"Surgery, Department of"@en . "Medicine, Faculty of"@en . "DSpace"@en . "UBCV"@en . "Gow, Kenneth William"@en . "2009-03-25T19:45:44Z"@en . "1997"@en . "Master of Science - MSc"@en . "University of British Columbia"@en . "Many investigators advocate aggressive fluid therapy in sepsis, yet changes in\r\nthe microcirculation may make fluid counterproductive. Sepsis is characterized by a\r\ngeneralized \"leak\" in capillaries which may promote interstitial edema which in\r\nturn, may decrease diffusion of oxygen, increase the distance from capillaries to cells,\r\nand alter capillary density. Further, fluid administration' may result in capillary\r\nhemodilution. Therefore, the author's hypothesis was that crystalloid resuscitation\r\nwill impair the ability of tissue to extract oxygen.\r\nFour groups (n=8) of anesthetized pigs received either normal saline infusion\r\n(48 ml-kg-l-hr-1) or no saline, and E coli endotoxin (50 mg/kg i.v.) or no endotoxin.\r\nWhole body and gut oxygen delivery and consumption were measured during\r\nprogressive hemorrhage. Dual line regression analysis was used to determine the\r\nonset of ischemia (DO2C) and oxygen extraction ratio (ERc). At onset of ischemia,\r\ngut was removed to determine degree of interstitial volume and the capillary\r\nhematocrit. With use of radiolabelled microspheres as a marker of blood flow, the\r\ngut blood flow transit time was determined. Endotoxin significantly decreased ERc\r\nfor the whole body (0.82+0.06 to 0.55\u00B10.08, p < 0.05) and gut (0.77 \u00B1 0.07 to 0.52 \u00B1 0.06,\r\np < 0.05). Saline resuscitation also significantly decreased ERc in the control pigs for\r\nthe whole body (0.82 \u00B1 0.06 to 0.62 \u00B1 0.08, p < 0.05) and gut (0.77 \u00B1 0.07 to 0.67 \u00B1 0.06, p\r\n< 0.05) but did not significantly change the already decreased ERc in the endotoxin\r\ntreated pigs. Morphometric techniques revealed that saline resuscitation increased\r\ngut interstitial volume (p < 0.05), and lead to arterial hemodilution (p < 0.05) but not\r\ncapillary hemodilution (p > 0.05). Using radiolabeled microspheres, saline was\r\nshown to increase the relative dispersion of blood flow transit times from 0.33 \u00B1 0.08\r\nto 0.72 \u00B1 0.53 (p < 0.05). Thus, saline resuscitation impairs tissue oxygen extraction\r\npossibly due to interstitial edema or increased heterogeneity of microvascular blood\r\nflow. After endotoxin infusion, where ERc is already decreased, saline resuscitation\r\nhas a lesser effect. Therefore, the author questions the use of aggressive crystalloid\r\nresuscitation for treatment of sepsis in humans."@en . "https://circle.library.ubc.ca/rest/handle/2429/6495?expand=metadata"@en . "9654323 bytes"@en . "application/pdf"@en . "THE EFFECTS OF CRYSTALLOID RESUSCITATION O N OXYGEN EXTRACTION IN WHOLE BODY A N D GUT DURING ENDOTOXEMIA by KENNETH WILLIAM GOW B.Sc, The University of Manitoba, 1991 B.Sc.(Med), The University of Manitoba, 1991 M.D., The University of Manitoba, 1991 FRCS(C), The University of British Columbia, 1997 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE , .(SURGERY ) i n THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF SURGERY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October 1997 \u00C2\u00A9 Kenneth William Gow, 1997 In p resent ing this thesis in partial fulf i lment of the requ i rements for an a d v a n c e d d e g r e e at, the University of British C o l u m b i a , I agree that the Library shall make it freely available for re ference and study. . I further agree that pe rmiss ion for extensive c o p y i n g of this thesis for scholar ly p u r p o s e s may b e granted by the h e a d of m y depar tment or by his o r her representat ives. It is u n d e r s t p b d t h a t c o p y i n g o r pub l ica t ion , of this thesis for f inancial gain shall not b e a l l o w e d wi thout m y writ ten p e r m i s s i o n . D e p a r t m e n t of S u C & g R Y T h e Universi ty of British C o l u m b i a Vancouver; C a n a d a Date QGTQfeEf* \ Z \ 9 9\"7 \u00E2\u0080\u00A2 D E - 6 (2/88) ABSTRACT Many investigators advocate aggressive fluid therapy in sepsis, yet changes in the microcirculation may make fluid counterproductive. Sepsis is characterized by a generalized \"leak\" in capillaries which may promote interstitial edema which in turn, may decrease diffusion of oxygen, increase the distance from capillaries to cells, and alter capillary density. Further, fluid administration' may result in capillary hemodilution. Therefore, the author's hypothesis was that crystalloid resuscitation will impair the ability of tissue to extract oxygen. Four groups (n=8) of anesthetized pigs received either normal saline infusion (48 ml-kg-l-hr\"1) or no saline, and E coli endotoxin (50 mg/kg i.v.) or no endotoxin. Whole body and gut oxygen delivery and consumption were measured during progressive hemorrhage. Dual line regression analysis was used to determine the onset of ischemia (DO2C) and oxygen extraction ratio (ERc). At onset of ischemia, gut was removed to determine degree of interstitial volume and the capillary hematocrit. With use of radiolabelled microspheres as a marker of blood flow, the gut blood flow transit time was determined. Endotoxin significantly decreased ERc for the whole body (0.82+0.06 to 0.55\u00C2\u00B10.08, p < 0.05) and gut (0.77 \u00C2\u00B1 0.07 to 0.52 \u00C2\u00B1 0.06, p < 0.05). Saline resuscitation also significantly decreased ERc in the control pigs for the whole body (0.82 \u00C2\u00B1 0.06 to 0.62 \u00C2\u00B1 0.08, p < 0.05) and gut (0.77 \u00C2\u00B1 0.07 to 0.67 \u00C2\u00B1 0.06, p < 0.05) but did not significantly change the already decreased ERc in the endotoxin treated pigs. Morphometric techniques revealed that saline resuscitation increased gut interstitial volume (p < 0.05), and lead to arterial hemodilution (p < 0.05) but not ii capillary hemodilution (p > 0.05). Using radiolabeled microspheres, saline was shown to increase the relative dispersion of blood flow transit times from 0.33 \u00C2\u00B1 0.08 to 0.72 \u00C2\u00B1 0.53 (p < 0.05). Thus, saline resuscitation impairs tissue oxygen extraction possibly due to interstitial edema or increased heterogeneity of microvascular blood flow. After endotoxin infusion, where ERc is already decreased, saline resuscitation has a lesser effect. Therefore, the author questions the use of aggressive crystalloid resuscitation for treatment of sepsis in humans. iii TABLE OF CONTENTS Abstract ii Table of Contents iv List of Tables vii List of Figures viii Acknowledgments x Chapter 1. INTRODUCTION 1 1.1 OVERVIEW 1 1.2 SEPSIS A N D SYSTEMIC INFLAMMATORY RESPONSE SYNDROME (SIRS) 3 1.2.1 Definitions 3 1.3 MEDIATORS OF SEPTIC TOXICITY 6 1.3.1 Mediation of septic toxicity ' 6 1.3.2 Bacterial endotoxin 7 1.3.3 Host defenses 10 1.3.4 Inflammatory Mediators 13 1.3.4.1 Peptide cytokines 13 1.3.4.2 Platelet activating factor 15 1.3.4.3 Eicosanoids 15 1.3.4.4 Nitric oxide 15 1.3.4.5 Oxygen radicals 16 1.3.5 Ischemia and Tissue Injury 17 1.4 SEPSIS AND TISSUE ISCHEMIA 18 1.4.1 Central hemodynamic changes in sepsis 18 1.4.2 Microvascular changes in sepsis 19 1.4.2.1 Alterations in erythrocyte deformability 19 1.4.2.2 Viscosity alterations 21 1.4.2.3 Increased vascular permeability and interstitial edema 21 iv 1.4.2.4 Alterations in microvascular flow and capillary heterogeneity 22 1.5 SPLANCHNIC ISCHEMIA IN CRITICAL ILLNESS 26 1.5.1 Protective mucosal barrier 26 1.5.2 Splanchnic Metabolic Model of O 2 control 26 1.5.3 Counter-current shunting (CCS) 28 1.5.4 Gut Ischemia during Septic Shock 30 1.6 AEROBIC A N D ANAEROBIC METABOLISM 31 1.6.1 Aerobic metabolism 31 1.6.2 Anaerobic metabolism 34 1.6.3 Oxygen Consumption and Delivery 34 1.6.4 Oxygen Extraction Ratio 37 1.6.5 Oxygen Consumption and Delivery in Sepsis 37 1.6.6 Oxygen Extraction in Sepsis 37 1.7 TREATMENT OF SEPSIS 41 1.7.1 Treatment Options 41 1.7.2 Augmentation of Oxygen Delivery 43 1.7.3 Fluid resuscitation in septic shock - crystalloids and colloids 44 1.7.4 Diffusion vs. Filtration 46 1.7.5 Starling's Law of the Capillary 47 1.7.6 Tonometry ' 49 1.7.7 Lactate 52 Chapter 2. HYPOTHESIS 54 Chapter 3. OBJECTIVES OF THE THESIS 55 Chapter 4. RESEARCH PLAN 57 4.1 THE PORCINE MODEL OF SEPTIC SHOCK 57 v 4.2 MORPHOMETRIC ANALYSIS OF GUT TISSUE 59 4.3 INFRA-SCAN ANALYSIS OF GUT TISSUE 59 4.4 DETERMINATION OF GUT CAPILLARY HEMATOCRIT 61 4.5 RELATIVE DISPERSION OF BLOOD TRANSIT TIMES 63 4.6 TONOMETRY 64 Chapter 5. EFFECT OF CRYSTALLOID RESUSCITATION O N WHOLE BODY A N D GUT IN ENDOTOXEMIA 65 5.1 INTRODUCTION 65 5.2 METHODS 68 5.2.1 Experimental Design and Protocol 68 5.2.2 Surgery Preparation and Instrumentation 69 5.2.3 Hemodynamic Measurements and Calculations 73 5.2.4 Transit Times 74 5.2.5 Interstitial Volume 75 5.2.6 Capillary Hematocrit 77 5.2.7 Ileal Tonometry 77 5.2.8 Statistical Analysis 78 5.3 RESULTS 80 5.3.1 Effect of endotoxin 80 5.3.2 Effect of Fluid Resuscitation 81 5.4 DISCUSSION 99 Chapter 6. CONCLUSIONS 106 Chapter 7. SUMMARY: CLINICAL RELEVANCE, FUTURE DIRECTIONS 108 REFERENCES 115 vi LIST OF TABLES Table 1. Onset of Ischemia and Extraction Ratio 90 Table 2. Arterial and Capillary Hematocrit 91 Table 3. Arterial and Portal Vein Leukocyte Counts 92 Table 4. Intestinal p H i at Baseline and Onset of Ischemia 93 Table 5. Arterial Lactate Values 94 Table 6. Onset of Rise of Arterial Lactate 95 Table 7. Oxygen Transport Variables 96 Table 8. Hemodynamic Variables 97 Table 9. Cl in ica l Variables 98 v i i LIST OF FIGURES Figure 1. Sepsis and SIRS 4 Figure 2. Mediators of Septic Toxicity 8 Figure 3. Endotoxin (Lipopolysaccharide (LPS)) 9 Figure 4. Host defenses 12 Figure 5. Toxic Mediators 14 Figure 6. Toxicity 20 Figure 7. Capillary heterogeneity 24 Figure 8. Metabolic model of O 2 control in the gut 27 Figure 9. Countercurrent Shunting (CCS) 29 Figure 10. Gut as \"Motor of Multiple Organ Dysfunction\" 32 Figure 11. Aerobic Metabolism 33 Figure 12. Anaerobic Metabolism 35 Figure 13. Relationship between oxygen delivery and consumption 36 Figure 14. Oxygen Extraction Ratio 38 Figure 15 Relationship between oxygen delivery and consumption in sepsis 39 Figure 16. Oxygen Extraction Ratio during sepsis 40 Figure 17. Treatment options 42 Figure 18. Starling's Forces of the Capillary 48 Figure 19. Tonometric Analysis 51 -Figure 20. Morphometric Analysis 60 Figure 21. Infra-scan Analysis 62 viii Figure 22. Protocol Timeline 70 Figure 23. Surgery and Instrumentation 72 Figure 24. Processing jejunum for morphometric analysis 76 Figure 25. An example of C\u00C2\u00BB2 Consumption and Delivery Curve 79 Figure 26. Whole Body Ability to Extract Oxygen 84 Figure 27. Gut Ability to Extract Oxygen 85 Figure 28. Whole Body Onset of Ischemia 86 Figure 29. Gut Onset of Ischemia 87 Figure 30. Relative dispersion of Gut Blood Transit Times 88 Figure 31. Interstitial Volume in Gut 89 ix ACKNOWLEDGMENTS I dedicate this thesis to my mother and father, May and Stephen Gow, who have always helped and encouraged me throughout my years of study. To my brother; James Gow whose determination has always been a standard against which I set myself. I would like to thank my supervisors, Drs. Terry Phang and Keith Walley, for their constant guidance and support. It is because of their example that I have learnt so much. I also thank the other members of my Supervisory Committee, Drs. John MacFarlane, York Hsiang, and Mike Allard, who have provided much encouragement and constructive criticism as my experiments and thesis developed. I am grateful for the thoughts and technical assistance of Drs. Blair Rudston-Brown, Oliver Bathe, Robert Granger, Suzy Tebbutt-Speirs, Christopher Goddard, Diane Minshall, Lynn Carter, Dean English, Lubos Bohunek, Betty Poon, and Peter Whitehead without whom this project could not have been completed. Chapter 1 I N T R O D U C T I O N 1.1 O V E R V I E W Sepsis is an all too frequent occurrence in critically ill patients . Overall, the cost to health care is staggering (1) with an estimated cost of $5 to $10 billion in the US annually. Sepsis will likely be a problem for clinicians not only now but for the foreseeable future (2). The increasing incidence of septic shock continues to be associated with a mortality rates ranging from 40% to 60% despite advances in critical care medicine. Clinicians have long sought the optimal treatment for patients suffering from sepsis. However, sepsis, by its very nature is a systemic illness whereby many of the toxic effects are the result of the patient's own immune system as it mounts a defense to the infection. The immune system reaction to infection focuses on various aspects of the microorganism but the lipopolysaccharide moiety known as endotoxin has been best described in literature. The immune system through its various leukocytes mounts a defense to the organisms. This defense is in turn mediated by numerous cytokines which have beneficial effects as well as side effects. While the immune system overall is beneficial in its role, if the effects are allowed to go unhindered, the immune system mediators will themselves lead to significant illness. Various toxicity effects occur from these cytokines such that further damage may ensue. Ultimately, the damage done to cells occurs at a microvascular level with ischemia as the final end-point. 1 While sepsis is a systemic insult, the splanchnic circulation is considered to be particularly affected. There are various reasons why the gut is predisposed to injury. Two explanations are the high metabolic demands which make the gut susceptible to minor deficits and the countercurrent organization of the vessels which leads to possible shunting from arterioles to venules in times on decreased flow. Whi le ischemia in any tissue is potentially harmful, ischemia to the gut is particularly concerning. With the high volume of bacteria that reside in the gut, damage to the mucosal protective layer may result in further bacterial load to the host. Since this translocation of bacteria may ultimately lead to systemic toxicity, the gut has been termed the \"Motor of Multiple Organ Failure\" (3). The toxicity that leads to multiple organ dysfunction is ultimately due to tissue ischemia. Ischemia may be defined by the decrease in blood flow. In turn, ischemia results in hypoxia which may be described as the lack of oxygen for aerobic metabolism. This results in anaerobic metabolism which is a less efficient mechanism for energy production and leads to metabolic byproducts. Techniques have been described which allow investigators to determine the onset of tissue ischemia and the tissue's ability to extract oxygen. These studies allow one to determine the potential benefit of various treatments of sepsis. Treatment of sepsis may be aimed at any step of the pathway from initial insult to the final endpoint of tissue ischemia. While new techniques such as blocking endotoxin or modulating the immune system are being studied (4), time honored therapeutic interventions such as antibiotics and fluid resuscitation are considered the mainstay of therapy (2). While the use of antibiotics is intuitive to 2 treat the infection, the use of fluid therapy is less than straight forward. Fluid therapy is thought to mediate its benefit by augmentation of filling pressures, replacement of intravascular deficits, decreased viscosity of blood, and a more uniform distribution of blood flow to tissues which is known as capillary blood flow homogeneity. However, since the septic state may be described as a generalized leak in the capillaries, the potential benefit of crystalloid resuscitation is questionable (5). The potential side effects from fluid resuscitation include interstitial edema, hemodilution, and a less uniform distribution of blood flow to tissues which is known as capillary blood flow heterogeneity. Therefore, while some consider aggressive crystalloid resuscitation as essential in the resuscitation of septic patients, physiologic changes that accompany sepsis may not only lead to a limited benefit but could make fluid resuscitation detrimental. In this thesis, the effect of crystalloid resuscitation is studied on the main outcome variable, oxygen extraction in a porcine model of endotoxemia. The literature pertaining to these topics has been reviewed and is discussed in detail below. 1.2 SEPSIS AND SYSTEMIC INFLAMMATORY RESPONSE SYNDROME (SIRS) 1.2.1 Definitions The following definitions were established by the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference (ACCP/SCCM) in 1991 (6) (Figure 1). 3 Figure 1. Sepsis and SIRS. This pictogram illustrates the differences between the terms described by the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference (ACCP/SCCM). 4 Systemic inflammatory response syndrome (SIRS) - Describes a continuous process, and describes an abnormal host response that is characterized by a generalized activation of the inflammatory reaction in organs remote from the initial insult. When the process is due to infection, the terms sepsis and SIRS are synonymous. SIRS can be seen following a wide variety of insults and includes, but is not limited to, more than one of the following clinical manifestations: (1) a body temperature > 38\u00C2\u00B0c or < 36\u00C2\u00B0c; (2) a HR >90 bpm; (3) tachypnea, with a RR > 20 bpm, or hyperventilation, as indicated by a PaC02 < 32 mmHg; and (4) an alteration in the WBC, being either > 12,000/mm3/ or <4,000/mm3, or the presence of more than 10% immature neutrophils. Infection - is a microbial phenomenon characterized by the presence of microorganisms or their invasion of normally sterile host tissue by those organisms. Bacteremia - presence of viable bacteria in the blood. The presence of viruses, fungi, parasites, and other pathogens in the blood should be described in a similar manner. Septicemia - has been defined in the past as the presence of microorganisms or their toxins in the blood. However, this term has been used clinically and in the medical literature in a number of ways leading to confusion and difficulty in interpreting data. Therefore, it was recommended to abandon this term. Sepsis - the systemic inflammatory response to infection. In association with infection, manifestations of sepsis are the same as those previously defined for SIRS. 5 Severe sepsis - sepsis associated with organ dysfunction, hypoperfusion abnormality, or sepsis-induced hypotension. Hypoperfusion abnormalities include lactic acidosis, oliguria, and acute alteration of mental status. Sepsis-induced hypotension is defined by the presence of a systolic blood pressure of less than 90 mmHg or its reduction by 40 mmHg or more from baseline in the absence of other causes for hypotension. Septic shock - a subset of severe sepsis and is defined as sepsis-induced hypotension, persisting despite adequate fluid resuscitation, along with the presence of hypoperfusion abnormalities or organ dysfunction. Multiple organ dysfunction syndrome (MODS) - the previous term multiple organ failure has been replaced with this term. The main change is the use of \"dysfunction\" used in place of the term \"failure\" to represent a continuum of physiologic derangements. This can evolve in the absence of an untreated focus of invasive infection and can be reproduced experimentally by the infusion of a diverse spectrum of endogenously derived mediators of inflammation. The term syndrome refers to a pattern of multiple and progressive symptoms and signs that are thought to be related. 1.3 MEDIATORS OF SEPTIC TOXICITY 1.3.1 Mediation of septic toxicity The body's immune system is paramount in importance in defending against infections. There are various mechanisms at work which comprise a fine balance of stimulatory and inhibitory effects. While for the most part, success results in the 6 host surviving, the mediators of defense may have toxic effects themselves. The various mechanisms important for immune defense have been shown to affect host tissues as well. The end result of these toxic effects is ischemia at the microvascular level (7) (Figure 2). 1.3.2 Bacterial Endotoxin While various aspects of bacteria are potentially toxic, the best described has been that of endotoxin. Gram-negative bacteria contain within their cell wall a macromolecular glycolipid termed lipopolysaccharide (LPS) (Figure 3). It comprises two components: the O-specific chain and the core. The O-specific chain is a polymer of oligosaccharides and accounts for the antigenic variability among species and strains of bacteria. The core is composed of an oligosaccharide covalently bound to a molecule lipid called lipid A. Most of the toxicity of LPS resides in the lipid A moiety. The structure of the core region is relatively constant across species and strains of gram-negative bacteria. The terms LPS and endotoxin are not strictly synonymous as LPS refers to the purified glycolipid whereas endotoxins contain small amounts of cell wall proteins, lipids, lipoproteins, and polysaccharides in addition to LPS (8). There are several proposed mechanisms by which endotoxin effects damage. Endotoxin has been shown to activate complement, the fibrinolytic pathway, neutrophils, and other mediator systems (9) (10). Furthermore, endotoxin has been shown to stimulate macrophages to release cytokines, including tumor necrosis factor (TNF) and interleukin-1 (IL-1), which, in turn, amplify the systemic response 7 F i g u r e 2. M e d i a t o r s o f Sep t i c T o x i c i t y . This figure illustrates the overall mediation of the toxicity involved with sepsis. It starts at the interaction between bacteria and the host-defense mechanisms. This interaction is meditated by various mediators leading to various toxicity's ultimately leading to ischemia. 8 Figure 3. Endotoxin. This figure illustrates endotoxin (lipopolysaccharide) as part of the cell wall of a gram negative bacteria. Endotoxin is comprised of two components: the O-specific chain and the core. The O-specific chain is a polymer of oligosaccharides and accounts for the antigenic variability among species and strains of bacteria. The core is composed of an oligosaccharide covalently bound to a molecule lipid called lipid A. Most of the toxicity of LPS resides in the lipid A moiety. 9 to endotoxin by stimulating neutrophils, endothelial cell, platelets, and the release of other cytokines, eicosanoids, platelet activating factor (PAF), endorphins, endothelial relaxing factor, and a variety of other plasma and cellular mediators (11) (10) (12). Some have speculated that endotoxin leads to mucosal hypoxia as a mechanism of damage. However, VanderMeer et al (13) showed that there is significant mucosal acidosis despite absence of mucosal hypoxia. Other mechanisms by which LPS may cause mucosal acidosis include uncoupling of oxidative phosphorylation (14) (15), inhibiting mitochondrial respiration (16) (17) (18) (17), decreasing availability of substrates (19), increasing metabolism, or decreasing the clearance of H + or C02-1.3.3 Host Defenses Baker and Huynh (7) divides host defenses into nonspecific and specific factors. Nonspecific factors include skin, mucus membranes, as well as complement, hemolytic factors, and phagocytosis. Phagocytosis involves interactions with a number of other systems, particularly the complement system. The specific immune system is responsible for recognizing bacteria and creating antibodies to help opsonize them. The specific immune system is a delicately balanced system with cells that have both stimulatory and inhibitory (or regulatory) function. While, for the most part, the immune system is important in preventing infection, the host's response to injury is often worse than the injury itself and can develop into a malignant, uncontrolled SIRS (7). This occurrence represents a state 10 of disseminated activation of the host's inflammatory systems. While any of the cells of immunity may be involved (Figure 4), it is thought that the macrophage is central to the development of this process (20). Macrophages play a crucial role in regulating the immune response after injury (21). They reside at strategic positions throughout the body, armed to eliminate foreign and infective agents. It is thought that it is the excessive, unregulated, prolonged stimulation of the macrophage in conjunction with other leukocytes and the endothelial system that leads to a vicious cascade of inflammatory mediator release (3). The macrophage is capable of producing directly toxic substances in addition to numerous cytokines and mediators that act synergistically to augment the inflammatory response (7). The various substances including TNF, interleukins, PAF, arachidonic acid metabolites, and nitric oxide are discussed below. Further, a number of investigators have studied the role of neutrophils in the development of SIRS and MODS. Much like the macrophage, it is thought that unregulated activation of neutrophils and the endothelial system are important factors (22) (23). The neutrophils are thought to adhere to the endothelium, creating a microenvironment in which neutrophil-derived proteases and toxic-reactive oxygen species are present at high concentrations. These toxic products can directly injury the endothelial l ining, leading to intercellular gap formation, altered vascular permeability, and further neutrophil migration (24). Together, the interaction among macrophages, neutrophils, and the endothelial system augments the inflammatory cascade and results in bystander organ failure (7). 11 Macrophages B-cells Neutrophils T-cells Figure 4. Host Defenses. This figure illustrates the different leukocytes thought to be important in immunity. 12 1.3.4 Inflammatory Mediators There are a multitude of toxic mediators thought to have importance in mediating the damages seen in SIRS and sepsis including TNF, IL-1, PAF, prostaglandins, thromboxanes, leukotrienes, nitric oxide, O 2 radicals, interferon, and complement (Figure 5). 1.3.4.1 Peptide cytokines Some degree of inflammatory response is due to the biological effects and balance of peptide cytokines (22). Both tumor necrosis factor (TNF) and interleukin-1 (IL-1) are important mediators in the pathogenesis of SIRS and MODS. These peptides are synthesized in nearly all organs containing phagocytes and blood mononuclear cells in response to a large number of stimuli including endotoxin, complement, eicosanoids, interferon, other interleukins, and viral antigens. The mechanism of action is thought to be the induction of subsequent effector molecules of SIRS. Infusion of either TNF or IL-1 wil l lead to lung injury characterized by increased permeability and neutrophil sequestration (22). TNF also causes neutrophil degranulation and superoxide production (22). It appears that TNF and IL-1 act synergistically in inducing shock (25), prostaglandin synthesis, and neutrophil chemotaxis (22). Inhibition of either TNF or IL-1 independently has been shown to prevent lung injury, shock, and death following endotoxin or E. coli administration (22). 13 Interleukins ^ \u00C2\u00B0 i radicals ^ ^ O Q Complement Interferon 0\u00C2\u00ABQ^ O * \u00C2\u00A9 \u00C2\u00AE \u00C2\u00A9 Leukotrienes T N F ^ A \u00C2\u00A9 O O O Q \u00C2\u00B0 0 P% r% Thromboxanes Nitric oxide \u00C2\u00B0 PAF Prostaglandins F i g u r e 5. I n f l a m m a t o r y M e d i a t o r s . This figure illustrates the various mediators thought to be important in the development of toxicity in sepsis. 14 1.3.4.2 Platelet activating factor Platelet activating factor (PAF) is a labile ether lipid first seen to be released by platelets in the presence of antigen and leukocytes. Cipolle et al (22) points to several reasons why PAF is considered an important mediator in SIRS. Infusion of PAF mimics many of the effects of sepsis and endotoxin. Further, PAF is produced during endotoxemia in animals and severe sepsis. Finally, several structurally different antagonists to PAF have been reported to inhibit endotoxin-induced hypotension, lung injury, and mortality. 1.3.4.3 Eicosanoids Eicosanoids are derivatives of the substrate arachidonic acid (AA), a cell membrane phospholipid. They include prostaglandins (PG), thromboxane (Tx), and leukotrienes (LT). A variety of inflammatory stimuli can directly or indirectly activate phospholipase leading to release of A A from phospholipid pools. Once released, A A is oxidatively metabolized via two major pathways leading to production of the eicosanoids. These derivatives all have varying functions which are beyond the scope of this thesis. 1.3.4.4 Nitric oxide Nitric oxide (NO) is formed via arginine metabolism and vascular endothelium appears to be the most important source (22). Endotoxin has been shown to stimulate N O formation in vitro (26). The vasodilation and vascular unresponsiveness that occur during sepsis and endotoxemia appear to be mediated 15 largely via the L-arginine-NO pathway (27). Inhibition of the L-arginine-NO pathway blocks TNF-mediated hypotension in rats (28). While N O may have negative effects, some argue that N O may in fact serve an important role. N O may be protective by inhibiting platelet aggregation and vasodilation, thereby preserving blood flow to important vascular beds to maintain critical oxygen delivery (22). Therefore, inhibiting N O may be detrimental during sepsis in that blood flow to already ischemic areas could be further reduced. 1.3.4.5 Oxygen radicals Under normal physiologic conditions, 80%-90% of cellular oxygen is consumed at the level of the mitochondrial respiratory chain and is directly reduced to water via the cytochrome oxidase complex (22). In pathologic conditions, fundamental oxygen (3C>2) can be reduced one electron at a time, giving rise to a cascade of activated oxygen species. The most important species is a free radical, the superoxide anion ( O 2 ) . Active oxygen metabolites may react to a variety of biologic substrates. They promote D N A scission and base modification, inactivate plasma proteins, cross-link membrane proteins, and most importantly, induce lipid peroxidation in the membrane which may lead to disorganization of its structure and lead to cell death (29). Oxygen free radicals are especially important during ischemia-reperfusion type injuries (22). During the reperfusion phase, severe tissue injury occurs following massive production of oxygen free radicals generated by several mechanisms. 16 1.3.5 Ischemia and Tissue Injury The effects described above lead to an alteration in the microvascular flow of. blood to tissue the mechanisms of which are described later. Therefore, the inflammatory mediators ultimately results in ischemia as the final endpoint of SIRS and sepsis. Ischemia not only precipitates cellular ATP degradation and impairs energy-requiring cellular function, but it may also activate processes that lead to an increase in oxygen-derived, free-radical production, which can augment or accelerate tissue injury (30). Ischemia also promotes the formation of xanthine oxidase from xanthine dehydrogenase. Furthermore, there is activation of cellular proteases which converts the xanthine dehydrogenase by proteolysis to xanthine oxidase which uses molecular oxygen as its electron acceptor leading to the generation of superoxide anion and/or hydrogen peroxide (31). Circulating levels of xanthine oxidase have been demonstrated in patients with ARDS (32). It has been hypothesized that l iver or gut tissue releases xanthine oxidase, thereby allowing it to serve as a toxic, oxygen-metabolite-generating system in the lung capillary bed (33). Furthermore, inflammation may be part of the ischemic process or may serve to augment and accelerate the ischemic process. Toxic oxygen metabolites, formed by xanthine oxidase during ischemia, may increase capillary permeability and increase the influx of neutrophils. The arrival of neutrophils and their subsequent activation by factors released from injured cells are likely to augment tissue injury. Local inflammation may also accelerate the ischemic process by causing granulocyte plugging and tissue edema, leading to a further impairment of oxygen delivery (31). 17 1.4 SEPSIS A N D TISSUE ISCHEMIA The hemodynamic changes that are described in sepsis may be divided into central and microvascular vascular changes (34). 1.4.1 Central hemodynamic changes in sepsis With regards to the central effects in sepsis, there is an initial stage during which homeostasis is maintained involving various changes including both tachycardia, tachypnea, and fever (12). Ultimately, homeostatic mechanisms fail and signs of circulatory shock ensue and is described as having both a hyperdynamic and a preterminal hypodynamic phase. The hyperdynamic phase of septic shock is typically seen and consists of a high cardiac output, low pulmonary artery occlusion pressure (PAOP), and a low peripheral vascular resistance (12). The hypodynamic stage is manifested by a low cardiac output and an increased systemic vascular resistance. Despite the increased cardiac output, many investigators have described a global myocardial dysfunction in septic shock (35) (36) (37). The increased cardiac output reflects the increase in heart rate while the stroke volume is usually low (35) (38). Plotted on a Frank-Starling curve against the PAOP, the left ventricular stroke work index was typically reflective of left ventricular failure (12) (39) (40). Furthermore, the magnitude of the myocardial depression correlates with survival as patients capable of maintaining their cardiac output by increasing their heart rate and left ventricular end-diastolic volume are more likely to survive (12). There have been several substances that have been implicated in contributing to this 18 myocardial depression (41) including TNF, low-molecular-weight water-soluble molecules, and lipid-soluble substances (42) (43) (44). The presence of these depressant factors has been shown to correlate with the severity of cardiac dysfunction and systemic hypoperfusion (45). 1.4.2 Microvascular changes in sepsis While initial investigations focused on measurable hemodynamic changes, more recent studies have emphasized the role of the microcirculatory changes in sepsis (46) (47). The changes described include: alterations in erythrocyte deformability, viscosity alterations, increased vascular permeability with development of interstitial edema, and changes in microvascular flow with development of capillary heterogeneity. A l l these changes wi l l go to alter the overall microcirculatory flow of blood and thereby alter the delivery of oxygen from capillaries to cells (Figure 6). 1.4.2.1 Alterations in erythrocyte deformability The normal state of deformability is the ability of the red cell to alter its biconcave, discoid shape to allow passage through a capillary smaller in diameter than itself (46). Driessen et al (48) emphasized the importance of normal erythrocyte deformability for the maintenance of adequate perfusion of the microcirculation, especially when perfusion pressure is reduced. They further explained that red cell deformability depends on the viscoelastic properties of the cell membrane, the viscosity of the cytoplasm, and the surface area/volume ratio, all of which may 19 Sludging ARegulation Edema Microthrombi 1 Permeability Figure 6. Toxicity. This figure illustrates the various microcirculatory effects described in sepsis. This includes alteration in autoregulation in the vessels with the ellipses representing the vascular tone, leukocyte sludging and microthrombi formation which leads to vascular obstruction, and increased capillary permeability which leads to interstitial edema as depicted by the shaded area surrounding the cell. 20 change during shock. When erythrocyte deformability is decreased, the time required for red cell passage through capillaries is prolonged and the cells may themselves impede blood flow and may lead to an impaired ability to extract oxygen as has been demonstrated by Powell et al (49) in septic humans. Furthermore, the increased rigidity of red cells promotes arteriovenous shunting of blood, which further decreases microcirculatory flow during sepsis (50). 1.4.2.2 Viscosity alterations Blood viscosity is an important determinant of microcirculatory flow. According to Poiseuille's equation, blood flow varies directly with vessel radius to the fourth power and inversely with the length of blood vessels and blood viscosity. Chien (51) studied the rheological factors in the microcirculation during low-flow states and concluded that venular blood viscosity increases due to a decrease in shear stress. The result of increased venular viscosity would be increased post-capillary resistance leading to diminished blood flow and increased transcapillary leakage (51). 1.4.2.3 Increased vascular permeability and interstitial edema Sepsis has often been characterized as a diffuse increase in microvascular permeability (46) (52) (12). Solomon and Hinshaw (53) showed that endotoxin increases capillary permeability in the skin and muscle tissues independently from changes in hydrostatic and colloid pressures, thereby implying the mechanism to be an alteration in the cell membrane structure. According to Starling's Forces, an 21 increase vascular permeability wil l predispose to interstitial edema. Hersch et al (54) studied the histologic and ultrastructural changes in a rat model of hyperdynamic sepsis and showed that despite a preservation of hemodynamic variables, there was a widespread degree of lesions from sepsis including interstitial and intracellular edema. Interstitial edema may in turn lead to an impaired diffusion of oxygen from vessels to cells (55) (34). Heughan et al (56) showed that saline loading wil l lead to interstitial edema and this was associated with a decrease in tissue oxygenation. Not only can interstitial fluid itself limit the diffusion of oxygen but interstitial edema in sepsis is known to result in an increased distance from cells to capillaries (57). Knisely et al (58) have theorized that such an increased distance wil l significantly limit the diffusion of oxygen. Other possible changes promoting edema formation include protein leakage (59), separation of tight junctions between endothelial cells (60), dysfunctional rather than destructive changes of vascular endothelium (61), and release of vasoactive agents (62). Finally, tissue edema formation may compress capillaries and thereby limit oxygen delivery (34). 1.4.2.4 Alterations in microvascular flow and capillary heterogeneity The ability of tissue to alter blood flow to actively metabolizing tissue is of paramount importance in survival. The organism adjusts oxygen extraction in response to changes in oxygen delivery through a balance between vasoconstrictor tone among organ systems and local metabolic vasodilation within tissues (63). However, various components required for normal adaptation to oxygen demands are altered in sepsis including development of vascular obstruction, development of 22 arteriovenous shunts/and endothelial damage. All these effects lead to an inability to provide normal homogenous blood distribution and thereby leads to a deleterious heterogenous blood flow (Figure 7). Various components of blood may lead to obstruction of vascular flow. Knisely et al (64) studied patients and animals and found large rigid erythrocyte aggregates in all vessels, and concentrated \"sludge\" in plugged vessels which lead to obstructed blood flow. Clotting factors may also be activated leading to disseminated intravascular coagulation (DIC) during sepsis (65) whereby fibrin and microthrombosis are deposited in vessels. Finally leukocytes have been shown to cause capillary obstruction during sepsis (65) (41) (66) which not only contribute to vascular obstruction but also has effects on red cell deformability and capillary cross-sectional area (65). Various studies have shown the occurrence of shunts in various tissue beds in sepsis. Cronenwett and Lindenauer (67) used microspheres to demonstrate arteriovenous shunting of blood in the septic canine. Further, Archie (68) described arteriovenous shunting in a cecal ligation shock model occurring in the splanchnic and renal circulation. Whether these changes represent the effects on the changes of red cell deformability (see above) or represent true anatomical arteriovenous shunts is unclear. While previously, the endothelial layer was thought of as only a passive tissue, more recent work has described its importance in modulating vascular tone, controlling local blood flow, influencing the rates of leakage of fluids and plasma proteins, modulating the accumulation and extravasation of leukocytes into tissue, 23 Homogenous distribution Heterogeneous distribution F i g u r e 7. C a p i l l a r y H e t e r o g e n e i t y . Homogeneous distribution leads to adequate blood flow to the cells. Heterogeneous distribution leads to maldistribution of blood flow to the cells. 2 4 and influencing leukocyte activation (69). Therefore, damage to the endothelium and the underlying smooth muscle from hypoxia, inflammation, complement activation, O 2 radical production, and lipid peroxidation, all postulated in the sepsis syndrome, would interfere with the matching of O 2 delivery and metabolic needs by interfering with vascular smooth muscle tone and possible interactions between the endothelium and the smooth muscle (55) (70). Schumacker and Samsel theorize that alteration in endothelial function leads to the loss of autoregulation of the microvasculature which in turn results in the inability of tissue to adjust to reduced oxygen delivery (63) (71). Indeed, sepsis typically is described as having a decreased systemic vascular resistance (72) which reflects the decreased arteriolar and venular tone (36). Dantzker (55) explains that a loss of autoregulatory ability could explain several of the features seen in sepsis. The effects explained above, all point to changes that wil l affect the ability to adapt to increased oxygen demands. Delivery of oxygen to tissue is ideally done by both an adequate and a homogenous distribution of capillaries. However, Drazenovic et al (73) found endotoxemia in a canine model reduced capillary density in mucosal vil l i and crypts. Further, Lam et al (57) using intravital microscopy to analyze changes in muscle capillaries in a rat model of sepsis not only found a decreased density of perfused capillaries, but also an increase in capillary heterogeneity, an increase in the inter-capillary distance, an increase in capillary red blood cell velocity, and a decrease in the peak hyperemic response. Thus, while homogenous capillary flow promotes oxygen extraction, all the microcirculatory changes described wil l result in a heterogenous flow of blood 25 which may impair oxygen diffusion and extraction by tissue (74) (75) (76) (63) (77) (78) (79). 1.5 SPLANCHNIC ISCHEMIA IN CRITICAL ILLNESS 1.5.1 Protective mucosal barrier A n intact mucosal layer is important to act as a defense against infection. It has been shown that an intact mucosal layer maintains an effective barrier even during states of immunosuppression (80). This implies that the cell-mediated immuni ty provided by the intestinal wall 's lymphocytes, macrophages, and the Peyer's patches, and the mesenteric lymph nodes serves a secondary or supportive role to the epithelium. While the mucosa is an important barrier, the gut is susceptible to ischemia during sepsis or endotoxemia due to the manner that the blood flow is organized. 1.5.2 Splanchnic Metabolic Model of O2 Control Metabolic demand theory in the splanchnic circulation proposes that the intestines control local microvascular smooth muscle tone, based on the prevai l ing tissue PO2, and independent of nervous or humoral influences. (81). There appears to be two separate microvascular mechanisms by which the intestine can regulate the rate of oxygen delivery (82) (Figure 8). First, local arteriolar tone governs the amount of total blood flow that is available to the individual capillary beds. Cont ro l of the precapillary sphincters modulates the number of capillaries that are perfused within a capillary bed, thereby effecting changes in the surface area and the capillary-26 feedback Figure 8 . Metabolic model of O 2 control in the gut. This figure illustrates theoretical model of oxygen control at the microvascular level in the gut. Local arteriolar tone governs total blood flow. Small and moderate decreases in oxygen or build-up of metabolites leads to alteration in the tone of precapillary sphincters. 27 to-cell diffusion distance (81). Second, wi th small and moderate decreases in oxygen delivery, the main compensatory mechanism for maintaining a stable oxygen consumption are the precapillary sphincters (83). This mechanism causes an increase in capillary surface area and results in increased oxygen extraction. Only wi th larger decreases in oxygen delivery does the arteriolar resistance decrease, thereby increasing blood flow to the tissue (84). However, despite this fine control (55), there is a loss of autoregulatory ability in sepsis resulting in a reduced capillary density in mucosal v i l l i and crypts (73). In fact, the reduction in splanchnic blood volume is disproportionately greater than that seen in other tissue beds (85). 1.5.3 Counter-current shunting (CCS) The microcirculation of the intestinal vi l lus is arranged in a counter-current system of arterioles and venules which is designed to improve absorption. However, the countercurrent shunting has been hypothesized as another reason for the susceptibility of the mucosa to ischemia during sepsis (82) (86) (55) (87) (88). This theory proposes that since the vessels are arranged in a vil lus wi th both the arteriole and the venule side by side, O2 may diffuse from the arteriole straight across to the venule without the vil lus ever benefiting from the O2, thereby effectively shunt ing oxygen away from the vi l lus (Figure 9). Furthermore, it has been shown that this effect is more exaggerated during times of decreased blood flows (89). 28 Villus Arteriole Venule Figure 9. Counter-current shunting (CCS). This figure illustrates theoretical concept of counter-current shunting in the intestinal villus. It describes how oxygen may shunt across from the arterioles to the venules due to the vessels being arranged side by side. 29 1.5.4 Gut Ischemia during Septic Shock While the mucosal barrier serves a particularly important role in preventing infection, its own fine microvascular balance places it at risk in sepsis (90). Fink et al (52) showed that even when the hemodynamic variables were maintained in sepsis, there are significant levels of gut mucosal acidosis. Furthermore, the high metabolic demands of the intestinal surface places the gut at further risk during times of limited oxygen supply (91) (92) (93) (94). Since the gut is susceptible to the effects of sepsis, the protective mucosal barrier is thereby placed in jeopardy (91) (95) (96). Salzman et al (97) have shown that endotoxin infusion leads to increased intestinal permeability to macromolecular hydrophilic solutes. Similarly, O'Dwyer et al (98) injected endotoxin in healthy humans and found increased permeability to nonmetabolizable sugars. Bacterial translocation is the process by which microorganisms migrate across the mucosal barrier and invade the host. Rush et al in a series of studies (99) have shown that shock in animal models and humans leads to the development of bacteremia and endotoxemia. Further, they have shown that most of the bacteria found in the blood were enteric organisms and by using radiolabeled E. coli and demonstrating them escaping from the gastrointestinal tract and into the blood stream, conclude that bacterial translocation is a true phenomenon. Deitch et al (100) had similar conclusions in a mouse model and further determined that the mechanism may be via increased gut permeability via activation of xanthine oxidase leading to development of oxygen free radicals (101). Mainous et al (102) studied the route of bacteremia and determined that bacterial translocation is 30 primarily via portal blood as opposed to mesenteric lymphatics and occurs in a dose dependent fashion to the degree of insult. It has been theorized that injury to the gut wi th the resultant bacteremia and endotoxemia may perpetuate a continuous cycle of injury leading to the development of multiple organ dysfunction syndrome (85). Hence, the term, the gut as the \"motor of multiple organ dysfunction\" (3) (Figure 10). 1.6. AEROBIC A N D ANAEROBIC METABOLISM 1.6.1 Aerobic metabolism The routine means to derive energy by cells is by aerobic metabolism. Energy is derived from amino acids, glucose, and fatty acids via the Kreb's (Tricarboxylic Acid) Cycle and oxidative phosphorylation. The Kreb's cycle is a series of controlled oxidation-reduction reactions during which the energy, released from the transfer of electrons, is captured. This involves the reduction of N A D + to N A D H , a shutt l ing of the electrons from N A D H into the mitochondria, transferring the electrons along a series of electron-carrier enzymes, and the capture of energy in the high-energy phosphate bonds of adenosine triphosphate (ATP). Oxygen is the terminal electron acceptor in this scheme and sufficient amounts are required if optimal use of substrate for the generation of A T P is to continue. When oxygen is not present i n adequate amounts, the organism relies on the less efficient anaerobic metabolism which generates less energy per substrate when compared to aerobic metabolism (55) (Figure 11). 31 o2 ^ Bacteria Endotoxin Figure 10. Gut as \"Motor of Multiple Organ Dysfunction\". It has been theorized that injury to the gut with the resultant bacteremia and endotoxemia may perpetuate a continuous cycle of injury leading to the development of multiple organ dysfunction syndrome. Hence, the term, the gut as the \"motor of multiple organ dysfunction\" 32 Figure 11. Aerobic Metabolism. Energy is derived from amino acids, glucose, and fatty acids via the Tricarboxylic Acid Cycle (TCA) and oxidative phosphorylation. The TCA cycle is a series of controlled oxidation-reduction reactions during which the energy, released from the transfer of electrons, is captured. This involves the reduction of N A D + to NADH, a shuttling of the electrons from NADH into the mitochondria, transferring the electrons along a series of electron-carrier enzymes, and the capture of energy in the high-energy phosphate bonds of adenosine triphosphate (ATP). Oxygen is the terminal electron acceptor. 33 1.6.2 Anaerobic metabolism When oxygen is not available, ATP is generated by glycolysis where pyruvate is the terminal electron acceptor. While the aerobic pathway results in the generation of 36 mmol of ATP per mol of glucose, the anaerobic pathway results i n only 2 mmol of ATP per mol of glucose. Not only does this produce much less energy per mol of substrate, but the pyruvate is reduced to lactate which may then contribute to the development of metabolic acidosis (55) (Figure 12). 1.6.3 Oxygen Consumption and Delivery Investigators have developed techniques whereby the oxygen consumption and delivery variables can be quantified to allow determination of the onset of ischemia and ability to extract oxygen. The technique of decreasing oxygen delivery by progressive hemorrhage and following the effects on oxygen consumption have been described (103) (104) (105). The ability of tissue to utilize O 2 during shock can be determined by plotting the relationship between oxygen consumption ( V O 2 ) and oxygen delivery ( D O 2 ) . With progressive decrease in oxygen delivery, the tissue compensates by increasing the amount of oxygen extracted. If the tissue is able to extract enough oxygen then aerobic metabolism is maintained. However, there is a point when O 2 delivery falls below sufficient O 2 requirements after which anaerobic metabolism ensues. This critical point is considered the onset of ischemia ( D 0 2 c ) (Figure 13). 34 4 ADP 4 ATP Glucose Glycogen Glyceraldehyde 3 - P 1^ * * \u00E2\u0080\u0094 NAD NADH Lactate Pyruvate F i g u r e 12. A n a e r o b i c M e t a b o l i s m . Anaerobic metabolism of ATP from glycolysis. When oxygen is not available, ATP is generated by glycolysis where pyruvate is the terminal electron acceptor. Not only does this produce much less energy per mol of substrate, but the pyruvate may be reduced to lactate possibly contributing to the development of metabolic acidosis 3 5 O z Consumption (V02) Onset of ischemia (D02c) 0 2 Delivery (D02) Figure 13. Relationship between Oxygen Delivery and Consumption. With progressive decrease in oxygen delivery, the tissue compensates by increasing the amount of oxygen extracted. If the tissue is able to extract enough oxygen then aerobic metabolism is maintained. However, there is a point when O2 delivery falls below sufficient O2 requirements after which anaerobic metabolism develops. This critical point is the onset of ischemia (DO2C). 36 1.6.4 Oxygen Extraction Ratio The ability to extract oxygen is calculated as the oxygen consumption at the onset of tissue ischemia ( V 0 2 c ) divided by the oxygen delivery at the onset of tissue ischemia ( D 0 2 c ) . This can also be plotted against oxygen delivery to represent the increases in the amount of oxygen extracted as oxygen is delivered (Figure 14). The ability to extract oxygen is based on three major factors: adequate oxygen delivery, unimpeded oxygen diffusion, and the ability to utilize oxygen. 1.6.5 Oxygen Consumption and Delivery in Sepsis Studies (106) (107) (108) (109) have shown two main changes in the oxygen consumption and delivery curves. First, the oxygen consumption at baseline is at a higher level, representing the increased overall metabolic rate. Secondly, tissues have been noted to have an earlier onset of ischemia (Figure 15). This earlier E)02C may be due to the second major change wi th sepsis, that of an impaired ability to extract oxygen by the tissue. 1.6.6 Oxygen Extraction in Sepsis A s wi th studies into the onset of ischemia, the ability to extract oxygen has also been documented to be impaired (106) (107) (108) (109). This can be documented wi th a lower oxygen extraction ratio curve for the same oxygen delivery (Figure 16). Dantzker suggested three possible reasons for an impaired oxygen extraction and utilization. First, the apparent utilization of anaerobic mechanisms for A T P generation despite high D O 2 may suggest an abnormality of the cells' ability to 37 0 2 Delivery (D02) F i g u r e 14. O x y g e n E x t r a c t i o n R a t i o . With progressive decrease in oxygen delivery, the tissue compensates by increasing the amount of oxygen extracted. If the tissue is able to extract enough oxygen then aerobic metabolism is maintained. However, there is a point when O2 delivery falls below sufficient O2 requirements after which anaerobic metabolism develops. This critical point is the onset of ischemia (DO2C). 38 Figure 15. Relationship between Oxygen Delivery and Consumption in Sepsis. There are two main changes in the oxygen consumption and delivery curves. The oxygen consumption a t baseline is at a higher level and tissues have an earlier onset of ischemia. 3 9 F i g u r e 16. O x y g e n Ex t r ac t ion R a t i o i n Seps i s . During sepsis, the ability to extract oxygen has been shown to be impaired. 40 utilize the oxygen that is delivered to it. Indeed, Hersch et al (54) in a study on the, histologic and ultrastructural changes in a rat model of hyperdynamic sepsis showed widespread mitochondrial destruction. Second, the finding of a high mixed venous PO2 in the face of tissue hypoxia may indicate a degree of shunting of blood around metabolizing tissue. Finally, there may be an impaired ability of oxygen to diffuse from the systemic capillaries to the cells requiring oxygen due to interstitial edema. 1.7 TREATMENT OF SEPSIS 1.7.1 Treatment Options Treatment of sepsis may be focused at any of the levels of septic induced toxicity (Figure 17). Certainly the prevention of infection would be the best strategy. Next would be to treat infection early to l imit the effects by either early drainage of infections or treatment wi th antibiotics (12) (2). N e w treatments include attempting to block the effects of endotoxin wi th antibodies and immunomodula t ion of the host defenses. However these new techniques have meet wi th conflicting results. Indeed, attempting to modulate the immune system is complicated at best as one needs to balance the advantages and disadvantages of the immune system. However, all too frequently, patients present far along their course of SIRS. Therefore, the treatments clinicians are left wi th are an attempt to improve oxygen delivery (47) and maintain adequate arterial perfusion pressure. Two general options are available, including inotropic support and fluid resuscitation. 41 Immunomodulate F i g u r e 17. T r e a t m e n t O p t i o n s . This figure illustrates the various treatment options currently being investigated. The currently advocated regimens include prevention, antibiotics, and augmenting oxygen delivery. Areas being investigated include immunomodulation and blocking the mediators of toxicity. 42 1.7.2 Augmentation of Oxygen Delivery Oxygen delivery is determined by the following equation: D O 2 = Oxygen content x Cardiac output where Oxygen content = Hgb x 1.39 x O 2 sat + 0.003 x P a 0 2 From this, one may see that oxygen delivery is dependent on hemoglobin, oxygen saturation, and cardiac output. Increases in any or all of these variable w i l l go to increase the oxygen delivery (47). To increase hemoglobin, one may transfuse packed red blood cells to increase the hemoglobin concentration (110). This is usually done if the hemoglobin is <100 g/1 (111). Many studies showed no benefit in oxygen consumption w h e n transfusing patients above these levels (112) (113). To increase oxygen saturation, one may increase the oxygen concentration that is given. In many situations, this may require intubation and ventilation to maximize the F i 0 2 - A second benefit of intubation and paralysis is a reduction i n the work of breathing. Hussain and Roussos (114) showed that early use of mechanical ventilation has been shown beneficial in the management of septic shock. To increase cardiac output, both vasopressors and fluids are used. Vasopressors are used to try to augment cardiac output by increasing cardiac contractility (47). Of the current inotropes, P-agonists including dopamine, dobutamine, norepinephrine, and epinephrine have been used most often. A l l but dobutamine also have a-adrenergic action that increases uneven arteriolar 43 vasoconstriction and may intensify microcirculatory flow maldistributions. Dobutamine has both p i and P2 actions; the latter relaxes previously vasoconstricted arterioles and may improve small vessel blood flow i n peripheral tissues (115). Vasodilators play a limited role in the care of septic patients since hypotension is invariably already present (7). Overall , the clinical efficacy and optimal therapy for pressor treatment has not been wel l documented in prospective randomized trials. Studies have described important side effects that would caution against their use in any situation including increased overall metabolic rate and systemic oxygen demands (116). Further, Ruiz et al report an increased mortali ty wi th the use of inotropes in such situations (117). Therefore, since the benefits of vasoactive agents are at present unclear, augmentation of cardiac output is primarily achieved by the administration of fluid resuscitation. 1.7.3 Fluid resuscitation in septic shock For many clinicians, fluid resuscitation is one of the prime therapies for the treatment of patients in septic shock (118) (2). The rationale for use of fluids has centered on re-establishing normal hemodynamic variables including the augmenting of filling pressures (2) and replacing relative and absolute intravascular volume deficits (2) (119) (120) (121) to augment cardiac function (122). F lu id therapy in septic shock is thought to increase venous return and cardiac output. When fluid therapy is associated wi th an increase in DO2 in patients wi th lactic acidosis, systemic oxygen consumption VO2 increases and lactic acid levels decrease (123) (124). Other investigators have proposed benefits at a microvascular level . 44 Hemodilut ion of blood has been shown to promote survival in critically i l l patients (111). The potential benefits include a decreased viscosity of blood (125) leading to an increased velocity of red cells (126) and possibly improved entry of erythrocytes into channels wi th smaller diameter, thereby decreasing the heterogeneity of blood flow (126) (79). Whi le both crystalloid and colloidal solutions are frequently used for resuscitation, the choice between these two fluids is controversial. Crystalloids have been generally advocated in view of availability, the lack, of potentially infectious complications, and overall costs. However, since only about 1/3 to 1/4 of the saline solution infused remains in the vascular space, large volumes are required in order to achieve therapeutic goals (127). Furthermore, wi th a significant amount going to the interstitial space, some have raised concerns regarding the development of interstitial edema (128). Wi th these concerns, many have looked to the use of colloids for resuscitation. Col lo id fluids may be advantageous because the sustained increases in plasma colloid osmotic pressure (COP) by these fluids w i l l aid in the retention of fluid in the intravascular space (129). Further, evidence shows that sepsis is associated wi th reprioritization of hepatic protein synthesis wi th decreased albumin production which contributes to a decreased colloid oncotic pressure (130). Demling et al (131) suggest that it is the hypoproteinemia which may be responsible for the early edema in soft tissues wi th sepsis. O n the other hand, the maintenance of an oncotic gradient by colloid infusion may be difficult in systemic areas in w h i c h microvascular permeability is increased. Further, there is concern that egress of larger molecules into the extravascular space may also increase the colloid oncotic 45 pressure in the interstitium, thereby limiting the effects of the colloid and in fact, may worsen the severity of the interstitial edema and hinder its resorption (127). In view of this complex issue, there is no clear agreement as to the ideal fluid to use in sepsis. Possible side-effects of crystalloid resuscitation include hemodilution that may lead to compromise of systemic oxygen delivery (132) (118), pulmonary edema (133) (5) (134) (135) (47) or interstitial edema secondary to the increased permeability (136), and hyperchloremic acidosis due to a large volume of saline required for stabilization (137). Indeed, the diffusion of oxygen from red blood cell to the tissue is dependent on surface area for diffusion and the length of the pathway from red cell to mitochondria, both functions of microvascular control, and the diffusion characteristics of oxygen in the tissue (55). Therefore, the effect of tissue edema or inflammation on the diffusing characteristics of the tissue may be significant (55). Further, while some argue that hemodilution leads to a decreased hematocrit which may reduce viscosity and improve blood flow and may thereby be beneficial for some organs like the heart and brain, other organs accommodate poorly to the decrease in oxygen carrying capacity (55). 1. 7.4 Diffusion vs filtration Diffusion refers to movement of substances in either direction across the capillary. Furthermore, diffusion rate depends on various factors including solubility of the substance in the tissues, the temperature, and the surface area available; it is inversely related to molecular size and the distance over which 46 diffusion occurs. Filtration refers to the net movement of fluid out of the capillaries and is governed by the Starling law of the capillary. 1.7.5 Starling's Law of the Capillary Flu id balance between the intravascular and interstitial fluid compartments is determined by the forces operative in the Starling law of the capillary (Figure 18): Qf = K f { P c - P i - 8 (7 ic -7 t i ) } in which Qf is the total flow of fluid across the capillary membrane; K f is the f lu id filtration coefficient; P c is the capillary hydrostatic pressure; P i is the interstitial hydrostatic pressure; 8 is the reflection coefficient; Ttc is the capillary colloid osmotic pressure; and K{ is the interstitial colloid osmotic pressure (138) (see below for details). The interaction of these four Starling forces determines fluid, flux between interstitial and intravascular compartments. Capillary hydrostatic pressure is the dominant dr iving force favoring fluid filtration across the capillaries into the interstitium. Interstitial hydrostatic pressure is usually negative, but can become positive if large amounts of edema fluid accumulate. The plasma colloid osmotic pressure is the only force acting to retain fluid wi th in the intravascular space. In contrast, interstitial colloid osmotic pressure favors fluid retention in the interstitial space and may be diluted by the accumulation of protein-sparse edema f luid. Increases in interstitial hydrostatic pressure and reductions in interstitial col loid osmotic pressure serve to limit edema formation (137). 47 Q f = K f ( P c - Pj - 8 [ 7 T C - 7X;]) F i g u r e 18. S t a r l i n g ' s Forces o f the C a p i l l a r y . The Starling's Forces determines the fluid flux between interstitial and intravascular compartments. Qf is the total flow of fluid across the capillary membrane; Kf is the fluid filtration coefficient; P c is the capillary hydrostatic pressure; Pj is the interstitial hydrostatic pressure; 8 is the reflection coefficient; 7tc is the capillary colloid osmotic pressure; and TCJ is the interstitial colloid osmotic pressure. 48 The fluid filtration coefficient represents the net amount of fluid crossing the capillary membrane for a given level of Starling's forces.. This value varies w i t h changes in the surface area of the functional microcirculation at any one t ime. Dilat ion of arterioles, and especially dilation of precapillary sphincters, can increase the number of functional capillaries. Normally, only a fraction of capillaries exhibit blood flow at any one time. The reflection coefficient is a measure of the ability of the capillary membrane to exclude large particles and l imit the movement of protein. A membrane completely impermeable to protein would demonstrate a reflection coefficient of 1. The average reflection coefficient for systemic capillaries is approximately 0.9 and for pulmonary capillaries is approximately 0.7 (139) (140). Under conditions of increased capillary permeability, the reflection coefficient may decrease to 0.4 (141), thereby favoring flow of fluid into the interstitium. The colloid osmotic pressure of plasma is determined by the number of particles in solution that are impermeable to the capillary membrane. A l b u m i n is responsible for approximately 80% of the plasma osmotic pressure (142). The normal plasma colloid osmotic pressure is 21-25 m m H g , and in a critical care population ranges from 18-20 m m H g (143) which w i l l also tend to favor flow of fluid into the interstitium. 1.7.6 Tonometry As discussed earlier, the splanchnic circulation is particularly vulnerable i n sepsis. Since some investigators have shown that the gut becomes ischemic earlier than whole body (86), many have sought a mechanism by which one may detect 49 changes early enough to prevent mucosal injury. The use of tonometry has been an evolving one. Bergofsky (144) estimated PO2 and P C O 2 in gallbladders and urinary bladders tonometrically by instil l ing the organs wi th saline and after equi l ibr ium, measuring P O 2 and P C O 2 in the intraluminal fluid. Dawson et al (145) carried this idea into use in the intestinal tract wi th small bowel. Then, rather then placing fluid wi th in the lumen, Kivisaari and Ni in ikosk i (146) measured PO2 and P C O 2 i n tissues by determining the partial pressures of O2 and C O 2 in saline contained within a Silastic tube which was permeable to these gases. Fiddian-Green et al (147) proposed the idea that tonometry could be used to estimate intramucosal p H (pHi) wi th the idea that HCO3\" concentrations in tissue and arterial blood are sufficiently similar to permit substituting the latter value into the Henderson-Hasselbalch equation (Figure 19). Finally, Antonsson et al (148) validated this technique by using microelectrodes to directly measure the ileal p H i and correlating these values to those derived by tonometry during endotoxemia and mesenteric occlusion in pigs. Clinicians have studied the role of tonometry recently and have showed some encouraging results (149) (150) (151). Gutierrez et al (152) have shown that tailoring therapy to p H i improved overall outcome in a group of critically i l l patients. Maynard et al (153) showed that gastric tonometry was the most reliable indicator of adequacy of tissue oxygenation in a group of critically i l l patients. Finally, studies have shown such a good correlation to oxygenation of tissue that tonometry may be used to determine the onset of anaerobic metabolism (154) (155). 50 Figure 19. Tonometric Analysis. This figure illustrates the manner that the tonometer is used to determine the mucosal pH (pHi). The pCC>2 of the mucosal wall equilibrates across the semipermeable membrane with the saline in the balloon. This can then be inserted into the Henderson-Hasselbalch equation using the arterial HCO3\" to determine the pHi. 51 1.7.7 Lactate Several investigators have used lactate as a marker for anaerobic metabolism (74) (106). When there is insufficient oxygen to sustain oxidative phosphorylation, increased glycolysis is called upon to maintain tissue A T P levels. Dur ing glycolysis, pyruvate acts as the terminal electron acceptor and leads to the production of lactate (55). However, Danztker (55) points out that there may be some confounding factors which may make the interpretation of serum lactate levels less than straightforward. Serum lactate levels reflect balance in production and the metabolism of lactate. In sepsis there are several reasons for increased production of lactate (156) including inflammatory mediators, cytokines, and other vasoactive substances that impair vasomotor tone, increase microvascular permeability, and facilitate aggregation of leukocytes and platelets. Capillary leakage causes decreased circulating blood volume and cardiac output that is further impaired by the direct effects of sepsis on ventricular function. Ultimately these changes lead to a fall i n perfusion and ischemia wi th the resultant anaerobic metabolism. However, other factors may either raise or lower the level of lactate. Factors raising the levels include a decreased clearance due to decreased perfusion to the liver and kidneys, the main sites of metabolism of lactate (157) (156), and a decrease in the activity of pyruvate dehydrogenase in sepsis (156). Factors leading to an underestimation of anaerobic metabolism include some tissue only producing lactate once all other alternatives to oxidative phosphorylation are exhausted, and as tissue p H falls, glycolytic flux is inhibited thereby providing a feedback on the level of acidosis (55). 52 Therefore, since lactate levels in the blood reflect the balance of all these complex effects, lactate levels are difficult values to interpret in sepsis (149) (158). 53 Chapter 2 HYPOTHESIS Sepsis has been characterized by increased capillary permeability. In view of known changes wi th respect to the Starling's forces at the capillary level, one w o u l d expect crystalloid resuscitation to result in a significant degree of interstitial edema. Interstitial edema may in turn impair oxygen diffusion from capillary to cells and thereby limit the ability of tissue to extract oxygen. Furthermore, interstitial edema along wi th endothelial edema may worsen an already heterogeneous distribution of capillaries and thereby further impair the ability of tissue to properly match perfusion wi th oxygen demands. Finally, hemodilut ion alone may lead to a decrease in the capillary hematocrit such that the delivery of oxygen w i l l be lowered. Therefore, my hypothesis is that crystalloid resuscitation w i l l impair the ability of tissue to extract oxygen in endotoxemia. 5 4 Chapter 3 OBJECTIVES OF THE THESIS The use of fluid resuscitation in sepsis is a generally accepted therapeutic maneuver. However, in studying the physiological implications of fluid therapy, many questions arise as to the potential disadvantages of fluid resuscitation. In this thesis, an attempt is made to address several important questions as to the effects of crystalloid fluids in an endotoxemic porcine model of sepsis. Specifically, the following questions w i l l be addressed: 1. Does crystalloid resuscitation impair the ability to extract oxygen? In the clinical scenario, the use of crystalloids has generally been the first l ine of therapy and often result in improved clinical parameters. However, measuring the oxygen extraction may be a more important parameter as it w i l l assess tissue ability to adapt to sepsis. To elaborate on the effects of crystalloid resuscitation, several potential mechanisms of action were studied: 2. Does crystalloid resuscitation lead to interstitial edema? In view of the described generalized leak in sepsis, one would anticipate that aggressive fluid loading w i l l lead to edema and thereby prevent oxygen extraction by tissue. Since it is difficult in the clinical scenario to quantify the degree of interstitial edema, our experiments w i l l examine this possibility. 55 3. Does crystalloid resuscitation lead to capillary blood flow heterogeneity? While some argue that fluid resuscitation reduces blood viscosity and allows for improved flow of blood, others argue that there may be endothelial and interstitial edema which may prevent flow of blood and thereby increase an already capillary heterogeneity and thereby impair oxygen extraction. 4. Does crystalloid resuscitation lead to decreased capillary hematocrit? While crystalloid resuscitation leads to hemodilution at a large vessel level, there have been no papers documenting the capillary red blood cell concentration. This is a possible side effect of fluids which may impair the overall concentration of oxygen delivered to tissues. Using morphometric techniques, one may determine these values. 56 Chapter 4 RESEARCH PLAN 4.1 THE PORCINE MODEL OF SEPTIC SHOCK Various considerations for the model of septic shock included the animal to be used and the type of sepsis produced. Pigs are quite similar to humans w i t h respect to renal, cardiovascular, and digestive anatomy and physiology (159) and therefore serve as a useful experimental animal. W i t h regards to the septic model , various types have been described including intravascular infusion of bacteria, peritonitis by bacterial inoculum, cecal ligation and perforation, or infusion of endotoxin. We chose the later for several reasons. Endotoxin, although not a sine qua non of sepsis, is considered fundamental in the development of sepsis (8). Further, endotoxin is a stable and relatively pure compound which simplifies some aspects of experimental design. In contrast, bacteria are typically stored frozen, grown in culture and washed several times to remove culture medium and solubilized bacterial products, and later require quantification (8). Finally, Ca in and Curtis point out that endotoxin may provide a better experimental model of hyperdynamic sepsis in acute animal experiments than live bacteria as it more closely matches changes in oxygen consumption wi th a rise by 15% to 20% above the critical D O 2 (160). Our study also examined the role of fluid resuscitation in the endotoxemic model. Whi le fluid administration in the clinical setting is not standardized, we needed a method of administration which was consistent and rational. A fixed 57 volume of crystalloid has been described by Fink and Heard (8). A fixed volume as opposed to one titrated to hemodynamic variables has several advantages. First, myocardial performance is impaired by endotoxemia, thus preload may need to be elevated to supranormal levels to maintain cardiac output. Second, sepsis may lead to alterations in ventricular diastolic compliance, thus, left ventricular end-diastolic and pulmonary capillary wedge pressures may be poor estimates for left ventr icular end-diastolic volume, the relevant variable for assessing preload. Finally, central venous and pulmonary capillary wedge pressures are typically small numbers, and using conventional catheter-transducer-amplifier systems, there is considerable imprecision (5-15%) (8) in the measurement of these variables; thus, it wou ld lead to imprecise titration of fluid administration. A porcine model of endotoxemia in which hemodynamic stability is achieved by an aggressive fluid resuscitation protocol has been previously described (52) (161). The model adequately reproduces several features of septic shock in humans including profound systemic arterial hypotension and low SVRI (96). Further, the volumes of fluid used were 25 cc/kg/hr for maintenance and 48 cc /kg/hr for resuscitation. This model has been used in recent studies of sepsis in our laboratory (74). Thus, we used a similar porcine model of endotoxemia wi th fluid resuscitation being crystalloid administered as a continuous infusion at the same fixed rates. 58 4.2 MORPHOMETRIC ANALYSIS OF GUT TISSUE Morphometry is the measurement of structure. It's aim is often to answer questions such as \"how many\" and \"how large\" something is. Whi le this is often the goal of morphometry, it is only wi th stereology that one is able to answer these questions. Stereology is the three-dimensional interpretation of flat images by the criteria of geometric probability. It can also be defined as extrapolation from two-dimensional to three-dimensional space. Stereology is practiced by measuring and counting profiles in sections (Figure 20). For most stereological problems, this can be done by superimposing grids of lines or squares and circles over an image. One may adjust sampling wi th in the ind iv idua l animal to a specific level of acceptable error. Various formulae are available (162) to allow determination of numbers of counts to l imit error rates. Counting involves laying the grid on the area of interest and then counting the grid intersections which hit the component of interest. The number of points that hi t profiles of the component, divided by the total number of test points equals the volume fraction of that component in the entire specimen. We apply these techniques in the determination of the gut interstitial volume and the capillary hematocrit. 4.3 INFRA-SCAN ANALYSIS OF GUT TISSUE The infra-scan method of tissue analysis is a recent development i n microscopic analysis. It utilizes a microscope attached to a camera which sends 59 Capillaries Slide to count F i g u r e 20. M o r p h o m e t r i c A n a l y s i s . Morphometry attempts to determine the number of a certain area interest, in this case capillaries measuring and counting profiles in sections 60 images through to a computer. The investigator is able to choose the image of interest and to select the area in question by use of a cursor. The imaging software (Infrascan, Richmond B.C.) w i l l analyze the entire field for pixels wi th s imi lar characteristics as the one chosen. The software w i l l then quantify the total images matching the area of interest and determine the fraction area that it represents of the total field (Figure 21). By using visual inspection, the automated analysis may be confirmed. Various areas of interest may be looked at. We were interested in the area fraction of interstitial space as this could be used to estimate the degree of interstitial edema that was present within the gut. 4.4 DETERMINATION OF GUT CAPILLARY HEMATOCRIT While the effects of fluid loading result in hemodilut ion at an arterial and venous level, the hematocrit in the microvessels may very wel l be different (55). Indeed, some investigators believe that the ability of the microvascular hematocrit to change independent of the systemic hematocrit in response to alterations in the microenvironment may account for some of the conflicting observations made during alterations in systemic hematocrit (55). Therefore, since changes in the systemic hematocrit may not reflect the results at a capillary level, we sought to determine this by morphometric technique. Using previously established techniques of morphometric analysis (163) (164), random samples w i l l be cut, fixed, and stained for cutting and processing for photography under a t ransmission electron microscope. Point counting technique w i l l be used to determine the 61 Microscopic Appearance Infrascan Analysis F i g u r e 21. In f ra - scan A n a l y s i s . Infrascan is a technique whereby the slide is placed on a microscope attached to a camera which feeds the image to a computer. The investigator chooses the image of interest and select the area in question. The imaging software analyzes the entire field for pixels with similar characteristics as the one chosen. The software will then quantify the total images matching the area of interest and determine the fraction area that it represents of the total field. 62 relative volume fraction occupied by erythrocyte and total capillary volume, thereby determining the capillary hematocrit. 4.5 RELATIVE DISPERSION OF RED BLOOD CELL TRANSIT TIME A major determinant of the transport of oxygen from blood to tissue is the time spent by the red blood cell passing through the vessels, called the red blood cell transit time. The red cell transit time in a given region is dependent on the blood volume and blood flow in that region, both of which may vary. The blood v o l u m e and blood flow may each be determined as per studies performed on the lung by Hogg et al (163) and the heart by Al l a rd et al (164). Regional blood volume w i l l be determined using Technetium (^mTc) labelled red blood cells and regional blood flow w i l l be measured using a reference flow technique and a left atrial injection of 15 urn microspheres labelled wi th Strontium (85Sr). The different radioactive signals can be used to quantify red cells and microspheres. Since the microspheres are also removed wi th a reference flow technique, one may correlate the counts of microspheres wi th the flow of these microspheres. Red blood cell transit time for the entire portion of tissue was determined by the quotient of blood volume and blood flow. By analyzing the red blood cell transit times in various segments of intestine, one may determine the average transit time and the variability in these values. The relative dispersion or coefficient of variation is the quotient of the standard deviation and the mean. As such, it provides one wi th a measure of the heterogeneity of the microcirculation in question. 63 4.6 TONOMETRY Fiddian-Green et al ( 1 4 7 ) introduced the concept that tonometry could be used to estimate intramucosal p H (pHi) wi th the idea that H C O 3 \" concentrations in tissue and arterial blood are sufficiently similar to permit substituting the latter value into the Henderson-Hasselbalch equation (Figure 1 9 ) . Antonsson et al ( 1 4 8 ) have validated this technique by using microelectrodes to directly measure the ileal p H i and correlating these values to those derived by tonometry during endotoxemia and mesenteric occlusion in pigs. The gastrointestinal tonometer consists of a saline-filled silicone balloon that is placed in the gut lumen. The silicone balloon is highly permeable to oxygen and carbon dioxide. Normal saline is injected into the ileal tonometer balloon and left to equilibrate for 4 0 minutes. A t the end of the equilibration period, the saline w i l l be removed under anaerobic conditions. The first mil l i l i ter of saline removed is discarded, as this is not in direct contact wi th the balloon. The C O 2 content of the subsequent volume of saline removed, which reflects the mucosal CO2, w i l l be measured. Simultaneously, the arterial bicarbonate (assumed to be equal to the bicarbonate content of the gut mucosa) w i l l be measured. Using a conversion table, measured PCO2 w i l l be transposed to a steady state PCO2 ( P C C ^ s s ) , depending on the exact duration of equilibration. Gut mucosal p H w i l l be calculated by substituting the PCO2SS and the simultaneously measured arterial bicarbonate in the Henderson-Hasselbalch equation (ie: p H i = 6 . 1 + Log{(HC03)/Pcc>2}) . 6 4 Chapter 5 EFFECT OF CRYSTALLOID RESUSCITATION O N WHOLE BODY A N D GUT IN ENDOTOXEMIA 5.1 INTRODUCTION In addition to antibiotic therapy, fluid resuscitation is a key component of the management of sepsis and septic shock (2). The goal of fluid resuscitation is to restore intravascular volume to maintain an adequate blood pressure and cardiac output. Restoring a normal cardiac output requires intravascular replacement of third space losses, compensation for venous pooling of blood, and a sufficiently high left ventricular filling pressure to compensate for decreased ventricular contractility during sepsis (72). Importantly, even a normal cardiac output may not necessarily be adequate during sepsis k>ecause sepsis is accompanied by impaired tissue oxygen extraction (55) (74) (106) (12). That is, higher oxygen delivery (equals cardiac output times arterial oxygen content) is required to prevent evidence of unmet oxygen demand (76) (12), including decreasing oxygen consumption (74) (106), rising lactate levels (74) (106), low gastric intramucosal p H (152), and organ system dysfunction (55) (76) (12). The causes of impaired tissue oxygen extraction during sepsis have not been fully elucidated. However, impaired oxygen uptake, due to mismatching of oxygen delivery to demand (74) (78) and due to impaired oxygen diffusion from erythrocytes to tissue mitochondria (165), appear to play important roles. Mismatching of oxygen delivery to demand and inadequate oxygen diffusion are due to several microvascular phenomenon; capillary plugging wi th leukocytes and 65 other debris (41), capillary endothelial edema and tissue edema (54), and altered microvascular flow (74) (57). While normal response to impaired oxygen delivery is met wi th a homogeneous capillary recruitment, the microvascular changes associated wi th sepsis lead to a heterogeneous capillary distribution (74) (57). W h i l e some argue that microvascular flow distribution and tissue oxygen extraction may be improved by fluid resuscitation (125), others argue that fluid resuscitation may lead to interstitial and endothelial edema (136) which may further impa i r microvascular flow and lead to a higher degree of capillary heterogeneity. Considering the multiple effects of sepsis and fluid resuscitation on the intravascular volume, on cardiac function, on the microvascular distribution of oxygen delivery and on tissue oxygen demand, fluid resuscitation in sepsis is a complex intervention. Our goal was to determine whether fluid resuscitation i n sepsis impairs tissue oxygen extraction, possibly due to interstitial edema and altered microvascular flow distribution (126) (78). The splanchnic circulation has been thought to be particularly susceptible to ischemic injury (52) (86) (90). This susceptibility has been attributed to specific dysregulation of organ perfusion (90), to increased overall oxygen requirements by the metabolically active splanchnic organs (86), and to the countercurrent flow i n the intestinal v i l l i whereby there is shunting of oxygen away from the tips of the v i l l i (55) (87) (89). Furthermore, it has been proposed that ischemia of the splanchnic circulation may predispose to decreased mucosal integrity and lead to bacterial translocation which may then lead to multiple organ dysfunction 66 syndrome (166) (52). Thus, we were particularly interested in intestinal oxygen extraction and the splanchnic circulation. Most describe sepsis as characterized by an increased capillary permeability (136) such that tissues are more prone to develop interstitial edema (136) (54). To further determine whether interstitial edema occurs in tissue in endotoxemia, we determined the volume of interstitial space by morphometric techniques. Some investigators have suggested that microvascular flow may be improved wi th hemodilution by reducing hematocrits (111) (125) leading to a decrease in blood viscosity (125) thereby increasing blood velocity and favoring entry into channels wi th smaller diameters (126). However, hemodilut ion also may lead to a decrease in the oxygen carrying capacity of blood with a resulting decrease in oxygen delivery at a microvascular level. Further, since hematocrits have only been determined at a large vessel level, this may not reflect the effects of hemodilut ion at a tissue leve l (55). Therefore, to further elucidate the possible effects of fluid resuscitation, we sought to determine the capillary hematocrits by morphometric analysis. Accordingly, we sought to determine whether crystalloid resuscitation in a porcine model of septic shock would impair oxygen extraction capacity of the whole body and specifically of the splanchnic bed. To better understand the mechanism of this effect we measured the dispersion of blood flow transit times in the gut, the volume fraction of interstitium, and the gut capillary hematocrit. 67 5.2 METHODS 5.2.1 Experimental Design and Protocol The experimental design was similar to that performed previously in our laboratory (74). Sample size was calculated by analyzing the results of a pre l iminary study using a sample size of 8 animals per group in endotoxic and control groups. To determine a difference in critical D O 2 of 4 m l 0 2 / k g . m i n using a standard deviation of 2 m l 0 2 / k g . m i n , required four groups of eight animals, for an alpha = 0.05 and beta = 0.10 (\"N\" Calculator for Macintosh, v 0.9, Compuserve #70721,3243). Therefore, a size of 8 animals per group were used as a reasonable estimate of required sample size. Dur ing instrumentation (see below), surgical preparation, and stabilization, all animals received 0.9% Sodium Chloride solution (Baxter, Toronto, O N ) infused via the left external jugular catheter at 25 ml'kg~l*hr~l. Fol lowing this, the animals were randomized to one of four groups: Control / F l u i d (n=8), C o n t r o l / N o f luid (n=8), Endotoxin /F lu id (n=8), and Endotoxin/No-f lu id (n=8) groups. Endotoxin groups received E. coli endotoxin 50 | ig /kg (0111:B4, Sigma, St. Louis, M O ) in 60 m l normal saline over 30 minutes immediately after the baseline data set. Cont ro l groups received an infusion of 60 m l normal saline without endotoxin. F l u i d groups received an infusion of normal saline at 48 ml*kg~l*hr~l from the baseline measurement set unti l the end of the experiment. No-f lu id groups d id not receive any further saline infusion. Fol lowing randomization to respective treatment groups, progressive hemorrhage was undertaken at 3 m l per minute using a constant withdrawal pump from the left carotid catheter unt i l the animal died. 68 Prior to death, when whole body oxygen consumption had fallen by 25%, suggesting that oxygen delivery was inadequate to meet demand, we infused radiolabeled red blood cells, injected radiolabeled microspheres (see below), and rapidly excised the previously prepared segment of jejunum (Figure 22). 5.2.2 Surgical Preparation and Instrumentation This study was approved by the Animal Care Committee of the Universi ty of British Columbia and conforms to National Institute of Health (NIH) standards for animal experimentation. Thirty-two pigs, weighing 25.7 \u00C2\u00B1 2.8 kg, were fasted overnight and then sedated wi th 0.5 m g / k g midazolam i.m. (Hofman la Roche, Mississauga, ON) . Thirty minutes later the animals were anesthetized using ketamine 500 mg i.m. (MTC Pharm, Cambridge, ON) followed by thiopentol 125 to 250 mg i.v. (Abbott, Montreal, PQ) titrated to effect. Anesthesia was maintained throughout the experiment using ketamine 5 ml*kg\"l*hr\"l i.v. infusion and 0.5% inspired isoflurane (Anaquest, Mississauga, ON) . To avoid changes in whole body oxygen demand, skeletal muscle relaxation was maintained wi th intravenous pancuronium bromide infusion (Organon, Scarborough, O N ) at 6 mg/hr , titrated to effect. A tracheostomy was performed and an 8.0 m m endotracheal tube (Portex, Wi lming ton , M A ) was inserted and secured. Dur ing instrumentation and experimentation the animals\" were mechanically ventilated (Harvard Apparatus dual phase control respirator pump, model 613, Mi l l s , M A ) wi th 30% oxygen. A l o w compliance catheter was inserted into the right carotid artery for arterial pressure 69 Endotoxi n or WB Ischemia Sham Infusion (Gut harvest) Surgical Prep. Stablizel Hemorrhage i 1 hr4 Orrin [ 3cc/rrin t TREATMENT GROUPS: ^p1 : 25cc/kg.hr (s) NonRuid resuscitated: Occ/kg.hr (n=8) Qp2: 25ccVkg.hr (s) Ruid resuscitated: 48cc/kg.hr (n=8) ^ p a 2 5 ^ h r ( E ) NonRuid resuscitated: 0cc/kg.hr (n=8) O p 4- 25cdkg.hr (\u00C2\u00A3) \^ : y : \u00E2\u0080\u00A2 Ruid resuscitated: 48cc/kg.hr (n=8) Figure 22. Protocol Timeline. This figure illustrates the protocol for all four groups of animals. S = saline sham infusion, E = endotoxin infusion. 70 measurement and arterial blood sampling (Figure 23). A catheter was inserted into the left external jugular vein for saline infusion and administration of medications. A pulmonary artery catheter (Criticath model DSP5105H, Ohmeda Medical Devices, Oxnard, C A ) was placed via the right external jugular vein for measurement of right atrial pressure, pulmonary artery occlusion pressure, for mixed venous blood sampling, and for cardiac output measurement in triplicate (Thermodi lu t ion Cardiac Output monitor Edwards Mode l 9250, Baxter Health Care, Irvine, C A ) . A catheter was inserted into the left carotid artery for hemorrhage. A n anterolateral thoracotomy was performed through the fifth or sixth intercostal space on the left. The pericardium was entered and a catheter inserted and secured in the left atrial appendage. This catheter was used for injection of radiolabelled red blood cells and radioactive microspheres for gut blood volume and flow measurements, respectively. Through a midline laparotomy, the pancreaticoduodenal vein at the second part of the duodenum and the superior rectal vein at the promontory of the sacrum were tied off. Fol lowing this, the splenic artery and vein were tied off to prevent autotransfusion and a catheter to sample portal vein blood was inserted via the splenic vein stump. Ligation of these vessels ensured that the gastrointestinal circulation was isolated such that venous drainage passed through the portal v e i n . Portal venous flow was measured by placement of a 1.5 cm ultrasonic flow probe connected to a Transonic T201 Ultrasonic Blood Flow Meter (Transonic Inc., Ithaca, N Y ) around the portal vein. A n orogastric tube was inserted to allow drainage of the gastric secretions. A tonometric catheter (Tonometries, Inc., 373 Plantation 71 72 Street, Worcester, MA) was inserted into the lumen of the ileum 30 cm proximal to the ileocecal valve through a small wound in the antimesenteric border and held in place with a suture. Thirty cm distal to the duodenal-jejunal junction a 60 cm length of jejunum was isolated for later resection for morphometric analysis. Umbilical tapes were brought around the mesentery to allow ligation of the vasculature following injection of radiolabelled microspheres (see below). The mesentery was otherwise not disrupted. The abdomen was loosely closed to facilitate later removal of the jejunal segment. 5.2.3 Hemodynamic Measurements and Calculations We measured arterial, mixed venous, and portal vein pH, Pc0 2' a n < ^ ^02 (ABL30, Radiometer, Copenhagen, Helsinki), C\u00C2\u00BB2 content (IL 482 co-oximeter, Instrumentation Laboratories, Lexington, MA) and lactate concentration (YSI 2300STAT lactate analyzer, Yellow Springs Instruments, Yellow Springs, OH) at baseline and every 20 minutes during progressive hemorrhage. Heart rate, mean arterial pressure, pulmonary artery pressure, central venous pressure, pulmonary capillary occlusion pressure, cardiac output, portal vein flows, and whole body oxygen consumption using a metabolic cart (mean of 5 measures at 1 minute intervals, MBM-1000, Deltatrac, Helsinki, Finland) were also recorded at 20 minute intervals. A complete blood count (Coulter counter, Coulter, Miami Lakes, FL) and tonometer measurements was measured every 40 minutes (see below). Blood oxygen content was calculated as hemoglobin x 1.39 x blood oxygen saturation + 0.003 x PO2. Whole body oxygen delivery (WB-DO2) w a s calculated as 73 cardiac output mult ipl ied by arterial oxygen content. Gut oxygen consumpt ion (Gut-VC>2) was calculated as portal vein flow mult ipl ied by the difference between arterial and portal venous oxygen content. Gut oxygen delivery (Gut-DC^) was calculated as portal vein flow mult ipl ied by arterial oxygen content. The oxygen extraction ratio for both whole body (WB-ER) and gut (Gut-ER) was calculated as oxygen consumption divided by oxygen delivery. From the multiple oxygen delivery - oxygen consumption points obtained during progressive hemorrhage the whole body and gut critical oxygen extraction ratio was determined using Samsel and Schumacker's dual-line regression analysis (167). 5.2.4 Transit times Gut blood flow and blood volume were determined in the isolated segments of gut after the lumenal contents were removed and after the tissue had been fixed for 24 hours in 6% phosphate buffered gluteraldehyde (163). Prior to removal , Technetium (99mj c ) labelled red blood cells were injected into the left atrial appendage and allowed to distribute in the vascular compartment for 10 minutes. 15 | i m microspheres labelled wi th Strontium (85Sr) were then injected rapidly (10 seconds) into the left atrium. A t the time of microsphere injection, blood was withdrawn from the left common carotid artery at 10 ml /minute for 2 minutes into weighed vials for later radiation counting. Next, the arterial and venous vasculature of the gut segment were simultaneously cross clamped and the segment of gut was rapidly excised and immersed in 6% phosphate buffered gluteraldehyde. The 60 cm segments of gut were divided into 30-two cm pieces and each placed into 74 preweighed vials containing 6% phosphate buffered gluteraldehyde. Each segment was weighed and counted using a Beckman 8000 gamma counter for 3 minutes (164). The blood volume of each 2 cm piece of gut was calculated as counts per minute of each piece divided by counts per minute per m l of the reference blood sample. Blood flow to each 2 cm piece of gut was calculated as microsphere counts per minute from that piece of gut divided by microsphere counts per minute per m 1 of the reference arterial withdrawal sample. Average transit time for each piece of gut was calculated as blood volume divided by blood flow giving units of time (163). Fol lowing calculation of individual transit times, a distribution of transit times for all of the pieces of gut taken together was then determined for each gut segment. 5.2.5 Interstitial Volume Jejunum from 5 animals in each group was randomly selected (Figure 24). From each of these animals, six of the thirty 2 cm sections of jejunum were randomly selected and the tissue was embedded in glycol methacrylate, sectioned at 2 jxm and stained using methenamine silver. Slides were coded so that the morphometric analysis was blinded. The fraction of bowel wa l l that was interstitial space was quantitated at 400X using an automated image analysis system (Infrascan, Richmond B.C.) for 5 random fields on each slide. O n each slide an area of interstitial space was identified manually. The analysis program then identified the fractional area on the slide that had matching characteristics. In each case the area fraction of interstitial space identified by the analysis package was confirmed by visual inspection. 75 60 cm Jejunum V V V V V V Figure 24. Processing jejunal specimen for morphometric analysis. Jejunum from 5 animals in each group was randomly selected. From each of these animals, six of the thirty 2 cm sections of jejunum were randomly selected. A portion of each piece selected was embedded in glycol methacrylate. Interstitial volume was determined for various areas while capillary hematocrit was determined by analyzing 5 capillaries/block. 76 5.2.6 Capillary Hematocrit Jejunum from 5 animals in each group was randomly selected. From each of these animals, six of the thirty 2 cm sections of jejunum were randomly selected and fixed overnight in 2.5% gluteraldehyde in 0.1M cacodylate buffer. The tissue was postfixed for 2 hours in 1% osmium tetroxide, then stained en bloc for 1 hour in 5% aqueous uranyl acetate and embedded in Effapoxy resin. Thin sections, selected from 1 | im toludine blue-stained section, were cut on a Reichert Ultracut ultramicrotome, mounted on 200 mesh copper grids, and stained wi th lead citrate. Five capillaries (<10 Jim) per section (30 capillaries per jejunal segment) were photographed at 5000X magnification on a Zeiss 10CR transmission electron microscope. Capillary plasma and erythrocyte volume fractions were determined by point-counting using a 400 point grid. The capillary hematocrit was calculated as the quotient of erythrocyte volume fraction and total capillary volume fraction. 5.2.7 Ileal Tonometry The gastrointestinal tonometer consists of a saline-filled silicone balloon that is placed in the gut lumen. The silicone balloon is highly permeable to oxygen and carbon dioxide. Normal saline (5.0 m l , at room temperature) was injected into the ileal tonometer balloon and was left for 40 min. A t the end of the equilibration period, the saline was removed under anaerobic conditions. The first mil l i l i ter of saline removed was discarded, as this was not in direct contact wi th the bal loon (since the tonometer tube has a residual volume of one mill i l i ter) . The C O 2 content of the subsequent volume of saline removed, which reflects the mucosal CO2, was 77 measured using the A B L Blood Gas Analyzer. Simultaneously, the arterial bicarbonate - assumed to be equal to the bicarbonate content of the gut mucosa - was measured. Using a conversion table supplied by the manufacturer (Tonometries Inc.), measured PCO2 was transposed to a steady state PCO2 (PCO2SS), depending o n the exact duration of equilibration. Gut mucosal p H was calculated by substituting the PCO2SS and the simultaneously measured arterial bicarbonate in the Henderson-Hasselbalch equation (ie: p H i = 6.1 + Log{(HC03)/PCC\u00C2\u00BB2}). The use of tonometry to measure changes in ileal intramucosal p H induced by endotoxin infusion in pigs has previously been validated (148). 5.2.8 Statistical Analysis The relationship between V O 2 and D O 2 for whole body and gut was determined for each animal by finding two best-fit linear regression lines (167) f rom a plot of V O 2 and D O 2 . The critical oxygen extraction ratio, ERc, was determined at the point of intersection of the two lines (Figure 25). The relationship between lactate and D O 2 for whole body and gut was determined for each animal by f inding two best-fit linear regression lines as above from a plot of lactate and D O 2 . To determine whether intravascular volume expansion altered ERc dur ing endotoxemia, we used a two-way A N O V A testing for effect of volume and endotoxin taking p < 0.05 as significant. We calculated the mean (|i), second moment (a2) and relative dispersion (o7|4,) of the distribution of gut blood flow transit times using standard formulas (74). We also used a 2 way A N O V A to test for 78 Critical Extraction Ratio 0.0 L i I I i i i i i 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 D0 2 (micymin.kg) F i g u r e 25. A n example o f O2 C o n s u m p t i o n a n d D e l i v e r y C u r v e . Typical whole body oxygen consumption (VO2) and delivery (DO2) data points, measured during progressive hemorrhage, are illustrated. During progressive hemorrhage aerobic metabolism maintains oxygen consumption at approximately a constant value (7.60 ml C\u00C2\u00BB2/min.kg) until a critical point is reached (10.00 ml 02/min.kg), after which oxygen consumption falls indicating anaerobic metabolism. Dual regression lines are fit to the data points. The intersection of the two lines identifies the critical oxygen extraction ratio (=7.60 / 10.00 = 0.76). 79 differences due to fluid resuscitation and endotoxin in the relative dispersion of gut blood flow and in the area fraction of interstitial space. A one-way A N O V A was used to test for effect of saline resuscitation on arterial and capillary hematocrit. When a significant difference was found we used a sequentially rejective Bonferroni test procedure to identify individual differences. Results are presented as mean \u00C2\u00B1 standard deviation throughout the text, tables, and figures. 5.3 RESULTS 5.3.1 Effect of Endotoxin Endotoxin infusion significantly reduced the critical oxygen extraction ratio in the whole body (Endotoxin 0.55 \u00C2\u00B1 0.08 versus Control 0.82 \u00C2\u00B1 0.06, p < 0.05) (Figure 26) (Table 1) and gut (Endotoxin 0.52 \u00C2\u00B1 0.05 versus Control 0.77 \u00C2\u00B1 0.07, p < 0.05) (Figure 27) (Table 1). Endotoxin infusion resulted in an earlier onset of ischemia in whole body (Endotoxin 12.2 \u00C2\u00B1 3.8 versus Control 9.1 + 1.7 m l 0 2 / m i n . k g , p < 0.05) (Figure 28) (Table 1) but not in gut (Endotoxin 24.3 \u00C2\u00B1 9.0 versus Control 25.3 \u00C2\u00B1 14.8 m l 0 2 / m i n . k g , p > 0.05) (Figure 29) (Table 1). Relative dispersion of R B C flow transit times between segments of jejunum was not affected by endotoxin (Figure 30). This measurement of blood flow distribution is not a measure of the capillary distribution as it includes transit time through all jejunal wal l vessels. Endotoxin had no significant effect on the volume fraction of interstitium (p > 0.05) (Figure 31). Endotoxin d id not affect arterial hematocrit (p > 0.05) or capillary hematocrit (p > 0.05) (Table 2). Endotoxin significantly decreased the arterial and portal v e i n leukocyte counts compared to controls at onset of ischemia (Endotoxin arterial 2.2 \u00C2\u00B1 80 0.5, portal 2.3 \u00C2\u00B1 0.8 wbc/mm 3 versus Control arterial 25.0 \u00C2\u00B1 11.7, portal 25.2 \u00C2\u00B1 12.1 wbc/mm 3 , p < 0.05) (Table 3). Endotoxin reduced the pHi as determined by tonometry at onset of ischemia (Endotoxin 6.91 \u00C2\u00B1 0.15 versus Control 7.06 \u00C2\u00B1 0.10, p < 0.05) (Table 4). Endotoxin did not change the arterial lactate values compared to controls (p > 0.05) (Table 5). Endotoxin also had no effect on the onset of rise of lactate compared to controls (p > 0.05) (Table 6). Initial oxygen transport (Table 7) and hemodynamic (Table 8) variables did not differ between the Endotoxin and Control groups (p > 0.05). At the critical oxygen delivery the Endotoxin groups had lower mean arterial blood pressure (Endotoxin 44 \u00C2\u00B1 14 versus Control 59 \u00C2\u00B1 11 mmHg, p < 0.05), lower systemic vascular resistance (Endotoxin 516 + 376 versus Control 1651 + 408 dynes^sec'l'cm\"^, p < 0.05), and higher pulmonary artery occlusion pressure (Endotoxin 7.7 \u00C2\u00B1 2.5 versus Control 4.3 + 1.8 mmHg, p < 0.05). In addition the Endotoxin groups survived for a shorter period of hemorrhage (Endotoxin 128 \u00C2\u00B1 43 versus Control 202 + 60 minutes, p < 0.05) (Table 9). 5.3.2 Effect of Fluid Resuscitation Fluid resuscitation significantly reduced the critical oxygen extraction ratio in the whole body (Fluid 0.62 \u00C2\u00B1 0.08 versus No-Fluid 0.82 \u00C2\u00B1 0.06, p < 0.05) ) (Figure 26) (Table 1) and gut (Fluid 0.67 \u00C2\u00B1 0.06 versus No-Fluid 0.77 \u00C2\u00B1 0.07, p < 0.05) ) (Figure 27) (Table 1). Fluid resuscitation did not alter the onset of ischemia in whole body (Figure 28) or gut (Figure 29) (p > 0.05) (Table 1). Fluid resuscitation significantly increased the relative dispersion of RBC flow transit times in both the control (Fluid 81 0.81 \u00C2\u00B10.59 versus No-fluid 0.36 \u00C2\u00B1 0.11, p < 0.05) and the endotoxin groups (Fluid 0.60 \u00C2\u00B1 0.46 versus No-fluid 0.31 \u00C2\u00B1 0.04, p < 0.05) (Figure 30). Fluid resuscitation significantly increased the area fraction of interstitial space to the same extent in both the Endotoxin (Fluid 63 + 20% versus No-Fluid 50 \u00C2\u00B1 20%, p < 0.05) and Control groups (Fluid 70 \u00C2\u00B1 21% versus No-Fluid 45 \u00C2\u00B1 19%, p < 0.05) (Figure 31). Arterial hematocrit at the onset of ischemia decreased compared to baseline with fluid resuscitation in Control (Ischemic 16 \u00C2\u00B1 4% versus Baseline 22 \u00C2\u00B1 3%, p < 0.05) and Endotoxin (Ischemic 20 \u00C2\u00B1 4 versus Baseline 23 \u00C2\u00B1 3%, p < 0.05) groups (Table 2). However, capillary hematocrit at the onset of ischemia was unchanged from baseline by fluid resuscitation in Control and Endotoxemic animals (p > 0.05) (Table 2). Fluid resuscitation decreased the arterial and portal vein leukocyte counts in the control groups at onset of ischemia (Fluid arterial 14.9 + 10.5, portal 14.9 + 9.9 wbc/mm3 versus Non-fluid arterial 25.0 \u00C2\u00B1 11.7, portal 25.2 \u00C2\u00B1 12.1 wbc/mm3, p < 0.05) (Table 3). Fluid resuscitation did not alter the arterial or portal vein leukocyte counts in the endotoxemic groups at onset of ischemia (p > 0.05) (Table 3). Fluid had no effect on the pHi as determined by tonometry at onset of ischemia when compared to non-fluid groups (p > 0.05) (Table 4). Fluid did not change the arterial lactate values when compared to non-fluid groups (p > 0.05) (Table 5). Fluid also had no effect on the onset of rise of lactate compared to non-fluid groups (p > 0.05) (Table 6). The initial oxygen transport (Table 7) and hemodynamic (Table 8) variables did not differ between the Fluid and No-fluid groups (p > 0.05). At the critical oxygen delivery the Fluid groups had a higher mean arterial pressure (Fluid 77 \u00C2\u00B1 12 82 versus No-fluid 59 \u00C2\u00B1 11 mmHg, p < 0.05), higher cardiac output (Fluid 5.7 \u00C2\u00B1 2.7 versus No-fluid 2.8 \u00C2\u00B1 0.7 L/min, p < 0.05), lower systemic vascular resistance (Fluid 1100 \u00C2\u00B1 600 versus No-fluid 1650 \u00C2\u00B1 410 dynes*sec~l*cm~5, p < 0.05), higher pulmonary artery occlusion pressure (Fluid 6.9 \u00C2\u00B1 1.1 versus No-fluid 4.3 \u00C2\u00B1 1.8 mmHg), and a lower hemoglobin (55 \u00C2\u00B1 11 versus No-Fluid 83.6 \u00C2\u00B113.0 g/1, p < 0.05). The fluid resuscitated groups had a greater time of survival after onset of hemorrhage (368 \u00C2\u00B1 105 minutes) compared to the non-fluid resuscitated group (202 + 60 minutes, p < 0.05) (Table 9). 83 1.00 p 0.90 . 0.80 \u00E2\u0080\u00A2 Whole Body 0 7 0 \" Extraction 0.60 R a t i o 0.50 0.40 0.30 0.20 0.10 0.00 \u00E2\u0080\u00A2 0.82 \u00C2\u00B1 0.06 \u00E2\u0080\u00A2 I . II i i 0.62 \u00C2\u00B1 0.08 Non-Fluid Fuid Control Si 0.55 \u00C2\u00B10.08 0.48 \u00C2\u00B10.04 Non-fluid Fluid Endotoxin Groups F i g u r e 26. W h o l e B o d y A b i l i t y to Extract O x y g e n . The whole body critical oxygen extraction ratio for all four groups is illustrated. Fluid-resuscitation reduced the critical oxygen extraction ratio in the Control groups but not significantly in the Endotoxin groups. Endotoxin reduced the critical oxygen extraction ratio (ANOVA, p <0.001). Values are expressed as mean \u00C2\u00B1 SD. 84 Gut Extraction Ratio 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 i : i 0.77 \u00C2\u00B10.07 Non-fluid 0.67 \u00C2\u00B1 0.06 Fluid Control * ; i 0.52 \u00C2\u00B10.05 0.52 \u00C2\u00B10.09 Non-fluid Fluid Endotoxin Groups Figure 27. Gut Ability to Extract Oxygen. The gut critical oxygen extraction ratio for all four groups is illustrated. Like the whole body, fluid-resuscitation reduced the gut critical oxygen extraction ratio in the Control groups but not in the Endotoxin groups. Endotoxin reduced the gut critical oxygen extraction ratio (ANOVA, p <0.001). Values are expressed as mean \u00C2\u00B1 SD. 85 25 20 r Whole Body Critical D0 2 15 (ml02/min.kg) 10 \u00E2\u0080\u00A2 15.3 \u00C2\u00B13 .1 \u00E2\u0080\u00A2 12.2 \u00C2\u00B13 .8 10.2 \u00C2\u00B12 .2 .9.1 \u00C2\u00B11.7 \u00E2\u0080\u00A2 s i a Non-fluid Fluid Control Non-fluid Fluid Endotoxin Groups F i g u r e 28. W h o l e B o d y O n s e t o f I s c h e m i a . Critical DO2 is not altered by fluid resuscitation in the whole body in nonseptic or the septic groups. However the critical DO2 is higher in the pooled septic group compared to the pooled nonseptic group. 86 Gut Critical D0 2 (mlO^min.kg) 50 r 45 -40 -35 -30 -25 20 15 10 5 0 \u00E2\u0080\u00A2 25.3 + 14.8 24.3 \u00C2\u00B1 9.0 \u00E2\u0080\u00A2 22.9 \u00C2\u00B1 11.7 \u00E2\u0080\u00A2 20.5\u00C2\u00B16.8 \u00E2\u0080\u00A2I\" \u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2 :l :I :\u00E2\u0080\u00A2 Non-fluid Fluid Non-fluid Fluid Control Endotoxin Groups F i g u r e 29. G u t O n s e t o f I schemia . Critical DO2 is not altered by fluid resuscitation in the gut in nonseptic or the septic groups. 87 Relative dispersion 1.60 1.40 r 1.20 1.00 0.80 0.60 0.40 0.20 0.00 0.81 \u00C2\u00B1 0 . 5 9 0.36 \u00C2\u00B1 0 . 1 1 X 0.60 \u00C2\u00B1 0.46 0.31 \u00C2\u00B1 0.04 Nonfluid Fluid Nonfluid Fluid Control Endotoxin Groups F i g u r e 30. R e l a t i v e d i s p e r s i o n s o f gut b l o o d t ransi t t imes . The average relative dispersion of total gut blood flow transit times in thirty ~2g segments of jejunum is shown for the four experimental groups. Fluid resuscitation significantly increased relative dispersion (ANOVA, p < 0.05). Values are expressed as mean \u00C2\u00B1 SD. 88 % 100 \u00E2\u0080\u009E 90 80 Interstitial 7 0 Space 60 50 40 30 20 10 0 70\u00C2\u00B121% 45\u00C2\u00B119% N6nfluid Fluid Control i 63\u00C2\u00B120% 50\u00C2\u00B120% 3 Nonfluid 'Fluid Groups Endotoxin Figure 31. Interstitial Volume in Gut. Area of interstitial space is illustrated for the four groups. Fluid resuscitation significantly increases the area of interstitial space (p > 0.05) however, endotoxin has no significant effect (ANOVA, p > 0.05). Values are expressed as mean \u00C2\u00B1 SD. 89 Table 1. Onset of Ischemia and Extraction Ratios CONTROL ENDOTOXIN Non-Resuscitated (n=8) Fluid-Resuscitated (n=8) Non-Resuscitated (n=8) Fluid-Resuscitated (n=8) WB D 0 2 vs WB V 0 2 Onset of Ischemia - WB D0 2c (ml02/min.kg of WB wt) 9.1 + 1.7 10.3 \u00C2\u00B1 2.2 12.2 \u00C2\u00B1 3.8+ 15.3 + 3.2+ Critical V 0 2 - WB VO zc (ml02/min.kg of WB wt) 7.5 \u00C2\u00B1 1 . 3 6.3 \u00C2\u00B1 1.3 6.5 + 1.5 7.4 \u00C2\u00B1 1 . 2 0 2 Extraction Ratio WB ERc (=WBV02c/WBD02c) 0.82 + 0.06 0.62 \u00C2\u00B1 0 .08* 0.55 \u00C2\u00B1 0 . 0 8 t 0.49 \u00C2\u00B1 0.04+ 0 2 Extraction Ratio WB ERmax (02ER at final measurement) 0.86 \u00C2\u00B1 0 .03 0.77 \u00C2\u00B1 0 .08* 0.74 + 0.10 0.81 + 0.11+ GUT D 0 2 vs GUT V 0 2 Onset of Ischemia - GutD02c (ml02/min.kg of gut wt) 25.3 \u00C2\u00B1 14.8 20 .5 \u00C2\u00B1 6.8 24 .3 + 9.0 22 .9 \u00C2\u00B1 11.7 Critical V O z - Gut VO zc (ml02/min.kg of gut wt) 19.6 \u00C2\u00B1 12.0 14.0 \u00C2\u00B1 5.6 12.9 + 6.0 11.8 \u00C2\u00B1 5.2 0 2 Extraction Ratio Gut ERc (=GutV02c/GutD02c) 0.77 + 0 .07 0.67 \u00C2\u00B1 0 .06* 0.52 \u00C2\u00B1 0 .05 t 0.52 + 0.09+ 0 2 Extraction Ratio Gut ERmax (02ER at final measurement) 0.87 + 0.03 0.72 + 0.08* 0.68 + 0.07 0.65 + 0 .08 t WB D 0 2 vs GUT V 0 2 Onset of Ischemia - GutD02c (ml02/min.kg of gut wt) 10.6 \u00C2\u00B1 4.4 13.6 \u00C2\u00B1 6.1 13.5 + 7.1 15.3 \u00C2\u00B1 5 . 9 Critical VO z - Gut VO zc (ml02/min.kg of gut wt) 16.7 \u00C2\u00B1 11.7 14.0 \u00C2\u00B1 5.1 10.2 \u00C2\u00B1 5.7 14.0 \u00C2\u00B1 6.2 * indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in control groups indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in endotoxin groups t indicates difference between the control and endotoxin groups (p<0.05) Arterial and Capillary Hematocrit Arterial Hct(%) @ Baseline Arterial Hct(%) @ Onset of Ischemia Capillary Hct(%) Onset oMschemia CONTROL Non-Resus. 23.0 \u00C2\u00B1 3.2 25.0 \u00C2\u00B1 3.5 27.5 \u00C2\u00B1 8.6 Fluid-Resus. 22.0 \u00C2\u00B1 3.3 16.1 \u00C2\u00B1 3.5*+ 29.2 \u00C2\u00B1 14.3 ENDOTOXIN Non-Resus. 24.0 \u00C2\u00B1 3.5 24.4 \u00C2\u00B1 4.2 25.3 \u00C2\u00B1 7.7 Fluid-Resus. 22.6 \u00C2\u00B1 2.6 19.8 \u00C2\u00B1 3.8**+ 25.3 \u00C2\u00B1 5.3 * indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in control groups indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in endotoxin groups + indicates difference between Arterial Hematocrit at Baseline and at Onset of Ischemia (p<0.05) Table 3. Arterial and Portal Vein Leukocyte Counts Arterial WBC Baseline Arterial WBC Critical Portal WBC Baseline Portal WBC Critical CONTROL Non-Resus. 13.0 \u00C2\u00B1 3.8 25.0 \u00C2\u00B1 11.7* 12.0 \u00C2\u00B1 4.2 25.2 \u00C2\u00B1 12.1** Fluid-Resus. 10.4 \u00C2\u00B1 7.4 14.9 + 10.5*\u00C2\u00A5 8.8 \u00C2\u00B1 6.6 14.9 \u00C2\u00B1 9.9**\u00C2\u00A5 ENDOTOXIN Non-Resus. 10.1 \u00C2\u00B1 3.6 2.2 + 0.5*t 9.0 \u00C2\u00B1 3.3 2.3 + 0.8**+ Fluid-Resus. 10.5 \u00C2\u00B1 7.9 1.7 \u00C2\u00B1 0.9*+ 9.6 \u00C2\u00B1 7.3 2.0 \u00C2\u00B1 1.4**+ WBC are expressed as leukocytes X 109/L * indicates difference between Baseline and Critical Arterial WBC (p<0.05) ** indicates difference between Baseline and Critical Portal WBC (p<0.05) + indicates difference between control and endotoxin groups (p<0.05) \u00C2\u00A5 indicates difference between Non-Resuscitated and Fluid-Resuscitated (p<0.05) Table 4. Intestinal pHi at Baseline and Onset of Ischemia pHi at Baseline pHi at Onset of Ischemia CONTROL Non-Resus. 7.25 \u00C2\u00B1 0.09 7.06 \u00C2\u00B1 0.10* Fluid-Resus. 7.19 \u00C2\u00B1 0.09 7.07 + 0.11* ENDOTOXIN Non-Resus. 7.21 + 0.10 6.91 \u00C2\u00B1 0.15*+ Fluid-Resus. 7.12 \u00C2\u00B1 0.26 6.90 + 0.24*+ * indicates difference between pHi at Baseline and Onset of Ischemia (p< 0.05) + indicates difference between pHi between control and endotoxin (p < 0.05) Table 5. Arterial Lactate Values Art. Lactate CONTROL Non-Resus. Initial 2.51 \u00C2\u00B1 1.06 Critical 3.94 \u00C2\u00B1 1.24 Fluid-Resus. Initial 1.76 \u00C2\u00B1 0.70 Critical 2.41 \u00C2\u00B1 0.58* ENDOTOXIN Non-Resus. Initial 1.93 \u00C2\u00B1 0.63 Critical 3.20 \u00C2\u00B1 0.97 Fluid-Resus. Initial 1.53 + 0.50 Critical 3.07 \u00C2\u00B1 1.07 Lactate is expressed as mmol/L * indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in control groups indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in endotoxin groups + indicates difference between the control and endotoxin groups (p<0.05) Table 6. Onset of Rise of Arterial Lactate CONTROL ENDOTOXIN Non-Resuscitated (n=8) Fluid-Resuscitated (n=8) Non-Resuscitated (n=8) Fluid-Resuscitated (n=8) WB DO z vs Arterial Lactate Onset of Ischemia - WB D0 2c (ml02/min.kg of WB wt) 13.1 \u00C2\u00B1 3.3 12.0 \u00C2\u00B1 4.7 14.9 \u00C2\u00B1 5.7 12.3 \u00C2\u00B1 0.8 U l * indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in control groups indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in endotoxin groups + indicates difference between the control and endotoxin groups (p<0.05) Table 7. Oxygen Transport Variables Pa0 2 PaCO z Arterial pH Art. Hgb ArtOzcontent Temp (mmHg) (mmHg) (g/D (g o2) (\u00C2\u00B0C) CONTROL Non-Resus. Initial 151.0 \u00C2\u00B1 20.1 29.9 \u00C2\u00B1 2.9 7.46 \u00C2\u00B1 0.05 77.3 \u00C2\u00B1 10.0 103.5 \u00C2\u00B1 12.3 37.0 \u00C2\u00B1 0.8 Critical 134.6 \u00C2\u00B1 19.5 32.1 \u00C2\u00B1 4.4 7.40 \u00C2\u00B1 0.06 83.6 \u00C2\u00B1 13.0 109.7 \u00C2\u00B1 16.0 38.4 \u00C2\u00B1 0.5 Fluid-Resus. Initial 137.4 \u00C2\u00B1 31.9 31.2 + 3.9 7.41 \u00C2\u00B1 0.05 73.9 \u00C2\u00B1 9.8 98.3 \u00C2\u00B111.9 36.9 \u00C2\u00B1 0.7 Critical 109.4 \u00C2\u00B1 38.6 36.0 \u00C2\u00B1 11.0 7.28 \u00C2\u00B1 0.08* 55.0 \u00C2\u00B1 11.2* 68.6 +16.1* 37.9 \u00C2\u00B1 0.8 ENDOTOXIN Non-Resus. Initial 159.0 \u00C2\u00B1 30.3 31.2 + 6.1 7.45 \u00C2\u00B1 0.05 78.9 \u00C2\u00B1 11.1 107.0 \u00C2\u00B1 14.5 38.0 \u00C2\u00B1 1.2 Critical 107.2 \u00C2\u00B1 35.8 35.3 \u00C2\u00B1 9.9 7.30 \u00C2\u00B1 0.07 83.4 \u00C2\u00B1 17.5 105.2 \u00C2\u00B1 27.5 38.7 \u00C2\u00B1 0.7 Fluid-Resus. Initial 129.7 \u00C2\u00B1 36.7 33.5 + 4.3 7.43 \u00C2\u00B1 0.05 79.1 + 13.5 102.91 \u00C2\u00B1 17.2 37.2 + 0.5 Critical 74.3 \u00C2\u00B1 19.9**+ 43.8 \u00C2\u00B1 5.8** 7.21 \u00C2\u00B1 0.08**+ 67.9 \u00C2\u00B1 10.1**+ 76.5 \u00C2\u00B1 15.6** 38.5 \u00C2\u00B1 0.5 * indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in control groups indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in endotoxin groups + indicates difference between the control and endotoxin groups (p<0.05) Table 8. Hemodynamic Variables Wedge Pressure (mmHg) Mean Art.Pres. (mmHg) Cardiac Output (L/min) Sys.Vasc.Res. (dynes/sec/cm5) Portal Flow (L/min) CONTROL Non-Resus. Initial 6.17 \u00C2\u00B1 2.36 77.2 \u00C2\u00B1 13.8 7.79 \u00C2\u00B1 1.93 784.30 \u00C2\u00B1 256.95 0.75 \u00C2\u00B1 0.17 Critical 4.32 \u00C2\u00B1 1.78 59.2 \u00C2\u00B1 10.6 2.76 \u00C2\u00B1 0.68 1651.76 \u00C2\u00B1 407.98 0.32 + 0.21 Fluid-Resus. Initial 7.94 \u00C2\u00B1 1.12 83.3 \u00C2\u00B1 10.4 8.23 \u00C2\u00B1 3.60 882.37 \u00C2\u00B1 407.10 0.86 \u00C2\u00B1 0.22 Critical 6.89 \u00C2\u00B1 1.10* 76.7 \u00C2\u00B1 11.6* 5.74 \u00C2\u00B1 2.67* 1183.38 \u00C2\u00B1 604.19* 0.66 \u00C2\u00B1 0.32* ENDOTOXIN Non-Resus. Initial 8.19 \u00C2\u00B1 2.37 75.7 \u00C2\u00B1 19.1 6.65 \u00C2\u00B1 2.45 833.16 \u00C2\u00B1 483.85 0.70 \u00C2\u00B1 0.24 Critical 7.71 \u00C2\u00B1 2.45+ 43.5 \u00C2\u00B1 13.9+ 3.64 \u00C2\u00B1 1.07 515.90 \u00C2\u00B1 375.47+ 0.32 \u00C2\u00B1 0.14 Fluid-Resus. Initial 9.13 \u00C2\u00B11.74 83.5 \u00C2\u00B1 13.9 6.78 + 1.93 920.24 \u00C2\u00B1 412.09 0.73 \u00C2\u00B1 0.26 Critical 9.49 + 1.91+ 49.3 \u00C2\u00B1 9.6+ 5.90 \u00C2\u00B1 2.11** 380.14 \u00C2\u00B1 415.89+ 0.53 \u00C2\u00B1 0.35 * indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in control groups indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in endotoxin groups + indicates difference between the control and endotoxin groups (p<0.05) Table 9. Clinical Variables Animal wts (kg) Total fluid (cc/kg) Total urine (cc/hr) Time of survival (min) CONTROL Non-Resuscitated 25.4 \u00C2\u00B1 1.8 69 \u00C2\u00B1 23 59 \u00C2\u00B130 202 + 60 Fluid-Resuscitated 27.6 \u00C2\u00B1 3.9 353 \u00C2\u00B1 80* 267 \u00C2\u00B1 160* 368 +105* ENDOTOXIN Non-Resuscitated 24.3 + 1.5 89 \u00C2\u00B1 18 145 + 93 128 \u00C2\u00B1 43+ Fluid-Resuscitated 25.8 \u00C2\u00B1 2.9 161 \u00C2\u00B1 61**+ 178 \u00C2\u00B1 115t 171 \u00C2\u00B1 81+ 0 0 * indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in control groups ** indicates difference between Non-fluid and Fluid-resuscitated groups (p<0.05) in endotoxin groups + indicates difference between the control and endotoxin groups (p<0.05) 5.4 D ISCUSSION The major finding from our study is that fluid resuscitation significantly impairs whole body O 2 extraction in control pigs but d id not significantly change the already decreased O 2 extraction in the endotoxin treated animals. Impairment of O2 extraction is associated wi th increased interstitial edema and increased heterogeneity of transit times but not to altered capillary hematocrit. Sepsis has substantial effects on the microcirculation that may account for impaired tissue oxygen extraction (74) (126). The host defense response against invading microorganisms can cause the development of the systemic inflammatory response syndrome (SIRS) (6). Endotoxin and other bacterial products trigger macrophages and other inflammatory cells to release TNF-oc, IL-1, IL-6 and many other pro-inflammatory mediators (166) (6). These endogenous mediators of the septic inflammatory response have important effects on the microvasculature including leukocyte slowing and microthrombi (168) (52) (4i) that may lead to decreased capillary flows, increased capillary permeability - that is, \"leaky capillaries\" (52) (12) wi th the resultant interstitial edema (136) and, an alteration in vascular tone (76) (74) (169). Altered microcirculatory regulation of blood flow may lead to mismatching of oxygen supply to demand resulting in impaired tissue oxygen extraction (78). L a m et al. demonstrate that after cecal ligation and puncture in rats, there was a 36% reduction in perfused capillary density wi th an increase in heterogeneity of the spatial distribution of perfused capillaries (57). These results are similar to the observations by Humer et al. that the relative dispersion of gut capillary blood flow 99 increases after endotoxin infusion in pigs and this was associated wi th impaired oxygen extraction (74). Some time ago, Honig and Odoroff suggested that the dispersion of capillary transit times would have an impact on oxygen exchange by the capillary bed (170). A subsequent theoretical analysis suggested that an increased dispersion of capillary transit times would impair oxygen exchange by increasing mismatch between oxygen supply and demand wi th in small regions of a capillary bed (78). Conversely, Morff has suggested that increased capillary recruitment improves tissue oxygenation in rat cremaster muscle (171). A number of investigators have suggested that to improve microvascular flow, reduced hematocrits may be useful (111) (125). Hemodi lu t ion decreases blood viscosity, may improve capillary red blood cell flux (125), and may improve blood flow distribution wi th in the capillary bed (125) (126). V a n der Linden et al found that hemodilut ion to a hematocrit of 20 or 30% using colloid infusion resulting i n an increased critical oxygen extraction ratio during progressive hemorrhage compared to a hematocrit of 40% (125). Tyml has demonstrated that the heterogeneity of microvascular flow in rat skeletal muscle is reduced by hemodilut ion (126). However, while some argue that microvascular flow distribution and tissue oxygen extraction may be improved by fluid resuscitation, others argue that fluid resuscitation may lead to interstitial and endothelial edema (136) which may further impair microvascular flow and lead to a higher degree of capillary heterogeneity. Therefore, our hypothesis was that not only w o u l d endotoxemia increase capillary heterogeneity, but that fluid resuscitation may also increase capillary heterogeneity and thereby impair oxygen extraction. 100 As per the hypothesis, the author found that fluid resuscitation resulted i n impaired O2 extraction associated increased heterogeneity of transit times. In addition, fluid resuscitation resulted in increased interstitial edema. Therefore, a possible mechanism for impaired O 2 extraction associated wi th fluid resuscitation is increased heterogeneity of transit times possibly resulting from decreased capillary recruitment secondary to altered morphology of the capillary bed due to interstitial edema. Since sepsis is characterized by a generalized increase in capillary permeability (52) (12), tissues are predisposed to development of interstitial edema (136). Since crystalloid resuscitation is known to redistribute according to Starling's forces, one wou ld anticipate that interstitial edema would be more prone to occur in sepsis. Further, an increased O2 diffusion distance associated wi th interstitial edema could by itself result in decreased interstitial O2 tension (56), but whether O2 uptake is affected by interstitial edema is unsupported in studies by others (172). A key new observation in the current study is that fluid resuscitation significantly increases the relative dispersion of blood flow transit times throughout the gut wal l . This may indicate an increased relative dispersion of blood flow transit times in the capillary bed as wel l as wi th in the larger arterioles and venules w i t h i n the gut wal l . Conceivably, the resultant decrease in capillary diameter due to endothelial edema may impair microvascular blood flow or contribute to leukocyte retention and plugging wi th in the capillary bed (164). In addition, red blood cell rheology may be altered in the septic state (49) and conceivably could be altered by fluid resuscitation resulting in more heterogeneous microvascular flow. As a 101 result, impaired oxygen extraction may be accounted for by mismatching of oxygen supply to demand. Thus, the potential detrimental effects of fluid resuscitation appear to contribute to overall impaired oxygen extraction in the current set of experiments. In our previous study (74) impaired O2 extraction in endotoxin pigs was associated wi th increased heterogeneity of transit times. In that study endotoxin pigs also received fluid resuscitation whereas controls did not. Similarly, in this study impaired O 2 extraction associated wi th increased heterogeneity of transit times was observed in endotoxic pigs which received fluid resuscitation. But impaired O2 extraction in endotoxic pigs was not associated wi th increased heterogeneity of transit times without fluid resuscitation. Therefore, increased heterogeneity of transit times is not the sole mechanism of impaired O2 extraction i n endotoxemia. Similarly, interstitial edema was increased in association wi th fluid resuscitation rather than wi th endotoxemia. Therefore, while interstitial edema wi th increased heterogeneity of transit times could result in impaired O2 extraction from f lu id resuscitation this is not sufficient to account for impaired O2 extraction from endotoxemia. One of the side-effects of massive saline infusion is the development of a metabolic acidosis (137). Acidosis is known to alter the oxygen dissociation curve by shifting the curve to the right. This would lead to a higher P50 and therefore a reduced affinity of hemoglobin for oxygen; thus hemoglobin releases oxygen at higher P O 2 , which would tend to raise tissue PO2. Therefore, while saline infus ion 102 may have shifted the dissociation curve, it would have done so in favor of improved tissue oxygenation. This would not explain the impaired oxygen extraction that was found. Capillary transit times are more closely related to O2 extraction than transit times in the whole gut wal l . Capillary transit times was not measured in this study but in a study by Humer et al (74), capillary transit times accounted for about one-half of whole gut wal l transit times. In that study altered heterogeneity of capillary transit times associated wi th endotoxemia and fluid resuscitation was paralleled by altered heterogeneity in whole gut wal l transit times. Whi le it is possible that increased heterogeneity of capillary transit times wi th unchanged heterogeneity of whole gut wal l transit times could occur in endotoxic pigs which did not receive fluid resuscitation, one does not have basis for suggesting that this possibility could be a mechanism of impaired O2 extraction in endotoxic non-fluid resuscitated animals. Interestingly, the author notes that although arterial hematocrit decreased as anticipated wi th saline resuscitation, microvascular hematocrit determined using morphometric measurements was not altered by saline resuscitation. Thus, the putative beneficial effect of crystalloid resuscitation on blood viscosity may not extend to the microvasculature. Furthermore, hemodilut ion does not explain the decreased oxygen extraction seen with fluid resuscitation. However, it is important to point out that the technique of determining capillary hematocrits only included those vessels that had erythrocytes present. As mentioned, there is heterogeneity of blood flow which implies that there w i l l be some vessels which were not perfused at 103 all and therefore, these would not have been counted as part of the morphometr ic analysis. The leukocyte counts in the endotoxemic groups showed a dramatic decrease compared to controls. Since the effect was seen in not only the arterial but also the portal vein samples, one would anticipate not only systemic but splanchnic trapping of leukocytes. Indeed Barroso-Aranda et al (168) showed that endotoxemia is associated wi th activation of polymorphonuclear leukocytes such that it results i n microvascular trapping. They further have shown that obstruction occurs exclusively at the capillary level without involvement of arterioles or venules and suggest that leukocyte capillary plugging may be an important contributor in the fatal outcome after endotoxin administration. Indeed capillary plugging may explain the heterogeneity in capillary blood flow described by some investigators (168) (74) (57) which is thought to lead to the impaired oxygen extraction by tissue (78). Tonometry has been used by many as an indicator of mucosal ischemia (148) (150) (151) (152). Our results show that tonometry can be used to indicate changes i n mucosal acidosis and showed that endotoxemic animals have a larger degree of mucosal acidosis at onset of ischemia. Furthermore, some have used tonometry to indicate onset of ischemia (154) (155). Whi le the author attempted to perform dual line regression between p H i values and oxygen delivery, there were not enough p H i data points to obtain useful numbers to determine onset of ischemia. Whi le more data points would have been ideal, the short period of time of the entire experiment wou ld have required a shortening of the time interval which would perhaps not 104 allow enough time to adequately equilibrate the gases wi th the saline in the tonometric balloon. Whi le many investigators use lactate as a marker of onset of ischemia (74) (106), others consider lactate is a difficult value to interpret (55). Whi le the arterial lactate levels are higher in at onset of ischemia, there appears to be too m u c h variability to make these differences significant. Further, our values show that rise in lactate as determined by a dual line regression wi th whole body oxygen delivery corresponds to the onset of ischemia as determined wi th a dual line regression between whole body oxygen delivery and consumption. The only group not to correspond was that of the control non-fluid group. The reasons for this difference are not clear. Possible reasons include an aerobic source of lactate production, early anaerobic production by some tissues, or an inability to metabolize lactate in this group (156). Whi le current recommendations include early and aggressive use of crystalloids in sepsis, the data suggests that saline resuscitation does not improve the oxygen extraction defect seen in sepsis. Indeed, the saline was shown to lead to increased interstitial edema and a maldistribution of blood flow which may contribute to the oxygen extraction defect. Therefore, the author questions the use of crystalloid resuscitation for treatment of sepsis in humans. 105 Chapter 6 CONCLUSIONS A review of the current literature on the microvascular changes associated wi th sepsis and endotoxemia lead us to our hypothesis that crystalloid resuscitation would impair the ability of tissue to extract oxygen. We found that there is a significant impairment in oxygen extraction capabilities wi th the addition of crystalloid resuscitation which would therefore support our hypothesis. Potential reasons for this impairment were investigated. First, we studied the potential role of the described increase in capillary permeability leading to interstitial edema. W e found an increase in interstitial volume by use of morphometric analysis. This wou ld support the theory that interstitial edema may be a reason for the impaired ability to extract oxygen. Further, alterations in capillary distribution has been theorized as a potential reason for decreased oxygen extraction in shock states. W e found there was an increase in blood flow heterogeneity by use of microspheres as markers of blood flow. This effect was seen not only in endotoxemia but wi th f lu id resuscitation. Therefore, an alteration in the capillary flow either by interstitial fluid hindering capillary flow or by fluid impairing endothelial control may be possible. Finally, some reports suggest that fluid resuscitation may lead to hemodilut ion at a capillary level which may thereby decrease oxygen delivery to tissue beds. W e found that while arterial hematocrit is significantly reduced wi th saline loading, capillary hematocrit was not altered compared to baseline arterial values. Therefore, an alteration in capillary hematocrit would not explain the decreased oxygen 106 extraction found for the gut and whole body. Thus, while aggressive and early saline resuscitation is considered by many to be a fundamental therapeutic maneuver in sepsis, our study has shown that this therapy leads to significant impairment in oxygen extraction and thereby may prove detrimental. 107 Chapter 7 SUMMARY: CLINICAL RELEVANCE, FUTURE DIRECTIONS The ideal therapy for sepsis has still to be defined. There are numerous levels at which therapy may be initiated as was described earlier. We sought to determine the effects of saline resuscitation in our model and found a significant impai rment in oxygen extraction. We based our findings on the likely development of interstitial edema as wel l as an impairment in the microvascular distribution leading to capillary heterogeneity. We would anticipate that if these findings were to be similar to what one sees clinically, that saline resuscitation may worsen an already critical situation. There are several areas of possible further study. Areas would include further studies in the current project, logical follow-up studies, and finally likely future endeavors which may prove beneficial in the therapy of sepsis. First, there may be other areas of investigations in the current study w h i c h may shed further light into the mechanisms of impaired oxygen extraction. W h i l e we have shown interstitial edema and heterogeneity of total gut blood flow, there are other aspects which have not yet been studied. Whi le we have looked at the total gut blood flow transit times, we have not yet subdivided these into large, medium, and small vessel as per Humer et al (74). This may give further information as to which vessels are contributing to the heterogeneity. Furthermore, while we have demonstrated heterogeneity in transit times, there are other mechanisms for impaired oxygen extraction which might be investigated. A s described earlier, leukocytes have been shown, to cause capillary obstruction dur ing 108 sepsis (65) (41) (66) which may contribute to vascular obstruction. U s i n g morphometric techniques, one may attempt to quantify the number of leukocytes present in capillaries to determine whether this is a contributing factor in the impaired oxygen extraction in endotoxin and saline resuscitated animals. Further, not only has development of capillary heterogeneity been shown to decrease oxygen extraction, but an alteration in the capillary density may also be involved (57). A l s o wi th morphometric techniques, the number of capillaries may be quantified. Another technique involves the use of intravital microscopy as has been used to study the microvasculature by others (126) (173). One would expect that increased interstitial edema would lead to decreased number of open capillaries and therefore a decreased capillary density (34). Finally, some investigators have theorized that the inability to utilize oxygen may be an important factor in the oxygen extraction problem seen in sepsis. Mechanisms include uncoupling oxidative phosphorylation (14) (15), inhibit ion of mitochondrial respiration (16) (17) (18) (17) and mitochondrial destruction (54). One may look at the possibility of mitochondrial destruction as a contributing factor to impaired oxygen extraction. Hersch et al (54) utilized electron microscopy to analyze mitochondrial features and describe significant mitochondrial destruction in a sheep model of sepsis. Whi le a l l these studies would add further information to our current study, they go no further in suggesting what other therapies may be more successful in i m p r o v i n g oxygen extraction in sepsis. Follow-up studies based on our current investigation includes using the same model wi th different types of fluid resuscitation. The aim would be to preserve the 109 beneficial effects of fluid resuscitation but to avoid the side-effects demonstrated i n our current study. Possible fluid include other crystalloids and the colloids. A s for crystalloids, two other common fluids used include Ringers Lactate and hypertonic saline. W i t h respects to Ringers Lactate (RL), advantages of its use wou ld include less development of hyperchloremic acidosis in view of the decreased concentration of sodium chloride as wel l as the buffering action of sodium lactate. However, since lactate requires conversion by the liver to pyruvate and then to bicarbonate to act as a buffer, there may be concern during sepsis where the liver has been shown to have decreased function that lactate may accumulate (174). The other potential crystalloid to study would be hypertonic saline. Because sodium is primari ly an extracellular ion, infusion of hypertonic saline would be expected to expand the extracellular fluid (ECF) space by a greater amount than the volume infused, because theoretically, water should enter the E C F from the intracellular fluid (128). Further, there may be potential positive inotropic effects as wel l as systemic and pulmonary vasodilation (128). Horton and Walker (175) have studied hypertonic saline (HS) i n a canine model of endotoxemia. They have shown that the addition of H S to R L reduces the total volume of R L required to maintain hemodynamic variables. Further, the net fluid gain was five times less than wi th R L alone. This w o u l d potentially l imit the degree of interstitial edema that accompanies most crystalloid fluid resuscitation as occurred in our study. Despite the increased chloride content, it has not been shown to lead to hyperchloremic acidosis as does normal saline (174). However, if HS is rapidly infused, it may precipitate pontine myelinolysis (174) (128). Whi le other crystalloids may be a potential area of investigation, 110 another area wi th great potential is the use of colloids. Sepsis has been characterized by a diffuse increase in microvascular permeability (46) (52) (12). Thus, the use of colloids which have been shown to stay in the intravascular space for longer durations (118) would seem obvious. A study invo lv ing in v ivo microscopy and surface oxygen partial pressure electrodes by Funk et al (173) was performed to compare crystalloids to colloids. They showed that colloids had improved capillary perfusion and tissue oxygenation compared to crystalloid resuscitation. Further, sepsis also is accompanied by a reprioritization of hepatic protein synthesis w i t h decreased albumin production which leads to a decreased colloid oncotic pressure (130). This would , according to Starling's forces, add to the leak of fluid out of capillaries leading to pulmonary edema (176) (177) (178) (179) and interstitial edema (180) (178). Morisaki et al (181) showed that when compared to crystalloid, col loid therapy decreases the progression of extrapulmonary tissue injury in septic sheep and based this on the preservation of microvascular surface area for tissue O2 exchange. The different types of colloids include albumin and hydroxyethyl starch. Some important properties of albumin include antioxidant activity, binding free fatty acids and endotoxin (182) (128). Because heat treating albumin at 60\u00C2\u00B0C for 10 hours inactivates hepatitis virus and other infectious agents, there is no concern regarding infections wi th this colloid (118). Adverse reactions to albumin are rare (137). Intravenous administered albumin distributes init ially to the intravascular space, but gradually redistributes to the interstitial space. Its intravascular half-life is 16 hours, longer than crystalloids and similar to endogeneous albumin (128). Hydroxyethyl starches (HES) are synthetic colloid derivatives from corn starch. The 111 agent is prepared by incorporating hydroxyethyl ether into the glucose residues of amylopectin (174). They have been shown to effectively expand plasma volume by 70% at 3 hours and 40% by 12 hours after infusion (174). Zikr ia et al (183) have suggested that one of the beneficial roles of HES is to alter the microvascular permeability of an ischemic limb, thereby reducing the development of interstitial edema and eventual compression ischemia. Hydroxyethyl starches have been shown to have anaphylactic reactions in 0.8% as wel l as decreases in factor VIII activity and prolongation of partial thromboplastin time (137) however there have been no increased frequency of bleeding reported. Concerns about the possible immunosuppressive effect of hydroxyethyl starch have also been raised (184). Higher molecular weight particles of hydroxyethyl starch are deposited in the reticuloendothelial system and may affect phagocytic function. However, in a study by Shatney and Chaudry (185) demonstrated that reticuloendothelial clearance rates of labeled l ip id emulsions were unchanged after hydroxyethyl starch, and similar ly, mortality rates from peritonitis were not altered in animals receiving hydroxyethyl starch. Pentastarch is a modification of the hetastarch formulation as it is has a lower average molecular weight, a more homogeneous size of particles, and less hydroxyethyl substitution. These changes allow for a more predictable excretion of pentastarch compared wi th hetastarch. Its volume-expanding effect lasts 12 hours. Because of its higher colloid oncotic pressure (about 40 m m Hg), it has a greater degree of volume expansion (1.5 times the volume infused) than either 5% albumin or 6% hetastarch (128). Further, it appears to have less effect on coagulation than hetastarch. 112 Thus, wi th the potential benefits of colloids over crystalloids, and wi th the advantages of pentaspan over some other colloids, it places pentaspan as the most rationale fluid to study. One may study the effects in a similar experiment as described wi th use of pentaspan as the resuscitative fluid and study the oxygen extraction at onset of ischemia. Further, by studying the interstitial volume as an indicator of interstitial volume, one may determine whether colloids improve oxygen delivery by limiting interstitial edema. While the use of fluid resuscitation to augment oxygen delivery is directed a relatively phase of SIRS, others have studied novel approaches at earlier stages i n the inflammatory response. These therapies may be divided into specific and nonspecific. Specific therapies include monoclonal antibody derived against gram-negative bacterial endotoxin has been tested in septic patients in a multicenter, prospective, randomized, double bl ind fashion (186). Adminis t ra t ion of this antibody to I C U patients wi th suspected gram-negative sepsis early i n their septic course resulted in a significant reduction in the mortality in the group who proved to be bacteremic on blood culture. However, when looking at the entire group, there was no significant reduction in mortality. Polymyxin B is another compound which has the property of efficiently binding endotoxin (187). Further, a po lymyxin B-impregnated filter is being developed as part of hemofiltration systems to chelate circulating endotoxin (182). Trials are also underway on monoclonal antibodies designed to interfere wi th neutrophil adhesion (anti-CD18 or CD11), IL-1 receptor, T N F , and P A F (22). M u c h work needs to be done not only on its potential effectiveness but the timing, potential adverse effects, and the specificity of therapy. 113 Less specific therapies include eicosanoid inhibitors. Increased activity of phospholipase A 2 , an important enzyme in the release of arachidonic acid f rom membrane phospholipids, is correlated wi th the severity of septic shock (188). Blocking the synthesis or activity of these substances results in increased survival i n experimental models (189). Other nonspecific therapies being studied include free radical scavengers, xanthine oxidase inhibitors, and superoxide dismutase to name a few (182). Whether the crystalloids, colloids, or the novel approaches w i l l have potential in reducing the overwhelming mortality rates remains to be seen. However, it is unlikely that any one therapy w i l l be the \"magic bullet\" that everyone seeks. 114 R E F E R E N C E S 1. Rackow EC, Ast iz M E . Mechanisms and Management of Septic Shock. Crit Care C l i n 1993;9:219-237. 2. Parrillo JE, Parker M M , Natanson C, et al. Septic Shock in Humans : Advances in the Understanding of Pathogenesis, Cardiovascular Dysfunction, and Therapy. A n n Int M e d 1990;113:227-242. 3. Carrico DJ, Meakins JL, Marshall JC, et al. Mul t ip le organ failure syndrome. A n n Surg 1986;121:196-208. 4. Light RB. Sepsis Syndrome. In: H a l l JB SG, Wood L D H , ed. Principles of Critical Care. N e w York: McGraw H i l l , 1992:645-655. 5. Ketai L . F lu id Loading in Septic Shock. A n n Int M e d 1990;113:991. 6. Bone R C , Balk R A , Cerra FB, et al. Definitions for Sepsis and Organ Failure and Guidelines for the Use of Innovative Therapies in Sepsis. Chest 1992;101:1644-1655. 7. Baker C C , H u y n h T. Sepsis in the Critically 111 Patient. Curr Probl Surg 1995;32:1018-1083. 8. Fink M P , Heard SO. Current Research Review: Laboratory Models of Sepsis and Septic Shock. J Surg Res 1990;49:186-196. 9. Morrison D C , Ryan JL. Endotoxins and disease mechanisms. A n n u Rev M e d 1987;38:417-432. 10. Morr i son D C , Ulevi tch RJ. The effects of bacterial endotoxins on host mediation systems. A m J Pathol 1987;93:527-618. 11. Dinarello C. The proinflammatory cytokines interleukin-1 and tumor necrosis factor and treatment of septic shock. J Infect Dis 1991;163:1177-1184. 12. Tuchschmidt J, Obltas D, Fried J C Oxygen consumption in sepsis and septic shock. Crit Care M e d 1991;19:664-671. 13. VanderMeer TJ, Wang H , Fink M P . Endotoxemia causes ileal mucosal acidosis in the absence of mucosal hypoxia in a normodynamic porcine model of septic shock. Crit Care M e d 1995;23:1217-1226. 14. Spitzer JA, Deaciuc IV. Effect of endotoxicosis and sepsis on intracellular calcium homeostasis in rat liver. Circ Shock 1986;18:81-93. 115 15. Tavakoli H , Mela L . Alterations of mitochondrial metabolism and protein concentrations in subacute septicemia. Infect Immun 1982;38:536-541. 16. Stadler J, Bill iar TR, Curran RD. Effect of exogenous and endogenous nitric oxide on mitochondrial respiration in rat hepatocytes. A m J Physiol 1991;260:C910-C916. 17. Ast iz M , Rackow EC, W e i l M H , Schumer W. Early Impairment of Oxidative Metabolism and Energy Production in Severe Sepsis. Circ Shock 1988;26:311-320. 18. Simonson SG, Welty-Wolf K , Huang Y-CT, et al. Altered Mi tochondr ia l Redox Responses in Gram Negative Septic Shock in Primates. Circ Shock 1994;43:34-43. 19. Souba W W , Herskowitz K , Klimberg VS. The effect of sepsis and endotoxemia on gut glutamine metabolism. A n n Surg 1990;21:543-551. 20. Deitch E A . Mult ip le organ failure: pathophysiology and potential future therapy. A n n Surg 1992;216:117-134. 21. Unanue ER, A l l e n P M . The basis for the immunoregulatory role of macrophages and other accessory cells. Science 1987;236:551-557. 22. Cipolle M D , Pasquale M D , Cerra FB. Secondary Organ Dysfunction: F r o m Clinical perspective to Molecular Mediators. Crit Care C l i n 1993;9:261-298. 23. Hefl in A C , Jr., Brigham K L . Prevention by Granulocyte Depletion of Increased Vascular Permeability of Sheep Lung following Endotoxemia. J C l i n Invest 1981;68:1253-1260. 24. Mi lesk i WJ. Sepsis: what it is and how to recognize it. Surg C l i n Nor th A m 1991;71:749-763. 25. Okusawa S, Gelfand JA, Ikejima T, et al. Interleukin-1 induces a shock-like state in rabbits. Synergism with tumor necrosis factor and the effect of cyclooxygenase inhibition. J C l i n Invest 1988;81:1162-1170. 26. Fleming I, Gray G A , Julou-Schaeffer G , et al. Incubation wi th endotoxin activates the L-arginine pathway in vascular tissue. Biochem Biophys Res C o m m u n 1990;171:562-570. 27. Julou-Schaeffer G , Gray G A , Parratt JR, et al. Act ivat ion of the L-arginine pathway is involved in vascular hyporeactivity induced by endotoxin. J Cardiovasc Pharmacol 1991;17:S207-S212. 28. Ki lbourn R G , Gross SS, Jubran A , et al. NG-methyl-L-arginine inhibits tumor necrosis factor-induced hypotension: Implications for the involvement of nitric oxide. Proc Nat l Acad Sci U S A 1990;87:3629-3635. 116 29. Deby C , Goutier R. New perspective on the biochemistry of superoxide an ion and the efficiency of superoxide dismutase. Biochem Pharmacol 1990;39:399-405. 30. M c C o r d J M . Oxygen-derived free radicals in postischemic tissue injury. N Engl J M e d 1985;312:159-163. 31. G r u m C M . Tissue oxygenation in low flow states and during hypoxemia. Crit Care M e d 1993;21:S44-S49. 32. G r u m C M , Ragsdale R A , Ketai L H . Plasma xanthine oxidase activity i n patients wi th A R D S . J Crit Care 1987;2:22-26. 33. Ba ldwin SR, Simon. R H , G r u m C M , Ketai L H , Boxer L A , Deval l LJ. Oxidant activity in expired breath of patients wi th adult respiratory distress syndrome. Lancet 1986;1:11-14. 34. Thijs L G , Schneider A J , Groeneveld ABJ . The haemodynamics of septic shock. Intensive Care M e d 1990;16(Suppl 3):S182-S186. 35. Rackow EC, Kaufman BS, Falk JL, et al. Hemodynamic response to f lu id repletion in patients wi th septic shock: Evidence for early depression of cardiac performance. Circ Shock 1987;22:11-22. 36. Ellrodt A G , Riedinger M S , K imch i A , et al. Left ventricular performance i n septic shock: Reversible segmental and global abnormalities. A m Heart J 1985;110:402-409. 37. Parker M M , McCarthy K E , Ognibene FP, Parrillo JE. Right Ventr icular Dysfunction and Dilation, Similar to Left Ventricular Changes, Characterize the Cardiac Depression of Septic Shock in Humans. Chest 1990;97:126-131. 38. ' Nishijma H , Wei l M H , Shubin H , et al. Hemodynamic and metabolic studies on shock associated wi th gram-negative bacteremia. Medicine 1973;52:287-294. 39. Jafri S, Lavine S, Field B, et al. Left ventricular diastolic function in sepsis. Crit Care M e d 1990;18:709-714. 40. Ognibene FP, Parker M M , Natanson C, et al. Depressed left ventricular performance: response to volume infusion in patients wi th sepsis and septic shock.. Chest 1988;93:903-910. 41. Goddard C M , Al l a rd M F , Hogg JC, Walley K R . Myocardial morphometr ic changes related to decreased contractility after endotoxin. A m J Physiol 1996;270(4 Pt 2):H1446-H1452. 42. Baumgartner J, McCutchan J, Melle G, et al. Prevention of Gram-Negat ive shock and death in surgical patients by, antibody of endotoxin core glycolipid. Lancet 1985;2:54-63. 117 43. Natanson C, Eichenholz P, Danner R, et al. Endotoxin and tumor necrosis factor challenges in dogs simulate the cardiovascular profile of human septic shock. J Exp M e d 1989;169:823-832. 44. Parrillo JE, Burch C, Shelhamer J H , et al. A circulating myocardial depressant substance in humans wi th septic shock. J C l i n Invest 1985;76:1539-1553. 45. Reily J, Cunnion R, Burch-Whitman C, et al. A circulating myocardial depressant substance is associated wi th cardiac dysfunction and peripheral hypoperfusion in patients wi th septic shock. Chest 1989;95:1072-1080. 46. Hinshaw L B . Sepsis/septic shock: participation of the microcirculation: an abbreviated review. Crit Care M e d 1996;24:1072-1078. 47. Shoemaker W C , Appe l PL, Kram H B , Bishop M , Abraham E. Hemodynamic and Oxygen Transport Moni tor ing To Titrate Therapy in Septic Shock. N e w Horizons 1993;1:145-159. 48. Driessen G K , Haist C W M , Heidtmann H , et al. Effect of reduced red cell \"deformability\" on flow velocity in capillaries of rat mesentery. Pflugers A r c h 1980;388:75-78. 49. Powel l RJ, Machiedo G W , Rush BF, Jr. Decreased Red Blood Ce l l Deformability and Impaired Oxygen Uti l izat ion during H u m a n Sepsis. A m Surg 1993;59:65-68. 50. Bellary S, Arden W W , Schwartz R W , et al. Effect of lipopolysaccharide, leukocytes and monoclonal anti-lipid A antibodies on erythrocyte membrane elastance. Shock 1995;3:132-136. 51. Chien S. Rheology in the microcirculation in normal and low flow states. Circ Shock 1982;8:71-80. 52. Fink M P , Cohn S M , Lee PC, et al. Effect of lipopolysaccharide on intestinal intramucosal hydrogen ion concentration in pigs: Evidence of gut ischemia in a normodynamic model of septic shock. Crit Care M e d 1989;17:641-646. 53. Solomon L A , Hinshaw L B . Effect of endotoxin on isogravimetric capillary pressure in the forelimb. A m J Physiol 1968;214:443-447. 54. Hersch M , Gnidec A A , Bersten A D , Troster M , Rutledge FS, Sibbald W J . Histologic and ultrastructural changes in nonpulmonary organs during early hyperdynamic sepsis. Surgery 1990;107:397-410. 55. Dantzker D. Oxygen Delivery and Uti l izat ion in Sepsis. Crit Care C l i n 1989;5:81-98. 118 56. Heughan C, Ni in isoski J, Hunt TK. Effect of Excessive Infusion of Saline Solution on Tissue Oxygen Transport. Surg Gynecol Obstet 1972;135:257-260. 57. Lam C, Tyml K , Mart in C, Sibbald W. Microvascular perfusion is impaired i n a rat model of normotensive sepsis. J C l i n Invest 1994;94:2077-2083. 58. Knisely M H , Reneau D D , Binely DF. The development and use of equations for predicting the limits on the oxygen supply to the cells of l iv ing tissues and organs. Angiology 1969;20(Suppl.ll):l. 59. Schutzer K - M , Larsson A , Risberg B, et al. Lung protein leakage in feline septic shock. A m Rev. Respir Dis 1993;147:1380. 60. Timesvari P, Abraham GS, Spur CP, et al. Escherichia coli 0111 B4 lipopolysaccharide given intracisternally induces blood-brain barrier opening during experimental neonatal meningitis in piglets. Pediatr Res 1993;34:182-186. 61. Voss BL , DeBault L E , Blick K E , et al. Sequential renal alterations in septic shock in the primate. Circ Shock 1991;33:142-155. 62. Bradley JR, Wilks D, Rubenstein D. The vascular endothelium in septic shock. J infect 1994;28:1-10. 63. Schumacker PT, Samsel R W . Oxygen Delivery and Uptake by Peripheral Tissues: Physiology and Pathophysiology. Crit Care C l i n 1989;5:255-269. 64. Knisely M H , Cowley R A , Hawthorne I, et al. Experimental and cl inical separation of hypovolemic and septic shock. Angiology 1970;21:728-744. 65. Ast iz M E , DeGent GE, L i n RY, et al. Microvascular function and rheologic changes in hyperdynamic sepsis. Crit Care M e d 1995;23:265-271. 66. Schmid-Schonbein G W . Capillary plugging by granulocytes and the no-reflow phenomenon in the microcirculation. Federation Proc 1987;46:2397-2401. 67. Cronenwett JL, Lindenauer S M . Direct measurement of arteriovenous anastomotic blood flow in the septic canine hindlimb. Surgery 1979;85:275-282. 68. Archie JP. Anatomic arterial-venous shunting in endotoxic and septic shock in dogs. A n n Surg 1977;186:171-176. 69. Krzanowski JJ- Vascular endothelium in sepsis and endotoxemia. J Fla M e d Assoc 1994;81:119-122. 70. W y l a m M E , Samsel R W , Umans JG, Mitchel l R W , Leff A R , Schumacker PT. Endotoxin In V i v o Impairs Endothelium-dependent Relaxation of Canine Arteries In Vitro . A m Rev Respir Dis 1990;142:1263-1267. 119 71. Samsel R W , Schumacker FT. Systemic hemorrhage augments local O2 extraction in canine intestine. J A p p l Physiol 1994;77:2291-2298. 72. Parker M M , Shelhamer J H , Bacharach SL, et al. Profound but Reversible Myocardial Depression in Patients wi th Septic Shock. A n n Int M e d 1984;100:483-490. 73. Drazenovic R, Samsel R W , W y l a m M E , Doerschuk C M , Schumaker PT. Regulation of perfused capillary density in canine intestinal mucosa, dur ing endotoxemia. J A p p l Physiol 1992;72:259-265. 74. Humer M F , Phang PT, Friesen BP, Al la rd M F , Goddard C M , Walley K R . Heterogeneity of gut capillary transit times and impaired gut oxygen extraction i n endotoxemic pigs. J A p p l Physiol 1996;81:895-904. 75. Herbertson M J , Werner H A , Russell JA, Iversen K , Walley K R . Myocardial oxygen extraction ratio is decreased during endotoxemia in pigs. J A p p l Physiol 1995;79:479-486. 76. Gutierrez G . Regional blood flow and oxygen transport: Implications for the therapy of the septic patient. Crit Care M e d 1993;21:1263-1264. 77. Nelson DP, Samsel R W , Wood L D H , Schumacker PT. Pathological supply dependence of systemic and intestinal O2 uptake during endotoxemia. J A p p l Physiol 1988;64:2410-2419. 78. Walley K R . Heterogeneity of oxygen delivery impairs oxygen extraction by peripheral tissues: theory. J A p p l Physiol 1996;81:885-894. 79. Zhang H , Vincent J-L. Oxygen Extraction Is Altered by Endotoxin Dur ing Tamponade-Induced Stagnant Hypoxia in the Dog. Circ Shock 1993;40:168-176. 80. Maddaus M A , Wells C L , Simmons RL. Role of cell-mediated immuni ty i n preventing the translocation of intestinal bacteria. Surg Forum 1986;37:107-109. 81. Granger HJ , Nyhof R A . Dynamics of intestinal oxygenation: Interactions between oxygen supply and uptake. A m J Physiol 1982;243:G91-G96. 82. Shepherd A P . Metabolic control on intestinal oxygenation and blood flow. Federation Proc 1982;41:2084-2089. 83. Granger HJ , Shepherd A P , Jr. Intrinsic microvascular control of tissue oxygen delivery. Microvasc Res 1973;5:49-72. 84. Granger HJ , Norris CP. Intrinsic regulation of intestinal oxygenation in the anesthetized dog. A m J Physiol 1980;238:H836-H843. 120 85. Mythen MG, Webb AR. The role of gut mucosal hypoperfusion in the pathogenesis of post-operative organ dysfunction. Intensive Care Med 1994;20:203-209. 86. Nelson DP, King CE, Dodd SL, Schumacker PT, Cain SM. Systemic and intestinal limits of O2 extraction in the dog. J Appl Physiol 1987;63:387-394. 87. Schacterle RD, Adams JM, Ribando RJ. A Theoretical Model of Gas Transport between Arterioles and Tissue. Microvasc Res 1991;41:210-228. 88. Falk A, Redfors S, Myrvold H, Haglund U. Small Intestinal Mucosal Lesions in Feline Septic Shock: A Study on the Pathogenesis. Circ Shock 1985;17:327-337. 89. Shepherd AP, Kiel JW. A model of countercurrent shunting of oxygen in the intestinal villus. Am J Physiol 1992;262:H1136-H1142. J ' 90. Vallett B, Lund N, Curtis SE, Kelly D, Cain SM. Gut and muscle tissue P O 2 in endotoxemic dogs during shock and resuscitation. J Appl Physiol 1994;76:793-800. 91. Wilmore DW, Smith RJ, ODwyer ST, Jacobs DO, Ziegler TR, Wang X-D. The gut: A central organ after surgical stress. Surgery 1988;104:917-923. 92. Takala J, Ruokonen E. Oxygen transport in septic shock. Schweiz med Wschr 1992;122:776-779. 93. Ruokonen E, Takala J, Kari A, Saxen H , Mertsola J, Hansen EJ. Regional blood flow and oxygen transport in septic shock. Crit Care Med 1993;21:1296-1303. 94. Dahn MS, Lange P, Lobdell K, Hans B, Jacobs LA, Mitchell RA. Splanchnic and total body oxygen consumption differences in septic and injured patients. Surgery 1987;101:69-80. 95. Bulkley GB, Kvietys PR, Parks DA, Perry MA, Granger DN. Relationship of Blood Flow and Oxygen Consumption to Ischemic Injury in the Canine Small Intestine. Gastroenterology 1985;89:852-857. 96. Fink MP, Antonsson JB, Wang H , Rothschild HR. Increased Intestinal Permeability in Endotoxic Pigs. Arch Surg 1991;126:211-218. 97. Salzman Al , Wang H , Wollert S, et al. Endotoxin-induced ileal mucosal hyperpermeability in pigs: role of tissue acidosis. Am J Physiol 1994;266:G633-G646. 98. O'Dwyer ST, Michie HR, Ziegler TR, Revhaug A, Smith RJ, Wilmore DW. A Single Dose of Endotoxin Increases Intestinal Permeability in Healthy Humans. Arch Surg 1988;123:1459-1464. 121 99. Rush BF, Jr., Sori AJ, Murphy TF, Smith S, Flanagan JJ, Jr., Machiedo GW. Endotoxemia and Bacteremia During Hemorrhagic Shock: The Link Between Trauma and Sepsis? Ann Surg 1988;207:549-554. 100. Deitch EA, Berg R, Specian R. Endotoxin Promotes the Translocation of Bacteria From the Gut. Arch Surg 1987;122:185-190. 101. Deitch EA, Specian RD, Berg RD. Endotoxin-induced bacterial translocation and mucosal permeability: Role of xanthine oxidase, complement activation, and macrophage products. Crit Care Med 1991;19:785-791. 102. Mainous MR, Tso P, Berg RD, Deitch EA. Studies of the Route, Magnitude, and Time Course of Bacterial Translocation in a Model of Systemic Inflammation. Arch Surg 1991;126:33-37. 103. Schumacker PT, Cain SM. The concept of a critical oxygen delivery. Intensive Care Med 1987;13:223-229. 104. Leach RM, Treacher DF. The Relationship Between Oxygen Delivery and Consumption. Dis Mon 1994;40:306-368. 105. Stein JC, Ellsworth ML. Capillary oxygen transport during severe hypoxia: role of hemoglobin oxygen affinity. J Appl Physiol 1993;75:1601-1607. 106. Nelson DP, Beyer C, Samsel RW, Wood LDH, Schumacker PT. Pathological supply dependence of O2 uptake during bacteremia in dogs. J Appl Physiol 1987;63:1487-1492. 107. Samsel RW, Nelson DP, Sanders WM, Wood LDH, Schumacker PT. Effect of endotoxin on systemic and skeletal muscle O2 extraction. J Appl Physiol 1988;65:1377-1382. 108. Vincent J-L. Oxygen transport in severe sepsis. Acta Anaesthesiol Scand 1993;37(Suppl 98):29-31. 109. Kreymann G, Grosser S, Buggisch P, Gottschall C, Matthaei S, Greten H. Oxygen consumption and resting metabolic rate in sepsis, sepsis syndrome, and septic shock. Crit Care Med 1993;21:1012-1019. 110. Steffes CP, Bender JS, Levison MA. Blood transfusion and oxygen consumption in surgical sepsis. Crit Care Med 1991;19:512-517. 111. Czer LSC, Shoemaker WC. Optimal hematocrit value in critically ill post-operative patients. Surg Gynecol Obstet 1978;117:363-368. 122 112. Conrad S, Dietrich K , Herbert C, et al. Effect of red cell transfusions on oxygen consumption following fluid resuscitation in septic shock. Circ Shock 1990;31:419-420. 113. Fortune J, Feustel P, Saif J, et al. Influence of hematocrit on cardiopulmonary function after acute hemorrhage. J Trauma 1987;27:243-249. 114. Hussain S N A , Roussos C. Distribution of respiratory muscle and organ blood flow during endotoxic shock in dogs. J A p p l Physiol 1985;59:1802-1810. 115. Shoemaker W C , Appe l PL, Kram H B . Oxygen transport measurements to evaluate tissue perfusion and titrate therapy. Crit Care M e d 1991;19:672-688. 116. Ruttimann Y , Schutz Y , Jequier E, et al. Thermogenic and metabolic effects of dopamine in healthy men. Crit Care M e d 1991;19:1030-1036. 117. Ruiz C , Wei l M H , Carlson R. Treatment of circulatory shock wi th dopamine. J A M A 1979;242:165-168. 118. Haupt M T , Kaufman BS, Carlson R W . Flu id resuscitation in patients w i t h increased vascular permeability. Crit Care C l i n 1992;8:341-353. 119. MacLean L D , W e i l M H . Hypotension in drugs produced by Escherichia col i endotoxin. Circ Res 1956;4:546. 120. MacLean L D , Wei l M H , Spink W W , et al. Canine intestinal and liver weight changes induced by E. Co l i endotoxin. Proc Soc Exp M e d 1956;92:602. 121. Teule GJJ, Hollander W D , Bronsveld W, et al. Effect of volume loading and dopamine on hemodynamics and red-cell redistribution in canine endotoxin shock. Circ Shock 1983;10:41. 122. W e i l M H , Nishijma H . Cardiac output in bacterial shock. A m J M e d 1978;64:920-927. 123. Ast iz M E , Rackow EC, Falk JL, Kaufman BS, W e i l M H . Oxygen delivery and consumption in patients wi th hyperdynamic septic shock. Crit Care M e d 1987;15:26-28. 124. Kaufman BS, Rackow E C , Falk JL. The Relationship Between Oxygen Delivery and Consumption during Flu id Resuscitation of Hypovolemic and Septic Shock. Chest 1984;85:336-340. 125. van der Linden P, Gilbart E, Paques P, Simon C, Vincent J-L. Influence of hematocrit on tissue O2 extraction capabilities during acute hemorrhage. A m J Physiol 1993;264:H1942-H1947. 123 126. T y m l K . Heterogeneity of microvascular flow in rat skeletal muscle is reduced by contraction and by hemodilution. Int J Microcirc C l i n Exp 1991;10:75-86. 127. Vincent JL. The Colloid-Crystalloid Controversy. K l i n Wochenschr 1991;Suppl XXVL104-111. 128. Griffel M I , Kaufman BS. Pharmacology of Colloids and Crystalloids. Crit Care C l i n 1992;8:235-253. 129. Carlson R W , Rattan S, Haupt M T . Flu id resuscitation in conditions of increased permeability. Anesthesiol Rev 1990;17:14. 130. Spagna G , Siegel J, Brown G , et al. Reprioritization of hepatic plasma protein in trauma and sepsis. A r c h Surg 1985;120:187-197. 131. Demling R H , Wong C, Wenger H . Effect of Endotoxin on the Integrity of the Peripheral (Soft Tissue) Microcirculation. Circ Shock 1984;12:191-202. 132. Rackow EC, Falk JL, Fein IA, et al. F lu id resuscitation in circulatory shock: a comparison of the cardiorespiratory effects of albumin, hetastarch, and saline solutions in patients wi th hypovolemic and septic shock. Crit Care M e d 1983;ll:839-850. 133. Karzai W, Reilly J M , Hoffman W D , et al. Hemodynamic effects of dopamine, norepinephrine, and fluids in a dog model of sepsis. A m J Physiol 1995;268:H692-H702. 134. D 'Orio V , Mendes P, Carlier P, Fatemi M , Marcelle R. Lung fluid dynamics and supply dependency of oxygen uptake during experimental endotoxic shock and volume resuscitation. Cri t Care M e d 1991;19:955-962. 135. Mul l i n s RJ, Tahamont M V , Bell DR, Ma l ik A B . Effect of f luid resuscitation from endotoxin shock on lung transvascular fluid and protein exchange. A m J Physiol 1991;260:H1415-H1423. 136. El lman H . Capillary permeability in septic patients. Cri t Care M e d 1984;12:629-633. 137. Ast iz M E , Galera-Santiago A , Rackow EC. Intravascular Volume and F l u i d Therapy for Severe Sepsis. N e w Horizons 1993;1:127-136. 138. Starling E H . O n the absorption of fluids from the connective tissue spaces. J Physiol (Lond) 1896;9:312-326. 139. Taylor A E , Garrl C A . Estimation of equivalent pore radii of pu lmonary capillary and alveolar membranes. A m J Physiol 1970;218:1133-1141. 124 140. Wittmers L E , Barlett M , Johnson JA. Estimation of capillary permeability coefficient of insulin in various tissues of the rabbit. Microvasc Res 1976;11:67-74. 141. Brigham K , Bowers RE, Haynes J. Increased sheep lung vascular permeability caused by E. coli endotoxin. Circ Res 1979;45:292-300. 142. Tullis JL. Albumin . J A M A 1977;237:355-360. 143. W e i l M H , Michaels S, Pur i V , et al. The stat laboratory. A m J C l i n Pathol 1981;76:34-42. 144. Bergofsky E H . Determination of tissue O2 tensions by hol low visceral tonometers: effect of breathing enriched O 2 mixtures. J C l i n Invest 1964;43:193-200. 145. Dawson A M , Trenchard D, Guz A . Small bowel tonometry: assessment of small gut mucosal oxygen tension in dog and man. Nature Lond 1965;206:943-944. 146. Kivisaari J, Ni in ikoski J. Use of Silastic tube and capillary sampling technique in the measurement of tissue P O 2 and P C O 2 . A m J Surg 1973;125:623-627. 147. Fiddian-Green R G , Pittenger G , Whitehouse W M . Back-diffusion of C O 2 and its influence on the intramucosal p H in gastric mucosa. J Surg Res 1982;33:39-48. 148. Antonsson JB, Boyle C C , Kruithoff K L , et al. Val idat ion of tonometric measurement of gut intramural p H during endotoxemia and mesenteric occlusion in pigs. A m J Physiol 1990;259:G519-G523. 149. Fiddian-Green R G , Haglund U , Gutierrez G , Shoemaker W C . Goals for the resuscitation of shock. Cri t Care M e d 1993;21:S25-S31. 150. Groeneveld ABJ , Ko lkman JJ. Splanchnic Tonometry: A Review of Physiology, Methodology, and Clinical Applications. J Crit Care 1994;9:198-210. 151. Clark C H , Gutierrez G . Gastric Intramucosal p H : A Noninvas ive Method for the Indirect Measurement of Tissue Oxygenation. A m J Cri t Care 1992;2:53-60. 152. Gutierrez G , Palizas F, Doglio G , et al. Gastric intramucosal p H as a therapeutic index of tissue oxygenation in critically i l l patients. Lancet 1992;339:195-199. 153. Maynard N , Bihari D, Beale R, et al. Assessment of Splanchnic Oxygenation by Gastric Tonometry in Patients Wi th Acute Circulatory Failure. J A M A 1993;1993:1203-1210. 154. Schlichtig R, Bowles SA. Distinguishing between aerobic and anaerobic appearance of dissolved C O 2 in intestine during low flow. J A p p l Physiol 1994;76:2443-2451. 125 155. Gutierrez G , Bismar H , Dantzker DR, Silva N . Comparison of gastric intramucosal p H wi th measures of oxygen transport and consumption in critically i l l patients. Crit Care M e d 1992;20:451-457. 156. Stacpoole PW. Lactic Acidosis. Endocrinol Metab C l i n Nor th A m 1993;22:221-245. 157. Lupo M A , Cefalu WT, Pardridge W M . Kinetics of lactate transport into rat liver in vivo. Metabolism 1990;38:374-377. 158. Ronco JJ, Fenwick JC, Tweeddale M G , et al. Identification of the Crit ical Oxygen Delivery for Anaerobic Metabolism in Critically 111 Septic and Nonseptic Humans. J A M A 1993;270:1724-1730. 159. Dodds WJ. The pig model for biomedical research. Fed Proc 1982;41:247-250. 160. Ca in S M , Curtis SE. Experimental models of pathologic oxygen supply dependency. Crit Care M e d 1991;19:603-612. 161. Breslow MJ , Mi l le r CF, Parker SD, Walman A T , Traystman RJ. Effect of vasopressors on organ blood flow during endotoxin shock in pigs. A m J Physiol 1987;252:H291-H300. 162. Elias H , Hyde D M . A n elementary introduction to stereology (quantitative microscopy). A m J Anat 1980;159:412-461. 163. Hogg JC, McLean T, Mar t in B A , Wiggs B. Erythrocyte transit and neut rophi l concentration in the dog lung. J A p p l Physiol 1988;65:1217-1225. 164. A l l a rd M F , Kamimura CT, English DR, Henning SL, Wiggs BR. Regional myocardial capillary erythrocyte transit time in the normal resting heart. Circ Res 1993;72:187-193. 165. Schumacker PT, Samsel R W . Analysis of oxygen delivery and uptake relationships in the Krogh tissue model. J A p p l Physiol 1989;67:1234-1244. 166. Beal A L , Cerra FB. Mul t ip le Organ Failure Syndrome in the 1990s: Systemic Inflammatory Response and Organ Dysfunction. J A M A 1994;271:226-233. 167. Samsel R W , Schumacker PT. Determination of the critical O2 delivery from experimental data: sensitivity to error. J A p p l Physiol 1988;64:2074-2082. 168. Barroso-Aranda J, Schmid-Schonbein G W , Zweifach B W , Mathison JC. Polymorphonuclear neutrophil contribution to induced tolerance to bacterial lipopolysaccharide. Circ Res 1991;69:1196-1206. 169. Piiper J, Haab P. Oxygen supply and uptake in tissue models wi th unequal distribution of blood flow and shunt. Respir Physiol 1991;84:261-271. 126 170. Hon ig CR, Odoroff CL . Calculated dispersion of capillary transit times: significance for oxygen exchange. A m J Physiol 1981;240:H199-H208. 171. Morff RJ. Contribution of capillary recruitment to regulation of tissue oxygenation in rat cremaster muscle. Microvasc Res 1988;36:150-161. 172. Ostgaard G , Reed R K . Interstitial fluid accumulation does not influence oxygen uptake in the rabbit small intestine. Acta Anaesthesiol Scand 1995;39:167-173. 173. Funk W , Baldinger V . Microcirculatory Perfusion during Volume Therapy: A comparative Study Using Crystalloid or Col lo id in Awake A n i m a l s . Anesthesiology 1995;82:975-982. 174. Wagner BKJ, D 'Amel io LF . Pharmacologic and clinical considerations i n selecting crystalloid, colloidal, and oxygen-carrying resuscitation fluids, part 1. C l i n Pharm 1993;12:335-346. 175. Horton JW, Walker PB. Small-volume Hypertonic Saline Dextran Resuscitation from Canine Endotoxin Shock. A n n Surg 1991;214:64-73. 176. Guyton A C , Lindsey A W . Effect of elevated left atrial pressure and decreased plasma protein concentration on the development of pulmonary edema. Circ Res 1959;7:73-84. 177. Zarins C K , Rice C L , Peters K M , et al. L y m p h and pulmonary response to isobaric reduction in plasma oncotic pressure. Circ Res 1978;43:925-930. 178. Kramer G C , Harms B A , Bodai BI, et al. Effects of hypoproteinemia and increased vascular pressure on lung fluid balance in sheep. J A p p l Physio l 1983;55:1514-1522. 179. Rackow E C , Wei l M H , MacNei l A , et al. Effect of crystalloid and colloid fluids on extravascular lung water in hypoproteinemic dogs. J A p p l Physiol 1987;42:2421-2425. 180. Demling R H , Harms B, Kramer G . Acute versus sustained hypoproteinemia and post-traumatic pulmonary edema. Surgery 1982;92:79-86. 181. Mor isaki H , Bloos F, Keys J, Mart in C, Neal A , Sibbald WJ. Compared w i t h crystalloid, colloid therapy slows progression of extrapulmonary tissue injury i n septic sheep. J A p p l Physiol 1994;77:1507-1518. 182. Demling R, LaLonde C, Saldinger P, Knox J. Mult iple-Organ Dysfunction i n the Surgical Patient: Pathophysiology, Prevention, and Treatment. Curr Prob Surg 1993;30:345-424. 127 183. Zikr ia B A , Subbarao C, Oz M C , et al. Macromolecules reduce abnormal microvascular permeability in rat limb ischemia-reperfusion injury. Cri t Care M e d 1989;17:1306-1309. 184. Mauer P H , Bernadinelli B. Immunologic studies wi th hydroxyethyl starch - a proposed plasma expander. Transfusion 1968;8:265-268. 185. Shatney C H , Chaudry IH. Hydroxyethyl starch administration does not depress reticuloendothelial function or increase mortality from sepsis. Circ Shock 1984;13:21-25. 186. Ziegler EJ, Fisher CJ, Jr.,, Sprung C L , et al. Treatment of Gram-negative bacteremic and septic shock wi th H A - 1 A human monoclonal antibody against endotoxin. A randomized, double-blind, placebo-controlled trial. N Engl J M e d 1991;324:429-436. 187. Munster A , Winchurch R, Thupari J , . et al. Reversal of postburn immunosuppression wi th low dose polymyxin B. J Trauma 1986;26:995-1000. 188. Vadas P, Pruzanski W , Stefanski E, et al. Pathogenesis of hypotension i n septic shock: Correlation of circulatory phospholipase A 2 levels wi th circulatory collapse. Crit Care M e d 1988;16:1-7. 189. Lefer A . Significance of l ip id mediators in shock states. Circ Shock 1989;27:3-12. 128 "@en . "Thesis/Dissertation"@en . "1997-11"@en . "10.14288/1.0087799"@en . "eng"@en . "Surgery"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "The effects of crystalloid resuscitation on oxygen extration in whole body and gut during endotoxemia"@en . "Text"@en . "http://hdl.handle.net/2429/6495"@en .