Share this article on:

Epidural Analgesia Prevents Endotoxin-Induced Gut Mucosal Injury in Rabbits

Kosugi, Shizuko MD*; Morisaki, Hiroshi MD*; Satoh, Tomoyuki MD*; Ai, Kimiaki MD*; Yamamoto, Michiko BA*; Soejima, Junko MD; Serita, Ryohei MD*; Kotake, Yoshifumi MD*; Ishizaka, Akitoshi MD; Takeda, Junzo MD*

doi: 10.1213/01.ANE.0000153863.95598.08
Regional Anesthesia: Research Report

In the present study, we evaluated the effect of epidural analgesia on the alterations of gut barrier function elicited by endotoxin in rabbits. After the placement of an epidural catheter, 28 male rabbits were randomized into either 0.5% lidocaine (group E) or saline (group C) group. The solutions (0.4 mL/kg) were epidurally injected, followed by continuous infusion (0.1 mL · kg−1 · h−1) throughout the study period. Under a continuous infusion of lipopolysaccharide (15 μg · kg−1 · h−1), mean arterial blood pressure, intramucosal pH, and plasma thrombomodulin concentrations were measured. At 4 h, mean arterial blood pressure was lower (P < 0.05), intramucosal pH was higher (P < 0.01), and the progression of hemodilution more profound (P < 0.05) in group E versus group C, whereas plasma thrombomodulin levels were increased to a similar extent between the groups. With less wet-to-dry weight ratio of ileum, histopathological injury scores of gut mucosa were significantly less in group E versus group C (P < 0.01). In a separate series of experiments (n = 10 each group), mucosal permeability in group E was significantly less compared with group C (P < 0.05). Collectively, these studies showed that despite a significant decrease of perfusion pressure and arterial oxygen content, epidural analgesia minimized endotoxin-induced functional and structural injury of gut mucosa possibly through endothelium-independent mechanisms.

IMPLICATIONS: Epidural analgesia prevented both endotoxin-induced functional and structural alterations of gut mucosa in rabbits without modulating injured endothelial cells. Although further investigation is warranted, epidural analgesia could be applied to critically ill patients at risk of gut barrier dysfunction as a therapeutic option to preserve its functional integrity.

Departments of *Anesthesiology and †Medicine, Keio University School of Medicine, Tokyo, Japan

Supported, in part, by Grant-in-Aid for the Scientific Research from the Ministry of Education, Science and Culture (#14770790), Tokyo, Japan.

Accepted for publication December 2, 2004.

Address correspondence and reprint requests to H. Morisaki, MD, Department of Anesthesiology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan. Address e-mail to

Intestinal mucosa is anatomically vulnerable to any type of oxygen deficit because of its low oxygen tension, right angle branching microvessels, and countercurrent blood supply (1). Through the loss of its barrier function, the gut becomes a significant supplier of microorganisms and toxins to the systemic circulation, evoking the discharge of proinflammatory mediators and the development of multiple organ dysfunction syndrome (2). Whereas the preservation of gut integrity has become a therapeutic goal for critically ill patients, few approaches are clinically relevant in preventing the progression of gut injury (2,3).

Epidural analgesia has been demonstrated, in both clinical and experimental examinations, to attenuate the decrease of intramucosal pH (pHi) in patients undergoing major abdominal surgery (4) and to augment intramucosal microcirculation of the gut in rats (5). In addition, epidural analgesia increased splanchnic venous capacitance by depressing sympathetic nerve activity in a rabbit model (6). We previously demonstrated that thoracic epidural anesthesia and analgesia retarded the progression of intramucosal acidosis and prevented endotoxin influx to the portal vein during acute hypoxia in rabbits (7). However, a question remains whether epidural anesthesia and analgesia are still protective for gut mucosa in clinically relevant disease conditions such as sepsis, where tissue hypoxia is more complicated and substantial. Among several pathophysiological profiles, sepsis is characterized by massive discharges of proinflammatory cytokines, such as tumor necrosis factor-α (TNF-α) (8), by pathologic releases of nitric oxide (NO) to counteract oxygen radicals (9), and by considerable liberation of thrombomodulin, which acts as a cofactor of thrombin for protein C activation from injured endothelium (10). With a focus on the alterations of these mediators, we designed the present study to examine whether epidural analgesia preserved functional and structural integrity of the gut in a rabbit model of endotoxemia.

Back to Top | Article Outline


This protocol was approved by the Keio University School of Medicine Council on Animal Care in accordance with the guidelines of the National Institutes of Health.

Fifty healthy rabbits (New Zealand White, male; SEASCO, Saitama, Japan), weighing 2.0–2.5 kg (average, 2.3 kg) and fasted for 24 h, underwent instrumentation under general anesthesia. With sevoflurane 3%–4% inhaled in oxygen (3–4 L/min) via a face mask, the rabbits underwent tracheostomy and IV line access on the marginal ear vein. The rabbits were then mechanically ventilated to maintain normocapnia (fraction of inspiratory oxygen, 0.3; inspiratory pressure, 12–15 cm H2O; and 10–12 breaths/min) using an intensive care unit type ventilator (New Port E-100; New Port Medical Inc, Newport Beach, CA). An epidural catheter was placed via T11-12 interspace, as previously described (11), and an indwelling arterial catheter (22-gauge) was inserted into the right carotid artery. After a midline abdominal incision, a silastic catheter was inserted through the mesenteric vein to the distal portion of the portal vein. A perivascular probe was attached around the portal vein for measurement of portal blood flow (Transit-Time Ultrasound Flowmeter, T206; Transomic System Inc, Ithaca, NY) (12). A sigmoid tonometer catheter (Tonometrics, Worcester, MA) was surgically inserted into the terminal ileum via the ileocecal portion. To obviate the effects of inhaled anesthesia and to mimic the rabbits the condition of sepsis in the intensive care unit, inhaled anesthesia was discontinued, and a mixture of buprenorphine (0.1 mg/mL) and midazolam (2 mg/mL) was continuously infused at a rate of 1 mL/h. Rectal temperature was monitored and maintained at approximately 37°C. Rabbits were observed for 30 min before baseline measurements were made.

Back to Top | Article Outline

Study Protocol 1

After baseline measurements (baseline), 30 rabbits were randomly assigned to a control (group C; n = 14) or epidural (group E; n = 16) group using computer-generated random numbers. All rabbits in group E received a 0.4-mL/kg bolus injection of 0.5% lidocaine through the epidural catheter, followed by a continuous infusion of 0.1 mL · kg−1 · h−1, as described previously (13). Group C received the same doses of normal saline alone epidurally. After an equilibration period, the measurements of systemic and splanchnic variables were described in the measurement of pHi, and the specific measurements and calculations were performed (0 h). Thereafter, both groups received continuous infusion of 15 μg · kg−1 · h−1 of lipopolysaccharide (LPS) (Escherichia coli serotype 055:55B5; Sigma Chemical Co, St Louis, MO), accompanied by Ringer's acetate solution infusion at a rate of 25 mL · kg−1 · h−1, throughout the study periods. We chose the identical fluid volume between the groups to obviate a confounding factor for edema formation. In our pilot study using periaortic flow measurements, as previously described (12), the animals demonstrated a hyperdynamic circulatory pattern by manifesting an approximately 25% increase of cardiac output for 5 h, indicating that the animal model could be clinically relevant to septic patients throughout the 4-h study periods. The measurements of systemic and splanchnic variables were repeated at 2- and 4-h time periods. At the completion of the experiment, 0.4 mL/kg of indocyanine green was injected through epidural catheter, and if the cephalic spread of dye did not cover the range between the T4 and L1 level of the spine, the rabbit was excluded from the data collection. After tissue sampling was performed to determine wet-to-dry weight ratio and histological analysis of terminal ileum, the rabbits were killed with an IV pentobarbital overdose.

Gut pHi was monitored using automated air tonometry (Tonocap, Datex Ohmeda, Helsinki, Finland) (14). The measured regional Pco2 (Prco2), together with simultaneously obtained arterial [HCO3 ], were applied in the Henderson-Hasselbalch equation for calculation of pHi according to the manufacture's instruction:

[HCO3 ] being the arterial bicarbonate concentration, 6.1 the dissociation constant of HCO3 , and 0.03 the solubility of CO2 in plasma.

Arterial TNF-α plasma concentrations were measured using enzyme-linked immunosorbent assay that was developed in our laboratory. The assays were performed by using a combination of purified polyclonal goat anti-rabbit TNF antibody as a capture antibody and biotinylated polyclonal goat anti-rabbit TNF antibody for detection (15). Standard material, which was used in the rabbit TNF-conditioned medium (PharMingen, San Diego, CA), and obtained samples were run in duplicate. The detection limit in this assay was 13.7 pg/mL, and linear standard curves were obtained that ranged from 123 to 10,000 pg/mL.

The arterial thrombomodulin plasma level was measured by using enzyme-linked immunosorbent assay, as described elsewhere (16). Briefly, each well of a microtiter plate was coated with 1 μg/mL of goat anti-rabbit thrombomodulin immunoglobulin G dissolved in 0.1 mol/L of NaHCO3 (pH value of 9.6), and the plate was incubated overnight at 4°C. After blocking with Block Ace (Dainippon Seiyaku, CO, Ltd, Osaka, Japan), rabbit standard thrombomodulin (American Diagnostica Inc, Greenwich, CT) and test specimens were loaded into the wells, and the plate was incubated for 60 min at room temperature. Biotylated anti-thrombomodulin immunoglobulin G was added to the wells, followed by a 30-min incubation. After a 15-min incubation with avidin-peroxidase complex, a substrate solution containing 0.01% H2O2 and 0.4 mg/mL of o-phenyl-enediamine was added. The reaction was stopped using 4.5 N of H2SO4, and the absorbance at 490 nm was measured using the plate analyzer (ETY-3A, Toyo Sokki, Zama, Japan).

At the completion of experiments in Study Protocol 1, mucosal edema and microstructure of the terminal ileum were examined. Wet weights of five 2-cm parts of ileal tissues were measured and then dried in a vacuum oven (DP22; Yamato Scientific, Tokyo, Japan) at 95°C and −20 cm H2O for 48 h. The dry weights were determined, and the wet-to-dry weight ratio was calculated (12). Ileal samples for microscopic examination were obtained from the terminal ileum, which was distant from the tonometer catheter placement. Histological sections were evaluated in a blinded manner using light microscopy. Twenty-five random fields from each tissue were examined, and the degree of mucosal damage was graded on a scale of 0–4, with a modification of grading system previously described (17). In this classification, normal villi were graded as 0, mucosal edema limited to the apex of the villous tip and development of the subepithelial space as 1, extension of subepithelial space as 2, localized area of mucosal destruction or extensive submucosal edema was 3, and severe cell disruption was grade 4.

Arterial and portal pH, Pco2, Po2, and lactate concentrations were determined by using a blood gas analyzer (Chiron 860 series, Chiron Diagnostics Corp, East Walpole, MA). Hemoglobin and hemoglobin oxygen saturation were measured using a co-oximeter (OSM3, Radiometer, Copenhagen). Splanchnic oxygen extraction ratio was calculated using standard formulae: splanchnic oxygen extraction ratio (%) = 100 × (CaO2 − CpO2)/CaO2. Arterial and portal blood samples were centrifuged, and the plasma was stored at −80°C until analysis. NO release was assessed by the determination of stable NO metabolites (NO2 + NO3 ; NOx) in plasma using spectrophotometric assay (Cayman, Detroit, MI). Plasma lidocaine concentrations at 4 h in both groups were determined by fluorescence polarization immunoassay (TDX system, Abbot, North Chicago, IL). Measurements of all variables were performed in duplicate, and mean values were used for results.

Back to Top | Article Outline

Study Protocol 2

In a separate series of experiments, we determined the changes of gut permeability using fluorescein isothiocyanate-conjugated dextran with a molecular weight of 4000 Da (FD4) (13). After the same preparatory surgery, excluding the placement of the tonometer catheter and perivascular flow probe, 20 rabbits (New Zealand White, male; SEASCO) were randomized into groups C or E (n = 10 each group). The rabbits received the same dose of endotoxin infusion as those in Study Protocol 1. At 4 h, the abdomen was opened for preparation of an in situ loop of the gut. Briefly, double ligatures at both ends were made on the 10-cm length of the terminal ileum. Through a cannula placed into this segment of terminal ileum, FD4 (50 mg) was injected. After 30 min, blood samples from both the portal vein and artery were taken and centrifuged, and plasma FD4 concentrations were measured using fluorescence spectrometry (Spectrofluorophotometer:RF-1500; Shimadzu, Kyoto, Japan). Results were corrected for the plasma protein contents measured by the Lowry method.

Data are expressed as mean ± sd unless otherwise specified. Analysis of variance with repeated measures was used to evaluate the differences as shown using SPSS/11.0J for Windows (SPSS Inc, Chicago, IL). Separate analysis was performed if the interaction was statistically significant. When P < 0.05, the Scheffe multiple-comparison test was applied to distinguish differences between measurement variables. If the data were not normally distributed, the Friedman test was used to evaluate pair-wise comparisons. The histological scoring data were analyzed by χ2 test. Differences were considered statistically significant if P < 0.05.

Back to Top | Article Outline


Plasma lidocaine concentrations in both groups were all less than the undetectable level (<1.0 μg/mL). Because of misplacement of the epidural catheter, 2 rabbits in group E in Study Protocol 1 were excluded from the data collection.

Compared to the 0-h period, LPS infusion decreased mean arterial blood pressure (MAP) in both study groups (P < 0.05). The extent of hypotension in group E was significantly more than group C (P < 0.05) (Table 1). Heart rate was significantly depressed to a similar extent in both study groups compared with 0 h (P < 0.05). Whereas arterial Po2 remained constant throughout the study periods in both study groups, arterial oxygen content was significantly depressed to a larger extent in group E compared with group C (P < 0.05), mainly because of the progressive reduction of hemoglobin levels. Macroscopic hemolysis and persistent hemorrhage were not observed in either study group. Arterial pH values remained unchanged, whereas arterial lactate showed a mild increase at 4 h of LPS infusion in both groups (P < 0.05).

Table 1

Table 1

Portal blood flow remained constant during the study periods in both study groups (Table 2). Portal pH values decreased, and lactate concentration increased in both groups (P < 0.05). Splanchnic oxygen extraction ratio remained constant during LPS infusion despite a slight increase of portal lactate at 4 h (P < 0.05). The pHi values of group E were preserved within the normal range (>7.32), whereas those of group C decreased significantly after LPS administration (P < 0.05) (Table 2). The wet-to-dry weight ratio of terminal ileum in group E was significantly smaller versus group C (3.40 ± 1.99 versus 6.15 ± 1.27; P < 0.01), and the plasma FD4 concentrations in group E were also significantly less (3.26 ± 0.47 versus 4.13 ± 0.98 mg/mg-protein; P < 0.05), indicating that the increase of endothelial and intestinal wall permeability induced by LPS infusion were significantly reduced by epidural analgesia.

Table 2

Table 2

Figure 1 illustrates representative microscopic pictures of villi of distal ileum in normal rabbits (n = 3), group C rabbits, and group E rabbits using light microscopy. The villi of the rabbits in group C showed definitive evidence of mucosal injury, such as disruption of microvilli, lifting of the epithelium from the basal lamina, and submucosal edema (Panel B), whereas those in group E seemed near normal (Panel C). The mucosal injury of distal ileum observed in group C was significantly attenuated in group E (Fig. 2) (P < 0.01).

Figure 1

Figure 1

Figure 2

Figure 2

Figure 3 illustrates the changes of plasma NOx and TNF-α concentrations at the 2- and 4-h study periods. Both arterial and portal NOx levels were not changed during endotoxin infusion in either study group (Fig. 3A). The level of TNF-α in group C was significantly reduced in group E after 2 h of endotoxin infusion (Fig. 3B). The difference of TNF-α levels found at the 2-h period between the groups disappeared at 4 h. The portal thrombomodulin level was comparable at 0 h and increased significantly to the similar extent in both group C and E at 4 h after LPS infusion (213 ± 46 ng/mL to 1037 ± 991 ng/mL versus 225 ± 28 ng/mL to 1395 ± 878 ng/mL, respectively; P < 0.01), indicating that endotoxin-induced endothelial injury was not modulated by epidural analgesia.

Figure 3

Figure 3

Back to Top | Article Outline


The current study indicates that the application of epidural analgesia with lidocaine in endotoxemic hosts attenuates the progression of intramucosal acidosis, the increase of intestinal permeability, and the structural alterations of intestinal villi, possibly through the restoration of microcirculation. In addition, the protective effects are independent from modulation of endotoxin-induced endothelial cell activation to liberate thrombomodulin. These beneficial effects were noted, even though the epidural block was accompanied by a significant decrease of perfusion pressure and arterial oxygen content, which could be potent confounding factors to deteriorate gut mucosal oxygenation.

Our previous study using an acute hypoxia model showed the similar progression of moderate hemodilution (arterial hemoglobin, 7.7 ± 0.8 g/dL) under approximately one-fifth fluid volume resuscitation compared with this endotoxemia model (7). Because persistent bleeding was not observed after the preparatory surgery, aggressive fluid resuscitation and frequent blood sampling were the primary contributors to this progressive hemodilution in both groups. Thus, the augmentation of hemodilution only observed in group E could be attributed to the presence of epidural analgesia per se. A previous study demonstrated that thoracic epidural anesthesia with extra intravascular fluid administration induced significant hemodilution but not with basic fluid infusion in healthy pigs (18). Interestingly, such additional fluid volume did not restore epidural anesthesia-induced hypotension, which was consistent with our findings. Another study showed that intravascular fluid administration with crystalloid solution resulted in more pronounced hemodilution in patients who developed hypotension during epidural anesthesia (19). Although the precise mechanisms remain unclear (19), either hypovolemia or hypotension caused by epidural anesthesia is able to recruit a considerable quantity of fluid from skin and skeletal muscle (20,21) in combination with its restoring effects of increased vascular permeability during endotoxemia. Some may argue that moderate hemodilution per se rather than epidural lidocaine protects mucosal microcirculation of the gut. In a porcine model, gut mucosal oxygen supply was well maintained to systemic hematocrit values of approximately 10%, whereas serosal tissue oxygen supply decreased (22). Conversely, another study reported that acute severe hemodilution from 20% to 14% of hematocrit, close to the level observed in group E, exhausted the compensatory mechanisms of splanchnic circulation (23). In normal rats, the critical hematocrit for intestinal tissue oxygenation was approximately 16% (24). Therefore, the hemoglobin concentration observed in group E seems close to the level of critical hematocrit, not providing further beneficial aspects of hemodilution on the microcirculation.

To explain the initial host response against endotoxin exposure, it is postulated that prototypic cytokines like TNF-α are discharged first from macrophages and monocytes, subsequently initiating a second level of inflammatory cascades such as interleukin-6 and damaging gut mucosa by modulating splanchnic oxygen transport and intramucosal microcirculation (25). The TNF-α levels in endotoxemia are characterized by a rapid increase within 30 min, followed by a one-hour peak and a decrease during the next two hours (26). Considering these kinetics, epidural analgesia may have been able to modulate the process to discharge TNF-α at the early stage of endotoxemia. Simultaneously, the decrease of plasma TNF-α levels at a later stage, accompanied by a loss of significant difference between the groups, seems to be rational, although the serum TNF-α profile in human and animal sepsis models is an issue of much debate regarding its time course and magnitude (8,9,26–28). However, a local increase of such cytokines and oxygen radicals could, in turn, upregulate inducible NO synthase expression, leading to a prolonged increase of NOx levels in tissue and plasma. A previous study showed that endogenous production of NO played an important role in the modulation of gut permeability in rats (29). Although we were unable to show any significant alterations of the NOx levels in this model (Fig. 3), it was consistent with a previous animal study that demonstrated a species-specific modulation of NO pathway after endotoxin injection in rabbits (9). Previous studies showing a significant increase of NOx in endotoxemia were applied to a shock model with a single injection of large-dose endotoxin (9,30). However, in the present study, we intended to mimic a more clinically relevant model of sepsis by using small-dose endotoxin infusion, presenting a normotensive hyperdynamic circulatory state. Another important finding of this study was to show a marked increase of plasma thrombomodulin in both study groups. Thrombomodulin, which plays a crucial role in hemostasis by binding thrombin and subsequently converting protein C to its active form, has been recognized as a sensitive marker of endothelial injury (10). In the present study, such endothelial injury was not modulated by epidural analgesia. However, intestinal edema through an increased vascular permeability was significantly reduced by application of epidural analgesia. These conflicting results suggest that (a) upregulation of NO in this model might not be involved so significantly as is detected, (b) protein leakage through separation of endothelial tight junctions, rather than endothelial cell activation liberating thrombomodulin, is a major element to develop tissue edema, and (c) plasma thrombomodulin does not always mirror the whole profile of endothelial cell injury.

The clinical implications of the present study should be interpreted with caution because endotoxemia or severe sepsis is likely accompanied by the risk of infection such as meningitis or coagulation abnormalities such as low platelet count and coagulation factors that limit the indication of epidural analgesia. Furthermore, application of epidural anesthesia and analgesia may worsen the stability of systemic hemodynamics in critically ill patients who are treated with aggressive fluid resuscitation and vasoactive drugs. Indeed, MAP in group E was significantly depressed versus MAP in group C. Even with such a confounding factor, epidural analgesia shows a protective property on gut mucosa. Although it remains to be determined whether epidural analgesia can restore gut barrier dysfunction when it has already been established, epidural analgesia may be indicated for patients with functional obstruction of gastrointestinal tract or major vascular surgery who may develop increased mucosal permeability and subsequent bacterial translocation (31,32). In addition, potent analgesic effects of epidural blockade in this acutely instrumented model might provide more optimal analgesia and subsequently blunt stress responses, resulting in the reduction of discharges of vasoconstrictive mediators and the preservation of gut mucosal microcirculation. Finally, small plasma concentrations of lidocaine, absorbed from the epidural space in group E, could have modified the results, although they were all less than the detectable limit of the assay we applied. A previous study showed that lidocaine (plasma concentration, 1.4–2.5 μg/mL) reduced the extravasation of albumin in the lungs of endotoxemic rabbits (33). Another study reported that IV pretreatment with the same dose of lidocaine in rats attenuated endotoxin-induced increases in leukocyte adhesion and transvascular leakage against albumin (34). However, there is a possibility that small-dose lidocaine infusion, inducing plasma lidocaine at less than the detectable range, via the parenteral route might have produced the similar results.

In conclusion, the present study demonstrated that epidural analgesia prevented endotoxin-induced functional and structural alterations of gut mucosa in rabbits without modulating injured endothelial cells. Although further investigation is warranted, this study suggests that epidural anesthesia may be given to critically ill patients who are at risk of gut barrier dysfunction as a therapeutic option to preserve functional integrity.

The authors gratefully thank Dr. Etsuo Yoshida, Associate Professor, Department of Physiology, Miyazaki Medical College, Miyazaki, Japan, for his valuable instruction and Mr. Hirotaka Ishimori, Laboratory Technician, Tokyo Electric Power Company Hospital, Tokyo, Japan, for his expert technical assistance to this experiment.

Back to Top | Article Outline


1. Casley-Smith R, Gannon BJ. Intestinal microcirculation: spatial organization and fine structure. In: Shepherd AP, Granger DN, eds. Physiology of the intestinal circulation. New York: Raven Press, 1984:9–31.
2. Meakins JL, Marshall JC. The gastrointestinal tract: the motor of multiple organ failure. Arch Surg 1986;121:197–201.
3. Silva E, DeBacker D, Creteur J, Vincent JL. Effects of vasoactive drugs on gastric intramucosal pH. Crit Care Med 1998;26:1749–58.
4. Kapral S, Gollmann G, Bachmann D, et al. The effects of thoracic epidural anesthesia on intraoperative visceral perfusion and metabolism. Anesth Analg 1999;88:402–6.
5. Sielenkämper AW, Eicker K, van Aken H. Thoracic epidural anesthesia increases mucosal perfusion in ileum of rats. Anesthesiology 2000;93:844–51.
6. Hogan GH, Stekiel TA, Stadnicka A, et al. Region of epidural blockade determines sympathetic and mesenteric capacitance effects in rabbits. Anesthesiology 1995;83:604–10.
7. Ai K, Kotake Y, Satoh T, et al. Epidural anesthesia retards progression of intestinal acidosis and increase of portal endotoxin concentrations during acute hypoxia in rabbits. Anesthesiology 2001;94:263–9.
8. Damas P, Reuter A, Gysen P, et al. Tumor necrosis factor and interleukin-1 serum levels during severe sepsis in humans. Crit Care Med 1989;17:975–8.
9. Bachetti T, Pasini E, Suzuki H, Ferrari R. Species-specific modulation of the nitric oxide pathway after acute experimentally induced endotoxemia. Crit Care Med 2003;31:1509–14.
10. Krafte-Jacobs B, Brilli R. Increased circulating thrombomodulin in children with septic shock. Crit Care Med 1998;26:933–8.
11. Taguchi H, Murao K, Nakamura K, et al. Percutaneous chronic epidural catheterization in the rabbit. Acta Anaesthesiol Scand 1996;40:232–6.
12. Satoh T, Morisaki H, Ai K, et al. Olprinone, a phosphodiesterase III inhibitor, reduces gut mucosal injury and portal endotoxin level during acute hypoxia in rabbits. Anesthesiology 2003;98:1407–14.
13. Wang W, Small N, Wang P, Chaudry IH. Increased gut permeability after hemorrhage is associated with upregulation of local and systemic IL-6. J Surg Res 1998;79:39–46.
14. Bennett-Guerrero E, Panah MH, Bodian CA, et al. Automated detection of gastric luminal partial pressure of carbon dioxide during cardiovascular surgery using the Tonocap. Anesthesiology 2000;92:38–45.
15. Kajikawa O, Goodman RB, Johnson MC, et al. Sensitive and specific immunoassays to detect rabbit IL-8 and MCP-1: cytokines that mediate leukocyte recruitment to the lungs. J Immunol Methods 1996;197:19–29.
16. Edagawa M, Yoshida E, Matsuzaki Y, et al. Reduction of post-ischemic lung reperfusion injury by fibrinolytic activity suppression. Transplantation 1999;67:944–9.
17. Chiu CJ, McArdle AH, Brown R, et al. Intestinal mucosal lesion in low-flow states. Arch Surg 1970;101:478–83.
18. Vagts DA, Iber T, Puccini M, et al. The effects of thoracic epidural anesthesia on hepatic perfusion and oxygenation in healthy pigs during general anesthesia and surgical stress. Anesth Analg 2003;97:1824–32.
19. Drobin D, Hahn RG. Time course of increased haemodilution in hypotension induced by extradural anaesthesia. Br J Anaesth 1996;77:223–6.
20. Haisjackl M, Luz G, Sparr H, et al. The effects of progressive anemia on jejunal mucosal and serosal tissue oxygenation in pigs. Anesth Analg 1997;84:538–44.
21. Länne T, Lundvall J. Very rapid net transcapillary fluid absorption from skeletal muscle and skin in man during profound hypovolemia circulatory stress. Acta Physiol Scand 1989;136:1–6.
22. Hahn RG. Origin of intravascular fluid recruited by vasodilation during epidural anaesthesia. Eur Surg Res 1996;28:70–74.
23. Nöldge GFE, Priebe H-J, Bohle W, et al. Effects of acute normovolemic hemodilution on splanchnic oxygenation and on hepatic histology and metabolism in anesthetized pigs. Anesthesiology 1991;74:908–18.
24. van Bommel J, Siegemund M, Henny P, et al. Critical hematocrit in intestinal tissue oxygenation during severe normovolemic hemodilution. Anesthesiology 2001;94:152–60.
25. Sair M, Etherington PJ, Curzen NP, et al. Tissue oxygenation and perfusion in endotoxemia. Am J Physiol 1996;271:H1620–5.
26. Losser MR, Bernard C, Beaudeux JL, et al. Glucose modulates hemodynamic, metabolic, and inflammatory responses to lipopolysaccharide in rabbits. J Appl Physiol 1997;83:1566–74.
27. Mathiak G, Szewczyk D, Abdullah F, et al. An improved clinically relevant sepsis model in the conscious rat. Crit Care Med 2000;28:1947–52.
28. Kragsbjerg P, Holmberg H, Vikerfors T. Dynamics of blood cytokine concentrations in patients with bacteremic infections. Scand J Infect Dis 1996;28:391–8.
29. Kubes P. Nitric oxide modulates epithelial permeability in the feline small intestine. Am J Physiol Gastrointest Liver Physiol 1992;262:G1138–42.
30. Gardiner SM, Kemp PA, March JE, et al. Cardiac and regional haemodynamics, inducible nitric oxide synthase (NOS) activity, and the effects of NOS inhibitors in conscious, endotoxaemic rats. Br J Pharmacol 1995;116:2005–16.
31. Deitch EA. Simple intestinal obstruction causes bacterial translocation in man. Arch Surg 1989;124:699–701.
32. Roumen RM, Frieling JT, van Tits HW, et al. Endotoxemia after major vascular operations. J Vasc Surg 1993;18:853–7.
33. Nishina K, Mikawa K, Takao Y, et al. Intravenous lidocaine attenuates acute lung injury induced by hydrochloric acid aspiration in rabbits. Anesthesiology 1998;88:1300–9.
34. Schmidt W, Schmidt H, Bauer H, et al. Influence of lidocaine on endotoxin-induced leukocyte-endothelial cell adhesion and macromolecular leakage in vivo. Anesthesiology 1997;87:617–4.
© 2005 International Anesthesia Research Society