Chapter 37: Acute Respiratory Distress Syndrome

ARDS is an acute process of noncardiogenic pulmonary edema due to lung inflammation leading to hypoxemia. The American-European Consensus Conference (AECC)¹ defines ARDS as (1) acute onset (< 7 days); (2) partial pressure of oxygen in arterial blood (Pao₂)/fraction of inspired oxygen (Fio₂) less than or equal to 200 mm Hg (regardless of positive end-expiratory pressure [PEEP] level); (3) diffuse bilateral infiltrates on frontal chest radiograph, consistent with pulmonary edema; and (4) pulmonary capillary arterial occlusion pressure (PAOP) less than or equal to 18 mm Hg if measured, or in the absence of capability to measure PAOP, no clinical evidence of left atrial hypertension.

Acute lung injury (ALI) carries the same definition except a lesser degree of hypoxemia (Pao₂/Fio₂ ≤ 300 mm Hg for ALI). The definitions of ARDS and ALI do not require mechanical ventilation and exclude chronic lung diseases.

Etiologies of ARDS are divided into primary (initial insult comes from alveolar side of lung interstitium) and secondary or extrapulmonary (insult comes from blood side of lung interstitium) (Table 37-1).² The diagnosis of ARDS does not require a pulmonary artery (PA) catheter. Echocardiography, however, is often relied on in diagnosing systolic or diastolic dysfunction as the cause of pulmonary edema and in ruling out ALI/ARDS. Patients presenting with cardiogenic edema may have normal PAOP by the time a pulmonary artery catheter is placed (therapy had already been initiated).

Acute lung injury is quite common in the United States, perhaps affecting more than 200,000 people per year, with an associated mortality rate of 30% to 35% and a burden of subsequent morbidity and disability on survivors.³

The clinical presentation follows an inflammatory response involving neutrophils, macrophages/monocytes, and lymphocytes undergoing adhesion, chemotaxis, and activation, as well as non- cellular events, such as complement activation, coagulopathy, and kinin elaboration. Cytokines, lipid mediators, oxidants, proteases, nitric oxide, growth factors, neuropeptides, and protein synthesis may be involved in lung injury.

Clinical Manifestations

Gas exchange derangements

Following endothelium or epithelium injury, plasma moves into the lung interstitium and alveolar space, creating shunt and reducing lung compliance. The hallmark of ARDS is shunting, where inspired gas never reaches flooded alveoli and a substantial fraction of mixed venous blood, at low oxyhemoglobin saturation, traverses the pulmonary circulation without being exposed to oxygen, causing arterial hypoxemia that does not respond to increasing the Fio₂. Dead space ventilation (decreased perfusion relative to ventilation) is also raised in ARDS. In addition to the adverse consequences on gas exchange, diseases characterized by interstitial and alveolar fluid accumulation reduce the compliance of the respiratory system, imposing a mechanical load on the patient with a resulting increase in the work of breathing.

Radiographic findings

The plain chest radiograph is useful in distinguishing ARDS from focal lung processes such as pneumonia or atelectasis. It is less helpful in separating ARDS from hydrostatic (high pulmonary capillary pressure) edema (usually, but not always, caused by heart disease).

Imaging patients with ARDS with chest computed tomography (CT) scans, although not frequently done clinically, often reveals the lungs to be composed of three compartments: one is relatively normal; another one has mild consolidation or atelectasis; and the third, in the more dependent area, has dense consolidation. The striking nonhomogeneity of lung flooding and consolidation on chest CT scans in patients with ARDS have important implications; the relatively normal part of the lung should be considered small (“baby” lung) rather than stiff in terms of the alveolar population receiving machine-delivered tidal volumes.⁴ This has implications for mechanical ventilation as discussed below since the remaining aerated alveoli are subject to overdistention when excessive tidal volumes are used.

The cornerstone of current management of ARDS consists of supportive therapy, including lung protective ventilation, and cardiovascular manipulation as needed. Supportive management of ARDS should be joined by aggressive diagnosis and treatment of the predisposing condition since the underlying problem (eg, sepsis) typically requires specific therapy.

Mechanical Ventilation

The majority of patients with ARDS will require intubation and mechanical ventilation to decrease the work of breathing and to apply PEEP. Some causes of ARDS, such as tocolytic-associated pulmonary edema, postictal pulmonary edema, and pulmonary edema associated with air emboli, tend to have a benign course and may respond simply to supportive therapy. Also, some patients will have rather mild ALI and may be successfully ventilated noninvasively (alert patients with near-drowning).^{5, 6}

Mechanical ventilation has been demonstrated to be injurious to the lung by both (1) overdistention of alveoli and (2) repeated recruitment (opening of closed lung during inspiration) and decruitment (closure of that same lung at end expiration), which amplifies lung injury.^{7, 8}

Lung Overdistention

Volutrauma is defined as overdistention of alveoli (rather than pressure, per se) and causes histopathologic, radiologic, and clinical findings of acute lung injury. It is determined by the transpulmonary pressure in the alveoli of concern. Computed tomographic studies show that in most patients, the ARDS lung is remarkably inhomogeneous, with some normally compliant and likely fragile alveoli (nondependent) and others that are flooded or collapsed (dependent). The loss of functional alveoli necessitates that, if the tidal volume is not reduced, alveoli that are open will be overdistended. Indeed, the apparently “stiff” lung of ARDS is better understood as gross overdistention of a few relatively normal alveoli rather than as caused by generalized parenchymal “stiffness.” The mechanisms by which alveolar overstretching produces damage remain controversial, but inflammatory cytokines are found in lavage fluid and systemic blood following ventilation with excessive tidal volumes.⁹

The lung pressure-volume relationship in ARDS has a sigmoidal shape, with a lower inflection point (LIP) and an upper inflection zone (UIZ), and exhibits hysteresis (better compliance on the descending limb than the ascending limb) when the inflation and deflation limbs are compared (Figure 37-1). The UIZ signals progressively falling compliance, indicating that many alveoli have reached their normal limits of expansion and overdistention has become the predominant influence on compliance. It has been proposed that when the inflation curve reaches the UIZ, that lung is at risk.

The ARDS Network (ARDSNet) tested the relevance of overdistention in a clinical study using a simple method: limiting tidal volume to 6 mL/kg ideal body weight (IBW).¹⁰ This important study entered 861 patients with both ALI and ARDS and randomized them to 6 versus 12 mL/kg IBW. The primary outcome measure, hospital mortality, markedly favored the lower tidal volume approach (39.8% versus 31.0%). A consequence of limiting the tidal volume in this manner is that the partial pressure of carbon dioxide in arterial blood (Paco₂) may rise above 40 mm Hg. Tolerating this respiratory acidosis is called permissive hypercapnia and is preferable to alveolar overdistention. Other ARDSNet ventilator strategies include reducing the tidal volume further, if needed, to keep the plateau pressure (P_plat) less than 30 cm H₂O; protocolized PEEP and Fio₂ settings to maintain the arterial saturation between 88% and 95%; and rate increases to as high as 35 breaths/min.

The ARDSNet protocol including low tidal volume ventilation has become the benchmark against which alternative approaches are compared.¹⁰ The existing data support a dose-response relationship in which the lower the tidal volume, the lower the mortality rate.¹¹ However, very low tidal volumes are poorly tolerated in most patients and require significant sedation, which is problematic.

Recruitment/Derecruitment and the Lower Inflection Point

The LIP of the pressure-volume curve (See Figure 37-1) is the point where lower applied pressures will result in significant lung collapse. Many studies suggest that an appropriate level of PEEP (minimal PEEP) applied at the LIP can protect the injured ARDS lung by preventing repeated deflation-inflation across the LIP and the associated shear force injury. Locating the LIP requires constructing a compliance curve by measuring compliance at various PEEP settings following recruitment maneuvers.⁷ The ARDS Network has tested a higher PEEP approach (superimposed on low tidal volume ventilation) and found no meaningful benefit or harm in this approach. It remains possible, however, that increasing PEEP only in the presence of improving compliance would be an appropriate, yet unconfirmed strategy.

Permissive Hypercapnia

Historically, physicians first attempted to ventilate patients with ARDS to a normal Paco₂. In patients with severe acute lung injury, however, this arbitrary goal has a mechanical cost because of inappropriately sized tidal volume and amplification of lung injury, as described above. Over the past decade, there has been increasing evidence that points to the safety and efficacy of allowing the Paco₂ to rise well above 40 mm Hg.¹² When patients with severe ARDS are ventilated with a lung protective strategy, Paco₂ may rise to 50 to 70 mm Hg (occasionally higher) and the pH falls accordingly. Despite the adverse effects of respiratory acidosis in the nonmechanically ventilated patient, very high levels of Paco₂ seem remarkably well tolerated by adequately sedated patients. One may allow the Paco₂ to rise gradually (5-10 mm Hg/h) in order to allow time for some cellular compensation. In patients without contraindications, sodium bicarbonate may be infused to raise the blood pH. Sodium bicarbonate delivers a substantial Co₂ load to a patient already marginally able to excrete what is produced endogenously and further risks volume overload and potassium depletion. Bicarbonate should be considered when ventilator strategy results in very low pH (ie, < 7.20).

Contraindications to permissive hypercapnia are the presence of increased intracranial pressure or active cardiac ischemia; also, the inspired gas must be oxygen enriched to prevent hypoxemia, and patients would be expected to require more sedation. Over time, in patients with good renal function, the kidney will retain bicarbonate in response to the respiratory acidosis as an endogenous compensation mechanism, making permissive hypercapnia easier to maintain.

Approach to Mechanical Ventilation

The principles that guide our ventilatory management of ARDS are (1) avoid alveolar overdistention targeting an initial tidal volume of 6 mL/kg IBW, tolerating hypercapnia if necessary¹⁰; (2) obtain and maintain an Fio₂ of less than or equal to 0.6 as soon as feasible; and (3) apply PEEP to avoid deflation/reinflation injury. Either volume- or time-cycled assist-control ventilation (ACV) is used. Since with pressure-control ventilation changes in lung elastance or airway resistance over time cause tidal volume to vary, it is typically deployed in a pressure-regulated volume-control (PRVC) or volume-guaranteed mode. In these circumstances the ventilator software adjusts pressure up or down in an attempt to keep tidal volume constant. PRVC ventilation in the distressed ARDS patient may require significant sedation, which, if necessary, in addition to aiding patient-ventilator synchrony, reduces the oxygen consumption and Co₂ production.

One of the management goals is to reduce the Fio₂ over the first 24 to 48 hours to less than 0.6 to avoid the potential for oxygen toxicity, and this can usually be accomplished by recruitment and increasing PEEP, using pulse oximetry as a guide. Positive end-expiratory pressure exploits the hysteresis of the inflation and deflation limbs of each breath to recruit alveoli on inspiration (by expanding collapsed units and translocating fluid from flooded units to the interstitial space) and then preventing derecruitment at end expiration. Recruitment maneuvers such as applying 40 cm H₂O continuous positive airway pressure for 40 seconds may increase lung opening, which can be held open by increased PEEP settings. The result is increased respiratory system compliance as a result of increased functional residual capacity and improved gas exchange because of oxygenation of perfused recruited air space.^{13, 14} Positive end-expiratory pressure may be initially set at 12 to 15 cm H₂O in a single step and then lowered over time as tolerated, or set lower after initial recruitment maneuvers and sequentially increased if recruitment-induced improvement in oxygenation is not maintained. Both methods require re-recruitment and return to previous PEEP setting if desaturation occurs. PEEP raises pleural pressure, which will produce an increase in juxtacardiac pressure; this raises right atrial pressure (downstream pressure for filling of the heart) and decreases the transmural filling pressure of the left ventricle, and hypotension may result. This, however, is uncommon, perhaps because the pleural pressure increment is slight because of a decrease in lung compliance due to ARDS. Significant hypotension can be treated by reducing the PEEP temporarily and by infusing fluids or vasoactive drugs to restore cardiac output. Just as shunt lesions are poorly responsive to increased Fio₂, so are they to reduced Fio₂, and it is usually possible to achieve the target of 0.6 early in the management of ARDS. If initial PEEP was set at 15 cm H₂O following recruitment and adequate oxygenation was achieved, PEEP can be decreased in steps of 2 cm H₂O, allowing at least 15 minutes between steps and seeking the least PEEP that maintains the recruitment-induced oxygenation improvement. An alternative to sustained higher inflation pressure over time for recruitment is incremental increases in PEEP with fixed tidal volume. This is a one-step maneuver that simultaneously increases end- expiratory lung volume (recruiting closed lung), with the increase in PEEP keeping some portion of the newly recruited lung from collapsing at end expiration. Reducing PEEP significantly, even for short periods of time, is often associated with alveolar derecruitment and rapid arterial hemoglobin desaturation. Thus, airway disconnections should be kept to a minimum, and when required or occurring, they should be followed by re-recruitment if recruitment had been used as part of the PEEP-setting strategy. When severely hypoxemic patients fail to respond to PEEP, one should suspect a focal lung process, such as lobar pneumonia or atelectasis. In some patients, PEEP must be raised higher than 15 cm H₂O (rarely much higher in order to get the Fio₂ below 0.6).

Once the PEEP is adjusted and the Fio₂ is lowered to 0.6, the tidal volume and rate should be reevaluated. A 0.3-second end-inspiratory pause should be used to determine P_plat. If P_plat exceeds 30 cm H₂O, the tidal volume (initially set at 6 cm H₂O IBW) should be reduced progressively, in decrements of 0.5 mL/kg IBW, until this target value of less than or equal to 30 cm H₂O is achieved. The initial respiratory rate of 20 breaths/min may need to be raised to reduce the work of breathing and help with Co₂ elimination, but the expiratory flow waveform should be inspected for the presence of end-expiratory flow (signaling auto-PEEP), which should be avoided. Respiratory rate should typically not exceed 35 to 40 breaths/min.

Prone Positioning

Multiple studies have shown that a substantial fraction of patients with ARDS exhibit improved oxygenation with prone positioning, because of a redistribution of lung perfusion and a change in regional diaphragm motion.¹⁵ Prone positioning improves oxygenation by (1) shifting atelectatic lung from dependent higher blood flow to nondependent lesser blood flow areas, (2) facilitating secretion drainage from atelectatic lung, (3) relieving cardiac compression of lung, (4) optimizing the lung-tethering effect on functional residual capacity, and (5) facilitating the dorsal chest protective effect against gravity influence. There are, however, practical challenges. 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. One controlled trial failed to show a survival benefit to prone ventilation, but a subset of more severely hypoxic patients may have benefited.¹⁶ A subsequent study suggested benefit of prone positioning when initiated early in more severe ARDS and sustained for longer periods of time each day.^{17, 18} We recommend initial prone positioning for 14 to 16 hours before return to the supine position in patients with the most severe form of ARDS (Fio₂ of 0.8 or greater or P_plat > 30 cm H₂O despite low tidal volume strategy).

High-Frequency Ventilation

High-frequency ventilation is the application of mechanical ventilation with a respiratory rate of greater than 100 breaths/min with a small tidal volume. This type of ventilation may improve oxygenation in patients with refractory hypoxemia.¹⁹ However, there is not sufficient evidence to conclude that high-frequency oscillatory ventilation reduces mortality or long-term morbidity rates in patients with ALI or ARDS.

Extracorporeal Gas Exchange

The use of venovenous extracorporeal membrane oxygenation (ECMO) has been an attractive strategy for the management of patients with acute lung injury. The Conventional Ventilatory Support versus Extracorporeal Membrane Oxygenation for Severe Adult Respiratory Failure (CESAR) trial has shown evidence of survival improvement in patients with severe but potentially reversible acute lung injury.²⁰ This strategy carries multiple technical difficulties; however, it is still a heroic salvage therapy for patients with isolated severe respiratory failure of reversible cause, in whom all other supportive measures have failed.

Inhaled Selective Pulmonary Vasodilator

Inhaled selective pulmonary vasodilators such as nitric oxide or prostacyclin have several potential salutary effects in ARDS.²¹ They selectively vasodilate pulmonary capillaries and arterioles that subserve ventilated alveoli, diverting blood flow to these alveoli (and away from areas of shunt). The beneficial effects on oxygenation of inhaled nitric oxide (INO) initially take place at inspired concentrations of 1 to 10 ppm, but roughly one-third of patients will not respond at all. There are ample studies showing beneficial physiologic effects of inhaled nitric oxide, but two large, controlled studies failed to show a clinical outcome benefit.^{22, 23} Potential adverse effects of INO include met-hemoglobinemia and increased levels of nitrogen dioxide, neither of which is common unless high doses of INO (80 ppm) are used. It is not clear whether there is any role for inhaled nitric oxide or inhaled prostacyclin, but it is reasonable to try it in patients whose hypoxemia remains refractory to mechanical ventilation optimization and with contraindications or failure of prone positioning.

Steroids

As the pathophysiology of acute lung injury seems to be unregulated inflammatory response of the lung, corticosteroid therapy has been studied as a possible effective treatment of ARDS. However, several trials of high-dose steroids for a short period at the onset of ARDS failed to demonstrate benefit.

A subset of patients with ARDS will develop disordered healing with subsequent lung fibrosis. This so-called unresolving fibroproliferative phase will manifest as persistent infiltrates on chest radiography without evidence of infection, increasing airway pressures, a further fall in lung compliance, less responsiveness to PEEP, and rising minute ventilation requirements.²⁴

Subsequent studies have shown that corticosteroids can ameliorate pulmonary inflammation at this phase. The current data on steroid use in patients with severe ARDS, although encouraging, need further validation in larger randomized trials before this therapy should be considered for general use in clinical practice. Decisions for steroid use in the circumstance of unresolving ARDS need to be individualized, and steroids should not routinely be used in this circumstance as no survival advantage was demonstrated in the ARDSNet trial.²⁵

If methylprednisolone is administered for unresolving ARDS, it should be avoided after day 13 from the onset of ARDS, the dose should be modest (0.5 mg/kg every 6 hours) and tapered slowly, active infection should be ruled out, and the use of neuromuscular blocking agents concomitantly should be avoided.

Neuromuscular Blocking Agents

The administration of neuromuscular blocking agents in patients with ARDS improves oxygenation primarily when there is persistent patient-ventilator dyssynchrony.²⁶ When these agents are used, the lowest effective dose is recommended, and the duration should be as brief as possible. Intermittent dosing is preferred. If continuous infusion is used, a twitch monitor (ulnar or temporal) should be applied, and a minimum of “two out of four” responses after a train-of-four electrical stimulus maintained. The titration of neuromuscular blocking agents should be toward unassisted breathing. If this can be accomplished with “four out of four,” that is ideal. Concomitant administration of corticoste-roids should be avoided because of the increased risk of myopathy.

Hemodynamic and Circulatory Management of ARDS

If pulmonary edema is diminished in acute lung injury, ventilator, PEEP, and oxygen therapy can be reduced. Several studies initially reported data showing a correlation between survival and net diuresis or reduction in PAOP, and a prospective trial showed that reducing extravascular lung water decreased ventilator and intensive care unit (ICU) days in patients with ARDS.²⁷ However, diuresis and fluid restriction could compromise perfusion, potentially contributing to organ dysfunction, so the target should be to seek the lowest intravascular volume compatible with an adequate cardiac output. A large clinical trial supported that a fluid-conservative versus a fluid-liberal approach improved outcome.²⁸ Use of central venous pressure monitoring to achieve this goal was as equally effective as a pulmonary artery catheter.²⁹ In the absence of hypotension or tissue hypoperfusion (oliguria), central venous pressure should be maintained at less than 5 mm Hg.

Increased intrathoracic pressure due to applied PEEP could theoretically decrease cerebral venous outflow, which could lead to increased intracranial pressure. Whether this actually occurs is uncertain. Observational studies with patients who had acute stroke or acute subarachnoid hemorrhage found that incremental elevation of applied PEEP was associated with diminished cerebral perfusion pressure due to decreased mean arterial pressure.^{30, 31} Another study that investigated 21 comatose patients with severe head injury or subarachnoid hemorrhage receiving ICP monitoring found that in patients with poor pulmonary compliance (as in ARDS), PEEP has no significant effect on cerebral and systemic hemodynamics, probably because the PEEP-induced increase in intrathoracic pressure is poorly transmitted to the pleural (juxtacardiac) space.³²

Given the uncertainty that exists, it is prudent to monitor both the mean arterial blood pressure and the intracranial pressure (if available) whenever applied PEEP is used in a patient with an intracranial abnormality. Reduction of applied PEEP should be considered if the mean arterial blood pressure decreases or intracranial pressure increases.

Pathophysiology and Manifestation of Neurogenic Pulmonary Edema

Neurogenic pulmonary edema (NPE) needs special attention in the neurology ICU setting as a hybrid form of ARDS, although its pathophysiology and prognosis are significantly different. This entity characteristically presents within minutes to hours of a severe central nervous system insult and usually resolves within several days. It is believed to be caused by a severe acute generalized systemic rise in vascular resistance, producing a sudden surge in pulmonary capillary pressure, resulting in pulmonary capillary endothelial injury. By the time of diagnosis, the initial insult has typically abated but endothelial injury persists.

Neurogenic pulmonary edema presents with the same clinical and radiologic findings of acute pulmonary edema without the evidence of left ventricular dysfunction. This picture may be confused with aspiration pneumonitis because they both commonly occur in settings of altered consciousness. A clinical diagnosis is based largely on the occurrence of pulmonary edema in the appropriate setting and in the absence of another obvious cause. NPE usually occurs in association with elevated ICP, but elevated ICP is not a necessary condition.

The major causes of NPE are cerebral hemorrhage, epileptic seizure, and head injury. Less commonly it may occur in nonhemorrhagic stroke, brain tumors, Guillain-Barré syndrome, vertebral artery ligation, and hydrocephalus.³³

Although NPE has been recognized for more than a century, its pathophysiologic mechanisms remain incompletely understood. Various theories have centered on potential roles for the hypothalamus and medulla,³⁴ elevated intracranial pressure, and activation of the sympathoadrenal system. α-Adrenergic blockers such as phentolamine can prevent the development of NPE, suggesting an important role for sympathetic activation. Pulmonary venoconstriction can occur with intracranial hypertension or sympathetic stimulation and can elevate capillary hydrostatic pressure and produce pulmonary edema without affecting left atrial, systemic, or pulmonary capillary wedge pressures.

The outcome of patients with NPE is usually determined by the course of the neurologic insult that produced NPE, and specific treatment should focus on the underlying disorder. NPE is generally managed in a supportive and conservative fashion because many episodes of NPE are well tolerated and the majority of cases resolve within 48 to 72 hours. Positive pressure mechanical ventilation is commonly required, and the pulmonary edema is PEEP responsive; however, special attention should be given to the hypercapnia because it carries the risk of cerebral vasodilatation and exacerbation of elevated intracranial pressure. Maintenance of low cardiac filling pressures may decrease edema genesis, but care must be taken to avoid compromise of cardiac output and cerebral perfusion; pulmonary artery catheterization may be helpful in guiding therapy. NPE is generally considered a contraindication to lung harvesting for transplantation.

Prognosis and Outcome of ARDS

Negative cognitive, emotional, and quality-of-life outcomes have been reported in survivors of ARDS.^{35, 36} They found that these outcomes occurred in over 70% of patients at hospital discharge and over 40% at 1 and 2 years after discharge. The etiology of these impairments is likely multifactorial and includes hypoxemia, hypotension, sedative use, delirium, and inflammatory mediators.

We wish to acknowledge Dr. Gregory A. Schmidt (University of Iowa) and the Society of Critical Care Medicine. Dr. Schmidt co-authored with RPD the syllabus chapter on Acute Respiratory Distress Syndrome published in the Adult Multiprofessional Critical Care Review, eds, JE Szalados and CG Rehm, Society of Critical Care Medicine, August 2007. Revision and update of some portions of that chapter were used in the preparation of the current manuscript.