Trauma, 7th Ed.

CHAPTER 57. Respiratory Insufficiency

Jeffrey L. Johnson and James B. Haenel

The maintenance of gas exchange may be tenuous in the injured patient because of dysfunction in three key elements of the respiratory system. First, the central nervous system may be impaired, resulting in inadequate respiratory drive, or inability to maintain patent proximal airways. Second, injury to the torso can produce changes in compliance, ineffective respiratory effort, and pain that impact the patient’s ability to complete the work of breathing. Third, primary and secondary insults to the lung result in ineffective gas exchange.

In practice, it is common for patients to suffer simultaneous insults, affecting all three elements. Impaired airway patency (e.g., diminished level of consciousness), increased work of breathing (e.g., multiple rib fractures), and impaired gas exchange (e.g., pulmonary contusion, fat emboli syndrome) often coexist in the same patient. Respiratory failure that relates primarily to CNS injury is discussed at length in other chapters and will not be extensively covered here. This chapter will focus on insults that affect work of breathing and gas exchange. The syndrome of postinjury acute respiratory distress syndrome (ARDS) is a major focus.


The neurohormonal response to injury (Chapter 61) results in a remarkable increase in cellular metabolism. This creates a substantial increase in carbon dioxide (CO2) production that must be matched by increased elimination from the lungs. While a resting adult eliminates 200 cm3/(kg min) of CO2, postinjury hypermetabolism results in CO2 production in the range of 425 cm3/(kg min).1 Thus, the minute ventilation required to maintain eucapnia may rise from a resting rate of approximately 5 L/min to more than 10 L/min. This represents a 100% increase in the work of breathing simply to meet metabolic demands.

Additionally, injured patients typically have an increase in physiologic and anatomic dead space—ventilated regions that do not participate in gas exchange. In a normal adult, the proportion of each breath that is dead space (Vd/Vt) is approximately 0.35. In the intubated, ventilated patient, the Vd/Vt can be calculated by a number of techniques including the Bohr–Enghoff method (Vd/Vt = [{PaCO2 – mean expired CO2}/PaCO2]).2 For practical purposes (since mean expired CO2 is not commonly measured), this is a reflection of the minute volume required to achieve a given PaCO2. In ventilated patients with pulmonary failure, the Vd/Vt often exceeds 0.6. Simply put, this extra dead space is a burden because each breath is less effective at eliminating CO2, and therefore minute ventilation requirements in the 12–20 L/min range are not uncommon in the postinjury setting.

The above increase in respiratory demand might be met by a healthy adult; however, the injured patient faces several challenges in completing this additional work. CNS dysfunction from injury impairs respiratory drive, as do many medications routinely used for sedation and analgesia. Decreased thoracic compliance from abdominal distension (e.g., as part of the abdominal compartment syndrome, Chapter 41), chest wall edema, and recumbent positioning increases the energy required to complete a respiratory cycle. Decreased pulmonary compliance from an increase in extravascular lung water and pleural collections (effusions/hemothorax) also contribute. Muscular weakness from impaired energetics (acidosis, cardiovascular failure, mitochondrial dysfunction, oxidant stress) or fatigue may be an insurmountable challenge. Finally, pain from torso injuries or operative interventions make the increased ventilatory demand a substantial burden to the patient.

The net effect of increased demand and diminished capacity to execute the work of breathing is hypercapnic respiratory failure. In the early postinjury period, patients most commonly present with a mixed acid–base picture where ventilation is inadequate to maintain physiologic pH in the face of a metabolic acidosis. This frequently occurs in the absence of major hypoxia, and therefore caution is warranted in relying on oximetry alone to assess adequacy of pulmonary function. Particularly in patients with tachypnea, blood gas analysis is optimal to quickly identify patients who are not meeting their ventilatory demands. While efforts to diminish the work of breathing should be routine, most patients with hypercapnic failure require some form of mechanical ventilation to meet their demands. Approaches to invasive and noninvasive ventilation are covered extensively in Chapter 55.

One approach to addressing the challenge of increased CO2 production in the injured patient is to blunt the hypermetabolic response. Simple maneuvers such as maintenance of euthermia, delivery of nutrition, and minimizing pain-induced stress are warranted but may not affect the incidence or outcome of hypercapnic respiratory failure. A more novel approach is delivery of beta-adrenergic blockade, which, in a retrospective analysis, improves outcomes in injured patients. There are reasonably compelling data in the burn population and the acutely brain injured that blunting of hypermetabolism is beneficial.3 A prospective safety and efficacy evaluation of this approach in injured adults is needed before it can be recommended as routine practice.

Two injury patterns that precipitate hypercapnic respiratory failure are worthy of special mention: spinal cord injury and flail chest/pulmonary contusion. In spinal cord injury, conventional wisdom asserts that lesions below C5 should not result in pulmonary failure, because innervation to the diaphragm remains intact. In practice, however, most complete cord lesions in the cervical and upper thoracic regions routinely result in failure requiring mechanical ventilation.4,5 The genesis is multifactorial, including delays in patient mobilization, ineffective cough due to loss of innervation of intercostal and abdominal musculature, pneumonia in the setting of multiple injuries, and aspiration at the time of the initial insult. Furthermore, unopposed vagal stimulation results in early bronchorrhea and bronchospasm. This patient population requires aggressive mobilization and pulmonary care as recurrent lobar collapse and pneumonia are the rule.

Early operative stabilization of the spine can be recommended as it has been shown to decrease the need for mechanical ventilation and ICU stay.6 Other adjuncts such as noninvasive ventilation, bronchodilators, mucolytics, and percussion should be considered, although most of these have not been studied in a fashion that permits a firm recommendation. Laparoscopic placement of diaphragmatic pacers is a benefit in select patients.7 Specifically, improvement in spirometry and some liberation from mechanical ventilation occurs in most patients evaluated thus far. Largely, this has been studied in the rehabilitation setting, and therefore a role for this approach in the acute setting remains to be defined.

Flail chest and pulmonary contusion can be thought of as a single entity. This is a challenging injury pattern, because it impacts both the patient’s ability to execute the work of breathing (from pain and mechanical instability of the thoracic) and gas exchange (from the pulmonary contusion). Isolated pulmonary contusion rarely requires mechanical ventilation. Since minor contusions are frequently identified by computed tomography (CT), it is important to realize that this tends to take a relatively benign course (see Chapter 26). It is clear that the number of rib fractures is strongly associated with pulmonary failure, ARDS, and mortality, and that this effect is more dramatic in the elderly population.8 This may in part be a marker of energy applied to the chest and injury to the underlying lung. Early pain control, preferably beginning in the emergency department with regional anesthesia, has been shown to be effective in reducing the impact of multiple rib fractures and should be routinely applied. Admission to a high-volume trauma center, use of patient-controlled analgesia, tracheostomy, and an algorithm-driven approach are associated with improved survival.9,10


Hypoxic pulmonary failure is a substantial co-contributor in the trauma setting. The etiologies are diverse including aspiration pneumonitis, pneumonia, pulmonary embolism, acute lung injury (ALI), and ARDS. Most of these entities are discussed elsewhere—ALI and ARDS will be the focus of this chapter. Indeed, the mechanisms at work in ALI and ARDS share many common features with these other processes that affect the alveolar–capillary interface.


ALI and ARDS are clinical syndromes of inflammatory lung injury that can be thought of as a common final pathway of diverse systemic processes. In the past two decades, there has been major progress in defining the underlying pathophysiology and optimal supportive care. Specific therapy for the underlying mechanisms, however, remains an elusive goal.


The recognition of ARDS as a distinct clinical entity resulted from the description by Ashbaugh et al. in 1967.11 Subsequent descriptions include five principal criteria: (1) hypoxemia refractory to oxygen administration; (2) diffuse, bilateral infiltrates on chest radiograph; (3) low static lung compliance; (4) absence of congestive heart failure; and (5) presence of an appropriate at-risk diagnosis.

The current standard definitions were developed in 1994, after a Consensus Conference of American and European investigators (AECC) agreed that ARDS should be viewed as the most severe end of a spectrum of ALI. It also recommended diagnostic criteria for ALI and ARDS (Table 57-1). The diagnostic criteria for ARDS include acute onset, the PaO2/FiO2 200 mm Hg or less (<300 for ALI), bilateral infiltrates on chest radiograph, and no evidence of left arterial hypertension (either clinical or with direct measurement). Moreover, the committee recognized that ARDS is most often associated with sepsis syndrome, aspiration, primary pneumonia, trauma, cardiopulmonary bypass, multiple blood transfusions, fat embolism, and pancreatitis. Although debate exists as to the usefulness of the AECC diagnostic criteria, they have promoted study of reasonably homogeneous populations, are largely accepted by clinical investigators, and are likely to be used for some time.12

TABLE 57-1 American–European Consensus Conference Definitions of Acute Lung Injury and Acute Respiratory Distress Syndrome


One concern about the standard definitions is substantial limitations to bedside application. For example, transient hypoxemia from mucus plugging is common in an ICU setting, and it is unclear that a patient who only transiently meets P/F criteria should be grouped with patients who have ongoing poor oxygenation. Additionally, recruitment of collapsed alveoli may result in a remarkable improvement in P/F ratio in a short period of time—does this patient no longer have ARDS? Lastly, while the AECC definition excludes patients with left atrial pressure (LAP) >18, Ferguson et al.13 showed that patients with no risk factors for congestive heart failure commonly had LAP >18 during a clinical course consistent with ARDS.

Whether the current definition is too broad also remains a matter of debate. For example, a 2004 study compared post-mortem analysis of lung tissue with the AECC definition. This investigation found the latter to be only 84% specific and 75% sensitive for the pathologic lung lesions characteristic of ARDS.14 It is clear, then, that the AECC definition, while useful for studies of populations, should be applied with caution to individual patients. It is also worthwhile to know its limitations and consider some alternative methods for objectively assessing lung injury.

An alternative that remains useful is the Murray lung injury score (LIS). This was proposed in 1988 and is based on four components: chest radiograph, hypoxemia, positive end-expiratory pressure (PEEP), and respiratory compliance (Table 57-2).15 Each component is scored from 0 to 4. The LIS is calculated by summing the scores of the available components and dividing by the number of components used. ARDS (or severe lung injury) is defined as an LIS greater than 2.5. Zero represents no lung injury and 0.1–2.5 represents mild to moderate lung injury.

TABLE 57-2 Lung Injury Scorea



A more recent definition makes a simple adjustment to the AECC definition and appears to approve diagnostic accuracy when compared with pathologic lesions found at autopsy. It may be more suitable for bedside evaluation of the individual patient (Table 57-3).16 Briefly, the authors include PEEP in the consideration of hypoxia and require either the absence of CHF or the presence of a recognized risk factor for ARDS. The degree of hypoxia required is a P/F less than 200 with PEEP ≥10; therefore, it excludes patients who are hypoxic purely because of derecruitment or suboptimal PEEP. By allowing patients with high filling pressures in the presence of a recognized ARDS risk factor, the definition recognizes the prevalence of high LAPs during the course of ARDS and includes patients with concomitant CHF and ARDS.

TABLE 57-3 The “Delphi” Definition of ARDS



The estimated incidence of ARDS in population-based studies is on the order of 40 cases per 100,000 person-years.1719 In the United States, this represents about 200,000 cases per year, and about 15% of ICU admissions. With modern mortality rates (see below), one can estimate this entity is responsible for about 50,000 deaths per year in the United States.


Several studies demonstrate that age is a risk factor for ARDS, although it is unclear whether this simply represents diminished physiologic reserve in older patients. Hudson et al. documented an increasing incidence of ARDS with increasing age. Subgroup analysis, however, showed that this was largely due to the group with postinjury ARDS who were significantly older (44 years vs. 36 years).20 These authors also observed that a higher Injury Severity Score (ISS) was a risk factor for ARDS.

Genetic variability between patients may also contribute to risk for ARDS (for a discussion of genomics relevant to trauma, see Chapter 53). Polymorphisms in genes encoding cytokines, vasomotor regulators, antioxidants, and surfactant proteins have all been associated with altered risks for either the development of ALI/ARDS or the outcome. Specific candidate genes have included angiotensin-converting enzyme, mannose-binding lectin, extracellular superoxide dismutase, surfactant protein B, and interleukin (IL)-10. Many of these studies are difficult to interpret because the reported effect is modified by gender, disease process (e.g., septic vs. traumatic ARDS), or the patient population studied (e.g., Caucasian vs. Asian).

Of potential particular interest in trauma is a polymorphism in the promoter region of the NADPH:qunione oxidoreductase 1 gene (NQO1). This is an inducible gene that when activated regulates generation of oxidant species. In 2009, Reddy et al. completed an analysis demonstrating that critically injured patients (ISS >16, ICU admission) with the variant gene were about half as likely to develop ALI compared with patients homozygous for the wild-type gene. This effect was independent of race, mechanism of injury, and severity of illness score (Apache III).21

Clinical risk factors for ARDS can be broadly categorized into direct and indirect groups (Table 57-4). Direct factors are those primarily associated with local pulmonary parenchymal injury and include pulmonary contusion, aspiration, and pulmonary infection. Indirect factors are those thought to be associated with systemic inflammation and resultant lung injury. These include severe sepsis, transfusion of banked red cells, transfusion of FFP, and multiple long bone fractures. Unless shock is associated with significant tissue injury or other known risk factors, it has not been shown to result in ARDS.22

TABLE 57-4 Clinical Risk Factors for ARDS



The current paradigm of systemic inflammation leading to ALI and ARDS posits that a variety of insults, both infectious and noninfectious, can result in an unbridled hyperinflammatory response. This leads to organ injury from indiscriminate activation of effector cells that subsequently release oxidants, proteinases, and other potentially autotoxic compounds. If the initial insult is severe enough, early organ dysfunction results (“one-hit” or single insult model). More often, a less severe insult results in a systemic inflammatory response that is not by itself injurious. These patients appear, however, to be primed such that they have an exaggerated response to a second insult, which leads to an augmented/amplified systemic inflammatory response and multiple organ dysfunction (“two-hit” or sequential insult model, see Chapter 61).23

Inflammatory models provide a unifying hypothesis for ARDS and MOF; however, the precise relationship between ARDS and MOF remains to be defined. MOF is a frequent occurrence and the most common cause of mortality in patients with ARDS.24 Indeed, postinjury ARDS appears to be an obligate precursor of other organ failures.25 This may be because the lung is a primary target of the inflammatory process, or because the resultant pulmonary damage impairs the lung’s ability to metabolize inflammatory mediators and control cellular effectors of injury.26 It is also now clear that ventilator strategies that inadvertently promote lung injury may produce systemic inflammation, perhaps leading to other organ failures.27

Inflammatory lung injury leads to the pathologic lesion of diffuse alveolar damage. This prototypic lesion of ARDS is at the alveolocapillary interface, which results in epithelial and endothelial damage as well as high-permeability pulmonary edema. The histologic appearance of this lesion can be divided into three overlapping phases: (1) the exudative phase, with edema and hemorrhage; (2) the proliferative phase, with organization and repair; and (3) the fibrotic phase.28

The exudative phase generally encompasses the first 3–5 days but may last up to a week. The initial histologic changes include interstitial edema, proteinaceous alveolar edema, and intra-alveolar hemorrhage. The exudative phase is characterized by the appearance of hyaline membranes, which are composed of plasma proteins mixed with cellular debris. Electron microscopy reveals endothelial injury with cell swelling, widening intercellular junctions, and increased pinocytotic vesicles. In addition, there is disruption of the basement membrane.

The alveolar epithelium usually exhibits extensive loss of type I cells, which slough and leave a denuded basement membrane. While some loss may be from necrosis, it appears that apoptosis contributes substantially. Activation of matrix metalloproteinases, Toll-like receptors, and oxidative stress pathways initiate programmed cell death in these cells. Demonstration of soluble Fas ligand in bronchoalveolar lavage (BAL) fluid early in ARDS supports this concept.29

Loss of the alveolar epithelial barrier results in alveolar edema, as the remaining cells are unable to drive sodium from the alveolar into the interstitial compartment.

During the proliferative phase, type II cells divide and cover the denuded basement membrane along the alveolar wall. This process may be seen as early as 3 days after the onset of clinical ARDS. Type II cells are also capable of differentiating into type I epithelial cells. Fibroblasts and myofibroblasts proliferate and migrate into the alveolar space in the third phase. Fibroblasts change the alveolar exudate into granulation tissue, which subsequently organizes and forms dense fibrous tissue. Eventually, epithelial cells cover the granulation tissue. This whole process is called fibrosis by accretion and is important in lung remodeling. Septal collagen deposition by fibroblasts and “collapse induration” also contribute to fibrous remodeling of the lung in ARDS.

The fibrotic stage is characterized by thickened, collagenous connective tissue in the alveolar septa and walls. Pulmonary vascular changes occur as well, with intimal thickening and medial hypertrophy of the pulmonary arterioles. Complete obliteration of portions of the pulmonary vascular bed is the end result.


Lung injury in ARDS involves components of inflammation, coagulation, vasomotor tone, and other systems (see Chapter 57). The pivotal cellular mediators appear to be leukocytes, with both local and humoral mediators orchestrating their function. Activation of these leukocytes results in release or activation of multiple cytokines, chemokines, oxidants, and proteases that result in the final common pathway of tissue injury in ARDS.

Image Neutrophils

A consistent histopathologic feature of ARDS is neutrophil infiltration of the pulmonary microvasculature, interstitium, and alveoli. Neutrophils are well equipped to cause damage through the release of reactive oxygen species and proteases.30 Furthermore, neutrophils may be an important source of proinflammatory cytokines. Persistence of neutrophils in serial BAL fluid samples from patients with ARDS suggests unbridled inflammation and portends poor prognosis. In animal models, neutrophil depletion prior to an insult markedly attenuates resulting lung injury.31

The lung normally contains a significant number of sequestered neutrophils, and their mere presence is not sufficient to cause tissue injury. A long-standing model suggests that after a “priming” stimulus, neutrophils firmly adhere to endothelium and accumulate in the lungs; however, lung injury does not occur unless a second activating stimulus is applied. Thus, for neutrophils to cause tissue damage, there must be adherence to the endothelium, transmigration to the interstitium, and subsequent activation. Adherence and transmigration create a toxic microenvironment that is protected from endogenous antioxidants and antiproteases normally present in the plasma.

Both cellular biomechanical and adhesive mechanisms are operative in the process of neutrophil sequestration. The initial phase is thought to result from a change in the cytoskeleton of the neutrophil that increases rigidity. This change impedes flow through the pulmonary microvasculature.32 A second, more prolonged phase is related to increased adhesive forces between neutrophils and endothelial cells. Initially, neutrophils “roll” and then “tether” to the endothelium as a result of the interaction of selections (L, E, and P) on the neutrophil and endothelial surfaces. This is followed by firm adhesion or “capture” of the PMN on the endothelial surface, which is mediated by β2 integrins on the neutrophil. These adhere to intracellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule (VCAM), and MADCAM. Once adherent, PMNS can transmigrate into the subendothelial space via paracellular or transcellular routes, again via surface integrins, adhering to platelet–endothelial cell adhesion molecule-1 (PECAM-1) or junctional adhesion molecules.33


The lung contains large numbers of fixed tissue macrophages that are a critical component of the inflammatory response in ALI. Activated macrophages can cause tissue injury by releasing the same toxic mediators as neutrophils (reactive oxygen species and proteases). Probably more important is the macrophage capability to synthesize multiple proinflammatory mediators, such as complement fragments, cytokines, and chemokines. Thus, macrophages are thought to have a major role in amplifying and perpetuating the inflammatory response. This is exacerbated by the long half-life of the macrophage, which is measured in days rather than hours as in the neutrophil. The alveolar macrophage has two additional key functions: control of local infection and modulation of fibrosis.34 Alveolar macrophage from ARDS patients demonstrates defective phagocytosis and bacterial killing, reflecting an increased risk for infection in these patients.


The pulmonary endothelium is not a passive bystander in the pathogenesis of ARDS, but actively participates in initiating and perpetuating the inflammatory response. Endothelial cells increase the expression of adhesion molecules (ICAM-1, ICAM-3, and E-selectin) following exposure to an activating stimulus. These ligands serve as tethering and signaling molecules by binding to their cognate leukocyte membrane proteins. Thus, the endothelial cell actively coordinates trafficking, firm adhesion, and transmigration. In the setting of systemic inflammation, inappropriate endothelial cell activation may lead to indiscriminate leukocyte recruitment and parenchymal inflammation. Moreover, endothelial cells produce and release vasoactive substances, such as prostacyclin, nitric oxide (NO), and endothelins. These substances may mediate much of the pulmonary vascular dysfunction characteristic of ARDS. Activated endothelium also expresses procoagulant activity, which contributes to intravascular coagulation and microvascular dysfunction.35 Thrombin, in turn, has proinflammatory effects on leukocytes. Endothelial injury, then, may be both a proximate cause and a marker for ALI.


It has long been understood that reperfusion of the ischemic gut can lead to lung injury.36 Because gut ischemia/reperfusion is an established phenomenon in the injured patient with hemorrhagic shock, this remains a tantalizing hypothesis for the development of inflammatory injury to the lung. Curiously, however, convincing evidence of inflammatory mediators leaving the gut into the portal circulation has been lacking in humans. This led Deitch37 to hypothesize that the egress of proinflammatory substances from the gut may be via lymph, not venous blood. This intriguing hypothesis has substantial support in animal models, including the finding that diversion of gut lymph abrogates lung injury. While the precise mediators of this phenomenon are as yet unknown, changes in posthemorrhagic shock lymph flow, lipid content, and protein content are an active area of investigation.38,39

Image Mediators and Markers of Acute Respiratory Distress Syndrome


Systemic complement activation secondary to trauma or sepsis is considered a major early factor in ARDS.40 C5a, a product of complement activation, is a powerful neutrophil chemoattractant. Moreover, C5a induces neutrophil aggregation and activation leading to pulmonary neutrophil sequestration and lung injury. Clinically, plasma and bronchoalveolar C3a levels correlate with the development of ARDS.41

Lipid Mediators

Phospholipids are potent inflammatory mediators that are formed by the action of phospholipase A2 (PLA2) on membrane phospholipids. PLA2 contributes to the inflammatory response by two separate pathways, catalyzing the production of both platelet-activating factor (PAF) and arachidonic acid. Arachidonic acid metabolism results in release of eicosanoids such as leukotrienes, thromboxane, and prostaglandins. Each of these is a pivotal mediator in the inflammatory cascade and has been implicated in the pathogenesis of ARDS.42

PAF is a phospholipid with potent vasoactive and inflammatory properties. It is produced by a number of cell types including macrophages, neutrophils, endothelial cells, and type II pneumocytes. PAF production is stimulated by endotoxin, tumor necrosis factor (TNF), and leukotrienes, and exerts diverse biologic actions, including neutrophil activation and adherence, platelet aggregation and degranulation, and macrophage production of inflammatory mediators. Infusion of PAF in animals results in increased vascular permeability and neutrophil-mediated ALI.43

Lysophosphatidyl cholines are another class of bioactive lipids that may play a significant role in lung injury after trauma. These compounds accumulate during routine storage of packed red blood cells and have been shown to cause lung injury in isolated perfused rodent lungs.44 These or related compounds from transfusion of banked red cells may provide a second insult leading to inflammatory organ injury in the injured patient.

As mentioned above, the pulmonary endothelium is recognized as an active participant in the development of ALI. As such, markers of endothelial activation or injury have been investigated as predictors of the development of ARDS. von Willebrand factor antigen (vWF:Ag) has been studied fairly extensively as a marker of endothelial dysfunction. vWF:Ag is synthesized largely by vascular endothelial cells and has been shown to be a sensitive marker of endothelial injury or activation.45 This antigen was studied prospectively in 45 patients to determine whether elevated levels of vWF:Ag are predictive for the development of ALI.46 Only patients with nonpulmonary sepsis were included, and one third developed ALI. Elevated plasma levels of vWF:AG (>45% above controls) were 87% sensitive and 77% specific for development of ALI. Positive predictive value was 65%.

Following activation, endothelial expression of adhesion molecules, including ICAM-1, VCAM-1, E-selectin, and P-selectin, is upregulated. These compounds are susceptible to proteolytic cleavage and may exist in the circulation in a soluble form. Therefore, these molecules represent a measure of endothelial activation or damage. We and others have demonstrated elevated ICAM-1 levels in severely injured patients who subsequently developed MOF.47,48 In contrast, plasma levels of soluble E- and P-selectins measured at admission were not useful in predicting ALI.

The maintenance of a functioning alveolar epithelium is important for recovery from ALI.49 Unlike the endothelium, there is a lack of specific histologic markers of alveolar epithelial injury. Surfactant abnormalities have been noted in the earliest reports of ARDS. Surfactant lipids and proteins are synthesized and released by alveolar epithelial type II cells. The surfactant-associated proteins SP-A and SP-B are decreased in BAL fluid from patients with ARDS and at risk for ARDS.50

Markers of leukocyte activation have also been measured in plasma and BAL fluid of patients in an attempt to predict development of ARDS. Gordon et al. noted markedly elevated plasma elastase levels very early after multisystem trauma.51 Subsequent studies supported a causative role for neutrophil elastase in ARDS,52 and have led to the clinical development of human elastase inhibitor.53 Increased expression of β2 integrins has been observed on the surface of circulating pulmonary artery neutrophils in patients at risk for postinjury ARDS.54

Circulating and BAL fluid levels of cytokines are also inherently attractive as predictors of ARDS; whether they are markers or mediators remains an important question. TNF-α, IL-1β, IL-6, and IL-8 have been the most intensely investigated in relation to the development of ARDS. Studies examining the predictive value of cytokines central to sepsis (TNF-α, IL-1β) levels in ARDS have had negative or mixed results.55 IL-8 is a neutrophil chemoattractant that has been studied in both plasma and BAL fluid. It is elevated in the plasma following injury but does not appear to consistently predict development of ARDS. Donnelly et al. studied 29 patients at risk for developing ARDS and observed that IL-8 levels in BAL fluid were significantly higher in patients who later progressed to ARDS.56


ARDS is characterized by diffuse, patchy, panlobular pulmonary infiltrates on plain chest radiograph (Fig. 57-1). CT of the chest will demonstrate that the parenchymal changes are inhomogeneous with the dependent lung regions most affected (Fig. 57-2). The inhomogeneous distribution of parenchymal densities led to the concept of a three-compartment model of the lung in ARDS.57 One compartment is substantially normal (healthy zone), one is fully diseased without any possibility of recruitment (diseased zone), and, finally, the third compartment is composed of collapsed alveoli potentially recruitable with increasing pressure (recruitable zone). Increased airway pressure is necessary to recruit collapsed alveoli. In the 1990s, however, it was already recognized that in the heterogeneously injured lung, airway pressure or stretch may be damaging to the healthy zone. This ventilator-induced lung injury was thought to be responsible for severe protracted ARDS, as well as perpetuation of systemic inflammation and MOF.58,59


FIGURE 57-1 (A) Normal admission chest x-ray in a 27-year-old trauma patient with multiple lower extremity fractures. (B) Chest x-ray from the same patient following onset of respiratory insufficiency. Note the bilateral, dense pulmonary infiltrates consistent with severe ARDS.


FIGURE 57-2 Chest CT scan of a patient with ARDS shows marked inhomogeneity of the process with areas of dense consolidation and essentially normal intervening pulmonary parenchyma.

A standard method used to describe the mechanical properties of the lung is to determine the static pressure–volume curve during inflation and deflation (Fig. 57-3). In early ARDS, the lower inflection point represents the airway pressure at which considerable alveolar recruitment occurs. The upper inflection point is where near maximal inflation occurs such that further increases in airway pressure result in alveolar overdistension and little change in volume. The “open lung approach” to mechanical ventilation in ARDS advocates setting PEEP at or just above the lower inflection point to avoid repetitive alveolar recruitment and collapse with each breath. In practice, however, inflection points in individual patients have been difficult to consistently measure.


FIGURE 57-3 Idealized static pressure–volume curve of the lungs in ARDS. At low pressures and volumes, derecruitment, atelectasis, and lung injury from repetitive opening and closing of gas exchange units are a concern. At high pressures and volumes, overdistention, increased shunt, and lung injury from excessive stretch predominate. In between these two extremes is the optimal zone where lung stays recruited but is not overstretched.

Image Pulmonary Edema

Increased pulmonary capillary permeability is a consistent feature. Increased permeability promotes alveolar flooding with protein-rich edema fluid, as well as release of proteins generally confined to the lung into the systemic circulation. Multiple studies have documented elevated total protein concentrations in BAL fluid from patients with established ARDS.60 One study observed higher BAL fluid protein concentrations in nonsurvivors, suggesting that more severe lung injury is associated with greater endothelial and epithelial permeability.61


Hypoxemia in ARDS results from ventilation–perfusion mismatching and intrapulmonary shunting of blood flow. Shunting results from blood that passes through the systemic venous to pulmonary arterial system without going through the normal gas exchange units in the lung (i.e., right–left shunt). Normally, the shunt fraction is less than 5%; however, in ARDS, it may exceed 25%. Because blood flowing through a shunt is not exposed to alveoli participating in gas exchange, supplemental oxygen is ineffective in increasing arterial oxygen concentration. Other techniques designed to restore ventilation to diseased lung regions, such as PEEP or continuous positive airway pressure (CPAP), are thus necessary to improve oxygenation. Hypoxic pulmonary vasoconstriction is a protective mechanism that limits perfusion to poorly ventilated alveoli and minimizes shunting. In ARDS, hypoxic pulmonary constriction is impaired, resulting in greater intrapulmonary shunt. Multiple factors may contribute to loss of hypoxic pulmonary vasoconstriction in ALI, including local prostaglandin or NO production.62

In early ARDS, most patients are tachypneic from hypoxemia and are secondarily hypocapneic. With disease progression, hypercapnia may become a prominent feature because of physiologic dead space and CO2 production. Increased dead space ventilation is invariably present. As discussed earlier, the normal dead space to tidal volume ratio is approximately 35%, but in ARDS, it can approach 60%.


Lung compliance is defined as the change in lung volume per change in transpulmonary pressure. Normal lung compliance in a mechanically ventilated patient ranges from 60 to 80 mL/cm H2O. With ARDS, it is not unusual to see greatly diminished lung compliance on the order of 10–30 mL/cm H2O. Initially, reduced compliance is the result of interstitial and alveolar edema with alveolar flooding. Surfactant dysfunction and terminal bronchiolar spasm contribute to loss of ventilated alveoli. In later ARDS, interstitial fibrosis and parenchymal loss further reduce pulmonary compliance.


Pulmonary hypertension is a frequent finding in patients with ARDS and can lead to increased interstitial pulmonary edema, right ventricular dysfunction, and impaired cardiac output. When severe, pulmonary hypertension has been observed to be a marker of poor outcome. The etiology of pulmonary hypertension in ARDS is multifactorial. Early in ARDS, the predominant mechanism is most likely impaired vasorelaxation related to hypoxemia, acidosis, and vasoactive mediators, in concert with obstruction from microvascular coagulation. In late ARDS, fibrosis and obliteration of the pulmonary vascular bed are most likely responsible. Steltzer et al. noted markedly depressed right ventricular function in nonsurvivors of ARDS.63 The authors related this to reduced myocardial contractility and not to pulmonary hypertension. They also noted that oxygen delivery was more related to cardiac performance than to pulmonary gas exchange.


The diagnosis of ARDS is based on the criteria used to define the syndrome and has historically been divided into four clinical phases with radiographic, clinical, physiologic, and pathologic correlates (Table 57-5). In the initial phase, dyspnea and tachypnea are evident, with a remarkably normal chest examination (radiographically and clinically). Arterial oxygen saturation is preserved, and hypocapnia from hyperventilation is frequently noted.

TABLE 57-5 Pathophysiologic Changes of Modern Adult Respiratory Distress Syndrome (Low-Pressure Pulmonary Edema)



The second phase quickly follows (12–24 hours), with physiologic and pathologic evidence of lung injury. The chest x-ray now shows bilateral patchy alveolar infiltrates with hypoxemia evident on arterial blood gas (ABG) determination. If ARDS persists and progresses, the third clinical phase becomes evident. Acute respiratory failure necessitates mechanical ventilation with increasing inspired oxygen concentrations. There is an increase in physiologic dead space and rising minute ventilation. Patients at this point may develop sepsis syndrome with a hyperdynamic hemodynamic pattern. The radiographic picture worsens with more diffuse infiltrates, consolidation, and air bronchograms.

Without resolution, progressive pulmonary failure and fibrosis characterize the fourth phase. Pneumonia, often recurrent, is frequent (see Chapter 18). Hypercapnia may worsen and become more difficult to control. MOF commonly develops and is the most common cause of death (see Chapter 61).


Because there is no proven specific treatment for ARDS, therapy primarily involves supportive measures to maintain life while the lung injury resolves. Such measures include identifying and treating predisposing conditions, mechanical ventilatory support with oxygen, nutritional support, nonpulmonary organ support, and hemodynamic monitoring as necessary. Attention to detail is necessary to avoid nosocomial infection and iatrogenic complications.


Given the central role of alveolar flooding in ARDS, the appropriate use of fluids and diuretics in managing patients has been a matter of debate since the description of this syndrome. In large part, the discussion centers on whether hydrostatic forces, changes in membrane permeability, or dysfunction of the lung epithelial barrier are the prime contributors to pulmonary edema in ARDS.

Some investigators believe that hydrostatic intravascular forces contribute significantly to pulmonary edema in ARDS and favor early diuresis and fluid restriction to minimize interstitial and alveolar edema. If necessary, cardiac output and blood pressure are supported with vasopressors. Potential problems with this approach include decreased end-organ perfusion with precipitation of MOF, and an increased shunt fraction. Some studies observe that increased extravascular lung water does not correlate with oxygenation or outcome in ARDS patients, and calls into question the physiologic basis for this approach.64Because hypovolemia may be uniquely dangerous in the resuscitation of the injured patient, caution should be exercised before adopting such an approach.

Two landmark trials published by the ARDS network in 2006 attempted to address appropriate fluid management in this group of patients. In one study, a total of 1,000 patients were randomized to either liberal or conservative fluid strategies over a period of 7 days.65 The conservative group received approximately a net 1 L less per day, and spent 2.5 fewer days on the ventilator. There was no mortality difference, and no increase in other organ failures in the conservative group. In the second study, care was based on either CVP measurements or pulmonary artery catheter measurements. This study also showed no difference in mortality. Overall, these studies do not show a profound effect of a specific invasive monitor or the amount of fluid administered. It may be that the uniform application of pressure-limited ventilation trumps any major effect of fluid balance in this patient population.

With respect to colloids, there is no evidence that their use for acute resuscitation improves outcome. A recent study in hypoproteinemic patients suggests that gas exchange can be improved in late ALI by using colloid in concert with diuretics to mobilize interstitial fluid and promote diuresis.66 A 2006 randomized study also suggested that critically ill patients with profound hypoproteinemia have fewer organ failures (by SOFA score) when given colloid therapy.67 Short-term improvement in physiology is not, however, accompanied by an improved outcome in these investigations. The very large Australian SAFE study suggests that routine use of albumin cannot be supported.68Thus, outside of the profoundly hypoproteinemic patient, it is difficult to argue that colloids are of any benefit.


Respiratory support with a mechanical ventilator is a cornerstone in the supportive management of patients with or at risk for ARDS. It has become increasingly clear, however, that mechanical ventilation can perpetuate or worsen lung injury, as well as result in more readily recognized forms of barotrauma (mediastinal emphysema, pneumothorax, etc.). This recognition has led to the concept of lung-protective strategies in mechanical ventilation.69

Most patients with ARDS require endotracheal intubation for mechanical ventilation. Patients with mild ALI may occasionally be managed initially with noninvasive positive pressure mask ventilation. This technique reduces shunt physiology and administers high concentrations of oxygen. It requires an alert and cooperative patient to tolerate the tight-fitting mask and avoid complications (in particular, vomiting and aspiration). Patients with multisystem injuries are frequently poor candidates for this approach. Emergent intubation is associated with significantly higher morbidity and mortality. Accordingly, early elective intubation should be considered in all patients with deteriorating gas exchange or mental status.

Historically, large tidal volumes (10–15 mL/kg) were used to achieve normal ABG values while avoiding microatelectasis and patient discomfort. With little modification and the addition of PEEP, this approach was applied as the standard for most critically ill patients. In the setting of ARDS, the conventional goals of mechanical ventilation have been to achieve adequate oxygenation using the least PEEP possible and a nontoxic FiO2 while at the same time maintaining normocarbia. This strategy was intended to minimize or avoid hyperoxic lung injury and the adverse effects of PEEP. The primary priority has shifted from maximizing tissue oxygen delivery to ensuring adequate lung protection (Table 57-6).

TABLE 57-6 Ventilatory Strategies in ARDS



PEEP is one of several methods of increasing mean airway pressure and improving oxygenation. It improves oxygenation by enhancing lung volume, and increasing functional residual capacity (FRC) through recruitment of collapsed alveoli. Lung compliance may also be improved. The use of PEEP in patients with ARDS has two primary goals: adequate tissue oxygen delivery and the reduction of FiO2 to nontoxic (generally below 0.6) levels. Increasing PEEP above a certain level, however, may have significant adverse effects. By raising intrathoracic pressure, PEEP may significantly reduce venous return and cardiac output.70 This may result in decreased tissue oxygen delivery despite improvement in arterial oxygen saturation. This effect is accentuated in the hypovolemic patient, and can usually be reversed with intravascular volume expansion. Cardiac depression is rarely seen with PEEP levels less than or equal to 10 cm H2O. If PEEP >10 cm H2O is required, assessment of intravascular volume status and cardiac function is warranted.

PEEP may also result in alveolar overdistention with compression and obliteration of surrounding pulmonary capillaries. This alveolar overdistention may actually worsen oxygenation by increasing the shunt fraction. Moreover, dead space ventilation may be increased, resulting in a higher minute ventilation requirement. Finally, PEEP may cause a maldistribution of tidal volumes and pressures, creating overdistension of normally aerated lung regions. This hyperinflated form of lung injury may result in barotrauma referred to as “volutrauma.”

The optimal approach to PEEP for ARDS remains a matter of debate. Early on, several strategies were proposed for determining optimal PEEP.71 “Best PEEP” sought the greatest static lung compliance. “Optimal” PEEP sought lowest shunt fraction. In “preferred” PEEP, the goal is maximal oxygen delivery. Today, many centers have adopted the algorithm-driven approach adopted by the ARDSNet investigators that is based on a ratio of PEEP to FiO2. In one trial by these investigators comparing “high” with “low” PEEP in a lung-protective strategy, there was no clear difference in outcomes.72

The advantage of a ratio approach is that it can be done without measurement of pulmonary compliance curves or invasive hemodynamic monitoring. One disadvantage is that it is not necessarily applicable to an individual patient with individual physiology.

In the individual patient, the application of PEEP needs to be balanced between providing a sufficient inspiratory plateau pressure (peak alveolar pressure) to maintain adequate oxygenation and at the same time avoiding derecruitment of the alveoli during exhalation. Fig. 57-3 demonstrates the area on the pressure–volume curve where optimal ventilation occurs. With recent ventilatory strategies emphasizing pressure-limiting techniques, the “least” PEEP method seemed logical. This approach has been associated with less barotrauma and cardiac compromise.73 The use of minimal PEEP, however, does not address the potential damage from repeated alveolar opening and collapse with each breath.74 This forms the basis for the “open lung” ventilation strategy. Static pressure–volume curves are used to select a PEEP level above the lower infection point to prevent alveolar end-expiratory collapse. Although this approach is intriguing, static pressure–volume curves can be difficult to obtain as well as to implement.

Practically speaking, bedside PEEP trials are frequently initiated in response to worsening hypoxemia. We identify the optimal or least PEEP in the following manner. Baseline PEEP is set at 10 cm H2O, with a Vt of 6 mL/kg ideal body weight. If the patient desaturates, then the patient is recruited for 2 minutes using pressure-controlled ventilation with a PIP of 20 cm H2O, a respiratory rate of 10 breaths/min, and an I:E of 1:1 with PEEP raised to 25–40 cm H2O. The patient is monitored for cardiopulmonary instability during recruitment because this maneuver results in substantial hypoventilation for the period of recruitment. The patient is returned to his or her previous settings and if desaturation reoccurs, the recruitment maneuver is repeated for 2 minutes and the patient is returned to his or her previous ventilator settings except the PEEP is increased by 2.5–5 cm H2O. This process is repeated as needed until oxygen saturation remains stable.

With improvement in the patient’s lung injury, PEEP should be withdrawn as the pulmonary compliance, shunt fraction, and dead space ventilation improve. Premature attempts at PEEP reduction, however, will only delay resolution of lung injury and prolong the need for ventilatory support. PEEP weaning should be performed in an orderly fashion. PEEP is reduced slowly in 2.5–5 cm H2O increments, with serial monitoring of arterial oxygen saturation using ABG or oximetry. In general, tidal volumes should be increased to both maintain mean airway pressure and allow respiratory rates associated with greater patient comfort. A significant decrease in PaO2 should prompt a quick return to previous levels of PEEP. Once alveolar collapse and loss of FRC occur, higher levels of PEEP for more prolonged periods may be needed. Generally, PEEP should not be reduced by more than 3–5 cm H2O in a 12-hour period.


In patients with ARDS, the aerated lung volume able to participate in gas exchange is markedly reduced to one third of the original volume. Thus, in nonfibrotic stages of ARDS, the lungs may be thought of as small rather than stiff. There is accumulating evidence that pulmonary gas exchange may be essentially normal in the aerated portion of the injured lung. The use of conventional tidal volumes in this setting would be expected to result in alveolar overdistention with further impairment of gas exchange, frequent barotrauma, and ventilator-induced lung injury.

One approach is to reduce tidal volume in order to limit airway pressures while maintaining adequate alveolar ventilation with an increased respiratory rate. While the safe upper limit of airway pressure is not precisely known, pressures in excess of 35–40 cm H2O are associated with lung injury in animal models. A benchmark clinical trial of low-volume, pressure-limited ventilation in established ARDS forms the basis for modern ventilator management in ALI and ARDS.75 In this study, patients ventilated using smaller (6 cm3/kg) tidal volume had a mortality of 31% versus a 40% mortality in a group ventilated with a larger (12 cm3/kg) tidal volumes. This study has solidified the basis for “low stretch” strategies that are thought to diminish both local and systemic inflammation.


Now that the importance of defending airway pressures has been highlighted, the maintenance of eucapnia is no longer stressed. Controlled hypoventilation (relative to CO2 production) with increased PaCO2is referred to as permissive hypercapnia. Gradual increases in PaCO2 are usually well tolerated, provided renal compensation is adequate (see Chapter 60) and severe acidosis (pH <7.1) does not occur.

Respiratory acidosis has a myriad of physiologic effects that may be relevant to the critical care physician. Because CO2 diffuses freely across cell membranes, an increase in extracellular PCO2 will result in intracellular acidosis. The pathologic effects of this hypercapnia are related to the severity and duration of any resultant intracellular acidosis. Cytosolic pH is normally tightly regulated between 6.9 and 7.2.76Three mechanisms are responsible for this regulation: (1) physiochemical buffering, mainly due to proteins and phosphates; (2) reduced intracellular generation of protons; and (3) changes in transmembrane ion exchange. Physiochemical buffering is immediate, while the other mechanisms require 1–3 hours. These regulatory mechanisms are remarkably powerful and efficient. As a result, normoxic hypercapnia has only limited potential for resulting in intracellular acidosis, and is generally well tolerated.

Hypercapnia has multiple cardiovascular effects. Acute respiratory acidosis results in reversible impairment of myocardial contractility. The principal mechanism for this decrease in cardiac contractility is intracellular acidosis, which interferes with myofilament responsiveness to activator calcium.77 Hypercapnia may lead to a substantial increase in coronary flow and a rise in coronary sinus oxygen tension. The overall hemodynamic response to acute hypercapnia in human experiments is increased cardiac output, heart rate, and stroke volume, with decreased systemic vascular resistance. This reflects the net result of the myocardial depressive effects of CO2, as well as indirect effects with stimulation of the sympathetic nervous system and catecholamine release. In some patients, however, hypercapnia may adversely affect right ventricular function. It results in pulmonary vasoconstriction and may lead to pulmonary hypertension with right ventricular dysfunction.78

Increased PCO2 produces a rightward shift of the oxygen–hemoglobin dissociation curve by two mechanisms. This increases the P50 and facilitates unloading of oxygen at the tissue level, which may be beneficial.

Hypercapnia has multiple effects on the central nervous system. Similar acid–base changes to those described above also occur in the brain. Preliminary studies suggest a decrease in the oxygen demand of the brain.79 However, it has long been recognized that hypercarbia increases cerebral blood flow and intracranial pressure. Increased intracranial pressure results from enlargement of cerebral blood volume secondary to diminished vascular tone.

Contraindications to permissive hypercapnia include increased intracranial pressure or other cerebral disorders in which intracranial hypertension may be detrimental. Uncorrected hypovolemia and significant cardiac disease are relative contraindications due to the negative inotropic effects of permissive hypercarbia.

Several reports document that permissive hypercapnia in the setting of ARDS is remarkably well tolerated.80 Ideally, this strategy should be implemented slowly over several hours to allow compensatory mechanisms to act. The role of sodium bicarbonate infusion is unclear. In most institutions, sodium bicarbonate would not be administered unless the pH was less than 7.2. With increased experience, however, many centers reserve bicarbonate infusion for pH less than 7.0. An alternative is to use tris-hydroxymethyl aminomethane (THAM), which does not increase CO2 production but is not specifically labeled for this use.81 Sedation is mandatory with permissive hypercapnia in mechanically ventilated patients in order to control respiratory drive and prevent discomfort. Even with heavy sedation, however, respiratory drive may be insufficiently suppressed, resulting in patient–ventilator dyssynchrony. Neuromuscular blockade is often necessary in these patients.


There continues to be great interest in mechanical ventilation of ARDS patients in the prone position. Bryan described the beneficial effects of the prone position on arterial oxygenation more than 30 years ago.82 Other investigators subsequently confirmed these findings.83 Although large controlled studies are lacking, it appears that at least 50% of patients show improved oxygenation with prone positioning.

Pappert et al. used the multiple inert gas elimination technique and showed that the improvement in oxygenation was the result of decreased intrapulmonary shunt fraction.84 Albert and coworkers have investigated the mechanism of prone ventilation in a canine oleic acid model of ARDS. The prone position consistently reduced shunt fraction compared with the supine position. The improvement in gas exchange was independent of changes in cardiac output or FRC between the two positions. Subsequently, pulmonary blood flow was shown to be distributed preferentially to dorsal lung regions in the supine and prone positions.85 In supine animals, pleural pressure increases from nondependent to dependent regions. This may lead to dependent atelectasis in the highly perfused dorsal lung regions, resulting in intrapulmonary shunt and hypoxemia. In the prone position, the pleural pressure gradient is less, and the dorsal (now nondependent) regions are exposed to a lower pleural pressure. This results in opening of previously atelectatic alveoli. Intrapulmonary shunting is reduced because perfusion of the dorsal lung regions is maintained. Lamm et al. recently confirmed this mechanism in a canine oleic acid lung injury model.86 The investigators used single photon emission computed tomography (SPECT) scanning to quantitate regional ventilation and perfusion. Supine animals had markedly reduced or absent ventilation to the dorsal lung regions while maintaining perfusion to those areas. With prone positioning, ventilation of dorsal regions improved significantly and perfusion was maintained. Using chest CT, Gattinoni et al. showed that in the supine position, gasless lung was found predominantly in the dorsal regions.87 With prone positioning, densities redistributed to the ventral areas and dorsal regions were well aerated. Thus, prone positioning results in recruitment of previously atelectatic dorsal lung regions.

Although prone positioning appears to improve oxygenation in many patients with ARDS, a significant number of patients have no response. In a small number of patients, gas exchange actually deteriorates. Currently, responders to prone positioning cannot be reliably predicted while in the supine positions.

Despite early reports documenting improved gas exchange in approximately two thirds of patients, it is remarkable that it took 20 years for the first prospective, randomized trial to be performed. The trial, reported by Gattinoni et al.,88 consisted of 152 patients randomized to a prone group and 152 patients to a supine group. Patients randomized to the prone arm were followed for the first 10 days and turned prone for at least 6 hours each day if they met the necessary criteria. Regrettably, no differences in mortality rates were noted between the two groups after a 10-day period (21.1% vs. 25%), and the study was underpowered to detect a statistical difference. A post hoc analysis revealed that in patients with severe hypoxemia, defined as a p/f ratio <90, an APACHE score >49, or having been exposed to high Vt(>12 mL/kg) mortality was significantly lower in the prone group than that in the supine group (20.5% vs. 40.00%). Retrospective analysis of these data in a separate report indicated that the best predictor of improved outcome during prone ventilation was a decrease in the PaCO2 and not the response to arterial oxygenation. Since the Gattinoni trial there have been three other randomized trials of prone ventilation. Unfortunately, these studies do not strongly suggest that prone positioning affects survival.8991

Prone positioning must be performed with care to avoid inadvertent extubation or loss of intravenous lines or chest tubes. Transient hemodynamic instability and desaturation also may occur during repositioning. Cardiopulmonary resuscitation is difficult, if not impossible, in the prone position. Placement of multifunction electrode pads, which allow defibrillation, cardioversion, and pacing, has been recommended to facilitate cardiopulmonary resuscitation in the prone position. Other areas of concern that accompany prone positioning include facial and eyelid edema, peripheral nerve injury, tongue injuries, and skin necrosis. Multiply injured patients may present unique problems due to the presence of incisions, drainage tubes, extremity fractures, cervical spine or facial fractures, and the like. Additional vexing questions remain. In patients who respond, how long should prone positioning be maintained? Some suggest that it be limited to 8–12 hours to allow for patient care; however, it is patients who are maintained prone for more than 20 hours who appear to benefit.92


Recognition of the impact of mechanical ventilation on furthering lung injury has been an impetus to revisit high-frequency modes of ventilation (HFV). After all, this might be considered the ultimate in “low tidal volume ventilation.”

Several modes of high-frequency ventilation have been available over the last 20 years, including high-frequency jet ventilation, high-frequency positive pressure ventilation, ultra-high-frequency ventilation, and high-frequency oscillatory ventilation. These techniques use very small tidal volumes (1–5 mL/kg) delivered at rates of 60–3,600 cycles/min. To date there are a number of prospective clinical studies that have failed to demonstrate any meaningful benefit of high-frequency ventilation over conventional modes of positive pressure ventilation in patients with ARDS.93 Although peak airway pressures are reduced compared with conventional modes, mean airway pressures, barotrauma, and hemodynamic compromise appear unchanged.

Proponents of high-frequency ventilation argue that it provides the ideal approach to lung-protective ventilation. By definition, HFV delivers an extremely low tidal volume in concert with a relatively high mean airway pressure using a piston pump. Mean airway pressures of 25–45 cm H2O are not unusual. Delivery of plateau pressures in this range with tidal volumes that may be just larger than dead space ventilation, perhaps as high as 5 mL/kg, may not be equally protective in all settings of lung injury. For example, will ARDS that is caused by a direct pulmonary insult such as aspiration respond the same as an injured lung from a nonpulmonary cause of ARDS such as pancreatitis?

In an effort to answer these types of questions recently investigators from the Multicenter Oscillatory Ventilation for Acute Respiratory Distress Syndrome Trial (MOAT) published the results of a randomized controlled study comparing HFV with a CV arm in adults with early ARDS.94 One hundred and forty-eight patients were enrolled with 75 randomized to HFV and 73 to CV. The primary study end point was safety and effectiveness of HFV compared with CV. Tidal volume in the CV group was based on actual weight and averaged 10.2 mL/kg, while peak airway pressures were 38 cm H2O. As expected, mean airway pressures were significantly higher at 29 cm H2O versus 23 cm H2O in the HFV group. Oxygenation, as evidenced by the PaO2/FiO2 ratio, was higher in the initial 16 hours in the HFV group but was subsequently decreased and the p/f ratio was no longer significantly different between the two groups. While not statistically significant, there was a clear trend toward a better outcome in the HFV group; mortality was 37% in the FHO group and 52% in the CV group (P = .12). It is regrettable, however, that this trial failed to provide a lung-protective approach for the CV group. So, until a true, randomized, prospective trial comparing HFV with CV using a lung-protective strategy is performed, all we really know is that HFV is safe and effective when compared with CV.

Currently, the primary clinical indications for high-frequency ventilation include treatment of neonatal respiratory distress syndrome and ventilatory support during proximal airway procedures.


Extracorporeal life support (ECLS) is the new term for what was previously called extracorporeal membrane oxygenation (ECMO). ECLS also encompasses the technique of extracorporeal carbon dioxide removal (ECCO2R). It is a modified form of cardiopulmonary bypass that allows for oxygenation and CO2 removal while “resting” the lungs. During ECLS, gas exchange occurs independent of the lungs. Therefore, the lungs are not exposed to potential barotrauma from mechanical ventilation or to oxygen toxicity from exposure to toxic FiO2 levels. Theoretically, the lung is better able to heal during this period of rest and support.

A multicenter, prospective trial of ECLS using venoarterial bypass in the late 1970s did not demonstrate improved survival in patients with severe ARDS.95 In contrast, this modality has achieved success in managing neonates with persistent pulmonary hypertension and is considered the standard of care in this group. The technology has changed considerably and more recent studies in adults with ARDS suggest better survival rates compared with historical controls.96 The lack of both concurrent control groups and randomization makes these studies difficult to interpret. The principal complication of ECLS is hemorrhage due to the need for systemic heparinization and frequent thrombocytopenia. Moreover, ECLS is extremely labor intensive and costly. Technical advancements, such as heparin-bonded circuits, may alleviate many of these problems. In a recent randomized study in the United Kingdom, transfer of ARDS patients to a regional center specializing in ECLS resulted in improved outcomes; however, almost one third of the patients did not receive ECLS after transfer and care delivery in patients not referred was not protocolized; it is therefore difficult to make firm conclusions from this study.97


Image Inhaled Nitric Oxide

NO is responsible for regulation of basal vascular tone. When delivered by the inhaled route, NO is a selective pulmonary vasodilator with no systemic side effects. Moreover, delivery of NO by the inhaled route exposes only ventilated alveoli to its vasodilatory effects. Selective vasodilation of pulmonary vessels in ventilated areas diverts blood away from nonventilated areas, improving ventilation–perfusion matching and hypoxemia. NO decreases mean pulmonary artery pressure, and in doing so, lessens the hydrostatic pressure for pulmonary edema.98 For all of these reasons, inhaled NO has seemed very attractive for use in severely hypoxemic patients.

At least 12 prospective randomized clinical trials of inhaled NO versus placebo or standard therapy have been reported. The results are remarkably similar and not encouraging. The largest trial was a multicenter study and was placebo controlled and blinded.99 Results from this study used fixed doses of NO at 0, 1.25, 5, 20, and 40 ppm delivered to patients with ARDS from causes other than sepsis. Only at doses at 5 ppm were there decreases (not statistically significant) in the oxygenation index and duration of mechanical ventilation. As a result of this trial, the investigators went on to perform a low-dose trial using 5 ppm of NO versus a placebo. The primary end points were days alive and off-assisted ventilation. Mortality was actually a secondary outcome variable, as was meeting extubation criteria. Not surprisingly, there was a statistically significant increase in PaO2, but this diminished after 48 hours. Unfortunately, inhaled NO at 5 ppm had no substantial effect on duration of mechanical ventilation or mortality. Currently, the role of NO should be limited to those patients with severe, refractory hypoxemia or pulmonary hypertension in whom inhaled NO may act as a “bridge” allowing short-term physiologic support. Application of NO in these situations may allow for possible patient survival until other therapies may be employed such as pronation or alternative modes of ventilation. Most studies observe that inhaled NO resulted in only modest improvements in oxygenation that were not sustained beyond 24 hours and frequently did not allow significant reduction in the intensity of ventilatory support. There were no differences in mortality in any of the studies. The multicenter trial also noted no differences in the number of days alive and the number of days off mechanical ventilation. Finally, a metaanalysis of published trials has suggested harm in the cohort of patients given NO.100


The ability of corticosteroids to attenuate the inflammatory response would seem to make them ideal treatment for ARDS. From the first description of ARDS, corticosteroids were suggested as possible therapeutic agents. Large prospective clinical trials, however, showed no survival benefit with high-dose steroids in early ARDS.101 These initial studies focused only on early ARDS and utilized a short course (less than 48 hours) of high-dose steroids. It has become increasingly clear that pathogenetic mechanisms initiating ARDS are different from those that perpetuate late ARDS. The fibroproliferative phase of ARDS is particularly critical. Why lung injury completely resolves in one patient and extensive fibrosis develops in another is unknown. The extent of initial injury, especially the amount of basement membrane disruption, may be an important factor.102 Another important factor is ongoing injury or inflammation. Histologic evidence of continued endothelial injury has been described in patients with advanced fibroproliferation.103 Meduri et al. provide evidence suggesting a link between ongoing fibroproliferation and persistent inflammation.104 The authors measured plasma and BAL fluid cytokine levels following initiation of steroid rescue treatment in patients with late ARDS. They noted a significant reduction in plasma TNF-α and IL-6 levels in patients who responded compared with those who did not.

Several anecdotal reports support the use of steroids in late ARDS.105108 Most of these investigators used methylprednisolone (or its equivalent) in doses ranging from 2 to 8 mg/kg per day for at least 2 weeks. Survival ranged from 76% to 83%, well exceeding that expected in this group of patients. Meduri et al. reported results from a prospective, randomized trial of steroids in late ARDS.109 Twenty-four patients were randomized to receive methylprednisolone 2 mg/kg per day for 32 days or placebo. Patients treated with steroids had reduced LISs, increased PaO2 to FiO2 ratios, and greater extubation success 10 days after treatment. ICU and hospital mortality were significantly reduced (0% vs. 62% and 12% vs. 62%, respectively). This study can be criticized for the small number of control patients (n = 8) and the crossover of patients from the placebo group to the steroid group (four out of eight). A subsequent publication demonstrated that patients treated with methylprednisolone have a rapid reduction in the proinflammatory cytokines IL-1, IL-6, and TNF, supporting the hypothesis that ongoing ARDS represents a pathologically persistent inflammatory state.

Recently, the NIH ARDS Clinical Trials Network reported the results of a multicenter, randomized controlled trial of corticosteroids in patients with persistent ARDS.110 One hundred and eighty patients with ARDS of at least 7-day duration were assigned either methylprednisolone or placebo. The primary end point was 60-day mortality, and secondary end points were ventilator-free days, organ failure, and various biomarkers. There was no difference in 60-day mortality except in those patients who received methylprednisolone after ARDS day 14, and then mortality significantly increased. Methylprednisolone did increase the number of ventilator-free days, improved oxygenation and respiratory compliance, and resulted in less vasopressor use. Interestingly, the rate of infectious complications was no different between the groups, but the methylprednisolone group had a higher incidence of neuromuscular weakness. Given the challenges of this patient group, any use of steroids for late ARDS must be individualized and, optimally, delivered before disease day 14.

Image Nutritional Support

Overfeeding patients or administering excess carbohydrates can lead to excess production of CO2. In the setting of marginal ability to execute work of breathing, this may, in theory, precipitate or prolong hypercapnic pulmonary failure. Careful monitoring of nutritional support may be necessary in patients with tenuous respiratory status. An indirect calorimeter can be helpful in providing estimates of CO2production and the respiratory quotient. The respiratory quotient should be kept below 0.9 by appropriate adjustment of the proportion of lipid and total calories administered. Our goals for nutritional support are to deliver 21–25 nonprotein cal/kg per day and 0.25–0.30 g of nitrogen/kg per day. The primary goal of carbohydrate administration should be a rate of less than 5 mg/(kg min).


Mortality associated with ARDS has historically been reported from 30% to 60%. The majority of deaths are related to sepsis and MOF (see Chapter 61). Respiratory failure is a cause of death in only 15% of patients.111 The principal therapy that improves outcome in patients with ARDS appears to be low tidal volume ventilation; however, a gradual improvement in crude mortality from 1996 to 2005 was observed in the ARDS network, with a low 26% in 2004–2005.112 This suggests that other improvements in critical care delivery are at work as well (Fig. 57-4).


FIGURE 57-4 Denver Health SICU mechanical ventilation protocol.


1. Uehara M, Plank LD, Hill GL. Components of energy expenditure in patients with severe sepsis and major trauma: a basis for clinic care. Crit Care Med. 1999;27:1295.

2. Tang Y, Turnery MJ, Baker AB. A new equal area method to calculate and represent physiologic, anatomical and alveolar dead spaces. Anesthesiology. 2006;104:696.

3. Williams FN, Jeschke MG, Chinkes DL, et al. Modulation of the hypermetabolic response to trauma: temperature, nutrition, and drugs. JACS. 2009;208:489.

4. Como JJ, Sutton ER, McCunn M, et al. Characterizing the need for mechanical ventilation following cervical spinal cord injury with neurologic deficit. J Trauma. 2005;59:912.

5. Harrop JS, Sharan AD, Scheid EH Jr, et al. Tracheostomy placement in patients with complete cervical spinal cord injuries: American Spinal Injury Association grade A. J Neurosurgery. 2004;100(suppl 1):20.

6. Schinkel C, Frangen TM, Kmetic A, et al. Timing of thoracic spine stabilization in trauma patients: impact on clinical course and outcome. J Trauma. 2006;61:156.

7. Onders RP, Elmo M, Khansarinia S, et al. Complete worldwide operative experience in laparoscopic diaphragm pacing: results and differences in spinal cord injured patients and amyotrophic lateral sclerosis patients. Surg Endosc. 2009;23:1433.

8. Bulger EM, Arneson MA, Mock CN, Jurkovich GJ. Rib fractures in the elderly. J Trauma. 2000;48:1040.

9. Harrington DT, Phillips B, Machan J, et al. Factors associated with survival following blunt chest trauma in older patients: results from a large regional trauma cooperative. Arch Surg. 2010;145:432.

10. Todd SR, McNally MM, Holcomb JB, et al. A multidisciplinary clinical pathway decreases rib fracture-associated infectious morbidity and mortality in high-risk trauma patients. Am J Surg. 2006;192:806.

11. Ashbaugh DG, Bigelow DB, Petty TL. Acute respiratory distress in adults. Lancet. 1967;2:319.

12. Bernard GR, Artigas A, Brigham KL. The American–European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994;149:818.

13. Ferguson ND, Meade MO, Hallett DC, Stewart TE. High values of the pulmonary artery wedge pressure in patients with acute lung injury and acute respiratory distress syndrome. Int Care Med. 2002;28:1073.

14. Esteban A, Fernandez-Segoviano P, Frutos-Vivar F, et al. Comparison of clinical criteria for the acute respiratory distress syndrome with autopsy findings. Ann Intern Med. 2004;141:440.

15. Murray JF, Matthay MA, Luce JM, et al. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis. 1988; 138:720.

16. Ferguson ND, Frutos-Vivar F, Esteban A, et al. Acute respiratory distress syndrome: underrecognition by clinicians and diagnostic accuracy of three clinical definitions. Crit Care Med. 2005;33:2228.

17. Frutos-vivar F, Ferguson ND, Esteban A. Epidemiology of acute lung injury and acute respiratory distress syndrome. Semin Respir Crit Care Med. 2006;27:327.

18. Goss CH, Brower RG, Hudson LD, et al. Incidence of acute lung injury in the United States. Crit Care Med. 2003;31:1607.

19. Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med. 2005;353(16):1685–1693.

20. Hudson LD, Milberg JA, Anardi D. Clinical risks for development of the acute respiratory distress syndrome. Am J Respir Crit Care Med. 1995; 151:293.

21. Reddy AJ, Christie JD, Aplenc R, et al. Association of human NAD(P) H:quinone oxidoreductase 1 (NQO1) polymorphism with development of acute lung injury. J Cell Mol Med. 2009;13:1784.

22. Norwood SH, Civetta JM. The adult respiratory syndrome. Surg Gynecol Obstet. 1985;161:497.

23. Moore EE, Moore FA, Harken AH, et al. The two-event construct of postinjury multiple organ failure. Shock. 2005;24(suppl 1):71.

24. Stapleton RD, Wang BM, Hudson LD, et al. Causes and timing of death in patients with ARDS. Chest. 2005;128:525.

25. Ciesla DJ, Moore EE, Johnson JL, et al. The role of the lung in postinjury multiple organ failure. Surgery. 2005;138:749.

26. Wiedemann HP, Matthay MA, Gillis CN. Pulmonary endothelial cell injury and altered lung metabolic function. Early detection of the adult respiratory distress syndrome and possible functional significance. Clin Chest Med. 1990;11:723.

27. Tremblay LN, Slutsky AS. Ventilator-induced lung injury: from bench to the bedside. Intensive Care Med. 2006;32:24.

28. Ware LB. Pathophysiology of acute lung injury and the acute respiratory distress syndrome. Semin Respir Crit Care Med. 2006;27:337.

29. Mancione AM. Role of the pulmonary epithelium and inflammatory signals in acute lung injury. Expert Rev Clin Immunol. 2009;5:63.

30. Sibille Y, Reynolds HY. Macrophages and polymorphonuclear neutrophils in lung defense and injury. Am Rev Respir Dis. 1990;141:471.

31. Heflin AC Jr, Brigham KL. Prevention by granulocyte depletion of increased vascular permeability of sheep lung following endotoxemia. J Clin Invest. 1981;68:1253.

32. Worthen GS, Schwab B III, Elson EL, et al. Mechanics of stimulated neutrophils: cell stiffening induces retention in capillaries. Science. 1989; 245:183.

33. Ley K, Laudanna C, Cybulsky MI, Nourshargh S. Getting to the site of inflammation: the leukocyte adhesion cascade updated. Nat Rev Immunol. 2007;7:678.

34. Demling RH. Adult respiratory distress syndrome: current concepts. New Horiz. 1993;1:388.

35. Welty-Wolf KE, Carraway MS, Ortel TL, Piantadosi CA. Coagulation and inflammation in acute lung injury. Thromb Haemost. 2002; 88:17.

36. Hassoun HT, Kone BC, Mercer DW, et al. Post-injury multiple organ failure: the role of the gut. Shock. 2001;15:1.

37. Magnotti LJ, Upperman JS, Xu DZ, et al. Gut-derived mesenteric lymph but not portal blood increases endothelial cell permeability and promotes lung injury after hemorrhagic shock. Ann Surg. 1998;228:518.

38. Peltz ED, Moore EE, Zurawel AA, et al. Proteome and system ontology of hemorrhagic shock: exploring early constitutive changes in postshock mesenteric lymph. Surgery. 2009;146:347.

39. Jordan JR, Moore EE, Sarin EL, et al. Arachidonic acid in postshock mesenteric lymph induces pulmonary synthesis of leukotriene B4. J Appl Physiol. 2008;104:1161.

40. Fosse E, Pillgram-Larsen J, Svennevig JL, et al. Complement activation in injured patients occurs immediately and is dependent on the severity of the trauma. Injury. 1998;29:509.

41. Gama de Abreu M, Kirschfink M, Quintel M, Albrecht DM. White blood cell counts and plasma C3a have synergistic predictive value in patients at risk for acute respiratory distress syndrome. Crit Care Med. 1998;26:1040.

42. Edelson JD, Vadas P, Villar J, et al. Acute lung injury induced by phospholipase A2. Structural and functional changes. Am Rev Respir Dis. 1991;143:1102.

43. Zimmerman GA, McIntyre TM, Prescott SM, Stafforini DM. The platelet-activating factor signaling system and its regulators in syndromes of inflammation and thrombosis. Crit Care Med. 2002;30(5 suppl):S294.

44. Silliman CC, Voelkel NF, Allard JD, et al. Plasma and lipids from stored packed red blood cells cause acute lung injury in an animal model. J Clin Invest. 1998;101:1458.

45. Pittet JF, Mackersie RC, Martin TR, et al. Biological markers of acute lung injury: prognostic and pathogenetic significance. Am J Respir Crit Care Med. 1997;155:1187.

46. Rubin DB, Wiener-Kronish JP, Murray JF, et al. Elevated von Willebrand factor antigen is an early plasma predictor of acute lung injury in nonpulmonary sepsis syndrome. J Clin Invest. 1990;86:474.

47. Partrick DA, Moore FA, Moore EE. The inflammatory profile of interleukin-6, interleukin-8, and soluble intercellular adhesion molecule-1 in postinjury multiple organ failure. Am J Surg. 1996;172:425.

48. Law MM, Cryer HG, Abraham E. Elevated levels of soluble ICAM-1 correlate with the development of multiple organ failure in severely injured trauma patients. J Trauma. 1994;37:100.

49. Matthay MA, Wiener-Kronish JP. Intact epithelial barrier function is critical for the resolution of alveolar edema in humans. Am Rev Respir Dis. 1990;142:1250.

50. Gregory TJ, Longmore WJ, Moxley MA, et al. Surfactant chemical composition and biophysical activity in acute respiratory distress syndrome. J Clin Invest. 1991;88:1976.

51. Gordon MW, Robertson CE, Dawes J. Neutrophil elastase levels and major trauma in man. Intensive Care Med. 1989;15:543.

52. Moraes TJ, Chow CW, Downey GP. Proteases and lung injury. Crit Care Med. 2003;31(4 suppl):S189.

53. Okayama N, Kakihana Y, Setoguchi D, et al. Clinical effects of a neutrophil elastase inhibitor, sivelestat, in patients with acute respiratory distress syndrome. J Anesth. 2006;20:6.

54. Simms HH, D’Amico R. Increased PMN CD11b/CD18 expression following post-traumatic ARDS. J Surg Res. 1991;50:362.

55. Donnelly TJ, Meade P, Jagels M, et al. Cytokine, complement, and endotoxin profiles associated with the development of the adult respiratory distress syndrome after severe injury. Crit Care Med. 1994;22:768.

56. Donnelly SC, Strieter RM, Kunkel SL. Interleukin-8 and development of adult respiratory distress syndrome in at-risk patient groups. Lancet. 1993;341:643.

57. Gattinoni L, Pesenti A, Avalli L, et al. Pressure–volume curve of total respiratory system in acute respiratory failure: computed tomographic scan study. Am Rev Respir Dis. 1987;136:730.

58. Slutsky AS, Tremblay LN. Multiple system organ failure. Is mechanical ventilation a contributing factor? Am J Respir Crit Care Med. 1998; 157:1721.

59. Finfer S, Rocker G. Alveolar overdistension is an important mechanism of persistent lung damage following severe protracted ARDS. Anaesth Intensive Care. 1996;24:569.

60. Holter JF, Weiland JE, Pacht ER, et al. Protein permeability in the adult respiratory distress syndrome: loss of size selectivity of the alveolar epithelium. J Clin Invest. 1986;78:1513.

61. Clark JG, Milberg JA, Steinberg KP, et al. Type III procollagen peptide in the adult respiratory distress syndrome. Association of increased peptide levels in bronchoalveolar lavage fluid with increased risk for death. Ann Intern Med. 1995;122:17.

62. Marshall BE, Hanson CW, Frasch F, et al. Role of hypoxic pulmonary vasoconstriction in pulmonary gas exchange and blood flow distribution. 2. Pathophysiology. Intensive Care Med. 1994;20:379.

63. Steltzer H, Krafft P, Fridrich P, et al. Right ventricular function and oxygen transport patterns in patients with acute respiratory distress syndrome. Anaesthesia. 1994;49:1039.

64. National Heart, Lung and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network. Comparison of two fluid-management strategies in acute lung injury. N Engl J Med. 2006; 354(24):2564.

65. Brigham KL, Kariman K, Harris TR, et al. Correlation of oxygenation with vascular permeability–surface area but not with lung water in humans with acute respiratory failure and pulmonary edema. J Clin Invest. 1983;72:339.

66. Martin GS, Moss M, Wheeler AP, et al. A randomized, controlled trial of furosemide with or without albumin in hypoproteinemic patients with acute lung injury. Crit Care Med. 2005;183:1681.

67. Dubois MJ, Orellana-Jimenez C, Melot C, et al. Albumin administration improves organ function in critically ill hypoalbuminemic patients: a prospective, randomized, controlled pilot study. Crit Care Med. 2006; 34:2536.

68. Finfer S, Bellomo R, Boyce N, et al. A comparison of albumin and saline for fluid resuscitation in the intensive care unit. N Engl J Med. 2004; 350:2247.

69. Marini JJ. New options for the ventilatory management of acute lung injury. New Horiz. 1993;1:489.

70. Van Hook CJ, Carilli AD, Haponik EF. Hemodynamic effects of positive end-expiratory pressure. Historical perspective. Am J Med. 1986; 81:307.

71. Stoller JK, Kacmarek RM. Ventilatory strategies in the management of the adult respiratory distress syndrome. Clin Chest Med. 1990; 11:755.

72. Brower RG, Lanken PN, MacIntyre N, et al. Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome. N Engl J Med. 2004;351:327.

73. Carroll GC, Tuman KJ, Braverman B, et al. Minimal positive end-expiratory pressure (PEEP) may be “best PEEP.” Chest. 1988; 93:1020.

74. Muscedere JG, Mullen JB, Gan K, et al. Tidal ventilation at low airway pressures can augment lung injury. Am J Respir Crit Care Med. 1994; 149:1327.

75. The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med. 2000;342:1301.

76. Dries DJ. Permissive hypercapnia. J Trauma. 1995;39:984.

77. Orchard CH, Kentish JC. Effects of changes of pH on the contractile function of cardiac muscle. Am J Physiol. 1990;258:C967.

78. Viitanen A, Salmenpera M, Heinonen J. Right ventricular response to hypercarbia after cardiac surgery. Anesthesiology. 1990;73:393.

79. Prough DS, Rogers AT, Stump DA, et al. Hypercarbia depresses cerebral oxygen consumption during cardiopulmonary bypass. Stroke. 1990; 21:1162.

80. McIntyre RC Jr, Haenel JB, Moore FA, et al. Cardiopulmonary effects of permissive hypercapnia in the management of adult respiratory distress syndrome. J Trauma. 1994;37:433.

81. Kallet RH, Jasmer RM, Luce JM, et al. The treatment of acidosis in acute lung injury with tris-hydroxymethyl aminomethane (THAM). Am J Respir Crit Care Med. 2000;161:1149.

82. Bryan AC. Conference on the scientific basis of respiratory therapy. Pulmonary physiotherapy in the pediatric age group. Comments of a devil’s advocate. Am Rev Respir Dis. 1974;110:143.

83. Douglas WW, Rehder K, Beynen FM, et al. Improved oxygenation in patients with acute respiratory failure: the prone position. Am Rev Respir Dis. 1977;115:559.

84. Pappert D, Rossaint R, Slama K, et al. Influence of positioning on ventilation–perfusion relationships in severe adult respiratory distress syndrome. Chest. 1994;106:1511.

85. Wiener CM, Kirk W, Albert RK. Prone position reverses gravitational distribution of perfusion in dog lungs with oleic acid–induced injury. J Appl Physiol. 1990;68:1386.

86. Lamm WJ, Graham MM, Albert RK. Mechanism by which the prone position improves oxygenation in acute lung injury. Am J Respir Crit Care Med. 1994;150:184.

87. Gattinoni L, Pelosi P, Vitale G, et al. Body position changes redistribute lung computed-tomographic density in patients with acute respiratory failure. Anesthesiology. 1991;74:15.

88. Gattinoni L, Tognoni G, Pesenti A, et al. Effect of prone positioning on the survival of patients with acute respiratory failure. N Engl J Med. 2001;345:568.

89. Curley MA, Hibbard PL, Fineman LD. Effect of prone positioning on clinical outcomes in children with acute lung injury. A randomized, controlled trial. JAMA. 2005;294:229.

90. Guerine E, Gaillard S, Lemasson S, et al. Effects of systematic prone positioning in hypoxemia acute respiratory failure: a randomized controlled trial. JAMA. 2004;19:2379.

91. Mancebo JM, Fernandez R, Blanch L, et al. A multicenter trial of prolonged prone ventilation in severe acute respiratory distress syndrome. Am J Respir Crit Care Med. 2006;173:1233.

92. Diaz JV, Brower R, Calfee CS, Matthay MA. Therapeutic strategies for severe acute lung injury. Crit Care Med. 2010;38:1644.

93. Hurst JM, Branson RD, Davis K Jr, et al. Comparison of conventional mechanical ventilation and high-frequency ventilation: a prospective, randomized trial in patients with respiratory failure. Ann Surg. 1990; 211:486.

94. Derdak S, Mehta S, Stewart T, et al. High-frequency oscillatory ventilation for acute respiratory distress syndrome in adults: a randomized, controlled trial. Am J Respir Crit Care Med. 2002;166:801.

95. Zapol WM, Snider MT, Hill JD, et al. Extracorporeal membrane oxygenation in severe acute respiratory failure: a randomized prospective study. JAMA. 1979;242:2193.

96. Anderson H III, Steimle C, Shapiro M, et al. Extracorporeal life support for adult cardiorespiratory failure. Surgery. 1993;114:161.

97. Peek GJ, Mugford M, Tiruvoipati R, et al. Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for sever adult respiratory failure (CESAR): a multicentre randomized controlled trial. Lancet. 2009;374:1351.

98. Benzing A, Brautigam P, Geiger K, et al. Inhaled nitric oxide reduces pulmonary transvascular albumin flux in patients with acute lung injury. Anesthesiology. 1995;83:1153.

99. Taylor RW, Zimmerman JL, Dellinger RP, et al. Low-dose inhaled nitric oxide in patients with acute lung injury: a randomized, controlled trial. JAMA. 2004;291:1603.

100. Adhikari NK, Burns KE, Friedrich JO, et al. Effect of nitric oxide on oxygenation and mortality in acute lung injury: systematic review and meta-analysis. BMJ. 2007;334:779.

101. Bernard GR, Luce JM, Sprung CL, et al. High-dose corticosteroids in patients with the adult respiratory distress syndrome. N Engl J Med. 1987;317:1565.

102. Crouch E. Pathobiology of pulmonary fibrosis. Am J Physiol. 1990; 259:L159.

103. Bachofen A, Weibel E. Alterations of the gas exchange apparatus in adult respiratory insufficiency associated with septicemia. Am Rev Respir Dis. 1977;116:589.

104. Meduri GU, Chinn AJ, Leeper KV, et al. Corticosteroid rescue treatment of progressive fibroproliferation in late ARDS: patterns of response and predictors of outcome. Chest. 1994;105:1516.

105. Ashbaugh DG, Maier RV. Idiopathic pulmonary fibrosis in adult respiratory distress syndrome: diagnosis and treatment. Arch Surg. 1985; 120:530.

106. Hooper RG, Kearl RA. Established ARDS treated with a sustained course of adrenocortical steroids. Chest. 1990;97:138.

107. Biffl WL, Moore FA, Moore EE, et al. Are corticosteroids salvage therapy for refractory acute respiratory distress syndrome? Am J Surg. 1995; 170:591.

108. Meduri GU, Headley AS, Golden E, et al. Effect of prolonged methylprednisolone therapy in unresolving acute respiratory distress syndrome: a randomized controlled trial. JAMA. 1998;280:159.

109. Meduri GU, Tolley EA, Chrousos GP, et al. Prolonged methylprednisolone treatment suppresses systemic inflammation in patients with unresolving acute respiratory distress syndrome: evidence for inadequate endogenous glucocorticoid secretion and inflammation-induced immune cell resistance to glucocorticoids. Am J Respir Crit Care Med. 2002;165:983.

110. Steinberg KP, Hudson LB, Goodman RB, et al. Efficacy and safety of corticosteroids for persistent acute respiratory distress syndrome. New Engl J Med. 2006;354:1671.

111. Ferring M, Vincent JL. Is outcome from ARDS related to the severity of respiratory failure? Eur Respir J. 1997;10:1297.

112. Erickson SE, Martin GS, Davis JL, et al. Recent trends in acute lung injury mortality: 1996–2005. Crit Care Med. 2009;37:1574.