Acute Respiratory Distress Syndrome: Integrated Lung Physiology

Acute respiratory distress syndrome (ARDS) combines injury to the gas-exchanging barrier with an uneven loss of aeration and changing mechanical and vascular conditions. The remaining ventilated lung must support gas exchange while bearing the loads generated by the ventilator and the patient. Its response becomes understandable when aeration, perfusion, distension, effort, and circulation are followed together.

75.1 A clinical syndrome contains several physiological states

ARDS develops in an acute clinical context with impaired oxygenation and bilateral pulmonary abnormalities. Pneumonia, aspiration, systemic inflammation, and other insults can injure the alveolar–capillary interface through different pathways. The syndrome identifies a shared clinical pattern without establishing one uniform histological lesion or one mechanical state.

The global definition published in 2024 broadened the earlier Berlin framework, including selected patients receiving high-flow nasal oxygen, additional oxygenation and imaging approaches, and adaptations for resource-limited settings. These are consensus classification choices, not measurements of recruitability, lung stress, or a demonstrated response to a particular ventilator strategy. Earlier trials must still be interpreted through their actual enrollment definitions. 1

Barrier injury permits increased movement of fluid and protein into interstitial and alveolar spaces. Epithelial injury, altered surfactant function, inflammatory material, and impaired fluid clearance further change airspace stability and gas exchange. Hydrostatic forces can contribute simultaneously; identifying permeability injury does not make vascular pressure irrelevant.

Fluid, collapse, consolidation, and regional overinflation can coexist. Their proportions vary between patients and over time. A dense region on imaging may contain fluid, cells, blood, or nonaerated tissue with different capacities to reopen. The visible opacity does not by itself specify its mechanical reversibility.

The physiological consequences also separate. Severe hypoxemia can occur with different compliance, dead-space burden, pulmonary vascular load, and respiratory drive. Conversely, similar compliance can coexist with very different distributions of injury and perfusion. No single severity ratio supplies all of these dimensions.

This explains why ARDS requires an integrated description. The initiating disease, gas exchange, amount and distribution of aeration, pressure–volume behavior, patient effort, and circulatory reserve must be considered together. The diagnosis frames the problem; it does not prescribe a universal mechanical solution.

75.2 The functional lung is smaller and unevenly distributed

When some regions become poorly aerated or nonaerated, the delivered tidal volume is distributed through fewer available gas-containing units. This reduced functional compartment is commonly called the baby lung. Its importance is the amount of lung participating in ventilation, rather than a literal fixed anatomical structure.

Early computed-tomography work by Gattinoni and colleagues demonstrated dependent densities and heterogeneous aeration in ARDS. The small study and limited sampled levels did not map every alveolus, but helped establish that a chest radiograph’s diffuse appearance can conceal very unequal regional gas content. 2

Edematous lung has greater mass. The weight of overlying tissue, the heart, abdominal influences, regional pleural pressure, and intrinsic disease distribution can reduce dependent aeration. Other regions remain open and receive a disproportionate share of each breath. Position and applied pressure can change which regions participate.

Reduced respiratory-system compliance can consequently reflect fewer ventilated units operating in parallel as well as altered properties within those units. It also includes the chest wall’s contribution. Compliance is therefore informative about the operating system without being an exact measurement of functional lung size.

Boundaries between differently inflated regions create another problem. Neighboring tissue shares forces, so an expanding region can impose disproportionate deformation near a collapsed or poorly expanding neighbor. The resulting local stress may exceed what an average airway pressure suggests.

Cressoni and colleagues quantified CT-based inhomogeneity in ARDS and found associations with severity and outcome. The imaging measure supports a structural basis for uneven loading, but it is not a direct measurement of every local tissue force or proof that eliminating a particular imaging index improves survival. 3

The functional compartment is dynamic. Recruitment, derecruitment, edema evolution, position, and regional injury change both its size and its internal distribution. A tidal volume scaled to predicted healthy body size therefore does not uniquely determine the distension experienced by the remaining ventilated lung.

75.3 Shunt and inefficient ventilation arise through different pathways

Blood traversing nonaerated regions receives little additional oxygen, producing an intrapulmonary shunt contribution. Poorly ventilated but perfused regions create low ventilation–perfusion ratios. Increasing inspired oxygen helps some low-ratio regions more than completely unventilated units, but the observed response does not uniquely identify one lesion.

At the opposite end, ventilation reaching regions with little perfusion contributes to inefficient carbon dioxide elimination. Vascular obstruction, microvascular injury, regional overdistension, and reduced pulmonary blood flow can increase this burden. Shunt and high-ratio regions can coexist within the same lung.

Nuckton and colleagues prospectively measured an elevated physiological dead-space fraction early in ARDS and associated higher values with mortality. The observation identifies prognostic information beyond hypoxemia in that cohort; it does not prove that lowering the measured fraction by any means improves survival. 4

As developed in Chapter 32, clinical carbon-dioxide-based estimates depend on their measurement convention and include effects beyond a simple count of unperfused alveoli.

Blood mixing must be followed by oxygen content. For a constructed two-stream model, assume 70% of flow leaves exchanging regions at 20 mL O₂/dL, while 30% retains a mixed venous content of 12 mL/dL. The arterial mixture is 0.70 × 20 + 0.30 × 12 = 17.6 mL/dL. Oxygen tensions cannot simply be averaged because their relationship with content is nonlinear.

If mixed venous content falls to 10 with the other assumptions unchanged, arterial content becomes 17.0 mL/dL. The shunt fraction has not changed, yet the resulting arterial oxygen content has. Cardiac output, systemic extraction, and demand can therefore alter the oxygenation consequences of the same assumed pulmonary mixing pattern.

Carbon dioxide adds further interactions. Greater metabolic production requires more effective alveolar ventilation, while increasing minute ventilation can add mechanical load. A rising arterial carbon dioxide tension can reflect increased wasted ventilation, reduced effective ventilation, greater production, or combinations. The arterial gas is an integrated result, not a direct regional map.

75.4 Global mechanics must be related to stress and strain

Pressure supplies information only when its location and measurement conditions are known. During passive ventilation with an interpretable end-inspiratory no-flow hold, plateau airway pressure approximates the communicating units’ equilibrated pressure. Total end-expiratory pressure includes applied and intrinsic components that can be measured under suitable conditions.

Let Pplat be plateau pressure and PEEPtot total end-expiratory pressure in cm H₂O. Let VT be tidal volume in mL and CRS respiratory-system compliance in mL/cm H₂O:

For constructed VT 400 mL, Pplat 26, and PEEPtot 10 cm H₂O, driving pressure is 16 and compliance 25 mL/cm H₂O. This is a global respiratory-system measurement. Effort, incomplete equilibration, airway closure, and regional communication can make the simple interpretation invalid.

Lung stress concerns distending forces within the lung, commonly represented globally by transpulmonary pressure. Strain describes deformation relative to an explicitly chosen reference. A simple tidal expansion ratio of 0.40 L over an end-expiratory gas volume of 0.80 L is 0.50. The same tidal volume over 1.20 L gives approximately 0.33. These are constructed global ratios, not complete local strain measurements.

The reference matters. End-expiratory volume under PEEP includes static inflation and possibly recruited units; it is not automatically the relaxed functional residual capacity used in another strain definition. Regional deformation can also diverge from the global ratio.

Chiumello and colleagues demonstrated considerable variability in stress and strain at comparable airway pressures or tidal volumes. Amato and colleagues subsequently associated driving pressure with survival in a secondary mediation analysis of previous trials. Together these studies support relating breath size to the functional operating system, while neither establishes a universal local-stress correction or proves one driving-pressure target improves outcome. 5 6

75.5 Repeated mechanical energy can amplify injury

Inflation transfers energy to move gas, deform tissue, and overcome resistance. Part is stored elastically and returned during expiration; part is dissipated. Where energy is applied, how rapidly it is delivered, and how often the cycle is repeated affect the biological consequence.

Pressure–volume area represents mechanical work when the pressure reference, volume path, and units are appropriate. Airway work includes contributions from airways and chest wall, so it is not identical to energy reaching the lung parenchyma. Patient muscle activity can add work that a ventilator-only estimate misses.

Mechanical power expresses energy per unit time. Let Ebreath be a specified energy component in J/breath and f breathing frequency in breaths/min:

If an illustrative component is 0.30 J per breath, repetition at 20 breaths/min gives 6 J/min; at 30 it gives 9 J/min. The assumed per-breath energy is unchanged, but exposure per minute increases by 50%. Neither value is a human safety threshold or a complete power estimate.

Cressoni and colleagues varied mechanical power in piglets and demonstrated injury under the tested combinations of volume, pressure, and frequency. The experiment supports a repeated-energy mechanism, while its animal preparation and experimental boundary cannot supply a universal clinical threshold. Equal global power can also be distributed across different lung sizes and regions. 7

The ARDS Network trial supplied a separate kind of evidence. Its lower-tidal-volume and pressure-limiting strategy reduced mortality compared with the tested traditional strategy. This is a clinical outcome result for a defined ventilatory approach, not proof that one calculated energy term alone explains the benefit or that a patient’s local strain is known from predicted body weight. 8

Injury mechanisms therefore include excessive deformation, repeated instability, regional concentration, inflammation, and the cumulative exposure imposed on vulnerable tissue. A more normal blood gas does not establish that the mechanical cost of obtaining it was acceptable.

75.6 Recruitment and overdistension can occur simultaneously

Increasing end-expiratory pressure can maintain aeration in unstable regions and sometimes open additional units. This may distribute the next tidal breath through more lung and improve ventilation–perfusion matching. The same pressure increase can further inflate already open regions, compress vessels, and increase static loading.

Recruitment describes regained aeration or participation under stated conditions. It differs from expansion of units that were already open. An increase in measured end-expiratory gas volume can contain both processes; a larger volume alone cannot distinguish them.

Opening and closing behavior can also differ. A region that opens at one pressure may remain aerated at a lower pressure afterward. Time, surface forces, airway closure, and surrounding tissue loading contribute to this hysteresis. A brief measurement can consequently miss slow opening or delayed derecruitment.

Gattinoni and colleagues’ CT study demonstrated wide variation in potentially recruitable lung and its relation to the response to PEEP. Its measured recruitability belonged to a particular pressure range and imaging definition; it is not a guarantee that all dense tissue will reopen safely or that every physiological improvement implies a net benefit. 9

Improved oxygenation can accompany recruitment, redistribution of perfusion, or changes in cardiac output and venous oxygen content. Improved respiratory-system compliance can suggest more favorable mechanics but also depends on operating volume and the chest wall. Neither observation independently proves absence of regional overdistension.

ART tested a recruitment-maneuver and compliance-titrated PEEP bundle against a lower-PEEP strategy in moderate-to-severe ARDS. The experimental strategy increased 28-day mortality and barotrauma. This outcome qualifies the assumption that opening more lung necessarily helps; it does not isolate the effect of every PEEP increase or establish that all higher-PEEP strategies have the same result. 10

The useful physiological question is how much additional participation is gained, where added distension occurs, and what circulatory cost accompanies it. The response must be assessed across these domains rather than judged by one oxygenation or compliance change.

75.7 Chest-wall and pleural pressures change the meaning of airway pressure

The respiratory system includes lung and chest wall. Airway pressure must overcome both, while pleural pressure forms the lung’s external boundary. Increased abdominal pressure, obesity, posture, edema, and chest-wall configuration can alter the partition between them.

At appropriately equilibrated conditions, transpulmonary pressure is alveolar pressure minus the relevant pleural pressure. A high airway pressure can therefore contain a substantial surrounding-pressure component. Conversely, a modest airway pressure during vigorous effort can coexist with a large distending difference.

For a constructed passive tidal change, let airway pressure rise by 16 cm H₂O while the representative pleural estimate rises by 6. The estimated transpulmonary change is 10 cm H₂O. With the same airway change and a pleural change of 2, it becomes 14. These values illustrate partitioning; they are not fixed correction factors or measured regional stresses.

Esophageal pressure can help estimate pleural behavior when placement, balloon filling, calibration, and interpretation are suitable. It samples a particular mechanical environment. Mediastinal effects, posture, regional pleural gradients, and the method of deriving transpulmonary pressure limit treating it as a uniform pressure surrounding every alveolus.

This distinction is especially relevant when interpreting end-expiratory support. Maintaining a favorable estimated pressure difference in one region does not prove that all dependent regions remain open or that nondependent regions avoid excessive distension. The entire pressure–volume and circulatory response still matters.

EPVent-2 compared esophageal-pressure-guided PEEP with an empirical higher-PEEP/oxygen strategy while both groups received low tidal volumes. Its primary ranked outcome combining death and ventilator-free days was not significantly improved. The comparison does not make esophageal measurement physiologically meaningless or establish a universal transpulmonary target; it tests the particular guidance strategy against its actual comparator. 11

Chapters 23 and 41 develop these pressure references in detail. Within ARDS, their essential role is to prevent airway pressure from being mistaken for a complete measure of lung distension or ventricular loading.

75.8 The injured lung loads the circulation as well as gas exchange

Pulmonary vascular load changes through several mechanisms. Hypoxic vasoconstriction, hypercapnia, acidemia, endothelial injury, microvascular obstruction, and altered vessel geometry can increase the opposition to right ventricular ejection. Regional blood-flow diversion can sometimes improve matching while increasing the burden on the remaining vascular pathways.

Lung volume adds competing effects. Inflation can open extra-alveolar vessels through tissue traction while compressing vessels within alveolar walls. Recruitment can relieve some hypoxic and geometrical constraints; overdistension can worsen vascular compression. The net pulmonary load cannot be inferred from airway pressure alone.

The right ventricle responds through activation and filling reserve, but these are finite. Dilation, tricuspid regurgitation, impaired coronary supply, and septal displacement can reduce effective pulmonary flow and left ventricular filling. A falling pulmonary artery pressure can even accompany deterioration when the ventricle can no longer sustain flow.

Mekontso Dessap and colleagues documented acute cor pulmonale despite protective ventilation in a large ARDS cohort. Severe right ventricular involvement was associated with worse outcome. The observational findings justify attention to the right heart without converting the study’s prediction variables into universal causal thresholds or proving benefit from a score-directed intervention. 12

Thoracic pressure also affects systemic venous return and left ventricular transmural load. A respiratory change can improve left ventricular ejection conditions while reducing the inflow reaching it. Oxygenation and output may consequently move in opposite directions.

These effects extend to other organs through oxygen delivery and congestion. Greater arterial oxygen content is helpful only in relation to flow and distribution. A rise in venous pressure can impair renal or hepatic drainage even if systemic arterial pressure remains acceptable. Added fluid may support useful filling in one state while aggravating pulmonary filtration and organ congestion in another.

ARDS support must therefore be interpreted as a cardiopulmonary perturbation. Gas exchange, forward flow, ventricular geometry, pressure references, and organ response belong in the same account.

75.9 Patient effort contributes both useful ventilation and mechanical exposure

Spontaneous effort can sustain ventilation, redistribute inflation, and preserve respiratory muscle activity. Its effect depends on strength, timing, lung mechanics, and assistance. The relevant distinction is the actual load generated, rather than whether a breath is labeled spontaneous or controlled.

Diaphragmatic contraction lowers pleural pressure. In injured heterogeneous lung, that change may be transmitted unevenly, producing strong dependent-region distension. Positive airway pressure from the ventilator can add to the transpulmonary difference at the same time.

Yoshida and colleagues combined a clinical observation with injured-animal experiments to demonstrate pendelluft: gas moving between lung regions during inspiration without an equivalent change in airway tidal volume. The experiments showed how dependent inflation could be concentrated despite a seemingly limited global breath. They establish a potential injury mechanism, not proof that every spontaneous effort causes additional clinical lung injury. 13

High drive can also increase tidal volume, respiratory frequency, inspiratory flow demand, and asynchrony. Breath stacking may increase delivered volume, while strong negative swings can alter vascular filtration and left ventricular load. Conversely, excessive assistance and inactivity can contribute to respiratory muscle dysfunction. Both the lung and the muscles require assessment.

Airway waveforms, occlusion measurements, esophageal signals, imaging, and examination provide different clues. None alone quantifies all regional effort or establishes a universal sedation threshold. Gas-exchange improvement after suppressing effort is also distinct from demonstrating a favorable overall outcome.

ROSE tested early continuous neuromuscular blockade with deep sedation against usual care with lighter sedation targets in moderate-to-severe ARDS managed with a higher-PEEP strategy. It did not improve its 90-day in-hospital mortality endpoint and produced more adverse cardiovascular events. The result concerns a complete strategy and comparator, rather than proving that temporary control of harmful effort is never useful or that routine paralysis follows from the pendelluft mechanism. 14

75.10 Proning changes the distribution of inflation and load

Changing posture changes the relationship between lung shape, chest-wall shape, gravity, the heart, and abdominal contents. In the prone position, these relationships can reduce the unevenness of regional inflation and improve access to previously poorly ventilated dorsal tissue.

The effect is more complex than moving edema from one surface to another. Tissue mass does not disappear when the patient is turned. Instead, external loading, regional pleural pressure, airway stability, and the distribution of gas change. Some regions gain aeration while others may lose it.

Perfusion does not simply mirror the new gravitational direction. Its anatomical and vascular determinants can preserve substantial dorsal flow. Better ventilation of that flow can reduce venous admixture, while more even inflation can distribute a breath’s mechanical burden across a larger or more uniformly participating compartment.

Chest-wall compliance may change in the opposite direction from lung compliance. A modest global compliance response therefore does not exclude improved regional homogeneity. Likewise, an immediate oxygenation response is not a complete measure of the resulting mechanical exposure or circulatory effect.

Proning can also alter pulmonary vascular load and right ventricular geometry. Vieillard-Baron and colleagues observed reduced right ventricular enlargement and septal abnormalities in selected severe ARDS patients with acute cor pulmonale. This supports an unloading mechanism in that population; the before–after physiological study was not an independent survival trial. 15

PROSEVA supplied randomized outcome evidence for an early, prolonged proning strategy in selected patients with trial-defined severe ARDS receiving protective ventilation. It reduced 28-day and 90-day mortality compared with supine care in experienced units. That result should retain the trial’s timing, exposure, population, and cointerventions; it does not establish equivalent benefit from every brief turn or every cause of hypoxemia. 16

The physiological and outcome accounts therefore reinforce each other without becoming interchangeable. Distribution, mechanics, gas exchange, cardiac response, and procedural consequences all contribute to the interpretation of a posture change.

75.11 Physiological personalization requires separate outcome validation

CT, bedside ultrasound, electrical impedance tomography, pressure measurements, and gas-exchange indices reveal different parts of the lung. Their spatial coverage, timing, assumptions, and response to changing conditions differ. A more detailed measurement does not automatically identify a beneficial treatment strategy.

Electrical impedance tomography, or EIT, tracks regional impedance changes within its sampled thoracic region. Ventilation distribution and algorithm-derived collapse or overdistension estimates can reveal responses hidden in a global compliance value. These estimates still depend on acquisition, reference conditions, regional mechanics, and the inference being made; they are not direct microscopic counts of closed or damaged alveoli.

EITVent randomized 190 patients with moderate-to-severe ARDS to EIT-guided PEEP or a lower-PEEP/oxygen-table strategy. It stopped early for futility, did not significantly improve primary 28-day mortality, and showed no significant first-week separation in PEEP. A favorable higher-recruitability subgroup finding does not replace the overall result and requires confirmation. The trial neither validates a universal EIT-derived setting nor establishes that regional measurement has no physiological value. 17

Extracorporeal gas exchange tests another link in the system. Venovenous support can remove carbon dioxide and add oxygen while allowing changes in ventilatory exposure, but requires adequate circulation to deliver the resulting blood to tissues. It does not directly replace the heart’s systemic pumping function, and its complications and full management are developed in Chapter 79.

EOLIA compared immediate venovenous support with a conventional strategy permitting rescue crossover in very severe ARDS. Its 60-day mortality difference did not meet conventional statistical significance, with substantial crossover and important bleeding-related costs. This comparison cannot be read as simple support-versus-no-support equivalence or a universal efficacy claim for every extracorporeal configuration. 18

Across these approaches, physiological information sharpens the proposed mechanism. Clinical evidence must then establish whether the complete strategy improves outcomes at an acceptable cost in the relevant population.

75.12 Integrate mechanics, gas exchange, and circulation over time

Return to the constructed passive ventilation example: VT 400 mL, total PEEP 10, and plateau pressure 26 cm H₂O. Driving pressure is 16 and respiratory-system compliance 25 mL/cm H₂O. Suppose a later state has the same tidal volume, total PEEP 16, and plateau pressure 30. Driving pressure is now 14 and compliance approximately 28.6 mL/cm H₂O.

If measured end-expiratory gas volume also rises from 0.80 to 1.20 L, the tidal expansion ratio falls from 0.50 to approximately 0.33. These findings are compatible with a more favorable global mechanical state, but the extra gas volume can include inflation of open regions as well as recruitment. Neither the ratio nor compliance establishes every region’s stress.

Now assume arterial oxygen content improves from 16 to 17 mL O₂/dL while forward cardiac output falls from 5 to 4 L/min. Oxygen delivery, calculated as flow × content × 10, changes from 800 to 680 mL O₂/min. That is a fall of 120 mL/min, or 15%, despite the improved arterial content and lower driving pressure.

The example does not identify why output fell. Venous-return limitation, pulmonary vascular loading, ventricular reserve, transmitted pressure, and unrelated disease changes remain possibilities. New right ventricular dilation or venous congestion would add useful evidence, but would still require measurement context and temporal interpretation.

Patient effort, posture, carbon dioxide production, and the underlying injury may change during the same interval. A later gas or compliance improvement cannot be assigned automatically to one intervention. Repeated observations should describe comparable conditions and separate measured quantities from calculations that reuse them.

ARDS physiology becomes coherent when the amount and distribution of participating lung are connected to mechanical exposure, gas exchange, right-heart loading, systemic flow, and tissue need. Support is then understood through its integrated consequences. A better isolated number is evidence to interpret, while benefit belongs to the patient’s full physiological and clinical trajectory.

References

1. Matthay MA, Arabi Y, Arroliga AC, et al. A new global definition of acute respiratory distress syndrome. Am J Respir Crit Care Med. 2024;209:37–47. Original consensus definition.

2. Gattinoni L, et al. Adult respiratory distress syndrome profiles by computed tomography. J Thorac Imaging. 1986;1:25–30. Original CT study.

3. Cressoni M, Cadringher P, Chiurazzi C, et al. Lung inhomogeneity in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med. 2014;189:149–158. Original CT analysis.

4. Nuckton TJ, Alonso JA, Kallet RH, et al. Pulmonary dead-space fraction as a risk factor for death in the acute respiratory distress syndrome. N Engl J Med. 2002;346:1281–1286. Original cohort.

5. Chiumello D, Carlesso E, Cadringher P, et al. Lung stress and strain during mechanical ventilation for acute respiratory distress syndrome. Am J Respir Crit Care Med. 2008;178:346–355. Original physiological study.

6. Amato MBP, Meade MO, Slutsky AS, et al. Driving pressure and survival in the acute respiratory distress syndrome. N Engl J Med. 2015;372:747–755. Original individual-data analysis.

7. Cressoni M, Gotti M, Chiurazzi C, et al. Mechanical power and development of ventilator-induced lung injury. Anesthesiology. 2016;124:1100–1108. Original animal experiment.

8. Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med. 2000;342:1301–1308. Original randomized trial.

9. Gattinoni L, Caironi P, Cressoni M, et al. Lung recruitment in patients with the acute respiratory distress syndrome. N Engl J Med. 2006;354:1775–1786. Original CT study.

10. Writing Group for the ART Investigators, Cavalcanti AB, Suzumura ÉA, et al. Effect of lung recruitment and titrated positive end-expiratory pressure (PEEP) vs low PEEP on mortality in patients with acute respiratory distress syndrome: a randomized clinical trial. JAMA. 2017;318:1335–1345. Original ART trial.

11. Beitler JR, Sarge T, Banner-Goodspeed VM, et al. Effect of titrating positive end-expiratory pressure (PEEP) with an esophageal pressure-guided strategy vs an empirical high PEEP-Fio₂ strategy on death and days free from mechanical ventilation among patients with acute respiratory distress syndrome: a randomized clinical trial. JAMA. 2019;321:846–857. Original EPVent-2 trial.

12. Mekontso Dessap A, Boissier F, Charron C, et al. Acute cor pulmonale during protective ventilation for acute respiratory distress syndrome: prevalence, predictors, and clinical impact. Intensive Care Med. 2016;42:862–870. Original cohort.

13. Yoshida T, Torsani V, Gomes S, et al. Spontaneous effort causes occult pendelluft during mechanical ventilation. Am J Respir Crit Care Med. 2013;188:1420–1427. Original physiological experiments.

14. National Heart, Lung, and Blood Institute PETAL Clinical Trials Network, Moss M, Huang DT, et al. Early neuromuscular blockade in the acute respiratory distress syndrome. N Engl J Med. 2019;380:1997–2008. Original ROSE trial.

15. Vieillard-Baron A, Charron C, Caille V, et al. Prone positioning unloads the right ventricle in severe ARDS. Chest. 2007;132:1440–1446. Original physiological study.

16. Guérin C, Reignier J, Richard JC, et al. Prone positioning in severe acute respiratory distress syndrome. N Engl J Med. 2013;368:2159–2168. Original PROSEVA trial.

17. Yuan X, Zhong M, Li Z, et al. Electrical impedance tomography-guided positive end-expiratory pressure and mortality of patients with the acute respiratory distress syndrome: the EITVent randomized clinical trial. Am J Respir Crit Care Med. 2026;212:440–451. Original EITVent trial.

18. Combes A, Hajage D, Capellier G, et al. Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome. N Engl J Med. 2018;378:1965–1975. Original EOLIA trial.