the lymph node in and in the lymph node and a postoperative septic in The authors that proximal gut colonization is associated with increased bacterial translocation and septic morbidity. The of organisms identified was as compelling evidence in support of the gut origin of sepsis and that bacterial translocation is associated with an increase in postoperative septic morbidity. The importance of the gut origin of sepsis in the or of multiple organ has been reviewed The data in support of from the gut are studies have demonstrated that the gut is the reservoir for in late In there is much circumstantial patients are often on TPN and no enteral for It has been not that this may be associated with increased rates of septic morbidity related to bacterial translocation. These patients also frequently gut with and all of which can to alterations in gastrointestinal microflora which may in predispose to translocation. have been shown to reduce Thus sucralfate, which gastric may reduce nosocomial has shown that of with and the in surgical patients can also significantly reduce postoperative which may be a of gut these data support the view that the gut origin of sepsis is as an explanation for some of sepsis in the are to for the development of systemic inflammatory response or There is increasing support for the concept of a to the clinical and metabolic changes that occur in multiple organ In this first cell trauma as a of or or cell from results in an inflammatory response that is and potentially It unlikely that bacterial translocation has any role in the of these In some patients, a second occurs as a result of surgery, or cell damage related to the The suggests that the inflammatory response results in an response or an between and In other an occurs which is the manifestation of severe and sometimes cell damage by Bacterial translocation has been as or this inflammatory as endotoxin is known to induce the of and There is evidence from animal studies to support this basis of and In there is good evidence to suggest that the mechanism of the gut is an ischaemia–reperfusion In humans, the at the of the villus is lower than that in As a epithelial cell injury may in any in which tissue is This injury may bacterial translocation, which as a of the inflammatory for this in human is et significantly higher of in the portal than in the systemic circulation in patients undergoing and the first of after into the These data are in with the possibility that, on reaching the endotoxin or bacteria which without it being necessary to first endotoxin or bacteria via the portal to the or mesenteric lymph nodes. This also for the of clinical studies to consistently endotoxin or bacteria in patients with or to the in is on a mechanism such that at the of the is lower than that of any in blood is to result in ischaemia–reperfusion blood as in shock and bowel is associated with translocation and septic complications in animal The importance of as a determinant of in with illness is by the findings of a recent study which demonstrated that the use of the was associated with a significant in this was a result of reduced translocation is It has been that results in the gut a major of The possibility exists that can to an ischaemia–reperfusion injury to the with a of gut barrier function and an gut inflammatory response without the for translocation of microbes as as the mesenteric lymph nodes or endotoxin or bacteria the mucosal barrier, if within the gut can an immune response such that the gut a and other which the local and the systemic immune to and has termed this of gut It ischaemia–reperfusion injury and the of gut barrier function in association with a increased In this it is not that the gut barrier is simply a barrier to the passage of intraluminal bacteria to distant but that bacteria or gut or cause the gut-associated lymphoid tissue to factors that contribute to distant organ The gut origin of sepsis is an and concept that that bacteria the intestinal barrier and cause sepsis at distant sites. There is now good evidence from human to support this theory, which that translocation predisposes to an increase in septic morbidity. Bacterial and endotoxin translocation probably also occurs to a on a basis in healthy In this translocation serves to an but normal barrier function is and morbidity does not Only if normal mechanisms of defence are overwhelmed does translocation occur. there doubt that the phenomenon of bacterial translocation is associated with septic morbidity. However, it is important to this in the of the surgical or morbidity is Bacterial translocation is to an in only a of Other factors are important in the of postoperative septic morbidity. These may to colonization of the upper or the presence of such as and and be directly to the phenomenon of translocation. Similarly, whilst it is that translocation is a in the of in those patients who the first of their it is in the present of our knowledge to the importance of this compared with other In it is to that the of the gut origin of sepsis hypothesis, whilst confirmed in certain patients has not been associated with any in patient In at gut the use of pre- or alterations in antibiotic and bowel have not translated into to patient care with to septic morbidity. Bacterial translocation a to those with an interest in the metabolic care of the surgical It is as yet, a to change clinical