Chapter 38. Acute Lung Injury and the Acute Respiratory Distress Syndrome

• Acute lung injury (ALI) and its more severe form, the acute respiratory distress syndrome (ARDS), are common causes of acute hypoxemic respiratory failure (AHRF).
• Both ALI and ARDS are characterized by hypoxemia that is resistant to oxygen therapy; this is due to widespread alveolar filling or collapse.
• Initial therapy for all patients with ALI and ARDS should be supplemental oxygen in the highest concentration available; failure to achieve 95% arterial saturation or greater confirms the presence of a large right-to-left shunt.
• Most patients with ALI and ARDS require early intubation because their hypoxemic respiratory failure is typically severe and may be prolonged.
• If a patient with severe hypoxemia as indicated by arterial blood gas analysis has a clear chest radiograph, consider a possible error (e.g., incorrect fractional inspired oxygen [Fio2] or arterial oxygen tension [Pao2]); in such situations, also consider the possibility of other types of right-to-left shunts (e.g., intracardiac shunts or pulmonary arteriovenous malformations) or the continued perfusion of an unventilated or poorly ventilated lung (e.g., due to acute mucous plugging of one main bronchus).
• The acute phase of ALI and ARDS is characterized by an exudative alveolar flooding due to pulmonary capillary leak and by extensive alveolar collapse due to loss of normal surfactant activity; while interventions directed at modulating inflammatory or other pathways of lung injury or restoring surfactant function hold theoretical promise, at present no specific pharmacologic therapy has been shown to improve outcomes; in this phase one provides mechanical ventilation and other supportive care while identifying and treating the precipitating cause or causes of ALI or ARDS.
• Ventilator management of patients with ALI and ARDS should use a strategy with low tidal volumes and limits to end-inspiratory pressure (i.e., plateau pressures [Pplat]), to reduce the risk of ventilator-induced lung injury (VILI); such a strategy gives higher priority to the goal of decreasing the risk of VILI by limiting end-inspiratory volume and pressure than the traditional goal of keeping arterial carbon dioxide tension (Pco2) and pH in the normal range.
• The target for oxygenation should be a Pao2 between 55 and 80 mm Hg (88% to 95% saturation); one should achieve this by adjusting Fio2 and positive end-expiratory pressure (PEEP) with the goal of decreasing Fio2 to 0.5 to 0.6 (or less), concentrations that are less concerning for pulmonary oxygen toxicity.
• In general, the ventilatory strategy should use a tidal volume of 6 mL/kg of predicted body weight (PBW) with a Pplat target that does not exceed 30 cm H2O; if Pplat exceeds 30 cm H2O with a 6 mL/kg PBW tidal volume, the latter should be decreased to 5 mL/kg PBW; if Pplat still exceeds 30 cm H2O, tidal volume should be further decreased to 4 mL/kg PBW.
• When changing to low tidal volume ventilation, one should increase the respiratory rate (RR) up to 35/min to maintain minute ventilation; the target for ventilation should be a pH of 7.30 to 7.45; if the pH is between 7.15 and 7.29, one may correct the respiratory acidosis by giving intravenous sodium bicarbonate (NaHCO3) (if RR is 35/min and arterial carbon dioxide tension [Paco2] is <25 mm Hg); if the pH remains <7.15 with RR of 35/min and NaHCO3 is considered or given, one can increase the tidal volume in 1–mL/kg PBW increments until pH exceeds 7.15 (i.e., the Pplat target may be exceeded).
• During assisted ventilation, use sedation to synchronize the patient's respiratory efforts with the ventilator and to decrease oxygen consumption V̇O2); one may also need to induce muscle paralysis by use of neuromuscular blocking agents, which despite their clinical utility, likely increase the risk of diffuse weakness in survivors.
• Despite the lack of results of a randomized clinical trial (RCT) to support its use, some clinicians favor a “dry" fluid management strategy to reduce edemagenesis and the duration of assisted ventilation; however, this approach generally lowers cardiac output and risks worsening or inducing nonpulmonary organ dysfunction.
• If one uses a “dry" fluid management strategy, one should be guided by daily and cumulative net intake and output volumes and laboratory indices of hypovolemia (blood urea nitrogen [BUN]:creatinine ratio and serum total bicarbonate) while monitoring for adequate urine output (>0.5 mL/kg PBW/h), effective circulation (by physical signs, or measurements of cardiac index or venous co-oximetry or both), adequate perfusing pressure (i.e., mean arterial pressure [MAP] ≥60 mm Hg), and electrolyte abnormalities (serum sodium and potassium).
• Caution! Although this strategy is supported by observational studies, no prospective RCT to date has established whether the “dry" fluid management strategy improves outcomes compared to a “non-dry" strategy; nor has the efficacy or safety of using a pulmonary artery or central venous catheter for this purpose been established.
• Late-phase or fibroproliferative-phase ARDS (corresponding to persistent ARDS of 1 week duration or more) is characterized by subacute inflammation, proliferation of alveolar lining cells and interstitial cells, and varying degrees of fibrosis.
• In severe late-phase ARDS, a prolonged course of high-dose methylprednisolone sodium succinate (MPSS) can improve gas exchange and mechanics in some patients; however, use of MPSS increases the risk of diffuse weakness that may be prolonged; preliminary results from one RCT found that MPSS administration when added to a low-tidal-volume ventilation strategy that limited Pplat as described above did not significantly improve mortality at 60 days (or ICU or hospital lengths of stay) compared with ventilation using a low tidal volume and limited plateau pressure strategy alone.

Severe arterial hypoxemia that is resistant to supplemental oxygen is a common reason for admission to the ICU. This form of respiratory failure, termed acute hypoxemic respiratory failure (AHRF), arises from widespread flooding or collapse of alveoli. As a result, a substantial fraction of mixed venous blood traverses nonventilated alveoli (i.e., alveoli with ventilation-perfusion ratio of 0). This in turn results in a large right-to-left intrapulmonary shunt (Fig. 38-1). In addition to its adverse effects on oxygenation, fluid that accumulates in alveoli and interstitial tissues increases lung stiffness, which decreases lung compliance. This imposes a larger mechanical load on the patient's respiratory system, resulting in an increased elastic work of breathing. Without treatment with supplementary oxygen and assisted ventilation, the gas exchange derangement and abnormal lung mechanics can result in progressive hypoxemia, respiratory muscle fatigue, and eventual respiratory arrest and death.

This chapter describes two notable forms of AHRF: acute lung injury (ALI) and its more severe subgroup, the acute respiratory distress syndrome (ARDS) (Fig. 38-2). ALI and ARDS affect an estimated 50,000 to 150,000 patients each year in the United States alone.1,2 Despite recent advances in ventilatory management and improvements in nonrespiratory supportive ICU care, the mortality rate remains high (e.g., ∼25% to over 50%, depending on the characteristics of the population studied and the ventilatory strategy utilized).3–9

The overall goal of this chapter is to provide a comprehensive and updated clinical approach to ALI and ARDS that is grounded in current concepts of pathogenesis and pathophysiology. Objectives include: (1) to present the definitions, epidemiology, and precipitating causes of ALI and ARDS; (2) to describe the pathogenesis and pathophysiology of ALI and ARDS; (3) to describe the differential diagnosis of ALI and ARDS and clinical approaches to distinguish among various causes of acute hypoxemic respiratory failure; (4) to review the current status of treatment for ALI and ARDS and their precipitating causes; and (5) to provide current recommendations for ventilatory management of ALI and ARDS. The last objective includes a detailed description of the low-tidal-volume ventilatory strategy used by the ARDS Clinical Trials Network (ARDSNet) investigators in the clinical trials sponsored by the National Heart, Lung and Blood Institute (NHLBI) of the National Institutes of Health (NIH).3,4

ARDS was initially described in 1967 by Ashbaugh and coauthors as a syndrome characterized by the acute onset of dyspnea, severe hypoxemia, diffuse lung infiltrates, and decreased compliance of the respiratory system.10 Following this initial report, different authors utilized varying definitions that incorporated the elements of acute timing of onset, hypoxemia, radiographic infiltrates, and absence of overt clinical congestive heart failure.11–15 In 1988, Murray and coauthors proposed using a Lung Injury Score (LIS), in part to take into account the effects of respiratory system compliance and the pressures delivered using the mechanical ventilator.14 The LIS combined elements of severity of chest radiograph infiltrates, hypoxemia, respiratory system compliance, and amount of administered positive end-expiratory pressure (PEEP) on the mechanical ventilator.14

American European Consensus Conference Definitions

In 1994, the first American European Consensus Conference (AECC) published these consensus-derived definitions of ALI and ARDS: acute onset, bilateral pulmonary infiltrates on chest radiograph consistent with pulmonary edema, poor systemic oxygenation, and the absence of evidence of left atrial hypertension (Table 38-1).16 The ratio of arterial oxygen tension (Pao2) to the fraction of inspired oxygen (Fio2), Pao2:Fio2, was chosen to reflect the degree of hypoxemia even when measured at different Fio2.17 The syndrome is called ALI when this ratio is ≤300 and ARDS when ≤200. The AECC coined the term acute lung injury in order to identify patients who are early in the course of their ARDS and those who may have a form of AHRF that is milder than ARDS. The AECC definitions of ALI and ARDS were intentionally quite broad in order to encompass different types of AHRF occurring in a wide variety of settings. Most patients with ALI progress to ARDS, prompting some to use the use the term “ALI/ARDS" to describe all patients with a Pao2:Fio2 ratio ≤300 who also meet the other AECC criteria (see Table 38-1).

Several clinical trials have used these standardized definitions of ALI and ARDS to specify the trial's inclusion criteria for their study populations.3,4,8 Using the AECC definition of ALI and ARDS in these and other clinical trials contributes to the comparability of their study populations and facilitates generalizability of their results. For example, in order for clinicians to extrapolate the results of a trial to their own patients, they must first be reasonably sure that their patients resemble those in the clinical trial. This requires that they meet the same inclusion criteria for ALI or ARDS as those used in the clinical trial. However, having the same inclusion criteria is necessary but not sufficient, since the clinician must also take into account the exclusion criteria used in the clinical trial to assess comparability.

Validity and Reliability of Aecc Definitions

Despite standardization, these definitions have not been formally validated beyond their face validity. In addition, problems remain with the reliability of the various components of the definitions.1 For example, interpretation of chest radiographs can be inaccurate and variable among different observers.18,19 However, formal training sessions can improve this variability.19 Likewise, the Pao2:Fio2 criterion also suffers from variability since it can be influenced by the level of PEEP used in mechanical ventilation and other transient factors, such as airway secretions or inadequate sedation. For example, higher PEEP generally increases Pao2 at a given Fio2. This in turn may raise a patient's Pao2:Fio2 >300 so that the patient no longer meets the inclusion criteria for ALI. Conversely, in the absence of PEEP a Pao2:Fio2 <300 may reflect the presence of simple basilar atelectasis rather than ALI or ARDS. In such a case, adding PEEP may recruit enough atelectatic lung to raise Pao2:Fio2 >300, so that the patient no longer meets this criterion for ALI.20

Finally, exclusion of congestive heart failure (left atrial hypertension) also presents problems with reliability. Diagnostic criteria for left atrial hypertension on purely clinical grounds may be inaccurate.1,21 Insertion of a pulmonary artery catheter for this purpose may also be inconclusive since the pulmonary artery occlusion pressure may be higher than 18 mm Hg due to intravascular volume loading (e.g., in patients with shock), rather than due to congestive heart failure.21 Conversely, many patients with pulmonary edema due to congestive heart failure and high left atrial pressures may have normal pulmonary artery occlusion pressures by the time the catheter is inserted and measurements taken.22

Further refinement of the reliability and validity of definitions of ALI and ARDS are desirable goals. More reliable definitions will not only improve estimates of the public health impact of these syndromes, but also decrease misclassification errors that are especially problematic for research aimed at clarifying molecular and genetic mechanisms in ALI and ARDS.23

Incidence

Inconsistent definitions for ALI and ARDS in large databases of diagnoses for hospital admissions or complications, variations in the application of the AECC definition in different studies, and variations in different at-risk populations (such as sepsis, trauma, surgeries, and bone marrow transplants) have hampered obtaining accurate estimates of the incidence of ALI and ARDS. Recent studies have used data from the ARDSNet clinical trials, as well as regional cohort studies to arrive at estimates in the United States in the range of 25 to 65 cases per 10,000 person-years.1,2,24 Thus the number of patients with ARDS and ALI in the U.S. each year is estimated to be between 53,000 and 140,000. If one assumes a short-term mortality of ∼30% to 40%, the annual number of deaths in the U.S. directly attributable to ALI and ARDS is between 16,000 and 56,000. These estimates of attributable mortality are greater than from the acquired immune deficiency syndrome (AIDS), asthma, or cervical cancer.1,2 In addition, ARDS and ALI contribute significantly to long-term mortality, morbidity, and diminished quality of life. (see Chap. 19).25–28 In addition to these estimates in the U.S., two non-U.S. studies arrived at similar incidence rates for ARDS in Berlin29 and Scandinavia.29 Several useful reviews have summarized the important public health impact of ALI and ARDS.1,2,30

Precipitating Causes

ALI and ARDS can be considered to be a “final common pathway" reaction of the lung to a large variety of precipitating causes. Some authors have classified these causes as representing direct (pulmonary) or indirect (extrapulmonary or systemic) injury to the lung.31,32 (Table 38-2). Although the methodology for applying these labels is not standardized and thus may be inconsistent between studies, this construct may have pathogenetic and pathophysiologic underpinnings since indirect causes may have different mechanisms of injury compared to direct injury to the lung.33 Not all patients with these precipitating conditions develop ALI/ARDS. Indeed, the frequency of ALI/ARDS is quite variable. Depending on the precipitating cause, it ranges from ∼10% to 40%. If patients have more than one of these precipitating causes, they are at an even higher increased risk for development of ALI or ARDS. Although most studies of the incidence of ARDS among different at-risk subgroups were performed before the AECC standardization of the definitions, they nonetheless provide broad estimates of the risk of development of ARDS in these subgroups.11−13,34

Risk Factors for the Development of ALI and ARDS

Why certain individuals with the same precipitating cause develop ALI and ARDS while others don't is not known. Some of this differential risk can be attributed to other acquired risk factors, such as chronic alcoholism.35 In addition, certain individuals may have an inherent predisposition to developing ALI or ARDS while others may have inherent protections against ALI or ARDS. These considerations are discussed below.

ALI and ARDS arise from diverse populations. Within these populations, specific clinical variables may affect both the risk of ARDS and ultimate outcomes if ARDS develops.34,36,37 Clinical variables and outcomes associated with an increased risk of ARDS include chronic alcohol abuse;38 lack of diabetes;39 hypoproteinemia;40 age and gender;13 severity of injury and illness as measured by injury severity score (ISS) or Acute Physiology, Age, and Chronic Health Evaluation (APACHE) score;13 hypertransfusion of blood products;13 and possibly cigarette smoking.41 The mechanistic underpinnings of these associations are the subject of ongoing research. Nonetheless, these observations highlight the influence of the heterogeneity of a diverse source population on the risks of ALI and ARDS. Human studies aimed at investigating genetic or molecular mechanisms that predispose or protect individuals from developing ALI and ARDS must take into account the effects of these population and environmental differences.

Determinants of ARDS Outcome

Clinical variables with reported associations with increased mortality among patients with ALI and ARDS include advanced age, low Pao2:Fio2, high plateau pressure (i.e., low compliance), extent of pulmonary infiltrates, chronic liver disease, nonpulmonary organ dysfunction, severity of illness, hypoproteinemia, and length of hospitalization prior to ARDS.4,38,42−48 In addition, magnitude of the dead space fraction has been identified as a risk factor for mortality, possibly indicating the importance of loss of the pulmonary vascular bed in disease severity.49

Different precipitating causes of ALI and ARDS carry different prognoses. For example, trauma-associated ALI or ARDS has a better prognosis than other causes, even after adjustment is made for other baseline variables such as age.48 Although response to some therapeutic interventions in ALI or ARDS varies according to direct or indirect cause of lung injury,50,51 low-tidal-volume ventilation is equally efficacious in both of these subgroups.52

Despite recent advances in therapy, the mortality rates reported in clinical trials in the past decade have ranged from ∼25% to more than 50%.3–9 The lowest mortality rates have been achieved with low-tidal-volume ventilation, which also limited end-inspiratory (plateau) pressure.3,4 Although no recent studies have explicitly examined the cause of death in nonsurvivors, previously the majority of deaths were found to be due to multiorgan failure or sepsis, while progressive respiratory failure accounted for only a small minority of deaths (e.g., ∼15%).53,54

ALI and ARDS impact the health and lives of survivors beyond the ICU. Long-term functional, neuropsychiatric, and cognitive dysfunction following ALI and ARDS has been described (see Chap. 19). In the aggregate, these deficits represent an important aspect of the impact of ARDS and its associated morbidity on the general population.

In cases of ALI and ARDS, it is useful to distinguish between the early phases of lung injury and subsequent events.36,55,56 By light microscopy, the early appearance is of interstitial and alveolar edema, capillary congestion, and intra-alveolar hemorrhage with minimal evidence of cellular injury (Fig. 38-3). By electron microscopy, changes of endothelial cell swelling, widening of intercellular junctions, increased numbers of pinocytotic vesicles, and disruption and denudation of the basement membrane are prominent. Inflammatory cell infiltration of the lung interstitium may be seen (particularly in ARDS complicating sepsis or trauma) as well as neutrophil sequestration in alveolar capillaries, but it usually a subtle finding. During this early exudative phase of diffuse alveolar damage (DAD), pulmonary edema and its clinical effects are most pronounced (see Fig. 38-3). It is also a time when manipulations to decrease the rate of edemagenesis are most likely to have impact as discussed below.

Over the ensuing days, hyaline membrane formation in the alveolar spaces becomes prominent. Hyaline membranes contain condensed fibrin and plasma proteins. Intra-alveolar activation of the coagulation system results in the formation of the fibrin, while plasma proteins are deposited in the alveolar space as part of the inflammatory exudate that leaks across the alveolar-capillary membrane. Inflammatory cells become more numerous within the lung interstitium. As the process of DAD progresses, there is extensive necrosis of type I alveolar epithelial cells. If the patient with ALI and ARDS does not recover or die during the first week, they may have a prolonged course of illness, termed late-phase ARDS.

The late phase of ARDS is dominated by disordered healing. This can occur as early as 7 to 10 days after initial injury and may eventually result in extensive pulmonary fibrosis. This has been termed the proliferative or fibroproliferative phase. Type II alveolar cells proliferate along alveolar septae and the alveolar walls; fibroblasts and myofibroblasts become more numerous. Evidence of lung flooding is less prominent and may be minimal at this point. Changes in the clinical manifestations of ALI and ARDS parallel the changes in pathology. One study found that patients in late-phase ARDS typically had a large dead space fraction, a high minute ventilation requirement, progressive pulmonary hypertension, slightly improved intrapulmonary shunt that is less responsive to PEEP, and a further reduction in lung compliance.57

A number of closely inter-related pathophysiologic mechanisms and systems contribute to the development of ALI and ARDS (Fig. 38-4). Inflammatory cytokines, oxygen radicals, activation of coagulation and complement, platelet and immune cell activation, proteases, leukotrienes, and eicosanoids have been hypothesized to play a role in the early stages of ALI and ARDS.33,36,37,58 In addition to ongoing inflammation and oxidation, factors specific to apoptosis, edema fluid resolution, fibrosis, and repair are likely to be important in the resolving and late phases of ALI and ARDS.33,36,37,58,59

Inflammatory Cytokines and Chemokines

Cytokines are small signaling proteins that participate in host defense by binding to cell surface receptors. Many types of cells produce cytokines in response to a variety of stimuli. Cytokines teleologically must provide a net benefit in terms of contributing to host defense. However, when inflammation is overly exuberant, they have been implicated in disease states, including ALI and ARDS.60,61

Two of the major early proinflammatory cytokines are tumor necrosis factor-α (TNF-α) and interleukin-1 (IL-1), both of whose production can be increased by hypoxia.62 These cytokines have similar effects in initiating and propagating inflammation.60,63−66 Their actions include: (1) recruitment, differentiation, amplification, and localization of macrophages to the lung parenchyma; (2) stimulation of other inflammatory cytokines such as interleukins-6 and -8 (IL-6 and IL-8); and (3) adherence of neutrophils to endothelium (see Fig. 38-4).58,67−70

In clinical studies, the most convincing evidence of the involvement of TNF-α and IL-1 in ARDS has been from direct measurement in bronchoalveolar lavage (BAL) or pulmonary edema fluid. Both fluids have consistently been reported to contain elevated levels of TNF-α and IL-1 in ARDS patients.71–75 Furthermore, some suggest that much of the proinflammatory state of BAL in ARDS may be mediated through IL-1.76 More recent studies suggest that an imbalance between inflammatory cytokines and their soluble receptors are an important pathogenetic mechanism in ARDS.77

Not only do TNF-α and IL-1 play an important role in the early pathogenesis and risk of developing ALI and ARDS, but persistently high levels of TNF-α and IL-1 in plasma and in BAL have also been associated with an increased risk of death in patients with ALI and ARDS.61,78

In addition to IL-1 and TNF-α, interleukin-6 (IL-6) has been suggested as an important mediator of lung injury. IL-6, a product of macrophages that have been stimulated by TNF-α and IL-1, has diverse functions. In mediating acute lung injury, IL-6 has traditionally been viewed as a proinflammatory cytokine. This view is based largely on the clinical observations of elevated levels in sepsis and ARDS.79,80 However, laboratory experiments indicate that IL-6 may also have anti-inflammatory effects.81

Like TNF-α and IL-1, persistently elevated IL-6 levels have been associated with an increased risk of death in ARDS patients.78 In addition, in the ARDSNet randomized clinical trial (RCT) that compared lower- and higher-tidal-volume ventilatory strategies, the lung protective treatment arm that had decreased mortality showed a significantly steeper fall in IL-6 levels over the first 72 hours compared to the higher-tidal-volume arm.3 In a separate RCT comparing low-tidal-volume versus traditional-tidal-volume ventilation in early ARDS, the former had lower concentrations of inflammatory mediators (TNF-α, IL-1, and IL-6) in BAL fluid and plasma over the first 36 hours of the trial.82 Thus although the exact mechanisms of IL-6 in mediating acute lung injury are uncertain, it may have an important role in clinical ALI and ARDS.

Interleukin-10 (IL-10) is a potent anti-inflammatory cytokine produced by monocytes as well as by both T and B lymphocytes. IL-10 is a potent inhibitor of proinflammatory molecules including TNF-α and IL-1.83 Lower levels of IL-10 in blood and lung lavage fluid have been association with development of ARDS.77,84,85 Furthermore, lower circulating levels of IL-10 have been associated with death in patients with established ARDS.86

Chemokines are small polypeptides that certain cells elaborate in response to a variety of proinflammatory mediators. Chemokines act locally on modulating leukocyte immune response. The most studied chemokine in ARDS is interleukin-8 (IL-8), which has been implicated in neutrophil chemotaxis. In a variety of lung injury models, blockage of IL-8 with directed antibodies has led to attenuation of the injury.87–89 IL-8 has also been associated with greater numbers of neutrophils in BAL fluid of patients with ARDS as well as with mortality.61,90−92

Given the interrelationship and overlapping regulatory mechanisms of cytokine and chemokine molecules, future studies will need to focus on cytokine and chemokine imbalances, (i.e., their interactions with antagonist molecules and soluble receptors), as well as time-dependent differences. Finally, many studies have measured cytokine concentrations, whereas given the existence of inactivating molecules, measuring biological activity may be more productive.

Neutrophils and Oxidant Stress

Neutrophil migration into the airspaces is evident in the early histopathology of ARDS.93,94 BAL fluid of subjects with ARDS reveals a predominance of neutrophils.95 Conversely, recovery is associated with resolution of neutrophilia.96 Normally, neutrophils become apoptotic and are then removed by macrophages. Recent evidence suggests that impaired neutrophil apoptosis may be an important mechanism in early ARDS,97 and that elevated levels of granulocyte-macrophage colony-stimulating factor (GM-CSF) may play a role in this mechanism.98–100

Neutrophils in the alveolar and interstitial spaces may result in lung injury due to neutrophil-mediated release of reactive oxidant species (ROS) and resulting oxidant stress. Oxidant stress refers to an imbalance between endogenously produced oxidants and endogenous antioxidants. Products of oxidant stress include superoxide, hydroxyl radicals, and peroxynitrite.101 Although some levels of ROS may be important in normal homeostasis, ROS are highly unstable and react quickly with surrounding proteins, DNA, and lipids, resulting in molecular damage.101

A widely held hypothesis is that excessive oxidant stress is a major early contributor in the pathogenesis of acute lung injury.102–109 ROS can be generated by neutrophils (polymorphonuclear [PMN] cells),102,104,110 or by the pulmonary endothelium.111,112 Both PMN-mediated and endothelial generation of ROS may be important to initiation and development of lung injury.

Evidence from clinical studies supporting excess ROS in ALI and ARDS include findings of: increased hydrogen peroxide in exhaled breath of ARDS patients;113,114 decreased levels of glutathione in lung lavage fluids of ARDS patients;115,116 increased nitrotyrosine and chlorotyrosine in lavage fluids of ARDS patients;117 increased lipid peroxidation products in plasma;118,119 increased protein carbonyl levels;120 increased plasma hypoxanthine;121 increased levels of nitrated fibrinogen in the plasma of ALI/ARDS patients;122 and increased serum ferritin levels in ARDS patients compared with at-risk patients.123,124 Although some of these findings may be nonspecific for ARDS,58 taken together, they make a strong case for the role of oxidant stress in the pathogenesis of ARDS. In addition, evidence of higher levels of oxidant stress in nonsurvivors of ARDS is illustrated by lower levels of plasma thiol groups,125 and higher levels of hydrogen peroxide in urine of ARDS patients who are nonsurvivors.126

Coagulation and Fibrinolysis

As noted above, the histopathology of the early exudative phase of ALI and ARDS is notable for diffuse alveolar fibrin deposition in the form of hyaline membranes (see Fig. 38-3).55,56,94,127,128 In addition, a shift in the procoagulant and anticoagulant balance to favor coagulation has been demonstrated in the BAL fluid of patients with ARDS.129–133 Furthermore, there are likely important links between the coagulation system activation and activation of the inflammatory system. For example, TNF-α and IL-1 can act synergistically to produce a procoagulant state through effects on tissue factor, thrombomodulin, and plasminogen activator inhibitor.134–136 Given the success of intravenous activated protein C in treating sepsis (see Chap. 46),137 as well as the observation that the mortality benefit in this trial was also evident in subjects with severe sepsis and pulmonary dysfunction as a single organ system dysfunction,138 interventions aimed at correcting abnormalities in coagulation and fibrinolysis in ALI and ARDS are a promising future research direction.

Alveolar Edema Clearance

Considerable evidence supports the active transport of sodium and water by the pulmonary epithelium from the distal airspaces of the lung.139 This process appears to be regulated by epithelial sodium channels predominantly on alveolar type I cells, potentially stimulated by catecholamines.140,141 In addition, recent evidence suggests a potential role for the cystic fibrosis transmembrane conductance regulator (CFTR) in alveolar fluid transport.142 Furthermore, the directional flow of fluid across the alveolus may be altered in conditions of hypoxia,143,144 as well as by reactive nitrogen and oxygen species.145

In ALI and ARDS, the pulmonary epithelium is injured, resulting in altered fluid transport. Clinical studies in ventilated patients have illustrated impaired fluid transport in ARDS when compared with hydrostatic pulmonary edema.36,146,147 In addition, subjects who could rapidly remove alveolar fluid had a lower hospital mortality.147,148 Whether these findings are simply a result of disruption of the alveolar epithelium, or whether they represent a finding potentially amenable to therapy (such as with β-adrenergic receptor agonists), awaits the results of interventional clinical trials.

Fibroproliferation and Repair Mechanisms

Although there is considerable temporal overlap with the exudative and inflammatory phases (see Fig. 38-3), the late phase of ARDS is characterized by accumulation of matrix and cells in the interstitial and alveolar spaces, contributing to disruption of the alveolar architecture.55,56 Recovery from ARDS requires resolution of this phase of the tissue injury–repair spectrum with some degree of the normal lung architecture. Clinical studies of ARDS patients indicate that the alveolus is repaired under conditions of ongoing inflammation.61 In addition, soluble collagen precursor, type III procollagen peptide (PCP III), in the edema fluid of early ARDS can differentiate between subjects who will have longer or shorter courses of ARDS.149 This suggests that fibroproliferation begins early in the course of lung injury. In addition to these findings with PCP III, other evidence of the prognostic significance of fibroproliferation in ARDS includes the finding that fibroblast mitogenic activity in BAL fluid is lower in survivors of ARDS than nonsurvivors.150,151

A complex balance of different growth factors, cytokines, and chemokines occurs in the fibroproliferative and reparative phases of acute lung injury.58,152 A variety of animal, basic, and clinical studies have demonstrated the importance of transforming growth factor-β (TGF-β) to fibroproliferation in the lung.153,154 A key step in fibroproliferation involves the activation of TGF-β from its latent form. This can be accomplished through interaction with a variety of mediators, including matrix metalloproteinases (MMPs), thrombospondin, plasmin, acid environments, α2-macroglobulin, and reactive oxygen species.153 In addition, TGF-β has extensive interactions with inflammatory cytokines, chemokines, and interferon-γ in regulating fibrosis.

Other important modulators of the remodeling process include matrix and cell surface proteoglycans and glycosaminoglycans;155–157 matricellular proteins that affect cell adhesion;158,159 matrix metalloproteases;160 and the balance of coagulation and fibrinolysis (particularly through the actions of plasminogen activator inhibitor-1 [PAI-1]).129 The interplay of these mediators with the potential for targeted therapies is the focus of current investigations.

Clinical Presentation and Differential Diagnosis

Acute hypoxemic respiratory failure has many etiologies besides ALI and ARDS (Table 38-3). However, the bedside appearance of patients with various forms of AHRF is remarkably similar. Marked tachypnea and dyspnea are invariably present. Physical examination reveals diffuse crackles in cases of cardiogenic pulmonary edema and focal findings of consolidation in cases of lobar pneumonia. Cardiogenic pulmonary edema may be accompanied by evidence of airflow obstruction, including wheezing and hypercapnia.161 The presence of crackles, a radiologic appearance of high-pressure edema (see below), and hypoxemia refractory to oxygen therapy, all suggest cardiogenic pulmonary edema as the primary process. Cough and purulent sputum are hallmarks of infectious processes, while copious clear or pink-colored airway secretions result from fulminant (“flash") cardiogenic pulmonary edema.

Distressed patients with AHRF typically have initial room air arterial blood gas results with Pao2 in the 30 to 55 mm Hg range and pulse oximetry less than 85% of arterial O2 saturation. If supplemental oxygen by mask or cannula raises arterial saturation to above 95%, a large intrapulmonary shunt is unlikely. Other causes of respiratory distress should then be considered, including airways disease, pulmonary embolus, or severe metabolic acidosis. Failure to achieve >95% saturation of arterial blood with supplemental oxygen indicates the presence of a large right-to-left shunt. The specific process should be investigated via physical examination and chest radiograph. In the rare instances that the chest radiograph is entirely clear of alveolar infiltrates, one should consider that the blood gas data are erroneous, that there is an anatomic right-to-left shunt at another site (e.g., pulmonary arteriovenous malformations or intracardiac shunt), or that there is continued perfusion of an unventilated lung due to recent complete or nearly complete occlusion of its main bronchus (but before the lung has collapsed due to absorption atelectasis) (see Table 38-3).

The differential diagnosis of ALI and ARDS (i.e., AHRF with diffuse pulmonary infiltrates consistent with pulmonary edema in the absence of a cardiac etiology) includes a variety of disorders and etiologies. Identifying the etiologies of the diffuse infiltrates is important because specific treatments exist for several of these conditions (e.g., acute eosinophilic pneumonia or diffuse alveolar hemorrhage). Table 38-4 lists the major clinical and diagnostic characteristics of these disorders.

Clinical Setting

The clinical setting in which the disorder develops can provide important diagnostic information. Cardiogenic edema is most often accompanied by systolic left ventricular or valvular dysfunction, and the abnormal heart sounds and murmurs associated with each should be sought. Electrocardiographic (ECG) and serum enzyme evidence of ischemia should be considered and suggest an obvious cause for cardiogenic edema. Review of intravascular volume administration often will supply information suggesting the explanation for pulmonary edema in patients with left ventricular or renal dysfunction.

ALI and ARDS commonly arise in a typical clinical context (see Table 38-2). Sepsis, pneumonia, trauma, transfusion of blood products, and acid aspiration account for the majority of cases of ALI and ARDS.3,4,13,36 Less common causes include pancreatitis, near-drowning, leukoagglutination reactions, lung infections with viral agents or Pneumocystis carinii, fat embolism syndrome, and drug toxicities.34

Chest Radiograph

The chest radiograph is a simple and widely available test used to assess patients with AHRF. Unfortunately, the accuracy of the routine radiograph in distinguishing hydrostatic from increased permeability edema is not high.162,163 Criteria that have been suggested to support a diagnosis of hydrostatic edema include increased heart size, increased width of the vascular pedicle, vascular redistribution toward upper lobes, septal lines, and a centrifugal pattern of spread with a perihilar bat's-wing distribution of the edema. The lack of these findings and patchy peripheral infiltrates that extend to the lateral lung margins suggest ARDS. However, all these signs overlap, and in the best of hands this test is unlikely to yield better than a 60% to 80% accuracy of diagnosis when applied without other diagnostic tools.162

Echocardiography

Echocardiography is a useful source of noninvasive information regarding cardiovascular function.164,165 Left ventricular dilatation, regional or global wall motion abnormalities, and substantial mitral regurgitation on Doppler imaging support a diagnosis of cardiogenic edema. A heart with echocardiographically normal dimensions and function (both systolic and diastolic) in a patient with pulmonary edema suggests pulmonary vascular leakage, although prior ventricular or valvular dysfunction with intercurrent resolution of the high pulmonary vascular pressures predisposing to cardiogenic edema must be kept in mind.

Right Heart Catheterization

Right heart catheterization is often performed in patients with pulmonary edema, although the benefits of this invasive procedure continue to be debated;166–168 it has even been suggested that this technology itself contributes to a poor patient outcome.169,170 An observational study suggested that routine right heart catheterization was harmful, with a 33% increased risk of death if used in seriously ill patients with acute hypoxemic respiratory failure.169 In addition, several investigations have demonstrated that physician understanding of and facility with the information generated are remarkably poor.169,171 Thus until results of the current NIH NHLBI ARDSNet clinical trial that is testing the safety and efficacy of pulmonary artery catheters versus central venous catheters in patients with ALI and ARDS are available, right heart catheterization cannot be recommended as a routine procedure in patients with ALI/ARDS.

However, it is recommended that clinicians use invasive hemodynamic methods, including right heart catheterization, to address specific questions regarding ventricular function, the adequacy of volume resuscitation, the degree of intrapulmonary shunt, and the adequacy of cardiac output and oxygen saturation of mixed venous or central venous blood. The specific pulmonary capillary wedge pressure (Ppw) that the AECC definition (see Table 38-1) used as the criterion to distinguish noncardiogenic from cardiogenic pulmonary edema (i.e., 18 mm Hg), was an arbitrary decision based on physiologic experiments, tradition, and volume resuscitation practices circa 1992. Since then, reports have indicated that some patients who originally met criteria for ALI or ARDS as listed in Table 38-1, but who have been volume loaded to increase their cardiac output, have periods when measurements of their Ppw are above 18 mm Hg.21 In the mechanically ventilated patient with normal lung function and serum oncotic pressure, cardiogenic edema is typically associated with a Ppw of 28 mm Hg or above.172 However, lower plasma oncotic pressure (e.g., due to hypoalbuminemia) will result in pulmonary edema at somewhat lower values of Ppw.173 Finally, caution should be used when trying to rule in cardiogenic pulmonary edema by simply measuring the Ppw after a bout of acute pulmonary edema, since such patients often have Ppw less than 18 mm Hg by the time the catheter has been passed.22

Bronchoalveolar Lavage

Acute eosinophilic pneumonia is a rare disorder that is characterized by diffuse AHRF due to eosinophilic infiltrates in the lungs.174–176 It is notable for its responsiveness to corticosteroid therapy. When the precipitating cause for ARDS is unclear, it is recommended to perform a bronchoalveolar lavage and measure the percentage of eosinophils in the lavage fluid.174 Lavages can generally be done safely in many patients with ALI and ARDS except those with the lowest values of Pao2: Fio2 or hemodynamic instability.177,178

Likewise, a bedside bronchoscopy with BAL can be diagnostic for diffuse alveolar hemorrhage (DAH). In this case, the bronchoscopy may or may not reveal fresh blood in the trachea and major bronchi. However, BAL generally produces a bloody return, which may deepen in red color as the lavage continues. DAH occurs commonly in the first week or two post–bone marrow transplantation.179,180 DAH also occurs in association with a variety of vasculitic disorders. These include Goodpasture's syndrome, Wegener's granulomatosis, systemic lupus erythematosus, and antiphospholipid antibody syndrome (see Chap. 104).181–186 Finally, DAH also may result from inhalation of crack cocaine.181 For this cause of DAH, careful history taking and sending the patient's urine for toxicology analysis for cocaine may help determine the etiology.

Treat the Precipitating Cause of ALI and ARDS and Other Serious Comorbidities

A key early step in treating patients with ALI and ARDS is to identify and treat the precipitating cause or causes of the ALI and ARDS as well as any other serious and life-threatening comorbidities (Fig. 38-5). The ventilatory and other supportive management of ALI and ARDS is inadequate if not accompanied by aggressive attempts at diagnosis and treatment of the precipitating cause(s) (Table 38-2). Because ARDS is a syndrome based on nonspecific radiographic and physiologic criteria (Table 38-1), making the diagnosis of ALI or ARDS is not equivalent to diagnosing the patient's underlying problem. Not appreciating this seemingly obvious fact will delay diagnostic procedures in these patients and may delay therapy of a potentially treatable underlying disorder (Table 38-5).

For example, although appropriate supportive therapy may transiently stabilize a patient with ARDS due to sepsis from an abdominal abscess, if clinicians delay performing diagnostic tests such as abdominal CT scan or ultrasonography of the biliary tract in a timely manner, the underlying source of sepsis will go undiagnosed and the patient will eventually deteriorate. Likewise, the timely start of empiric antimicrobial therapy in patients with ALI or ARDS associated with severe sepsis or septic shock is as important as a timely diagnostic work-up (see Chaps. 45 and 46). Finally, if the precipitating cause of ALI and ARDS is unclear, one should consider performing early fiberoptic bronchoscopy to obtain bronchial washings for cytologic and microbiologic analyses, or in selected cases, surgical lung biopsy.

Ventilator Management of Respiratory Abnormalities

Maintaining Adequate Arterial Oxygenation

The hallmark respiratory abnormality of ALI and ARDS is hypoxemia that is resistant to oxygen therapy. This is due to the presence of a large right-to-left intrapulmonary shunt arising from fluid-filled and collapsed alveoli (see Fig. 38-1). Maintaining adequate arterial oxygenation is a goal given high priority by both traditional and more recent approaches to ventilator management, as noted below. One should use sufficient PEEP to reduce the right-to-left shunt to oxygenate the patient. By this approach, clinicians should be able to avoid prolonged exposure of such patients to potentially toxic concentrations of high inspired oxygen (e.g., Fio2 of 0.7 and above).

PEEP improves arterial oxygenation, primarily by recruiting collapsed and partially fluid-filled alveoli and thereby increasing the functional residual capacity (FRC) at end-expiration.187,188 PEEP also redistributes alveolar fluid into the interstitium,189 which should also improve oxygenation.

Noninvasive Ventilation

Assisted ventilation is generally done invasively via an endotracheal tube, but in selected cases noninvasive ventilation (NIV) may be successful190,191 (see Chap. 33). Although NIV seems to be useful in respiratory failure in immunocompromised hosts,190 we believe it is generally not a good choice for most patients with ALI and ARDS. This is because ARDS typically has a long course and is often associated with hemodynamic instability, coma, and multiorgan system failure (including ileus).192

Comparison of Traditional and Current Approaches to Ventilator Management

Goals and Priorities of the Traditional Approach

The traditional approach to ventilator management of patients with ALI and ARDS gave high priority to these goals: (1) to maintain arterial O2 saturation (O2sat) above 88% to 90% to provide for adequate tissue oxygenation while trying to minimize lung injury due to high concentrations of inspired oxygen (oxygen toxicity); (2) to provide sufficient ventilation to keep arterial pH and Paco2 within normal limits193 (Fig. 38-6).

To achieve the first goal, clinicians applied various levels of PEEP. This use of PEEP was first described in the initial description of ARDS by Ashbaugh and coworkers in 1967.10 Clinicians increased levels of PEEP in order to decrease Fio2 to below 70% while monitoring for adverse circulatory effects of PEEP.194 Since arterial oxygenation was found to be determined in part by mean airway pressure, they also used relatively large tidal volumes of 10 to 15 mL/kg. These were double to triple spontaneous tidal volumes, which are of the order of 5 mL/kg. Both the use of PEEP and traditional large tidal volumes (delivered at high flow rates) generally result in relatively high peak and end-inspiratory pressures (plateau pressure [Pplat]) in patients with ALI or ARDS, whose lungs typically have decreased compliance.

To achieve the second goal, clinicians ventilated patients with ALI and ARDS with relatively large tidal volumes of 10 to 15 mL/kg at high respiratory rates if needed. The resulting high minute ventilation was needed to produce a normal alveolar ventilation. This is because patients with ALI and ARDS typically have increased physiologic dead spaces and elevated dead space:tidal volume ratios (Vds:Vt) (e.g., up to 75%).195,196 In some cases, patients were exposed to high peak and end-inspiratory pressures due to the large tidal volumes in order to maintain normal arterial blood Pco2 and pH. Clinicians also used large tidal volumes and high inspiratory flow rates as a supplement to sedation to decrease patient discomfort while receiving assisted ventilation.

For example, if a patient with ARDS had a normal CO2 production (e.g., 200 mL/min), and a normal Vds:Vt of 0.3, then a minute ventilation of ∼7 L/min is needed to keep the patient's Paco2 at 40 mm Hg.197 However, when the patient's Vds:Vt increases, then additional minute ventilation is needed to keep Paco2 at 40 mm Hg. In this example, 14 L/min of minute ventilation is needed if the Vds:Vt is 0.66, and 18 L/min is needed if the Vds:Vt is 0.75,197 both of which occur in patients with ARDS.195,196 For a patient with a predicted (lean) body weight (PBW) of 60 kg, one can easily achieve a minute ventilation of 18 L/min with tidal volumes of 600 to 900 mL at rates of 20 to 30 per minute.

Goals and Priorities of Lung-Protective Approach

The lung-protective approach to ventilator management has the same goal for oxygenation as the traditional approach (i.e., to maintain an arterial saturation greater than 88% to 90%). However, it gives higher priority to protection from ventilator-induced lung injury (VILI) (see Chap. 37) than to normalization of arterial Pco2 and pH.193 The lung-protective approach's goal to decrease risk of VILI often conflicts with the traditional approach's goal to provide a high minute ventilation to keep arterial pH and Pco2 within normal limits (see Fig. 38-6). This conflict arises since the current lung-protective approach reduces the risk of VILI by decreasing the size of the tidal volume from the traditional 10 to 15 mL/kg body weight to tidal volumes of 4 to 6 mL/kg predicted body weight (PBW). Even with respiratory rates up to 35/min, such low tidal volumes will limit the resultant minute ventilation. This may result in a degree of permissive hypercapnia in some patients with ALI and ARDS.

For example, for a 60-kg PBW patient, a tidal volume of 6 mL/kg PBW (360 mL) with a respiratory rate of 35/min produces a minute ventilation of only 12.6 L. If one needs to reduce the tidal volume for the same 60-kg PBW patient to 4 mL/kg PBW (240 mL), in order to keep the Pplat from exceeding the threshold of 30 cm H2O, at the same respiratory rate of 35/min it provides only 8.4 L of minute ventilation. It is likely that a patient with ARDS ventilated with tidal volumes of 4 to 6 mL/kg PBW has a Vds:Vt of 0.66 or greater. This is due to the combined effects of an increased physiologic dead space in ARDS195,196 and the fact that ventilating with a lower tidal volume ventilation decreases the denominator of the patient's Vds:Vt. If the patient had a Vds:Vt of 0.66, low tidal volume ventilation, which provides 8.4 to 12.6 L/min, will result in permissive hypercapnia since, as described above, 14 L/min is needed to maintain this patient's Paco2 at 40 mm Hg. For example, for a patient with a Vds:Vt of 0.66, a minute ventilation of 12.6 L/min would result in a Paco2 of ∼45 mm Hg, while 8.4 L/min of minute ventilation would result in a Paco2 of ∼65 mm Hg.197

Decreasing the Risk of Ventilator-Induced Lung Injury

The important change in priority of the lung protective approach of ventilator management of patients with ALI and ARDS compared to the traditional approach is the result of a remarkable confluence of two lines of scientific research that culminated in a landmark confirmatory randomized controlled clinical trial3 (Fig. 38-7). The first line of basic research initially studied effects of mechanical forces (high pressure or high volume or both) in animal models of lung injury and then extended these observations to isolated lungs in situ or in vitro, and eventually to isolated lung cells. The second line of research involved careful clinical observations of patients with ALI and ARDS that examined the effects of systematic changes in selected ventilatory parameters with their physiologic effects and radiographic changes.187

Basic Research Related to Ventilator-Induced Lung Injury

Although Chap. 37 provides a more comprehensive description of this research, review of some of the early reports may prove useful since it specifically relates to this chapter's recommendations for ventilating patients with ALI and ARDS. In 1974 Webb and Tierney reported that mechanical ventilation using large tidal volumes and high inflation pressures could cause a fatal lung injury (similar morphologically to ARDS) in rats with otherwise normal lungs.198 In 1985, Dreyfuss and colleagues199 reproduced these experiments and carefully studied the changes that occurred within the lung. They observed that the injury that occurred was morphologically and pathophysiologically similar to ARDS and hypothesized that mechanical ventilation with large tidal volumes or high inflation pressures might exacerbate or perpetuate the lung injury in patients suffering from ARDS. Two major questions were raised by this early work: Did high inflation pressures or large tidal volume excursions cause the lung injury? Did positive end-expiratory pressure (PEEP) worsen or attenuate this injury?

Dreyfuss and associates200 then designed a set of experiments to answer these questions. They studied animals with normal lungs subjected only to varied protocols of mechanical ventilation. Some were ventilated with high pressures and large tidal volumes. Others had chest banding to limit chest wall and lung excursion during ventilation at high airway pressures. Another group was subjected to negative-pressure ventilation to assess the effect of large tidal volume excursions in the absence of high airway pressures. PEEP (10 cm H2O) was applied to the lungs of some animals undergoing high-pressure/large-tidal-volume ventilation. Finally, control animals were ventilated using parameters typical of conventional ventilation. These investigators found that large tidal volumes were associated with lung injury, but that high inflation pressure in the absence of large tidal volume excursion was not. Surprisingly, PEEP had the effect of limiting the injury that occurred in animals ventilated with large tidal volumes and high pressures. The application of PEEP not only preserved gas exchange, but also prevented the morphologic progression of the lesion caused by this pattern of ventilation. These observations have been verified in larger animals.201–204 (see Chap. 37 for more details).

A summary of these and subsequent animal experiments indicated that: (1) high-tidal-volume ventilation results in a lung injury morphologically similar to ARDS in humans, (2) PEEP is protective to some degree against this injury, (3) high-tidal-volume ventilation also can result in multiorgan system failure in otherwise healthy animals,205,206 and (4) high-tidal-volume ventilation results in the release of inflammatory cells and proinflammatory cytokines.207,208

Clinical Research Utilizing Lung Imaging

The second parallel line of research, noted above, studied patients with ARDS by use of CT scans (see Fig. 38-7). The initial observations indicated that despite a homogeneous “white-out" pattern of diffuse pulmonary edema infiltrates on chest radiographs of patients with ALI and ARDS, the CT scans of the same patients often proved the infiltrates to have remarkable heterogeneity.188,209 Based on strata of Hounsfïeld units to define different densities, the lung in ARDS could be partitioned into various compartments: nonaerated, poorly aerated, and normally aerated.187 Moreover, these compartments could be tracked before and after PEEP was added or subtracted and before and after tidal volumes of different sizes were delivered to the lung.

The end result was a reconceptualization of the mechanical changes in the ARDS lung.187 The traditional interpretation was that the stiffness of lungs in ARDS (e.g., low static compliances of 20 to 40 mL/cm H2O) represented many alveoli with similarly low specific compliances (i.e., essentially a single compartment in terms of its mechanical properties). Instead, studies using the CT scans and their pathophysiologic correlations indicated that a more accurate picture is that the lung in ARDS is multicompartmental and that there is a small part of the ARDS lung that has relatively normal compliance. Gattinoni and coworkers coined the term “baby lung" when referring to this compartment and its vulnerability to overdistension.187 In this revised conception of the acutely injured lung, some alveoli are normally compliant and vulnerable to overdistension while others are flooded or collapsed. The loss of functional alveoli necessitates that the tidal volume be distributed to far fewer aerated alveoli than in a healthy lung. Indeed, the apparent stiffness of the lungs of ARDS patients is regarded as the result of a small fraction of the lung containing relatively normal alveoli that becomes stiff as those alveoli reach their limits of distension, rather than due to generalized parenchymal “stiffness."

Based on the consistent results from the animal and in vitro studies referred to above and the insights provided by the results of the CT scans, it was hypothesized that traditional tidal volumes (e.g., 10 to 15 mL/kg) caused overdistension of alveoli in the lungs of patients with ALI and ARDS. This in turn not only resulted in exacerbation and perpetuation of their lung injury, but also, through the possible release of proinflammatory cytokines and other mechanisms, possibly contributed to the development and worsening of multiorgan system dysfunction and failure.

Intersection of Basic and Clinical Research

The intersection of these two lines of research, one basic and one clinical, resulted in two hypotheses of how patients with ALI and ARDS should be ventilated (see Fig. 38-7). First, the end-inspiratory lung volume should be limited to avoid alveolar overdistension (so-called “volutrauma") and second, sufficient PEEP should be applied so as to prevent cycles of end-expiratory derecruitment followed by inspiratory recruitment (Chap. 37 discusses the basis for both of these ventilatory recommendations in detail).

The next step was to test these hypotheses in prospective randomized controlled clinical trials (RCT) (see Fig. 38-7). Four large multicenter RCTs were conducted (Table 38-6). Three RCTs tested lower- vs. higher-tidal-volume strategies.3,8,9 The fourth RCT tested a multifactorial lung-protective strategy (including lower tidal volume, higher PEEP based on static pressure-volume curves of the respiratory system, and recruitment maneuvers) against conventional ventilation.7

The strategies tested were based on interpretations of the static pressure-volume (P-V) curve of the respiratory system in ARDS (Fig. 38-8). The curve in Fig. 38-8 has a lower inflection point (LIP) and an upper inflection point (UIP). One of the RCTs used such curves, which were obtained by use of a super-syringe on paralyzed patients, to set PEEP above the LIP in the group receiving the lung protective strategy,7 but examination of the schematicized curve in Fig. 38-8 can be useful for understanding the strategies for trying to prevent VILI used in all four of the RCTs.

It was hypothesized that the LIP indicated the point at which most of the collapsed or partially fluid-filled alveoli in the lung became recruited.210,211 At pressures higher than the LIP these recruited alveoli exhibited near-normal specific compliance. It was also hypothesized that the UIP represented the point above which the ventilated alveoli in the lung were overdistended, with a resulting low specific compliance of aerated but overdistended alveoli. Based on such P-V curves, it was suggested that PEEP above LIP should prevent the recruitment-derecruitment cycles of alveoli, and as such prevent the lung injury resulting from these cycles (“atelectrauma" [see Chap. 37]). Based on this interpretation, tidal volumes that resulted in end-inspiratory (plateau pressures) below the UIP should decrease alveolar overdistension and thus prevent VILI from this cause.

However, despite the attractiveness of this interpretation of the static pressure-volume curve, subsequent reports indicate that the situation is more complex and that recruitment of alveoli in experimental models and in patients with ALI and ARDS extends beyond the pressure at the LIP and continues over a wide range of airway pressures up to 45 cm H2O.212–214

Nhlbi ARDS Clinical Trials Network Low-Tidal-Volume Ventilatory Strategy Clinical Trial

The NHLBI ARDS Clinical Trials Network conducted the landmark RCT that investigated the use of lung-protective ventilation in patients with ALI and ARDS3 This ARDSNet RCT compared a ventilator strategy with lower tidal volumes designed to limit stretch of the lungs during mechanical ventilation with a strategy that utilized traditional larger tidal volumes. The randomized clinical trial enrolled 861 patients at multiple centers. Briefly, one study arm received a tidal volume of 6 mL/kg PBW if the Pplat did not exceed 30 cm H2O, and tidal volumes of 4 or 5 mL/kg PBW if Pplat did exceed 30 cm H2O (see Table 38-9 for complete details of the protocol of this arm). The other arm received tidal volumes of 12 mL/kg PBW if the Pplat did not exceed 50 cm H2O, and tidal volumes as low as 4 mL/kg PBW if the Pplat did exceed 50 cm H2O. There was a 9% absolute mortality reduction (22% relative mortality reduction) in the group receiving the lower-tidal-volume ventilation strategy (see Table 38-6). This corresponds to a number needed to treat of ∼11 patients in order to prevent one death. Importantly, plasma levels of IL-6 were lower among the 6-mL/kg group, as were the number of organ-failure-free days, indicating that the lower tidal volume strategy was associated with a faster clearance of IL-6, a proinflammatory cytokine, from the plasma, and less organ dysfunction.

This ARDSNet RCT differed from two prior, smaller studies that showed no apparent benefit to the lower-tidal-volume strategies8,9 (see Table 38-6; Table 38-7). These differences may have been due to chance alone since the earlier studies were of limited sample size, with an associated lack of statistical power to detect this degree of difference in mortality. Other possible reasons for these differences include the fact that the ARDSNet RCT used smaller tidal volumes in the low-tidal-volume group than the others (see Table 38-7). In addition, the protocols differed in how they dealt with respiratory acidosis due to lower tidal volumes and permissive hypercapnia. Related to respiratory acidosis, the ARDSNet RCT protocol required increases in ventilator rate as the tidal volume was initially decreased, or if the pH fell below normal limits. In addition, bicarbonate infusions were allowed. In the other two earlier RCTs, similar respiratory rate increases were not mandated. Possibly as a result of these differences in protocols, there were smaller differences in Paco2 and arterial pH between study groups in the ARDSNet RCT compared to the other two RCTs (see Table 38-7).

Although the ARDSNet RCT was subsequently criticized for its design in relying on strict ventilator protocols for the higher-tidal-volume group,215–217 the ARDSnet lower-tidal-volume ventilatory strategy has become accepted as the standard for ventilator management of ALI-ARDS patients. This is based both on the results of the ARDSnet RCT, but also on the plethora of basic and clinical studies, as described above and in Chap. 37, relating to ventilator-induced lung injury that support its hypothetical mode of efficacy.

Using Higher Levels of PEEP to Decrease the Risk of Vili

There is controversy about whether higher-than-traditional levels of PEEP can decrease the risk of ventilator-induced lung injury in patients with ALI and ARDS. In addition, if higher PEEP is effective against VILI, the question of what level of PEEP should be used clinically remains. When the pressure-volume relationship is measured in patients with ARDS, the lower inflection point (LIP) (see Fig. 38-8) is in the range of 8 to 15 cm H2O.7 In an earlier CT study of patients with ARDS,211,218 the amount of reopening-collapsing tissue became insignificant only when PEEP reached 20 cm H2O (although the greatest reduction was seen between 10 and 15 cm H2O of PEEP).

However, as noted above, more recent animal and human studies have shown that the LIP is not the originally hypothesized simple threshold above which no further recruitment occurs. Instead, recruitment continues from below the LIP to inflation pressures of 45 cm H2O.212,213 In other words, there is a broad “inflection zone" from 0 to 45 cm H2O over which there is progressive recruitment with progressively higher PEEP. The implication of these findings is that preventing cycles of recruitment and derecruitment in most of the alveoli necessitates that many open alveoli will be overdistended. Hence, prevention of injury to alveoli from cycles of recruitment-derecruitment will increase the risk of VILI from overdistention (see Figure 38-9).

Clinical Studies of Higher PEEP Ventilator Strategies

Proponents of a high-PEEP strategy (also called “open lung" strategy) were encouraged by a relatively small but statistically significant RCT performed by Amato and colleagues and published in 19987 (see Table 38-6). As noted above, in this RCT the group of subjects with ARDS in the “open lung" arm were treated by a combination of three interventions: (1) a low-tidal-volume ventilatory strategy, (2) higher-than-traditional levels of PEEP as dictated by the patient's LIP as described below, and (3) recruitment maneuvers. The control group of subjects was treated by a conventional ventilatory strategy for the participating ICUs, which did not include any of the three interventions of the open lung group and was not protocolized. In this study, the higher level of PEEP was determined by using a super-syringe to derive a static pressure-volume curve on paralyzed subjects at the start of the study. From inspection of the static pressure-volume curve, the LIP (see Fig. 38-8) was identified and the PEEP was set at 2 cm H2O above the LIP. If a sharp LIP could not be determined on the pressure-volume curve, then the PEEP was set empirically at 16 cm H2O.7

Although this study found that the group receiving the open lung approach had significantly lower mortality than the group receiving conventional ventilation (see Table 38-6), it was unclear which intervention or combination of interventions was responsible for the observed improvement. Was the lower mortality due to the low-tidal-volume ventilatory strategy, the higher-than-traditional levels of PEEP, or the recruitment maneuvers, or a combination of two or all three?

To try to answer these questions, which had also been raised in an earlier report by Amato and colleagues,219 the NHLBI ARDSNet investigators decided to study one intervention at a time in separate RCTs. As noted above, the first ARDSNet RCT compared a lower-tidal-volume ventilatory strategy to a strategy using traditional tidal volumes.3 This RCT found a significant decrease in mortality in the group treated with the lower-tidal-volume ventilatory strategy (see Table 38-6). After completion of this RCT, the ARDSNet investigators conducted a second RCT (“ALVEOLI") to try to answer the question: When used in addition to the lower-tidal-volume strategy of ventilation, do higher PEEP levels improve survival?

The results of the ALVEOLI study4 are presented in Table 38-8. In summary, ALVEOLI found that mortality rates were similar in those treated with higher and traditional levels of PEEP despite significant increases in Pao2: Fio2. The higher-PEEP group had a higher mean plateau pressure despite a lower mean tidal volume (see Table 38-8). Although hypothetical, it is possible that the benefits of the higher PEEP in reducing the acute lung injury due to shear stress created by recruitment-derecruitment were negated by its adverse effects, such as worsening lung injury by overdistension (see Figure 38-9).

Recommended Core Ventilator Management

As the core ventilator management in ALI and ARDS, we recommend that clinicians use the lower-tidal-volume ventilatory strategy (“ARDSnet lung-protective strategy") that the ARDSNet investigators showed to be superior to traditional tidal volume strategy (Table 38-9). Because this strategy that used traditional levels of PEEP was shown to yield similar outcomes compared to using higher PEEP levels4 (see Table 38-8), we also recommend using the same combinations of PEEP and Fio2 that were an integral part of the ventilator protocol for the lower-tidal-volume strategy (see Table 38-9). Clinicians should be cautious in utilizing this protocol strictly for patients with ALI and ARDS who have conditions for which respiratory acidosis due to permissive hypercapnia are contraindicated (Tables 38-10 and 38-11).

In addition to this core ventilator management, some clinicians use various interventions as adjuncts (Table 38-12). Finally, clinicians may choose to use alternative ventilatory strategies whose efficacy is unproven (e.g., salvage therapies for patients in dire clinical circumstances) (Table 38-13).

Adjuncts to Core Ventilator Management

Clinicians may use one or more interventions as adjuncts to “customize" the recommended core ventilator management to try to improve pulmonary physiology and otherwise benefit individual patients (see Table 38-13). Most of these adjuncts hold the possibility of benefit based on extrapolation from animal or clinical research that generally uses physiologic end points as suggestions of efficacy. However, it is inaccurate to extrapolate from an improvement in physiologic outcomes to efficacy in terms of clinically meaningful outcomes (e.g., survival or days of mechanical ventilation). For example, in the first ARDSNet RCT, the lower-tidal-volume group had better survival than the higher-tidal-volume group, despite having lower mean values for Pao2: Fio2.3 Furthermore, the safety of these adjuncts is generally uncertain. Clinicians who want to use high-level evidence to guide care of patients with ALI and ARDS should be warned that to date, all of the following adjuncts fall short of that level of scientific evidence.

Restrictive (“Dry”) Fluid Management

The optimal fluid management in patients with ALI and ARDS remains uncertain. The rationale for restricting fluids in ALI and ARDS suggests that if edemagenesis could be diminished early after the lung injury, the duration of potentially dangerous ventilator, PEEP, and oxygen therapy could be reduced and outcome conceivably improved. In this regard it is interesting that most patients with ARDS do not die during the early phase of disease as a consequence of severe hypoxemia, but rather over days to weeks, frequently with evidence of hypermetabolism, nosocomial infection, and multiple organ system failure.53,54 On the other hand, some have argued that maximizing oxygen delivery (so-called “supercharging") to peripheral tissues is a necessary strategy in critically ill patients such as those with sepsis and ARDS, and advocate approaches such as volume loading to achieve this goal.220,221 This latter position has fallen out of favor because the oxygen extraction defect purported to exist in patients with sepsis and ARDS now seems to have been artifactual, or at least not clinically relevant.222,223 Moreover, controlled clinical trials of such goal-oriented hemodynamic therapy showed no improvement in survival224 or worse survival.225 In patients with ALI and ARDS, the consequence of an approach of maximizing oxygen delivery could be escalating requirements for mechanical ventilation, oxygen, and PEEP. To the extent that intensity and duration of supportive therapy are major determinants of complications and organ dysfunctions, the net result could be detrimental. Fluid management strategies in ALI and ARDS are currently the focus of an ongoing ARDSNet investigation (the Fluids and Catheter Treatment Trial, or FACTT). FACTT also is comparing use of central venous catheters to pulmonary artery catheters to guide fluid management. Until the results of this study are known, we favor an approach that aims for a cautious reduction in central venous pressure (CVP) or pulmonary artery wedge pressure (Ppw) while maintaining adequate circulation early in the course of ARDS as a strategy to minimize edemagenesis and to avoid adverse effects of its treatment. Although this approach is controversial, its justification derives from both animal experiments226–233 and clinical data (see below).

Several retrospective or noninterventional studies have reported data showing a correlation between survival and net diuresis or reduction in Ppw.234–236 Prospective data collection also has demonstrated that titration of therapy to minimize extravascular lung water in patients with ARDS results in decreased ventilator and ICU days.237 Until a prospective, randomized trial of a strategy of conservative fluid management in ALI and ARDS patients provides evidence regarding efficacy and safety, the circulatory management of these patients can be described as a therapeutic dilemma.238,239 On the one hand, judicious volume reduction may improve lung function. In the extreme, however, hypoperfusion and other organ failure could occur.

In order to reduce Ppw safely in patients with ALI and ARDS, the clinician's attention should be focused on the parameters of organ function that should be followed in all critically ill patients (e.g., mental status, urine output and concentration, circulatory adequacy, and metabolic evidence of anaerobic metabolism). This should not be misinterpreted as trading reduced pulmonary edema for a hypoperfused state. Rather, the aim of this approach is to find the lowest Ppw or CVP compatible with adequate functioning of the circulatory system. Preload can be reduced in part by meticulous attention to limit all extraneous fluid administration. Furosemide is often used to achieve diuresis and net negative fluid balance. However, this may confound interpretation of renal function as an indicator of the adequacy of perfusion. In selected patients, particularly those with manifest systolic dysfunction, dobutamine can be used to maintain perfusion at a reduced Ppw.

Interpreting right heart catheterization data during mechanical ventilation with PEEP can be challenging.240 The usual measurement made at right heart catheterization is the Ppw. Measurements of Ppw may fail to reflect left ventricular filling pressure for a number of reasons. First, if alveolar pressure is greater than left atrial pressure at the site of measurement (West zone 1 or 2), then the downstream pressure is not measured, and the right heart catheter reflects alveolar pressure. Existence of zone 1 or 2 conditions can be inferred from the excursion of the pulmonary arterial and wedge pressure waveforms during positive pressure ventilation.241  This possibility must be considered if one attempts preload reduction in patients with ARDS requiring high PEEP levels.

Second, it is often difficult to identify end-inspiration in many patients with ALI or ARDS. This is especially problematic if patients are spontaneously breathing while receiving assisted ventilation. To improve accuracy in making routine measurements of Ppw as well as CVP, it is recommended that the airway pressure be transduced and its transducer's output be printed simultaneously with the pressure tracing. The transduced airway pressure tracing clearly identifies the start of inspiration by the start of its positive deflection in patients who are not assisting the ventilator. Likewise the airway pressure tracing also easily identifies the start of an assisted breath by the occurrence of the associated negative pressure deflection. This simple change can significantly decrease interobserver variability.242

Some have suggested that the nadir in Ppw after an airway disconnect might better reflect left ventricular filling pressure.243,244 This approach may be dangerous and is not recommended since even relatively short discontinuation of PEEP can have adverse effects on oxygenation that may take hours to recover.

Permissive Hypercapnia

Traditionally physicians have attempted to ventilate patients to a normal arterial Pco2 (see Fig. 38-6). In patients with severe lung disease, however, this arbitrary goal has a mechanical cost: the probable amplification of lung injury (i.e., VILI). Increasing evidence points to the safety and efficacy of allowing the arterial Pco2 to rise above 40 mm Hg when used in combination with a ventilatory strategy that uses low tidal volumes and low plateau pressures. When patients with ALI and ARDS are ventilated with volume- and pressure-limited ventilation as described above, the mean Paco2 may rise modestly (into the low 40s mm Hg) or higher (into the mid-to-high 50s mm Hg) with corresponding falls in arterial pH (see Table 38-7). Occasional patients may have a Paco2 above 100 mm Hg245,246 (see Table 38-7).

Respiratory acidosis has many physiologic effects, including cellular metabolic dysfunction, depression of myocardial contractility, coronary vasodilation, systemic vasodilation, pulmonary vasoconstriction, enhanced hypoxic pulmonary vasoconstriction, cerebral vasodilation, increased intracranial pressure, and renal vasoconstriction, among others247,248 (see Table 38-10). Yet even very high levels of Pco2 seem remarkably well tolerated by adequately sedated patients. Perhaps this is related to highly efficient and rapidly acting cellular compensatory mechanisms that tend to defend intracellular pH. Because respiratory acidosis raises intracranial pressure, permissive hypercapnia should not be used in patients with cerebral edema, trauma, or space-occupying lesions. This and other contraindications are listed in Table 38-11.

Prone Positioning

Multiple studies have shown that about two-thirds of patients with ARDS exhibit improved oxygenation with prone positioning (“proned").249–252 Hypotheses offered to explain the improvement in oxygenation include (1) increased FRC, (2) change in regional diaphragm motion, (3) redistribution of perfusion, and (4) better clearance of secretions.253 FRC has been shown to be increased in the prone position in intubated, mechanically-ventilated patients without lung injury who are undergoing general anesthesia for surgery.254 Animal models of ventilation-perfusion distribution have suggested that gravity has far less influence on the distribution of perfusion in the prone position, and that the distribution of blood flow to regions of the lung is relatively unaffected by the change from the supine to the prone position.255 This, coupled with the observation that turning to the prone position is associated with a migration of the edema fluid to the dependent portions of the lung (as demonstrated by CT scan), has suggested to some investigators that ventilation-perfusion relationships might be favorably altered by the prone position.256 In patients managed in the prone position, special attention is necessary to prevent pressure necrosis of the nose, face, and ears, and to ensure maintenance and patency of the endotracheal tube. Pressure on the eye could lead to retinal ischemia, especially in hypotensive patients. Some patients experience cardiac arrhythmias or hemodynamic instability on being turned.

These considerations led to a large clinical trial performed by Gattinoni and colleagues.252 In this study, subjects were placed in the prone position for six or more hours daily for 10 days. The results, published in 2001, revealed that although oxygenation was transiently improved, prone positioning offered no survival advantage over routine supine positioning.252 Further post-hoc analyses indicated that a patient's response to prone positioning may have prognostic value. Patients whose Paco2  fell by 1 mm Hg or more when placed in the prone position had a lower mortality rate than those whose Paco2 didn't fall or rose (mortality of 35.1% vs. mortality of 52.2%).257

Prone positioning may have a role in the future, when combined with lower tidal volume ventilation, such as the ARDSNet protocol (see Table 38-9), ventilator sighs,258 or when employed using a different treatment protocol. Clinicians may reserve use of prone positioning as salvage therapy for severe hypoxemia, but they should understand that there are no controlled studies that support its efficacy in this circumstance, even in those patients whose oxygenation improves (see below). Finally, although some advocate the routine use of prone positioning as part of a multifactorial lung protective strategy,259 no controlled clinical trials support such routine use, especially given its potential adverse effects listed above.

Recruitment Maneuvers

Recruitment maneuvers evolved from traditional “sighs," which are extra-large breaths of the order of two or three normal-sized tidal volume breaths. Sighs normally occur 4 to 10 times per hour and increase the surfactant's surface-tension-lowering properties, thus stabilizing small alveoli and resisting atelectasis.

Recruitment maneuvers were part of the “open lung" strategy in the clinical trial of low tidal volume ventilation by Amato et al7 (see Tables 38-6 and 38-7). In the Amato trial, recruitment maneuvers consisted of application of continuous positive airway pressure (CPAP) of 35 to 40 cm H2O for 30-second periods. Others have advocated longer periods at the same or higher airway pressures.259 The justification for recruitment maneuvers is to “recruit" or open totally or partially collapsed alveoli, which then would be kept inflated during expiration by a high level of PEEP.259

Evidence is lacking that recruitment maneuvers alone improve clinically significant outcomes such as mortality or ventilator-free days. Most studies of recruitment maneuvers have used physiologic end points, such as improvement in oxygenation. The ARDSNet studied recruitment maneuvers as a substudy of 96 subjects in the higher-PEEP group in the ALVEOLI study260 (see Table 38-8). There were no clinically relevant improvements in arterial saturation, but complications occurred, such as transient hypotension and slight drops in arterial saturation during the recruitment maneuver. Other studies have shown more consistent improvement in oxygenation after recruitment maneuvers if relatively low levels of PEEP were being used,261–263 if larger tidal volumes were used,264 or if the patients are paralyzed.265

Given the lack of controlled clinical trials that demonstrate efficacy in clinically relevant end points and the potential adverse effects, we don't recommend routine use of recruitment maneuvers. Furthermore, because they exceed the threshold of 30 cm H2O used in the ARDSnet clinical trial that showed improved survival, and the lack of studies that demonstrate improved outcomes,266–268 we don't recommend routine use of “sighs" either.

Salvage Interventions

When treating subjects with severe ARDS, some clinicians may try unproven interventions if the patient is deteriorating with severe hypoxemia (e.g., Pao2 less than 45 to 50 mm Hg) or needing an Fio2 of 0.9 or more to maintain Pao2 above 55 mm Hg. These may be referred to as “salvage" interventions. These clinicians justify their use of these interventions on two grounds: (1) the dire condition of the patient, and (2) a hope of clinical efficacy. The latter is based on results from basic science studies suggesting a reasonable rationale, from their use in animal models, and from clinical usage that showed improvements in certain physiologic parameters (e.g., Pao2: Fio2).

Despite the failure to improve survival in phase III clinical trials of patients with ALI/ARDS who were not necessarily in such dire straits, these clinicians may feel ethically obligated through the “Rule of Rescue269" to provide an intervention that may help as long as the risk is acceptable. Because of expense, lack of proven efficacy, and potential for harm, we do not advocate routine use of any of these “salvage interventions." Rather, we support a management strategy guided by evidence and including active observation of critically ill patients with ARDS. We urge those who advocate for their use to conduct clinical trials in the targeted population of patients with severe ARDS to assess their safety and efficacy.

Nonetheless, when there are severe problems with oxygenation in an otherwise salvageable patient, some clinicians will want to utilize certain adjunctive therapies. It is important to realize that these therapies should not distract caregivers from the fundamentals of good critical care, including nutrition, aspiration precautions, hygiene and prevention of nosocomial infections, appropriate sedation practices, and careful vigilance for complications of critical care.

Tracheal Gas Insufflation

Tracheal gas insufflation (TGI), involves introducing fresh gas near the carina through a modified endotracheal tube. This added flow washes CO2-rich gas out of the trachea (and, through turbulence, out of smaller airways as well), reducing anatomic dead space.270 The Paco2-reducing effect of TGI is lessened by acute lung injury, but this is partially counterbalanced by the higher Paco2 values used during permissive hypercapnia.271 In patients with ARDS, TGI with 100% humidified oxygen, delivered throughout the respiratory cycle at a flow of 4 L/min, successfully lowered Pco2 from 108 to 84 mm Hg.272 Potential risks of TGI include tracheal erosion, oxygen toxicity related to the unknown Fio2, hemodynamic compromise or barotrauma due to the occult presence of auto-PEEP, and a larger tidal volume than the ventilator is set to deliver (i.e., potentially increasing the risk of VILI).

Inhaled Nitric Oxide and Inhaled Prostacyclin (Epoprostenol)

Since Roissant and colleagues published their initial experience using inhaled nitric oxide as a therapy for ARDS, there has been a rapid expansion of interest and literature in this field.273–278 Given via inhalation, NO has several potentially salutary effects in ARDS. It selectively vasodilates pulmonary capillaries and arterioles that subserve ventilated alveoli, diverting blood flow to these alveoli (and away from areas of shunting). The vasodilating effect, signaled by a fall in pulmonary artery pressure and pulmonary vascular resistance, appears maximal at very low concentrations (0.1 ppm) in patients with ARDS.277 The beneficial effects on oxygenation take place at somewhat higher inspired concentrations of NO (1 to 10 ppm).277 The rapid inactivation of NO via hemoglobin binding prevents unwanted systemic hemodynamic side effects, but also mandates the continuous delivery of gas to the ventilator circuit. Thus, if continuous delivery of NO is interrupted (e.g., during patient transport or due to supply exhaustion), precipitous and life-threatening hypoxemia and right-sided heart failure may occur.279 A recent large clinical trial compared the use of NO to placebo in subjects with ALI not due to sepsis, with no other organ failures.280 The trial showed no benefit in survival with use of NO despite some patients having a transient improvement in oxygenation. Trials in other subgroups of ALI may be the focus of future investigations, but there is no consensus evidence for routine use of NO in ALI.

Even as a salvage intervention, inhaled NO is unattractive since there is a reasonable alternative available at much less expense. Whereas the cost of inhaled NO for 1 day is in the thousands of dollars, the daily cost of inhaled prostacyclin (epoprostenol) is in the hundreds of dollars. Inhaled prostacyclin, although less well studied, appears to provide the same degree of improvement in oxygenation in a majority of patients with ALI and ARDS at much less expense.281–284

Corticosteroids

Although it is commonly accepted that steroids have little or no role to play in treating the early acute phase of ARDS,285,286 their role in later phases remains controversial. A number of anecdotal reports and small series have suggested that high-dose corticosteroids may be of some benefit during the proliferative phase of ARDS.287–289 The rationale behind this therapy is that much of the scarring that occurs during this phase of the illness is a consequence of unattentuated inflammation that can cause severe damage to the affected alveoli.290 There is an obvious risk, however, of immunosuppressing already debilitated patients who are still in an environment in which they are exposed to multiple resistant organisms (and frequently have multiple indwelling appliances), as well as a potential risk for long-term neuromuscular sequelae. The NHLBI ARDSNet recently completed a double-blind RCT (Late Steroid Rescue Study or LaSRS) designed to evaluate the benefits and risks of this therapy in 180 patients with ARDS lasting 7 to 21 days. The ARDSNet presented preliminary results at the 2004 International Meeting of the American Thoracic Society. These indicated that there was no difference in 60-day mortality, or ICU or hospital length of stay between those treated with high-dose methylprednisolone and those treated with placebo.291 In addition, investigators reported more adverse events related to weakness in the steroid-treated group than in the placebo-treated group. Pending publication of the final results of this study, the routine use of steroids, specifically for late-phase ARDS, cannot be advocated on firm scientific grounds.

Pressure-Control Ventilation and Inverse-Ratio Ventilation

Pressure-control ventilation (PCV) is favored by some clinicians because it limits the maximal peak airway pressure. It also limits static end-inspiratory or alveolar pressure. However, some intensivists and respiratory care providers may not appreciate what that limit is. For example, if a patient with ARDS is being ventilated with PCV with an inspiratory pressure of 30 cm H2O and PEEP of 10 cm H2O, then the total end-inspiratory pressure is the sum of 30 cm H2O plus 10 cm H2O, or 40 cm H2O (Fig. 38-10). Some clinicians may mistakenly believe that the alveoli are being exposed to only 30 cm H2O, and thus do not decrease the inspiratory pressure such that the end-inspiratory pressure does not exceed 30 cm H2O, the threshold used in the pivotal ARDSNet study.3

Although one could possibly mimic the tidal volumes and end-inspiratory pressures that were used in the ARDSnet low-tidal-volume ventilatory strategy (see Table 38-9), it would be challenging. It is possible that the low tidal volume per se in the ARDSnet trial was a factor in decreasing the mortality, apart from the benefits of keeping the plateau pressures limited to 30 cm H2O or less.

Inverse ratio ventilation (IRV) entails the use of prolonged inspiratory times (inspiratory:expiratory ratio >1) with either a volume-cycled or pressure-cycled mode of mechanical ventilation (see Fig. 38-10). A subset of patients with hypoxia refractory to conventional modes of mechanical ventilation responded to IRV.292,293 Unfortunately, there is no way to prospectively identify which patients will respond to IRV. The exact mechanisms by which IRV improves gas exchange in some patients with ARDS remain obscure, but are believed to involve both alveolar recruitment at lower airway pressures and more optimal distribution of ventilation.294,295 Although it is tempting to attribute the beneficial effect of IRV to intrinsic PEEP, anecdotal reports have excluded intrinsic PEEP or gas trapping as the mechanism by which gas exchange improves in at least some patients.295 An important caution when using this mode is that both auto-PEEP and the higher mean alveolar pressure typical of IRV tend to reduce cardiac output. In one study that examined PCV with or without IRV, cardiac output fell with IRV so that systemic oxygen delivery actually worsened.296 Some investigators have noted a very gradual (over several hours) but progressive tendency for oxygenation to improve following a change to IRV.297 This phenomenon has led some to suggest that a subset of lung units may be recruitable only through the combined effects of prolonged inspiration and time. Further studies are needed to shed light on this interesting aspect of IRV.

Inverse ratio pressure-control ventilation has been employed as part of the open lung strategy for ventilating patients with ARDS, with specific attention to keeping the tidal volume at or less than 6 mL/kg and limiting the driving (inspiratory) pressure.7 The role of volume-control IRV remains poorly defined, and it is best regarded as a salvage therapy for patients with hypoxia refractory to more conventional approaches. Whichever approach is used, the intensivist should monitor auto-PEEP regularly, since the shortened expiratory times of IRV predispose to this effect.

Clinicians have also used other modalities of ventilation that depend on computerized ventilators, none of which have been compared to current low-tidal-volume strategies in RCTs (see Chap. 32).

High-Frequency Oscillatory Ventilation

If excessive lung excursion during tidal volume breathing is associated with injury to the lung, then it seems reasonable that ventilation with very small tidal volumes at high frequencies would be associated with the least possible ventilator-induced lung injury and potentially with improved outcome. Although the FDA has approved a ventilator for adults that provides high-frequency oscillatory ventilation (HFOV), its role in clinical practice remains unclear.298 A study of HFOV299 published in 2002 demonstrated a trend towards decreased mortality compared to conventional mechanical ventilation. However, the conventional ventilation was not based on a low-tidal-volume strategy such as ARDSNet (see Table 38-9). Because the mean airway pressure (± SD) was higher in the group of patients treated with HFOV than the conventionally-treated group (e.g., 29 ± 6 cm H2O vs. 23 ± 6 cm H2O during the first 24 hours), caution must be raised about the possibility of VILI due to the high distending pressures. Until a RCT is done comparing HFOV to the ARDSNet protocol, clinicians will not have evidence to guide them about when or if to employ HFOV.

High-frequency jet ventilation is different from HFOV. High-frequency jet ventilation typically employs tidal volumes of 1 to 5 mL/kg (or higher) and respiratory rates of 60 to 300 breaths per minute. Multiple trials of high-frequency ventilation have failed to demonstrate any benefit compared with conventional mechanical ventilation.300–302 Nor has high-frequency jet ventilation has been associated with either improved oxygenation, reduced barotrauma, or decreased days of mechanical ventilation. On the basis of these negative results and the potential promise of HFOV, we recommend avoiding the use of high-frequency jet ventilation, even as a salvage intervention.

Extracorporeal Membrane Oxygenation and Extracorporeal Co2 Removal

The use of extracorporeal gas exchange, such as extracorporeal membrane oxygenation (ECMO) or extracorporeal CO2 removal (ECCO2R), to adequately oxygenate and ventilate the patient while allowing the lung to remain at rest was viewed as an attractive strategy for the management of patients with acute lung injury. However, this promise has not been supported by clinical outcome studies. The earliest large-scale attempt to use ECMO in patients with severe ARDS in the 1970s demonstrated no survival benefit to its use, although it did generate a large database and a great deal of insight into patients with this problem.303 Unfortunately, enrollment criteria in this study were such that the mortality among all patients entered into it was certain to be high (e.g., ∼90% in both groups). Hence it was unlikely that any difference would be demonstrated between groups. Some believe that more careful patient selection, earlier randomization of patients, and better technology might have demonstrated a benefit to ECMO. A second wave of studies using ECMO or ECCO2R was reported throughout the 1980s.304–306 A number of techniques have been described, including venovenous ECMO, to assist in the elimination of carbon dioxide. However, the most recent report of ECMO in adult patients with ARDS also failed to demonstrate any survival benefit.307 Despite these results, some specialized centers have continued to offer ECMO to adults with severe ARDS, based on their opinion that it is a relatively safe life-saving salvage intervention.308

Experimental (Non-FDA Approved) Interventions

Partial Liquid Ventilation

Partial liquid ventilation using perfluorocarbons instilled into the trachea of adults and children with the respiratory distress syndrome has been described.309–312 Preliminary results from adult usage313 and the more extensive experience in pediatric patients suggest that this mode of therapy may be both safe and efficacious in improving gas exchange. Partial liquid ventilation may allow oxygenation in patients who might otherwise be quite difficult to oxygenate with conventional modes of ventilation, in part because the perfluorocarbon is able to recruit dependent alveoli (by virtue of the hydraulic column) that PEEP is not. A practical problem is that perflubron is radiodense, making the lungs appear white, so it is impossible to use chest radiographs to detect infection or to follow the progress of healing. Currently, perfluorocarbons are available only as experimental agents.

Exogenous Surfactant

It has long been known from both animal models and human studies that surfactant levels are decreased or that the ratios of the surfactants are abnormal in humans and animals with ARDS.314,315 Intensivists caring for adults have been encouraged by the dramatic results of surfactant therapy in infants with the respiratory distress syndrome (RDS) of prematurity. Surfactant therapy of RDS improves gas exchange and lung mechanics, decreases the requirement for CPAP, and lessens barotrauma.316–318 Anzueto and associates6 reported the first large prospective RCT of surfactant in ARDS. Their results were disappointing: There was no benefit associated with the exogenous surfactant delivered by inhalation. Because there were concerns about the appropriate dose, alternative modes of delivery, timing of therapy, and the precise surfactant formulation studied, investigators did not view this study as definitive evidence against the use of exogenous surfactant. Since then a number of RCTs, large and small, have been carried out and more are underway. Results so far have been mixed.319–323 Large phase III trials with different surfactant preparations are underway. Like perflubron, currently exogenous surfactant for adults is available only as an experimental agent.

Supportive Care and Monitoring Patients with ALI and ARDS

Supportive Therapy

Current management of ARDS does not benefit from proven pharmacologic interventions to limit acute lung injury or restore physiologic function. The largest strides in the recent management of ARDS have come from therapies aimed at the delivery of mechanical ventilation.3 While it is possible that further explication of the mechanisms of lung injury will provide new avenues for pharmacologic intervention, at present, management of these patients relies on application of proven mechanical ventilation strategies, combined with meticulous supportive therapy. Appropriate management includes timely diagnosis and treatment of underlying diseases, nosocomial infections, and other problems. Indeed, even if new pharmacologic agents become available, the same supportive therapy will be necessary to maintain a viable patient to benefit from treatment. For a detailed description of elements of supportive care, please refer to the relevant other chapters in this text.

Reducing PEEP, even for short periods of time, is often associated with alveolar derecruitment and hence rapid arterial hemoglobin desaturation. Thus once endotracheal tube suctioning has been accomplished for diagnostic purposes, nursing and respiratory therapy staff should be instructed to keep airway disconnections to a minimum or to use an in-line suctioning system that maintains sterility and positive pressure, usually via the suctioning catheter residing in a sterile sheath and entering the endotracheal tube via a tight-sealing diaphragm. These suctioning systems generally are effective for lesser levels of PEEP (<15 cm H2O) but often leak if higher levels are attempted.

Monitoring

Patients with ARDS, by virtue of their serious gas exchange (and sometimes hemodynamic) impairment, combined with the effects of therapy (sedation, therapeutic paralysis, and PEEP), are at risk of sudden and life-threatening deterioration. Changes in intrapulmonary shunt, oxygen consumption, and systemic perfusion are frequent, making arterial saturation and oxygen delivery volatile. Accordingly, careful monitoring for hypoxemia and the adequacy of oxygen delivery is advisable. Continuous pulse oximetry is generally reliable (barring gross hypoperfusion) and should be used routinely. In addition, frequent sampling of arterial blood gases is advisable throughout the first day of management, as well as following major interventions or changes in clinical appearance of the patient.

Monitoring the patient's airway, ventilator function, and the ventilator-patient interface are equally important, as is assessment for liberation from assisted ventilation, and if needed, weaning (see Chap. 44).

Finally, hemodynamic monitoring, including use of a pulmonary artery catheter, has been discussed earlier in this chapter (see Chap. 13).

Long-Term Sequelae of ARDS

Over the past two decades, as treatment for ARDS has decreased hospital mortality,324 clinicians and clinical investigators have become more interested in the long-term health problems of ARDS survivors. Pulmonary function is usually mildly impaired after hospital discharge from ARDS and improves slightly over the next year.325,326 Thus survivors with worsening dyspnea may have another superimposed respiratory lesion, such as tracheal stenosis, and should be evaluated as such. Despite their young age, ARDS survivors score well below the reference standards and other critical care controls on quality-of-life measures,27,326−329 and many have evidence of cognitive dysfunction,330 posttraumatic stress disorder,328 and physical debilitation,25 many months after hospital discharge. The long-term sequelae of critical illness and ARDS are an active area of ongoing research, covered in detail in Chap. 19.

Standardization of the criteria that define ALI and ARDS has aided in identification of specific at-risk groups. This in turn has spurred further research into the underlying reasons why certain risk groups (such as alcoholics331,332) are at greater risk for ALI. A recent NHLBI consensus statement summarized the important directions for future research, including functional response to injury and interaction between biochemical pathways and different cell types.23 The completion of the human genome project has led to characterization of many of the genes encoding mediators of lung injury.333 The effect of variation in these genes on predisposition to ALI in at-risk groups, such as sepsis, pneumonia, and trauma, may help to identify subgroups whose genotypes place them at unusually high risk or low risk for developing ALI and ARDS. Identifying these putative enabling and protective polymorphisms for developing ALI will provide hypotheses for interventions for prevention and for treatment of patients with ALI and ARDS in the future.

However, while waiting for those new genetically tailored therapies, much can be done in the present. The landmark ARDSNet low-tidal-volume ventilation strategy trial proved that ventilator therapy can be protocolized to reduce ventilator-induced lung injury.3 Arguably this simple and inexpensive strategy can save thousands of lives of ALI and ARDS patients if widely accepted and utilized. Unfortunately, studies since the publication of the ARDSnet study in 2000 have indicated that that there are challenges to the widespread and timely acceptance and implementation of this low-tidal-volume strategy.334 − 336 Thus the present challenges include not only improving on this therapy, but also overcoming the obstacles so that clinicians can consistently make a diagnosis of ALI and ARDS early, and then begin appropriate ventilatory support.