Bacterial translocation (BT) is defined by the passage of viable indigenous bacteria from the intestinal lumen to mesenteric lymph nodes (MLNs) and other territories, and its diagnostic criteria rely on the isolation of viable bacteria in MLNs. Small intestinal overgrowth, increased intestinal permeability and immunological alterations are the main factors involved in its pathogenesis. BT is obviously difficult to identify in patients with cirrhosis, and alternative methods have been proposed instead. Bacterial DNA detection and species identification in serum or ascitic fluid has been proposed as a reliable marker of BT. Bacterial products, such as endotoxin, or bacterial DNA can translocate to extra-intestinal sites and promote an immunological response similar to that produced by viable bacteria. Therefore, pathological BT plays an important role in the pathogenesis of the complications of cirrhosis, not only in infections, but by exerting a profound inflammatory state and exacerbating the haemodynamic derangement. This may promote in turn the development of hepatorenal syndrome, hepatic encephalopathy and other portal hypertension-related complications. Therapeutic approaches for the prevention of BT in experimental and human cirrhosis are summarized. Finally, new investigations are needed to better understand the pathogenesis and consequences of translocation by viable bacteria (able to grow in culture), or non-viable BT (detection of bacterial fragments with negative culture) and open new therapeutic avenues in patients with cirrhosis.
Bacterial infections are frequent complications in patients with cirrhosis, with an incidence on admission or development during hospitalization of about 32% [1, 2]. This incidence contrasts with that of nosocomial infections in the general population, which is 5 to 7%. Bacterial infections in patients with cirrhosis are associated with a poor prognosis and an increased risk of mortality . Moreover, bacterial infections or the use of antibiotics as a surrogate marker of infection are the most important independent variables associated with failure to control bleeding from oesophageal varices . Bacterial infections in addition to being an important prognostic factor for the development of variceal bleeding are clearly associated with its occurrence.
The most frequent infections in cirrhotic patients are Spontaneous Bacterial Peritonitis (SBP), Urinary Tract Infections and pneumonia, of which 80% are caused by Gram-negative bacilli (GNB), especially Escherichia coli. This fact suggests that most of the infection episodes in patients with cirrhosis are of enteric origin. Passage of viable bacteria from the intestinal lumen through the intestinal wall and to mesenteric lymph nodes (MLNs) and other sites is therefore the accepted pathogenic mechanism to explain the development of spontaneous infections, such as SBP or bacteraemia. This phenomenon, first described in 1979, is defined as bacterial translocation (BT) . BT is also present in other clinical scenarios different from liver cirrhosis such as haemorrhagic shock, intestinal obstruction, major trauma, burns and severe acute pancreatitis [6, 7].
Bacterial translocation occurs in 25–30% of patients with cirrhosis with liver dysfunction according to clinical studies  and up to 45–78% of rats with cirrhosis and ascites [9-11]. BT not only is the key factor leading to development of spontaneous bacterial infections but also plays a role in the immune and haemodynamic changes in advanced cirrhosis, either when translocation is caused by viable bacteria or their fragments, such as endotoxin or bacterial DNA, what we define as non-viable BT.
Bacterial translocation may caused by viable bacteria, thereby having the possibility of inducing ‘spontaneous’ bacterial infections, or by bacterial fragments, such as endotoxin or bacterial DNA, that induce release of pro-inflammatory cytokines and nitric oxide. In this manuscript, the term ‘pathological’ BT defines a morbid stage associated with clinical and pathophysiological implications in patients with cirrhosis, and we will discuss the mechanisms involved in the pathogenesis of pathological BT, the available diagnostic techniques for its detection and the consequences of the translocation of bacteria and their products in patients with cirrhosis.
Pathophysiology of bacterial translocation
Our knowledge of the pathogenesis of BT is based mainly on studies in experimental models because of the obvious difficulty of accessing to MLNs in patients with cirrhosis with the exemption of certain research environments . In experimental models, BT is defined as a MLN-positive culture, reaching 56% of rats with CCl4-induced cirrhosis and ascites had BT while this figure drops to 0–10% in animals without ascites . Patients with cirrhosis and BT had a higher degree of liver dysfunction (estimated by Child-Pugh score) in comparison with patients without BT . In another study, BT was investigated at two stages of experimental portal hypertension: acute (when shunting is minimal); and chronic (when shunting is extensive and mimics the portal hypertension of cirrhosis). The authors of this study found that BT is a frequent event in the acute portal hypertension model probably because of a greater mesenteric inflammation; however, there was no difference in the rate of BT in chronic portal hypertensive rats (15 days after portal vein ligation) compared with control rats . Hence, portal hypertension alone may not be a major factor in the development of BT in cirrhosis. Recently, our group showed that patients with cirrhosis and ascites and the presence of bacterial DNA in serum, a surrogate marker of BT, had the same degree of portal hypertension than patients without bacterial DNA . These studies suggest that the degree of liver failure and not portal hypertension per se plays an important role in the pathogenesis of BT.
The mechanisms that influence the pathogenesis of BT are at least three: intestinal bacterial overgrowth, immune alterations of cirrhosis and increased intestinal permeability.
Small intestinal bacterial overgrowth (SIBO)
Intestinal bacterial overgrowth is a very heterogeneous syndrome characterized by an increased number and/or abnormal type of bacteria in the small bowel . Most authors consider diagnostic of SIBO to be the finding of ≥105 colony-forming units per ml of proximal jejunal aspiration. Therefore, the gold standard for diagnosing SIBO is microbial investigation of jejunal aspirates. However, some limitations difficult its application in clinical practice: it is an invasive test, microbial investigation places high demands on the quality of laboratory work, distribution of bacterial overgrowth might be irregular and more than one sample is needed to detect it. Other non-invasive diagnostic methods more frequently used in clinical practice are hydrogen and methane breath tests after glucose or lactulose challenge . However, when breath tests are compared to properly obtained quantitative jejunal cultures, no tests performs perfectly.
Small intestinal bacterial overgrowth, which is likely related with a slowed intestinal transit, low acid gastric secretion, intestinal immunological factors and pancreatic and biliary secretions, is one of the main factors promoting BT. A direct relationship between numbers of a specific bacterial strain populating a segment of the intestine and numbers of viable bacteria of this strain present in MLN has been demonstrated in mice . Moreover, experimental studies have shown that cirrhotic rats with ascites and BT have a higher rate of SIBO compared with those without BT . The absence of SIBO is associated with a low rate of BT (0–11%) and the rate of BT is comparable to that observed in control rats . However, the fact that BT is not present in 50% of rats with SIBO suggests that other factors besides SIBO are involved in the pathogenesis of BT. Clinical studies have shown that SIBO is more common in patients with cirrhosis than in controls, especially in patients with advanced liver dysfunction  and in those with a history of previous episodes of SBP . However in these studies, SIBO was estimated by the breath hydrogen test, which is not a reliable method for diagnosis. In a study that estimated SIBO by quantitative cultures of jejunal aspirates, the occurrence of SBP did not correlate with the presence of SIBO. Interestingly enough, however, authors found a significant correlation among the incidence of SBP and protein levels in ascitic fluid, a marker of immunity local, and serum bilirubin levels . Therefore, other factors such as immunological abnormalities and liver insufficiency may play a relevant role in the pathogenesis of BT.
Increased intestinal permeability
The intestinal barrier is formed mainly by a mucinous component secreted by intestinal epithelial cells and intestinal epithelium per se, which forms a layer with intercellular junctions (‘tight junctions’) that allow selective passage of substances. Structural and functional alterations in the intestinal mucosa that increase intestinal permeability to bacteria and its products have been described in cirrhosis. Experimental studies have observed a mucous congestion, dilatation of intercellular spaces in the intestinal epithelium, oedema and submucosal inflammatory changes in the intestinal barrier [22-25]. However, it is not clear whether these structural changes are the cause or the result of BT.
The mucin secreted by epithelial goblet cells in large amounts (3 L/day) form a thick layer of glycoproteins that prevents a direct contact of the bacteria with enterocyte microvilli. Furthermore, intestinal mucous secretions contain immunoglobulin A, which neutralizes toxins and other bacterial products and binds to bacteria thereby preventing its adhesion and colonization in the intestinal epithelium.
Bile secretions also play a role in the prevention of BT by inhibiting bacterial overgrowth, exerting a trophic effect on intestinal mucosa and neutralizing endotoxin . Therefore, bile acids prevent BT and avoid the passage of bacterial products from the lumen of intestine [27, 28].
From a different perspective, studies have demonstrated increased intestinal permeability to macromolecules in patients with advanced cirrhosis, especially in those with previous episodes of SBP or hepatic encephalopathy  or sepsis . Oxidative damage of intestinal mucosa  and endotoxaemia, elevated levels of NO and inflammatory cytokines may play a role in increasing intestinal permeability in cirrhosis [32, 33]. Most studies of intestinal permeability are performed with probes that measure paracellular permeability, such as lactulose , while BT of living bacteria follows mostly a transcellular route. Therefore, results obtained in studies of intestinal permeability may not be associated with clinical consequences. This may explain why the increase in intestinal permeability alone does not seem to be a determining factor in the pathogenesis of BT. Experimental studies have demonstrated that BT occurs in up to 87% of rats with increased intestinal permeability and SIBO, whereas none of the animals with only increased intestinal permeability showed the presence of BT . Furthermore, only therapy against SIBO, without interfering in intestinal permeability, is able to decrease the rate of BT in cirrhotic rats .
The intestinal tract is an active immune organ, containing essentially every type of leucocyte involved in immune response. The intestinal immune system consists of the gut-associated lymphoid tissue, the largest immunological organ of the body, which comprises four lymphoid compartments: Peyer's patches, lamina propria lymphocytes [including dendritic cells (DCs)], intraepithelial lymphocytes and MLN, which are involved in both the adaptive and innate responses. Changes in local and systemic immunity are clinically relevant to promote BT in cirrhosis. Bacteria that translocate to MLNs or to portal blood are usually phagocyted and neutralized before they begin to grow and cause bacteraemia or infections in immunocompetent patients. Patients with cirrhosis have systemic immune alterations that may promote the development of infections and BT.
Some of these alterations are genetically driven. For example, nucleotide-binding oligomerization domain-containing protein 2 (NOD2) plays an important role in the local immune system, recognizing bacterial molecules (peptidoglycans) and stimulating an immune reaction. Recently, Appenrodt et al. demonstrated that the occurrence of SBP was significantly increased in carriers of NOD2 variants, suggesting that local immune alterations might be implicated in BT .
Advanced cirrhosis is associated with a decrease in the cellular and humoral components of immune response [36, 37] being correlated with the likelihood of developing SBP . Cirrhosis has been also associated with a decreased activity of the reticuloendothelial system (RES) that is one of the most relevant defence systems against bacteraemia and other infections acquired haematogenously. The presence of porto-systemic shunts and decreased phagocytic capacity of Kupffer cells is associated with the development of bacteraemia and SBP . The altered clearance capacity of RES does not only affect viable bacteria but also bacterial products such as endotoxin or bacterial DNA. Translocation of bacterial products promotes a chronic inflammatory response and impairs further the haemodynamic changes observed in cirrhosis. To date, few studies have focused on the immunological alterations of the intestinal barrier favouring BT in cirrhosis. Inamura et al. have demonstrated a relationship among decreased cellular proliferative capacity and interferon-gamma synthesis by intraepithelial lymphocytes in cirrhotic mice and rate of BT .
The synthesis of cytokines, particularly TNF-alpha, interleukins and NO exacerbates oxidative damage in the intestinal mucosa , which in turn increases intestinal permeability probably favouring BT. From a different perspective, our group has demonstrated that administration of anti-TNF monoclonal antibodies to cirrhotic rats with ascites was associated with a significant decrease in the rate of BT . Therefore, the inflammatory response induced by BT also acts on the intestinal barrier permeability favouring bacteria and other bacterial products translocation, thus creating a feedback in which BT itself perpetuates the pathogenic mechanisms that cause it.
However, portal hypertension-associated splanchnic hyperaemia hinders the recruitment of leucocytes in response to inflammatory stimulus such as leucotriene administration , suggesting that the local immune response is not sufficient to prevent the passage of bacteria from intestinal lumen to the systemic circulation. Moreover, the ability of MLN to retain and destroy bacteria is reduced, possibly by the repetition of these events. As a logical consequence of these local and systemic immunological disorders, viable bacteria or their products (endotoxin or bacterial DNA) may access the systemic circulation and areas with low bactericidal capacity such as ascitic fluid. These immunological abnormalities as a whole play an important role in the pathogenesis of BT, spontaneous bacteraemia, SBP and the development of complications associated with BT.
In summary, these three disorders discussed previously (intestinal bacterial overgrowth, increased intestinal permeability and immune system disorders) play a relevant role in the pathogenesis of BT in cirrhosis, which explain at least in part the high rate of BT with respect to other clinical situations in which there is only one pathogenic factor involved (Fig. 1).
Diagnostic methods of bacterial translocation
In experimental studies, BT is defined as a positive-culture MLN. Studies of BT in humans are limited because of the need for surgery and the removal of MLN in non-optimal conditions (e.g. peri-operative antibiotics). Therefore, alternative approaches to diagnosing BT in humans have been postulated.
Lipopolysaccharide (LPS) is a major component of the gram-negative bacterial wall that when present in systemic circulation in a patient might be considered a surrogate marker of BT. However, there are several reasons for not considering LPS measurement as a sole tool in detecting BT: the Lymulus Amebocyte Lysate was originally designed to measure LPS in water, and protocols are not specifically directed to quantify LPS in biological samples. Furthermore, LPS levels have a short half-life and may be influenced by several factors, such as the concentration of LPS transporters, antibodies, HDL and other immunogenetic, microbiological and physiological variables , and because of the need of using endotoxin-free systems for samples collection. These drawbacks are reflected in the rate of endotoxin detection in cirrhosis, which ranges from 0 to 93% in different studies .
Lipopolysaccharide binding protein (LBP) measurement, a protein with a relatively long half-life synthesized by the liver in response to bacteraemia or endotoxaemia, has been proposed as a surrogate marker of BT. Patients with elevated LBP levels have been shown to have a pro-inflammatory state and haemodynamic derangement, which may be reversed after intestinal decontamination with norfloxacin . Interestingly, a prospective study in non-infected patients with cirrhosis and ascites shows a four-fold increase in the risk of developing bacterial infections in patients with an elevated LBP vs normal LBP . However, LBP only reflects GNB and not gram-positive translocation, which represents a major drawback in the investigation of BT .
Serum presence of peptidoglycan, a polymer consisting of sugars and amino acids that forms part of the cell wall of gram-positive bacteria, has been investigated as a marker of BT in only one experimental model of haemorrhagic shock . However, more studies are needed to confirm the utility of the detection of this bacterial product as a surrogate marker of BT in cirrhosis.
Over the last decade, polymerase chain reaction-based detection of bacterial DNA (bactDNA) has been proposed as a surrogate marker for BT because it has been detected in blood and ascites of approximately one-third of patients with cirrhosis and culture-negative ascites [50, 51], a value similar to that reported in other studies in humans . Detection of bactDNA in biological fluids in animals with experimental cirrhosis and ascites is associated with its simultaneous presence in MLNs, in either culture-positive or culture-negative MLNs . These data support the hypothesis that the detection of bactDNA in biological fluids in patients with advanced cirrhosis constitutes a surrogated marker for the diagnosis of BT (Table 1). However, detection of bacterial products does not imply the viability of the accessing bacteria, and therefore the clinical consequences of their presence may be different compared with the presence of viable bacteria. The rates of bactDNA detection in the literature in this setting are different; probably because of the fact that there is not a standard method of detection that probably justifies variations reported [35, 53, 54]. Therefore, uniformity of analytical methods is needed to ascertain its real value in clinical setting.
Table 1. Different biological markers to identify bacterial translocation
Influenced by several factors: LBP levels, serum lipoproteins
Variable detection: 0–90%
Only detects BT from GNB
Predicts clinical outcomes
Indirect marker of BT
Only detects BT from GNB
GPC (and GNB)
Not widely validated
GPC and GNB
Predicts clinical outcomes
Validated in experimental models
Not fully validated by other groups
Clinical implications of bacterial translocation
The immune system has traditionally been divided into the innate and adaptive components. The adaptive component is organized around two types of cells, B and T lymphocytes, and it is based on the specificity of each cell receptor for a particular antigen. Once a lymphocyte recognizes an antigen, specific clonal expansion occurs in the lymphocyte, which is necessary to generate an effective immune response. However, the process by which adaptive immune system produces a clonal expansion of lymphocytes and their differentiation into effector cells usually lasts 3–4 days, more than enough time for host's damage. By contrast, the innate immune system component, which includes antimicrobial peptides, macrophages, DCs and the alternative complement pathway, is activated immediately after infection, and can rapidly control the replication of the pathogens.
Toll-like receptors (TLR) are a type of membrane proteins involved in innate immune responses that recognize various conserved molecular patterns of pathogens (PAMPs). To date, 10 types of TLR have been identified, each of which recognizes conserved structures of various pathogens or PAMPs. The most studied TLR in cirrhosis are TLR-4, which recognizes LPS, TLR-2, which is activated with the presence of the peptidoglycan wall component of GPC, and TLR-9, which recognizes bactDNA. Activation of these receptors is followed by the release of pro-inflammatory cytokines that can worsen the haemodynamic changes of cirrhosis, especially splanchnic vasodilation and hyperdynamic syndrome. Riordan et al. investigated TLR expression in peripheral blood mononuclear cells in patients with cirrhosis, and they found that TLR-2 expression is upregulated, whereas TLR-4 expression is unaltered or downregulated, suggesting an important stimulatory role for GPC, but not for GNB . However, other bacterial products than LPS, such as bactDNA, lipoproteins and heat-shock protein 60, could stimulate different TLRs. All known TLR agonists have been shown to induce tumour necrosis factor (TNF) secretion by monocytes. Therefore, it might be speculated that in cirrhosis, priming of mononuclear cells and associated release of pro-inflammatory cytokines is mediated by an upregulation of TLR expression .
Bacterial DNA, unlike the DNA of vertebrates, contains multiple sets of unmethylated dinucleotides (so-called ‘CpG motifs’) that through the interaction with TLR-9 induce an inflammatory response . Our group has previously shown that bactDNA activates cell-mediated immune response and NO overproduction by peritoneal macrophages from patients with cirrhosis and ascites . Moreover, the presence of bactDNA in patients with non-infected ascites is associated with a soluble immune response similar to that in patients with SBP. The bactDNA-associated inflammatory response was directly proportional with the serum concentrations of bactDNA, confirming the role of the translocation of bacterial DNA in the immune response .
Bacterial translocation can also exacerbate the hepatic and systemic haemodynamic abnormalities of liver cirrhosis. Experimental studies also suggested that BT is associated with further deterioration of the hyperdynamic circulation of cirrhosis. Cirrhotic rats with BT have an increased activity of endothelial nitric oxide synthase (eNOS) in mesenteric vasculature mediated by increased levels of TNF-alpha and eNOS cofactor tetrahydrobiopterin . Studies in humans have demonstrated that cirrhotic patients with increased levels of LBP, an indirect marker of BT, have a more profound systemic haemodynamic derangement, which could be reversed after selective intestinal decontamination with norfloxacin . Moreover, bacterial DNA translocation in patients with ascites and portal hypertension aggravates the systemic circulatory dysfunction, further exacerbating the peripheral vasodilation; which was related with increased inflammatory state estimated by the presence of higher plasma levels of TNF-α . It has been suggested that the worsening of the hyperdynamic circulation and the inflammatory state associated with BT may play a role in the complications of portal hypertension, especially the development of hepatorenal syndrome (HRS). Recently, experimental studies have demonstrated that kidneys in cirrhosis show an increased expression of TLR4, and the pro-inflammatory cytokine TNF-α, which makes them susceptible to a further inflammatory insult during BT . Angeli et al. investigated the presence of bactDNA in patients with refractory ascites, demonstrating that BactDNA translocation was associated with an impaired cardiovascular and renal functions and a higher risk of HRS and death . Recently, Kalambokis et al. showed that rifaximin, a non-absorbable antibiotic, improves haemodynamic abnormalities and renal function in patients with advanced cirrhosis . These results are in line with other randomized control trial that found a lower incidence of HRS in patients with advanced cirrhosis and ascites treated with norfloxacin .
Endothelin 1 (ET-1) is the most potent mediator of stellate cell contraction and in the liver, ET-1 receptors predominate in hepatic stellate cells, which have an important role in the regulation of intrahepatic portal hypertension in cirrhosis . Both endotoxin itself and cytokines released in response to BT are potent stimuli for the production of ET-1, which may act in combination with cyclooxygenase products to increase portal venous resistance during endotoxaemia [66, 67]. During BT, endotoxin (LPS) promotes an imbalance in the expression of vasoconstrictors (endothelin-1) in relation to the expression of vasodilator substances (NO, carbon monoxide), thereby increasing hepatic vascular tone . Therefore, both infections or BT may cause an acute increase in portal pressure in patients with cirrhosis, triggering a variceal bleeding . Recently, our group have demonstrated that bactDNA translocation is associated with a more profound abnormality of the intrahepatic circulation, in the sense of worse hepatic endothelial dysfunction, as suggested by a greater post-prandial increase in hepatic venous pressure gradient in bactDNA-positive patients with cirrhosis and ascites .
Other studies have suggested that BT may be implicated in hepatic encephalopathy (HE) pathogenesis. It is a well-known fact that bacterial infections constitute a trigger for HE in patients with cirrhosis. Previous studies on the pathogenesis of HE have focused on the deleterious role of toxins on the brain. However, in the last decade, several studies suggest how inflammation and infection exert synergistic effects with toxins such as ammonia in the pathogenesis of HE [70, 71]. Based on the aforementioned, BT-associated inflammatory response may have a role in the pathogenic mechanisms involved in HE. SIBO is associated with BT in patients with cirrhosis, estimated by the presence of bactDNA, and with a higher prevalence of HE in one recent study . Moreover, a relationship between minimal hepatic encephalopathy (MHE) and SIBO had been also demonstrated in patients with cirrhosis . Rifaximin, a non-absorbable antibiotic, improves MHE in some clinical studies [74, 75]; suggesting that BT, including translocation of bacterial products, plays an important role in the pathogenesis of HE.
Bacterial translocation, as a trigger for an abnormal inflammatory state and haemodynamic derangement, has prognostic implications in liver cirrhosis. In a multicentre study that included 156 patients with cirrhosis and non-neutrocytic ascites, we observed that patients with positive bactDNA in ascitic fluid and plasma had a worse prognosis than patients with negative bactDNA . In that study, acute-on-chronic liver failure was the main cause of mortality in bactDNA-positive patients; suggesting that BT may be associated with haemodynamic and clinical deterioration of the overall clinical situation of patients with cirrhosis. Bruns et al. also found that bactDNA detection in patients with severe liver dysfunction (MELD score >15) is associated with an increased mortality . These results are in line with other studies that showed how intestinal decontamination improves survival in patients with cirrhosis [64, 77], liver function parameters  and other portal hypertension-related complications like thrombocytopaenia .
Advanced cirrhosis behaves like a chronic inflammatory disease in which the immune system is exposed to a continuous activation by bacteria and bacterial products of intestinal origin, the so-called BT. The immunological abnormalities associated with cirrhosis and BT perpetuate the translocation of bacteria and their products while exerting an effect on other systems, such as the circulatory system. Therefore, BT is implicated in some of the complications of cirrhosis, worsening hyperdynamic circulation, renal function, intrahepatic endothelial dysfunction and HE (Fig. 2). It is noteworthy that much of the pro-inflammatory state of cirrhosis is owing to the release of bacterial products into the systemic circulation. Finally, note that the presence of bacterial DNA fragments is a direct and sensitive marker of BT and that seems to play a direct role in the inflammatory response associated with BT, as well as being a predictor of mortality in patients with cirrhosis and ascites. Finally, we consider that more studies are needed to increase understanding of the pathogenesis of BT to open new therapeutic avenues in cirrhosis and portal hypertension.