Chapter 73. The Pathophysiology of Critical Illness

1. Acute injuries of various etiologies (eg, trauma, infection, shock, etc) cause immediate local and systemic physiologic responses involving all major organ systems.

2. The timing and intensity of the physiologic response to injury is affected by the severity of the injury as well as host factors. An initial activation of the proinflammatory immune responses may be followed by late immunosuppression.

3. Nosocomial infections are a major cause of death of critically ill patients and can be reduced by adopting best practices of infection control.

4. The vascular endothelium is a major organ target of the initial response to injury that activates vasomotor, inflammatory, and procoagulant cascades.

5. Our increased understanding of the molecular biology of the immune response to injury will hopefully lead to new and effective therapies that may include oxygen or nitrogen free-radical scavenging, mitochondrial protection, prevention of apoptosis, and enhancement of the immune response.

The term critical illness defines a variety of clinical situations that are cared for primarily in intensive care units (ICUs) and have in common varying degrees of dysfunction of single or multiple organ systems and a guarded immediate prognosis. Anesthesiologists care for critically ill patients in the operating room, in the ICU, and during procedures in areas outside of the operating room, such as radiologic, endoscopic, and cardiac electrophysiology suites. Thus it is crucial that anesthesiologists have an in-depth understanding of the pathophysiologic processes encountered during critical illness to contribute effectively to the perioperative management of critically ill patients.

This chapter focuses on a number of basic elements of critical illness, including inciting factors, the transition to organ dysfunction, the role of the immune system, and current as well as potential future therapies for various aspects of critical illness. We analyze the progression of critical illness examining common types of primary injuries, the body's response to them, and the progression of the response either toward healing and recovery or to deterioration and death. Epidemiologic and therapeutic considerations are also reviewed. The chapter begins with a case report of a critically ill patient through different stages of her illness that are referred to in subsequent sections of the chapter.

A 77-year-old woman with a history of severe chronic obstructive pulmonary disease and hypertension was admitted to the surgical ICU from a general ward with acute respiratory distress. She had been recovering uneventfully from a pancreatic resection for cancer until that morning when she experienced nausea and malaise, then vomited a large amount of dark fluid, and immediately complained of shortness of breath.

Upon arrival at the surgical ICU, she was in obvious respiratory distress, with a marginal (87%-89%) arterial oxygen saturation (SpO2) on a nonrebreathing oxygen mask. Following a rapid-sequence tracheal intubation, a large amount of dark green material was suctioned from her airways, similar in appearance to that suctioned from her stomach. A bronchoscopy revealed diffuse staining of the tracheobronchial mucosa with biliary material, down to the lobar bronchi bilaterally. Over the next 4 to 6 hours, the patient's condition deteriorated. She required high levels of ventilatory support and large doses of intravenous (IV) infusions of vasopressors to maintain an adequate systemic arterial blood pressure. A pulmonary artery catheter revealed moderate to severe pulmonary artery hypertension and a low cardiac output, primarily a result of right ventricular dysfunction, which was confirmed by transthoracic echocardiography. She was treated with broad-spectrum antibiotics, recombinant human activated protein C, and inhaled nitric oxide (NO), and was ventilated using a "lung-protective" strategy of low tidal volumes at a moderate level of positive end-expiratory pressure (7-10 cm H2O). Over the next few days, she was extremely unstable from the cardiovascular and pulmonary standpoints: She had three episodes of pulseless electrical activity requiring cardiopulmonary resuscitation and developed severe hypoxemia with minimal stimulation, which greatly impeded basic care.

By day 10 her cardiovascular and respiratory function began to stabilize, and a computed tomogram of her chest was obtained (Fig. 73-1). Subsequently, hemodynamic and ventilatory support was tapered, and she was transitioned to assisted ventilation; vasopressors were discontinued, and she started to interact with ICU personnel. On ICU day 13, she had a bedside percutaneous dilatational tracheostomy. On ICU day 15 her bronchial secretions became purulent; she developed a fever and leukocytosis, and required increased ventilatory support. She was diagnosed with ventilator-associated pneumonia (VAP) from Pseudomonas aeruginosa based on clinical and bacteriologic criteria. She was again treated with antibiotics, and she gradually improved. She was discharged from the surgical ICU on day 24 and transferred to a semi-intensive (step-down) respiratory unit to complete the process of weaning from mechanical ventilation.

Thirty-eight days after ICU admission, and 45 days after her surgery, she was transferred to a rehabilitation facility. Although she was sustaining unassisted breathing, she was still extremely debilitated: she could not stand for more than a few minutes or walk more than a few steps and was unable to feed herself. In addition, she continued to require narcotic-based analgesics and antipsychotic medications to allow sleep at night and diminish her recurring episodes of delirium.

A variety of acute injuries can result in critical illness and the need for ICU care. In the following material, we classify these "stresses" as related to a few main categories: surgery and trauma, metabolic disorders, and infection. In many patients, as in the patient just described, multiple stresses (eg, surgery, inflammation, and infection) coexist.

Direct Tissue Injury during Surgery and Trauma

Direct tissue injury during surgery and trauma can produce major perturbations of the body's homeostasis, including hemodynamic instability, respiratory abnormalities, and marked fluid shifts. Pain, immobility, and prolonged bed rest can contribute to perioperative morbidity. Only a few decades ago, these phenomena constituted a formidable challenge for surgeons and anesthesiologists, and they made many surgical procedures complex and dangerous for the patient. Today, the physiologic changes associated with so-called "routine" surgery can be flawlessly managed by the proper administration of perioperative anesthesia care. However, certain situations still carry a significant risk of perioperative complications, including complex elective procedures (eg, pneumonectomy, thoracoabdominal aneurysm repair), major trauma, emergency surgery, and procedures performed in patients with significant comorbidity. These patients constitute most of the population admitted to a surgical ICU.

Trauma is the primary cause of death in individuals younger than 40 years in the United States.1 In underdeveloped countries, trauma victims may have limited or no access to the type of structured intensive care that could save thousands of lives every year. Traumatic injuries are classified as penetrating and blunt injuries (see Chapter 76). Penetrating injuries (stab and gunshot wounds) require immediate control of hemorrhage, drainage of blood or air under pressure, and rapid transfer to a trauma center. Immediate appropriate care increases survival and can reduce the need for intensive care.2 Blunt trauma (eg, motor vehicle accidents, falls) can produce a combination of injuries that may require immediate treatment, such as a ruptured spleen or an open bone fracture, and others that are best treated conservatively. A damage control approach minimizes the time of surgical intervention and emphasizes perioperative surgical ICU care, including volume resuscitation, temperature control, hemodynamic monitoring, nutrition, and protocols for the prevention of nosocomial infectious complications. The injuries associated with major surgery and trauma derive from three main mechanisms: hemorrhage, tissue injury, and hypoperfusion.

Hemorrhage causes an immediate physiologic response to preserve oxygen delivery to vital organs. Using the large venous reservoir of the splanchnic and cutaneous circulation, the body maintains venous return and thus cardiac output to a certain extent.3 In healthy individuals, symptoms of shock develop when acute blood loss exceeds 20% of the blood volume (approximately 1 L of blood in a 70-kg adult). Immediate resuscitation is necessary to prevent cardiovascular collapse and death. Even at the low hemoglobin (Hgb) levels that accompany this degree of bleeding, what limits the tolerability of hemorrhage is the reduced blood volume and cardiac output, not the decreased oxygen-carrying capacity of the blood. For example, a 30% to 40% blood loss would eventually reduce the plasma Hgb concentration to a level that still produces ample delivery of oxygen to the tissues. With an adequate cardiac output and arterial oxygen tension (PaO2), a Hgb concentration as low as 5 g/dL still provides ample oxygen delivery to tissues supplied by normal vessels (Table 73-1).

Implementing early and effective therapies carry great importance in determining the subsequent clinical course. Despite a long-standing belief that aggressive volume resuscitation should be included in the initial care of the bleeding patient, this approach is now being revised for several reasons. First, the infusion of large volumes of crystalloids can cause a dilutional coagulopathy, characterized by thrombocytopenia and decreased concentrations of circulating clotting factors.4 Counteracting these defects by infusing platelets and fresh-frozen plasma, in conjunction with massive crystalloid resuscitation,5 results in intravascular volume overload and in edema of organs. Second, there is evidence that massive volume resuscitation for penetrating trauma of the torso may exacerbate bleeding from the site of injury and actually increase mortality.6 Third, transfusion of blood products may have immunosuppressive effects leading to an increased risk of infection, cause acute lung injury/acute respiratory distress syndrome (ALI/ARDS), and death.7,8

Tissue injury produces local and systemic perturbations that are minimized during elective surgery by using proper anesthetic and surgical techniques but can produce major damage in the random, unprotected situation of a traumatic injury. Tissue damage, loss of perfusion, and cell necrosis result in the immediate release of tissue factors into the bloodstream, which trigger a systemic inflammatory response characterized by the synthesis and release of cytokines. Tumor necrosis factor (TNF)-α and interleukin (IL)-6 are among the earliest cytokines to appear in the systemic circulation immediately after tissue trauma, and they amplify the complex inflammatory response that is an integral part of critical illness.9

Tissue trauma of different types can produce both direct and systemic effects leading to critical illness. One well-documented example occurs in injuries to the lung. In our case report, the initial cause of the patient's complex critical illness was a chemical injury to the alveolar epithelium caused by the aspiration of gastric contents. A similar evolution to critical illness can occur with other lung injuries, such as a lung contusion from blunt chest trauma or a pneumonia. Despite the variety of inciting factors, the lung responds to acute injury in a stereotypical fashion, leading to the pathologic picture of diffuse alveolar damange10 and the clinical picture of ALI/ARDS, which is a syndrome combining local and systemic inflammation.11 Although not as well studied, a similar sequence of events appears to follow major acute injury to the kidneys12,13 and the gut.14

Hypoperfusion and ischemia may result from direct tissue trauma, impaired regional blood flow, or a generalized decrease of blood flow: a low-flow state. A severe reduction of oxygen supplied to the tissues is marked by the local production of lactate from anaerobic glycolysis and may cause or exacerbate metabolic acidosis. However, even when profound ischemia occurs, tissue viability can be reestablished. The effects of restoration of blood flow after ischemic injury has been best characterized in the heart, where reversal of arterial occlusion is common with advanced techniques of coronary revascularization.15 Myocardial stunning describes the acute ischemic dysfunction caused by acute cessation of regional coronary circulation and can be reversed by rapid restoration of blood flow.16 Myocardial hibernation, a chronic state of dysfunction caused by insufficient coronary perfusion, can also, but less predictably, be improved with coronary revascularization procedures.17 However, global hypoperfusion is often irreversible: In the "postresuscitation syndrome" that occurs following prolonged cardiopulmonary resuscitation, stunning of the cardiac muscle (stony heart)15 is largely responsible for the high fatality rate observed early after cardiopulmonary resuscitation, and it is characterized by severely depressed contractility and malignant arrhythmias.

Although restoration of tissue oxygen supply may prevent cell death and reestablish function, acute reperfusion of ischemic tissues can cause significant cellular injury, known as ischemia-reperfusion injury. In the absence of adequate oxygen supply, mitochondrial production of adenosine triphosphate (ATP) decreases, causing an increase in adenosine monophosphate and its by-product hypoxanthine, and the conversion of the enzyme xanthine dehydrogenase to xanthine oxidase. Upon reperfusion with oxygen-rich blood, hypoxanthine is oxidized by xanthine oxidase at a high rate, producing large quantities of cytotoxic oxygen free radicals, including molecular oxygen, superoxide, and hydrogen peroxide ions.18 Ischemia-reperfusion injury has been described in patients with the crush syndrome,19 organ transplantation,20 restoration of blood flow to limbs or splanchnic organs,21 and during experimental alterations of ventilation and perfusion of the lung.22,23

Metabolic Failure

Metabolic failure is a less common reason for admission to a surgical ICU. The most representative metabolic derangements include the consequences of severe malnutrition, such as extreme weight loss, kwashiorkor, and avitaminosis. These diseases are rarely seen in industrialized societies.

A different metabolic derangement, morbid obesity, has become more prevalent in the United States and other industrialized countries.24 Morbid obesity (defined as weight 30% in excess of ideal body weight, or a body mass index >40 [body mass index = weight in kg ÷ height in m2]) increases the risk of complications related to surgery and anesthesia. Morbidly obese patients are increasingly admitted to the ICU because of complications following bariatric surgery or unrelated events, such as trauma. Prevalent comorbidities of obesity include hypertension, diabetes mellitus, obstructive sleep apnea, chronic lung disease, and heart failure, which increase the rate of postoperative cardiovascular and respiratory complications. In addition, morbid obesity often presents logistical problems, such as the need for a special bed, a limited ability to perform important diagnostic tests such as magnetic resonance imaging and computed tomography scans, difficulty of vascular access and monitoring,25,26 and complex issues of proper drug dosing.27

Common metabolic derangements requiring ICU care include those related to chronic diseases such as diabetes mellitus, diabetic ketoacidosis, and hyperosmolar nonketotic coma. See Chapter 13 for management recommendations for diabetes mellitus and its complications.

Infection

Infection contributes importantly to both the onset and the progression of critical illness. In a survey of 198 European ICUs, sepsis was responsible for 25% of the ICU admissions, and 37% of ICU patients experienced sepsis and septic shock at some time during their stay.28 Infection is the leading cause of death in ICU patients. It is estimated that approximately 750 000 people develop sepsis annually in the United States, and 30% to 40% of these patients will die.29 The specific entity of severe sepsis following elective surgery has recently been reported to have an incidence of 0.5% to 1%, and a mortality rate that has decreased from 44% to 34% over the present decade.30 The presence of a proven infection in patients with ICU length of stays of more than 48 hours increases the mortality rate from 22% to 35%.31 The terms sepsis, severe sepsis, septic shock, and refractory septic shock describe the spectrum of severity of sepsis. Multiorgan dysfunction syndrome (MODS) is a common complication of sepsis and other inflammation-driven critical illnesses. MODS is associated with a high mortality. These terms were reviewed in a consensus conference of the Society of Critical Care Medicine and the American College of Chest Physicians32 and are thoroughly discussed in Chapter 72.

The most common community-acquired infections that require ICU admission include pneumonia, urosepsis, and peritonitis. Microorganisms involved in these infections are highly susceptible to antibiotic therapy when used properly. However, recent spread of multiresistant microorganisms outside health care institutions into the community has occurred, as exemplified by the case of methicillin- resistant staphylococcus aureus (MRSA). Community-acquired MRSA pneumonia is emerging as a particularly virulent infection due to the production of bacterial toxins, which is not well treated by vancomycin.33 Infections are also important because of their enormous impact on public health, current (tuberculosis, malaria, and AIDS) or potential (severe acute respiratory syndrome, avian influenza, and H1N1 influenza), and are summarized in Table 73-2.

Health-care-acquired, or "nosocomial" infections (Table 73-3), are prevalent in the ICU and increase the mortality rate of critically ill patients.34-39 In our case report, the development of a Pseudomonas aeruginosa pneumonia acquired in the ICU halted the course of recovery from ARDS and prolonged the patient's time on the ventilator and in the ICU. Nosocomial infections are often caused by multiresistant bacteria such as P. aeruginosa, MRSA, vancomycin-resistant Enterococcus, and various enteric gram-negative bacteria, many of which have become difficult to eradicate. Independent risk factors for developing nosocomial infections such as VAP include the severity of the patient's underlying illness, its endogenous microflora, and the performance of invasive procedures. In turn, the patient's own microflora is affected by general practices of antibiotic therapy and by the transmission of nosocomial bacteria from patient to patient through breaches in proper infection-control measures.38,39 The most frequent ports of entry of nosocomial bacteria are indwelling devices such as central venous lines and urinary catheters. It has been clearly demonstrated that the incidence and related morbidity of nosocomial infections can be significantly curtailed by adopting simple procedures of sterile insertion and maintenance, and by removing devices as soon as they are no longer necessary.42

The term shock describes the pathophysiologic manifestations of a catastrophic event that causes a profound, protracted reduction of perfusion to vital organs. Originally used to describe refractory hypotension, a more modern definition of shock indicates a state of tissue hypoperfusion caused by low blood flow, toxic agents, or both, where the circulation fails to meet the metabolic requirements of the cell. Initially, neurohumoral compensatory mechanisms maintain perfusion to vital organs. However, if appropriate supportive treatment is not promptly instituted, these compensatory mechanisms are overwhelmed, and progressive tissue ischemia leads to organ failure. Clinically, signs and symptoms of end-organ hypoperfusion include mental status changes, loss of skin turgor, cool extremities, oliguria, acidemia, a high serum lactate concentration, and a reduced mixed or central venous oxygen saturation.

The body responds to an acute injury—such as the aspiration of gastric contents into the lungs as in our case report, or a bacterial infection, or a surgical wound—in a reproducible way. The initial local inflammatory response aims to control the immediate effects of the injury and generates signals from various cells at the site of injury that activate a systemic response. Hence an injury of sufficient severity triggers a complex response involving multiple cell types, including leukocytes, endothelial cells, and even cells within the nervous system.43 When these homeostatic responses are coupled with appropriate interventions, they enable us to maintain or restore the function of vital organs and to expedite repair, which leads to healing and rehabilitation. When the injury hampers the host's regulatory processes or when our intervention is inadequate or delayed, the attempt to preserve the body's integrity is overwhelmed by the injury and organ dysfunction, and death ensues.

Timing

The complexity of the body's response to injury over time constitutes a major challenge to our ability to develop effective treatments that will inhibit the progression of traumatic, inflammatory, and infectious injuries to MODS and death. The classical view of an initial hypodynamic and a subsequent hyperdynamic phase of the cardiovascular and metabolic responses to injury was overly simplistic.44 The timing and intensity of this complex response appears to vary widely among patients and is influenced by multiple factors, including the magnitude of the injury, the patient's physiologic status, and therapeutic interventions. Figure 73-2 illustrates three hypothetical time sequences of the evolution of critical illness, showing different patterns of activation and suppression of the immune system following an infectious injury. Although studied thoroughly in the case of sepsis, this sequence of events is likely to be applicable to other noxious events, such as after major surgery, trauma, inflammation, and shock.45,46

Physical examination following a serious injury may reveal changes of skin color and turgor, body temperature, arterial blood pressure, breathing pattern, and mental function. Other easily measured parameters, including urine output, base deficit, and serum lactate level, may also be abnormal. Although we separately describe the characteristic responses of various organ systems, in the body these systems are highly interconnected and have certain similar characteristic responses, such as upregulation of inflammatory mediators. Dysfunction of one system often impairs the functions of other organs; a classic example is the development of functional renal failure because of decompensated hepatic failure (hepatorenal syndrome) as can arise in a cirrhotic patient with an intra-abdominal infection.

The Immune-Inflammatory Response

Over the last decade and a half, numerous advances have been made in the understanding of how the host recognizes and responds to threats, be they exogenous (infectious) or endogenous (tissue injury). It is clear that the immune-inflammatory response plays a central role in the evolution of critical illness. Both the innate immune system and the adaptive immune system are activated by injury and infection. The innate immune system is designed to recognize and initiate inflammatory responses to exogenous (ie, parts of microorganisms) or endogenous (ie, released by damaged tissues) factors. The innate immune system is also involved in the development of adaptive responses during infection and other inflammation-driven processes.47 The innate immune system uses a family of receptors, known as pattern recognition receptors (PRRs), to recognize conserved molecular motifs on microorganisms and on endogenous factors that seem to be released when there is damage to tissues. Thus far a number of PRRs have been identified, including toll-like receptors (TLRs), NODs, scavenger receptors and mannose binding lectins. The TLRs have been most extensively studied since their initial discovery in the late 1990s, and TLRs and innate immune pathways are being extensively examined as potential targets for sepsis. In particular, antagonism of TLR4, the host receptor for endotoxin (lipopolysaccharide), has undergone multiple clinical trials.

Activation of innate immune pathways leads to the upregulation of the inflammatory mediators that contribute to the pathophysiology of sepsis, including cytokines, chemokines, arachidonic acid metabolites, complement factors, acute-phase reactants, reactive oxygen species, coagulation pathway intermediaries, and so on. Although leukocytes, in particular monocytes and macrophages, have been most extensively studied as contributors to the so-called cytokine storm and other effects of sepsis, multiple other cell types express innate immune receptors/pathways and participate in inflammatory responses during sepsis and other inflammation-driven disorders. These include endothelial and epithelial cells, and fibroblasts. The downstream effects of innate immune activation are complex, involving multiple cell types, systemic cascades (inflammation, complement, coagulation), and multiple intracellular pathways. The complexity of the system and the fact that the components of the system are important in clearing infection likely explain the failure of anticytokine strategies in larger clinical sepsis trials. More recently, some degree of success has been achieved by supplementing individual substances or pathways that seem deficient during certain phases of critical illness, and treating these "relative insufficiencies" with natural or synthetic equivalents of the undersupplied substance. Examples of such strategies are described in the following sections.

Immediately following acute tissue injury, a variety of inflammatory mediators is released by activated leukocytes, endothelial cells and fibroblasts. These mediators include cytokines, arachidonic acid metabolites, complement fractions, various acute-phase proteins, and oxygen and nitrogen free radicals.46 Cytokines are polypeptide messengers of this huge cascade of reactions that have been extensively studied and identified as critical mediators of physiologic and laboratory manifestations of inflammation. Unlike hormones, which are produced by specialized endocrine tissues, cytokines are not stored, and their acute increase following tissue injury reflects upregulated gene transcription, translation, and synthesis by activated immune and/or endothelial cells. A characteristic of this type of response is that it tends to modulate itself by upregulating the expression of both pro- and anti-inflammatory cytokines and acute-phase reactants.

Over the past 3 decades there has been a prevailing view that the pathogenesis of traumatic and infectious injury is based on an excessive and uncontrolled response to the initial insult ("cytokine storm") that turns a beneficial defense system into the actual offender, and, ultimately, causes critical illness. This appealing view, brought forward by Lewis Thomas in the 1970s48 and further championed by Roger Bone in the 1990s,49 led to the development and testing of many therapies that targeted or modulated inflammatory mediators, including high-dose corticosteroids, ibuprofen, antibodies against bacterial endotoxin (lipopolysaccharide [LPS]), TNF-α, and IL-1 receptor antagonists.50-52 However, human trials with these agents were largely unsuccessful. Given the complexity of the inflammatory response, it is not surprising that inhibiting the activity of a single mediator may not provide sufficient benefit.

The Central Nervous System

The central nervous system response to major injury and stress involves activation of the autonomic system, as well as interactions with the endocrine and the immune systems. The classical neurohumoral response to acute injury includes early activation of the autonomic sympathetic system resulting in tachycardia, increased myocardial contractility, tachypnea, splanchnic and cutaneous vasoconstriction, and increased availability of metabolic substrates. This "fight-or-flight" response is closely linked to activation of the hormonal pathways of the adrenal medulla, the pituitary-adrenal axis, antidiuretic hormone (ADH), and angiotensin-aldosterone (see "The Hormonal Response"); all these physiologic pathways, in various ways, tend to preserve perfusion to vital organs.53

The brain also interacts with the immune system, and recent experimental observations have identified a vagal-mediated neural pathway that participates in the regulation of inflammation. Figure 73-3 shows a schematic of this "cholinergic anti-inflammatory pathway."54 Inflammatory stimuli including LPS, TNF-α, and IL-1 activate visceral vagal afferents that reach multiple vagal nuclei in the medulla and hypothalamus. This neural input elicits an immediate anti-inflammatory response mediated by the same autonomic parasympathetic system. Experimental activation of the cholinergic/anti-inflammatory pathway inhibits function and reduces the expression of TNF during endotoxemia through the binding of nicotinic receptors on macrophages.55 Knowledge of this cholinergic/anti-inflammatory pathway has possible therapeutic implications: Experimental activation of vagal output by electrical, as well as pharmacologic, vagal stimulation has inhibited macrophage activation and cytokine release.54

The Hormonal Response

The hormonal response to acute injury tends to increase vascular pressures, preserve circulating blood volume, and provide substrates for tissue metabolism and regeneration. These hormonal pathways include the synthesis and release of adrenal corticosteroids and catecholamines, ADH and aldosterone-angiotensin, prolactin, growth hormone, and other anti-insulin hormones. The role of these hormones in the acute response is complex, and significant understanding of their role has been gained over the past several decades.56 Recent evidence reveals that some of these pathways may be key determinants of survival for critically ill patients and suggests they may be successfully modulated pharmacologically.57

The output of the pituitary-adrenal axis increases several-fold as a normal response to acute illness.58 Any malfunction of this axis, whether at the level of the precursors, the end product (cortisol) or the cellular receptors, leads to an ineffective response and may be associated with a reduced chance of survival. The term relative adrenal insufficiency refers to a state in which the pituitary-adrenal axis fails to mount a sufficient response in the presence of an acute and severe stress, despite functioning adequately in health.59 Based on this concept, patients who fail to mount an appropriate increase in the activity of the pituitary-adrenal axis in response to a severe stress like sepsis may benefit from exogenous replacement of corticosteroid hormones. However, clinical data concerning the effectiveness of physiologic doses of corticosteroids for the treatment of severe sepsis and septic shock remain controversial.60,61 Although a physiologic benefit (increased arterial blood pressure, earlier resolution of shock) has been shown consistently, the most recent and largest controlled trial to date61 failed to find any effect on survival. It is possible that corticosteroids may be beneficial in selected patients that we are still unable to identify and/or treat properly. In particular, our standard clinical evaluation of the function of the pituitary-adrenal axis in critically ill patients through the administration of adrenocorticotropic hormone (ACTH) and subsequent dosage of circulating cortisol is probably insensitive.62

Similarly, ADH is insufficiently synthesized in a significant fraction of critically ill patients, particularly those suffering from vasodilatory shock.63 The administration of a low dose of vasopressin, a synthetic analog of ADH, appears to improve general systemic hemodynamics and reduce the need for infusing other vasopressor medications such as norepinephrine. However, a recent trial in vasodilatory shock patients did not demonstrate a benefit of adding vasopressin to norepinephrine when compared with increasing the dose of epinephrine alone.64 As in the case of corticosteroids, it is clear that this therapy does not benefit all patients, and we still do not have adequate means to evaluate the complex physiologic response to stress and injury at the bedside. Further post hoc analyses of this large study may reveal possible benefits of vasopressin administration in subgroups of critically ill patients with vasodilatory shock, such as those with concomitant adrenal insufficiency64 or acute renal insufficiency.65

The hormonal response to acute stress includes a surge in the synthesis and release of anti-insulin hormones such as epinephrine, glucocorticoids, glucagon, and growth hormone. This proglycemic response increases the availability of glucose as a ready source of energy for vital organs during a time of starvation and massive catabolism. However, the resulting hyperglycemia of critical illness, which is independent of any coexisting diabetic state, may affect the course of critical illness in a number of ways, including an increased susceptibility to infections, immunosuppression, and neurologic damage.66 A provocative clinical trial from a single surgical ICU suggested that a morbidity and mortality benefit occurs if the blood sugar level is maintained between the very narrow range of 80 and 110 mg/dL. Such an effect seemed to be mediated through a decrease in the incidence of bacteremia, critical illness polyneuropathy, and the need for blood transfusions.67 Over the ensuing years, the strength of this finding was significantly mitigated by the repeated observation that the intensive insulin therapy required to achieve this goal often resulted in severe hypoglycemia, which in turn was independently associated with an increased mortality.68,69

The Hemodynamic Response

The hemodynamic response to injury has been extensively investigated by monitoring central vascular pressures, cardiac output, and indices of oxygen delivery and consumption.70 Systemic vasoconstriction commonly predominates in the immediate period following an acute injury, most likely as a reflex response to intravascular hypovolemia from hemorrhage or sequestration of body fluids outside the vascular compartment. After normovolemia is restored, a hyperdynamic pattern of high cardiac output and normal-to-low blood pressure ensues. The extent of this hyperdynamic response varies with the type and magnitude of the injury. It is important to note that hyperdynamic states occur in other conditions, such as hyperthyroidism, liver cirrhosis, and with large arteriovenous fistulae, even if acute injury and/or inflammation are not present.

This hyperdynamic response may be part of a systemic inflammatory response syndrome (SIRS) that is associated with acute injury of various etiologies and caused by the release of cytokines and other inflammatory mediators such as prostaglandins, bradykinins, and complement fractions.71 The cardinal manifestations of SIRS include hyper- or hypothermia, tachycardia, leukocytosis or leukopenia, and hyperventilation. In addition, a number of other signs and symptoms, such as hypotension, platelet and coagulation abnormalities, and changes in mental status, are characteristic of SIRS, although each taken individually lacks specificity both as a diagnostic sign and as an end point for defining therapy.

The Endothelium: Systemic Vasodilatation, Inflammation, and Coagulation

The endothelium, through its effects on inflammation, coagulation pathways, blood flow, and vascular barrier function, contributes to the pathophysiology of organ dysfunction in sepsis and other inflammation-driven disorders.72-75 The endothelium produces proinflammatory cytokines such as IL-6 and IL-8, and participates in leukocyte recruitment to organs. It is also involved in modulating the coagulation pathways. Endothelial activation is a critical contributor to vascular leak during sepsis.76 Endothelial activation, coagulopathy, and vascular leak occur at locations removed from sites of infection, causing inflammation, edema, and ultimately to "bystander organ injury." Finally, the endothelium participates in the cardiovascular response to systemic inflammation, including decreased systemic vascular resistance.

Immediately following injury there is upregulation of adhesion molecule expression on endothelial cells and leukocytes leading to increased endothelial cell–leukocyte interactions and eventual migration of leukocytes across the vessel wall into the adjacent tissues (Fig. 73-4).77,78

Direct injuries and inflammatory mediators cause the endothelium to lose its anticoagulant properties. Plasminogen activator inhibitor (PAI)-1 is produced by endothelium and is the principal inhibitor of tissue plasmin activator and urokinase. When PAI-1 is released in increased amounts into the bloodstream, fibrinolysis is severely suppressed. This mechanism is postulated to contribute to sepsis-induced disseminated intravascular coagulation (DIC).79-81 The imbalance between thrombin and plasmin and the increased platelet stickiness result ultimately in clot formation on the endothelial surface. Thus widespread activation of the coagulation system in response to stress may lead to DIC, intravascular formation of fibrin, decreased fibrinolytic activity, and widespread thrombosis.82 Although the full syndrome of diffuse vascular thrombosis is rare, DIC often presents a considerable clinical challenge because of the concomitant presence of a consumptive coagulopathy and thrombosis. The coagulation system has been targeted therapeutically in patients with sepsis and septic shock. Antithrombin III and protein C are both anticoagulants maintained in an active state by endothelial activity, and they are decreased in conditions of stress such as SIRS and sepsis. Although a clinical trial of the therapeutic administration of antithrombin III was unsuccessful,83 the infusion of recombinant human activated protein C (drotrecogin-α) modestly increased the survival rate of critically ill patients with severe sepsis at high risk of death.84

The endothelium produces multiple mediators that have diverse functions. Some mediators have multiple effects, such as protein C, vasodilator mediators like NO and prostacyclin (both also inhibit platelet aggregation), and the vasoconstrictor endothelins, and all have potent vasoactive and immunologic properties. NO is produced constitutively in endothelial cells by the isozyme nitric oxide synthase (NOS-3, endothelial nitric oxide synthase [eNOS]). Following stimulation of various kinds, there is upregulation of the inducible form of NOS (NOS-2, iNOS), which results in increased production of NO and contributes to the profound vasodilatation that occurs with severe systemic inflammatory processes.85 Activation of endothelial cells also results in the production of mediators with opposing effects, such as the endothelins, which can cause vasoconstriction, whereas NO causes vasodilatation.86 When the induction of iNOS is overwhelming, such as in septic shock, the massive vasodilatation of the extreme hyperdynamic syndrome results in profound hypotension that may compromise vital tissue perfusion and lead to MODS and death (see Chapter 74). NO-induced vasodilatation has been reversed by the administration of nonspecific NOS inhibitors such as methylene blue,87 NG-nitro-L-arginine methyl ester (L-NAME),88 and N-methyl-L-arginine hydrochloride.89 However, this clinical strategy has been unsuccessful, possibly as a result of the importance of the constitutive isoenzymes of NO that are inhibited by these nonselective compounds.

The Metabolic Response

The metabolic response to traumatic injury is characterized by hypercatabolism and increased urinary nitrogen excretion. Once initial hemodynamic resuscitation and stabilization are achieved, substrates are used to enhance tissue repair and preserve vital organ function. Tissues that do not have a high resting energy requirement, such as fat and skeletal muscle, release nutrients from storage (glucose and fatty acids) or proceed to break down their structural proteins to provide fuel for visceral organs. High levels (3- to 4-fold increase) of circulating catecholamines coexist with this metabolic response. Epinephrine, in particular, promotes glycogenolysis in muscle and liver, lipolysis and ketogenesis in adipose tissue, and hepatic gluconeogenesis.90,91 These processes provide energy for the high metabolic requirements during recovery and repair. In addition, epinephrine induces insulin resistance in skeletal muscle, which contributes to the development of hyperglycemia. Like many homeostatic responses, the high catecholaminergic state has dual and opposing effects; although catecholamines promote healing and repair, high catecholamine levels may cause excessive catabolism, hyperglycemia, muscle wasting, and reduced host defenses. One clinical study of children with acute burn injuries has shown the benefits of blocking the catecholaminergic response to acute burn injury. Because children with burns have a particularly high metabolic demand,92 it remains uncertain whether this benefit can be extrapolated to critically ill adults without burns.

At the tissue level, the ability to live or die can be reduced to the ability of the body to supply adequate amounts of oxygen and nutrients (eg, glucose) for cellular respiration and to remove metabolic by-products (carbon dioxide [CO2] and toxins). These two seemingly simple end points are reached through enormously complex phenomena. This section reviews several aspects of this process that are relevant to acute injury, whether traumatic, inflammatory, or infectious.

The Cellular Response to Traumatic Injury

The extent of the damage caused by a traumatic injury (eg, amount of energy received on impact, size of the wound) often determines the immediate outcome of a patient. An uncontrollable hemorrhage will result in early death, with the possible salvage strategy being very simple: a tourniquet and immediate transport, not ICU care. A complex but not immediately lethal injury will then evolve over time, and the outcome will depend on the adequacy of the endogenous response to injury and the ability to support it successfully.

The body's immediate response to traumatic injury requires the generation of signals that communicate the presence of a threat to the rest of the body. These danger signals, which may include necrotic tissue by-products and local changes in pH and temperature, activate endothelial cells, monocytes, and macrophages to produce and release inflammatory mediators. Production and release of cytokines begins within a few minutes of injury and facilitates a systemic response that has protean clinical manifestations. Early measurements of plasma cytokine levels in trauma and surgical patients indicate that the degree of increase in IL-6, one of the earliest responders, correlates with the degree of injury.92,93 In addition to the immediate synthesis and release of cytokines, soluble receptors and antagonist mediators are released following trauma, further modulating the effects of the cytokines and initiating a competitive anti-inflammatory, immunosuppressive response.94,95 This balance between the activation and suppression of inflammation continues throughout the course of critical illness, and it likely affects the ultimate evolution toward recovery or death.96 Better understanding of the determinants of this balance between the pro- and anti-inflammatory responses could guide the generation of more effective therapeutic strategies. A state of immunosuppression or anergy is well described after severe trauma and surgery and may increase susceptibility to nosocomial infections, a common determinant of further morbidity and a longer ICU stay.97

Innate Immune Receptors in Infection

Like traumatic injury, in which the extent of damage determines the bodily response, a sufficient inoculum of pathogenic microorganisms triggers the host's systemic inflammatory response, manifesting in its most severe form as septic shock. The host response to infection is initiated by interactions between structural components of the invading organism (pathogen-associated molecular patterns [PAMPs]) and the immune cells of the host. All classes of microbes, including gram-positive and gram-negative bacteria, fungi, viruses, and parasites, have PAMPs, although their exact PAMPs vary. For instance, gram-negative bacteria have endotoxin (LPS) in the outer membrane of their cell walls. LPS is a potent bacterial toxin that is thought to play an important role in gram-negative sepsis. Gram-positive bacteria do not have LPS, but they have other PAMPs that have inflammatory effects, such as lipoproteins, peptidoglycan, and lipoteichoic acid.

LPS is the best characterized and most extensively studied PAMP. A variety of host mediators, including soluble mediators in the blood, cell surface receptors, and intracellular signaling pathways, are involved in PAMP-induced activation of inflammation. For instance, both LPS-binding protein and the monocyte surface antigen CD1498 bind LPS, and they are involved in initiating inflammatory responses to LPS. CD14 is vital for LPS-induced activation of immune cells and has both membrane-bound and soluble forms. Soluble CD14 is required for LPS-induced activation of CD14-negative cells.99 CD14 also seems to play a role in cell signaling by gram-positive bacterial PAMPs, such as peptidoglycan.95

Over the last decade, a great deal has been discovered about the mechanisms of inflammation induced by microorganisms. Despite considerable structural heterogeneity, inflammatory responses to different microbial components are mediated through common receptors and intracellular pathways. TLRs recognize various classes of PAMPs and are critical proximal mediators of inflammation during infection. TLRs are evolutionarily conserved innate immune receptors that are present in both insects and mammals. They sense common motifs in components of microorganisms (PAMPs). Mammalian TLRs were discovered in 1998 when TLR4 was identified as the LPS receptor.100 The same year, human TLRs 1-5 were cloned.101 Since that time, multiple mammalian TLRs have been identified, each recognizing a different PAMP.

As with all biologic systems, intercellular signaling in early inflammation and infection has an intrinsic redundancy. In addition to the TLRs, other innate immune receptors and signaling pathways seem to be important in the recognition of and response to infection, including NOD receptors, scavenger receptors, mannose binding lectins, and dectin. Refer to recent reviews for a more inclusive discussion.102-105

From Inflammatory Response to Immunosuppression

Within the first 30-90 minutes after exposure to endotoxin, mononuclear cells release proinflammatory cytokines such as TNF-α, IL-6, and IL-1β. This activates inflammatory cascades that stimulate further production and release of cytokines, chemokines (chemoattractant cytokines), lipid mediators, and free radicals, which are involved in leukocyte migration and adhesion and in tissue injury.77,78 A similar inflammatory response can be triggered by all of the stressors or insults described in earlier sections of this chapter. The failure of early therapies aimed at blocking isolated aspects of the inflammatory response, such as infusing high-dose corticosteroids, TNF antagonists, IL-1 receptor antagonists, and non–isozyme-specific inhibitors of NO synthesis, has led investigators to search for a deeper understanding of this extraordinarily complex and redundant system.

There is mounting evidence to support the concept of immunomodulation, that is, an equilibrium between activation and suppression of the immune system, during critical illness. It is postulated that a state of immunoparalysis can occur, whereby the host has inadequate immune responses to microorganisms and plays an important role in subsequent morbidity and mortality following injury or infection. Anti-inflammatory mediators and altered cellular responses are believed to predispose patients to the development of nosocomial infections and subsequent MODS.106,107 At various times after the initial proinflammatory state of traumatic and infectious injury, counterregulatory cytokines, soluble decoy receptors for cytokines, and antagonists of proinflammatory cytokines are secreted. Immunosuppression is believed to occur when this response supersedes its function of restoring immunologic balance. This is exemplified by studies that indicate that blood from septic patients has reduced endotoxin-induced cytokine production, unlike blood from control patients.108 Anergy, or the inability of T cells to proliferate and secrete lymphokines upon stimulation, is often observed in severe trauma and septic patients.109-111 In trauma, a decrease in the number of T cells probably results from interactions between T cells and immunosuppressive mediators such as IL-10.112,113 Regulator T cells (T-reg), which are important elements of the normal immunologic equilibrium, may play a role in the immunosuppression resulting from burn injury.112,114

In acute infection, adaptive immunity may be impaired because of apoptosis of T lymphocytes115 rather than cell necrosis. The term necrosis refers to accidental cell death, whereas the term apoptosis refers to programmed cell death. In contrast to necrosis, which generally has a stimulating effect on the innate immune system, apoptosis can induce secretion of anti-inflammatory cytokines leading to immunosuppression and anergy. It has been shown that the B lymphocytes, CD4 T lymphocytes, and dendritic cells undergo apoptosis but that CD8 T lymphocytes or natural killer cells do not undergo apoptosis.115 After the initial decrease in the number of CD4 T cells, the subset of these cells that recovers are selectively T-reg and not T-helper (Th) cells.116 T-reg cells have multiple negative regulatory effects on immune function and can contribute to immunosuppression.112 The importance of immune cell apoptosis in sepsis has also been demonstrated in animal models. Strains of mice that produce the antiapoptotic protein Bcl-2 are protected from death after intra-abdominal sepsis induced by cecal ligation and puncture.117 With late immunosuppression as the predominating state after the initial injury, critically ill patients become a target for nosocomial infections, which can lead to vital organ failure and death.

Metabolic Disturbances

Critically ill patients may develop metabolic disturbances that hinder the ability to generate sufficient energy to maintain body homeostasis. ATP is the primary fuel source for cellular respiration. The mitochondria are the principal sites of aerobic ATP production. Here, pyruvate, a product of glucose metabolism, is used in the citric acid cycle to generate the reduced forms of nicotinamide adenine dinucleotide (NADH) or flavin adenine dinucleotide (FADH) and provide the flow of electrons necessary to synthesize ATP by oxidative phosphorylation. Ultimately, 36 or 38 mol of ATP are generated from the oxidation of 1 mol of glucose. Conversely, anaerobic glycolysis occurs in the cytoplasm, and generates only 2 mol of ATP per mole of glucose, with the production of lactate or ethanol as end products. During critical illness, the failure to generate adequate amounts of ATP can lead to cellular and organ dysfunction.

Several studies describe a decrease in ATP production in septic patients118-120 and lower ATP levels in the tissues of critically ill septic patients who died when compared with tissues of those who survived.121 One reason for the failure to generate ATP during critical illness is inadequate delivery of oxygen to vital tissues, secondary to impaired oxygenation and perfusion. During SIRS and septic shock, multiple factors may decrease tissue perfusion, including inadequate cardiocirculatory performance, hypoxemia from acute lung injury, mediator- or vasopressor-induced vasoconstriction, and/or coagulation derangements causing microvascular obstruction.

Cytopathic Hypoxia

Cytopathic hypoxia is postulated to contribute to the failure to generate adequate ATP because of the inability of cells to consume available oxygen.122 Experimental and clinical studies show that tissue oxygen tension during sepsis is often higher than in nonseptic states.123-125 Mitochondrial dysfunction is hypothesized to underlie the apparent disconnection between cellular oxygen tension and the low ATP levels. Mitochondrial dysfunction is reported in cells exposed to macrophages, inflammatory cytokines, LPS, and infection.126-128 Multiple factors may contribute to mitochondrial dysfunction in acute injury, including the production of NO and reactive oxygen species. Low doses of LPS can increase iNOS messenger RNA activity in the intestinal tissues of experimental animals,127 and physiologic concentrations of NO can reversibly inhibit the activity of cytochrome oxidase by competition with oxygen.129,130 In addition, NO can damage mitochondria by generating peroxynitrite when reacting with the molecular oxygen of the superoxide radical anion. Peroxynitrite is a potent oxidizing and nitrating that can inhibit multiple mitochondrial functions.131 The presence of peroxynitrite is associated with mitochondrial inhibition, DNA breakage, and activation of poly ADP-ribose polymerase (PARP), an enzyme responsible for repair of single-strand DNA breaks.132 PARP activation can lead to energetic failure by causing a decrease in the oxidized form of nicotinamide adenine dinucleotide (NAD+) and NADH, an important electron donor in mitochondrial oxidative phosphorylation (see earlier).133

The transition from organ failure to recovery or to further damage and death is still incompletely understood. An autopsy study of patients who died from sepsis showed discordance between the clinical manifestations and histologic findings in various organs.117 In the heart, there was no histologic evidence of injury to myocytes despite reduced contractility. In the kidneys, there was only focal injury with preservation of normal glomerular and tubular architecture despite the acute impairment of renal function. Furthermore, it is a common observation that most patients who survive septic shock recover the baseline function of their heart, lungs, and kidneys, suggesting that reversible factors, rather than cell death, are responsible for organ dysfunction.134 What is known about the mechanisms of the organ failure of critical illness is discussed in Chapter 74. Here we provide a schema of pathophysiologic events occurring in the progression of critical illness. At successive steps, different therapeutic approaches may be useful for influencing the progression of illness toward survival and away from death (Fig. 73-5).

Initially, the inciting "stress" should be countered with early and appropriate etiologic therapy (eg, antimicrobials for sepsis) and early resuscitation directed to the physiologic goals that we use as surrogates of adequate tissue perfusion (eg, arterial blood pressure, central venous pressure, urine output, and mixed venous oxygen saturation or PaO2, lactate).135 Intervening the cascade of tissue hypoperfusion, systemic inflammation and dysregulation of the immune response may prevent subsequent organ damage. However, even this principle is not universal, as exemplified by the successful use of limited fluid resuscitation in selected trauma patients.2 In our case report, the presence of pulmonary hypertension and right ventricular insufficiency further complicated the management of resuscitation via goal-directed therapy, prompting the use of invasive monitoring with a pulmonary artery catheter, which may not be indicated routinely for the management of ALI/ARDS.136

Subsequently, therapies supporting the natural response to acute stress, such as an infusion of activated protein C (as was done in our case report), may improve survival of the most critically ill patients. In the future, therapies targeting mitochondrial protection, such as the administration of PARP inhibitors, superoxide and peroxynitrite scavengers, or selective NOS-2 inhibitors, can attenuate the cytopathic effects of oxygen and nitrogen free radicals, and decrease mitochondrial dysfunction. Immunosuppression, a persistent catabolic state, and poor nutritional status can all decrease host defenses and cause increased susceptibility to infectious complications. Inappropriate use of antibiotics and lack of adequate attention to basic procedures to prevent transmission of antibiotic-resistant bacteria may further increase the chances of nosocomial infections at this stage. In our case report, Pseudomonas pneumonia hindered the progression of our patient's course toward recovery, although ultimately our patient survived. Future therapies may target various aspects of immunosuppression, including apoptotic mechanisms (eg, by manipulation of Bcl-2), and may provide an appropriate balance of nutrients that can stimulate immune function. In addition, stimuli to increase mitochondrial energy generation can potentially play a role, so that cellular energy production regains baseline levels.

As the basic mechanisms of critical illness are further defined through basic and clinical research, it is likely that the therapeutic algorithms for treating critical illness will evolve, from a gross physiologic approach supporting respiration and circulation to a more targeted mechanistic approach.

Acute injuries of various etiologies, such as the aspiration of gastric contents into the lungs (our case report), cause immediate local and systemic physiologic responses that are mediated by well-characterized inflammatory and humoral phenomena. Despite the complexity of these phenomena, our understanding of the evolution of critical illness has greatly increased in recent years as investigators have begun to elucidate the cellular and molecular mechanisms of the host's response to injury. The interaction between the initial "stress" (usually trauma, surgery, or infection) and the response of the host determines the evolution of a critical illness toward either healing and recovery or complications and death. Although a robust inflammatory response is necessary to fend off the most severe injuries, a subsequent decrease in its intensity and the onset of a hypoimmune state favor the onset of late infections that may lead to dysfunction of organs and death.

Critically ill patients undergo complex therapies and invasive interventions. The effectiveness of our therapies depends on a number of factors, including our ability to intervene properly and at the optimal time. Recent therapeutic interventions have improved the probability of survival of many critically ill patients. As knowledge of the basic mechanisms that underlie critical illness evolves, additional effective therapeutic strategies should emerge that will improve the care and outcome of critically ill patients.

• Fink MP. Bench-to-bedside review: cytopathic hypoxia. Crit Care. 2002;6:491-499.
• Gelman S. Venous function and central venous pressure. Anesthesiology. 2008;108:735-738.
• Hotchkiss RS, Karl IE. The pathophysiology and treatment of sepsis. N Engl J Med. 2003;348:138-150.
• Lipshutz AKM, Gropper MA. Perioperative glycemic control. Anesthesiology. 2009;110:408-421.
• Ochoa JB, Makarenkova V. T lymphocytes. Crit Care Med. 2005;33:S510-513.
• Peleg AY, Hooper DC. Hospital- acquired infections due to gram-negative bacteria. N Engl J Med. 2010;362:1804-1813.
• Rittirsch D, Flierl MA, Ward PA. Harmful molecular mechanisms in sepsis. Nat Rev Immunol. 2008; 8:776-787.
• Russell J, Phang P. The oxygen delivery/consumption controversy. Approaches to management of the critically ill patient. Am J Respir Crit Care Med. 1994;149:533-537.
• Tracey K. The inflammatory reflex. Nature. 2002;420:853-859.