Multiple organ dysfunction syndrome (MODS) is the final and often lethal stage that can occur as a progression of septic or hemorrhagic shock. The mechanism of MODS is thought to be excessive systemic inflammation (systemic inflammatory response syndrome [SIRS]) that affects a complex array of molecular pathways (1) causing alterations in cell metabolism (2), apoptosis, and necrosis (3). Other key components in MODS pathophysiology are alterations in microvascular perfusion (4), coagulopathy (5), and increased vascular permeability (6). These pathologic changes can cause dysfunction in all of the major organ systems, including respiratory, cardiovascular, hepatic, renal, gastrointestinal, hematological, immune, endocrine, and the central nervous system (7).
Not all patients that develop septic or hemorrhagic shock progress to MODS. The severity of the sepsis or hemorrhage, comorbidities, and patient genetics all play a role. An important factor in the development of MODS is exposure of the patient to a second insult or "hit." For example, it has been shown that hemorrhagic shock in rats causes neutrophil priming but no lung or liver damage (8). If, however, the intra-abdominal pressure (IAP) is raised to levels simulating the abdominal compartment syndrome (ACS), these primed neutrophils become activated, causing both lung and liver damage (8). A corollary to this second-hit theory is that there is often an underlying driving force or "motor" that leads to the development of MODS.
It was initially postulated that bacterial translocation from the gut was the motor of MODS (9). Other studies have suggested that the lung (10) and altered microcirculation may be the motor of sepsis and MODS (11). Although there is no consensus as to the organ system that actually generates MODS, it is well established that injury in one organ system or the microvasculature can cause injury to a second, distant, organ system (9).
From an etiological perspective, it is likely that the gut is the initial motor of MODS. Microcirculation in the gut is preferentially disrupted in both septic (12, 13) and hemorrhagic shock (I/R) (14, 15). This can lead to tissue hypoxia and inflammation-induced alteration in both endothelial (16) and epithelial function (17). Loss of barrier function can increase permeability and result in intestinal edema and ascites formation (11), which can contribute to intra-abdominal hypertension (IAH) and ACS (18, 19). In addition, the damaged gut can be a continual source of inflammation propagating SIRS, which my lead to MODS (5, 7, 9, 17, 19).
In this study, we used a clinically applicable animal model of sepsis and I/R-induced organ injury and hypothesized that increased gut capillary permeability would cause an inflammatory peritoneal ascites that would perpetuate damage to multiple organs; application of peritoneal negative pressure therapy (NPT) would remove ascites and reduce both morbidity and mortality. We further hypothesized that the mechanism of protection would be the attenuation of excessive inflammation in the peritoneal cavity, mitigating SIRS and eliminating the second hit necessary to propagate organ damage.
MATERIALS AND METHODS
The experiment was performed in compliance with the National Institutes of Health's Guidelines on the Use of Laboratory Animals. The CHUA Committee at Upstate University Hospital approved the study protocol.
Animals and preparations
Female Yorkshire pigs (21-38 kg) were pretreated with glycopyrrolate (0.01 mg kg−1, i.m.), telazol (5 mg kg−1 tiletamine hydrochloride and zolazepam hydrochloride, i.m.), and xylazine (2 mg kg−1, i.m.). A ketamine (3 mg mL−1) plus xylazine (0.3 mg mL−1) continuous infusion using an infusion pump (3M model 3000) was used to maintain anesthesia throughout the experiment. Ketamine and xylazine were injected into a 1-L bag of Ringers lactate and infused at a rate of 100 mL h−1 for the duration of the experiment. This rate could be adjusted to maintain adequate anesthesia, which was determined by lack of eye and pinch reflexes with no dramatic increases in heart rate or arterial blood pressure. An open tracheotomy was performed and the animal connected to a G5 ventilator (Hamilton Medical, Reno, Nev) with initial settings during the surgical preparation as follows: tidal volume (Vt) of 12 mL kg−1, respiratory rate (RR) of 15 breaths min−1, titrated to maintain PaCO2 within the normal range (35-45 cmH2O), FiO2 of 21%, and positive end-expiratory pressure (PEEP) of 3 cmH2O. Lung volume history was standardized by recruiting the lung using the PV TOOL (Hamilton Medical). The lung was inflated to a peak pressure of 30 cmH2O, held for 5 s, and deflated back to 5 cmH2O of PEEP. The lung was automatically sighed (50% increase in Vt every 50 breaths) throughout the experiment. Low Vt protective mechanical ventilation was not used because we did not want to "protect" the lung with the ventilator but rather assess the development of lung injury if it occurred. Under sterile conditions, a left carotid artery catheter was placed for blood chemistry and gas content measurements (Roche Cobras b211; Roche Diagnostics, Indianapolis, Ind), and systemic arterial pressure monitoring. A 4-cm right lateral neck incision was made, and a veinotomy was performed on the right internal jugular vein for placement of a triple lumen catheter, allowing anesthesia, fluid, and antibiotic administration. A right internal jugular Swan-Ganz catheter (7 French) was placed for measurement of pulmonary artery (PAP) and wedge pressures (PAW), sampling of mixed venous blood gases, and cardiac output (CO; Agilent CMS-2001, Boebingen, Germany). A Foley catheter was inserted into the bladder for measurement of urine output (UOP), collection of urine samples, and was connected to a pressure transducer leveled at midline to measure bladder pressure. The health of the animal was determined by normal hemodynamic and lung function parameters following instrumentation and normal blood gases and chemistries. No animal had to be killed due to disease at baseline (BL).
A computerized number generator was used for randomization with even-numbered animals going into the NPT group and odd-numbered animals going into the passive drain (PD) group. For the first 12 h of the protocol, the animals were treated in an identical fashion. All animals received a regimen of intravenous fluids and antibiotics at a dose and quantity established in our initial experiments (Fluids and antibiotics). Pulmonary and arterial blood pressure and gases, cardiac output, UOP, heart rate, IAP, and temperature were continually monitored and recorded every 60 min. A midline laparotomy was performed, and the superior mesenteric artery (SMA) was isolated and clamped for 30 min to induce intestinal ischemia. This was confirmed by the loss of the mesenteric pulse and discoloration of the bowel. After 30 min, the clamp was removed, and reperfusion was confirmed by the reappearance of the mesenteric pulse and the return of normal color to the bowel. At this point, the cecum was identified and brought out of the peritoneum, and an enterotomy of 2 cm was performed to harvest feces (0.5 mL kg−1). Feces was combined with 2 mL kg−1 of the pig's blood to create a fecal-blood clot. The cecum was left open and returned to the abdominal cavity, and the clot was implanted into the right lower quadrant of the abdominal cavity. A catheter was placed in Morrison pouch between the liver and right kidney and brought out through the skin for collection of peritoneal ascites. This catheter was sutured to the skin in a purse-string fashion. Collected ascites was flash-frozen for measurement of inflammatory mediators. The abdomen was then closed with sutures and the time recorded as T0 (i.e. 0 h after injury; Fig. 1).
The entire abdomen was reopened at T12, and the V.A.C. Abdominal Dressing System (KCI, Inc., San Antonio, Tex) was applied to the open wound as per packet instructions (Fig. 1). The fenestrated bioinert plastic that housed the sponge material was in direct contact with the intestine. However, to prevent iatrogenic bowel injury, care was taken to ensure that the dressing sponge did not touch the bowel. At this point, animals were randomly assigned to treatment groups as follows: control group (PD; n = 6; 26.5 ± 1.2 kg) had the dressing placed, but the vacuum was not activated (i.e. negative pressure was not applied); however, the drain line was left open to allow passive drainage of ascites. Experimental group (NPT; n = 6; 28.3 ± 2.5 kg) had the dressing placed and the vacuum activated so that negative pressure (−125 mmHg) was applied continuously for the remainder of the experiment.
Hemodynamic measurements and calculations
Elecrocardiogram monitoring, pulse oximetry, MAP, central venous pressure, PAP, and PAW were measured (Agilent, CMS-2001 System M1176A, with Monitor M1094B) using Edwards transducers (Pressure Monitoring Kit PXMK1183; Edwards Lifesciences, Irvine, Calif). Urine output was recorded at BL (after surgical preparation), immediately after injury (T0), and every hour for 48 h. Cardiac output was measured by thermodilution (CMS-2001 System M1176A with Monitor M1094B; Agilent). Three separate boluses of cold solution (dextrose 5% and sodium chloride 0.45%) were injected at end-expiration and the average of the three measurements recorded. Physiologic measurements were made hourly (BL, T0 - T48; Fig. 1). Pulmonary parameters such as RR, peak airway pressure (Ppeak), mean airway pressure (Pmean), plateau pressure, Vt, PEEP, auto-PEEP, expiratory minute volume (EMV), and static compliance (Cstat) were measured or calculated by the G5 ventilator (Hamilton Medical) at BL and every hour after injury for 48 h or until death.
Pulmonary mechanics were measured by the following methods: Ppeak, by definition, is the highest airway pressure measured during the breath cycle. Plateau pressure was measured if a stable plateau pressure (a pressure change <1 cmH2O over 100 ms) was present during the late inspiratory phase. Mean airway pressure was the moving average of the pressures measured during eight consecutive cycled breaths. Static compliance and auto-PEEP were calculated using the least squares fit method to the entire breath. Lung function was also assessed via arterial blood gases (PaO2 and PaCO2) and oxygenation expressed as a P-F ratio (PaO2/FiO2) every hour for the first 6 h (T0 - T6) and every 6 h thereafter (T6 - T48). Lung function parameters were measured hourly (BL, T0 - T48; Fig. 1).
Management: fluids, antibiotics, and mechanical ventilation
Fluids and antibiotics-
During the surgical procedure, pigs received a fluid bolus of Ringers lactate (1 L, i.v.) over 30 min. After T0 measurements, broad-spectrum antibiotics (2 g ampicillin, i.v. [Bristol Myers Squibb, Princeton, NJ] and 500 mg Flagyl, i.v. [Baxter, Deerfield, Ill]) were delivered over 15 min. This antibiotic regimen was repeated at 12, 24, and 36 h postinjury (Fig. 1). All fluid was warmed to 36°C in a water bath (Precision 280 Series, Thermo Electronic Corp., Waltham, Mass). Ketamine + xylazine anesthesia mixed in Ringers (see Anesthesia for details) was infused at a rate of 100 mL h−1 for a total of 4.8 L in 48 h. This infusion rate was adjusted to maintain proper anesthesia, and all changes were recorded on the data sheet. A second intravenous dose of Ringers with a drip sufficient to keep the vein open was used to maintain adequate hydration determined by physiology parameters (UOP and MAP). Hydration was deemed inadequate if UOP decreased to less than 0.5 mL kg−1 h−1 or if MAP decreased to less than 60 mmHg. If either of these effects occurred, pigs received an additional fluid bolus of Ringers lactate (500 mL, i.v.). All fluids infused or withdrawn were recorded and used to analyze fluid balance.
If arterial desaturation occurred (SaO2 fell below 92%), the FiO2 was increased to maintain adequate oxygenation. If 100% FiO2 did not maintain adequate oxygenation, the PEEP was increased in 2-cmH2O increments (maximum PEEP allowed was 15 cmH2O) until adequate oxygenation was obtained or hemodynamics were compromised (i.e. significant fall in MAP).
Assessment of organ function
Kidney function was assessed by measuring UOP and arterial blood urea nitrogen (BUN) level (Roche Cobras b221) at BL, every hour for the first 6 h (BL, T0 - T6), and every 6 h thereafter (Fig. 1). Urine was flash-frozen for measurement of protein concentration. Liver function was assessed by plasma concentrations of alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin, albumin, and total protein. Measurement of blood gases and chemistries were made with a Roche Blood gas analyzer (Cobras b221). The parameters listed below were measured at BL, every hour for the first 6 h (BL, T0 - T6), and every 6 h thereafter after injury (Fig. 1). Both arterial and mixed venous samples were measured for pH, PcO2, PO2, SO2%, hematocrit, hemoglobin, sodium, potassium, chloride, ionized calcium, glucose, BUN, and lactate. Prothrombin time (PT), international normalization ratio (INR), and activated partial thromboplastin time were measured. A complete blood count (CBC) with differential, including white blood cell (WBC) count, hemoglobin, hematocrit, and platelets were measured every 12 h (Fig. 1). All measurements not described were made by the Clinical Pathology Department at Upstate University Hospital using standard procedures. The criteria for organ dysfunction were as follows: acute renal failure, UOP less than 0.5 mL kg−1 h−1 for over 6 h plus BUN greater than 60 mg dL−1 (20) and significant histopathology; shock liver, ALT greater than 80 IU L−1, AST greater than 80 IU L−1, INR greater than 1.5, and total bilirubin greater than 2 mg dL−1 (21) with significant histopathology; and acute lung injury (ALI), P/F ratio less than 250 with a Cstat less than 10 mL cmH2O−1 and significant histopathology.
Inflammatory mediators sampled
Inflammatory mediators were measured in the plasma, peritoneal ascites fluid (Pfluid) and bronchoalveolar lavage fluid (BALF) (Fig. 1). Bronchoalveolar lavage fluid was collected at necropsy by lavage (60 mL saline) of the right middle lobe. All samples were spun at 3,500 RPM at 15°C for 10 min, snap-frozen in liquid nitrogen, and stored at −80°C.
TNF-α, IL-1β, IL-6, IL-8, IL-12, and transforming growth factor beta (TGF-β; R&D Systems, Minneapolis, MN), IL-10, C5a (Immuno-biological Laboratories, Minneapolis, Minn), von Willebrand factor (vWf; American Diagnostics Inc. Stanford, Conn), and C-reactive protein (CRP) (Immunology Consultant's Laboratories, Inc., New Burg, Ore) were measured using pig-specific enzyme-linked immunosorbent assays according to the manufacturer's assigned specifications. Prostaglandin E2 (PGE2; Oxford Biomedical Research, Oxford, Mich) and antioxidants (Cayman Chemical Company, Ann Arbor, Mich) were tested using commercially available EIA kits. Endotoxin was tested using an end-point chromogenic limulus amebocyte lysate assay (Lonza Group Ltd., Basel, Switzerland). Total protein was measured in plasma, Pfluid, BALF, and urine using a bicinchoninic colormetric assay (Pierce Biotechnology, Rockford, Ill). Urine proteins were first precipitated using technical resource 0049.0 (Pierce Biotechnology). Blood cultures (aerobic and anaerobic) and Gram staining were performed by the Upstate Medical University Clinical Pathology Department using standard techniques.
The heart and lungs were removed en bloc. Gross photographs were obtained after the lungs were inflated to peak airway pressure of 25 cmH2O. The heart was then removed, and the bronchus to the right middle lobe was exposed, cannulated with a small endotracheal tube, and the balloon was inflated to secure the endotracheal tube. The lobe was lavaged with 60 mL isotonic sodium chloride solution (three 20-mL injections) and collected. The BALF was centrifuged and frozen for measurement of inflammatory mediators. The right mainstem bronchus was clamped, and the left lung was filled to a level of 25 cm to standardize fixation pressure with 10% neutral buffered formalin, and the tracheal tube was clamped. The lung was then immersed in formalin for a minimum of 48 h before histologic sectioning. Histologic analyses from both dependent and nondependent lung areas were made. The dependent area was obtained by measuring 3 cm from the aortic groove, and two samples were harvested from the most medial section. The nondependent sample was selected by measuring 3 cm medial from the lateral lung margin. The kidney was cut in half along the central axis and a section cut from the center of the interior of the kidney containing both cortex and medulla. Representative samples of small intestine were harvested from the duodenum, mid-jejunum, and ileum; these were sectioned for histologic analysis. A 3-cm section of liver was harvested from the center of the left lobe. All tissues were fixed in 10% buffered formalin. Standard 5-μm paraffin sections were made and stained with hematoxylin and eosin. Predetermined, randomized, and unbiased quantitative histometric analysis of the tissue was then performed. High-resolution photomicrographs were taken of random sampling areas from each section and analyzed using a predetermined scoring system. The lung was assessed for atelectasis, fibrinous deposits, blood in the airspaces, vessel congestion, alveolar wall thickness, and leukocyte infiltration. The kidney was assessed for interstitial edema, epithelial changes, tubular degeneration, capillary congestion, and leukocyte infiltration. Liver samples were assessed for interstitial edema, sinusoid congestion, hepatocellular necrosis, hepatocyte vacuolation, and leukocyte infiltration. The small intestine was assessed for shortening of the villi, Gruenhagen spaces, epithelial lesions, denuding of mucosa, capillary congestion, lesions to the glands, and edema (see Histological Parameters, Supplemental Digital Content 1, https://links.lww.com/SHK/A49).
Organ edema was measured using the W-D ratio = (wet weight − dry weight) for all four organs. Tissue was excised, minced, immediately weighed, and placed in an oven at 60°C and allowed to dry. Dry weight was determined when the weight did not change over a 24-h period. If the animal died before T48, the tissues were collected for W-D at the time of death.
Animals were treated identically until they were randomized into groups at T12; thus, figures show data BL to T12 presented with all 12 animals in aggregate. Continuous parameters were summarized by randomized group assignment and time using descriptive statistics; categorical parameters were summarized using counts and percentages. Survival rates between the two treatment groups were compared using the event time distribution functions and a log-rank test. The primary analysis was a two-factor (treatment group and time [T12 - T48]) repeated-measures analysis of covariance (RM ANCOVA), using T12 as the covariate based on predicted-imputed data (the interactions term Group × Time remained in the statistical model if the probability value was <0.05). Significant results from the model were further examined at each time point after T12 for the intergroup examination (PD vs. NPT; times = 12, 24, 36, 48) using a Bonferroni correction to adjust for multiplicity. For animals that failed to survive to T48, a least-squares regression model was used to calculate the intra-animal predicted value for the RM analyses. This technique allowed the RM analyses to be performed using all of the animals randomized and treated in the study. For intragroup examination (PD and NPT), we used a one-factor (time) RM analysis from T12 to T48, followed by a Dunnett test to compare each time point to T48. The same analysis was used on data from BL to T12 on the aggregate animal population. Quantitative histology data were analyzed using a Mann-Whitney U test after testing for normality. W-D ratios and fluid balances were analyzed using an unpaired Student t test. Probability values less than 0.05 were considered significant. Data are presented as mean ± SEM. All analyses were performed using version 9.13 of the Statistical Analysis System from the SAS Institute (Cary, NC), version 188.8.131.52. of JMP (SAS Institute), or version 5.0a of GraphPad Prism (La Jolla, Calif).
Mortality, hemodynamics, pulmonary, kidney, and liver function
Lung compliance was significantly improved with NPT as compared with PD (Table 1). Animals in the NPT group had significantly lower Pmean, peak inspiratory pressures, and plateau pressures than animals in the PD group (Table 1). PaO2, PaCO2, SaO2, and SvO2 were not different between groups (Table 1). Oxygenation expressed as a P-F ratio was higher in the NPT group, albeit not significantly different from PD (P = 0.0812). The mortality rate in the NPT was 17% versus 50% in the PD, but this difference was not statistically significant (P = 0.1859). Negative pressure therapy demonstrated a significant improvement in cardiac output and PAW (Table 1). Minimal differences were seen between groups in HR, MAP, or PAP (Table 1). Urine output was significantly lower in the PD as compared with the NPT group (Table 1) despite more fluids given to this group (Fig. 2). A significantly elevated IAP (measured via the bladder pressure) in the PD group as compared with the NPT may have been one of the mechanisms contributing to the reduced UOP (Table 1). There was a significant difference in AST and albumin by treatment × time in the NPT group compared with the PD, with little change in alkaline phosphatase (Table 2). Alanine aminotransferase decreased to a similar degree in each group (Table 2). Unlike other organ systems, there are few clinically measured parameters that describe intestinal function. Thus, we described intestinal injury solely on necropsy results. There was a significant reduction in intestinal edema (W-D) in the NPT as compared with the PD group (PD = 7.44 ± 0.46; NPT = 5.97 ± 0.45, P < 0.05). There were no significant differences in W-D in the other organs (lung, liver, kidney). There was also a significant improvement in histopathology in the NPT as compared with the PD group (see Histology results below).
Similar volumes of fluids were infused in both groups, and both groups had a positive fluid balance (Fig. 2). The NPT group had a significantly higher UOP, which translated into a significant decrease in the volume of positive fluid as compared with the PD group (Fig. 2). Application of NPT removed a larger volume of ascites (864.64 ± 13.09) than did PD (88.3 ml ± 56.30).
Blood chemistry, coagulation, leukocytes, and platelets
Blood lactate concentration rose dramatically in both groups in response to the removal of the SMA clamp, subsequently returned to normal values, and then rose slightly toward the end of the study (Table 2). Animals in both groups had evidence of coagulopathy as shown by elevations in prothrombin time and INR (Table 2). White blood cell counts decreased in both groups between BL and T48; however, a significantly greater amount of WBCs were observed in the NPT group as compared with PD (Table 2). In terms of WBC composition, the percentage of lymphocytes was significantly depressed in the NPT group compared with PD, whereas neutrophils (Table 2) were significantly elevated. Platelets differed significantly by Treatment × Time (Table 2). There were significant differences between NPT and PD in blood levels of Mg2+ and Ca2+ (Table 2).
Inflammatory mediators: Pfluid and plasma
Peritoneal ascites fluid-
Negative pressure therapy significantly reduced IL-6 and IL-8 (Fig. 3). There were no significant differences between groups in vWf, CRP, TGF-β, PGE2, endotoxin, antioxidants, C5A, or IL-10 (data not shown).
There was significant reduction in TNF-α, IL-12, IL-6, and IL-1β with NPT as compared with the PD group (Fig. 4). No significant differences were found between groups in vWf, CRP, TGF-β, PGE2, endotoxin, and antioxidants (data not shown). In addition, IL-6 and IL-1β differed significantly by Treatment × Time (Fig. 4).
Both dependent and nondependent sections of the lung showed a significant decrease in atelectasis, fibrinous deposits, and leukocyte infiltration when NPT was used (Table 3). Negative pressure therapy also caused a significant decrease in all kidney and intestine mucosal parameters assessed, with the exception of Gruenhagen spaces, as well as a significant decrease in edema in the intestinal wall (Table 3). In the liver, NPT led to significant decreases in interstitial edema, hepatocellular necrosis, and interstitial WBCs (Table 3). Histology of the intestine was internally consistent with no sign of localized evidence of injury. This was in agreement with the gross observation that there were no signs of damage or fistula formation to the bowel that was in contact with the fenestrated bioinert plastic VAC dressing (Fig. 5).
All animals in the study became bacteremic as a result of injury. Multiple species of gram-positive and gram-negative bacteria (Serratia maecescans, Pseudomonas aeruginosa, Klebsiella pneumoniae, Escherichia coli, Enterococcus species, and Enterobacter cloacae) were cultured from blood collected at T24 and T48.
The most important finding of this study was that NPT reduced histologic damage to the lungs, intestine, kidney, and liver. The results suggest that the mechanism for this protection involved removal of inflammatory peritoneal ascites causing a moderation of systemic inflammation (SIRS), which diminished end organ damage. Although NPT significantly reduced the histopathology in all organs measured, a concomitant improvement in organ function cannot be conclusively asserted. The mean values for lung function did meet our criteria for ALI in the PD group; however, there was variability within both groups, with 66% of the animals in the PD group meeting ALI criteria and 33% of the animals in the NPT group. This variability combined with the small number of animals does not allow us to draw the conclusion that NPT reduced ALI. These data also suggest that histologic injury is a very sensitive predictor that is manifest before the organ becomes clinically dysfunctional. We speculate that some organ failure would have occurred if the experiment had been performed for another 24 h.
The disparity in fluid balance may have contributed to the difference in physiologic and histologic response between groups. It has been shown that aggressive fluid resuscitation significantly increases mortality in both endotoxin and peritonitis animal models (22). In both of these sepsis models, lung, liver, and kidney histopathology were significantly elevated in the group with aggressive fluid resuscitation (high fluid volume, 15 mL kg−1 h−1 Ringers lactate + 5 mL kg−1 h−1 hydroxyethyl starch vs. moderate fluid volume, 10 mL kg−1 h−1 Ringers lactate) (22). In the Brandt study, the postinjury treatment was identical except for the fluid resuscitation (22). In the current study, postinjury treatment varied (PD vs. NPT), and the fluids given were in response to physiologic changes (i.e. UOP and MAP). The animals in the PD group required more fluid to maintain UOP and MAP, suggesting that kidney and cardiac function were more severely impaired than in the NPT group. Thus, in our study, it is impossible to determine if the increase in histopathology was the result of the increased fluids or the absence of NPT treatment. The mechanism of NPT-induced UOP increase is not known. Intra-abdominal pressure was lower in the NPT group, which may have improved renal perfusion. The reduction in systemic inflammation in the NPT group may also have reduced renal histopathology, which would also improve renal function and UOP.
We hypothesize that the combination of ascites removal and reduction of the inflammatory milieu in the peritoneum are the most likely mechanisms for the reduced morbidity observed with NPT. Application of NPT removed a larger volume of ascites (864.64 ± 13.09) than did PD (88.3 ± 56.30 mL), thereby reducing the intestinal surface in contact with the ascites in addition to reducing both IL-6 and IL-8 concentrations in the ascites that remained. We think that this reduction in peritoneal inflammation was responsible, in part, for the blunted systemic inflammatory response in the NPT group, with plasma concentrations of TNF-α, IL1-β, IL-6, and IL-12 all diminished. Negative pressure therapy reduced IAP, which may also play a role in the reduction of systemic inflammation and organ damage (8). Negative peritoneal pressure could also contribute to NPT-induced protection by other mechanisms such as improving capillary and lymph flow (23). Although the precise mechanisms demand further study, these data clearly demonstrate that NPT altered inflammation in the gut and reduced the organ injury associated with our porcine model.
Our data further suggest that inflammatory ascites, rather than bacterial translocation, is the motor driving organ histopathology, at least in this animal model (24). Although NPT undoubtedly removed more bacteria from the peritoneal cavity than did PD, the fact remains that all pigs had polymicrobial bacteremia at T24 and T48. This suggests that bacterial translocation was not the mechanism driving organ failure because organ injury was attenuated with NPT despite the presence of polymicrobial systemic bacteremia. It would seem, based on these data, that organ injury can be prevented even in the presence of significant bacteremia by reducing the inflammatory milieu in the gut.
The ability of inflammatory ascites to perpetuate systemic inflammation is a deduction based on certain longstanding clinical observations as well as recent research. It is well known that intestinal capillaries can exchange molecules with fluid in the peritoneal cavity because this is the mechanism used routinely in peritoneal dialysis for renal failure (25). In fact, it has been shown that dextrose-based direct peritoneal resuscitation in animal models of hemorrhagic shock prevents persistent splanchnic vasoconstriction, improves perfusion to all organ systems, down-regulates gut-associated inflammation, prevents MODS, and improves survival (26, 27). Similarly, if the peritoneum were to fill with a proinflammatory ascites, it is reasonable to deduce that such fluid could act as a reservoir rich in inflammatory mediators and provide a steady source of inflammation to the systemic circulation.
Although inflammatory ascites being a motor of MODS is a novel concept, it is closely linked to the established "gut-lymph" hypothesis (19, 28). The difference between the "gut-ascites" and "gut-lymph" hypotheses is that the former postulates that inflammatory ascites functions as a motor of organ injury, whereas the latter postulates that the motor is toxic mesenteric lymph. It is quite probable that these two hypotheses actually exist along a continuum of the pathologic process and thus are not separate entities. The initial sequence of events is identical in both hypotheses: septic or hemorrhagic shock causes a hypoxia and inflammation-induced injury to the intestine. This injury causes increased vascular permeability, resulting in third space fluids rich in inflammatory mediators that either are removed by lymphatic flow, become interstitial edema, or leak into the peritoneal space as ascites. Lymph coming from the injured gut, containing high concentrations of toxic and inflammatory mediators, enters the circulation via the thoracic duct immediately cephalad to the lung causing adult respiratory distress syndrome (ARDS) (29). Ascites rich in inflammatory mediators surround the intestine, thus perpetuating systemic inflammation and promoting end organ injury (Table 3). It is likely that both toxic third space fluids (lymph and ascites) are working synergistically to cause multiple organ injury.
Trauma and sepsis-induced injury to the intestine causing increased vascular permeability, edema, and ascites is a significant clinical problem. In a recent study by Vidal et al. (30), it was shown that 31% of patients had IAH (bladder pressure >12 mmHg) upon admission to the intensive care unit (ICU), and another 33% developed IAH while in the ICU. The sentry criteria for this study were all consecutive patients admitted to the ICU between November 1, 2004 and July 31, 2005, and expected to stay for more than 24 h. Thus, in the study of Vidal et al., 64% of the patients in the ICU potentially had an intestinal injury causing edema or ascites sufficient to generate intra-abdominal hypertension. If this study accurately represents the ICU population, well more than half of all ICU patients potentially could have a gut injury that results in accumulation of inflammatory ascites that may result in MODS. Early diagnosis and removal of ascites accumulation would, in theory, significantly reduce the morbidity and mortality associated with MODS.
Malbrain and De Laet (31) recently stated that the concept of increased vascular permeability with accumulation of extravascular fluid in tissues secondary to sepsis or I/R is a readily accepted mechanism for injury in the lung and kidney, and the same pathologic processes must be occurring in the gut. They add that the gut as the motor of MODS cannot be denied and further state that difficulties in assessing gut function should not prevent the realization of this concept. The injury to the gut microvascular is identical to that which causes ARDS. Thus, the increased IAP, which is the hallmark of ACS, is only a symptom of this intestinal injury and is not itself a disease. Indeed, this acute intestinal permeability syndrome is as complex a syndrome as ARDS; however, due to the difficulty of identifying it, this syndrome is virtually unknown.
In summary, this study used a sophisticated, clinically applicable porcine model of multiple organ injury and showed that application of peritoneal NPT 12 h after injury significantly reduced lung, kidney, liver, and intestinal pathology and improved pulmonary mechanics. We postulate that reducing the peritoneal inflammatory milieu by NPT was the likely mechanism of protection. Negative pressure therapy might be a viable modality to treat complex septic and trauma patients and may be considered as a therapeutic intervention to be initiated at the earliest stages of care in critically ill patients at risk of developing ARDS, ACS, and MODS.
Critique of the study
Before experimentation began, we determined that the appropriate statistical analysis would be RM ANCOVA to detect differences between groups and the effect of treatment as a function of time. Because of the high mortality rate observed in the PD group, data points were imputed based on an animal-specific regression analysis that extrapolated values from the time of death until T48. Animals in the PD group all developed MODS, with all parameters correlating and indicating profound physiologic derangement. The imputed values were based on very conservative assumptions and were more likely to underestimate the severity of disease in the nonsurviving animals group than to exaggerate any differences.
Furthermore, to assess significance at each time point, we subjected the results of the RM ANCOVA to a post-hoc test and corrected the P value to Bonferroni correction to adjust for the effect of multiple testing. In a study with such a small sample size, achieving significance is anticipated to be a challenge. We subjected our data to rigorous statistical scrutiny for this reason. We think that many more of the observed parameters would have met significance criteria had our sample size been larger. Although future, high-powered studies will be needed to elucidate the mechanisms of NPT protection, we feel strongly that these data support our hypothesis that NPT ameliorates the severity of organ dysfunction in this clinically applicable, large animal model.
In response to a procedural criticism, the cecum from which feces was harvested was not closed and the peritoneal cavity was not cleaned as would have been done in a patient with a traumatic bowel perforation. Although this deviates from clinical practice, this injury model causes a very consistent pathogenesis and, if anything, is more severe than what would be seen in human patients (32). In addition, an open perforation is similar to the widely accepted cecal ligation and puncture model. It is logical to assume that if a treatment regimen such as NPT is effective in this animal model, it should also be effective in patients in which the initiating source of infection is eliminated. We are confident that the data clearly demonstrate the protective effect of NPT in this animal model.
1. Jean-Baptiste E: Cellular mechanisms in sepsis. J Intensive Care Med
2. Protti A, Singer M: Bench-to-bedside review: potential strategies to protect or reverse mitochondrial dysfunction in sepsis-induced organ failure. Crit Care
3. Tang PS, Mura M, Seth R, Liu M: Acute lung injury and cell death: how many ways can cells die? Am J Physiol Lung Cell Mol Physiol
4. Garrison RN, Spain DA, Wilson MA, Keelen PA, Harris PD: Microvascular changes explain the "two-hit" theory of multiple organ failure. Ann Surg
5. Marshall JC: Inflammation, coagulopathy, and the pathogenesis of multiple organ dysfunction syndrome
. Crit Care Med
6. Czabanka M, Peter C, Martin E, Walther A: Microcirculatory endothelial dysfunction during endotoxemia-insights into pathophysiology, pathologic mechanisms and clinical relevance. Curr Vasc Pharmacol
7. Johnson D, Mayers I: Multiple organ dysfunction syndrome
: a narrative review. Can J Anaesth
8. Rezende-Neto JB, Moore EE, Masuno T, Moore PK, Johnson JL, Sheppard FR, Cunha-Melo JR, Silliman CC: The abdominal compartment syndrome as a second insult during systemic neutrophil priming provokes multiple organ injury. Shock
9. Suliburk J, Helmer K, Moore F, Mercer D: The gut in systemic inflammatory response syndrome and sepsis. Enzyme systems fighting multiple organ failure. Eur Surg Res
10. Imai Y, Parodo J, Kajikawa O, de Perrot M, Fischer S, Edwards V, Cutz E, Liu M, Keshavjee S, Martin TR, et al.: Injurious mechanical ventilation and end-organ epithelial cell apoptosis and organ dysfunction in an experimental model of acute respiratory distress syndrome. JAMA
11. Mayberry JC, Welker KJ, Goldman RK, Mullins RJ: Mechanism of acute ascites formation after trauma resuscitation. Arch Surg
12. Ince C: The microcirculation is the motor of sepsis. Crit Care
9(suppl 4):S13-S19, 2005.
13. van Haren FM, Sleigh JW, Pickkers P, Van der Hoeven JG: Gastrointestinal perfusion in septic shock. Anaesth Intensive Care
14. Matheson PJ, Wilson MA, Garrison RN: Regulation of intestinal blood flow. J Surg Res
15. Zakaria el R, Li N, Garrison RN: Mechanisms of direct peritoneal resuscitation-mediated splanchnic hyperperfusion following hemorrhagic shock. Shock
16. Farand P, Hamel M, Lauzier F, Plante GE, Lesur O: Review article: organ perfusion/permeability-related effects of norepinephrine and vasopressin in sepsis. Can J Anaesth
17. Fink MP, Delude RL: Epithelial barrier dysfunction: a unifying theme to explain the pathogenesis of multiple organ dysfunction at the cellular level. Crit Care Clin
18. Balogh Z, McKinley BA, Cox CS Jr, Allen SJ, Cocanour CS, Kozar RA, Moore EE, Miller IC, Weisbrodt NW, Moore FA: Abdominal compartment syndrome: the cause or effect of postinjury multiple organ failure. Shock
19. Deitch EA: Role of the gut lymphatic system in multiple organ failure. Curr Opin Crit Care
20. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P: Acute renal failure-definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care
21. Strassburg CP: Gastrointestinal disorders of the critically ill. Shock liver. Best Pract Res Clin
22. Brandt S, Regueira T, Bracht H, Porta F, Djafarzadeh S, Takala J, Gorrasi J, Borotto E, Krejci V, Hiltebrand LB, et al.: Effect of fluid resuscitation on mortality and organ function in experimental sepsis models. Crit Care
23. Venturi ML, Attinger CE, Mesbahi AN, Hess CL, Graw KS: Mechanisms and clinical applications of the vacuum-assisted closure (VAC) device: a review. Am J Clin Dermatol
24. Moore FA: The role of the gastrointestinal tract in postinjury multiple organ failure. Am J Surg
25. Ponferrada LP, Van Stone JC: Peritoneal dialysis kinetics. Adv Ren Replace Ther
26. Garrison RN, Zakaria el R: Peritoneal resuscitation. Am J Surg
27. Garrison RN, Conn AA, Harris PD, Zakaria el R: Direct peritoneal resuscitation as adjunct to conventional resuscitation from hemorrhagic shock: a better outcome. Surgery
28. Deitch EA, Xu D, Kaise VL: Role of the gut in the development of injury- and shock induced SIRS and MODS: the gut-lymph hypothesis, a review. Front Biosci
29. Magnotti LJ, Xu DZ, Lu Q, Deitch EA: Gut-derived mesenteric lymph: a link between burn and lung injury. Arch Surg
134:1333-1340; discussion 1340-1331, 1999.
30. Vidal MG, Ruiz Weisser J, Gonzalez F, Toro MA, Loudet C, Balasini C, Canales H, Reina R, Estenssoro E: Incidence and clinical effects of intra-abdominal hypertension in critically ill patients. Crit Care Med
31. Malbrain ML, De laet I: It's all in the gut: introducing the concept of acute bowel injury and acute intestinal distress syndrome. Crit Care Med
32. Steinberg J, Halter J, Schiller H, Gatto L, Carney D, Lee HM, Golub L, Nieman G: Chemically modified tetracycline prevents the development of septic shock and acute respiratory distress syndrome in a clinically applicable porcine model. Shock