Explore this free Rotation Guide to see what Rotation Prep has to offer.
Unlock all of Rotation Prep by linking your NEJM Resident 360 profile
to your personal or institutional subscription to NEJM, NEJM Journal Watch, or NEJM Knowledge+.
First described in 1967, acute respiratory distress syndrome (ARDS) has had many names — double pneumonia, shock lung, post-traumatic lung, respirator lung, and Da Nang lung — reflecting the heterogeneity of the syndrome. Treatment of ARDS requires correcting the underlying cause as quickly as possible while supporting the lungs with mechanical ventilation in a way that minimizes injury from mechanical ventilation. Advances in the treatment of underlying causes and ventilation methods account for most of the reduction in mortality in patients with ARDS. In this section, we review the following topics related to ARDS:
Curbside Consults: Listen to an interview with Dr. Patricia Kritek where we take a deep dive into ARDS, beyond low tidal-volume ventilation, the importance of PEEP, neuromuscular blockade, and prone positioning to highlight key clinical trials and discuss in detail how these therapies may work.
(02:27 history and overview of ARDS; 04:47 diagnosis; 10:22 ARMA trial, low tidal volume ventilation, and ventilator associated lung injury; 19:16 PEEP and recruiting lung; 28:59: optimizing PEEP and assessing pleural pressure; 37:30: additional therapies; 40:52: neuromuscular blockade; 46:07: prone positioning; 51:52: rescue therapies; 55:02: summary and concluding remarks)
Dr. Kritek is Professor of Medicine in the Division of Pulmonary and Critical Care Medicine and the Associate Medical Director of Critical Care at the University of Washington Medical Center.
Pathophysiology and Diagnosis
ARDS is a disorder of oxygenation that is secondary to diffuse alveolar damage. The damage can be scattered and nonhomogenous throughout the lungs.
The causes of the inciting injury are broad and include pneumonia, sepsis, and aspiration (most cases), as well as trauma (lung contusion and nonthoracic), pancreatitis, inhalation injury, transfusion-related acute lung injury (TRALI), drowning, hemorrhagic shock, major burn, cardiopulmonary bypass, and reperfusion edema after lung transplantation or embolectomy.
The subsequent inflammatory response to the underlying injury leads to damage to epithelial barriers (exacerbated by mechanical stretch) and accumulation of protein-rich edema fluid in alveoli. Over time, epithelial integrity is reestablished and alveolar fluid is reabsorbed. Fibrosis can follow and increase the risk for mortality. Physiologically, the alveolar damage results in ventilation-perfusion mismatch (V/Q mismatch), as evidenced by observations of increased shunting (alveoli unable to exchange oxygen) and dead space (microvascular injury leading to lack of perfusion).
The following schematic illustrates ARDS pathophysiology during the early injury phase:
Figure 2. The Healthy Lung and the Exudative Phase of ARDS. The healthy lung is shown on the left, and the exudative phase of ARDS is shown on the right. Injury is initiated by either direct or indirect insults to the delicate alveolar structure of the distal lung and associated microvasculature. In the exudative phase, resident alveolar macrophages are activated, leading to the release of potent proinflammatory mediators and chemokines that promote the accumulation of neutrophils and monocytes. Activated neutrophils further contribute to injury by releasing toxic mediators. The resultant injury leads to loss of barrier function, as well as interstitial and intraalveolar flooding. Tumor necrosis factor (TNF)–mediated expression of tissue factor promotes platelet aggregation and microthrombus formation, as well as intraalveolar coagulation and hyaline-membrane formation. AECI denotes type I alveolar epithelial cell, AECII type II alveolar epithelial cell, Ang-2 angiopoietin-2, APC activated protein C, CC-16 club cell (formerly Clara cell) secretory protein 16, CCL chemokine (CC motif) ligand, DAMP damage-associated molecular pattern, ENaC epithelial sodium channel, GAG glycosaminoglycan, HMGB1 high-mobility group box 1 protein, KL-6 Krebs von den Lungen 6, LPS lipopolysaccharide, LTB4 leukotriene B4, MMP matrix metalloproteinase, MPO myeloperoxidase, mtDNA mitochondrial DNA, Na+/K+ ATPase sodium–potassium ATPase pump, NF-κB nuclear factor kappa light-chain enhancer of activated B cells, NET neutrophil extracellular trap, PAMP pathogen-associated molecular pattern, PRR pattern recognition receptor, ROS reactive oxygen species, sICAM soluble intercellular adhesion molecule, SP surfactant protein, sRAGE soluble receptor for advanced glycation end products, VEGF vascular endothelial growth factor, and vWF von Willebrand factor. (Source: Acute Respiratory Distress Syndrome , N Engl J Med 2017.)
Because reliable biomarkers for the underlying injury of ARDS do not exist, diagnosis is based on clinical criteria. In 2012, the criteria for ARDS were revised in the Berlin Definition with the goal of identifying patients with evidence of alveolar edema on chest imaging caused by intrinsic lung injury rather than increased hydrostatic force (e.g., heart failure; see examples) and with hypoxemia (defined by the PaO2/FiO2 ratio) that requires some ventilation support (positive end-expiratory pressure [PEEP] ≥5). The severity categories also correlate with 90-day mortality.
Diagnostic Criteria for ARDS (Berlin Definition)
Chest x-ray or Computed Tomography
Bilateral opacities that are not fully explained by pleural effusions, lung collapse, or nodules
Etiology of Edema
Not fully explained by heart failure or volume overload
≤1 week since:
new or worsening respiratory symptoms and/or
known clinical insult
Oxygenation* (with PEEP ≥ 5 cm H2O)
PaO2/FiO2 200–300 mm Hg
PaO2/FiO2 100–200 mm Hg
PaO2/FiO2 ≤100 mm Hg
27% (95% CI: 24%–30%)
32% (95% CI: 29%–34%)
45% (95% CI: 42%–48%)
*Positive pressure can be delivered noninvasively in the mild ARDS group with continuous positive airway pressure (CPAP) FiO2 = fraction of inspired oxygen, PaO2 = partial pressure of arterial oxygen, PEEP = positive end-respiratory pressure.
A limitation of the Berlin Definition is the use of blood gas measurement for partial pressure of arterial oxygen (PaO2). When blood gas measurement is not available, oxygen saturation by pulse oximetry can be used as a surrogate to avoid underdetection.
Note: Mild ARDS was referred to as acute lung injury (ALI) in some literature before the publication of the Berlin Definition.
After addressing the underlying cause of ARDS, the next step is supportive care that limits further lung injury. Over time, physicians began to realize that ventilators can cause harm through the various mechanisms described below.
Causes of Ventilator-Induced Lung Injury
volutrauma (barotrauma): Delivering too much pressure leads to overdistention of alveoli. Because the compliance (Δ volume / Δ pressure) of the ARDS lung is heterogenous, the same airway pressure may cause underdistention of a more affected lung region with low compliance and overdistention of a less affected region.
atelectrauma: Allowing alveoli to collapse completely during each breath cycle with too little airway pressure leads to shear stress and denaturation of surfactants.
biotrauma: The physical force and trauma of ventilation (such as those described above) leads to release of mediators that sustain inflammation and translocation of proinflammatory products and bacteria through already permeable barriers, causing systemic damage.
The following figure provides more details of lung damage associated with ventilation:
Figure 2. Lung Injury Caused by Forces Generated by Ventilation at Low and High Lung Volumes. When ventilation occurs at low lung volumes, lung injury can be caused by the opening and closing of lung units (atelectrauma) as well as by other mechanisms. This injury is magnified when there is increased lung inhomogeneity, as shown on computed tomography (Panel A), especially in patients with the acute respiratory distress syndrome (ARDS) who have surfactant dysfunction, pulmonary edema, and atelectasis. In addition, ventilation may be very inhomogeneous, a status that may be partially or fully reversed using positive end-expiratory pressure (PEEP), as shown in a ventilated ex vivo rat lung. At high lung volumes, overdistention can lead to gross barotrauma (air leaks) (Panel B). Overdistention can also lead to increased alveolar–capillary permeability and gross pulmonary edema. Ventilation at both high and low lung volumes has structural, physiological, biologic, and systemic effects (Panel C). Mediators that are released into the lung can cause further lung injury, recruit neutrophils to the lung, or set the stage for the development of pulmonary fibrosis. In addition, the increased alveolar–capillary permeability associated with ventilator-induced lung injury can lead to translocation of mediators, lipopolysaccharides, and bacteria into the systemic circulation, potentially leading to multiple-organ dysfunction and death. PaCO2 denotes partial pressure of arterial oxygen to the fraction of inspired oxygen. (Source: Ventilator-Induced Lung Injury , N Engl J Med 2013.)
Low Tidal Volume Ventilation
In 2000, the landmark ARMA trial (also referred to as the ARDSNet trial) showed that a ventilation strategy with tidal volume of 6 mL/kg of ideal body weight and a plateau pressure ≤30 cm water (H2O) resulted in 9% lower mortality than a strategy with 12 mL/kg of ideal body weight and a plateau pressure ≤50 cm H2O (31.0% vs. 39.8%). Although the significance of tidal volume is often emphasized, it is important to remember that the ARMA trial also limited plateau pressure.
ARDSNet ventilation is now standard of care. The ARDSNet pocket card is a useful reference for calculating the ideal tidal volume and provides some general guidelines for titrating ventilator parameters.
For more on ventilator settings see the Ventilation section in this rotation guide.
Important notes about the ARDSNet strategy:
The increased dead space (ventilated but not perfused lung) in ARDS limits the fraction of each tidal breath that contributes to ventilation, leading to carbon dioxide (CO2) retention and subsequently acidemia. Although increasing the respiratory rate is helpful, a certain amount of hypercapnia (i.e., permissive hypercapnia) can prevent injury from increasing tidal volume.
The pH goal is >7.30; a pH <7.15 may require additional treatment (e.g., bicarbonate).
Normoxemia is not necessary, and trying to achieve it may cause more harm, often through the high PEEP required. The oxygenation goal is PaO2 of 55–80 mmHg or peripheral capillary oxygen saturation (SpO2) of 88%–95%.
To achieve adequate oxygenation, PEEP is helpful for opening diseased and collapsed alveoli for oxygen exchange (i.e., recruitment).
Recruitment maneuvers (maneuvers to hold a high PEEP for a period of time) are sometimes used to improve oxygenation, but the evidence for benefit is not definitive. Too much PEEP can cause overdistention and pressure on pulmonary circulation, leading to increased pulmonary resistance, decreased left heart preload, and hypotension. One goal of adjusting the ventilator is to optimize the lung’s pressure-volume curve to stay between the ends of atelectrauma and volutrauma (see figure below; A and B on the graph correspond to the CT images used as examples).
Figure 1. Schematic Diagram of a Pressure-Volume Curve of a Lung in a Patient with the Acute Respiratory Distress Syndrome. The inflation limb (lower curve) and deflation limb (upper curve) differ from one another. The lower inflection point defines the onset of alveolar recruitment from a state of substantial collapse; the lung below this point is illustrated in the axial computed tomographic (CT) scan in Panel A. The upper inflection point is thought to reflect the point at which recruitment is no longer occurring and overdistention may start to occur; the lung in this condition is illustrated in the axial CT scan in Panel B. CT scans adapted from Gattinoni et al. The CT scans in Panels A and B correspond to the areas marked A and B in the upper panel. (Adapted from High-Frequency Oscillatory Ventilation on Shaky Ground , N Engl J Med 2013.)
Sometimes such a high PEEP is needed that the plateau pressure exceeds the typical 30 cm H2O threshold for safety. In these situations, the high airway pressure may not be harmful because much of the pressure is needed to expand the tissue surrounding and compressing the lungs (as with severe obesity, massive ascites, pleural effusions, or a stiff chest wall). The transpulmonary pressure (Ptp) stresses and damages the alveoli; Ptp is the difference between alveolar pressure (Palv), measured by airway pressure on the ventilator, and pleural pressure (Ppl) (see figure below). An esophageal balloon can simulate the pressure in the pleural space and help titrate PEEP. One small, single-center trial showed that use of esophageal balloons was associated with improved oxygen and compliance and a promising nonsignificant reduction in mortality.
In addition to protective lung ventilation, the following treatments may also be helpful:
conservative fluid management
While many patients with ARDS have concurrent hypotension or shock and require fluid resuscitation, too much added fluid to increased capillary permeability leads to pulmonary edema that exacerbates lung injury. You might hear attendings and respiratory therapists say, “Dry lungs are happy lungs.”
The FACTT trial showed that a conservative fluid strategy decreased duration of mechanical ventilation, compared with a liberal strategy.
Synchrony between the patient’s respiration and the ventilator improves oxygenation by ensuring the right tidal volume (rather than the patient trying to exhale when the ventilator is delivering a breath) and prevents injury (e.g., panel E in the figure above depicting high transpulmonary pressure generated by the patient trying to inhale on top of the ventilator delivering a breath). Synchrony can be enhanced with the use of neuromuscular blocking agents (NMBA).
In 2010, the ACURASYS trial showed that the use of the NMBA cisatracurium within 48 hours of mechanical ventilation in patients with a P/F ratio <150 reduced 90-day mortality, as compared with placebo (31.6% vs. 40.7%), and increased the number of ventilator-free days.
Much of the benefit of cisatracurium in the ACURASYS trial is thought to be from minimizing ventilator-induced lung injury from dyssynchrony, once again illustrating the key principle of avoiding harm in the treatment of ARDS. Other benefits include the possible anti-inflammatory effects of NMBA and decreased oxygen requirement by muscle paralysis (see figure below). One negative feature of NMBA use is heavy sedation, which is associated with definite adverse effects.
In 2019, the ROSE trial challenged the mortality benefit of neuromuscular blockade reported in the ACURASYS trial. The ROSE trial found no significant difference in 90-day mortality between cisatracurium with deep sedation and light sedation with no neuromuscular blockade. The results of this trial have been controversial given the limited inclusion criteria (of 1004 patients screened, only 340 were included) and use of deep sedation in the early paralysis group.
The difference in mortality between the two trials is attributed to differences in sedation levels. One explanation offered in an editorial is “reverse triggering” — a phenomenon that describes additional gas delivery and overinflation in deeply sedated, but not paralyzed, patients after a mechanically assisted breath (breath delivered by the ventilator triggers a contraction of the diaphragm, which initiates a spontaneous breath). In the ACURASYS trial, the negative physiological consequences of reverse triggering might have disadvantaged the “control” patients and led to the observed mortality benefit in the paralyzed patients.
In the ROSE trial, patients in the cisastracurium group also had serious cardiovascular events, providing another reason to avoid this treatment. In general, routine use of neuromuscular blockade (NMB) in patients with moderate-to-severe ARDS is not recommended. However, NMB can be considered in patients at risk of reverse triggering or in those with increased respiratory drive or other factors that could cause marked transpulmonary pressure swings.
Patients typically lay supine in the intensive care unit (ICU); this position is associated with negative gravitational effects on the posterior lung regions, causing the heart to compress the left lung and more dependent atelectasis from interstitial edema. Placing patients in the prone position allows more lung regions to be functional and improves V/Q mismatch. (View a video of prone positioning in a patient with ARDS.)
The PROSEVA trial showed that, compared with supine positioning, prone positioning within 36 hours of mechanical ventilation in patients with a P/F ratio <150 reduced 28-day (16.0% vs. 32.8%) and 90-day mortality.
Prone positioning requires an experienced nursing team to move the patient safely and prevent subsequent complications (e.g., pressure ulcers, extubation, intravenous decannulation), thus limiting its widespread adoption.
In 2020, the DEXA-ARDS trial demonstrated a 60-day mortality benefit in mechanically ventilated patients receiving dexamethasone (20 mg IV for 5 days followed by 10 mg for another 5 days) versus controls (21% vs. 36%). Patients receiving dexamethasone also had more ventilator-free days.
How NMBAs Might Lead to Improved Survival in Patients with ARDS
Figure 1. Possible Mechanisms by Which Neuromuscular Blocking Agents (NMBAs) Might Lead to Improved Survival in Patients with the Acute Respiratory Distress Syndrome (ARDS). Respiratory physiological features of a patient with ARDS are illustrated before (top) and after (bottom) paralysis induced with the use of NMBAs. Before paralysis, increased respiratory drive from multiple causes can lead to increased tidal volumes, active exhalation, and patient–ventilator asynchrony, all of which can potentially worsen various forms of ventilator-induced lung injury. In addition, muscle activation may divert blood flow away from vital organs and lead to a lower mixed venous partial pressure of oxygen (PO2). These mechanisms may lead to increased organ dysfunction and ultimately death. After paralysis, the administered NMBAs prevent patient-initiated generation of high and low lung volumes and also prevent active expiration, allowing for better lung-protective ventilation and less ventilator-induced lung injury. Ventilator-induced lung injury may also be lessened by less pulmonary blood flow due to decreased oxygen consumption. NMBAs may also indirectly improve arterial oxygenation by decreasing blood flow to active muscle groups (because of decreased oxygen requirements) and by improving the distribution of ventilation relative to perfusion (V̇/̇Q). (Arterial PO2 may also be decreased through this mechanism if V̇/̇Q is worsened.) Finally, NMBAs may have a direct anti-inflammatory effect. The relative effect of NMBAs on many of these mechanisms depends on the state of muscle activation before paralysis, which is dependent on several factors, including the patient's level of sedation. (Source: Neuromuscular Blocking Agents in ARDS. N Engl J Med 2010.)
After exhausting the established therapies described above, the following additional treatments may be attempted for refractory hypoxemia, although strong evidence of benefit is lacking.
airway pressure release ventilation (APRV): APRV is a mode of ventilation that inverts the pressure settings; a continuous high positive airway pressure is applied and intermittently released, allowing ventilation with the goal of sustaining lung recruitment.
inhaled nitric oxide: Inhaled nitric oxide can decrease pulmonary vascular resistance locally in ventilated areas of the lung and shunt more blood to that area, thus improving V/Q mismatch and oxygenation. Small trials (e.g., Rossaint R et al.) have shown benefit, but the effect may be transient.
supplemental oxygen: Ventilated patients in the ICU often receive supplemental oxygen, but the appropriate target for arterial oxygen saturation remains controversial. Two trials (ICU-ROX and Liberal or Conservative Oxygen Therapy for ARDS) addressed this issue with somewhat differing results. ICU-ROX did not show a benefit from conservative-oxygen therapy, as compared with usual care in ICU patients, and the LOCO2 trial suggested potential harm from a conservative oxygen strategy, compared with a liberal strategy in patients with ARDS. Until we have more data, we believe that targeting an SpO2 of 92%−94% is reasonable.