Pathophysiology of Sepsis and Multiple Organ Dysfunction

Published on 26/03/2015 by admin

Filed under Critical Care Medicine

Last modified 22/04/2025

Print this page

This article have been viewed 8159 times

129 Pathophysiology of Sepsis and Multiple Organ Dysfunction

Konrad Reinhart, Frank Bloos

Pathophysiology of Sepsis

The term sepsis is derived from a Greek word meaning “putrid.” It was believed that putrefaction of a wound was caused by contact with air and that death occurred when the process of putrefaction reached the blood (septicemia). In the 19th century, the concept of infection as a cause of sepsis was introduced by the Austrian obstetrician, Ignaz Philipp Semmelweis and the English surgeon, Joseph Lister. From then on, the term sepsis was closely connected to bacterial infection. However, as the understanding of human immune physiology improved, the importance of the host response to infection in the pathophysiology of sepsis was recognized.

Sepsis has been defined as an invasion of microorganisms or their toxins into the bloodstream, together with the host response to this invasion.¹ Thus, the pathophysiology of sepsis combines the impact of infection with the host response of generalized inflammation, which finally leads to multiorgan dysfunction and death. This definition has been extended by the addition of several terms to more carefully describe the disease and its pathophysiology (Table 129-1). The American College of Chest Physicians/Society of Critical Care Medicine (ACCP/SCCM) Consensus Conference defined sepsis as a systemic inflammatory response syndrome (SIRS) caused by infection.² More recently it has been recognized that SIRS is counteracted by a hypoinflammatory state that also plays an important role in the further development of organ dysfunction.³

TABLE 129-1 Definitions

Term	Definition
Bacteremia	Presence of viable bacteria in the blood
Systemic inflammatory response syndrome (SIRS)	Generalized hyperinflammatory response to several impacts
Sepsis	SIRS caused by infection
Severe sepsis	Sepsis associated with organ dysfunction
Septic shock	Sepsis associated with arterial hypotension

Data from ACCP/SCCM Consensus Conference Committee. Definition for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 1992;20:864–74.

Sepsis is characterized by loss of hemostatic balance and endothelial dysfunction, which in turn severely compromise the cardiocirculatory system as well as intracellular homeostasis. Cellular hypoxia and apoptosis (programmed cell death) then contribute to organ dysfunction and death. The network of organ systems affected by sepsis is depicted in Figure 129-1.

Figure 129-1 Pathophysiology of sepsis. HMGB, high-mobility group B protein; IL, interleukin; MIF, migration inhibitory factor; PAF, platelet-activating factor; TNF, tumor necrosis factor.

Microbiological Stimulus

By definition, infection is a fundamental part of the pathophysiology of sepsis. Any microorganism able to induce infection in humans may be complicated by sepsis. Bacteria as well as fungi, parasites, and to a lesser degree viruses can trigger the mechanisms that lead to sepsis. Although SIRS is the final common pathway of this process, the signal transduction pathway from infection to complex host response differs with the microbiological stimuli. Induction of an innate immune response is triggered by specific microbial molecules (e.g., bacterial wall components, exotoxins, bacterial DNA, viral RNA) called pathogen-associated molecular patterns (PAMPs). Damage-associated molecular patterns (DAMPs) are the noninfectious equivalents to PAMPs. DAMPs are released after cellular injury of the host (i.e., trauma) and can also induce the innate immune response.⁴

The presence of PAMPs is sensed by recognition molecules called pattern-recognition proteins (PRR), which are able to initiate a host response. These proteins may be categorized into secreted, transmembrane, and cytosolic PRRs. The Toll-like receptors (TLRs) represent the membrane PRRs. Eleven different TLRs have been discovered in mammals, whereas TLR-11 is not expressed in humans (Table 129-2). Retinoic acid-inducible gene I (RIG-I)–like receptors (RLRs) and the nucleotide-binding domain and leucine-rich repeat-containing receptors (NLRs) are cytosolic PRRs. RLRs recognize viral RNA and some double-stranded DNA. NLRs represent a large family of intracellular sensors that can detect pathogens and stress signals. NLRs detect microbiological products such as peptidoglycans and other degradation products of microorganisms as well as stress-related substances.^5,⁶

TABLE 129-2 Human Toll-Like Receptors and Their Natural Ligands

TLR Type	Related Pathogen-Associated Molecular Pattern	Location
TLR1 (via TLR2)	Bacterial products such as tri-acyl lipopeptides	Cell surface
TLR2	Gram-positive bacterial products, including peptidoglycans; some virus-related proteins	Cell surface
TLR3	Viral double-stranded RNA	Endosomal
TLR4	Endotoxin, other bacterial products, some fungal products	Cell surface
TLR5	Flagellin	Cell surface
TLR6 (via TLR2)	Some bacterial products	Cell surface
TLR7	Single-stranded RNA	Endosomal
TLR8	Viral single-stranded RNA	Endosomal
TLR9	Viral and bacterial DNA	Endosomal
TLR10	Unknown

TLR, Toll-like receptor.

Gram-Negative Sepsis

In gram-negative bacteremia, initiation of the immune response is mediated primarily by lipopolysaccharide (LPS), a bacterial cell wall product. In plasma, LPS is bound to the LPS binding protein (LBP). Bound LPS is transported to the opsonic receptor, CD14, which is located on several cell membranes including on monocytes.⁷ A soluble form of CD14 interacts with CD14-negative cells (e.g., dendritic cells). However, CD14 alone cannot explain the actions of LPS, because CD14 does not have an intracellular tail.

Another binding site of LPS is the transmembranous receptor, TLR4, which exists in combination with the accessory protein, MD2.⁸ The binding of LPS to CD14 and TLR4 induces, via other molecules, activation of the transcription factor, nuclear factor kappa-B (NF-κB). Activated NF-κB migrates into the nucleus where it binds to and activates gene promoters, resulting in the transcription and expression of genes for cytokines and other proinflammatory mediators.⁹ In monocytes, LPS also induces cytokine transcription via the triggering receptor expressed on myeloid cells-1 and the myeloid DAP12-associated lectin.¹⁰ Intracellular pattern-recognition proteins in monocytes for LPS have recently been identified as another pathway of cytokine expression and include nucleotide-binding oligomerization domain 1 and 2 as LPS binding sites.¹¹

Gram-Positive Sepsis

During the last decade, gram-positive bacteria have gained greater importance as causative organisms for sepsis.¹² Gram-positive bacteria lack endotoxin and are recognized by cell wall components such as peptidoglycans and released bacterial toxins (exotoxins). Recently, lipoteichoic acid (LTA), a component of the cell wall in all gram-positive bacteria, has been recognized as the main pattern for recognition of gram-positive bacteria.¹³ TLR2 has been identified as the only pattern-recognition protein for gram-positive bacteria.¹⁴ The relationship between LTA and TLR2 is not completely clarified. Although LTA clearly interacts with TLR2, TLR2 is not a specific receptor for LTA, because TLR2 can recognize several other components of gram-positive bacteria.¹⁵ Gram-positive and gram-negative sepsis are indistinguishable clinically, suggesting a similar pathway of signal transduction. Indeed, peptidoglycans and LTA stimulate the release of tumor necrosis factor alpha (TNF-α), interleukin (IL)-6, and IL-10. It has been speculated that CD14 is also involved in the signaling of gram-positive infections.

Some exotoxins cause a special type of septic shock called the toxic shock syndrome (TSS). TSS may be caused by the exotoxin, TSS toxin-1, staphylococcal enterotoxins from Staphylococcus aureus, or streptococcal pyogenic exotoxins.¹⁶ These toxins are capable of acting as so-called superantigens which deploy their effects via the T-cell antigen receptor (TCR). The TCR consists of five variable elements: Vβ, Dβ, Jβ, Vα, and Jα. Conventionally, the T cell is activated if the major histocompatibility complex (MHC) of an antigen-presenting cell matches all five elements. Thus, T cells are activated by proper antigen contact only. This results in the stimulation of about 1 in 10,000 T cells. However, a superantigen such as TSS toxin-1 works as a bridge between the MHC and the Vβ chain of the TCR only (Figure 129-2).¹⁷ Because T-cell activation now occurs independently of a match between the MHC and TCR, about 20% of the entire T-cell pool may be activated at once. Besides further T-cell proliferation, T-cell activation causes the release of several cytokines (i.e., interferon gamma [IFN-γ], IL-2, TNF-α) from T cells, as well as IL-1β and TNF-α from macrophages. Thus, the presence of superantigens results in a release of cytokines similar to gram-negative sepsis. It is assumed that actions other than cytokine production may be responsible for actions of superantigens in TSS; for example, superantigens may amplify the effects of LPS.

Figure 129-2 Pathophysiology of superantigen action. Superantigens work as a bridge between the T-cell receptor and the class 2 major histocompatibility complex (MHC) molecules.

Other Microbiological Stimuli of Sepsis

Sepsis can also be induced by fungi, viruses, and parasites. Signal transduction by nonbacterial products, however, is not as well characterized as bacterial sepsis. In part, this may be due to the fact that induction of cytokine release differs markedly not only among different microorganisms but also among species. Nevertheless, the release of proinflammatory mediators has been demonstrated during infections with nonbacterial infections such as Candida albicans¹⁸ and Plasmodium falciparum.¹⁹ The signal transduction in viral infections is complicated by the fact that viruses can interfere with TNF-related cytokine release to avoid the host’s antiviral activities.²⁰ Pattern recognition of viruses occurs mainly via endosomal TLR receptors which detect single- and double-stranded RNA or DNA (see Table 129-2).

The Immune Response in Sepsis

The cytokines TNF-α and IL-1β are released by activated macrophages and CD4 T cells within the first hour after infection. These primary mediators induce the release of several secondary mediators that amplify inflammation (Table 129-3). An important step in signal amplification is activation of the complement system. Besides being activated by antigen-antibody complexes, the complement system may be stimulated by bacterial surface sugars and endotoxin. The complement fragment C5a, a cleavage product of the complement cascade, is a strong chemoattractant. C5a appears about 2 hours after the initiation of sepsis and stimulates macrophages to further produce proinflammatory mediators. Another mediator that amplifies the immune response is macrophage migration inhibitory factor (MIF), which is produced by T cells, macrophages, monocytes, and pituitary cells in response to an infectious stimulus. MIF appears about 8 hours after the onset of sepsis and activates T cells and macrophages to produce proinflammatory mediators. About 24 hours after the initiation of sepsis, levels of high-mobility group box 1 (HMGB1) protein increase and appear to play a role in endotoxin-related sepsis. HMGB1 is a nuclear binding protein that, among other things, is capable of activating NF-κB. As a rather late mediator in sepsis, it is produced by macrophages and neutrophils and stimulates phagocytic cells.²¹

TABLE 129-3 Macrophage Mediators Involved in the Pathogenesis of Sepsis

Mediator	Typical Effects
Cytokines: IL-1, IL-6, IL-12, IL-15, IL-18, TNF, MIF, HMGB1, IL-10	Activate neutrophils, lymphocytes, and vascular endothelium; up-regulate cellular adhesion molecules; induce prostaglandins, nitric oxide synthase, and acute-phase proteins; induce fever
IL-10 is predominantly a negative regulator of these effects.
Chemokines: IL-8, MIP-1α, MIP-1β, MCP-1, MCP-3	Mobilize and activate inflammatory cells, especially neutrophils; activate macrophages
Lipid mediators: platelet-activating factor, prostaglandins, leukotrienes, thromboxane, tissue factor	Activate vascular endothelium; regulate vascular tone; activate extrinsic coagulation cascade
Oxygen radicals: superoxide and hydroxyl radicals, nitric oxide	Antimicrobial properties; regulation of vascular tone

HMGB, high-mobility group B protein; IL, interleukin; MCP, monocyte chemoattractant protein; MIF, migration inhibitory factor; MIP, macrophage inflammatory protein; TNF, tumor necrosis factor.

Data from Cohen J. The immunopathogenesis of sepsis. Nature 2002;420:885–91.

Normally, the inflammatory process is well balanced and is necessary for the host to overcome the infectious impact. However, under certain conditions, the amplification process of inflammation is not limited to the site of infection and becomes generalized. As noted earlier, this phenomenon has been called the systemic inflammatory response syndrome, or SIRS. SIRS is not restricted to infectious stimuli; it is present in a variety of other conditions such as pancreatitis, burns, multiple trauma, and in patients undergoing heart surgery with cardiopulmonary bypass.

It is not completely understood why inflammation becomes generalized in some patients but stays localized in others. Genetic variants of cytokines may play a role in this issue. Single nucleotide polymorphisms (SNPs) are single-base changes in the DNA which do not cause obvious changes in the function of the respective cytokine. However, SNPs in some cytokines are associated with a worse outcome from septic shock or an increased risk for developing sepsis.²²^–²⁴ Among several others, such variants have been described in TNF-α, IL-6, and CD14.²⁵ However, results from these studies are difficult to interpret because of contradictory results and differences in populations of different ethnicities.

The immune response in sepsis does not involve only proinflammatory mediators. As in many other physiologic processes, the organism produces inhibitors to control certain reactions. Proinflammatory mediators are counteracted by antiinflammatory molecules such as IL-4 and IL-10 because CD4 T cells can switch from the production of inflammatory cytokines (type 1 helper T cells [T_H1]) to the production of antiinflammatory cytokines (type 2 helper T cells [T_H2]). Soluble TNF receptors and IL-1 receptor antagonists (IL-1Ra) are released to inhibit the actions of TNF and IL-1 in their roles as primary mediators of sepsis. T cells, neutrophils, and macrophages also may become unresponsive to infectious stimuli (anergy).²⁶ Another mechanism of the antiinflammatory response is the onset of apoptosis, a genetically programmed autodestructive release of proteases that induces cell death. In sepsis, enhanced apoptosis causes loss of immune effector cells, including CD4 and CD8 T cells, B cells, and dendritic cells.²⁷ Absolute lymphocyte counts are significantly decreased in patients with sepsis.²⁸ Further, apoptotic cells impair the function of surviving immune cells.²⁹

Results from animal studies suggest that the autonomic nervous system is also involved in suppression of cytokine release during sepsis. In the experimental setting, vagal stimulation can inhibit TNF expression. It is hypothesized that an inflammatory reflex, with the afferent vagal nerve sensing cytokine release and an efferent immunosuppressing cholinergic arm, exists.³⁰ The importance of such a reflex in humans merits further investigation.

The antiinflammatory response in sepsis has been termed the compensatory antiinflammatory response syndrome (CARS).³ It has been suggested that the first response to infection is hyperinflammation, which is followed by a hypoimmune state. From there, recovery would be possible, but the prolonged inability to eradicate microorganisms might result in the death of the patient.³¹ However, serum levels of antiinflammatory cytokines are increased in parallel with the increase of proinflammatory mediators.^32,³³ Thus, diminished inflammation develops at the same time as the process of hyperinflammation. Although the persistence of high levels of antiinflammatory mediators may contribute to mortality in septic patients, the clinical role of change between hyperinflammatory and hypoinflammatory states remains unclear.

Loss of Hemostatic Balance

Under normal conditions, the vascular luminal surface has anticoagulant properties. Tissue factor is a 4.5-kD protein that is bound to cell membranes which are normally not in contact with blood. Expression of tissue factor mainly depends on release of IL-6.^34,³⁵ Tissue factor expression occurs on mononuclear cells, but endothelial cells, polymorphonuclear cells, and other cell types may be additional sources. The expression of tissue factor induces intravascular thrombin formation initiated by the extrinsic coagulation pathway. Because this process is not restricted to a local area, it is called disseminated intravascular coagulation (DIC); DIC causes a consumption of coagulation factors.

Physiologically, excessive coagulation is counteracted by several natural anticoagulants including antithrombin, the thrombomodulin/protein C/protein S system, and tissue factor pathway inhibitor. In addition to the activation of tissue factor-dependent thrombin generation, anticoagulant function is attenuated in sepsis. Patients with sepsis demonstrate reduced levels of protein C and antithrombin due to consumption and reduced synthesis.³⁶ Thus, the physiologic balance between procoagulant and anticoagulant substances is altered in sepsis as there is a shift of the hemostatic balance toward a procoagulant state (Figure 129-3).

Figure 129-3 Shift of hemostatic balance toward a procoagulant state in sepsis. PAI, plasminogen activator inhibitor; TFPI, tissue factor pathway inhibitor.

(Modified from Levi M, Ten Cate H. Disseminated intravascular coagulation. N Engl J Med 1999;341:586–92.)

Besides its anticoagulant actions, the protein C pathway is an important link between coagulation and inflammation, because activated protein C has antiinflammatory properties. Protein S can bind to receptors that mediate an antiinflammatory regulatory loop of dendritic cell and monocyte inflammatory function. Thrombomodulin has been described to prevent excessive complement activation.³⁷ Thus, there is considerable cross-talk between inflammation and coagulation which is impaired in sepsis as a result of depletion of the thrombomodulin/protein C/protein S system. There is also evidence that antithrombin blunts activation of several cytokines, suggesting that low antithrombin levels also affect the relationship between coagulation and inflammation.³⁸

Endothelial Dysfunction

Besides separating blood from tissue, endothelial cells have multiple physiologic functions involving the regulation of vascular tone, coagulation, and immune response.

The endothelium produces several vasoactive mediators, including nitric oxide (NO), prostacyclins, and endothelin. NO is a potent vasodilator produced by NO synthase (NOS) from the amino acid, L-arginine. NO directly relaxes the vessel’s smooth muscle. There are two different forms of endothelial NOS: the constitutional form (cNOS) and the inducible form (iNOS). Physiologically, cNOS—also referred to as endothelial NOS (eNOS)—produces only small amounts of NO, and iNOS is expressed at low levels.³⁹ In sepsis, iNOS expression is stimulated by cytokines such as IL-1β and TNF-α.⁴⁰ This is followed by massive NO production and profound vasodilatation. Whether increased activity of cNOS also plays a role in sepsis is currently a matter of debate.

During inflammation, endothelial cells express adhesion molecules on their surface, which causes the adherence of leukocytes. These adhesion molecules include endothelial leukocyte adhesion molecule-1, intracellular adhesion molecule-1, and vascular cell adhesion molecule-1. Endothelial leukocyte adhesion molecule-1 is a selectin that mediates the initial step of leukocyte adhesion, followed by leukocyte rolling along the endothelial surface. The leukocyte finally migrates through the endothelial layer into the tissue, mediated by intracellular adhesion molecule-1 and vascular cell adhesion molecule-1 expression on both endothelial cells and leukocytes.⁴¹

Migration of leukocytes into the tissue is a physiologic mechanism to move immune cells to the site of infection. However, in generalized inflammation such as in sepsis, endothelial cells in several organs remote from the site of infection express adhesion molecules, inducing a generalized rolling and sticking of circulating leukocytes to the vascular surface. Adherence to endothelial cells activates leukocytes and induces a respiratory burst.⁴² The respiratory burst involves the release of cytotoxic substances such as elastase, myeloperoxidase, and reactive oxygen species. These products are capable of damaging endothelial cells and the surrounding tissue. Endothelial cell damage causes capillary leakage whereby intravascular fluid penetrates the extracellular space, leading to tissue edema.

Endothelial cells have an anticoagulant surface produced by the expression of heparan sulfate on the cell membrane, release of plasminogen activator, and production of protein C. However, sepsis shifts the hemostatic balance toward a procoagulatory state. Endothelial cells share in this process by expressing tissue factor (see the previous section, Loss of Hemostatic Balance).

Cardiac and Circulatory Dysfunction

Sepsis is frequently complicated by organ dysfunction and shock. Shock occurs when the cardiovascular system is unable to transport sufficient amounts of oxygen to the tissues. In fact, sepsis compromises all levels of the cardiovascular system, resulting in cardiac dysfunction, vascular dysregulation, and microcirculatory damage. Impairment of the cardiovascular system causes a characteristic hemodynamic pattern that in cases of adequate fluid loading and the absence of severe preexisting cardiac dysfunction, consists of a high cardiac output, arterial hypotension, and low systemic oxygen (O₂) extraction. In early sepsis, O₂ consumption is increased owing to higher metabolic needs (i.e., tachypnea, fever, increased cardiac work, increased rate of protein synthesis), further compromising the relationship between O₂ supply and demand (Figure 129-4). The hepatic and splanchnic region is markedly affected by these changes associated with sepsis. Hepatosplanchnic O₂ uptake increases markedly during fever and bacteremia.⁴³

Figure 129-4 The cardiocirculatory system of a septic patient is altered on systemic, regional, and microregional levels. At the same time, sepsis increases O₂ consumption, further deteriorating the O₂ supply/demand relationship. ARDS, acute respiratory distress syndrome.

Cardiac Dysfunction

In experimental septic shock, myocardial contractility is compromised shortly after the induction of sepsis.⁴⁴ This finding is confirmed in septic patients when a reduced ejection fraction is observed by echocardiography, especially in patients with elevated troponin levels.⁴⁵ The drop in myocardial contractility is accompanied by diastolic dilatation of the left ventricle, which causes the left ventricular end-diastolic volume to rise. This mechanism allows the heart to maintain a sufficient stroke volume despite impaired contractility. Clinically, a rightward shift of the Frank-Starling curve occurs. Thus, compared with healthy humans, patients with sepsis require greater cardiac filling pressures to maintain a similar stroke volume.⁴⁶ Septic patients without compensatory left ventricular dilatation have a significantly greater risk of death.⁴⁷ Cardiac dysfunction is reversible if the patient recovers from sepsis.

The presence of myocardial depressant substances was initially proposed in the 1980s because the serum of septic patients was able to suppress the contractility of rat myocytes in vitro.^48,⁴⁹ Cytokines induce increased activation of iNOS, with subsequent enhanced NO production. NO affects myocytes in several ways: NO stimulates guanylate cyclase, and its product, 3′,5′-cyclic guanosine monophosphate, interferes with intracellular myocardial calcium metabolism. This includes a reduction in calcium’s affinity to the contractile apparatus and inhibition of the α-adrenergic-mediated increase in the slow inward calcium current. NO may directly damage myocardial cells by the formation of peroxynitrite via combination with superoxide ions. Peroxynitrite deploys toxic effects on many intracellular molecules by means of oxidation.⁵⁰

Sepsis is associated with alterations of regional and microregional blood flow, which results in a mismatch between regional O₂ supply and demand and, subsequently, multiple organ dysfunction. It was therefore hypothesized that the heart shares in this type of injury. Although there were hints from experimental work that the coronary circulatory reserve is altered in sepsis,^51,⁵² clinical studies did not show a compromised coronary blood flow.^53,⁵⁴ However, more recently it was demonstrated that patients with sepsis show elevated serum levels of troponin.⁵⁵ Elevated troponin values are associated with a higher incidence of regional wall motion abnormalities and death.

Vascular Dysfunction and Hypovolemia

In cardiogenic or hypovolemic shock, vasoconstriction is a common mechanism to avoid arterial hypotension. In sepsis, however, profound arterial vasodilatation occurs. Endothelial cells play an important role in the regulation of vascular tone because they release several vasoactive substances such as NO and endothelin. Sepsis shifts the balance of these substances toward a vasodilatory state by uncontrolled NO production, as discussed earlier. Severe arterial hypotension due to profound systemic vasodilatation is one of the characteristic hemodynamic features of sepsis. The mechanism by which NO induces vasodilatation is complex. Important pathways in which NO participates include the activation of potassium channels and hyperpolarization of the plasma membrane of smooth muscle cells. These mechanisms in turn inhibit the actions of vasopressors such as norepinephrine and angiotensin II, so that vasoconstriction does not occur despite high serum concentrations of these substances.⁵⁶

Endothelial cells regulate vascular tone not only to maintain systemic blood pressure but also to control blood flow to single organs. Several mechanisms to preserve organ blood flow are impaired in sepsis. For example, there is a loss of coupling between the hepatic artery flow and the portal blood flow in endotoxic shock.⁵⁰ Similarly, the coronary circulatory reserve necessary to quickly adjust myocardial O₂ supply based on changes in myocardial O₂ requirements is reduced in sepsis.⁵⁷ The autoregulation of perfusion of the intestinal mucosa is also depressed in experimental models of sepsis.⁵⁸

Hypovolemia is another characteristic of sepsis. Sepsis is accompanied by the development of significant tissue edema. The underlying mechanism is capillary leakage, which is another effect of endothelial damage.⁵⁹ This leakage also allows for the extravasation of albumin,⁶⁰ which reduces the intravascular oncotic pressure. Under conditions of capillary leakage, the Starling forces cannot counteract the development of tissue edema or reduce existing edema. Because endothelial damage affects all parts of the capillary network throughout the body, large amounts of intravascular fluid are shifted into the extravascular space.

Microcirculatory Dysfunction

Severe sepsis and septic shock may be associated with high lactate levels and metabolic acidosis despite a low systemic O₂ extraction. These signs of tissue hypoxia, which may be observed despite adequate fluid resuscitation, are interpreted as microcirculatory failure. Some parts of the microcirculation are extremely sensitive to physiologic stress, including hypoxia or ischemia. These areas are referred to as weak microcirculatory units.^61,⁶²

An increase in the number of weak microcirculatory units—and therefore increased microcirculatory shunting—is thought to play a major role in the O₂ extraction deficit in sepsis. This hypothesis has been confirmed in experimental sepsis. By using intravital microscopy, an increased number of capillaries with lack of flow was observed in septic animals.⁶³ More recently, capillary red blood cell oxygenation was measured in vivo by a spectrophotometric functional imaging system.⁶⁴ In the presence of sepsis, an increased proportion of perfused capillaries showed very high red blood cell velocities. These high-flow capillaries were interpreted as microcirculatory shunts. The remaining capillaries with normal flow had a fivefold increase in O₂ extraction. However, this increase in O₂ extraction was insufficient to maintain O₂ supply to all regions, given the number of capillaries without flow.

Several factors may be responsible for microcirculatory shunting in sepsis. The underlying mechanism is the hindrance of blood flow by microvascular obstruction, which has several causes: (1) the onset of intravascular coagulation due to a shift to a procoagulant state causes the development of microthrombi; (2) activation and damage of endothelial cells lead to endothelial cell swelling, narrowing the capillary lumen; (3) activated leukocytes hinder red blood cell flow by rolling and sticking to endothelial cells; and (4) red blood cells have reduced deformability in sepsis, which causes them to be captured in capillaries.⁶⁵

Although there is good evidence from experimental work supporting the hypothesis of microcirculatory dysfunction in sepsis, there is considerable debate whether the O₂ extraction deficit is due to a derangement of intracellular metabolic pathways rather than to microcirculatory dysfunction.⁶³ Some capillaries are able to increase their O₂ extraction,⁶⁶ which argues against this hypothesis. However, the assessment of tissue oxygenation on the cellular level is problematic, even in the experimental setting. Currently it is thought that microcirculatory dysfunction shares in the development of tissue hypoxia in sepsis. Clinically this hypothesis is supported by the finding that early resuscitation guided by central venous O₂ saturation improves survival in these patients.⁶⁷

Endocrine Dysfunction

As depicted in Table 129-4, critical illness is associated with alterations in several endocrine functions. It is not clear whether these changes represent a physiologic response to critical illness or reflect a complex picture of endocrine dysfunction that needs diagnostic and treatment strategies. In sepsis, adrenal insufficiency and vasopressin deficiency might contribute to the loss of vasomotor control. However, current studies do not support treatment of endocrine dysfunction in severe sepsis or septic shock.

TABLE 129-4 Changes in Hormone Concentrations in Critically Ill Patients

Hormone	Acute Critical Illness	Prolonged Critical Illness
Catecholamines	++	+
Cortisol	++	+
Adrenocorticotropic hormone	Ø +	Ø −
Growth hormone	Ø −	−
Thyroid hormones	Ø −	−
Thyroid-stimulating hormone	Ø −	−
Androgen hormones	−	−
Prolactin	–	Unknown

Data from Ligtenberg JJ, Girbes AR, Beentjes JA et al. Hormones in the critically ill patient: to intervene or not to intervene? Intensive Care Med 2001;27:1567–77.

Adrenal Insufficiency

Adrenal corticosteroids are involved in several physiologic pathways in the human body, including maintenance of vascular tone, vascular permeability, and distribution of total body water. In the clinical setting, corticosteroids also augment the effects of vasopressors.⁶⁸ Under normal conditions, corticosteroids are secreted by the adrenal cortex in a diurnal pattern. Corticosteroid secretion is tightly controlled by a feedback mechanism involving the hypothalamic-pituitary-adrenal axis. However, several mechanisms may impair the physiologic stress response of the hypothalamic-pituitary-adrenal axis in critically ill patients (Figure 129-5), resulting in an inadequate increase of serum cortisol levels. This condition is referred to as relative adrenal insufficiency.⁶⁹

Figure 129-5 Hypothalamic-pituitary-adrenal axis. Physiologic control of cortisol release and its feedback mechanism are illustrated. This axis may be impaired on all levels in critically ill patients. +, stimulation; −, inhibition.

(Modified from Cooper MS, Stewart PM. Corticosteroid insufficiency in acutely ill patients. N Engl J Med 2003;348:727–34.)

The concept of relative adrenal insufficiency as a clinically relevant condition is a matter of debate. It has been demonstrated that an inadequate rise of cortisol after the corticotropin stimulation test is associated with increased mortality in patients with septic shock.⁷⁰ However, most studies measured total cortisol levels, which are reduced due to hypoalbuminemia, while the concentrations of free cortisol may be adequate.⁷¹ Administration of low-dose corticosteroids does not affect survival from septic shock.⁷²

Vasopressin Deficiency

Vasopressin is excreted from the neurohypophysis in response to arterial hypotension or hypovolemia. Because septic shock is characterized by both arterial hypotension and hypovolemia, one would expect plasma vasopressin levels to be high; however, they are low in patients with septic shock ⁷³ and do not adequately respond when arterial hypotension occurs.⁷⁴ Indeed, the administration of vasopressin or analogs can quickly restore blood pressure in these patients.⁷⁵ Inadequate vasopressin levels may be caused by depression of the baroreflex, increased metabolism of vasopressin, and depletion of vasopressin stores in the pituitary gland.

Vasopressin is metabolized by plasma vasopressinase and by renal and hepatic clearance. Increased vasopressin metabolism seems unlikely in sepsis, because renal and hepatic functions are often compromised in this setting, and there is no evidence of increased vasopressinase activity in this disease. Depression of the baroreflex may play a role in vasopressin depletion. The baroreflex is mediated by sympathetic stimulation, and there is some evidence that sympathetic function might be impaired in sepsis.⁷⁶ Depletion of pituitary vasopressin stores was found by magnetic resonance imaging in a case series of three septic shock patients with low plasma vasopressin levels.⁷⁷ The clinical relevance of the low vasopressin response to septic shock remains unclear, except in a subgroup with less severe septic shock, because administration of low-dose vasopressin was not associated with better survival.⁷⁸

Insulin Deficiency

Hyperglycemia is a common feature in critically ill patients, such as those with severe sepsis or septic shock, and is caused by the endocrine response to stress. This includes activation of the hypothalamic-pituitary-adrenal axis, with the release of cortisol. Additionally, secretion of epinephrine, glucagon, and growth hormone is increased. All these hormones counteract effects of insulin. Besides the endocrine stress response, several cytokines additionally increase insulin resistance by inhibiting intracellular pathways normally activated by the insulin receptor.⁷⁹ There is also evidence that pancreatic insulin secretion is impaired in sepsis.^80,⁸¹

Endocrine stress response, insulin resistance, and impaired insulin secretion are responsible for the hyperglycemic state of patients with sepsis. Besides hyperglycemia, insulin deficiency may also be unfavorable in sepsis, since insulin has several beneficial effects such as potent antiinflammatory and anabolic actions. However, insulin administration to achieve normal glucose levels is not supported by clinical trials; the risk of hypoglycemia seems to outweigh any possible beneficial effects of insulin.^82,⁸³

Pathophysiology of Multiorgan Dysfunction

Multiorgan dysfunction is the parallel or sequential failure of at least two organs. It is a frequent complication of sepsis. Clinically, multiorgan dysfunction is termed the multiple organ dysfunction syndrome (MODS). MODS can involve any organ of a critically ill patient, even a remote organ that was not originally affected by the underlying disease. The development of MODS significantly contributes to ICU mortality. Scoring systems such as the Sequential Organ Failure Assessment (SOFA) or the Multiorgan Dysfunction Score, which assess the severity of MODS, correlate well with mortality.⁴

Some major components of the pathophysiology of MODS are depicted in Figure 129-6. The development of MODS includes a complicated network of inter- and intracellular actions. Inflammation seems to be a major trigger for the induction of the processes that lead to MODS. Both infectious and noninfectious stimuli may be responsible for activation of the innate immune response. As previously discussed, PAMPs activate the immune response and induce complex metabolic and circulatory changes in the host. Correspondingly, cellular trauma causes the release of DAMPs such as mitochondrial peptides and mitochondrial DNA into the circulation. DAMPs may induce systemic inflammation similar to sepsis.⁸⁴

Figure 129-6 Pathophysiology of multiple organ dysfunction syndrome (MODS).

Because MODS can involve a variety of pathologic changes, different concepts of the pathophysiology of MODS have been generated (Table 129-5).^2,⁸⁵ Cellular dysfunction due to tissue hypoxia is likely an important factor in the onset of MODS. However, other factors also play a role, including the onset of programmed cell death (apoptosis) and the direct toxic effects of substances such as endotoxin and reactive oxygen species. The development of SIRS does not require the presence of infection; severe trauma, burns, pancreatitis, and cardiac surgery with cardiopulmonary bypass are also associated with SIRS and increase the risk for MODS.⁸⁶

TABLE 129-5 Conceptual Models of Multiple Organ Dysfunction

Pathologic Process	Manifestation
Uncontrolled infection	Persistent infection, nosocomial acquired infection, endotoxemia
Systemic inflammation	Cytokinemia (particularly IL-6, IL-8, TNF), leukocytosis, increased capillary permeability
Immune paralysis	Nosocomial infection, increased antiinflammatory cytokine levels (IL-10), decreased HLA-DR expression; shift from type 1 to type 2 helper T-cells
Tissue hypoxia	Increased lactate, low central venous O₂ saturation
Microvascular coagulopathy and endothelial dysfunction	Increased procoagulant activity, decreased anticoagulant activity (antithrombin III ↓, protein C ↓), high levels of fibrin derivatives, increased von Willebrand factor, soluble thrombomodulin, increased capillary permeability
Dysregulated apoptosis	Increased epithelial and lymphoid apoptosis, decreased neutrophil apoptosis
Gut-liver axis	Increased infection with gut organisms, endotoxemia, Kupffer cell activation

HLA, human leukocyte antigen; IL, interleukin; TNF, tumor necrosis factor.

Modified from Marshall JC. Inflammation, coagulopathy, and the pathogenesis of multiple organ dysfunction syndrome. Crit Care Med 2001;29:S99–106.

Tissue Hypoxia

As discussed previously, the host response to infection severely impairs the cardiovascular system through the development of cardiac dysfunction, systemic and regional vascular dysregulation, and microcirculatory damage. Systemic hemodynamics can be restored by measures such as fluid resuscitation or treatment with catecholamines. However, no clinically available measures allow for a differential diagnosis of regional or microregional disturbances of blood flow, nor are therapies available to specifically address such disturbances. Thus, microcirculatory dysfunction and maldistribution of regional blood flow will persist or even progress if the underlying disease (e.g., sepsis) cannot be treated successfully. Tissue hypoxia may therefore be present even though treatment goals for adequate systemic hemodynamics have been achieved.

Tissue hypoxia is difficult to assess in a critically ill patient. It is therefore uncertain whether tissue hypoxia plays a leading role in organ dysfunction, because other mechanisms such as apoptosis have been identified as well.⁸⁷ In addition, it has been suggested that disturbance of mitochondrial O₂ utilization rather than tissue hypoxia is the motor of organ dysfunction. This hypothesis is supported by the observation that depletion of enzymes of the respiratory chain is associated with the development of MODS in septic patients.⁶⁶ Nevertheless, data are available that demonstrate the importance of tissue oxygenation in the development of MODS. For example, when central venous O₂ saturation was used to guide aggressive treatment to maintain O₂ delivery, there was a significant reduction in mortality,⁸⁸ supporting the concept of tissue hypoxia as an important mechanism in MODS. Likewise, patients with early lactate clearance have less severe organ dysfunction and improved outcome.⁸⁹

Apoptosis

Apoptosis is a physiologic mechanism whereby activation of a specific intracellular program induces cell death. Apoptosis is therefore a regulatory process for the proliferation and differentiation of cells. However, pathologic activation of apoptosis seems to be involved in the pathogenesis of MODS. Apoptosis is induced by a cascade system through either an extrinsic (receptor-dependent) or an intrinsic (receptor-independent) pathway. The extrinsic pathway is activated by the so-called death receptor superfamily, consisting of receptors such as the Fas receptor (CD95) or the TNF receptor. Thus, cytokines are able to induce apoptosis. The intrinsic pathway may be induced by DNA damage. The process of apoptosis is mediated by an enzymatic cascade system in which active caspase-3 is the executioner protein that finally starts apoptosis. The receptors of the extrinsic pathway mediate the activation of procaspase-8 to active caspase-8 via several signaling proteins (Figure 129-7). The intrinsic pathway works by altering the mitochondrial membrane potential through the signal protein, p53, which mediates the activation of caspase-9. Both active caspase-8 and active caspase-9 activate the final common step in the apoptosis pathway (caspase-3).⁹⁰

Figure 129-7 Extrinsic and intrinsic signaling process for apoptosis. FADD, Fas-associated death domain protein; RIP, receptor interacting protein; TRADD, TNF receptor 1–associated death protein.

(Modified from Lydon A, Martyn JA. Apoptosis in critical illness. Int Anesthesiol Clin 2003;41:65–77.)

The extent to which apoptosis-related cell death contributes to the development of MODS is difficult to assess because this process has been investigated primarily in the experimental setting. However, apoptosis of intestinal epithelial cells has been shown in trauma patients shortly after injury.⁹¹ The importance of increased apoptosis in the clinical setting warrants further elucidation.

The “Two-Hit” Theory

Any severe impact to the human body such as a traumatic or surgical injury or prolonged shock can directly induce the development of organ dysfunction. Possible mechanisms include ischemia, reperfusion injury, or immediate tissue destruction due to trauma. Such an event is called a “first hit.” This first hit may be severe enough to induce SIRS, with all the consequences to the cardiocirculatory system and cellular functions mentioned earlier.

Even if the first hit does not induce a primary MODS, a “second hit” such as an infectious insult (e.g., pneumonia or bacteremia due to catheter infection) could further activate an immune system that is already primed by the first hit. The “two-hit” theory hypothesizes that a second (or third) insult amplifies the inflammatory response to the first hit in such a way that SIRS occurs. If these events are followed by multiple organ dysfunction, the term secondary MODS is used.

The two-hit theory has been criticized for being somewhat arbitrary, because the differentiation between primary and secondary MODS is not always possible in the clinical setting.⁹² However, our current understanding of the inflammatory response to injury supports the concept of priming. Neutrophils are known to produce greater amounts of reactive oxygen species and have increased adhesion properties after they have been exposed to proinflammatory mediators. It has been demonstrated, at least in vitro, that priming followed by activation of neutrophils increases the extent of endothelial damage.