Skip to Content

Acute Hypoxemic Respiratory Failure (AHRF, ARDS)


Bhakti K. Patel

, MD, University of Chicago

Last full review/revision Mar 2020| Content last modified Mar 2020

Acute hypoxemic respiratory failure is severe arterial hypoxemia that is refractory to supplemental oxygen. It is caused by intrapulmonary shunting of blood resulting from airspace filling or collapse (eg, pulmonary edema due to left ventricular failure, acute respiratory distress syndrome) or by intracardiac shunting of blood from the right- to left-sided circulation . Findings include dyspnea and tachypnea. Diagnosis is by arterial blood gas measurement and chest x-ray. Treatment usually requires mechanical ventilation.

(See also Overview of Mechanical Ventilation.)

Etiology of Acute Hypoxemic Respiratory Failure

Airspace filling in acute hypoxemic respiratory failure (AHRF) may result from

  • Elevated alveolar capillary hydrostatic pressure, as occurs in left ventricular failure (causing pulmonary edema) or hypervolemia
  • Increased alveolar capillary permeability, as occurs in any of the conditions predisposing to acute respiratory distress syndrome (ARDS)
  • Blood (as occurs in diffuse alveolar hemorrhage) or inflammatory exudates (as occur in pneumonia or other inflammatory lung conditions)

Right-to-left intracardiac shunts, in which deoxygenated venous blood bypasses the lungs and enters the systemic circulation, usually occur as a long-term complication of large, untreated left-to-right shunts (eg, from patent foramen ovale, atrial septal defect). This phenomenon is termed Eisenmenger syndrome. This discussion focuses on refractory hypoxemia due to pulmonary causes.

Pathophysiology of Acute Hypoxemic Respiratory Failure


ARDS is divided into 3 categories of severity: mild, moderate, and severe based on oxygenation defects and clinical criteria (see table Berlin Definition of ARDS). The mild category corresponds to the previous category termed acute lung injury (ALI).

Berlin Definition of ARDS

ARDS Category


Level of severity


200 mm Hg < PaO2/FIO2 ≤ 300 mm Hg* with PEEP or CPAP ≥ 5 cm H2O


100 mm Hg < PaO2/FIO2 ≤ 200 mm Hg with PEEP ≥ 5 cm H2O


PaO2/FIO2 ≤ 100 mm Hg with PEEP ≥ 5 cm H2O

Clinical criteria


Onset within 1 week of known insult or of new or worsening respiratory symptoms

Imaging (x-ray or CT of chest)

Bilateral opacities not fully explained by effusions, lobar or lung collapse, or nodules

Origin of edema

Respiratory failure not fully explained by heart failure or fluid overload

*PaO2 in mm Hg; FIO2 in decimal fraction (eg, 0.5).

ARDS = acute respiratory distress syndrome; CPAP = continuous positive airway pressure; FIO2 = fraction of inspired oxygen; PaO2 = partial pressure of arterial oxygen; PEEP = positive end-expiratory pressure.

Adapted from ARDS Definition Task Force, Ranieri VM, Rubenfeld GD, et al: Acute respiratory distress syndrome: The Berlin definition. Journal of the American Medical Association 307:2526–2533, 2012. doi: 10.1001/jama.2012.5669.

In ARDS, pulmonary or systemic inflammation leads to release of cytokines and other proinflammatory molecules. The cytokines activate alveolar macrophages and recruit neutrophils to the lungs, which in turn release leukotrienes, oxidants, platelet-activating factor, and proteases. These substances damage capillary endothelium and alveolar epithelium, disrupting the barriers between capillaries and airspaces. Edema fluid, protein, and cellular debris flood the airspaces and interstitium, causing disruption of surfactant, airspace collapse, ventilation-perfusion mismatch, shunting, and pulmonary hypertension. The airspace collapse more commonly occurs in dependent lung zones. This early phase of ARDS is termed exudative. Later, there is proliferation of alveolar epithelium and fibrosis, constituting the fibro-proliferative phase.

Causes of ARDS may involve direct or indirect lung injury.

Common causes of direct lung injury are

Less common causes of direct lung injury are

Common causes of indirect lung injury include

  • Sepsis
  • Trauma with prolonged hypovolemic shock

Less common causes of indirect lung injury include

Sepsis and pneumonia account for about 60% of cases.

Refractory hypoxemia

Whatever the cause of airspace filling in AHRF, flooded or collapsed airspaces allow no inspired gas to enter, so the blood perfusing those alveoli remains at the mixed venous oxygen content no matter how high the fractional inspired oxygen (FIO2). This effect ensures constant admixture of deoxygenated blood into the pulmonary vein and hence arterial hypoxemia. In contrast, hypoxemia that results from ventilating alveoli that have less ventilation than perfusion (ie, low ventilation-to-perfusion ratios as occur in asthma or chronic obstructive pulmonary disease and, to some extent, in ARDS) is readily corrected by supplemental oxygen.

Symptoms and Signs of Acute Hypoxemic Respiratory Failure

Acute hypoxemia (see also Oxygen Desaturation) may cause dyspnea, restlessness, and anxiety. Signs include confusion or alteration of consciousness, cyanosis, tachypnea, tachycardia, and diaphoresis. Cardiac arrhythmia and coma can result.

Inspiratory opening of closed airways causes crackles, detected during chest auscultation; the crackles are typically diffuse but sometimes worse at the lung bases, particularly in the left lower lobe. Jugular venous distention occurs with high levels of positive end-expiratory pressure (PEEP) or right ventricular failure.

Diagnosis of Acute Hypoxemic Respiratory Failure

  • Chest x-ray and arterial blood gas (ABG) measurement
  • Clinical definition (see table Berlin Definition of ARDS)

Hypoxemia is usually first recognized using pulse oximetry. Patients with low oxygen saturation should have a chest x-ray and ABGs and be treated with supplemental oxygen while awaiting test results.

If supplemental oxygen does not improve the oxygen saturation to > 90%, right-to-left shunting of blood should be suspected. An obvious alveolar infiltrate on chest x-ray implicates alveolar flooding as the cause, rather than an intracardiac shunt. However, at the onset of illness, hypoxemia can occur before changes are seen on x-ray.

Once AHRF is diagnosed, the cause must be determined, considering both pulmonary and extrapulmonary causes. Sometimes a known ongoing disorder (eg, acute myocardial infarction, pancreatitis, sepsis) is an obvious cause. In other cases, history is suggestive; pneumonia should be suspected in an immunocompromised patient, and alveolar hemorrhage is suspected after bone marrow transplantation or in a patient with a connective tissue disease. Frequently, however, critically ill patients have received a large volume of IV fluids for resuscitation, and high-pressure AHRF (eg, caused by ventricular failure or fluid overload) resulting from treatment must be distinguished from an underlying low-pressure AHRF (eg, caused by sepsis or pneumonia).

High-pressure pulmonary edema due to left ventricular failure is suggested by a 3rd heart sound, jugular venous distention, and peripheral edema on examination and by the presence of diffuse central infiltrates, cardiomegaly, and an abnormally wide vascular pedicle on chest x-ray. The diffuse, bilateral infiltrates of ARDS are generally more peripheral. Focal infiltrates are typically caused by lobar pneumonia, atelectasis, or lung contusion. Although echocardiography may show left ventricular dysfunction, implying a cardiac origin, this finding is not specific because sepsis can also reduce myocardial contractility.

When ARDS is diagnosed but the cause is not obvious (eg, trauma, sepsis, severe pulmonary infection, pancreatitis), a review of drugs and recent diagnostic tests, procedures, and treatments may suggest an unrecognized cause, such as use of a radiographic contrast agent, air embolism, or transfusion. When no predisposing cause can be uncovered, some experts recommend doing bronchoscopy with bronchoalveolar lavage to exclude alveolar hemorrhage and eosinophilic pneumonia and, if this procedure is not revealing, a lung biopsy to exclude other disorders (eg, hypersensitivity pneumonitis, acute interstitial pneumonitis).

Prognosis for Acute Hypoxemic Respiratory Failure

Prognosis is highly variable and depends on a variety of factors, including etiology of respiratory failure, severity of disease, age, and chronic health status. Overall, mortality in ARDS was very high (40 to 60%) but has declined in recent years to 25 to 40%, probably because of improvements in mechanical ventilation and in treatment of sepsis. However, mortality remains very high (> 40%) for patients with severe ARDS (ie, those with a PaO2:FIO2 < 100 mm Hg). Most often, death is not caused by respiratory dysfunction but by sepsis and multiorgan failure. Persistence of neutrophils and high cytokine levels in bronchoalveolar lavage fluid predict a poor prognosis. Mortality otherwise increases with age, presence of sepsis, and severity of preexisting organ insufficiency or coexisting organ dysfunction.

Pulmonary function returns to close to normal in 6 to 12 months in most ARDS patients who survive; however, patients with a protracted clinical course or severe disease may have residual pulmonary symptoms, and many have persistent neuromuscular weakness.

Treatment of Acute Hypoxemic Respiratory Failure

  • Mechanical ventilation if oxygen saturation is < 90% on high-flow oxygen

Underlying conditions must be addressed as discussed elsewhere. AHRF is initially treated with high flows of 70 to 100% oxygen by a nonrebreather face mask. If oxygen saturation > 90% is not obtained, mechanical ventilation probably should be instituted. Specific management varies by condition.

Mechanical ventilation in cardiogenic pulmonary edema

Mechanical ventilation (see also Overview of Mechanical Ventilation) benefits the failing left ventricle in several ways. Positive inspiratory pressure reduces left and right ventricular preload and left ventricular afterload and reduces the work of breathing. Reducing the work of breathing may allow redistribution of a limited cardiac output away from overworked respiratory muscles. Expiratory pressure (expiratory positive airway pressure [EPAP] or PEEP) redistributes pulmonary edema from alveoli to the interstitium, allowing more alveoli to participate in gas exchange.

Noninvasive positive pressure ventilation (NIPPV), whether continuous positive pressure ventilation or bilevel ventilation, is useful in averting endotracheal intubation in many patients because drug therapy often leads to rapid improvement. Typical settings are inspiratory positive airway pressure (IPAP) of 10 to 15 cm H2O and EPAP of 5 to 8 cm H2O.

Conventional mechanical ventilation can use several ventilator modes. Most often, assist-control (A/C) is used in the acute setting, when full ventilatory support is desired. Initial settings are tidal volume of 6 to 8 mL/kg ideal body weight, respiratory rate of 25/minute, FIO2 of 1.0, and PEEP of 5 to 8 cm H2O. PEEP may then be titrated upward in 2.5-cm H2O increments while the FIO2 is decreased to nontoxic levels.

Pressure support ventilation can also be used (with similar levels of PEEP). The initial inspiratory airway pressure delivered should be sufficient to fully rest the respiratory muscles as judged by subjective patient assessment, respiratory rate, and accessory muscle use. Typically, a pressure support level of 10 to 20 cm H2O over PEEP is required.

Mechanical ventilation in ARDS

Nearly all patients with ARDS require mechanical ventilation, which, in addition to improving oxygenation, reduces oxygen demand by resting respiratory muscles. Targets include

  • Plateau alveolar pressures < 30 cm H2O (factors that potentially decrease chest wall and abdominal compliance considered)
  • Tidal volume 6 mL/kg predicted body weight to minimize further lung injury
  • FIO2 as low as is allowed to maintain adequate oxygen saturation to minimize possible oxygen toxicity

PEEP should be high enough to maintain open alveoli and minimize FIO2 until a plateau pressure of 28 to 30 cm H2O is reached. Patients with moderate to severe ARDS are the most likely to have mortality reduced by use of higher PEEP.

NIPPV is occasionally useful with ARDS. However, compared with treatment of cardiogenic pulmonary edema, higher levels of support for a longer duration are often required, and EPAP of 8 to 12 cm H2O is often necessary to maintain adequate oxygenation. Achieving this expiratory pressure requires inspiratory pressures > 18 to 20 cm H2O, which are poorly tolerated; maintaining an adequate seal becomes difficult, the mask becomes more uncomfortable, and skin necrosis and gastric insufflation may occur. Also, NIPPV-treated patients who subsequently need intubation have generally progressed to a more advanced condition than if they had been intubated earlier; thus, critical desaturation is possible at the time of intubation. Intensive monitoring and careful selection of patients for NIPPV are required.

Conventional mechanical ventilation in ARDS previously focused on normalizing arterial blood gas values. It is clear that ventilating with lower tidal volumes reduces mortality. Accordingly, in most patients, tidal volume should be set at 6 mL/kg ideal body weight (see sidebar Initial Ventilator Management in ARDS). This setting necessitates an increase in respiratory rate, even up to 35/minute, to produce sufficient alveolar ventilation to allow for adequate carbon dioxide removal. On occasion, however, respiratory acidosis develops, some degree of which is accepted for the greater good of limiting ventilator-associated lung injury and is generally well tolerated, particularly when pH is ≥ 7.15. If pH drops below 7.15, bicarbonate infusion or tromethamine may be helpful. Similarly, oxygen saturation below "normal" levels may be accepted; target saturation of 88 to 95% limits exposure to excessive toxic levels of FiO2 and still has survival benefit.

Because hypercapnia may cause dyspnea and cause the patient to breathe in a fashion that is not coordinated with the ventilator, analgesics (fentanyl or morphine) and sedatives (eg, propofol initiated at 5 mcg/kg/minute and increasing to effect up to 50 mcg/kg/minute; because of the risk of hypertriglyceridemia, triglyceride levels should be checked every 48 hours) may be needed. Sedation is preferred to neuromuscular blockade because blockade still requires sedation and may cause residual weakness.

PEEP improves oxygenation in ARDS by increasing the volume of aerated lung through alveolar recruitment, permitting the use of a lower FIO2. The optimal level of PEEP and the way to identify it have been debated. Routine use of recruitment maneuvers (eg, titration of PEEP to maximal pressure of 35 to 40 cm H2O and held for 1 minute) followed by decremental PEEP titration was found to be associated with an increased 28-day mortality (1). Therefore, many clinicians simply use the least amount of PEEP that results in an adequate arterial oxygen saturation on a nontoxic FIO2. In most patients, this level is a PEEP of 8 to 15 cm H2O, although, occasionally, patients with severe ARDS require levels > 20 cm H2O. In these cases, close attention must be paid to other means of optimizing oxygen delivery and minimizing oxygen consumption.

The best indicator of alveolar overdistention is measurement of a plateau pressure through an end-inspiratory hold maneuver; it should be checked every 4 hours and after each change in PEEP or tidal volume. The target plateau pressure is < 30 cm H2O. If the plateau pressure exceeds this value and there is no problem with the chest wall that could be contributing (eg, ascites, pleural effusion, acute abdomen, chest trauma), the physician should reduce the tidal volume in 0.5- to 1.0-mL/kg increments as tolerated to a minimum of 4 mL/kg, raising the respiratory rate to compensate for the reduction in minute ventilation and inspecting the ventilator waveform display to ensure that full exhalation occurs. The respiratory rate may often be raised as high as 35/minute before overt gas trapping due to incomplete exhalation results. If plateau pressure is < 25 cm H2O and tidal volume is < 6 mL/kg, tidal volume may be increased to 6 mL/kg or until plateau pressure is > 25 cm H2O. Some investigators believe pressure control ventilation protects the lungs better, but supportive data are lacking, and it is the peak pressure rather than the plateau pressure that is being controlled. With pressure control ventilation, because the tidal volume will vary as the patient's lung compliance evolves, it is necessary to continually monitor the tidal volume and adjust the inspiratory pressure to ensure that the patient is not receiving too high or too low a tidal volume.

Initial Ventilator Management in ARDS

Generally, the following approach is recommended for ventilator management in ARDS:

  • Assist-control mode is used initially with a tidal volume 6 mL/kg ideal body weight, respiratory rate 25/minute, flow rate 60 L/minute, FIO2 1.0, and PEEP 15 cm H2O.
  • Once oxygen saturation is > 90%, FIO2 is decreased.
  • Then, PEEP is decreased in 2.5-cm H2O increments as tolerated to find the least PEEP associated with an arterial oxygen saturation of 90% on an FIO2 of ≤ 0.6.
  • The respiratory rate is increased up to 35/minute to achieve a pH of > 7.15, or until the expiratory flow tracing shows end-expiratory flow.

Ideal body weight (IBW) rather than actual body weight is used to determine the appropriate tidal volume for patients with lung disease receiving mechanical ventilation:


Prone positioning improves oxygenation in some patients by allowing recruitment of nonventilating lung regions. One study suggests this positioning substantially improves survival (2, 3). Interestingly, the mortality benefit from prone positioning is not related to the degree of hypoxemia or the extent of gas exchange abnormality but possibly to mitigating ventilator-induced lung injury (VILI).

Optimal fluid management in patients with ARDS balances the requirement for an adequate circulating volume to preserve end-organ perfusion with the goal of lowering preload and thereby limiting transudation of fluid in the lungs. A large multicenter trial has shown that a conservative approach to fluid management, in which less fluid is given, shortens the duration of mechanical ventilation and length of stay in the intensive care unit when compared with a more liberal strategy. However, there was no difference in survival between the 2 approaches, and use of a pulmonary artery catheter also did not improve outcome (4). Patients not in shock are candidates for such an approach but should be monitored closely for evidence of decreased end-organ perfusion, such as hypotension, oliguria, thready pulses, or cool extremities.

A definitive pharmacologic treatment for ARDS that reduces morbidity and mortality remains elusive. Inhaled nitric oxide, surfactant replacement, activated protein C (drotrecogin alfa), and many other agents directed at modulating the inflammatory response have been studied and found not to reduce morbidity or mortality. Some small studies suggest that systemic corticosteroids may be beneficial in late-stage (fibroproliferative) ARDS, but a larger, prospective, randomized trial found no reduction in mortality. Corticosteroids may be deleterious when given early in the course of the condition.

Treatment references

  • 1. Writing Group for the Alveolar Recruitment for Acute Respiratory Distress Syndrome Trial (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 318(14):1335–1345, 2017. doi: 10.1001/jama.2017.14171
  • 2. Guérin C, Reignier J, Richard JC, et al: Prone positioning in severe acute respiratory distress syndrome. N Engl J Med 368(23):2159–2168, 2013. doi: 10.1056/NEJMoa1214103
  • 3. Scholten EL, Beitler JR, Prisk GK, et al: Treatment of ARDS with prone positioning. Chest 151:215–224, 2017. doi: 10.1016/j.chest.2016.06.032. Epub 2016 Jul 8
  • 4. National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network, Wiedemann HP, Wheeler AP, et al: Comparison of two fluid-management strategies in acute lung injury. N Engl J Med 354(24):2564–2575, 2006. doi: 10.1056/NEJMoa062200

Drugs Mentioned In This Article

Drug Name Select Trade
tromethamine THAM
propofol DIPRIVAN
aspirin No US brand name

Copyright © 2022 Merck & Co., Inc., known as MSD outside of the US, Kenilworth, New Jersey, USA. All rights reserved. Merck Manual Disclaimer