Abhijit Duggal, MD, MPH

An 85-year-old woman was brought to our intensive care unit because of worsening hypoxemia over the past day. About 3 weeks earlier she had been diagnosed with acute bilateral pulmonary emboli in the distal branches of the left and right lower lobes and right middle lobe, for which she was receiving anticoagulation therapy.

At presentation she had generalized fatigue and dyspnea at rest that was worse with exertion, but she denied having fever, chest pain, or cough. Her medical history included hypertension, hyperlipidemia, hypothyroidism, stage 1 breast cancer in remission, thromboembolic stroke, and myasthenia gravis. Before her hospital admission, she had been taking rosuvastatin, metoprolol tartrate, pyridostigmine, prednisone, furosemide, levothyroxine, and rivaroxaban. She did not smoke, she was retired, and she had not traveled recently.

Her blood pressure was 135/66 mm Hg, pulse 73 beats per minute, respiratory rate 16, temperature 35.4ºC (95.7ºF), and oxygen saturation 88% while receiving oxygen at 6 L/min via nasal cannula. Physical examination revealed mild edema in the lower extremities and basilar decreased breath sounds. She had no finger clubbing or cyanosis and was not using accessory muscles to breathe. Of note, her oxygen saturation remained more than 93% when she was recumbent but sharply dropped to less than 85% when she was upright.

Laboratory values

Results of initial laboratory testing were as follows:

• Sodium 138 mmol/L (reference range 132–148)
• Potassium 4.2 mmol/L (3.5–5.0)
• Chloride 99 mmol/L (98–111)
• Bicarbonate 29 mmol/L (23–32)
• Creatinine 0.52 mg/dL (0.7–1.4).
• White blood cell count 11.06 × 10⁹/L (3.7–11.0)
• Hemoglobin 12.6 g/dL (12–16)
• Platelet count 211 × 10⁹/L (150–400).
• International normalized ratio 1.4.

Electrocardiography and imaging studies

Standard 12-lead electrocardiography  showed normal sinus rhythm with left axis deviation and left ventricular hypertrophy.

Chest radiography showed bilateral interstitial opacities and small pleural effusions.

Computed tomography (CT) of the chest with contrast, compared with a CT scan done 20 days earlier, showed that the pulmonary emboli had resolved.

Arterial blood gases

In view of her positional hypoxemia, blood for arterial blood gas measurements was drawn in the supine and upright positions.

Supine, with a fraction of inspired oxygen (Fio₂) via high-flow nasal cannula of 45%, her values were:

• pH 7.45 (reference range 7.35–7.45)
• Pco₂ 34 mm Hg (36–46)
• Po₂ 81 mm Hg (85–95)
• Bicarbonate 23 mmol/L (22–26).

Upright, her hypoxemia was significantly worse:

• pH 7.46
• Pco₂ 33 mm Hg
• Po₂ 57 mm Hg
• Bicarbonate 23 mmol/L.

The methemoglobin level was normal on both measurements.

During her stay in the intensive care unit, she required up to 100% Fio₂ because of persistent hypoxemia.

CAUSES OF HYPOXEMIA

1. So far, the patient’s laboratory tests and imaging studies point to which of the following as the most likely cause of her severe hypoxemia?

• Ventilation-perfusion (V/Q) mismatch
• Diffusion abnormality
• Hypoventilation
• Shunting
• None of the above

The arterial blood gas measurements suggested the possibility of shunting as the cause, although further imaging would be needed to confirm that.

V/Q mismatch can occur in respiratory failure due to pulmonary embolism, pulmonary edema, or shunting. If ventilation is preserved but perfusion is impaired, the V/Q ratio approaches infinity (dead-space ventilation), a situation that can be seen in pulmonary embolism. If perfusion is preserved and ventilation is impaired, the V/Q ratio approaches zero, which is consistent with a physiologic shunt.

Hypoxemia may improve in less severe forms of V/Q mismatch. In our patient, the repeat CT with contrast showed that her pulmonary embolism had resolved, so this is probably not the cause of her severe hypoxemia.

Diffusion abnormalities are due to defects in the lung parenchyma, such as in chronic obstructive pulmonary disease, interstitial lung disease, and lung fibrosis.

Hypoxemia from diffusion defects is usually aggravated by a precipitating factor that increases oxygen demand, and it usually improves with oxygen supplementation. This is unlikely in our patient, as she did not have a history of chronic interstitial lung disease and CT showed no evidence of severe lung parenchymal disease.

Hypoventilation is usually due to drugs that cause respiratory depression, to stroke, or to neuromuscular diseases such as myasthenia gravis that can cause respiratory muscle weakness. It results in elevation of Pco₂ and, if not corrected, respiratory acidosis.

Our patient had a diagnosis of myasthenia gravis, though hypoventilation is unlikely in her case because she had a normal respiratory rate and low Pco₂ values.

Shunting can be physiologic or anatomic and can occur in the heart or the lungs. In physiologic shunting, severe V/Q mismatch can occur when ventilation is affected, as in severe pulmonary edema and pneumonia. In anatomic shunting, a defect such as an atrial septal defect or a pulmonary arteriovenous malformation allows blood to bypass areas of ventilation from the venous to the arterial circulation, preventing it from being oxygenated. In true anatomic shunting, supplemental oxygen with 100% Fio₂ has little effect, whereas in V/Q mismatch it can raise the arterial oxygen saturation.

Our patient’s radiograph did not suggest severe pneumonia or pulmonary edema, which makes these unlikely causes of her hypoxemia. At this point, because of her positional hypoxemia, further evaluation with contrast-enhanced echocardiography was needed to evaluate for anatomic shunting in the heart or lungs.

FURTHER TESTING

Transthoracic echocardiography (TTE) with agitated saline with a Valsalva maneuver was performed. Normally, no bubbles are seen in  the left-sided chambers after intravenous injection of agitated saline contrast, whereas bubbles on the left side suggest an intracardiac or intrapulmonary shunt. In our patient, this test was negative, and her right ventricular systolic pressure was normal.

2. What further testing should be considered to evaluate our patient’s hypoxemia?

• High-resolution chest CT
• Transesophageal echocardiography (TEE)
• Pulmonary function testing
• Repeated arterial blood gas measurement
• Edrophonium testing

Repeat imaging with high-resolution CT would likely not provide additional information and would expose the patient to additional radiation without adding much clinical benefit.

TEE could help further evaluate the intracardiac anatomy and look for shunting, which may be missed on TTE because of suboptimal positioning or image quality.

Pulmonary function testing is useful in establishing the baseline function and impairment in respiratory volumes. If an acute myasthenic crisis is suspected, measuring the negative inspiratory force and the forced vital capacity can be useful in monitoring worsening respiratory muscle weakness and assessing the need for mechanical ventilation.

In our patient, it is unlikely that pulmonary function testing would help, since her acute respiratory failure was probably not caused by neuromuscular weakness.

Repeated arterial blood gas measurement would likely only confirm that the patient still has positional hypoxemia but would not help sort through the differential diagnosis.

Edrophonium testing is useful in diagnosing myasthenia gravis and differentiating it from other neuromuscular diseases, such as Lambert-Eaton syndrome. Edrophonium, a reversible acetylcholinesterase inhibitor, prevents degradation of acetylcholine and prolongs its effect at the synaptic cleft, thus improving muscle weakness.

Our patient has already been diagnosed with myasthenia gravis, so this test is not likely to uncover the cause of her hypoxemia.

Because we still strongly suspected a shunt, TEE was performed with intravenous injection of agitated saline. TEE with the patient upright revealed intracardiac right-to-left shunting through a patent foramen ovale. The midesophageal view after saline injection showed a large interatrial septal aneurysm with total excursion of 2 cm, and right-to-left shunting within the first beat, consistent with an intracardiac shunt (Figure 1). Color Doppler imaging (Figure 2) demonstrated turbulent flow through the patent foramen ovale, consistent with right-to-left shunting, and also showed the patent foramen ovale in a closed position (Figure 3).

3. Which is now most likely the cause of our patient’s hypoxemia?

• Chronic thromboembolic pulmonary hypertension
• Myasthenic crisis
• Platypnea-orthodeoxia syndrome due to the patent foramen ovale
• Methemoglobinemia

Chronic thromboembolic pulmonary hypertension is usually a long-term result of untreated or inadequately treated thromboembolic disease (eg, pulmonary emboli), which causes vascular remodeling and pulmonary arteriopathy, which in turn leads to increased pulmonary vascular resistance and pulmonary hypertension.

This is unlikely the cause of our patient’s acute hypoxemia, as her symptoms did not suggest it. Moreover, an elevated right ventricular systolic pressure on TTE would suggest pulmonary hypertension, but TTE did not show this, and repeat chest CT indicated that her pulmonary embolism had been adequately treated and had resolved. A V/Q scan and right heart catheterization would help rule out chronic thromboembolic pulmonary hypertension, although these were not done in our patient.

Myasthenic crisis is the progressive fatiguing and paralysis of respiratory muscles ultimately requiring mechanical ventilation to sustain life. It is often brought on by infection or drug therapy.

Our patient did not require intubation and she had no signs or symptoms of myasthenic crisis such as ptosis, dysphagia, or dysarthria. She had a negative inspiratory force of −21 cm H₂O, and pulmonary function testing 4 days before her hospital admission had shown a forced vital capacity of 1.84 L, making myasthenic crisis an unlikely cause of her respiratory failure.

Platypnea-orthodeoxia syndrome is a syndrome of dyspnea (platypnea) and hypoxemia (orthodeoxia) that is induced by sitting upright or standing and resolves when lying down. It is a result of right-to-left intracardiac or intrapulmonary shunting in the presence of an anatomic defect and a functional element causing redirection of shunt flow through the anatomic defect in an upright position.¹ It is associated with specific cardiac, pulmonary, and hepatic diseases, such as atrial septal defect, pulmonary arteriovenous malformation, and hepatopulmonary syndrome.² It can occur even if right-sided chamber pressures are normal, and several mechanisms of the underlying pathophysiology have been described.³

Platypnea-orthodeoxia syndrome can be triggered by an event that causes a spontaneous transient elevation of right atrial pressure and pulmonary hypertension, such as our patient’s acute pulmonary embolism. Increased right-to-left shunting occurs in an upright position, causing preferential redirection of flow from the inferior vena cava through the interatrial septum and the patent foramen ovale.⁴

Our patient was elderly and, like one in every four people in the world, she had had a patent foramen ovale since the day she was born. Never causing a problem, it had remained undiagnosed until complicated by platypnea-orthodeoxia syndrome after her recent pulmonary embolism.

Methemoglobinemia. Methemoglobin has a lower affinity for oxygen than normal hemoglobin. Elevations usually occur with medications such as anesthetics and nitrates and can be diagnosed through an elevated level on arterial blood gas testing.

Our patient did not have elevated methemoglobin on her blood gas measurements on admission; therefore, this is unlikely to be the diagnosis.

CASE CONCLUDED

Percutaneous closure of the patent foramen ovale with a 30-mm Amplatzer Cribriform Occluder brought significant improvement in our patient’s functional status and arterial oxygenation saturation, and 2 weeks later at follow-up she no longer needed  supplemental oxygen. TEE 6 months later showed an intact closure device and no interatrial shunting.

WHEN HYPOXEMIA DOES NOT RESPOND TO OXYGEN

In the intensive care unit, time is critical, and when hypoxia is refractory to high Fio₂, shunting should be considered.

In the acute-care setting, platypnea-orthodeoxia syndrome can be identified quickly by pulse oximetry and serial blood gas measurements in the upright and supine positions. A drop in arterial oxygenation in the upright position vs the supine position helps confirm the diagnosis.

Other conditions in the differential diagnosis of this syndrome include recurrent pulmonary embolism, acute respiratory distress syndrome, interstitial pulmonary fibrosis, intrapulmonary shunting due to arteriovenous malformation, and diaphragm paralysis due to neuromuscular disease.

In our patient, positional blood gas measurements demonstrated a significant drop in arterial oxygen saturation from the supine to the upright position, raising our suspicion of shunting. It helped narrow the differential diagnosis and guided our selection of additional diagnostic tests.

The initial chest radiograph in our patient was normal. TTE did not reveal shunting and showed a normal right ventricular systolic pressure. TTE with agitated saline also failed to reveal shunting. Because of suboptimal positioning and image quality, TTE may miss the shunting physiology, and that is why we proceeded to positional TEE, which can better evaluate the hemodynamic effects of positional changes on patent foramen ovale and shunting. 

MORE ABOUT PATENT FORAMEN OVALE

The prevalence of patent foramen ovale is estimated at 27% in the general population, but it is usually not symptomatic. It can be associated with atrial septal aneurysm and Chiari network malformations. When associated with atrial septal aneurysm, it carries a higher risk of stroke.⁵

Our patient had a large atrial septal aneurysm with a septal excursion of 2 cm as well as a history of thromboembolic stroke, which was likely associated with the patent foramen ovale and the atrial septal aneurysm.

Atrial septal aneurysm is rare, with a prevalence of 1% at autopsy and 1.9% by TTE. It is defined by a septal excursion of at least 10 mm and a base diameter of at least 15 mm and is more frequently detected on TEE than on TTE.⁶

Studies have shown that contrast and color Doppler TEE are superior to TTE for detecting patent foramen ovale.⁷ Tilt-table TEE  with contrast enhancement has also been used to better demonstrate the morphology of the interatrial septum and the degree of shunting due to the separation between the septum primum and septum secundum causing the patent foramen ovale.⁸ Contrast-enhanced transcranial Doppler has also been shown comparable to contrast TEE to detect interatrial shunting. However, TEE provides additional anatomic information.⁹

In our patient, atrial septal aneurysm and patent foramen ovale were exaggerated by upright positioning, which opened the aneurysm and increased the shunting through the patent foramen ovale.

The treatment of choice in symptomatic patients with platypnea-orthodeoxia syndrome is directed at the underlying cause, in this case closure of the foramen ovale. This treatment has been shown to be safe and effective in these patients,¹⁰ but caution should be used when considering foramen ovale closure in patients with pulmonary hypertension.¹¹

In patients with irreversible or severe pulmonary hypertension, closure of the patent foramen ovale can exacerbate right heart dysfunction and lead to right heart failure. There are situations when closure of a patent foramen ovale can be considered in pulmonary hypertension; however, each decision is individualized, and caution must be used. A detailed discussion is beyond the scope of this paper.

A thorough history and physical examination are important in differentiating the various causes of hypoxemia. Appropriate diagnostic testing is needed along with prompt treatment of the underlying cause of platyp­nea-orthodeoxia syndrome.

An 85-year-old woman was brought to our intensive care unit because of worsening hypoxemia over the past day. About 3 weeks earlier she had been diagnosed with acute bilateral pulmonary emboli in the distal branches of the left and right lower lobes and right middle lobe, for which she was receiving anticoagulation therapy.

At presentation she had generalized fatigue and dyspnea at rest that was worse with exertion, but she denied having fever, chest pain, or cough. Her medical history included hypertension, hyperlipidemia, hypothyroidism, stage 1 breast cancer in remission, thromboembolic stroke, and myasthenia gravis. Before her hospital admission, she had been taking rosuvastatin, metoprolol tartrate, pyridostigmine, prednisone, furosemide, levothyroxine, and rivaroxaban. She did not smoke, she was retired, and she had not traveled recently.

Her blood pressure was 135/66 mm Hg, pulse 73 beats per minute, respiratory rate 16, temperature 35.4ºC (95.7ºF), and oxygen saturation 88% while receiving oxygen at 6 L/min via nasal cannula. Physical examination revealed mild edema in the lower extremities and basilar decreased breath sounds. She had no finger clubbing or cyanosis and was not using accessory muscles to breathe. Of note, her oxygen saturation remained more than 93% when she was recumbent but sharply dropped to less than 85% when she was upright.

Laboratory values

Results of initial laboratory testing were as follows:

• Sodium 138 mmol/L (reference range 132–148)
• Potassium 4.2 mmol/L (3.5–5.0)
• Chloride 99 mmol/L (98–111)
• Bicarbonate 29 mmol/L (23–32)
• Creatinine 0.52 mg/dL (0.7–1.4).
• White blood cell count 11.06 × 10⁹/L (3.7–11.0)
• Hemoglobin 12.6 g/dL (12–16)
• Platelet count 211 × 10⁹/L (150–400).
• International normalized ratio 1.4.

Electrocardiography and imaging studies

Standard 12-lead electrocardiography  showed normal sinus rhythm with left axis deviation and left ventricular hypertrophy.

Chest radiography showed bilateral interstitial opacities and small pleural effusions.

Computed tomography (CT) of the chest with contrast, compared with a CT scan done 20 days earlier, showed that the pulmonary emboli had resolved.

Arterial blood gases

In view of her positional hypoxemia, blood for arterial blood gas measurements was drawn in the supine and upright positions.

Supine, with a fraction of inspired oxygen (Fio₂) via high-flow nasal cannula of 45%, her values were:

• pH 7.45 (reference range 7.35–7.45)
• Pco₂ 34 mm Hg (36–46)
• Po₂ 81 mm Hg (85–95)
• Bicarbonate 23 mmol/L (22–26).

Upright, her hypoxemia was significantly worse:

• pH 7.46
• Pco₂ 33 mm Hg
• Po₂ 57 mm Hg
• Bicarbonate 23 mmol/L.

The methemoglobin level was normal on both measurements.

During her stay in the intensive care unit, she required up to 100% Fio₂ because of persistent hypoxemia.

CAUSES OF HYPOXEMIA

1. So far, the patient’s laboratory tests and imaging studies point to which of the following as the most likely cause of her severe hypoxemia?

• Ventilation-perfusion (V/Q) mismatch
• Diffusion abnormality
• Hypoventilation
• Shunting
• None of the above

The arterial blood gas measurements suggested the possibility of shunting as the cause, although further imaging would be needed to confirm that.

V/Q mismatch can occur in respiratory failure due to pulmonary embolism, pulmonary edema, or shunting. If ventilation is preserved but perfusion is impaired, the V/Q ratio approaches infinity (dead-space ventilation), a situation that can be seen in pulmonary embolism. If perfusion is preserved and ventilation is impaired, the V/Q ratio approaches zero, which is consistent with a physiologic shunt.

Hypoxemia may improve in less severe forms of V/Q mismatch. In our patient, the repeat CT with contrast showed that her pulmonary embolism had resolved, so this is probably not the cause of her severe hypoxemia.

Diffusion abnormalities are due to defects in the lung parenchyma, such as in chronic obstructive pulmonary disease, interstitial lung disease, and lung fibrosis.

Hypoxemia from diffusion defects is usually aggravated by a precipitating factor that increases oxygen demand, and it usually improves with oxygen supplementation. This is unlikely in our patient, as she did not have a history of chronic interstitial lung disease and CT showed no evidence of severe lung parenchymal disease.

Hypoventilation is usually due to drugs that cause respiratory depression, to stroke, or to neuromuscular diseases such as myasthenia gravis that can cause respiratory muscle weakness. It results in elevation of Pco₂ and, if not corrected, respiratory acidosis.

Our patient had a diagnosis of myasthenia gravis, though hypoventilation is unlikely in her case because she had a normal respiratory rate and low Pco₂ values.

Shunting can be physiologic or anatomic and can occur in the heart or the lungs. In physiologic shunting, severe V/Q mismatch can occur when ventilation is affected, as in severe pulmonary edema and pneumonia. In anatomic shunting, a defect such as an atrial septal defect or a pulmonary arteriovenous malformation allows blood to bypass areas of ventilation from the venous to the arterial circulation, preventing it from being oxygenated. In true anatomic shunting, supplemental oxygen with 100% Fio₂ has little effect, whereas in V/Q mismatch it can raise the arterial oxygen saturation.

Our patient’s radiograph did not suggest severe pneumonia or pulmonary edema, which makes these unlikely causes of her hypoxemia. At this point, because of her positional hypoxemia, further evaluation with contrast-enhanced echocardiography was needed to evaluate for anatomic shunting in the heart or lungs.

FURTHER TESTING

Transthoracic echocardiography (TTE) with agitated saline with a Valsalva maneuver was performed. Normally, no bubbles are seen in  the left-sided chambers after intravenous injection of agitated saline contrast, whereas bubbles on the left side suggest an intracardiac or intrapulmonary shunt. In our patient, this test was negative, and her right ventricular systolic pressure was normal.

2. What further testing should be considered to evaluate our patient’s hypoxemia?

• High-resolution chest CT
• Transesophageal echocardiography (TEE)
• Pulmonary function testing
• Repeated arterial blood gas measurement
• Edrophonium testing

Repeat imaging with high-resolution CT would likely not provide additional information and would expose the patient to additional radiation without adding much clinical benefit.

TEE could help further evaluate the intracardiac anatomy and look for shunting, which may be missed on TTE because of suboptimal positioning or image quality.

Pulmonary function testing is useful in establishing the baseline function and impairment in respiratory volumes. If an acute myasthenic crisis is suspected, measuring the negative inspiratory force and the forced vital capacity can be useful in monitoring worsening respiratory muscle weakness and assessing the need for mechanical ventilation.

In our patient, it is unlikely that pulmonary function testing would help, since her acute respiratory failure was probably not caused by neuromuscular weakness.

Repeated arterial blood gas measurement would likely only confirm that the patient still has positional hypoxemia but would not help sort through the differential diagnosis.

Edrophonium testing is useful in diagnosing myasthenia gravis and differentiating it from other neuromuscular diseases, such as Lambert-Eaton syndrome. Edrophonium, a reversible acetylcholinesterase inhibitor, prevents degradation of acetylcholine and prolongs its effect at the synaptic cleft, thus improving muscle weakness.

Our patient has already been diagnosed with myasthenia gravis, so this test is not likely to uncover the cause of her hypoxemia.

Because we still strongly suspected a shunt, TEE was performed with intravenous injection of agitated saline. TEE with the patient upright revealed intracardiac right-to-left shunting through a patent foramen ovale. The midesophageal view after saline injection showed a large interatrial septal aneurysm with total excursion of 2 cm, and right-to-left shunting within the first beat, consistent with an intracardiac shunt (Figure 1). Color Doppler imaging (Figure 2) demonstrated turbulent flow through the patent foramen ovale, consistent with right-to-left shunting, and also showed the patent foramen ovale in a closed position (Figure 3).

3. Which is now most likely the cause of our patient’s hypoxemia?

• Chronic thromboembolic pulmonary hypertension
• Myasthenic crisis
• Platypnea-orthodeoxia syndrome due to the patent foramen ovale
• Methemoglobinemia

Chronic thromboembolic pulmonary hypertension is usually a long-term result of untreated or inadequately treated thromboembolic disease (eg, pulmonary emboli), which causes vascular remodeling and pulmonary arteriopathy, which in turn leads to increased pulmonary vascular resistance and pulmonary hypertension.

This is unlikely the cause of our patient’s acute hypoxemia, as her symptoms did not suggest it. Moreover, an elevated right ventricular systolic pressure on TTE would suggest pulmonary hypertension, but TTE did not show this, and repeat chest CT indicated that her pulmonary embolism had been adequately treated and had resolved. A V/Q scan and right heart catheterization would help rule out chronic thromboembolic pulmonary hypertension, although these were not done in our patient.

Myasthenic crisis is the progressive fatiguing and paralysis of respiratory muscles ultimately requiring mechanical ventilation to sustain life. It is often brought on by infection or drug therapy.

Our patient did not require intubation and she had no signs or symptoms of myasthenic crisis such as ptosis, dysphagia, or dysarthria. She had a negative inspiratory force of −21 cm H₂O, and pulmonary function testing 4 days before her hospital admission had shown a forced vital capacity of 1.84 L, making myasthenic crisis an unlikely cause of her respiratory failure.

Platypnea-orthodeoxia syndrome is a syndrome of dyspnea (platypnea) and hypoxemia (orthodeoxia) that is induced by sitting upright or standing and resolves when lying down. It is a result of right-to-left intracardiac or intrapulmonary shunting in the presence of an anatomic defect and a functional element causing redirection of shunt flow through the anatomic defect in an upright position.¹ It is associated with specific cardiac, pulmonary, and hepatic diseases, such as atrial septal defect, pulmonary arteriovenous malformation, and hepatopulmonary syndrome.² It can occur even if right-sided chamber pressures are normal, and several mechanisms of the underlying pathophysiology have been described.³

Platypnea-orthodeoxia syndrome can be triggered by an event that causes a spontaneous transient elevation of right atrial pressure and pulmonary hypertension, such as our patient’s acute pulmonary embolism. Increased right-to-left shunting occurs in an upright position, causing preferential redirection of flow from the inferior vena cava through the interatrial septum and the patent foramen ovale.⁴

Our patient was elderly and, like one in every four people in the world, she had had a patent foramen ovale since the day she was born. Never causing a problem, it had remained undiagnosed until complicated by platypnea-orthodeoxia syndrome after her recent pulmonary embolism.

Methemoglobinemia. Methemoglobin has a lower affinity for oxygen than normal hemoglobin. Elevations usually occur with medications such as anesthetics and nitrates and can be diagnosed through an elevated level on arterial blood gas testing.

Our patient did not have elevated methemoglobin on her blood gas measurements on admission; therefore, this is unlikely to be the diagnosis.

CASE CONCLUDED

Percutaneous closure of the patent foramen ovale with a 30-mm Amplatzer Cribriform Occluder brought significant improvement in our patient’s functional status and arterial oxygenation saturation, and 2 weeks later at follow-up she no longer needed  supplemental oxygen. TEE 6 months later showed an intact closure device and no interatrial shunting.

WHEN HYPOXEMIA DOES NOT RESPOND TO OXYGEN

In the intensive care unit, time is critical, and when hypoxia is refractory to high Fio₂, shunting should be considered.

In the acute-care setting, platypnea-orthodeoxia syndrome can be identified quickly by pulse oximetry and serial blood gas measurements in the upright and supine positions. A drop in arterial oxygenation in the upright position vs the supine position helps confirm the diagnosis.

Other conditions in the differential diagnosis of this syndrome include recurrent pulmonary embolism, acute respiratory distress syndrome, interstitial pulmonary fibrosis, intrapulmonary shunting due to arteriovenous malformation, and diaphragm paralysis due to neuromuscular disease.

In our patient, positional blood gas measurements demonstrated a significant drop in arterial oxygen saturation from the supine to the upright position, raising our suspicion of shunting. It helped narrow the differential diagnosis and guided our selection of additional diagnostic tests.

The initial chest radiograph in our patient was normal. TTE did not reveal shunting and showed a normal right ventricular systolic pressure. TTE with agitated saline also failed to reveal shunting. Because of suboptimal positioning and image quality, TTE may miss the shunting physiology, and that is why we proceeded to positional TEE, which can better evaluate the hemodynamic effects of positional changes on patent foramen ovale and shunting. 

MORE ABOUT PATENT FORAMEN OVALE

The prevalence of patent foramen ovale is estimated at 27% in the general population, but it is usually not symptomatic. It can be associated with atrial septal aneurysm and Chiari network malformations. When associated with atrial septal aneurysm, it carries a higher risk of stroke.⁵

Our patient had a large atrial septal aneurysm with a septal excursion of 2 cm as well as a history of thromboembolic stroke, which was likely associated with the patent foramen ovale and the atrial septal aneurysm.

Atrial septal aneurysm is rare, with a prevalence of 1% at autopsy and 1.9% by TTE. It is defined by a septal excursion of at least 10 mm and a base diameter of at least 15 mm and is more frequently detected on TEE than on TTE.⁶

Studies have shown that contrast and color Doppler TEE are superior to TTE for detecting patent foramen ovale.⁷ Tilt-table TEE  with contrast enhancement has also been used to better demonstrate the morphology of the interatrial septum and the degree of shunting due to the separation between the septum primum and septum secundum causing the patent foramen ovale.⁸ Contrast-enhanced transcranial Doppler has also been shown comparable to contrast TEE to detect interatrial shunting. However, TEE provides additional anatomic information.⁹

In our patient, atrial septal aneurysm and patent foramen ovale were exaggerated by upright positioning, which opened the aneurysm and increased the shunting through the patent foramen ovale.

The treatment of choice in symptomatic patients with platypnea-orthodeoxia syndrome is directed at the underlying cause, in this case closure of the foramen ovale. This treatment has been shown to be safe and effective in these patients,¹⁰ but caution should be used when considering foramen ovale closure in patients with pulmonary hypertension.¹¹

In patients with irreversible or severe pulmonary hypertension, closure of the patent foramen ovale can exacerbate right heart dysfunction and lead to right heart failure. There are situations when closure of a patent foramen ovale can be considered in pulmonary hypertension; however, each decision is individualized, and caution must be used. A detailed discussion is beyond the scope of this paper.

A thorough history and physical examination are important in differentiating the various causes of hypoxemia. Appropriate diagnostic testing is needed along with prompt treatment of the underlying cause of platyp­nea-orthodeoxia syndrome.

An 85-year-old woman was brought to our intensive care unit because of worsening hypoxemia over the past day. About 3 weeks earlier she had been diagnosed with acute bilateral pulmonary emboli in the distal branches of the left and right lower lobes and right middle lobe, for which she was receiving anticoagulation therapy.

At presentation she had generalized fatigue and dyspnea at rest that was worse with exertion, but she denied having fever, chest pain, or cough. Her medical history included hypertension, hyperlipidemia, hypothyroidism, stage 1 breast cancer in remission, thromboembolic stroke, and myasthenia gravis. Before her hospital admission, she had been taking rosuvastatin, metoprolol tartrate, pyridostigmine, prednisone, furosemide, levothyroxine, and rivaroxaban. She did not smoke, she was retired, and she had not traveled recently.

Her blood pressure was 135/66 mm Hg, pulse 73 beats per minute, respiratory rate 16, temperature 35.4ºC (95.7ºF), and oxygen saturation 88% while receiving oxygen at 6 L/min via nasal cannula. Physical examination revealed mild edema in the lower extremities and basilar decreased breath sounds. She had no finger clubbing or cyanosis and was not using accessory muscles to breathe. Of note, her oxygen saturation remained more than 93% when she was recumbent but sharply dropped to less than 85% when she was upright.

Laboratory values

Results of initial laboratory testing were as follows:

• Sodium 138 mmol/L (reference range 132–148)
• Potassium 4.2 mmol/L (3.5–5.0)
• Chloride 99 mmol/L (98–111)
• Bicarbonate 29 mmol/L (23–32)
• Creatinine 0.52 mg/dL (0.7–1.4).
• White blood cell count 11.06 × 10⁹/L (3.7–11.0)
• Hemoglobin 12.6 g/dL (12–16)
• Platelet count 211 × 10⁹/L (150–400).
• International normalized ratio 1.4.

Electrocardiography and imaging studies

Standard 12-lead electrocardiography  showed normal sinus rhythm with left axis deviation and left ventricular hypertrophy.

Chest radiography showed bilateral interstitial opacities and small pleural effusions.

Computed tomography (CT) of the chest with contrast, compared with a CT scan done 20 days earlier, showed that the pulmonary emboli had resolved.

Arterial blood gases

In view of her positional hypoxemia, blood for arterial blood gas measurements was drawn in the supine and upright positions.

Supine, with a fraction of inspired oxygen (Fio₂) via high-flow nasal cannula of 45%, her values were:

• pH 7.45 (reference range 7.35–7.45)
• Pco₂ 34 mm Hg (36–46)
• Po₂ 81 mm Hg (85–95)
• Bicarbonate 23 mmol/L (22–26).

Upright, her hypoxemia was significantly worse:

• pH 7.46
• Pco₂ 33 mm Hg
• Po₂ 57 mm Hg
• Bicarbonate 23 mmol/L.

The methemoglobin level was normal on both measurements.

During her stay in the intensive care unit, she required up to 100% Fio₂ because of persistent hypoxemia.

CAUSES OF HYPOXEMIA

1. So far, the patient’s laboratory tests and imaging studies point to which of the following as the most likely cause of her severe hypoxemia?

• Ventilation-perfusion (V/Q) mismatch
• Diffusion abnormality
• Hypoventilation
• Shunting
• None of the above

The arterial blood gas measurements suggested the possibility of shunting as the cause, although further imaging would be needed to confirm that.

V/Q mismatch can occur in respiratory failure due to pulmonary embolism, pulmonary edema, or shunting. If ventilation is preserved but perfusion is impaired, the V/Q ratio approaches infinity (dead-space ventilation), a situation that can be seen in pulmonary embolism. If perfusion is preserved and ventilation is impaired, the V/Q ratio approaches zero, which is consistent with a physiologic shunt.

Hypoxemia may improve in less severe forms of V/Q mismatch. In our patient, the repeat CT with contrast showed that her pulmonary embolism had resolved, so this is probably not the cause of her severe hypoxemia.

Diffusion abnormalities are due to defects in the lung parenchyma, such as in chronic obstructive pulmonary disease, interstitial lung disease, and lung fibrosis.

Hypoxemia from diffusion defects is usually aggravated by a precipitating factor that increases oxygen demand, and it usually improves with oxygen supplementation. This is unlikely in our patient, as she did not have a history of chronic interstitial lung disease and CT showed no evidence of severe lung parenchymal disease.

Hypoventilation is usually due to drugs that cause respiratory depression, to stroke, or to neuromuscular diseases such as myasthenia gravis that can cause respiratory muscle weakness. It results in elevation of Pco₂ and, if not corrected, respiratory acidosis.

Our patient had a diagnosis of myasthenia gravis, though hypoventilation is unlikely in her case because she had a normal respiratory rate and low Pco₂ values.

Shunting can be physiologic or anatomic and can occur in the heart or the lungs. In physiologic shunting, severe V/Q mismatch can occur when ventilation is affected, as in severe pulmonary edema and pneumonia. In anatomic shunting, a defect such as an atrial septal defect or a pulmonary arteriovenous malformation allows blood to bypass areas of ventilation from the venous to the arterial circulation, preventing it from being oxygenated. In true anatomic shunting, supplemental oxygen with 100% Fio₂ has little effect, whereas in V/Q mismatch it can raise the arterial oxygen saturation.

Our patient’s radiograph did not suggest severe pneumonia or pulmonary edema, which makes these unlikely causes of her hypoxemia. At this point, because of her positional hypoxemia, further evaluation with contrast-enhanced echocardiography was needed to evaluate for anatomic shunting in the heart or lungs.

FURTHER TESTING

Transthoracic echocardiography (TTE) with agitated saline with a Valsalva maneuver was performed. Normally, no bubbles are seen in  the left-sided chambers after intravenous injection of agitated saline contrast, whereas bubbles on the left side suggest an intracardiac or intrapulmonary shunt. In our patient, this test was negative, and her right ventricular systolic pressure was normal.

2. What further testing should be considered to evaluate our patient’s hypoxemia?

• High-resolution chest CT
• Transesophageal echocardiography (TEE)
• Pulmonary function testing
• Repeated arterial blood gas measurement
• Edrophonium testing

Repeat imaging with high-resolution CT would likely not provide additional information and would expose the patient to additional radiation without adding much clinical benefit.

TEE could help further evaluate the intracardiac anatomy and look for shunting, which may be missed on TTE because of suboptimal positioning or image quality.

Pulmonary function testing is useful in establishing the baseline function and impairment in respiratory volumes. If an acute myasthenic crisis is suspected, measuring the negative inspiratory force and the forced vital capacity can be useful in monitoring worsening respiratory muscle weakness and assessing the need for mechanical ventilation.

In our patient, it is unlikely that pulmonary function testing would help, since her acute respiratory failure was probably not caused by neuromuscular weakness.

Repeated arterial blood gas measurement would likely only confirm that the patient still has positional hypoxemia but would not help sort through the differential diagnosis.

Edrophonium testing is useful in diagnosing myasthenia gravis and differentiating it from other neuromuscular diseases, such as Lambert-Eaton syndrome. Edrophonium, a reversible acetylcholinesterase inhibitor, prevents degradation of acetylcholine and prolongs its effect at the synaptic cleft, thus improving muscle weakness.

Our patient has already been diagnosed with myasthenia gravis, so this test is not likely to uncover the cause of her hypoxemia.

Because we still strongly suspected a shunt, TEE was performed with intravenous injection of agitated saline. TEE with the patient upright revealed intracardiac right-to-left shunting through a patent foramen ovale. The midesophageal view after saline injection showed a large interatrial septal aneurysm with total excursion of 2 cm, and right-to-left shunting within the first beat, consistent with an intracardiac shunt (Figure 1). Color Doppler imaging (Figure 2) demonstrated turbulent flow through the patent foramen ovale, consistent with right-to-left shunting, and also showed the patent foramen ovale in a closed position (Figure 3).

3. Which is now most likely the cause of our patient’s hypoxemia?

• Chronic thromboembolic pulmonary hypertension
• Myasthenic crisis
• Platypnea-orthodeoxia syndrome due to the patent foramen ovale
• Methemoglobinemia

Chronic thromboembolic pulmonary hypertension is usually a long-term result of untreated or inadequately treated thromboembolic disease (eg, pulmonary emboli), which causes vascular remodeling and pulmonary arteriopathy, which in turn leads to increased pulmonary vascular resistance and pulmonary hypertension.

This is unlikely the cause of our patient’s acute hypoxemia, as her symptoms did not suggest it. Moreover, an elevated right ventricular systolic pressure on TTE would suggest pulmonary hypertension, but TTE did not show this, and repeat chest CT indicated that her pulmonary embolism had been adequately treated and had resolved. A V/Q scan and right heart catheterization would help rule out chronic thromboembolic pulmonary hypertension, although these were not done in our patient.

Myasthenic crisis is the progressive fatiguing and paralysis of respiratory muscles ultimately requiring mechanical ventilation to sustain life. It is often brought on by infection or drug therapy.

Our patient did not require intubation and she had no signs or symptoms of myasthenic crisis such as ptosis, dysphagia, or dysarthria. She had a negative inspiratory force of −21 cm H₂O, and pulmonary function testing 4 days before her hospital admission had shown a forced vital capacity of 1.84 L, making myasthenic crisis an unlikely cause of her respiratory failure.

Platypnea-orthodeoxia syndrome is a syndrome of dyspnea (platypnea) and hypoxemia (orthodeoxia) that is induced by sitting upright or standing and resolves when lying down. It is a result of right-to-left intracardiac or intrapulmonary shunting in the presence of an anatomic defect and a functional element causing redirection of shunt flow through the anatomic defect in an upright position.¹ It is associated with specific cardiac, pulmonary, and hepatic diseases, such as atrial septal defect, pulmonary arteriovenous malformation, and hepatopulmonary syndrome.² It can occur even if right-sided chamber pressures are normal, and several mechanisms of the underlying pathophysiology have been described.³

Platypnea-orthodeoxia syndrome can be triggered by an event that causes a spontaneous transient elevation of right atrial pressure and pulmonary hypertension, such as our patient’s acute pulmonary embolism. Increased right-to-left shunting occurs in an upright position, causing preferential redirection of flow from the inferior vena cava through the interatrial septum and the patent foramen ovale.⁴

Our patient was elderly and, like one in every four people in the world, she had had a patent foramen ovale since the day she was born. Never causing a problem, it had remained undiagnosed until complicated by platypnea-orthodeoxia syndrome after her recent pulmonary embolism.

Methemoglobinemia. Methemoglobin has a lower affinity for oxygen than normal hemoglobin. Elevations usually occur with medications such as anesthetics and nitrates and can be diagnosed through an elevated level on arterial blood gas testing.

Our patient did not have elevated methemoglobin on her blood gas measurements on admission; therefore, this is unlikely to be the diagnosis.

CASE CONCLUDED

Percutaneous closure of the patent foramen ovale with a 30-mm Amplatzer Cribriform Occluder brought significant improvement in our patient’s functional status and arterial oxygenation saturation, and 2 weeks later at follow-up she no longer needed  supplemental oxygen. TEE 6 months later showed an intact closure device and no interatrial shunting.

WHEN HYPOXEMIA DOES NOT RESPOND TO OXYGEN

In the intensive care unit, time is critical, and when hypoxia is refractory to high Fio₂, shunting should be considered.

In the acute-care setting, platypnea-orthodeoxia syndrome can be identified quickly by pulse oximetry and serial blood gas measurements in the upright and supine positions. A drop in arterial oxygenation in the upright position vs the supine position helps confirm the diagnosis.

Other conditions in the differential diagnosis of this syndrome include recurrent pulmonary embolism, acute respiratory distress syndrome, interstitial pulmonary fibrosis, intrapulmonary shunting due to arteriovenous malformation, and diaphragm paralysis due to neuromuscular disease.

In our patient, positional blood gas measurements demonstrated a significant drop in arterial oxygen saturation from the supine to the upright position, raising our suspicion of shunting. It helped narrow the differential diagnosis and guided our selection of additional diagnostic tests.

The initial chest radiograph in our patient was normal. TTE did not reveal shunting and showed a normal right ventricular systolic pressure. TTE with agitated saline also failed to reveal shunting. Because of suboptimal positioning and image quality, TTE may miss the shunting physiology, and that is why we proceeded to positional TEE, which can better evaluate the hemodynamic effects of positional changes on patent foramen ovale and shunting. 

MORE ABOUT PATENT FORAMEN OVALE

The prevalence of patent foramen ovale is estimated at 27% in the general population, but it is usually not symptomatic. It can be associated with atrial septal aneurysm and Chiari network malformations. When associated with atrial septal aneurysm, it carries a higher risk of stroke.⁵

Our patient had a large atrial septal aneurysm with a septal excursion of 2 cm as well as a history of thromboembolic stroke, which was likely associated with the patent foramen ovale and the atrial septal aneurysm.

Atrial septal aneurysm is rare, with a prevalence of 1% at autopsy and 1.9% by TTE. It is defined by a septal excursion of at least 10 mm and a base diameter of at least 15 mm and is more frequently detected on TEE than on TTE.⁶

Studies have shown that contrast and color Doppler TEE are superior to TTE for detecting patent foramen ovale.⁷ Tilt-table TEE  with contrast enhancement has also been used to better demonstrate the morphology of the interatrial septum and the degree of shunting due to the separation between the septum primum and septum secundum causing the patent foramen ovale.⁸ Contrast-enhanced transcranial Doppler has also been shown comparable to contrast TEE to detect interatrial shunting. However, TEE provides additional anatomic information.⁹

In our patient, atrial septal aneurysm and patent foramen ovale were exaggerated by upright positioning, which opened the aneurysm and increased the shunting through the patent foramen ovale.

The treatment of choice in symptomatic patients with platypnea-orthodeoxia syndrome is directed at the underlying cause, in this case closure of the foramen ovale. This treatment has been shown to be safe and effective in these patients,¹⁰ but caution should be used when considering foramen ovale closure in patients with pulmonary hypertension.¹¹

In patients with irreversible or severe pulmonary hypertension, closure of the patent foramen ovale can exacerbate right heart dysfunction and lead to right heart failure. There are situations when closure of a patent foramen ovale can be considered in pulmonary hypertension; however, each decision is individualized, and caution must be used. A detailed discussion is beyond the scope of this paper.

A thorough history and physical examination are important in differentiating the various causes of hypoxemia. Appropriate diagnostic testing is needed along with prompt treatment of the underlying cause of platyp­nea-orthodeoxia syndrome.

Continued progress in understanding the pathophysiology of acute respiratory distress syndrome (ARDS) is translating into changes in the way we diagnose and manage it. Over the past 20 years, low tidal volume,¹ positive end-expiratory pressure (PEEP),² and fluid restriction³ have become the standard of care. A multidisciplinary approach, including targeted use of sedatives, early mobilization, and protocols for weaning from the ventilator, has also brought about substantial changes in ARDS management and its outcomes.^4–6

In this article, we review the most relevant articles about ARDS in the last 5 years. We include the new definition of ARDS and studies of ventilatory and nonventilatory therapies that have implications in managing patients with ARDS.

A STANDARDIZED APPROACH

ARDS is characterized by damage to the alveolar architecture, severe hypoxemia, and bilateral parenchymal opacities.

The working definition of ARDS developed in 1994 by the American-European Consensus Conference (AECC) was the basis for enrollment in most of the landmark trials and observational studies over the past 20 years.^7,8 However, it was limited in its reliability and validity.

An updated definition

In 2011, the ARDS Definition Task Force, using a novel consensus process, updated the ARDS definition,⁹ focusing on its feasibility, reliability, and validity in predicting response to therapies and outcomes in ARDS. This new “Berlin” definition is not substantially different from the old, but defines the criteria more specifically:

• Bilateral opacities, unexplained by nodules, atelectasis, or effusion, on chest radiography or computed tomography
• New or worsening respiratory symptoms, or a clinical insult associated with ARDS within 7 days of diagnosis
• Objective assessment of cardiac function (eg, with echocardiography) to exclude cardiogenic pulmonary edema
• Hypoxemia, with a partial pressure of arterial oxygen divided by the percentage of inspired oxygen (PaO₂/FiO₂ ratio) of 300 mm Hg or less despite noninvasive or invasive mechanical ventilation with PEEP or continuous positive airway pressure (CPAP) of at least 5 cm H₂O.

In addition, the new definition classifies the severity of disease on the basis of the degree of hypoxemia, ie, the PaO₂/FiO₂ ratio:

• Mild: PaO₂/FiO₂ ratio > 200 and ≤ 300 mm Hg
• Moderate: PaO₂/FiO₂ ratio > 100 and ≤ 200 mm Hg
• Severe: PaO₂/FiO₂ ratio ≤ 100 mm Hg.

The term “acute lung injury” has been eliminated, as has the previous criterion of a pulmonary artery wedge pressure of 18 mm Hg or less.

The panel also evaluated four ancillary variables for predicting outcomes in severe ARDS:

• Compliance of the respiratory system less than or equal to 40 mL/cm H₂O
• Radiographic severity (involvement of three or four quadrants on chest radiography)
• PEEP of 10 cm H₂O or greater
• Corrected expired volume 10 L/min or greater.

The task force evaluated the reliability and validity of this definition in a meta-analysis of 4,400 patients previously enrolled in randomized controlled trials or observational studies.

Findings. The Berlin definition predicted the risk of death better than the AECC definition. The mortality rate increased with the severity of ARDS, from 27% with mild disease to 32% with moderate disease to 45% with severe disease. The four ancillary variables did not contribute to the predictive validity of severe ARDS for mortality and were removed from the definition.

Thille et al¹⁰ retrospectively reviewed autopsy findings from 712 patients and found that the new definition identified a homogeneous group who had severe ARDS.¹⁰

Conclusions. The new definition may overcome some of the limitations of the old one, but it needs to be validated in clinical practice, especially its ability to predict death.

VENTILATORY SUPPORT

Prompt recognition, lung-protective ventilation, and a conservative fluid strategy remain the cornerstones of ARDS management. However, other strategies are being tested.

Prone-position ventilation in severe ARDS: The right therapy in a specific population

Prone-position ventilation was first described almost 30 years ago, but it has been used inconsistently in clinical practice.

Physiologic and observational studies indicated that prone positioning might improve survival in patients with ARDS, but several randomized trials failed to demonstrate any positive effect on outcomes.^11,12 Some trials also reported a higher rate of complications with this intervention.¹³ However, meta-analyses suggested that prone-position ventilation might have a beneficial effect in patients with severe ARDS (defined as a PaO₂/FiO₂ ratio ≤ 100 mm Hg).¹⁴

In view of these findings, investigators conducted a trial of prone-position ventilation exclusively in patients with severe ARDS.

The PROSEVA study

The Proning Severe ARDS Patients (PROSEVA) study was a randomized controlled trial designed to determine whether prone-position ventilation, applied early, would improve outcomes in patients with severe ARDS.¹⁵

In PROSEVA, 466 patients with severe ARDS (defined as a PaO₂/FiO₂ ratio < 150 mm Hg, FiO₂ ≥ 60%, and PEEP ≥ 5 cm H₂O) underwent either at least 16 hours of prone positioning or were left in the supine position after 12 to 24 hours of initial conventional mechanical ventilation. The patients were recruited from centers in France and Spain where prone-position ventilation had been used in daily practice for more than 5 years.

The primary outcome studied was the rate of death at 28 days. The secondary end points were the death rate at day 90, rates of successful extubation, the length of stay in the intensive care unit, and complications.

Findings. At study entry, the patients in the supine group were sicker, more of them required a vasopressor, and fewer of them were receiving neuromuscular blocking agents than those in the prone group. These baseline differences may have influenced the outcomes; the unadjusted 28-day mortality rate was 16.0% in the prone group compared with 32.8% in the supine group (P < .001). However, the hazard ratio for death with prone positioning was 0.39 (95% confidence interval [CI] 0.25–0.63) even after adjusting for severity and the use of vasopressors and neuromuscular blocking agents. Prone-position ventilation was not associated with a higher incidence of complications, and the rate of successful extubation was higher.

Conclusions. In patients with severe ARDS, early use of prolonged prone positioning significantly decreased the 28-day and 90-day mortality rates. This trial has made prone positioning one of the strategies in managing patients with early severe ARDS. To minimize complications such as pressure ulcers and line or tube dislodgement, personnel caring for these patients must follow a protocol and undergo specific training.

These results were corroborated by a meta-analysis by Beitler et al¹⁶ that found a significant decrease in mortality rate with prone-position ventilation even in older studies when lung-protective ventilation strategies were separated from high-tidal-volume ventilation.

High-frequency oscillatory ventilation: No benefit in two trials

Observational data and experimental studies suggested that high-frequency oscillatory ventilation (HFOV) is superior to conventional mechanical ventilation in ARDS patients.^17,18 However, outdated and cumbersome equipment, lack of protocols, and a lack of high-quality evidence led to limited and inconsistent use of HFOV, mainly as a rescue therapy in ARDS.¹⁹

Over the last few years, HFOV has been gaining acceptance, especially earlier in the course of ARDS.²⁰ After preliminary clinical trials reported promising results, two trials conducted in Canada and the United Kingdom compared HFOV vs conventional mechanical ventilation in patients with ARDS.

The OSCAR study

The Oscillation in ARDS (OSCAR) study²¹ was a “pragmatic” trial²² (ie, it had minimal exclusion criteria) of the safety and effectiveness of HFOV as a primary ventilatory strategy for ARDS. It included 795 patients randomized to receive conventional ventilation (n = 397) or HFOV (n = 398). Research centers followed detailed algorithms for HFOV management and adopted their usual practice for conventional ventilation. Medical care was given according to the clinician’s judgment.

The primary outcome studied was survival at 30 days. The secondary outcomes were all-cause mortality in the intensive care unit and the hospital, duration of mechanical ventilation, and use of antimicrobial, sedative, vasoactive, and neuromuscular-blocking drugs.

Findings. The patient baseline characteristics were similar in both groups.

There was no significant difference in intensive care unit mortality rates, hospital mortality rates, or mortality rates at 30 days (41.7% in the HFOV group vs 41.1% in the conventional ventilation group; P = .85, 95% CI 6.1–7.5) even after adjustments for center or severity of illness.

The duration of mechanical ventilation was similar in both groups (14.9 ± 13.3 days in the HFOV group vs 14.1 ± 13.4 days in the conventional ventilation group, P = .41). However, sedatives and neuromuscular-blocking drugs were used more often and longer in the HFOV group than in the conventional ventilation group. There was no difference in the use of vasoactive or antimicrobial medications.

Conclusions. This multicenter randomized control trial did not demonstrate any benefit from using HFOV for routine management of ARDS. Its pragmatic design made it less likely to reach a firm conclusion,²² but it at least made a case against routinely using HFOV in patients with ARDS.

The OSCILLATE study

The Oscillation for Acute Respiratory Distress Syndrome Treated Early (OSCILLATE) study²³ assessed the safety and efficacy of HFOV as a treatment for early-onset moderate-to-severe ARDS.

The inclusion criteria were similar to those in the OSCAR trial except that pulmonary symptoms had to be present less than 2 weeks and ARDS assessment was done under standard ventilator settings. As this was an efficacy trial, it had more exclusion criteria than the OSCAR trial. A total of 548 patients were randomized to receive conventional ventilation (n = 273) or HFOV (n = 275). The baseline characteristics were similar between groups.

Conventional ventilation was given according to a protocol used in an earlier trial² and included recruitment maneuvers. HFOV was given in centers that had experience in this treatment, and there were protocols for ventilation management, hemodynamic optimization, and weaning. All other care was left to the clinician’s choice.

The primary outcome studied was in-hospital mortality. The investigators also evaluated whether there were interactions between the treatment and baseline severity of lung injury and center experience with HFOV.

Findings. The trial was stopped after an interim analysis found that HFOV might be harmful, although the statistical threshold for stopping was not reached. The in-hospital mortality rate was 47% in the HFOV group and 35% in the control group (relative risk of death with HFOV 1.33, 95% CI 1.09–1.64, P = .005). HFOV was worse than conventional ventilation regardless of the severity of disease or center experience. The HFOV group had higher mean airway pressures but similar FiO₂ compared with the conventional ventilation group.

The HFOV group received significantly more vasopressors, sedatives, and neuromuscular blockers. This group’s fluid balance was higher as well, but not significantly so. Refractory hypoxemia (defined as PaO₂ < 60 mm Hg for 1 hour with an FiO₂ of 1.0 and neuromuscular blockade) was more frequent in the conventional ventilation group, but the number of deaths in the subgroup with refractory hypoxemia was similar with either treatment.

Conclusions. This multicenter randomized controlled trial demonstrated that HFOV was harmful when used routinely to manage ARDS. The trial’s protocol was based on the results of a pilot study carried out by the same investigators, which provided the best evidence available regarding the safety of HFOV at that time.

The results of the OSCAR and OSCILLATE trials have quelled enthusiasm for early, routine use of HFOV in ARDS. Although there are concerns that the protocol (ie, the way HFOV was implemented) rather than HFOV itself may have led to worse outcomes, there is no signal to support its routine use. We need further studies to define if it remains a viable rescue therapy.

Extracorporeal membrane oxygenation: Is it a viable option in severe ARDS?

Extracorporeal membrane oxygenation (ECMO) uses cardiopulmonary bypass technology to provide gas exchange. In patients with severe hypoxemia, ECMO can ensure adequate oxygenation and ventilation while ensuring the optimization of lung-protective ventilation. But ECMO was never as successful in adults with ARDS as it was in children and neonates.²⁴

The first two trials of ECMO in ARDS^24,25 reported equal or worse survival rates compared with conventional ventilation, and the overall mortality rate in these studies was staggeringly high. However, these studies were carried out before the era of lung-protective ventilation and at a time when ECMO technology was relatively primitive.

With new technology such as venovenous circuits and smaller cannulas, ECMO has gained more acceptance. It was used in patients with severe or refractory hypoxemia associated with ARDS during the H1N1 pandemic.^26,27

The CESAR trial

The Conventional Ventilatory Support Versus Extracorporeal Membrane Oxygenation for Severe Adult Respiratory Failure (CESAR) trial²⁸ assessed the safety, clinical efficacy, and cost-effectiveness of ECMO in managing severe ARDS. It compared best standard practice vs a protocol that included ECMO. The trial was conducted from 2001 to 2006.

Patients with severe ARDS, as defined by a Murray score²⁹ greater than 3 or uncompensated hypercapnea, were prospectively randomized and recruited from an ECMO center and 148 tertiary intensive care units and referral hospitals in England. This was a pragmatic trial, with minimal exclusion criteria (essentially, mechanical ventilation with high pressures and high FiO₂ for more than 7 days, intracranial bleeding, or contraindication to heparinization).

A total of 180 patients were randomized in a one-to-one ratio to receive ECMO or conventional management. The ventilator management in the conventional treatment group was not done according to a protocol but in general was low-volume and low-pressure. All patients randomized to ECMO were transferred to the ECMO center and treated according to a standardized ventilation protocol. After 12 hours, if predefined goals were not reached, venovenous ECMO was started. Patients assigned to conventional management could not cross over to ECMO.

The primary outcomes were death or severe disability at 6 months after randomization, and cost-effectiveness. The secondary outcomes were hospital resource use (eg, rescue techniques, length of stay, duration of ECMO) and health status after 6 months.

Findings. The groups were similar at baseline. Sixty-eight (75%) of the 90 patients randomized to receive ECMO actually received it. Of the 22 patients who did not receive ECMO, 16 (18% of the 90) improved on conventional therapy, 5 (6%) died during or before transfer, and 1 could not receive heparin.

Two patients had severe complications in the ECMO group: one had an arterial puncture, and one had an oxygen delivery failure during transport. In each case, these events contributed to the death of the patient.

More patients in the ECMO group received lung-protective ventilation, 84 (93%) vs 63 (70%).

The primary outcome, ie, death or severe disability at 6 months, occurred in 33 (37%) of the 90 patients in the ECMO group and in 46 (53%) of the patients in the conventional management group (relative risk 0.69, 95% CI 0.05–0.97, P = .03). More patients in the ECMO group survived, but the difference was not statistically significant (relative risk of death 0.73, 95% CI 0.52–1.03, P = .07). The most common cause of death in the ECMO group was multiorgan failure (42%), whereas in the conventional management group, the most common cause of death was respiratory failure (60%).

Length of stay in the hospital and in the critical care unit and health care costs were double for patients in the ECMO group. There was no difference in quality-of-life markers at 6 months in the survivors.

Conclusions. This pragmatic trial demonstrated that a protocol that includes ECMO could improve survival rates in ARDS.

Of note, the ECMO group got care in regional centers that used protocols. Therefore, in interpreting the results of this trial, we have to consider that being in a center with protocol-specified care for ARDS could drive some of the difference in mortality rates.

Regardless, this trial demonstrated that ECMO is feasible and led to better outcomes than expected. The findings were encouraging, and spurred the use of ECMO in severe ARDS during the 2009 H1N1 pandemic. Two propensity-matched studies and a number of case series reported a survival benefit associated with the use of ECMO in patients with severe ARDS.^27,30

A recent meta-analysis also reported that ECMO might lower the mortality rate in ARDS; however, the patients in the H1N1 pandemic were younger and usually had isolated respiratory failure.³¹

The success of ECMO has opened new possibilities in the management of ARDS. As the technology improves and our experience increases, ECMO will likely gain more acceptance as a treatment for severe ARDS.

Airway pressure release ventilation

The use of airway pressure release ventilation and other ventilator modalities in ARDS is not supported by current evidence, though results of clinical trials may influence our practice in the future.

PHARMACOTHERAPY IN ARDS

The pathogenesis of ARDS includes damage to the alveolar-capillary membrane, with leakage of protein-rich edema fluid into alveoli. This damage is propagated by a complex inflammatory response including but not limited to neutrophil activation, free-radical formation, dysregulation of the coagulation system, and extensive release of inflammatory mediators.^32,33 As a consequence, there are multiple potential targets for pharmacologic therapy in ARDS.

A variety of drugs, including corticosteroids, anti-inflammatory agents, immune-modulating agents, pulmonary vasodilators, antioxidants, and surfactants, have been studied in patients with ARDS.³⁴ But effective pharmacotherapy for ARDS remains extremely limited.

Neuromuscular blockade in early severe ARDS

Mechanical ventilation can result in injurious stretching of the lung parenchyma, either from alveolar overdistention (volutrauma) or from continual recruitment and derecruitment of unstable lung units during the ventilator cycle (atelectrauma).³⁵ Ventilator-induced lung injury can be exacerbated by asynchronous breathing.

In theory, neuromuscular blockers could minimize patient-ventilator asynchrony and provide much better control of tidal volume and pressure in patients with ARDS. This may result in less volutrauma and atelectrauma associated with asynchronous breathing. Data also suggest that cisatracurium (Nimbex), a neuromuscular blocking agent, may have a direct effect on the amount of inflammation in lungs with ARDS.³⁶

The ACURASYS study

The ARDS et Curarisation Systématique (ACURASYS) study³⁷ was a randomized trial in 340 patients undergoing mechanical ventilation for severe ARDS to evaluate the impact of neuromuscular blockade within the first 48 hours in this population.

The primary outcome was the mortality rate before hospital discharge or within 90 days of study entry. Secondary outcomes included the 28-day mortality rate, the rate of intensive care unit-acquired paresis, and the number of ventilator-free days. To be included, patients had to have been mechanically ventilated for less than 48 hours and to meet the AECC criteria for severe ARDS, with a PaO₂/FiO₂ ratio less than 150 mm Hg.

The intervention group received a continuous infusion of cisatracurium for 48 hours, while the control patients received placebo. Muscle strength was evaluated by clinical scoring of strength in different muscle groups.

Findings. The study groups were similar at baseline.

The crude 90-day mortality rate was lower in the cisatracurium group (31.6% vs 40.7%, P = .08). Regression analysis showed an improved 90-day survival rate with the use of this neuromuscular blocker after adjustment for severity of illness and the severity of ARDS (based on degree of hypoxemia and plateau pressures) (hazard ratio for death at 90 days 0.68; 95% CI 0.48–0.98; P = .04). The rate of paresis acquired in the intensive care unit did not differ significantly between the two groups.

Conclusion. In patients with severe ARDS, giving a neuromuscular blocking agent early improved the survival rate and increased the time off the ventilator without increasing muscle weakness.

These data are in line with similar findings from two other studies published by the same group.^38,39 A meta-analysis of 432 patients showed that the use of neuromuscular blockade in early severe ARDS is associated with a statistically significant effect on early mortality (relative risk 0.66, 95% CI 0.50–0.87).⁴⁰ The pooled analysis of these trials did not show any statistically significant critical-illness polyneuropathy.

These results need to be interpreted carefully, as we have inadequate data to see if they generalize to different intensive care units, and the evaluation and categorization of critical-illness polyneuropathy remains to be defined.

Cisatracurium is a promising treatment for moderate to severe ARDS and merits investigation in a large confirmatory randomized controlled trial.

Other pharmacologic agents

A number of other drugs have been studied in ARDS patients, including both inhaled and intravenous beta agonists,^41,42 statins,⁴³ and nutritional supplements.⁴⁴ But as with other drugs previously studied in ARDS such as corticosteroids, N-acetylcysteine, and surfactant,³⁴ these agents showed no effect on outcomes. In fact, a recent trial of intravenous salbutamol in ARDS patients was stopped after an interim analysis because of a higher incidence of arrhythmias and lactic acidosis with this agent.⁴²

These findings reaffirm that pharmacologic therapy needs to be carefully considered, and potential harms associated with these therapies need to be addressed before they are introduced in the care of critically ill patients.

Continued progress in understanding the pathophysiology of acute respiratory distress syndrome (ARDS) is translating into changes in the way we diagnose and manage it. Over the past 20 years, low tidal volume,¹ positive end-expiratory pressure (PEEP),² and fluid restriction³ have become the standard of care. A multidisciplinary approach, including targeted use of sedatives, early mobilization, and protocols for weaning from the ventilator, has also brought about substantial changes in ARDS management and its outcomes.^4–6

In this article, we review the most relevant articles about ARDS in the last 5 years. We include the new definition of ARDS and studies of ventilatory and nonventilatory therapies that have implications in managing patients with ARDS.

A STANDARDIZED APPROACH

ARDS is characterized by damage to the alveolar architecture, severe hypoxemia, and bilateral parenchymal opacities.

The working definition of ARDS developed in 1994 by the American-European Consensus Conference (AECC) was the basis for enrollment in most of the landmark trials and observational studies over the past 20 years.^7,8 However, it was limited in its reliability and validity.

An updated definition

In 2011, the ARDS Definition Task Force, using a novel consensus process, updated the ARDS definition,⁹ focusing on its feasibility, reliability, and validity in predicting response to therapies and outcomes in ARDS. This new “Berlin” definition is not substantially different from the old, but defines the criteria more specifically:

• Bilateral opacities, unexplained by nodules, atelectasis, or effusion, on chest radiography or computed tomography
• New or worsening respiratory symptoms, or a clinical insult associated with ARDS within 7 days of diagnosis
• Objective assessment of cardiac function (eg, with echocardiography) to exclude cardiogenic pulmonary edema
• Hypoxemia, with a partial pressure of arterial oxygen divided by the percentage of inspired oxygen (PaO₂/FiO₂ ratio) of 300 mm Hg or less despite noninvasive or invasive mechanical ventilation with PEEP or continuous positive airway pressure (CPAP) of at least 5 cm H₂O.

In addition, the new definition classifies the severity of disease on the basis of the degree of hypoxemia, ie, the PaO₂/FiO₂ ratio:

• Mild: PaO₂/FiO₂ ratio > 200 and ≤ 300 mm Hg
• Moderate: PaO₂/FiO₂ ratio > 100 and ≤ 200 mm Hg
• Severe: PaO₂/FiO₂ ratio ≤ 100 mm Hg.

The term “acute lung injury” has been eliminated, as has the previous criterion of a pulmonary artery wedge pressure of 18 mm Hg or less.

The panel also evaluated four ancillary variables for predicting outcomes in severe ARDS:

• Compliance of the respiratory system less than or equal to 40 mL/cm H₂O
• Radiographic severity (involvement of three or four quadrants on chest radiography)
• PEEP of 10 cm H₂O or greater
• Corrected expired volume 10 L/min or greater.

The task force evaluated the reliability and validity of this definition in a meta-analysis of 4,400 patients previously enrolled in randomized controlled trials or observational studies.

Findings. The Berlin definition predicted the risk of death better than the AECC definition. The mortality rate increased with the severity of ARDS, from 27% with mild disease to 32% with moderate disease to 45% with severe disease. The four ancillary variables did not contribute to the predictive validity of severe ARDS for mortality and were removed from the definition.

Thille et al¹⁰ retrospectively reviewed autopsy findings from 712 patients and found that the new definition identified a homogeneous group who had severe ARDS.¹⁰

Conclusions. The new definition may overcome some of the limitations of the old one, but it needs to be validated in clinical practice, especially its ability to predict death.

VENTILATORY SUPPORT

Prompt recognition, lung-protective ventilation, and a conservative fluid strategy remain the cornerstones of ARDS management. However, other strategies are being tested.

Prone-position ventilation in severe ARDS: The right therapy in a specific population

Prone-position ventilation was first described almost 30 years ago, but it has been used inconsistently in clinical practice.

Physiologic and observational studies indicated that prone positioning might improve survival in patients with ARDS, but several randomized trials failed to demonstrate any positive effect on outcomes.^11,12 Some trials also reported a higher rate of complications with this intervention.¹³ However, meta-analyses suggested that prone-position ventilation might have a beneficial effect in patients with severe ARDS (defined as a PaO₂/FiO₂ ratio ≤ 100 mm Hg).¹⁴

In view of these findings, investigators conducted a trial of prone-position ventilation exclusively in patients with severe ARDS.

The PROSEVA study

The Proning Severe ARDS Patients (PROSEVA) study was a randomized controlled trial designed to determine whether prone-position ventilation, applied early, would improve outcomes in patients with severe ARDS.¹⁵

In PROSEVA, 466 patients with severe ARDS (defined as a PaO₂/FiO₂ ratio < 150 mm Hg, FiO₂ ≥ 60%, and PEEP ≥ 5 cm H₂O) underwent either at least 16 hours of prone positioning or were left in the supine position after 12 to 24 hours of initial conventional mechanical ventilation. The patients were recruited from centers in France and Spain where prone-position ventilation had been used in daily practice for more than 5 years.

The primary outcome studied was the rate of death at 28 days. The secondary end points were the death rate at day 90, rates of successful extubation, the length of stay in the intensive care unit, and complications.

Findings. At study entry, the patients in the supine group were sicker, more of them required a vasopressor, and fewer of them were receiving neuromuscular blocking agents than those in the prone group. These baseline differences may have influenced the outcomes; the unadjusted 28-day mortality rate was 16.0% in the prone group compared with 32.8% in the supine group (P < .001). However, the hazard ratio for death with prone positioning was 0.39 (95% confidence interval [CI] 0.25–0.63) even after adjusting for severity and the use of vasopressors and neuromuscular blocking agents. Prone-position ventilation was not associated with a higher incidence of complications, and the rate of successful extubation was higher.

Conclusions. In patients with severe ARDS, early use of prolonged prone positioning significantly decreased the 28-day and 90-day mortality rates. This trial has made prone positioning one of the strategies in managing patients with early severe ARDS. To minimize complications such as pressure ulcers and line or tube dislodgement, personnel caring for these patients must follow a protocol and undergo specific training.

These results were corroborated by a meta-analysis by Beitler et al¹⁶ that found a significant decrease in mortality rate with prone-position ventilation even in older studies when lung-protective ventilation strategies were separated from high-tidal-volume ventilation.

High-frequency oscillatory ventilation: No benefit in two trials

Observational data and experimental studies suggested that high-frequency oscillatory ventilation (HFOV) is superior to conventional mechanical ventilation in ARDS patients.^17,18 However, outdated and cumbersome equipment, lack of protocols, and a lack of high-quality evidence led to limited and inconsistent use of HFOV, mainly as a rescue therapy in ARDS.¹⁹

Over the last few years, HFOV has been gaining acceptance, especially earlier in the course of ARDS.²⁰ After preliminary clinical trials reported promising results, two trials conducted in Canada and the United Kingdom compared HFOV vs conventional mechanical ventilation in patients with ARDS.

The OSCAR study

The Oscillation in ARDS (OSCAR) study²¹ was a “pragmatic” trial²² (ie, it had minimal exclusion criteria) of the safety and effectiveness of HFOV as a primary ventilatory strategy for ARDS. It included 795 patients randomized to receive conventional ventilation (n = 397) or HFOV (n = 398). Research centers followed detailed algorithms for HFOV management and adopted their usual practice for conventional ventilation. Medical care was given according to the clinician’s judgment.

The primary outcome studied was survival at 30 days. The secondary outcomes were all-cause mortality in the intensive care unit and the hospital, duration of mechanical ventilation, and use of antimicrobial, sedative, vasoactive, and neuromuscular-blocking drugs.

Findings. The patient baseline characteristics were similar in both groups.

There was no significant difference in intensive care unit mortality rates, hospital mortality rates, or mortality rates at 30 days (41.7% in the HFOV group vs 41.1% in the conventional ventilation group; P = .85, 95% CI 6.1–7.5) even after adjustments for center or severity of illness.

The duration of mechanical ventilation was similar in both groups (14.9 ± 13.3 days in the HFOV group vs 14.1 ± 13.4 days in the conventional ventilation group, P = .41). However, sedatives and neuromuscular-blocking drugs were used more often and longer in the HFOV group than in the conventional ventilation group. There was no difference in the use of vasoactive or antimicrobial medications.

Conclusions. This multicenter randomized control trial did not demonstrate any benefit from using HFOV for routine management of ARDS. Its pragmatic design made it less likely to reach a firm conclusion,²² but it at least made a case against routinely using HFOV in patients with ARDS.

The OSCILLATE study

The Oscillation for Acute Respiratory Distress Syndrome Treated Early (OSCILLATE) study²³ assessed the safety and efficacy of HFOV as a treatment for early-onset moderate-to-severe ARDS.

The inclusion criteria were similar to those in the OSCAR trial except that pulmonary symptoms had to be present less than 2 weeks and ARDS assessment was done under standard ventilator settings. As this was an efficacy trial, it had more exclusion criteria than the OSCAR trial. A total of 548 patients were randomized to receive conventional ventilation (n = 273) or HFOV (n = 275). The baseline characteristics were similar between groups.

Conventional ventilation was given according to a protocol used in an earlier trial² and included recruitment maneuvers. HFOV was given in centers that had experience in this treatment, and there were protocols for ventilation management, hemodynamic optimization, and weaning. All other care was left to the clinician’s choice.

The primary outcome studied was in-hospital mortality. The investigators also evaluated whether there were interactions between the treatment and baseline severity of lung injury and center experience with HFOV.

Findings. The trial was stopped after an interim analysis found that HFOV might be harmful, although the statistical threshold for stopping was not reached. The in-hospital mortality rate was 47% in the HFOV group and 35% in the control group (relative risk of death with HFOV 1.33, 95% CI 1.09–1.64, P = .005). HFOV was worse than conventional ventilation regardless of the severity of disease or center experience. The HFOV group had higher mean airway pressures but similar FiO₂ compared with the conventional ventilation group.

The HFOV group received significantly more vasopressors, sedatives, and neuromuscular blockers. This group’s fluid balance was higher as well, but not significantly so. Refractory hypoxemia (defined as PaO₂ < 60 mm Hg for 1 hour with an FiO₂ of 1.0 and neuromuscular blockade) was more frequent in the conventional ventilation group, but the number of deaths in the subgroup with refractory hypoxemia was similar with either treatment.

Conclusions. This multicenter randomized controlled trial demonstrated that HFOV was harmful when used routinely to manage ARDS. The trial’s protocol was based on the results of a pilot study carried out by the same investigators, which provided the best evidence available regarding the safety of HFOV at that time.

The results of the OSCAR and OSCILLATE trials have quelled enthusiasm for early, routine use of HFOV in ARDS. Although there are concerns that the protocol (ie, the way HFOV was implemented) rather than HFOV itself may have led to worse outcomes, there is no signal to support its routine use. We need further studies to define if it remains a viable rescue therapy.

Extracorporeal membrane oxygenation: Is it a viable option in severe ARDS?

Extracorporeal membrane oxygenation (ECMO) uses cardiopulmonary bypass technology to provide gas exchange. In patients with severe hypoxemia, ECMO can ensure adequate oxygenation and ventilation while ensuring the optimization of lung-protective ventilation. But ECMO was never as successful in adults with ARDS as it was in children and neonates.²⁴

The first two trials of ECMO in ARDS^24,25 reported equal or worse survival rates compared with conventional ventilation, and the overall mortality rate in these studies was staggeringly high. However, these studies were carried out before the era of lung-protective ventilation and at a time when ECMO technology was relatively primitive.

With new technology such as venovenous circuits and smaller cannulas, ECMO has gained more acceptance. It was used in patients with severe or refractory hypoxemia associated with ARDS during the H1N1 pandemic.^26,27

The CESAR trial

The Conventional Ventilatory Support Versus Extracorporeal Membrane Oxygenation for Severe Adult Respiratory Failure (CESAR) trial²⁸ assessed the safety, clinical efficacy, and cost-effectiveness of ECMO in managing severe ARDS. It compared best standard practice vs a protocol that included ECMO. The trial was conducted from 2001 to 2006.

Patients with severe ARDS, as defined by a Murray score²⁹ greater than 3 or uncompensated hypercapnea, were prospectively randomized and recruited from an ECMO center and 148 tertiary intensive care units and referral hospitals in England. This was a pragmatic trial, with minimal exclusion criteria (essentially, mechanical ventilation with high pressures and high FiO₂ for more than 7 days, intracranial bleeding, or contraindication to heparinization).

A total of 180 patients were randomized in a one-to-one ratio to receive ECMO or conventional management. The ventilator management in the conventional treatment group was not done according to a protocol but in general was low-volume and low-pressure. All patients randomized to ECMO were transferred to the ECMO center and treated according to a standardized ventilation protocol. After 12 hours, if predefined goals were not reached, venovenous ECMO was started. Patients assigned to conventional management could not cross over to ECMO.

The primary outcomes were death or severe disability at 6 months after randomization, and cost-effectiveness. The secondary outcomes were hospital resource use (eg, rescue techniques, length of stay, duration of ECMO) and health status after 6 months.

Findings. The groups were similar at baseline. Sixty-eight (75%) of the 90 patients randomized to receive ECMO actually received it. Of the 22 patients who did not receive ECMO, 16 (18% of the 90) improved on conventional therapy, 5 (6%) died during or before transfer, and 1 could not receive heparin.

Two patients had severe complications in the ECMO group: one had an arterial puncture, and one had an oxygen delivery failure during transport. In each case, these events contributed to the death of the patient.

More patients in the ECMO group received lung-protective ventilation, 84 (93%) vs 63 (70%).

The primary outcome, ie, death or severe disability at 6 months, occurred in 33 (37%) of the 90 patients in the ECMO group and in 46 (53%) of the patients in the conventional management group (relative risk 0.69, 95% CI 0.05–0.97, P = .03). More patients in the ECMO group survived, but the difference was not statistically significant (relative risk of death 0.73, 95% CI 0.52–1.03, P = .07). The most common cause of death in the ECMO group was multiorgan failure (42%), whereas in the conventional management group, the most common cause of death was respiratory failure (60%).

Length of stay in the hospital and in the critical care unit and health care costs were double for patients in the ECMO group. There was no difference in quality-of-life markers at 6 months in the survivors.

Conclusions. This pragmatic trial demonstrated that a protocol that includes ECMO could improve survival rates in ARDS.

Of note, the ECMO group got care in regional centers that used protocols. Therefore, in interpreting the results of this trial, we have to consider that being in a center with protocol-specified care for ARDS could drive some of the difference in mortality rates.

Regardless, this trial demonstrated that ECMO is feasible and led to better outcomes than expected. The findings were encouraging, and spurred the use of ECMO in severe ARDS during the 2009 H1N1 pandemic. Two propensity-matched studies and a number of case series reported a survival benefit associated with the use of ECMO in patients with severe ARDS.^27,30

A recent meta-analysis also reported that ECMO might lower the mortality rate in ARDS; however, the patients in the H1N1 pandemic were younger and usually had isolated respiratory failure.³¹

The success of ECMO has opened new possibilities in the management of ARDS. As the technology improves and our experience increases, ECMO will likely gain more acceptance as a treatment for severe ARDS.

Airway pressure release ventilation

The use of airway pressure release ventilation and other ventilator modalities in ARDS is not supported by current evidence, though results of clinical trials may influence our practice in the future.

PHARMACOTHERAPY IN ARDS

The pathogenesis of ARDS includes damage to the alveolar-capillary membrane, with leakage of protein-rich edema fluid into alveoli. This damage is propagated by a complex inflammatory response including but not limited to neutrophil activation, free-radical formation, dysregulation of the coagulation system, and extensive release of inflammatory mediators.^32,33 As a consequence, there are multiple potential targets for pharmacologic therapy in ARDS.

A variety of drugs, including corticosteroids, anti-inflammatory agents, immune-modulating agents, pulmonary vasodilators, antioxidants, and surfactants, have been studied in patients with ARDS.³⁴ But effective pharmacotherapy for ARDS remains extremely limited.

Neuromuscular blockade in early severe ARDS

Mechanical ventilation can result in injurious stretching of the lung parenchyma, either from alveolar overdistention (volutrauma) or from continual recruitment and derecruitment of unstable lung units during the ventilator cycle (atelectrauma).³⁵ Ventilator-induced lung injury can be exacerbated by asynchronous breathing.

In theory, neuromuscular blockers could minimize patient-ventilator asynchrony and provide much better control of tidal volume and pressure in patients with ARDS. This may result in less volutrauma and atelectrauma associated with asynchronous breathing. Data also suggest that cisatracurium (Nimbex), a neuromuscular blocking agent, may have a direct effect on the amount of inflammation in lungs with ARDS.³⁶

The ACURASYS study

The ARDS et Curarisation Systématique (ACURASYS) study³⁷ was a randomized trial in 340 patients undergoing mechanical ventilation for severe ARDS to evaluate the impact of neuromuscular blockade within the first 48 hours in this population.

The primary outcome was the mortality rate before hospital discharge or within 90 days of study entry. Secondary outcomes included the 28-day mortality rate, the rate of intensive care unit-acquired paresis, and the number of ventilator-free days. To be included, patients had to have been mechanically ventilated for less than 48 hours and to meet the AECC criteria for severe ARDS, with a PaO₂/FiO₂ ratio less than 150 mm Hg.

The intervention group received a continuous infusion of cisatracurium for 48 hours, while the control patients received placebo. Muscle strength was evaluated by clinical scoring of strength in different muscle groups.

Findings. The study groups were similar at baseline.

The crude 90-day mortality rate was lower in the cisatracurium group (31.6% vs 40.7%, P = .08). Regression analysis showed an improved 90-day survival rate with the use of this neuromuscular blocker after adjustment for severity of illness and the severity of ARDS (based on degree of hypoxemia and plateau pressures) (hazard ratio for death at 90 days 0.68; 95% CI 0.48–0.98; P = .04). The rate of paresis acquired in the intensive care unit did not differ significantly between the two groups.

Conclusion. In patients with severe ARDS, giving a neuromuscular blocking agent early improved the survival rate and increased the time off the ventilator without increasing muscle weakness.

These data are in line with similar findings from two other studies published by the same group.^38,39 A meta-analysis of 432 patients showed that the use of neuromuscular blockade in early severe ARDS is associated with a statistically significant effect on early mortality (relative risk 0.66, 95% CI 0.50–0.87).⁴⁰ The pooled analysis of these trials did not show any statistically significant critical-illness polyneuropathy.

These results need to be interpreted carefully, as we have inadequate data to see if they generalize to different intensive care units, and the evaluation and categorization of critical-illness polyneuropathy remains to be defined.

Cisatracurium is a promising treatment for moderate to severe ARDS and merits investigation in a large confirmatory randomized controlled trial.

Other pharmacologic agents

A number of other drugs have been studied in ARDS patients, including both inhaled and intravenous beta agonists,^41,42 statins,⁴³ and nutritional supplements.⁴⁴ But as with other drugs previously studied in ARDS such as corticosteroids, N-acetylcysteine, and surfactant,³⁴ these agents showed no effect on outcomes. In fact, a recent trial of intravenous salbutamol in ARDS patients was stopped after an interim analysis because of a higher incidence of arrhythmias and lactic acidosis with this agent.⁴²

These findings reaffirm that pharmacologic therapy needs to be carefully considered, and potential harms associated with these therapies need to be addressed before they are introduced in the care of critically ill patients.

Continued progress in understanding the pathophysiology of acute respiratory distress syndrome (ARDS) is translating into changes in the way we diagnose and manage it. Over the past 20 years, low tidal volume,¹ positive end-expiratory pressure (PEEP),² and fluid restriction³ have become the standard of care. A multidisciplinary approach, including targeted use of sedatives, early mobilization, and protocols for weaning from the ventilator, has also brought about substantial changes in ARDS management and its outcomes.^4–6

In this article, we review the most relevant articles about ARDS in the last 5 years. We include the new definition of ARDS and studies of ventilatory and nonventilatory therapies that have implications in managing patients with ARDS.

A STANDARDIZED APPROACH

ARDS is characterized by damage to the alveolar architecture, severe hypoxemia, and bilateral parenchymal opacities.

The working definition of ARDS developed in 1994 by the American-European Consensus Conference (AECC) was the basis for enrollment in most of the landmark trials and observational studies over the past 20 years.^7,8 However, it was limited in its reliability and validity.

An updated definition

In 2011, the ARDS Definition Task Force, using a novel consensus process, updated the ARDS definition,⁹ focusing on its feasibility, reliability, and validity in predicting response to therapies and outcomes in ARDS. This new “Berlin” definition is not substantially different from the old, but defines the criteria more specifically:

• Bilateral opacities, unexplained by nodules, atelectasis, or effusion, on chest radiography or computed tomography
• New or worsening respiratory symptoms, or a clinical insult associated with ARDS within 7 days of diagnosis
• Objective assessment of cardiac function (eg, with echocardiography) to exclude cardiogenic pulmonary edema
• Hypoxemia, with a partial pressure of arterial oxygen divided by the percentage of inspired oxygen (PaO₂/FiO₂ ratio) of 300 mm Hg or less despite noninvasive or invasive mechanical ventilation with PEEP or continuous positive airway pressure (CPAP) of at least 5 cm H₂O.

In addition, the new definition classifies the severity of disease on the basis of the degree of hypoxemia, ie, the PaO₂/FiO₂ ratio:

• Mild: PaO₂/FiO₂ ratio > 200 and ≤ 300 mm Hg
• Moderate: PaO₂/FiO₂ ratio > 100 and ≤ 200 mm Hg
• Severe: PaO₂/FiO₂ ratio ≤ 100 mm Hg.

The term “acute lung injury” has been eliminated, as has the previous criterion of a pulmonary artery wedge pressure of 18 mm Hg or less.

The panel also evaluated four ancillary variables for predicting outcomes in severe ARDS:

• Compliance of the respiratory system less than or equal to 40 mL/cm H₂O
• Radiographic severity (involvement of three or four quadrants on chest radiography)
• PEEP of 10 cm H₂O or greater
• Corrected expired volume 10 L/min or greater.

The task force evaluated the reliability and validity of this definition in a meta-analysis of 4,400 patients previously enrolled in randomized controlled trials or observational studies.

Findings. The Berlin definition predicted the risk of death better than the AECC definition. The mortality rate increased with the severity of ARDS, from 27% with mild disease to 32% with moderate disease to 45% with severe disease. The four ancillary variables did not contribute to the predictive validity of severe ARDS for mortality and were removed from the definition.

Thille et al¹⁰ retrospectively reviewed autopsy findings from 712 patients and found that the new definition identified a homogeneous group who had severe ARDS.¹⁰

Conclusions. The new definition may overcome some of the limitations of the old one, but it needs to be validated in clinical practice, especially its ability to predict death.

VENTILATORY SUPPORT

Prompt recognition, lung-protective ventilation, and a conservative fluid strategy remain the cornerstones of ARDS management. However, other strategies are being tested.

Prone-position ventilation in severe ARDS: The right therapy in a specific population

Prone-position ventilation was first described almost 30 years ago, but it has been used inconsistently in clinical practice.

Physiologic and observational studies indicated that prone positioning might improve survival in patients with ARDS, but several randomized trials failed to demonstrate any positive effect on outcomes.^11,12 Some trials also reported a higher rate of complications with this intervention.¹³ However, meta-analyses suggested that prone-position ventilation might have a beneficial effect in patients with severe ARDS (defined as a PaO₂/FiO₂ ratio ≤ 100 mm Hg).¹⁴

In view of these findings, investigators conducted a trial of prone-position ventilation exclusively in patients with severe ARDS.

The PROSEVA study

The Proning Severe ARDS Patients (PROSEVA) study was a randomized controlled trial designed to determine whether prone-position ventilation, applied early, would improve outcomes in patients with severe ARDS.¹⁵

In PROSEVA, 466 patients with severe ARDS (defined as a PaO₂/FiO₂ ratio < 150 mm Hg, FiO₂ ≥ 60%, and PEEP ≥ 5 cm H₂O) underwent either at least 16 hours of prone positioning or were left in the supine position after 12 to 24 hours of initial conventional mechanical ventilation. The patients were recruited from centers in France and Spain where prone-position ventilation had been used in daily practice for more than 5 years.

The primary outcome studied was the rate of death at 28 days. The secondary end points were the death rate at day 90, rates of successful extubation, the length of stay in the intensive care unit, and complications.

Findings. At study entry, the patients in the supine group were sicker, more of them required a vasopressor, and fewer of them were receiving neuromuscular blocking agents than those in the prone group. These baseline differences may have influenced the outcomes; the unadjusted 28-day mortality rate was 16.0% in the prone group compared with 32.8% in the supine group (P < .001). However, the hazard ratio for death with prone positioning was 0.39 (95% confidence interval [CI] 0.25–0.63) even after adjusting for severity and the use of vasopressors and neuromuscular blocking agents. Prone-position ventilation was not associated with a higher incidence of complications, and the rate of successful extubation was higher.

Conclusions. In patients with severe ARDS, early use of prolonged prone positioning significantly decreased the 28-day and 90-day mortality rates. This trial has made prone positioning one of the strategies in managing patients with early severe ARDS. To minimize complications such as pressure ulcers and line or tube dislodgement, personnel caring for these patients must follow a protocol and undergo specific training.

These results were corroborated by a meta-analysis by Beitler et al¹⁶ that found a significant decrease in mortality rate with prone-position ventilation even in older studies when lung-protective ventilation strategies were separated from high-tidal-volume ventilation.

High-frequency oscillatory ventilation: No benefit in two trials

Observational data and experimental studies suggested that high-frequency oscillatory ventilation (HFOV) is superior to conventional mechanical ventilation in ARDS patients.^17,18 However, outdated and cumbersome equipment, lack of protocols, and a lack of high-quality evidence led to limited and inconsistent use of HFOV, mainly as a rescue therapy in ARDS.¹⁹

Over the last few years, HFOV has been gaining acceptance, especially earlier in the course of ARDS.²⁰ After preliminary clinical trials reported promising results, two trials conducted in Canada and the United Kingdom compared HFOV vs conventional mechanical ventilation in patients with ARDS.

The OSCAR study

The Oscillation in ARDS (OSCAR) study²¹ was a “pragmatic” trial²² (ie, it had minimal exclusion criteria) of the safety and effectiveness of HFOV as a primary ventilatory strategy for ARDS. It included 795 patients randomized to receive conventional ventilation (n = 397) or HFOV (n = 398). Research centers followed detailed algorithms for HFOV management and adopted their usual practice for conventional ventilation. Medical care was given according to the clinician’s judgment.

The primary outcome studied was survival at 30 days. The secondary outcomes were all-cause mortality in the intensive care unit and the hospital, duration of mechanical ventilation, and use of antimicrobial, sedative, vasoactive, and neuromuscular-blocking drugs.

Findings. The patient baseline characteristics were similar in both groups.

There was no significant difference in intensive care unit mortality rates, hospital mortality rates, or mortality rates at 30 days (41.7% in the HFOV group vs 41.1% in the conventional ventilation group; P = .85, 95% CI 6.1–7.5) even after adjustments for center or severity of illness.

The duration of mechanical ventilation was similar in both groups (14.9 ± 13.3 days in the HFOV group vs 14.1 ± 13.4 days in the conventional ventilation group, P = .41). However, sedatives and neuromuscular-blocking drugs were used more often and longer in the HFOV group than in the conventional ventilation group. There was no difference in the use of vasoactive or antimicrobial medications.

Conclusions. This multicenter randomized control trial did not demonstrate any benefit from using HFOV for routine management of ARDS. Its pragmatic design made it less likely to reach a firm conclusion,²² but it at least made a case against routinely using HFOV in patients with ARDS.

The OSCILLATE study

The Oscillation for Acute Respiratory Distress Syndrome Treated Early (OSCILLATE) study²³ assessed the safety and efficacy of HFOV as a treatment for early-onset moderate-to-severe ARDS.

The inclusion criteria were similar to those in the OSCAR trial except that pulmonary symptoms had to be present less than 2 weeks and ARDS assessment was done under standard ventilator settings. As this was an efficacy trial, it had more exclusion criteria than the OSCAR trial. A total of 548 patients were randomized to receive conventional ventilation (n = 273) or HFOV (n = 275). The baseline characteristics were similar between groups.

Conventional ventilation was given according to a protocol used in an earlier trial² and included recruitment maneuvers. HFOV was given in centers that had experience in this treatment, and there were protocols for ventilation management, hemodynamic optimization, and weaning. All other care was left to the clinician’s choice.

The primary outcome studied was in-hospital mortality. The investigators also evaluated whether there were interactions between the treatment and baseline severity of lung injury and center experience with HFOV.

Findings. The trial was stopped after an interim analysis found that HFOV might be harmful, although the statistical threshold for stopping was not reached. The in-hospital mortality rate was 47% in the HFOV group and 35% in the control group (relative risk of death with HFOV 1.33, 95% CI 1.09–1.64, P = .005). HFOV was worse than conventional ventilation regardless of the severity of disease or center experience. The HFOV group had higher mean airway pressures but similar FiO₂ compared with the conventional ventilation group.

The HFOV group received significantly more vasopressors, sedatives, and neuromuscular blockers. This group’s fluid balance was higher as well, but not significantly so. Refractory hypoxemia (defined as PaO₂ < 60 mm Hg for 1 hour with an FiO₂ of 1.0 and neuromuscular blockade) was more frequent in the conventional ventilation group, but the number of deaths in the subgroup with refractory hypoxemia was similar with either treatment.

Conclusions. This multicenter randomized controlled trial demonstrated that HFOV was harmful when used routinely to manage ARDS. The trial’s protocol was based on the results of a pilot study carried out by the same investigators, which provided the best evidence available regarding the safety of HFOV at that time.

The results of the OSCAR and OSCILLATE trials have quelled enthusiasm for early, routine use of HFOV in ARDS. Although there are concerns that the protocol (ie, the way HFOV was implemented) rather than HFOV itself may have led to worse outcomes, there is no signal to support its routine use. We need further studies to define if it remains a viable rescue therapy.

Extracorporeal membrane oxygenation: Is it a viable option in severe ARDS?

Extracorporeal membrane oxygenation (ECMO) uses cardiopulmonary bypass technology to provide gas exchange. In patients with severe hypoxemia, ECMO can ensure adequate oxygenation and ventilation while ensuring the optimization of lung-protective ventilation. But ECMO was never as successful in adults with ARDS as it was in children and neonates.²⁴

The first two trials of ECMO in ARDS^24,25 reported equal or worse survival rates compared with conventional ventilation, and the overall mortality rate in these studies was staggeringly high. However, these studies were carried out before the era of lung-protective ventilation and at a time when ECMO technology was relatively primitive.

With new technology such as venovenous circuits and smaller cannulas, ECMO has gained more acceptance. It was used in patients with severe or refractory hypoxemia associated with ARDS during the H1N1 pandemic.^26,27

The CESAR trial

The Conventional Ventilatory Support Versus Extracorporeal Membrane Oxygenation for Severe Adult Respiratory Failure (CESAR) trial²⁸ assessed the safety, clinical efficacy, and cost-effectiveness of ECMO in managing severe ARDS. It compared best standard practice vs a protocol that included ECMO. The trial was conducted from 2001 to 2006.

Patients with severe ARDS, as defined by a Murray score²⁹ greater than 3 or uncompensated hypercapnea, were prospectively randomized and recruited from an ECMO center and 148 tertiary intensive care units and referral hospitals in England. This was a pragmatic trial, with minimal exclusion criteria (essentially, mechanical ventilation with high pressures and high FiO₂ for more than 7 days, intracranial bleeding, or contraindication to heparinization).

A total of 180 patients were randomized in a one-to-one ratio to receive ECMO or conventional management. The ventilator management in the conventional treatment group was not done according to a protocol but in general was low-volume and low-pressure. All patients randomized to ECMO were transferred to the ECMO center and treated according to a standardized ventilation protocol. After 12 hours, if predefined goals were not reached, venovenous ECMO was started. Patients assigned to conventional management could not cross over to ECMO.

The primary outcomes were death or severe disability at 6 months after randomization, and cost-effectiveness. The secondary outcomes were hospital resource use (eg, rescue techniques, length of stay, duration of ECMO) and health status after 6 months.

Findings. The groups were similar at baseline. Sixty-eight (75%) of the 90 patients randomized to receive ECMO actually received it. Of the 22 patients who did not receive ECMO, 16 (18% of the 90) improved on conventional therapy, 5 (6%) died during or before transfer, and 1 could not receive heparin.

Two patients had severe complications in the ECMO group: one had an arterial puncture, and one had an oxygen delivery failure during transport. In each case, these events contributed to the death of the patient.

More patients in the ECMO group received lung-protective ventilation, 84 (93%) vs 63 (70%).

The primary outcome, ie, death or severe disability at 6 months, occurred in 33 (37%) of the 90 patients in the ECMO group and in 46 (53%) of the patients in the conventional management group (relative risk 0.69, 95% CI 0.05–0.97, P = .03). More patients in the ECMO group survived, but the difference was not statistically significant (relative risk of death 0.73, 95% CI 0.52–1.03, P = .07). The most common cause of death in the ECMO group was multiorgan failure (42%), whereas in the conventional management group, the most common cause of death was respiratory failure (60%).

Length of stay in the hospital and in the critical care unit and health care costs were double for patients in the ECMO group. There was no difference in quality-of-life markers at 6 months in the survivors.

Conclusions. This pragmatic trial demonstrated that a protocol that includes ECMO could improve survival rates in ARDS.

Of note, the ECMO group got care in regional centers that used protocols. Therefore, in interpreting the results of this trial, we have to consider that being in a center with protocol-specified care for ARDS could drive some of the difference in mortality rates.

Regardless, this trial demonstrated that ECMO is feasible and led to better outcomes than expected. The findings were encouraging, and spurred the use of ECMO in severe ARDS during the 2009 H1N1 pandemic. Two propensity-matched studies and a number of case series reported a survival benefit associated with the use of ECMO in patients with severe ARDS.^27,30

A recent meta-analysis also reported that ECMO might lower the mortality rate in ARDS; however, the patients in the H1N1 pandemic were younger and usually had isolated respiratory failure.³¹

The success of ECMO has opened new possibilities in the management of ARDS. As the technology improves and our experience increases, ECMO will likely gain more acceptance as a treatment for severe ARDS.

Airway pressure release ventilation

The use of airway pressure release ventilation and other ventilator modalities in ARDS is not supported by current evidence, though results of clinical trials may influence our practice in the future.

PHARMACOTHERAPY IN ARDS

The pathogenesis of ARDS includes damage to the alveolar-capillary membrane, with leakage of protein-rich edema fluid into alveoli. This damage is propagated by a complex inflammatory response including but not limited to neutrophil activation, free-radical formation, dysregulation of the coagulation system, and extensive release of inflammatory mediators.^32,33 As a consequence, there are multiple potential targets for pharmacologic therapy in ARDS.

A variety of drugs, including corticosteroids, anti-inflammatory agents, immune-modulating agents, pulmonary vasodilators, antioxidants, and surfactants, have been studied in patients with ARDS.³⁴ But effective pharmacotherapy for ARDS remains extremely limited.

Neuromuscular blockade in early severe ARDS

Mechanical ventilation can result in injurious stretching of the lung parenchyma, either from alveolar overdistention (volutrauma) or from continual recruitment and derecruitment of unstable lung units during the ventilator cycle (atelectrauma).³⁵ Ventilator-induced lung injury can be exacerbated by asynchronous breathing.

In theory, neuromuscular blockers could minimize patient-ventilator asynchrony and provide much better control of tidal volume and pressure in patients with ARDS. This may result in less volutrauma and atelectrauma associated with asynchronous breathing. Data also suggest that cisatracurium (Nimbex), a neuromuscular blocking agent, may have a direct effect on the amount of inflammation in lungs with ARDS.³⁶

The ACURASYS study

The ARDS et Curarisation Systématique (ACURASYS) study³⁷ was a randomized trial in 340 patients undergoing mechanical ventilation for severe ARDS to evaluate the impact of neuromuscular blockade within the first 48 hours in this population.

The primary outcome was the mortality rate before hospital discharge or within 90 days of study entry. Secondary outcomes included the 28-day mortality rate, the rate of intensive care unit-acquired paresis, and the number of ventilator-free days. To be included, patients had to have been mechanically ventilated for less than 48 hours and to meet the AECC criteria for severe ARDS, with a PaO₂/FiO₂ ratio less than 150 mm Hg.

The intervention group received a continuous infusion of cisatracurium for 48 hours, while the control patients received placebo. Muscle strength was evaluated by clinical scoring of strength in different muscle groups.

Findings. The study groups were similar at baseline.

The crude 90-day mortality rate was lower in the cisatracurium group (31.6% vs 40.7%, P = .08). Regression analysis showed an improved 90-day survival rate with the use of this neuromuscular blocker after adjustment for severity of illness and the severity of ARDS (based on degree of hypoxemia and plateau pressures) (hazard ratio for death at 90 days 0.68; 95% CI 0.48–0.98; P = .04). The rate of paresis acquired in the intensive care unit did not differ significantly between the two groups.

Conclusion. In patients with severe ARDS, giving a neuromuscular blocking agent early improved the survival rate and increased the time off the ventilator without increasing muscle weakness.

These data are in line with similar findings from two other studies published by the same group.^38,39 A meta-analysis of 432 patients showed that the use of neuromuscular blockade in early severe ARDS is associated with a statistically significant effect on early mortality (relative risk 0.66, 95% CI 0.50–0.87).⁴⁰ The pooled analysis of these trials did not show any statistically significant critical-illness polyneuropathy.

These results need to be interpreted carefully, as we have inadequate data to see if they generalize to different intensive care units, and the evaluation and categorization of critical-illness polyneuropathy remains to be defined.

Cisatracurium is a promising treatment for moderate to severe ARDS and merits investigation in a large confirmatory randomized controlled trial.

Other pharmacologic agents

A number of other drugs have been studied in ARDS patients, including both inhaled and intravenous beta agonists,^41,42 statins,⁴³ and nutritional supplements.⁴⁴ But as with other drugs previously studied in ARDS such as corticosteroids, N-acetylcysteine, and surfactant,³⁴ these agents showed no effect on outcomes. In fact, a recent trial of intravenous salbutamol in ARDS patients was stopped after an interim analysis because of a higher incidence of arrhythmias and lactic acidosis with this agent.⁴²

These findings reaffirm that pharmacologic therapy needs to be carefully considered, and potential harms associated with these therapies need to be addressed before they are introduced in the care of critically ill patients.