TISSUE CULTURES

Serosal scrapings and lymph nodes were separately homogenised in saline by means of a stomacher (Seward Medical, London, UK). Each homogenate was inoculated onto Columbia blood agar and cysteine lactose electrolyte deficient (CLED) media for aerobic incubation and incubated in air with 5% carbon dioxide and onto Wilkins-Chalgren blood agar with neomycin and Columbia blood agar for anaerobic culture. Anaerobic cultures were incubated in a Don Whitley anaerobic workstation in an atmosphere of 80% nitrogen, 10% hydrogen, and 10% carbon dioxide. Aerobic cultures were incubated at 37°C for two days and anaerobic cultures were incubated at 37°C for five days.

BLOOD CULTURES

Aerobic and anaerobic blood culture bottles were incubated constantly at 37°C for seven days. The aerobic bottles were additionally mechanically agitated for the first 48 hours. Bacterial growth in the blood culture bottles was detected on a Bactec NR730 machine (Becton Dickinson Europe, Meylan Cedex, France) which samples CO₂ in the head space of the culture vials twice daily for the first three days and then once daily for the last four days of the incubation period. The machine indicates a positive culture whenever the culture bottle head space CO₂ concentration exceeds baseline threshold levels. All blood culture vials with positive readings had 0.5 ml of media extracted which was examined by Gram staining and further inoculated onto Columbia blood agar, chocolate Columbia blood agar, and CLED media, and incubated in air with 5% carbon dioxide, and onto Columbia blood agar; incubation was for 48 hours in an anaerobic atmosphere as described previously.3

Isolates grown from mesenteric lymph nodes, serosal scrapings, and peripheral blood samples were identified by characteristic colonial and microscopic appearances and by using standard microbiological tests. Whenever required, full speciation of organisms belonging to the family Enterobacteriaceae was performed using API 20E strips (bioMerieux, France). Enterococcus faecalis and other streptococci were identified as follows: microscopic and colonial appearances; Lancefield grouping by the Streptex kit (Murex, UK); and where appropriate by using API 20 Strep strips (bioMerieux, France). Obligate anaerobic isolates were identified by characteristic colonial growth characteristics, senstivity to metronidazole, microscopic appearances, and rapid ID32A strips (bioMerieux, France). Obligate aerobic isolates and non-fermenting Gram negative bacilli were identified using API 20NE strips (bioMerieux, France). Yeasts were identified by characteristic colonial and microscopic appearances, the germ tube reaction, and API 20AUX strips (bioMerieux, France). Bacillus species and coryneforms were identified by standard microbiological tests.4

Organisms belonging to the genus Staphylococcus were identified by microscopic and macroscopic appearances, catalase reaction, coagulase reaction, and DNAse reaction.

Organisms were broadly grouped into indigenous gastrointestinal flora and bacteria which would not normally be found resident in the gut. Anaerobic organisms were identified by their ability to grow in a strict anaerobic atmoshere. Their Gram reaction susceptibility to metronidazole and biochemical reactions (API) were identified by their microscopic and macroscopic morphology germ tube reaction and by the Candida API biochemical reactions.

Postoperative sepsis was defined as the presence of clinical, haematological, or radiographic evidence of infection where positive bacterial culture was obtained from a septic focus.

STATISTICAL ANALYSIS

All data were expressed as median and interquartile range. Comparisons between non-parametric groups of data were made using the χ² test or the Mann-Whitney U test.

BACTERIAL OVERGROWTH

In this study Gram negative facultative anaerobes, and particularly E coli, were the most common organisms identified in mesenteric lymph nodes and serosal scrapings. Obligate anaerobes were cultured less frequently. In healthy subjects the distal ileum harbours aerobic and anaerobic bacteria at a concentration of approximately 10⁸ organisms. Distal to the ileocaecal valve the bacterial concentration is of the order of 10¹² colony forming units per ml, with anaerobic bacteria outnumbering aerobes by 1000-fold.15 ,16 The predominant isolates areBacteroides spp, Bifidobacterium spp,Eubacterium spp, Lactobacillus spp, andClostridium spp. Intraluminal population density of bacteria would therefore not appear to be the most significant influence on bacterial translocation as one would expect significantly greater numbers of obligate anaerobes to be identified in extraintestinal tissues.17 ,18 Previous authors have found a very low incidence of bacterial translocation by obligate anaerobes.19-21 This has prompted suggestions that obligately anaerobic bacteria are the principal inhibitors of bacterial overgrowth and translocation of E coli and other potentially pathogenic bacteria.17-19 These observations are based again on animal studies where it was noted that a relative increase in numbers of oxygen tolerant bacteria in the distal bowel was associated with an increase in translocation of these organisms without a concomitant increase in anaerobic bacterial translocation. We cultured obligate anaerobes from the mesenteric lymph nodes in 13% of patients (9/69). This is a high proportion of these organisms when compared with previous studies and would appear to refute the theory of an entirely protective role for these organisms in humans. This would be in keeping with the possibility that obligate anaerobes actually translocate at the same rate as facultative anaerobes across the intestinal epithelium but that the obligate anaerobes are more likely to be destroyed in microenvironments which are not oxygen deficient and which may well prevail within the mesenteric lymph nodes or en route to them. Our finding that whenever obligate anaerobes were isolated, they were part of a multiorganism culture involving facultative anaerobes, might reflect in some way poor oxygenation or perfusion of these tissues.

INTESTINAL PERMEABILITY

Migration of organisms across the bowel wall may occur by pinocytosis in epithelial cells and has been proposed as the principle mechanism of translocation in the presence of an intact mucosal barrier.22 ,23 However, many studies have identified alterations in intestinal permeability in critically ill patients. Using dual sugar probe techniques the most frequent finding is that lactulose absorption is increased while the absorption of the smaller molecules is unchanged.21 This suggests that changes in intestinal permeability are caused by loss of tight junction integrity or by areas of cell loss at villous tips. Theoretically this might predispose to passive movement of bacteria across the intestinal epithelium on the presumption that bacteria would follow the pathway of large molecules. In support of this Berg has shown in non-published data that non-viable latex particles of similar diameter to E coli translocate with similar efficiencies to viable indigenous bacteria.23 Such a mechanism could account for the diversity of microorganisms which we found in mesenteric lymph nodes.

IMMUNE STATUS

Under normal circumstances translocating bacteria should be phagocytosed within the mesenteric lymph nodes. This would explain the failure to culture viable bacteria from the MLN of healthy rodents.20 However, if the host is immunocompromised the normal defence mechanisms may fail, permitting egress and survival of these bacteria at distant extraintestinal sites, the inference being 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 cont- rol animals, suggesting a specific role for T cell mediated immunity in inhibiting translocation.24 Immunosuppressant agents and leukaemia have been associated with increased translocation to mesenteric lymph nodes in rodents25-27 and leukaemia has been associated with increased translocation to blood in humans.28Deficiencies in humoral immunity (serum immunoglobulins) have been implicated in the penetration of pathogenic Vibrio cholerae and Salmonella enteritidis into the lamina propria.29 ,30 The possibility that bacterial translocation may not be an all or none phenomenon has been previously suggested.31 Experimental evidence exists in human studies which attest 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.32 This would suggest translocation and subsequent control in the mesenteric nodes. The association between multiorganism colonies and an increasing incidence of postoperative sepsis as shown in this series is compatible with this speculative mechanism of translocation and is probably a manifestation of immunosuppression.33 Furthermore it is of some interest to note that our results do show that the incidence of bacterial translocation was greatest in patients aged over 70, in those undergoing urgent surgery, and in patients with intestinal obstruction, all of whom might be expected to have a degree of impaired immunity. Bacterial translocation may then serve as a promoter of septic morbidity but may not actually be the initiator.

An important finding of this study is the association between bacterial translocation and postoperative sepsis. Forty five per cent of patients in whom bacterial translocation occurred developed a septic focus compared with 19% of patients who showed no evidence of bacterial translocation. Furthermore, the commonest translocating organisms, the Enterobacteriaceae, were also the organisms responsible for postoperative septic complications in 67% (70/104) of cases. Of those patients in whom bacteria were cultured from the MLN and who developed postoperative sepsis, 32% (10/31) grew at least one phenotypically similar species type at both sites (table 7). This is the first compelling evidence from human studies that confirms a potential pathological significance for bacterial translocation. Despite this, we recognise that our data do not prove causality between translocation and septic morbidity and that there are some obfuscating findings in our data. For instance, the occasional culture of non-enteric organisms such as coagulase negative staphylococci from MLNs in patients subsequently identified as having organisms similar to the predominant organism in septic foci is difficult to explain on the basis of intestinal translocation. We accept that these organisms may be skin contaminants and thereby could reflect poor sampling technique. Our experimental protocol endeavoured to avoid such problems and the low incidence in this series of such growth suggests that this was not a major problem. We feel that on no account do these findings invalidate the overall observations made from this study, but rather reflect the inherent methodological problems associated with investigating a pathophysiological phenomenon in a clinical setting with an inevitably heterogeneous patient population. Confirmation of proof of causality between translocation and sepsis would necessitate specific species subtyping of bacteria from the lumen of the gut, MLNs, or other extaintestinal tissue and subsequent septic foci with all their respective temporal relationships in terms of sampling times recorded. This will be logistically very difficult to achieve but does emphasise the need for additional prospective studies in the clinical setting as well as in the controlled environment of the laboratory.

These data also provide a rationale for the use of selective gut decontamination. The principle behind this treatment is that by selective elimination of enteric Gram negative Enterobacteriaceae using specific antibiotic therapy, translocation rates and subsequent postoperative sepsis may be reduced. While some studies have shown significant reductions in septic morbidity the results remain controversial and there is no firm evidence to date of enhancement of survival.34 Our results suggest that selective gut decontamination may have a role to play in the prevention of gut derived sepsis. Our findings of significant translocation by anaerobic bacteria suggest a role for therapy directed against these organisms. We found no evidence of candidal translocation in any of our tissue samples. However, Candida species were cultured from 10% of postoperative septic foci in this study. We have previously reported the presence of Candida spp in gastric microflora35 and although others have shown thatCandida spp can translocate36 ,37 our data would suggest that opportunistic infection caused by Candidaspp probably occurs by a mechanism other than translocation. Consequently, treatment directed against intestinal yeast may be unnecessary. Finally, selective gut decontamination does not influence translocation of bacterial endotoxin which may itself be responsible for continued cytokine release and may influence host susceptibility to postoperative infection. The implication therefore is that manipulations of microflora alone are insufficient to overcome the effects of altered barrier function and diminished immune defences. It is interesting to speculate, therefore, that selective gut decontamination directed only at enteric aerobic and anaerobic bacteria, possibly in conjunction with immune enhancing agents, may provide an optimal treatment strategy for critically ill patients which may significantly decrease septic morbidity.