Secondary Logo

Journal Logo

Basic Science Aspects


Deitch, Edwin A*; Forsythe, Raquel*; Anjaria, Dev*; Livingston, David H; Lu, Qi*; Xu, Da-Zhong*; Redl, Heinz

Author Information
doi: 10.1097/01.shk.0000133592.55400.83
  • Free



The potential causative role of gut injury/inflammation in the pathogenesis of hemorrhagic shock-, trauma- or sepsis-induced acute lung injury and systemic inflammation has undergone several modifications over the last two decades and remains a topic of controversy (1, 2). Most recently, the concept of bacterial translocation as the mechanism by which gut injury promotes acute lung injury, multiple organ dysfunction (MODS) and/or an exaggerated systemic inflammatory state has been supplanted by the notion that gut-induced distant organ injury is secondary to factors produced and/or liberated by the ischemic gut that reach the systemic circulation via the intestinal lymphatics rather than the portal venous system (3–5). This work has resulted in the generation of the gut-lymph hypothesis of MODS (6). However, studies supporting the gut-lymph hypothesis of MODS have been limited to work carried out in small animal rodent models of trauma-hemorrhagic shock (T/HS) and thermal injury. Because the physiologic responses of even clinically relevant rodents shock and trauma models may not fully replicate the physiologic responses observed in injured humans (7), a critical next step in testing this hypothesis would be studies in man. However, the performance of these studies in injured patients are difficult to fully justify because they would require invasive procedures to be performed in critically injured patients who may not be hemodynamically stable. Thus, as a necessary prelude to the performance of clinical studies, we tested the hypothesis that T/HS-induced acute lung injury would be blunted by the drainage of thoracic duct lymph in nonhuman primates and that thoracic duct lymph from primates subjected to T/HS, but not trauma sham-shock (T/SS), would injure endothelial cells, increase endothelial cell monolayer permeability, and suppress bone marrow growth. Thoracic duct drainage of lymph was chosen over selective drainage of intestinal lymph because the drainage of thoracic duct lymph is clinically feasible. The results of this primate study validate and extend the major findings observed in previously published rodent models of T/HS and support the gut lymph hypothesis of MODS.



Fifteen adult male baboons (Papio ursinus) weighing between 23 and 33 kg were studied. The animals had been caged at the nonhuman primate unit of the Biocon Research LTD Laboratory in Pretoria, South Africa, for at least 3 months. The clinical condition of the baboons had been checked before their entry into quarantine, and at the beginning of the study, all of the animals were free of disease.

The experimental protocol was reviewed and approved by the Institutional Animal Care and Use Committee of Biocon Research Laboratories, a research facility in South Africa where the studies were performed. The animals were treated in accordance with National Institutes of Health guidelines.

Experimental design and T/HS model

The goals of this study were three-fold. First, we sought to determine whether diversion of thoracic duct lymph prevented or reduced T/HS-induced lung injury. Second, we sought to investigate the composition and biologic activity of thoracic duct lymph from baboons subjected to T/HS or T/SS. Last, we sought to compare the biologic activity of plasma from baboons subjected to T/SS, T/HS, and T/HS with thoracic duct drainage. To accomplish these goals, the following three groups of baboons were studied: T/SS plus lymph diversion (LD), T/HS, and T/HS-LD. The trauma component of this model consisted of a neck dissection with resection of the proximal clavicle plus a laparotomy and handling of the intestines. The decision to use a nonlethal trauma insult of modest severity in combination with HS was based on the notion that we wished to model the clinical situation of a severe, but survivable, T/HS insult.

All of the animals were anesthetized and placed on a ventilator as previously described (8). Throughout the experiment, the baboons were spontaneously ventilated with continuous positive airway pressure of 2 mmHg to keep the alveoli open. Continuous end-tidal CO2 measurements were used to monitor ventilation, and the inspired oxygen content (FiO2) was 0.25 ± 0.02. Subsequent anesthesia was maintained with intravenous pentobarbital (2–5 mg/kg/h). After adequate anesthesia was obtained, the following catheters were placed: a Swan-Ganz catheter for measuring hemodynamic variables, an arterial line for monitoring blood pressure and for obtaining blood samples, a femoral vein catheter for blood withdrawal, and a urinary catheter for measuring urine production.

After instrumentation placement was completed, the thoracic duct was exposed through a left neck dissection. Because of the significant musculature of the baboon, exposure of the thoracic duct required division of the neck musculature as well as a segmental resection of the proximal portion of the clavicle. At this point, the baboons were randomized into one of the three experimental groups. The baboons randomized to the T/SS or the T/HS-LD groups had their thoracic duct ligated at the point where it empties in the subclavian vein, after which it was catheterized just proximal to the ligature. The neck incision was then closed and the thoracic duct catheter was exteriorized through the wound. A laparotomy was then performed, after which the animals were bled down to a mean arterial pressure (MAP) of 40 mmHg in two stages. Over the first 30 min, the MAP was reduced to 60 mmHg. The MAP was then reduced from 60 to 40 mmHg over the next 30 min. Fifteen minutes after the induction of shock, the laparotomy was closed. The MAP was kept at 40 mmHg until the base excess reached −5 or the total shock period reached 3 h. Base excess was used as the criteria for ending the shock period and initiating volume resuscitation because this is a physiologically relevant marker of the magnitude of the shock state.

To better mimic the clinical situation, volume resuscitation was also carried out in stages. During the 1st hour of resuscitation, the MAP was brought to 60 mmHg by the infusion of Ringer’s solution. Over the next hour, the MAP was brought to 100 mmHg by the infusion of 25% of the shed blood plus Ringer’s solution, and during the 3rd hour of resuscitation, an additional 25% of the shed blood plus Ringer’s solution was administered to raise the MAP to 120 mmHg. Two hours later, the baboons were sacrificed. The T/SS group was instrumented but had no blood withdrawn or infused. Temperature was maintained between 36.8°C and 37.5°C in all of the groups.

Blood samples and lymph samples

Systemic blood samples were collected at the following time points: 30 min before laparotomy and hemorrhage, at the beginning of hemorrhage (time 0), at 1, 2, and, when applicable, 3 h posthemorrhage, as well as hourly for 5 h postresuscitation. Portal blood samples were obtained at 5 h postresuscitation. Serum and plasma samples were prepared and aliquots were frozen at −80°C for further analysis. Lymph was collected in heparinized tubes on ice before T/HS and hourly thereafter. The lymph samples were then centrifuged and the cell-free humoral fractions were aliquoted and frozen at −80°C for further study. Blood gas analysis and hematology measurements were performed on site.

Analytic methods

Plasma and lymph TNF, IL-6, soluble TNF receptor (sTNFR), and elastase levels were measured as previously described using commercial kits (9). Routine clinical chemistry parameters, as well as lactate, total protein, and albumin levels, were performed using commercially available kits.

Lung injury

Lung injury was assessed by measuring the wet-to-dry ratio of a segment of the left lower lobe of the lungs of each baboon. In this assay, the weight of a segment of lung was measured before and 48 h after it was placed in an oven set at 95°C. An increased wet-to-dry ratio reflects lung edema. An additional segment from a defined area of the right lower lobe of the lung of each animal was processed for histology. The degree of intra-alveolar edema was assessed in a blinded fashion, using a 4-point scale, with 0 being no edema, 1 being mild edema, 2 being moderate edema, and 3 being severe edema.

Human umbilical vein endothelial cell viability assays

Human umbilical vein endothelial cells (HUVEC) were seeded at 4 × 104 cells per well in 24-well plates and were grown to confluence over 24 h. Medium only or medium containing 20% T/HS lymph, T/SS lymph, or postshock portal vein plasma was added to each well. After an 18-h incubation period, cell viability was determined using the mitochondrial tetrazolium (MTT) test, and the supernatants were assayed for lactic dehydrogenase (LDH) as a marker of cell injury as previously described (10). The MTT test recognizes viable cells by detecting a color change in cells with active mitochondrial activity, and MTT activity is measured spectrophotometrically using a standard kit.

HUVEC permeability assay

HUVECS were seeded at 20,000 cells per insert (0.33 cm2) on type I rat-tailed collagen-coated membranes (pore size, 3 μm) contained on the apical chamber of a two-chambered Transwell system (Costar, Cambridge, MA). As previously described (3, 10), the HUVEC monolayers had become confluent by 72 h after seeding. The HUVEC monolayers then were exposed to medium only, 20% T/HS lymph, T/SS lymph, or postshock portal vein plasma for 1 h. At this point, a 0.05% solution of a 40-kD Dextran rhodamine permeability probe (Molecular Probes, Eugene, OR) was added to the apical chamber of the Transwell system. After an additional 1-h incubation period, medium in the basal chamber was removed and the amount of Dextran rhodamine present was determined spectrophotometrically at 570 nm. Permeability of the HUVEC monolayers was expressed as clearance (C; nanoliters per centimeter squared per hour), according to the equation: C = Fa−b/[tracer]a × S, where Fa−b is the flux of the dye probe from the apical to the basolateral compartment (light units per hour), [tracer]a is the concentration of Dextran rhodamine in the apical chamber at the beginning of the incubation period, and S is the surface area of the monolayer (centimeters squared).

Hematopoietic assays

Bone marrow was obtained from the posterior iliac crest of healthy human volunteers and hematopoietic progenitor cultures were established as previously described (12). Briefly, low-density bone marrow mononuclear cells were separated by density centrifugation and were plated in duplicate at a density of 105 cells in Iscove’s medium containing 30% fetal calf serum, 2% bovine serum albumin, 1% methycellulose, 2 × 10−4 M 2-mercaptoethanol, and 2% glutamine. Based on our previous studies (11, 12), 1% or 5% thoracic duct lymph or 5% plasma from the various baboon groups was added to the bone marrow cultures. Parallel cultures containing medium without baboon lymph or plasma served as controls.

Granulocyte-monocyte colony forming unit (CFU-GM) cultures were supplemented with 3 U/mL rhGM-CSF and erythroid burst-forming unit (BFU-E) cultures were supplemented with 2 U/mL rh-erythropoietin and 6 U/mL rhIL-3.

Cultures were incubated at 37°C in 5% CO2. CFU-GM colonies with greater than 20 cells were enumerated at Day 10 and BFU-E colonies were counted at Day 15. All plates were counted by an observer who was blinded to the groups.


Aliquots (100 μL) of the lymph and plasma samples were cultured for translocating bacteria on blood agar plates to determine whether the samples were sterile.


Comparisons were made with a one-way analysis of variance, with the Tukey-Kramer multiple comparisons test when multiple groups or time points were tested. When two groups were compared, the unpaired t test was used. Analysis of intra-alveolar edema on lung histology was performed using the nonparametric ranking test. All data are expressed as the mean ± SD. Statistical significance was considered to be reached when P ≤0.05.


As illustrated in Figure 1, the extent of the HS insult as well as the hemodynamic response were similar between the T/HS and the T/HS-LD groups. This included MAP, heart rate, cardiac index, systemic vascular resistance, pulmonary vascular resistance, and arterial oxygen levels (Fig. 1). Just before resuscitation, the serum lactate had peaked at a value of almost 8 mmol/L, whereas the maximal base deficit was approximately 6 mEq/L (Fig. 2). These values returned to normal over the period of resuscitation and were similar between the two T/HS groups. Maximal blood withdrawal in the T/HS-LD group (34.6 ± 1.4 mL/kg) and the T/HS group (35.2 ± 4.2 mL/kg) were similar as was the volume of shed blood reinfused during the period of volume resuscitation (16.4 ± 0.7 mL/kg vs. 16.8 ± 0.3 mL/kg). The T/HS and T/HS-LD groups received equal baseline volumes of crystalloid resuscitation (15 mL/kg) during the resuscitation period. Additional crystalloid resuscitation given during the 5-h postshock period was slightly, but not significantly, greater for the T/HS versus the T/HS-LD group (166 vs. 130 mL). As expected, the hematocrit of the two T/HS groups was lower than the T/SS-LD group (Fig. 2), but there were no significant differences in the platelet or white blood cell counts between any of the groups (data not shown).

Fig. 1.
Fig. 1.:
The mean arterial pressure (MAP), heart rate, cardiac index, systemic vascular (SVR) and pulmonary vascular resistance (PVR) responded similarly to the T/HS insult as did arterial oxygen levels in the lymph duct-ligated and not ligated groups. Data are expressed as mean ± SD with five baboons per group.
Fig. 2.
Fig. 2.:
The plasma lactate and base excess response to T/HS was similar between the T/HS and T/HS-LD groups. As compared with the T/SS-LD group, the hematocrits of the T/HS groups were decreased (*P < 0.05 at R1–R5 time periods). Data are expressed as mean ± SD with five baboons per group.

To assess the inflammatory response, plasma elastase, serum IL-6, TNF, and the sTNFR (R55) levels were measured. sTNF-R55 and IL-6 began to increase during the shock period, whereas elastase became elevated only after the end of the shock period (Fig. 3). The increase in sTNF-R55 was observed only in the two T/HS groups and slowly returned to baseline values during the 5-h postshock resuscitation period. The IL-6 levels peaked during the first 2 h after the end of the shock period and had begun to decrease by the third postshock hour. In contrast to the other inflammatory markers, IL-6 was modestly increased in the T/SS-LD group, with this increase beginning during the shock period and persisting throughout the experimental period. Plasma elastase levels began to rise in both of the T/HS groups during the early resuscitation period and remained elevated throughout (Fig. 3). Plasma elastase levels did not increase in the T/SS-LD group. In contrast to IL-6, TNF levels were not increased in any of the groups (data not shown). Thus, the systemic inflammatory response increased to a similar extent in the T/HS-LD and the T/HS groups, and both groups manifested a greater systemic response than the T/SS-LD group.

Fig. 3.
Fig. 3.:
The temporal pattern of change in plasma elastase, serum IL-6, and serum s-TNF-R55 levels were similar between the T/HS and T/HS-LD groups and was increased as compared with the T/SS-LD group. Data are expressed as mean ± SD with five baboons per group (*P < 0.05 T/SS-LD vs. T/HS and T/HS-LD groups).

Thoracic lymph flow was similar between the T/HS-LD and the T/SS-LD groups before the induction of HS, although due to a technical problem, lymph flow could not be reliably determined in one of the T/SS-LD baboons (Fig. 4). In the T/HS-LD group, lymph flow significantly increased with resuscitation, whereas lymph flow remained relatively constant in the T/SS-LD group. The thoracic lymph total white blood cell count increased in both groups, going from 3.7 ± 0.9 to 7.1 ± 2.7 × 103/μL in the T/HS-LD group and 3.2 ± 1.6 to 6.3 ± 1.7 × 103/μL in the T/SS-LD group. When the thoracic lymph content of IL-6 and sTNF-R55 was determined by multiplying the concentration of each factor by the lymph volume, it was found that the sTNF-R55, but not the IL-6, content of the lymph was increased to a greater extent in the T/HS than the T/SS-LD group during the postshock resuscitative period (Fig. 4). At the same time that lymph cytokine levels were rising, lymph total protein and albumin levels decreased over time (Fig. 5), with the T/HS-LD group showing a greater decrease.

Fig. 4.
Fig. 4.:
Lymph flow was increased in the T/HS-LD group after the end of the shock period, however, in spite of this increase in lymph flow, the total amount of IL-6 exiting the gut each hour was similar between the groups. However, the amount of sTNF-R55 exiting the lymph was increased in the T/HS-LD animals. Data are expressed as mean ± SD with four baboons in the T/SS-LD and five baboons in the T/HS-LD groups. *P < 0.05 vs. T/SS-LD.
Fig. 5.
Fig. 5.:
Lymph protein and albumin levels decreased over the experimental period with greater decreases being observed in the T/HS-LD groups beginning at the R1 time point. Data are expressed as mean ± SD with four baboons in the T/SS-LD and five baboons in the T/HS-LD group. *P < 0.05 vs. T/HS-LD group.

To determine whether lymph diversion protected against T/HS-induced lung injury, lung wet-to-dry weights were measured and lung samples were collected for histologic evaluation of alveolar edema. Lymph diversion prevented the T/HS-induced increase in lung edema, as reflected in lung wet-to-dry ratios (Fig. 6). Additionally, the histologic degree of intra-alveolar edema was significantly greater in the T/HS than the T/SS-LD or the T/HS-LD groups (Fig. 6).

Fig. 6.
Fig. 6.:
Lung injury as reflected in the lung wet-to-dry ratio was increased in the T/HS but not the T/HS-LD group. Data are expressed as mean ± SD with five baboons per group. *P < 0.05 vs. T/SS-LD. Lung injury as reflected in the histologic grading of intra-alveolar edema (grade 0, no evidence of edema to grade 3, evidence of extensive edema) also showed that the magnitude of alveolar edema was greatest in the T/HS group. *P < 0.05 vs. T/SS-LD and T/HS-LD groups.

To determine whether T/HS lymph injured endothelial cells and/or increased endothelial cell monolayer permeability, the cytotoxic and permeability-inducing properties of the lymph samples were tested. There was a peak of cytotoxic activity in the thoracic duct lymph collected from the T/HS baboons during the 1st h post-T/HS resuscitation that was not observed in the T/SS animals (Fig. 7). This cytotoxic activity was rapidly lost as the resuscitation period continued. The T/HS lymph collected during the 1st h of resuscitation also increased the permeability of HUVEC monolayers to a 40-kD permeability probe (Fig. 8). To begin to assess whether the blood samples also contained endothelial cell toxic factors, portal vein plasma (PPM) collected at the end of the resuscitation period (R5) from all three groups was tested. These blood samples did not kill endothelial cells or increase endothelial cell monolayer permeability (Figs. 7 and 8). Because the T/HS R0 but not R5 lymph samples had biologic activity, it is possible that R5 plasma samples might have missed the peak of biologic activity. Consequently, systemic blood samples collected during the R0 time period were tested for endothelial cell cytotoxicity. Neither the T/HS nor the T/HS-LD or the T/SS-LD R0 plasma samples killed HUVECs or increased endothelial cell monolayer permeability (data not shown). These effects of the lymph samples do not appear to be due to translocating bacteria because all of the lymph samples were sterile, as were all the plasma samples (data not shown).

Fig. 7.
Fig. 7.:
Thoracic duct lymph produced during the 1st hour of resuscitation was cytotoxic for HUVECs whether measured as viability by the MTT test or LDH release. MTT viability data expressed are as mean ± SD percentage viability of HUVECs incubated in medium, and LDH release into the supernatant is expressed as mean ±SD units per liter with four baboons in the T/SS-LD and five baboons in the T/HS-LD groups. PPM indicates R5 portal vein plasma. *P < 0.01 vs. T/SS-LD RO and all other T/SS-LD and T/HS-LD time points.
Fig. 8.
Fig. 8.:
Thoracic duct lymph produced during the 1st hour of resuscitation increased the permeability of the HUVEC monolayers. Data are expressed as mean ± SD with four baboons in the T/SS-LD and five baboons in the T/HS-LD groups. PPM indicates R5 portal vein plasma. *P < 0.01 vs. T/SS-LD RO and all other T/SS-LD and T/HS-LD time points.

T/HS lymph, but not T/SS lymph, collected during the R5 time period suppressed the hematopoietic response of normal human bone marrow as reflected in decreased CFU-GM and BFU-E growth (Fig. 9) Additionally, thoracic duct lymphatic drainage blunted the ability of R5 plasma from baboons subjected to T/HS to inhibit bone marrow CFU-GM colony growth but not BFU-E colony formation (Fig. 10).

Fig. 9.
Fig. 9.:
Thoracic duct T/HS, but not T/SS, lymph at 1% and 5% concentrations inhibits CFU-GM and BFU-E progenitor cell growth. Data are expressed as mean ± SD with four baboons in the T/SS-LD and five baboons in the T/HS-LD groups. *P < 0.05 vs. medium and T/SS lymph groups; #P < 0.05 vs. 1% T/HS lymph as well as medium and T/SS lymph groups.
Fig. 10.
Fig. 10.:
(A) Plasma collected at the R5 time point from both T/HS groups suppressed BFU-E growth compared with medium, preshock plasma, or R5 T/SS-LD plasma. *P < 0.05 vs. all other groups. (B) R5 T/HS plasma suppressed GFU-GM growth (*P < 0.05 vs. all other groups) as did R5 T/HS-LD plasma (#P < 0.05 vs. all other groups), however, the magnitude of CFU-GM by T/HS-LD plasma was less than T/HS plasma. Data are expressed as mean ± SD with five baboons in each group.


The central concept being tested in this primate model of T/HS is that gut-derived factors contained primarily in the mesenteric lymph, rather than in the portal blood, contribute to distant organ injury. This hypothesis is based on recent studies in rodents indicating that division of the mesenteric lymphatic ducts prevents lung injury after HS (3), significantly ameliorates lung injury after thermal injury (4), limits HS-induced red blood cell dysfunction (13), prevents shock-induced neutrophil activation in vivo (14), and abrogates bone marrow failure (12). In these rodent studies, the mechanism of HS-induced lung injury appears to be through mesenteric lymph-induced activation of neutrophils and activation/injury of endothelial cells. This notion is based on in vitro studies indicating that mesenteric lymph, but not portal vein plasma, collected after a nonlethal episode of HS activates neutrophils (5, 15), increases endothelial cell monolayer permeability (3, 10), upregulates adhesion molecule expression (16), and can even cause endothelial cell death (3, 10, 17). This concept that gut-derived factors contained primarily in the mesenteric lymph rather than the portal system potentiates the development of distant organ and cellular (lung, red blood cell, neutrophil, endothelial cell, and bone marrow) injury and dysfunction, if correct, would help clarify several important issues. For example, because the lung is the first organ exposed to mesenteric lymph (i.e., mesenteric lymph enters the subclavian vein via the thoracic duct), it would help explain the clinical observation of why the lung is generally the first organ to fail in severely injured patients. Second, this gut lymphatic hypothesis would provide new information on the pathophysiology of gut-induced lung and other organ dysfunction. Last, it would help explain the discordant results between experimental and some clinical studies on the role of gut injury and loss of gut barrier function in the development of a systemic inflammatory state and distant organ injury. Specifically, this concept that the mesenteric lymphatics rather than the portal blood is the major route by which tissue injurious factors exit the gut after gut injury helps explain how the gut can contribute to distant organ injury and yet the portal vein of trauma patients not contain bacteria or endotoxin (18).

Because the results of studies performed on lower species such as rodents may not accurately or fully reflect the physiologic responses observed in human patients, we used a baboon primate model of T/HS to extend our rodent work and to test the hypothesis that gut-derived factors contained in lymph contribute to distant organ injury. The observations that thoracic duct ligation attenuated T/HS-induced lung injury and that thoracic duct lymph from baboons subjected to T/HS, but not T/SS, suppressed human bone marrow CFU-GM and BFU-E growth in addition to being cytotoxic for human endothelial cells and increasing HUVEC permeability provide the first primate evidence supporting the gut-lymph hypothesis of MODS. The protective effect of thoracic duct ligation on T/HS-induced lung injury was not related to hemodynamic differences between the T/HS thoracic duct ligated and nonligated baboons because the physiologic responses of the two groups were similar, as was the maximal volume of blood withdrawn and reinfused during the HS and resuscitation phases of the model. Likewise, the thoracic duct-ligated and nonligated baboons subjected to T/HS had a comparable cytokine response, as reflected in plasma IL-6 and sTNF-R55 levels, which were increased to a greater extent after T/HS than after T/SS. Thus, although ligation of the thoracic duct and prevention of lymph from entering the systemic circulation did not blunt the systemic cytokine response or modify the hemodynamic response to T/HS, it attenuated acute lung injury. This comparable cytokine response between the T/HS and the T/HS-LD groups is similar to the results we observed in our rodent studies, where we found that the tissue injurious and activating factors in lymph are not cytokines (2, 19, 20). Based on the fact that lymph from the T/HS, but not the T/SS, baboons injured endothelial cells and activated neutrophils (unpublished results, E.A. Deitch), we believe that the mechanism by which thoracic duct ligation prevented lung injury was by limiting T/HS-induced neutrophil-endothelial cell-mediated tissue injury.

The exact gut-derived factors in postshock lymph responsible for bone marrow progenitor growth suppression, endothelial cell injury, and increased endothelial cell monolayer permeability remain to be determined. However, because these lymph samples were sterile, it appears that the proinflammatory and tissue injurious effects of postshock lymph are not mediated directly by translocating bacteria. This experimental observation that the thoracic duct lymph from the shocked animals was sterile is consistent with a study examining thoracic duct lymph in intensive care unit patients (21). In this clinical study of intensive care unit patients, the thoracic duct lymph was sterile and contained only low levels of endotoxin, and there was no difference in thoracic duct endotoxin levels between the patients who did and those who did not develop MODS. Although the thoracic duct lymph from the baboons subjected to T/HS had increased levels of IL-6 and sTNF-R55, these levels were not different from that observed in the baboons subjected to T/SS and thus do not appear to be responsible for the biologic activity of the T/HS thoracic duct lymph. This observation is consistent with our previous rodent studies (19, 20) as well as small animal studies from Dr. Moore’s laboratory (22) indicating that the biologic effects of intestinal lymph are not due to cytokines but are caused by lipid and large-molecular-weight proteins. On the other hand, the fact that a laparotomy plus manipulation of the intestines (T/SS) was sufficient to induce an intestinal cytokine response, as reflected in elevated lymph but not plasma cytokine levels, is consistent with studies indicating that surgical manipulation of the intestine can lead to a transient episode of oxidative stress and increased intestinal permeability (23, 24).

Although translocating bacteria were not found in the plasma or lymph samples in the present study and are rarely recovered from the systemic or portal circulation in clinical or experimental studies, loss of gut barrier function and intestinal bacteria do appear to play a role in augmenting the intestinal inflammatory response because alterations in the gut bacterial flora, leading to bacterial overgrowth, are associated with an exaggerated systemic cytokine response after an episode of shock or gut injury (25). Consequently, we believe that bacteria and/or endotoxin crossing the mucosal barrier potentiate distant organ failure by activating the gut inflammatory response even when these translocating bacteria are trapped within the gut wall or intestinal lymph nodes and do not reach the systemic circulation. This concept of locally translocated bacteria/endotoxin contributing to the intestinal inflammatory response is dependent upon failure of gut barrier function because if gut barrier function remains intact, luminal bacteria and their products remain within the gut. This notion is of potential clinical relevance because experimental conditions associated with splanchnic hypoperfusion, such as HS, burns, or sepsis, have been shown to cause intestinal mucosal injury and loss of gut barrier function (26). Furthermore, prevention of gut injury and preservation of gut barrier function in these conditions prevents bacterial translocation, thereby highlighting the role of the gut barrier in preventing or limiting the escape of intestinal bacteria or endotoxin (26).

Thus, at the current time, our view of the pathogenesis of gut inflammation leading to distant organ dysfunction is based on a three-hit model. In this three-hit model, the first insult is splanchnic hypoperfusion or ischemia, which occurs in response to systemic insults such as burns, trauma, sepsis, or acute blood loss. Restoration of intestinal blood flow (second hit) contributes to gut injury, inflammation, and loss of gut barrier function through an ischemia-reperfusion-mediated mechanism. Loss of gut barrier function (third hit) allows intestinal bacteria and endotoxin to cross the mucosal barrier where they come into contact with enterocytes and intestinal immune cells. The majority of these translocating bacteria are phagocytosed by immune and nonimmune intestinal cells and thereby contribute to an intestinal inflammatory response. A relatively small number of these translocating bacteria or their products escape from the intestine and subsequently are trapped in the intestinal lymph nodes, where they also induce an inflammatory reaction. The net result of this intestinal ischemia-reperfusion injury, which is exacerbated by locally translocating bacteria, is the production of gut-derived toxic and inflammatory products, which reach the systemic circulation via the intestinal lymphatics.


1. Deitch EA: The role of intestinal barrier failure and bacterial translocation in the development of systemic infection and multiple organ failure. Arch Surg 125:403–404, 1990.
2. Deitch EA, Sambol JT: The gut-origin hypothesis of MODS. In Deitch EA, Vincent JL, Windsor A (eds.): Sepsis and Multiple Organ Dysfunction: A Multidisciplinary Approach. Philadelphia, PA: W.B. Saunders, 2002, pp 105–116.
3. Magnotti LJ, Upperman JS, Xu DZ, Lu Q, Deitch EA: Gut-derived mesenteric lymph but not portal blood increases endothelial cell permeability and potentiates lung injury following hemorrhagic shock. Ann Surg 228:518–527, 1998.
4. Magnotti LJ, Xu DZ, Lu Q, Deitch EA: Gut-derived mesenteric lymph: a link between burn and lung injury. Arch Surg 143:1333–1341, 1999.
5. Zallen G, Moore EE, Johnson JL, Tamura DY, Ciesla DJ, Silliman CC: Posthemorrhagic shock mesenteric lymph primes circulating neutrophils and provokes gut injury. J Surg Res 83:83–88, 1999.
6. Deitch EA: Bacterial translocation or lymphatic drainage of toxic products from the gut: what is important in human beings? Surgery 131:241–244, 2002.
7. Deitch EA: Animal models of sepsis and shock: a review and lessons learned. Shock 9:1–11, 1998.
8. Schlag G, Redl H, Khakpour Z, Davies J, Pretorius J: Hypovolemic-traumatic shock models in baboons. In Schlag G, Redhl H (eds.): Physiology of Shock, Sepsis and Organ Failure. Berlin: Springer-Verlag, 1993, pp 384–402.
9. Redl H, Baharami S, Furst W, Schiesser A, Schlag G, Mollnes TE: Evidence for early bacteremia/endotoxemia and inflammatory responses after trauma-hemorrhage and complement activation in a nonhuman primate model. Eur J Trauma 26:300–307, 2000.
10. Deitch EA, Adams CA, Lu Q, Xu DZ: Mesenteric lymph from rats subjected to trauma-hemorrhagic shock are injurious to rat pulmonary microvascular endothelial cells as well as human umbilical vein endothelial cells. Shock 16:290–293, 2001.
11. Wu JC, Livingston DH, Hauser CJ, Deitch EA, Rameshwar P: Trauma inhibits erythroid burst-forming unit and granulocyte-monocyte colony forming unit growth through the production of TGF-β1 by bone marrow stroma. Ann Surg 234:224–232, 2001.
12. Anjaria DJ, Rameshwar P, Deitch EA, Xu DZ, Adams CA, Forsythe RM, Sambol JT, Hauser CJ, Livingston DH: Hematopoietic failure after hemorrhagic shock is mediated partially through mesenteric lymph. Crit Care Med 29:1780–1785, 2001.
13. Zaets SB, Berezina TL, Caruso J: Xu da Z, Deitch EA, Machiedo GW: Mesenteric lymph duct ligation prevents shock-induced RBC deformability and shape changes. J Surg Res 109:51–56, 2003.
14. Adams CA, Hauser CJ, Adams JM, Fekete Z, Xu DZ, Sambol JT, Deitch EA: Trauma-hemorrhage-induced neutrophil priming is prevented by mesenteric lymph duct ligation. Shock 18:513–517, 2002.
15. Adams JM, Hauser CJ, Adams CA Jr, Xu DZ, Livingston DH, Deitch EA: Entry of gut lymph into the circulation primes rat neutrophil respiratory burst in hemorrhagic shock. Crit Care Med 29:2194–2198, 2001.
16. Adams CA, Sambol JT, Xu DZ, Lu Q, Granger DN, Deitch EA: Hemorrhagic shock induced up-regulation of P-selectin expression is mediated by factors in mesenteric lymph and blunted by mesenteric lymph duct interruption. J Trauma 51:625–632, 2001.
17. Osband AJ, Deitch EA, Lu Q, Zaets S, Dayal S, Lukose B, Xu DZ: The role of oxidant-mediated pathways in the cytotoxicity of endothelial cells exposed to mesenteric lymph from rats subjected to trauma-hemorrhagic shock. Shock 20:269–273, 2003.
18. Moore FA, Moore EE, Poggetti R, McAnena OJ, Peterson VM, Abernathy CM, Parsons PE: Gut bacterial translocation via the portal vein: a clinical perspective with major torso trauma. J Trauma 31:629–638, 1991.
19. Adams CA, Xu DZ, Lu Q, Deitch EA: Factors larger than 100 kD in post-hemorrhagic shock mesenteric lymph are toxic for endothelial cells. Surgery 129:351–362, 2001.
20. Dayal SD, Hauser CJ, Feketeova E, Fekete Z, Adams JM, Lu Q, Xu DZ, Zaets S, Deitch EA: Shock mesenteric lymph-induced rat polymorphonuclear neutrophil activation and endothelial cell injury is mediated by aqueous factors. J Trauma 52:1048–1055, 2002.
21. Lemaire JC, van Lanschot JB, Stoutenbeck CP, van Deventer SJ, Dankert J, Oosting H, Gouma DJ: Thoracic duct in patients with multiple organ failure: no major route of bacterial translocation. Ann Surg 229:128–136, 1999.
22. Gonzalez RJ, Moore EE, Ciesla DJ, Meng X, Biffl WL, Silliman CC: Post-hemorrhagic shock mesenteric lymph prime neutrophils for enhanced cytotoxicity via phospholipase A2. Shock 16:218–222, 2001.
23. Anup R, Aparna V, Pulimmod A, Balasubramanian KA: Surgical stress and the small intestine: role of oxygen free radicals. Surgery 125:560–569, 1999.
24. Prabhu R, Anup R, Balasubramanian KA: Surgical stress induces phospholipid degradation in the intestinal brush border membrane. J Surg Res 94:178–184, 2000.
25. Guo W, Ding J, Deitch EA: Alterations in the bacterial flora modulate the systemic cytokine response to hemorrhagic shock. Am J Physiol 269:G827–G832, 1995.
26. Deitch EA: Multiple organ failure: pathophysiology and potential future therapy. Ann Surg 216:117–134, 1992.

MODS; sepsis; intestine; respiratory failure

©2004The Shock Society