Steven Larosa MD1
Steven M. Opal MD, FACP2
1Assistant Professor of Medicine, Brown University School of Medicine
2Professor of Medicine, Department of Medicine, Brown University School of Medicine
Steven LaRosa, M.D., has received clinical coordinating center grants from Eisai Inc. and Novartis.
Steven M. Opal, M.D., has received preclinical grant support from Wyeth and clinical coordinating center grants from Eisai Inc. and Novartis.
Sepsis, along with the multiorgan failure that often accompanies the systemic inflammatory response syndrome (SIRS), is a leading cause of mortality in the intensive care unit.1,2 Sepsis develops in 650,000 to 750,000 patients annually in the United States.2,3 Nearly half of these patients manifest severe sepsis and septic shock. The mortality for septic shock remains approximately 30% to 45%, despite a concerted effort to improve the treatment options and outcome.1,2,3
Septic shock has become a major focus of critical care research. Although modest improvements in the prognosis have been made over the past 2 decades and promising new therapies have appeared in the past few years, innovations in the management of septic shock are still required. This chapter reviews some of the remarkable advances achieved in the understanding of the molecular pathophysiology of sepsis, the diagnostic and therapeutic strategies emerging from this research, and the current management of septic shock.
Definitions of Sepsis
The definitions of sepsis, septic shock, SIRS, and multiple organ dysfunction syndrome (MODS) were updated and revised in 2003 from the original definitions proposed by the American College of Chest Physicians/Society for Critical Care Medicine (ACCP/SCCM) Consensus Conference [see Table 1].1,44 The definitions take into account the finding that sepsis may result from a multitude of infectious agents and microbial mediators and may or may not be associated with actual bloodstream infection. Despite their clinical logic and intrinsic simplicity, the clinical applicability of these consensus definitions has been justifiably criticized. The SIRS definition is so broad and nonspecific that many patients admitted to general medical services and most ICU patients have conditions that meet the definition of SIRS. The current definition of sepsis fits such innocuous illnesses as a febrile child seen in a pediatrician's office with acute otitis media. Fever and leukocytosis from otitis meets SIRS criteria, and the specific type of bacterial or viral causative microorganism can cause the infection to meet the sepsis criteria. Despite their shortcomings, these definitions remain useful in classifying patients for inclusion in clinical trials of sepsis, and they are widely utilized for this purpose.
Table 1 The Terminology of Sepsis
One of the basic tenets upon which these definitions are based is that the inflammatory response itself, not the infectious agent, underlies the pathophysiology of the septic process. Although this fundamental hypothesis is likely correct, the nature of the causative microorganism clearly contributes to the ultimate fate of the patient. Pathogens differ in their susceptibility to host defenses, their potential for developing antimicrobial resistance, and their ability to generate toxins—all of which affect their pathogenicity.5 Failure to account for these intrinsic differences in microbial virulence limits the utility of current sepsis definitions.
Many patients who present with sepsis have multiple predisposing factors, a variety of preexisting illnesses, and major underlying organ dysfunction from a myriad of diseases. The degree to which sepsis contributes to further disordered organ function may be difficult to determine with accuracy. The same can be said for the degree to which sepsis contributes to mortality in patients who suffer from other serious underlying diseases. All these factors further limit the discriminatory value of the current consensus definitions of sepsis.
Further refinements in sepsis terminology may be possible when rapid diagnostic techniques become available to assess the immune status of septic patients. Functional genomics and proteomics (the study of human gene sequences and protein sequences, respectively) may assist in characterizing septic patients in the future. An innovative alternative conceptual framework for classifying and analyzing the complex, clinical syndrome of sepsis, the PIRO staging system, has been proposed [see Table 2].1 PIRO is an acronym for the four major components that characterize sepsis: predisposing factors, infection, response, and organ failure. This system is patterned after the TNM system (i.e., tumor size, presence of nodal involvement, and occurrence of metastasis) developed for study of neoplasia. It is hoped that a similar analysis using PIRO will facilitate a greater understanding of sepsis pathophysiology. The internal and external validity and the ultimate clinical utility of PIRO for defining sepsis remain to be determined.
Table 2 PIRO Staging of Sepsis1
Between 1979 and 2000, the incidence of sepsis increased by 8.7% annually, from 82.7 to 240.4 per 100,000 population.2 This trend is likely to continue in the foreseeable future because sepsis is essentially a disease whose incidence is linked to medical progress.3,6Successful management of a variety of severe medical and surgical diseases has produced a large population of patients with critical illness and impaired host defenses; these patients have a greatly increased risk of developing sepsis. Innovations in organ transplantation, implanted prosthetic devices, and long-term vascular access devices continue to cause this patient population to expand. The gradual aging of the population in many developed countries and the increasing prevalence of antibiotic-resistant microbial pathogens also contribute to the rising incidence of septic shock.
Pathogenesis: Microbial Factors
The microbiology of sepsis has undergone a remarkable transition in the past 25 years. The predominant microbial pathogens responsible for sepsis in the 1960s and 1970s were enteric gram-negative bacilli and Pseudomonas aeruginosa, but these organisms have now been surpassed by gram-positive bacterial path o gens.2,6 The rapid evolution of antibiotic-resistance genes in gram-positive bacterial pathogens and the frequent occurrence of vascular catheter-related bacterial sepsis by these organisms contribute to the increasing prevalence of gram-positive pathogens as a cause of sepsis. Opportunistic fungal pathogens are also increasing in frequency as a cause of sepsis.2
Bacterial endotoxin, or lipopolysaccharide (LPS), is an intrinsic component of the outer membrane of gram-negative bacteria and is essential for the viability of enteric bacteria.7 The unique potency of endotoxin is illustrated by the isolation of an endotoxin-deficient strain ofNeisseria meningitidis that is at least 100-fold less potent an inducer of cytokine production than wild-type bacteria.7 Endotoxin functions as an alarm molecule that alerts the host to invasion by gram-negative bacteria.8 Its presence in the circulation provokes a vigorous systemic inflammatory response. The host response to the endotoxin, rather than the endotoxin itself, accounts for the endotoxin's potentially lethal properties. As a species, humans are especially susceptible to the profound immunostimulant properties of endotoxin; even minute doses may be lethal.
The Toll-like receptor (TLR) family is the most important cellular, pathogen-associated pattern recognition receptor system in humans. The TLRs are transmembrane receptors for the detection of endotoxin and many other microbial mediators, including bacterial, fungal, viral, and parasitic components.9,10,11
TLRs belong to an array of pattern-recognition molecules that alert the innate immune response system to the presence of a microbial invader. Such pattern-recognition molecules include alternative complement components, mannose-binding lectin,12 and CD14.13
The innate immune system is by nature a rather nonspecific antimicrobial defense system. It lacks the precision of the adaptive immune system (i.e., B cells and T cells), but its immediate action—phagocytosis and clearance of pathogens—in the initial stages of infection makes the innate immune system a critical survival mechanism [see 6:II Innate Immunity]. Activation of the innate immune system and its cellular components (neutrophils, monocytes, macrophages, and natural killer [NK] cells) is primarily responsible for the pathogenesis of septic shock.14
LPS forms microaggregates in biologic fluids and then rapidly interacts with a variety of receptors (e.g., CD14) found in the serum or bound to cell membranes. In human blood and body fluids, LPS signaling is mediated by interactions with a hepatically derived, acute-phase plasma protein known as LPS-binding protein (LBP). LBP functions primarily as a shuttle molecule that facilitates LPS delivery to the receptor complex where cellular activation occurs. A series of events then releases a signal into the intracellular space, which subsequently activates LPS-responsive genes. Expression of these genes activates clotting elements, complement, other acute phase proteins, cytokines, chemokines and nitric oxide synthase genes have NFκB binding sites in their regulatory elements. The outpouring of inflammatory cytokines and other inflammatory mediators after LPS exposure contributes to SIRS and is central to the pathogenesis of septic shock induced by gram-negative bacteria.9,10,15,16
Another important endotoxin-binding protein found in human plasma is bactericidal/permeability-increasing protein (BPI), which is produced by neutrophils and is found in greatest concentrations in the azurophilic (primary) granules.17 Despite its 45% primary amino acid sequence homology with LBP, BPI has distinctly antagonistic actions with respect to LPS handling. BPI binds with high affinity to LPS and inhibits LPS delivery to CD14. BPI competes with LBP for LPS binding in biologic fluids.17 The relative concentrations of these two endotoxin-binding proteins primarily determine the net effect of LPS release.
In human plasma, the concentration of LBP is two to three orders of magnitude higher than that of BPI, whereas the opposite relative concentrations are found in neutrophil-laden abscess cavities, where BPI is the dominant endotoxin-binding protein.17 This favors LPS-activating activity in the plasma and LPS-inhibitory activity in abscess cavities. Thus, BPI functions as an endogenous antiendotoxin molecule; it may become a component of treatment of endotoxin-induced injury. A recombinant fragment of human BPI has been shown to reduce amputation rates and improve neurologic recovery rates in children with severe meningococcal sepsis.18 Other potential clinical uses of BPI as an antiendotoxin therapy are under consideration, including its use for neutropenic sepsis and for early intervention in suspected meningococcemia.
The phenomenon of endotoxin tolerance (or reprogramming) has been well characterized in experimental models of sepsis and probably also occurs in human sepsis. The term endotoxin tolerance refers to the desensitization to endotoxin-induced lethality that occurs after a priming (small) dose of endotoxin is administered, before a challenge dose of endotoxin that would otherwise be lethal. Endotoxin tolerance is characterized by marked attenuation of innate immune responses and is primarily mediated at the transcriptional level, with down-regulation of genes encoding for inflammatory cytokines and other acute-phase proteins.19 The precise molecular mechanism is not fully characterized. The desensitizing dose of endotoxin induces endogenous corticosteroids and anti-inflammatory cytokines such as interleukin-10 (IL-10); decreases cell surface expression of TLRs and other pattern recognition receptors; alters nuclear translocation of signal transduction molecules; and decreases the stability of messenger RNA (mRNA) for cytokine genes.10
The development of agents that act as inhibitors of LPS signaling pathways and of agents that enhance the clearance of LPS from the circulation are potentially important therapeutic strategies for treatment of septic shock.
Another important microbial mediator in the pathogenesis of septic shock is bacterial superantigen. Superantigens comprise a diverse group of protein-based exotoxins from streptococci, staphylococci, and other pathogens that all share the capacity to bind to specific sites on major histocompatibility (MHC) class II molecules on antigen-presenting cells (APCs) and activate large numbers of CD4+ T cells, bypassing the usual mechanism of antigen processing and presentation.
Conventional bacterial antigens are internalized by APCs and undergo limited proteolysis followed by processing within the endosomal component of the macrophage or dendritic cell. Appropriate peptide sequences of these microbial antigens (epitopes) are inserted into the central groove of MHC class II molecules and are then presented on the cell surface of APCs. Specific CD4+ T cells that recognize the unique epitope are then activated. Clonal expansion of this small subset of T cells results in a physiologic immune response to the newly introduced antigen. Superantigens, in contrast, do not require intracellular processing by APCs. Superantigens bind directly to MHC class II molecules adjacent to the epitope-specific peptide groove on APCs. Superantigens also bind and crosslink the APCs to a limited number of Vβ regions of the T cell receptor on CD4+ T cells. This bridging complex brings CD4+ T cells and macrophages into close proximity, which activates both the monocyte-macrophage and T cell populations.
Whereas a conventional peptide antigen stimulates only about one in 105 circulating lymphocytes that can recognize each unique structural epitope, a superantigen (e.g., toxic shock syndrome toxin-1 from Staphylococcus aureus, which binds to the Vβ2 region of T cells) can stimulate up to 10% to 20% of the entire circulating lymphocyte population. This results in excessive activation of both lymphocytes and macrophages, which, in turn, leads to the uncontrolled synthesis and release of inflammatory cytokines. Superantigen-induced immune activation may terminate in septic shock if the process is left unchecked. Polymicrobial infections with pathogens that release both bacterial superantigens and endotoxin may be particularly injurious to the host; the toxicity of bacterial endotoxin is greatly enhanced by superantigens that prime the immune system to react to endotoxin in an overly sensitized manner [see Figure 1].20
Figure 1. Interactions between bacterial endotoxin and bacterial superantigens are shown. (ICE—interleukin-1β—converting enzyme; IFN-γ—interferon-gamma; IL—interleukin; LBP—LPS-binding protein; TNF-α—tumor necrosis factor-α)
Other Microbial Mediators
Peptidoglycan from the cell wall of bacteria, capsular antigens, lipoteichoic acid, lipopeptides, flagellin, microbial DNA, microbial toxins, and procoagulant substances produced by microbial pathogens all contribute to the pathogenesis of sepsis. Peptidoglycan, lipoteichoic acid, and lipopeptide from gram-positive bacteria activate inflammatory cells in a manner comparable to that observed with bacterial endotoxin.20Both gram-positive and gram-negative bacteria can activate mononuclear cells, although the level of activation is quantitatively less with gram-positive components.20 Recognition of the unmethylated CpG motifs found in bacterial DNA also results in a vigorous inflammatory response.9
Moreover, gram-positive bacterial and fungal pathogens induce systemic hypotension, resulting in redistribution of blood flow and in splanchnic vasoconstriction. The ischemia and subsequent reperfusion of the gastrointestinal tract disrupts the intestinal mucosal barrier to bacterial products. Translocation of microbial components such as bacterial endotoxin occurs from the GI tract to the circulation during periods of severe stress and hypoperfusion of the GI mucosa.19,20 Gut-derived microbial mediators may play a pathogenic role in the ongoing inflammatory process with systemic hypotension and are targets for research in the management of sepsis.
Pathogenesis: Host-Derived Mediators
Inflammatory cytokines play a pivotal role in the pathogenesis of sepsis. In animal studies, the administration of tumor necrosis factor-α (TNF-α), an endogenous monocyte-macro phage-derived protein, has been shown to have lethal consequences; in human volunteers, dramatic hemodynamic, metabolic, and hematologic changes are observed after administration of TNF-α. The injurious effects of systemic levels of IL-1β have also been demonstrated.
The major inflammatory cytokines, TNF-α and IL-1β, induce their hemodynamic and metabolic effects in concert with an expanding group of host-derived inflammatory mediators that work in a coordinated fashion to produce the systemic inflammatory response [see Figure 1 andTable 3].21 The cytokine system functions as a network of communication signals between neutrophils, monocytes, macrophages, and endothelial cells. Autocrine and paracrine activation results in synergistic potentiation of the inflammatory response once it is activated by a systemic microbial challenge (e.g., endotoxemia). Much of the inflammatory response is localized and compartmentalized in the primary region of initial inflammation (e.g., lung tissue or the GI tract). If left unchecked, the inflammatory response spills over into the systemic circulation, resulting in a generalized reaction and culminating in diffuse endothelial injury, coagulation activation, and septic shock. The endocrinelike effect of the systemic release of cytokines and chemokines drives the inflammatory process and causes coagulation activation throughout the body.9,22
Table 3 Host-Derived Inflammatory Mediators in Septic Shock
Perhaps the most compelling evidence in humans of the pathophysiologic significance of systemic cytokine release was confirmed by accident in an unfortunate group of healthy volunteers given an experimental anti-CD28 monoclonal antibody.22 This antibody was designed to specifically expand T helper type 2 cells and T regulatory cells. The antibody was well tolerated in animal studies but was nearly fatal when given to humans. All six normal volunteers given this antibody developed shock, disseminated intravascular coagulation (DIC), and multiorgan failure within a few hours of receipt of the antibody. They developed striking elevations in IL-1, TNF, IL-8, interferon gamma (IFN-γ), and a myriad of other cytokines and chemokines immediately after receiving the antibody. All the volunteers survived, but they required long stays in the intensive care unit. This episode attests to the clinical relevance of the cytokine storm that often accompanies septic shock.22
The multitude of inflammatory cytokines and chemokines found in excess quantities in the bloodstream in patients with septic shock is impressive and is matched only by an equally daunting group of anti-inflammatory mediators [see Table 3]. The inflammatory mediators tend to predominate locally and in the early phases of sepsis (the first 12 to 24 hours), whereas the endogenous anti-inflammatory components often prevail systemically in the later phases of sepsis.21 Monocyte-macrophage-generated cytokines and chemokines primarily drive the early septic process; the lymphocyte-derived cytokines and interferons become important in the regulation of later phases of sepsis and may ultimately determine the outcome in septic shock.
CD4+ T Cells
Important functional differences exist within CD4+ T cells. Activated, yet uncommitted, CD4+ T cells (TH0 cells) have two major pathways of functional differentiation. TH0 cells exposed to IL-12 in the presence of IL-2 are driven toward a TH1-type functional development. These cells produce large quantities of IFN-γ, TNF-α, and IL-2 and promote an inflammatory, cell-mediated immune response. In contrast, TH0 cells exposed to IL-4 will preferentially develop into a TH2-type phenotype; TH2 cells secrete IL-4, IL-10, and IL-13. These cytokines promote humoral immune responses and attenuate macrophage and neutrophil activity.
The T helper cell response tends to polarize to a TH1 or TH2 response over time. TH1-type cytokines suppress the expression of TH2-type cytokines (e.g., IFN-γ inhibits the synthesis of IL-10). Conversely, IL-10 from TH2 cells is a potent inhibitor of TNF-α and IFN-γ synthesis by TH1 cells. The nature of the initial lymphocyte response is critical in driving the system toward a TH2-type or TH1-type response. Similar forms of functional differentiation may exist for CD8+ cells, as well (i.e., CD8+ type 1 and type 2 cells).23
This process of functional differentiation is clinically relevant because sepsis is often accompanied by a late TH2-type response after an initial septic insult. The stress hormone response in septic shock—expression of adrenocorticotropic hormone (ACTH), corticosteroids, and catecholamines—promotes a TH2 response after systemic injury. CD4+ cells are selectively depleted by apoptosis in severe sepsis, further limiting cell-mediated immunity and T helper cell capacity.19 This results in a phase of relative immune refractoriness (immune paralysis) in which the patient may be at increased risk for secondary bacterial or fungal infection. This pathophysiologic state is associated with endotoxin tolerance; anti-inflammatory cytokine synthesis; and deactivation of monocytes, macrophages, and neutrophils. Methods to detect this immunosuppressed state and restore immune competence are under active clinical investigation. Patients with depressed expression of MHC class II antigens (e.g., HLA-DR) on the cell surface of macrophages may be in a functionally immunosuppressed state and could benefit from immune adjuvants such as IFN-γ treatment or similar strategies to support immune function.
The Coagulation System
Activation of the coagulation cascade and generation of a consumptive coagulopathy and diffuse microthrombi are well-recognized complications of severe sepsis. Studies of endotoxin challenge and TNF challenge in normal human volunteers indicate that the extrinsic pathway (tissue factor pathway) is the predominant mechanism by which the coagulation system is activated in human sepsis.24 The contact factors in the intrinsic pathway are also activated, which secondarily initiates vasodilation through the generation of bradykinin.25 Activation of intravascular coagulation results in microthrombi and may contribute to the multiorgan failure that occurs in septic patients. Depletion of coagulation factors and activation of plasmin, antithrombin III, and protein C may subsequently lead to a hemorrhagic diathesis. Depletion of these endogenous anticoagulants may secondarily lead to a procoagulant state and portends a poor prognosis.26
The current interest in the administration of tissue factor pathway inhibitor, activated protein C, and antithrombin III for the treatment of sepsis stems from the potential therapeutic value of regulation of the coagulation system in sepsis. A phase 3 clinical trial in which 1,690 patients were treated with recombinant human activated protein C (drotrecogin alfa activated) was stopped when an interim analysis revealed a survival benefit for patients receiving activated protein C; mortality was 24.7% in treated patients versus 30.8% in placebo recipients (P < 0.005).27 Despite initial disappointments with antithrombin III28 and tissue factor pathway inhibitor for sepsis,29 both therapeutic strategies are in clinical trials for severe infection and sepsis-induced coagulopathy at present.
Neutrophil-Endothelial Cell Interactions
The recruitment of neutrophils to an area of localized infection is an essential component of the host inflammatory response. Localization and eradication of invading microbial pathogens at the site of initial infection is the principal objective of the immune response to microbial pathogens. This physiologic process becomes deleterious if diffuse neutrophil-endothelial cell interactions occur throughout the circulation in response to systemic inflammation.
Complex mechanisms govern the migration of neutrophils from the intravascular space into the interstitium, where invasive microorganisms may reside [see Figure 2]. Activated neutrophils degranulate, exposing endothelial surfaces and surrounding structures to reactive oxygen intermediates, nitric oxide, and a variety of proteases. This process contributes not only to microbial clearance but also to diffuse endothelial injury in the setting of generalized systemic inflammatory responses. Regulation of neutrophil activity remains a potential target for therapeutic intervention in the management of sepsis.
Figure 2. Neutrophil—endothelial cell interactions in sepsis. (C—complement; ICAM-1—intercellular adhesion molecule-1; IL-1β—interleukin-1β; MCP-1—monocyte chemoattractant protein-1; PAF—platelet-activating factor; PECAM-1—platelet endothelial cell adhesion molecule-1; PSGL-1—P-selectin glycoprotein ligand-1; TNF-α—tumor necrosis factor-α)
Nitric oxide is a highly reactive free radical that plays an essential role in the pathophysiology of septic shock. It has a very short half-life (1 to 3 seconds), which tends to limit its activity to local tissues, where it is first generated by one of three isoforms of nitric oxide synthase. Regulation of the nitric oxide synthases is complex. Full expression of inducible nitric oxide synthase requires TNF-α, IL-1, LPS, and probably other regulatory elements.
Nitric oxide is the major endothelium-derived relaxing factor that initiates the vasodilation and systemic hypotension observed in septic shock. Within minutes of administration of an inhibitor of nitric oxide synthesis, blood pressure in hypotensive patients in septic shock moves toward normal levels.
The other major physiologic effects of nitric oxide in septic shock are increased intracellular killing of microbial pathogens and regulation of platelet and neutrophil adherence. Nitric oxide is a highly diffusible gas that does not require specific receptors to cross cell membranes. In the presence of superoxide anion, nitric oxide leads to the formation of peroxynitrite. The peroxynitrite subsequently decays into highly cytotoxic molecules such as hydroxyl radicals and nitrosyl chloride, which, in turn, initiate lipid peroxidation and cause irreversible cellular damage. Nitric oxide inhibits a variety of key enzymes in the tricarboxylic acid pathway, the glycolytic pathway, DNA repair systems, electron transport pathways, and energy-exchange pathways. Because of its potent reactivity, nitric oxide alters the function of many metalloenzymes, carrier proteins, and structural elements.
Like many other components of the host inflammatory response, nitric oxide may have both advantageous and disadvantageous properties in sepsis. Nitric oxide regulates microcirculation to vital organs and contributes to intracellular killing of microbial pathogens. However, excessive and prolonged release of nitric oxide results in generalized vasodilatation and the systemic hypotension of septic shock. For those reasons, nitric oxide has become a target for therapeutic strategies in the management of sepsis.30 Nonselective inhibitors of nitric oxide synthase, for example, have been shown to improve the hemodynamics of septic patients. Unfortunately, this finding was not confirmed in a phase 3 trial.31
Late Host-Derived Mediators
Of the myriad numbers of host-derived mediators induced in sepsis, at least two additional mediators deserve further comment when considering the pathogenesis of septic shock. Macrophage migration inhibitory factor (MIF) is a late mediator that activates immune cells, upregulates TLR4 expression, and contributes to lethal septic shock.32 This corticosteroid-regulated mediator promotes inflammation and has become a target for therapeutic agents in sepsis. High-mobility group box-1 (HMGB-1) protein also participates in late-onset inflammatory activities in septic shock. Inhibitors of HMGB-1 are under active investigation as new therapeutic agents in sepsis.33
Pathogenesis: Organ Dysfunction
The diffuse endothelial injury accompanying septic shock results in organ dysfunction distant from the original site of the septic insult. The signal that results in diffuse endovascular injury is thought to be relayed by plasma factors (e.g., inflammatory cytokines, complement, kinins, and other host-derived inflammatory mediators) or by cellular signals expressed by immune effector cells.
Inadequate blood supply to vital tissues produces MODS. The failure of the microcirculation to support tissue maintenance may be the result of hypoperfusion of capillary beds, redistribution of blood flow within vascular beds, functional arteriovenous shunting, obstruction of blood flow from microthrombi, plate let or white blood cell aggregates, or abnormal deformability of red blood cells. Direct endothelial injury from nitric oxide, reactive oxygen intermediates, inflammatory cytokines, and inducers of apoptosis may directly damage endothelial surfaces. Endothelial swelling from the movement of intravascular fluid into the extravascular and intracellular spaces mechanically obstructs the capillary lumen further limiting blood flow in the microvasculature.
Although the origin of multiorgan failure in sepsis is principally related to microvascular effects, myocardial performance and pulmonary function also diminish over the course of septic shock and may contribute significantly to the development of MODS. Myocardial contractility decreases in response to a variety of myocardial depressant factors found in the plasma of septic patients. TNF-α is a prominent cause of myocardial dysfunction; IL-1, IL-6, nitric oxide, and other host-derived inflammatory mediators may be contributing factors. Acute lung injury occurs in septic shock as a result of damage to the pulmonary vascular circulation and the alveolocapillary membranes. A supply-dependent dysoxia, along with altered capacity for oxidative phosphorylation (cytopathic hypoxia), likely contributes to tissue injury and multiorgan failure in sepsis.34
Diagnosis of Severe Sepsis
In his classic treatise on human nature (The Prince, circa 1505), Machiavelli states, “Hectic fever [i.e., sepsis by current consensus definitions] at its inception is difficult to recognize but easy to treat; left untended, it becomes easy to recognize but difficult to treat.” This statement is as true today as it was 500 years ago. Fully developed septic shock is a readily apparent clinical syndrome that is seldom confused with other pathologic states. However, the early phases of septic shock may be quite subtle even in carefully monitored patients. Although fever is characteristic, hypothermia may occur and connotes a poor prognosis. Unexplained tachycardia and tachypnea are often part of the systemic inflammatory response seen in sepsis. It is important to note that many noninfectious diseases may masquerade as sepsis—including, but not limited to, deep vein thrombosis, acute pancreatitis, pulmonary emboli, myocardial infarction, blood transfusion reactions, and organ transplant rejection. A systematic search for infection should include a thorough physical exam, review of pertinent radiographic studies, and microbiologic studies. The three most common sites of infection, in descending order of frequency, are the lung, abdomen, and genitourinary tract. In approximately 30% of cases a causative organism and focus of infection is never found.27
Laboratory Indicators of Severe Sepsis
Many laboratory value aberrations may lead the clinician to the consideration of severe sepsis [see Table 4]. Either leukocytosis or leukopenia may occur in sepsis. Thrombocytosis may occur early as an acute phase response; thrombocytopenia may occur as a late, ominous sign. Overt DIC (i.e., with a markedly prolonged prothrombin time, decreased fibrinogen level, elevated fibrin split products, and thrombocytopenia) occurs in severe sepsis, but more subtle findings of microvascular dysfunction, including decreases in antithrombin and protein C activity, may precede and portend the development of organ dysfunction.35 Unexplained lactic acidosis as a sign of global tissue hypoperfusion may occur. Unexplained hypophosphatemia has been described in sepsis. Animal studies have indicated an inverse correlation of high inflammatory cytokine levels and low serum phosphate levels.36 C-reactive protein (CRP) levels and the erythrocyte sedimentation rate (ESR) may be elevated in sepsis, but these findings are not specific.
Table 4 Standard Laboratory Values in Sepsis
Novel Diagnostic Techniques
Rapid Urinary Antigen Test for Streptococcus pneumoniae
Community-acquired pneumonia (CAP) is the most common cause of severe sepsis, with Streptococcus pneumoniae the most common pathogen.37 Obtaining high-quality sputum samples for etiologic diagnosis is quite difficult in CAP.38 An immuno chromatographic assay (Binax NOW) that detects the C-polysaccharide from the cell wall of S. pneumoniae on unconcentrated urine samples in as little as 15 minutes is commercially available. In clinical studies, the assay has demonstrated limited sensitivity but high specificity. The assay has increased sensitivity in high-risk patients and in bacteremic patients, who are likely to go on to develop sepsis.39 Its best use appears to be in critically ill patients from whom a good quality sputum sample cannot be obtained.
Endotoxin Activity Assay
Endotoxin is a mediator of severe sepsis [see Bacterial Endotoxin, above]. Endotoxin given in small doses to healthy human volunteers can generate the clinical, inflammatory, and coagulopathic features of sepsis.40 High levels of endotoxin in patients with severe sepsis correlate with the presence of hypotension and are associated with worse prognosis.41 Until recently, measurement of endotoxin levels has been with a Limulus amebocyte assay (LAL). This assay is time consuming, and is hampered by the problems of plasma protein inhibition and cross-reacting elevation by fungal elements.42,43 A more contemporary approach is a rapid endotoxin assay (Endotoxin Activity Assay [EAA]), which is approved by the Food and Drug Administration for assessing the risk of developing severe sepsis in ICU patients. Health Canada has also approved this assay for the detection and ruling out of gram-negative infections. The EAA relies on priming of the endogenous neutrophil population by circulating endotoxin. This 1-hour assay measures the oxidative burst that is produced by activated neutrophils and compares this degree of activity to a blank, as well as a sample that is maximally activated by endotoxin.44 In one study, elevated endotoxin levels were found in 58% of patients admitted to a mixed surgical-medical ICU, irrespective of the reason for admission, and in 85% of the patients with severe sepsis.45 Endotoxin levels were elevated to 0.4 EA units, which is approximately equivalent to more than 50 pg/ml. The EAA assay was superior to the LAL in the detection of sepsis and gram-negative bacteremia.46
Procalcitonin, the propeptide of calcitonin, is normally produced by C cells in the thyroid. In septic patients, procalcitonin is generated by numerous extrathyroidal tissues; its precise origin in this situation is unclear.47 Procalcitonin has attributes that make it a potential marker for sepsis. It has a long half-life (approximately 24 hours), and measured levels increase from undetectable to over 100 ng/ml during the course of septic shock. Procalcitonin levels do not become elevated as rapidly as IL-6 or IL-8 levels; elevated levels of procalcitonin are seen 4 to 6 hours after a systemic challenge with endotoxin or other septic stimuli. The FDA has approved the procalcitonin assay as an aid in the risk assessment of patients with sepsis. Assay time is less than 20 minutes, and results are available in 1 hour.
Procalcitonin measurement has been shown to distinguish between noninfectious and bacterial causes of inflammation and also between bacterial and viral causes of inflammation in diverse patient populations, including trauma patients, pediatric patients, patients with pancreatitis, and those with acute respiratory distress syndrome (ARDS).48,49,50,51 Meta-analysis of studies in medical as well as surgical patients has demonstrated that procalcitonin has greater sensitivity and specificity than CRP.48,52
In patients with CAP and other lower respiratory infections, procalcitonin levels may be used to safely shorten the duration of antibiotic exposure.53,54 Studies have used a procalcitonin level below 0.25 µg/L as an indication for discontinuing antibiotics.53,54
Activated Partial Thromboplastin Time Waveform Analysis
The complex of CRP and very low density lipoprotein produces an abnormality of the optical transmission waveform obtained during measurement of the activated partial thromboplastin time on a specific photometric hemostasis autoanalyzer.55 This biphasic waveform abnormality is associated with the clinical diagnosis of DIC.56 Three studies have confirmed the association of this abnormal waveform with the diagnosis of severe sepsis,57,58,59 and two of those studies demonstrated that the waveform can help predict mortality in severe sepsis. Although the abnormal waveform reliably identifies severe sepsis and septic shock, it does not distinguish between infectious SIRS (sepsis) and noninfectious SIRS, as does procalcitonin. The diagnostic accuracy of the abnormal waveform for severe sepsis was similar to that of procalcitonin and CRP, but the waveform had the advantage of a higher negative predictive value.
Triggering Receptor Expression on Myeloid Cells
Triggering receptor expression on myeloid cells (TREM-1) is a member of the immunoglobulin superfamily and is expressed on the surface of neutrophils and of certain subpopulations of monocytes. TREM-1 is upregulated in the setting of bacterial and fungal infection, and, by complexing with its ligand, it acts synergistically with LPS to induce cytokine production.60 A soluble form of TREM may be released during infection. Soluble TREM at a cutoff of 60 ng/ml appears to have a high sensitivity and specificity in differentiating noninfectious SIRS from sepsis.61 In one study, TREM had a slightly higher discriminative value than procalcitonin.62 In a prospective study that compared soluble TREM-1 with multiple clinical laboratory values in mechanically ventilated patients with suspected pneumonia, the presence of soluble TREM-1 proved the strongest independent predictor of pneumonia (odds ratio, 41.5).63 However, soluble TREM-1 measurement is not currently available as a clinical laboratory assay.
Severe Sepsis and Multiple Organ Dysfunction
The hallmark of severe sepsis is the development of organ dysfunction. Clinical and pathologic findings are specific to the organs involved [see Table 5]. Confusion or decreased sensorium, unexplained decrease in urinary output, sudden onset of jaundice, excess bleeding at venipuncture sites, worsening hypoxemia, or sudden unexplained hypotension may indicate the development of severe sepsis. It is essential that clinicians recognize these early signs and symptoms, because successful management of septic shock depends on early recognition and appropriate intervention. The mortality in severe sepsis increases with an increasing number of organ dysfunctions.2
Table 5 Multiple Organ Dysfunction Syndrome in Severe Sepsis
Septic shock or cardiovascular dysfunction occurs in approximately 60% to 70% of patients with severe sepsis. Hemodynamic findings of septic shock are summarized [see Table 6]. The most common hemodynamic findings in early septic shock are a high cardiac output and a low systemic vascular resistance state, with initial maintenance of the systolic blood pressure as the heart attempts to compensate for the loss of systemic vascular tone. Myocardial performance is diminished even in the early phases of septic shock.64 Without adequate intervention, circulating blood volume is continually lost into the interstitial space and intracellular locations. The heart can no longer compensate sufficiently, and systolic hypotension results. Deterioration of myocardial performance, accompanied by diffuse vasoconstriction, marks the late refractory state of septic shock.
Table 6 Hemodynamic Findings in Sepsis
Acute Respiratory Distress Syndrome
ARDS remains a major cause of morbidity and mortality in septic shock.65 Increased capillary permeability in ARDS patients results in pulmonary edema, which manifests clinically as dyspnea and cough; a standard anteroposterior chest x-ray typically shows bilateral, symmetrical alveolar opacities in all four quadrants. Additional diagnostic criteria for ARDS include the absence of clinical evidence of left atrial hypertension, the requirement for intubation and positive pressure ventilation, and a ratio of the partial pressure of oxygen in arterial blood to the fractional concentration of oxygen in inspired air (PaO2:FIO2) of 200 or less.66
Management of Severe Sepsis and Septic Shock
As soon as septic shock is recognized, the clinician must simultaneously initiate several forms of therapy. Appropriate antibiotic therapy must be started, and the source of infection controlled. To maximize oxygen delivery and support failing organs, the clinician should institute measures that include fluid resuscitation and use of vasopressors, blood transfusions, and inotropic agents. Patients who have ARDS or who are developing it are managed with low tidal volume ventilation to minimize damage. Consideration is also given to the use of recombinant human activated protein C and low-dose corticosteroids to reverse the pathophysiology of severe sepsis. Tight glycemic control is utilized in an effort to minimize septic complications, and specific nutrition strategies are used to prevent immunoparalysis or dampen inflammation.
In 2004, eleven societies involved in the care of the critically ill collaborated to produce guidelines for the care of patients with severe sepsis (the Surviving Sepsis Campaign).67 These guidelines are available online, through the National Guideline Clearinghouse (http://www.guideline.gov/summary/summary.aspx=doc_id=4911&nbr=3508). The individual guidelines contained in this document are not universally agreed upon. The effect on mortality of combining the different treatments is also not known but is currently being studied.68
Antibiotic therapy, delivered intravenously, should be started within an hour after severe sepsis has been recognized, once appropriate culture specimens have been obtained. Blood cultures should be drawn percutaneously and through all vascular access devices that have been in place for at least 48 hours; cultures of other body fluids or wounds should be taken as appropriate.67
Antibiotic therapy with appropriate agents, given at the appropriate time, provides a survival advantage for patients with severe sepsis.69,70Animal and human studies have demonstrated an incremental decrease in survival for each hour of delay in the administration of antibiotic therapy from the onset of septic shock.71,72
The most appropriate choice of antibiotics for sepsis depends on the site of infection; susceptibility patterns of microbial pathogens within a given institution; prior antimicrobial exposure; and the presence or absence of pregnancy, hepatic and renal function, and history of drug allergy. Empirical antibiotic regimens vary with the site of infection [see Table 7]. Although the combination of a β-lactam agent with an aminoglycoside has been used for serious gram-negative infections, meta-analyses of trials with carbapenems, third- and fourth-generation cephalosporins, and extended-spectrum penicillins with β-lactamases have shown these agents to have comparable efficacy and less nephrotoxicity.73 Consideration should be given to empirical treatment of methicillin-resistant S. aureus (MRSA), depending on the clinical context—especially in view of the increasing incidence of infections from community-acquired MRSA (CA-MRSA). Fungal infections are causative in approximately 5% of cases of severe sepsis.2 Empirical antifungal therapy in patients with severe sepsis is not routinely recommended but is indicated for patients with well-described risk factors for fungemia, including a history of receiving multiple antibiotics for multiple days, the presence of a central venous catheter for total parenteral nutrition or hemodialysis, or Candida previously isolated from multiple anatomic sites.74 In vitro studies have demonstrated differential endotoxin release and cytokine production with various classes of antibiotic therapy, but this has not been shown to affect outcome and should not factor into decisions regarding agent selection.75An exception to this rule would be the use of clindamycin in cases of streptococcal and staphylococcal toxic shock syndrome, to minimize toxin production.
Table 7 Suggested Empirical Antibiotic Choices in Severe Sepsis
Source Control of Infection
Controlling the source of infection in sepsis is as important as the administration of appropriate antimicrobial therapy. Source control of infection includes the drainage of intra-abdominal abscesses and empyemas, surgical management of gastrointestinal perforation, the removal of infected prosthetic devices, vascular and urinary catheters, and the debridement of necrotic bowel and soft tissue.76
Drainage of intra-abdominal collections can often be accomplished with catheters placed under computed tomographic guidance and does not always require an operative procedure. In cases of suspected necrotizing fasciitis, early operative diagnosis and debridement improve outcome.77
Optimizing Tissue Oxygenation
Resuscitation is one of the immediate first steps in treatment of severe sepsis. Inadequate oxygen delivery for meeting tissue oxygen demand can be manifested by decreased blood pressure or increased serum levels of lactic acid. Reasonable goals for the first 6 hours of resuscitation include the following: a central venous pressure (CVP) of 8 to 12 mm Hg; a mean arterial pressure of at least 65 mm Hg; urine output of at least 0.5 ml/kg/hr or higher; and a central or mixed venous oxygen saturation of 70% or higher.67
The use of an algorithmic protocol (early goal-directed therapy [EGDT]) to meet a resuscitation goal of a central venous oxygen saturation of 70% has been shown to decrease mortality from severe sepsis. In a clinical trial, patients with septic shock and lactic acidemia were randomized to receive either standard therapy or EGDT. Patients in the EGDT arm received crystalloid fluid boluses every 30 minutes in an attempt to achieve a CVP of 8 to 12 mm Hg. If the mean arterial pressure remained below 65 mm Hg, vasopressors were added. If after these maneuvers the central venous oxygen saturation remained below 70%, red blood cell transfusions were given to achieve a hematocrit of 30%. If targets were still not met, dobutamine was added. In-hospital mortality was significantly lower in the EGDT group than in the standard treatment group (30.5% versus 46.5%, P = 0.009).78 Subsequent work has confirmed the validity of EGDT with regard to end points, outcome results, and cost effectiveness.79 Whether this approach can be generalized to other hospitals, and which component (or components) of the EGDT protocol is protective, are questions still needing answers, however.
Although debate continues regarding the appropriateness of colloid versus crystalloid fluids, the lack of clear evidence of benefit of colloid agents (e.g., albumin, dextran, and plasma expanders) and their high cost have generally resulted in the use of saline solutions for volume expansion. A randomized clinical trial comparing 4% albumin with normal saline for resuscitation (the SAFE trial) in 7,000 ICU patients yielded similar 28-day mortalities in the treatment arms,80 and a 2004 meta-analysis comparing a variety of colloids with crystalloids for fluid resuscitation of critically ill patients found no survival benefit with the use of colloids.81
The optimal amount of fluid for resuscitation of patients in septic shock remains a source of controversy. A delicate balance is required between maintenance of tissue perfusion and prevention of fluid overload, with its attendant risk of lung injury. In the EGDT study, patients in the EGDT arm received statistically more volume than the standard treatment group in the first 6 hours of the emergency department stay but similar volumes over the first 72 hours of hospitalization.78 In a large study of sepsis in European ICUs (Sepsis Occurrence in the Acutely Ill Patient [SOAP] study), a positive fluid balance over the first 72 hours was associated with a higher mortality.82 In the 2006 Fluid and Catheter Treatment Trial (FACTT), patients with acute lung injury or ARDS were randomized to receive either a liberal (CVP of 10 to 14 mm Hg or pulmonary artery occlusion pressure [PAOP] of 14 to 18 mm Hg) or conservative (CVP of less than 4 mm Hg or PAOP of less than 8 mm Hg) fluid therapy regimen. Mortality in the two treatment arms was similar at 60 days, but lung function and the number of ICU and ventilator days were better in the conservative fluid management arm.83 The preponderance of data suggests that the best approach is to use liberal fluid resuscitation in the acute phase of sepsis, followed by conservative fluid therapy in the established phase of acute lung injury. Fluid requirements in the acute phase can be high; patients may require as much as 6 to 10 L of crystalloid fluids during the first 24 hours of resuscitation.84
When treatment with fluids alone fails to produce hemodynamic improvement, vasopressor agents are often employed to reestablish systemic arterial blood pressure. Vasopressors should be titrated to the minimum dosage required to support perfusion of major organs, as evidenced by urine flow; in some patients, this can be achieved with a mean arterial pressure of 60 or 65 mm Hg.84
Dopamine, epinephrine, norepinephrine, phenylephrine, and vasopressin have been used to reverse hypotension in the setting of septic shock. The use of any of these agents in septic shock carries with it certain risks and should be reserved for patients with significant hemodynamic instability that is unresponsive to fluid therapy.
Dopamine and norepinephrine are usually considered first-line agents in septic shock.67 Lower doses of dopamine increase the cardiac index, whereas higher doses increase the systemic vascular resistance through the drug's effects on alpha adrenergic receptors in the peripheral circulation. Norepinephrine is a potent vasoconstrictor with minimal effects on the cardiac index. Dopamine historically had been favored over norepinephrine because of its presumed favorable effects on renal perfusion and the concern that norepinephrine would lead to digital ischemia. These concerns have not been realized in human and animal trials. In one clinical trial of fluid-resuscitated patients with septic shock, norepinephrine reversed hypotension to a greater extent than dopamine and was associated with a survival advantage.85Norepinephrine is less likely than dopamine to cause tachyarrhythmia and to interfere with the hypothalamic-pituitary axis.86,87 A meta-analysis failed to show the benefit of so-called low-dose dopamine for the prevention of acute renal failure, and this practice is not recommended.88
Second-line agents that have been used in septic shock that is refractory to dopamine and norepinephrine have included epinephrine, phenylephrine, and vasopressin. Epinephrine increases arterial pressure by increasing the cardiac index and stroke volume. Its utility is hampered, however, by its negative effects on gastric blood flow and lactate levels.89 Phenylephrine, a pure alpha adrenergic agonist, has been used in refractory shock but also has negative effects on splanchnic blood flow and has the potential to decrease cardiac output.90Vasopressin has become the favored agent in cases of shock that are refractory to dopamine, norepinephrine, or both.67 The rationale for vasopressin use includes the finding of low levels of endogenous vasopressin observed in prolonged septic shock and the ability of vasopressin to increase vascular tone through vascular V1 receptors.91,92 Small studies in patients with septic shock have demonstrated vasopressin's ability to increase arterial pressure and to allow dopamine and norepinephrine to be tapered.93,94 A large head-to-head trial of vasopressin and norepinephrine, the Vasopressin And Septic Shock Trial (VASST), has been completed in Canada; results should be available in the near future.
Low-Dose Corticosteroid Therapy for Septic Shock
Corticosteroids have long been considered to be of potential value in the treatment of severe sepsis because of their anti-inflammatory properties and positive effects of vascular tone, as well as their protective effect in animal models of sepsis. Recent studies have also suggested that a state of relative adrenal insufficiency and glucocorticoid resistance occurs during sepsis and is associated with a poor outcome. Adrenal function can be assessed by giving a 250 µg dose of synthetic ACTH and measuring the cortisol level 30 to 60 minutes later; normally, cortisol levels will increase by 9 µg/dl or more. In one study, 54% of 189 patients with septic shock were found to have relative adrenal insufficiency on the basis of this test.95 Although many trials of short-course, high-dose corticosteroids in sepsis have not demonstrated benefit, a meta-analysis of trials involving long-course, low-dose corticosteroids (≤ 200 mg/day of hydrocortisone for 5 or more days) showed a statistically significant reduction in 28-day all-cause mortality and an improvement in time to shock reversal.96 In the largest of these studies, patients were randomized within 8 hours of the development of vasopressor-dependent septic shock to receive either 50 mg of hydrocortisone every 6 hours and 50 µg/day of fludrocortisone or placebo for 7 days. A survival advantage with corticosteroid therapy was observed only in patients with an impaired adrenal response to the ACTH test.97 In contrast, a meta-analysis of three trials with low-dose corticosteroids demonstrating benefit in mortality and shock reversal did not reveal a difference in benefit between ACTH test responders and nonresponders.98 Numerous questions remain regarding the use of corticosteroids, including the appropriate dose and duration of therapy, the necessity of giving fludrocortisone along with the glucocorticoid, and whether therapy should be given only to patients with relative adrenal insufficiency—and if so, how this condition is best defined. A large clinical trial on corticoid therapy of septic shock (CORTICUS) has been completed and should provide answers to these questions. Currently available data suggest that corticosteroid therapy be given to those patients with vasopressor-dependent septic shock, in dosages no greater than 300 mg/day of hydrocortisone or its equivalent. Treatment should be given for at least 5 days and should then be tapered to prevent rebound hypotension.99
Recombinant Human Activated Protein C
The interplay of inflammation and coagulation is important in the pathophysiology of severe sepsis. The body has endogenous anticoagulants, including activated protein C, antithrombin, and tissue factor pathway, that initially regulate excessive coagulation in sepsis but are rapidly depleted.100 Recombinant human activated protein C (rhAPC; drotrecogin alfa activated) is a molecule with antithrombotic, anti-inflammatory and profibrinolytic activities. This agent is given by continuous intravenous infusion at a rate of µg/kg/hr for a total of 96 hours. In a placebo-controlled trial of patients with severe sepsis, treatment with rhAPC decreased absolute mortality by 6% and the relative risk of dying by 20%.27 The treatment effect was largely confined to patients at high risk for dying, as indicated by an Acute Physiology and Chronic Health Evaluation II (APACHE II) score above 25 or by failure of two or more organs.101 The main adverse effect associated with rhAPC is bleeding. RhAPC is not approved for use in children, and it is not recommended in surgical patients with a single organ failure.
The role of blood transfusions in improving the oxygen-carrying capacity of blood has been the subject of considerable debate. Data from the Transfusion Requirements in Critical Care (TRICC) trial showed that maintaining a hemoglobin level of 7 to 9 g/dl and setting a transfusion threshold as low as 7 g/dl in volume-resuscitated patients is not associated with a worse outcome than maintaining a hemoglobin level above 10 g/dl.102 Blood transfusion may also increase the risk of nosocomial infection in the critically ill population.103This higher transfusion threshold stands in contrast to the EGDT transfusion threshold of a 30% hematocrit for maintaining venous oxygen saturation [see Optimizing Tissue Oxygenation, above]. Maintenance of a higher hemoglobin level may also be necessary in patients with severe coronary artery disease and those with severe hypoxemia.
A reasonable hypothesis is that tight regulation of blood glucose levels during critical illness would prevent complications, including severe infections, polyneuropathy, multiple organ failure, and death. In a study of surgical ICU patients, strict glycemic control (target glucose level of 80 to 110 mg/dl) was associated with improved survival, shorter ICU stays, and a lower incidence of bacteremia than in patients who received conventional care (target glucose level of 180 to 200 mg/dl).104 When this study was repeated in a medical ICU, however, a mortality benefit was observed only in the patients receiving strict glycemic control who remained in the ICU for more than 3 days.105Patients in the strict glycemic control arm did, however, have a lower incidence of renal injury and spent less time on mechanical ventilation than patients in the conventional arm. In a prospective observational study of ICU patients, regression modeling suggested a mortality benefit in maintaining a threshold glucose level of 145 mg/dl (8.0 mmol/L).106 Until more data are available, this more lenient threshold may be safer in a medical population with sepsis. Treatment of hyperglycemia in the hospitalized patient is discussed elsewhere [see 9:III Complications of Diabetes Mellitus].
Ventilator Management of ARDS
In the 1990s, a major finding in regard to ARDS management was the recognition of the hazard of providing excessive tidal volume through overly high ventilator settings. The resulting overdistention of airways can promote the progression of lung injury and the release of inflammatory mediators into the systemic circulation.107 The ARDS clinical trials network study demonstrated conclusively that low stretch tidal volume settings (6 ml/kg) are clearly superior to the previous conventional high tidal volume setting (12 ml/kg).108 A follow-up study attempted to determine the optimal level of positive end-expiratory pressure (PEEP) for patients receiving low tidal volume ventilation. In a clinical trial of 549 patients with acute lung injury or ARDS, clinical outcomes were similar in the low PEEP arm (< 14 cm H2O) and the high PEEP arm (> 14 cm H2O).109 A clinical study of the placing of patients in the prone position for ventilation in ARDS did not demonstrate a mortality benefit but found that prone positioning did lead to transient improvements in oxygenation.110
Nutritional support in septic shock has changed considerably over the past 2 decades. Reliance on total parenteral nutrition has given way to early and extensive use of enteral hyperalimentation. Enteral feeding of septic patients has been shown to benefit enterocyte function, help maintain the intestinal permeability barrier, and help prevent gut-derived endotoxin and cytokine generation.111 Nutritional supplementation with glutamine, arginine, and omega-3 fatty acids has experimental support and is increasingly being used in septic patients. Interpretation of clinical studies in the area of immunonutrition is hampered by the different formulas of ingredients being studied, as well as by the different patient populations being studied. Clinical studies of arginine-containing formulas have demonstrated a decreased rate of infectious complications but no mortality benefit. One study of an enteral formula containing eicosapentaenoic acid, gamma linoleic acid, and antioxidants (Oxepa) in patients with ARDS demonstrated that recipients had a shorter number of ventilator days and ICU days compared with patients receiving a control formula.112 The same formula was subsequently studied in a randomized, controlled trial of patients with sepsis and demonstrated an association with improved mortality as well as fewer ventilator and ICU days compared with the control formula.113 Larger studies will need to be conducted to confirm this benefit.
The provision of enteral care is discussed elsewhere [see 4:XIII Enteral and Parenteral Nutrition].
Additional Therapeutic Concerns
Sedation, renal replacement therapy, the prevention of deep vein thrombosis, stress ulcer prophylaxis, and the prevention of nosocomial infection are additional concerns in the care septic patients. Administration of sedatives in intermittent bolus doses may lead to a shorter duration of mechanical ventilation than uninterrupted continuous infusions. Patients with sepsis are at high risk for deep vein thrombosis and should receive prophylaxis with either heparin or pneumatic compression stockings. Prophylaxis against stress gastritis in the critically ill can be accomplished through the use of histamine2 blockers or proton pump inhibitors. Mechanically ventilated patients should be maintained semirecumbent at a 45° angle to prevent ventilator-associate pneumonia.67 Carts containing all the materials necessary for sterile insertion of central venous catheters should be used and maintained in all ICUs.114
Experimental Therapies for Septic Shock
Additional strategies for the treatment of septic shock are currently being studied, with numerous agents and devices in different stages of testing [see Table 8]. Target areas for these strategies include endotoxin, cell signaling, cytokines, complement, the coagulation cascade, immunoparalysis, apoptosis and mitochondrial dysfunction. The future therapy of sepsis will likely resemble cancer chemotherapy, with multiple agents used to target an individual patient's particular inflammatory phenotype and genotype.
Table 8 Experimental Therapies in the Treatment of Severe Sepsis
Figures 1 and 2 Hung. Assistance with tables: Gail Armstrong and Stacey Donfrancesco.
Assistanc with collection of cited literature: Phinnara Has.
Editors: Dale, David C.; Federman, Daniel D.