Evaluation and management of the nonventilated, hospitalized adult patient with acute hypoxemia

INTRODUCTION — Acute oxygen desaturation and hypoxemia events are commonly encountered in hospitalized patients. In some patients, acute hypoxemia may herald the onset of a serious illness, while in others it is easily treated and reversed.

This topic review will discuss the evaluation and management of the spontaneously breathing adult patient with acute hypoxemic respiratory failure. Management of patients with acute hypercapnic respiratory failure, the pathophysiology of hypoxemia, adverse effects of oxygen, indications for long-term oxygen therapy, and management of respiratory distress in the ventilated patient are discussed separately.

●(See "The evaluation, diagnosis, and treatment of the adult patient with acute hypercapnic respiratory failure".)

●(See "Measures of oxygenation and mechanisms of hypoxemia".)

●(See "Adverse effects of supplemental oxygen".)

●(See "Long-term supplemental oxygen therapy".)

●(See "Assessment of respiratory distress in the mechanically ventilated patient".)

DEFINITION, MECHANISMS, AND ETIOLOGY OF HYPOXEMIA

●Hypoxemia is defined as an abnormally low level of oxygen in the blood (ie, low partial oxygen tension [eg, PaO₂ <60 mmHg]).

●Hypoxia is defined as a condition where the oxygen supply is inadequate either to the body as a whole (general hypoxia) or to a specific region (tissue hypoxia).

Hypoxemia does not necessarily always indicate tissue hypoxia.

The causes of hypoxemia can be classified into six main mechanistic categories. These and their relevant etiologies are listed in the table and discussed in detail separately (table 1). (See "Measures of oxygenation and mechanisms of hypoxemia".):

●Ventilation/perfusion (V/Q) mismatch.

●Right to left shunt (mechanical, physiologic).

●Hypoventilation (table 2). (See "The evaluation, diagnosis, and treatment of the adult patient with acute hypercapnic respiratory failure".)

●Diffusion limitation.

●Reduced inspired oxygen tension. (See "High-altitude illness: Physiology, risk factors, and general prevention".)

●Low mixed venous oxygen.

Rare causes of hypoxemia include the following:

●Genetic causes of severe anemia or hemoglobinopathies (low hemoglobin values affecting oxygen delivery to tissue or altering oxygen carrying capacity). (See "Hemoglobin variants that alter hemoglobin-oxygen affinity".)

●Cyanide toxicity, carboxyhemoglobinemia, or methemoglobinemia. (See "Cyanide poisoning" and "Carbon monoxide poisoning" and "Methemoglobinemia".)

●Leukocyte or platelet larceny (spurious hypoxemia; also known as pseudohypoxemia). (See "Measures of oxygenation and mechanisms of hypoxemia", section on 'Spurious hypoxemia'.)

While in most patients, V/Q mismatch is the dominant mechanism, usually more than one mechanism is responsible for hypoxemia, particularly in patients with underlying cardiopulmonary disease.

INITIAL TRIAGE AND TREATMENT — When patients present with suspected hypoxemia (eg, dyspnea, desaturation on pulse oximetry, tachypnea, wheeze, chest pain, palpitations, respiratory distress), we simultaneously perform the following (algorithm 1) [1,2]:

●Assess and stabilize the airway, breathing, and circulation. (See 'Assess and stabilize airway, breathing, and circulation' below.)

●Immediately administer oxygen. (See 'Oxygen and respiratory support' below.)

●Initiate empiric treatment for the suspected underlying cause. (See 'Address presumed cause for hypoxemia' below.)

Once stable, we generally make a decision to triage the patient to a hospital ward (with or without telemetry), a high-dependency unit, or the intensive care unit (ICU). This decision is individualized and depends upon factors including the severity of the illness and hypoxemia, predicted trajectory of the illness, and resource availability.

Assess and stabilize airway, breathing, and circulation — For patients with suspected acute hypoxemia, we rapidly assess airway, breathing, and circulation and ensure that pulse oximetry is accurate:

●Airway, breathing, circulation – We evaluate the patient's airway for patency, their breathing pattern, pulse rate and rhythm, temperature, and blood pressure, all of which should be stabilized, if compromised. (See "Approach to the patient with dyspnea", section on 'Evaluation of acute dyspnea' and "Approach to the adult with dyspnea in the emergency department".)

•For spontaneously breathing patients who are comfortable, we apply supplemental oxygen, typically with a low-flow oxygen system, but occasionally a high-flow system will be needed. We administer the lowest flow to obtain a target peripheral oxygen saturation (SpO₂) greater than 90 percent. The upper limit of the target SpO₂ can then be adjusted on a case-by-case basis according to that needed for the suspected etiology. (See 'Oxygen and respiratory support' below.)

•For patients with respiratory distress or who are hemodynamically unstable, noninvasive or invasive mechanical ventilation are typically needed, the details of which are discussed separately. (See "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications" and "Direct laryngoscopy and endotracheal intubation in adults" and "Approach to the adult with dyspnea in the emergency department", section on 'Emergency stabilization of patients with danger signs'.)

•We also ensure that intravenous access is secured. Any life-threatening arrhythmias and/or hypotension are treated immediately at the bedside (eg, antiarrhythmic agents, cardioversion, intravenous fluids and/or vasopressors). (See "Advanced cardiac life support (ACLS) in adults" and "Evaluation of and initial approach to the adult patient with undifferentiated hypotension and shock".)

●Pulse oximetry – We also simultaneously apply a pulse oximeter (if not already in place) and check the waveform, if feasible, to ensure accuracy, bearing in mind conditions that can erroneously lower or raise the SpO₂ (table 3). If there is any concern about pulse oximetry accuracy, we obtain an arterial blood gas (ABG). (See "Pulse oximetry", section on 'Interpreting the results' and "Pulse oximetry", section on 'Troubleshooting sources of error'.)

Once the patient is stable, we perform a thorough history and examination, obtain laboratory and imaging studies, and continue therapy for the presumed cause as detailed below. (See 'History and examination' below and 'Laboratory testing' below and 'Arterial blood gas' below and 'Chest imaging (radiography and/or ultrasonography)' below and 'Electrocardiography' below and 'Address presumed cause for hypoxemia' below.)

Oxygen and respiratory support — There are several choices for oxygen delivery in the acute situation (figure 1). Our general approach is as follows:

●For most patients with acute nonhypercapnic, hypoxemic respiratory failure, we typically start with low-flow oxygen devices such as nasal cannulae (1 to 6 liters/minute [L/min]). (See 'Patients with minimal oxygen requirements (eg, ≤6 L/min)' below.)

●When oxygen requirements are greater than what can be provided with low-flow oxygen (eg, ≥6 L/min) or breathing is labored, advanced respiratory support is generally needed. Options are the following:

•High-flow oxygen delivered via a low-flow system

•Humidified high-flow oxygen delivered via nasal cannulae (HFNC)

•Oxygen delivered via a noninvasive ventilation (NIV) device or invasive mechanical ventilation (ie, an endotracheal tube)

Among the options, we prefer an initial trial of HFNC unless there is a separate indication for a different modality (eg, concomitant acute hypercapnia requiring bilevel positive airway pressure (BPAP) or respiratory distress requiring immediate mechanical ventilation). Choosing among these options is discussed below. (See 'Patients with requirements for advanced respiratory support' below.)

Patients with minimal oxygen requirements (eg, ≤6 L/min) — For patients with minimal oxygen requirements, we use the following general approach:

●For most patients with acute hypoxemia, we use low-flow nasal cannulae. They are the most common initial delivery device used and can accurately provide oxygen flow up to 6 liters/minute (L/min). (See "Continuous oxygen delivery systems for the acute care of infants, children, and adults", section on 'Nasal cannula'.)

●For patients who are mouth breathers or who cannot receive or tolerate nasal cannulae (eg, patients with nasal packing), we typically use a simple facemask to deliver low-flow oxygen. (See "Continuous oxygen delivery systems for the acute care of infants, children, and adults", section on 'Face masks'.)

●In rare circumstances, when patients have oxygen-induced hypercapnia, fine regulation of oxygen via nasal cannulae (ie, using small incremental changes of oxygen; eg, 1 L/min, sometimes less) or venturi masks (figure 2) (eg, 24, 28, or 35 percent) is appropriate. Details regarding oxygen titration in oxygen-sensitive patients are provided separately. (See "The evaluation, diagnosis, and treatment of the adult patient with acute hypercapnic respiratory failure", section on 'Titration of oxygen'.)

This approach is based upon the wide experience with low-flow systems in adults with respiratory failure. Data that compare low-flow oxygen with oxygen delivered via humidified high-flow systems are discussed below. (See 'Patients with requirements for advanced respiratory support' below.)

Patients with requirements for advanced respiratory support — Recommendations for advanced respiratory support of non-coronavirus disease 2019 (COVID-19) patients are provided in this section, while recommendations for patients with acute, hypoxemic respiratory failure due to COVID-19 are provided separately. (See "COVID-19: Respiratory care of the nonintubated hypoxemic adult (supplemental oxygen, noninvasive ventilation, and intubation)", section on 'Patients with requirements for advanced respiratory support'.)

Once oxygen requirements start to increase over 6 L/min or breathing becomes labored, options are HFNC, high-flow oxygen delivered through a low-flow system, and oxygen delivered via an NIV device. Among these options, we typically prefer an initial trial of HFNC, unless there is a separate indication for NIV (figure 1). In practice, either or both modalities are often trialed until the patient improves or deteriorates.

Our approach for the use of HFNC is for the most part aligned with guidelines issued by the European Respiratory Society [3].

Our approach — For patients who do not have COVID-19-related acute hypoxemic respiratory failure who require advanced respiratory support, the following is a reasonable approach:

●For patients with acute, hypoxemic respiratory failure who have a need for advanced respiratory support and in whom there is a known benefit and no contraindications to NIV (table 4), we generally treat with a trial of NIV. Data that support NIV in specific patient groups are discussed below. (See 'Patients with indications for noninvasive ventilation' below.)

●For patients with acute, nonhypercapnic, hypoxemic respiratory failure who have a need for advanced respiratory support, we generally treat with a trial of HFNC. Data that support this choice are discussed below. (See 'Humidified, high-flow oxygen delivered via nasal cannulae (HFNC)' below.)

For those who cannot tolerate or receive HFNC or if HFNC is not available, high flows of oxygen may be delivered through a low-flow system (eg, up to 10 L/min, sometimes more) using a simple mask (eg, face mask or scoop mask) or a non-rebreather mask. (See "Continuous oxygen delivery systems for the acute care of infants, children, and adults", section on 'Face masks'.)

Patients with indications for noninvasive ventilation — NIV, typically BPAP or continuous positive airway pressure (CPAP), is the application of ventilation through a noninvasive interface (eg, mask or nasal cannulae) rather than through an endotracheal or tracheostomy tube. In patients with acute hypoxemic respiratory failure, NIV can provide oxygenation and/or ventilation support with the goal of avoiding invasive mechanical ventilation.

NIV is suitable for select patients with acute respiratory failure, the details of which are provided separately:

●Patients with acute hypercapnic hypoxemic respiratory failure due to an acute exacerbation of chronic obstructive pulmonary disease (AECOPD). (See "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications", section on 'Acute exacerbation of chronic obstructive pulmonary disease (AECOPD) with hypercapnic respiratory acidosis'.)

●Patients with acute hypoxemic respiratory failure due to acute cardiogenic pulmonary edema. (See "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications", section on 'Acute cardiogenic pulmonary edema (ACPE)'.)

●Patients with acute hypercapnic hypoxemic respiratory failure due to neuromuscular or chest wall disease. (See "Respiratory muscle weakness due to neuromuscular disease: Management", section on 'Noninvasive ventilation'.)

Data comparing HFNC with NIV in patients with hypoxemic respiratory failure are also discussed separately. (See "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications", section on 'Hypoxemic nonhypercapnic respiratory failure not due to ACPE'.)

Humidified, high-flow oxygen delivered via nasal cannulae (HFNC) — The major advantages of HFNC compared with low-flow oxygen are that HFNC can provide a fraction of inspired oxygen (FiO₂) up to 1.0 and flows up to 60 L/min, sometimes higher, via nasal cannulae as well as a small amount of positive end-expiratory pressure (eg, <5 cm H₂O). Data that support the value of HFNC in patients with acute hypoxemic respiratory failure are discussed in this section. The practical application of HFNC is discussed separately. (See "Heated and humidified high-flow nasal oxygen in adults: Practical considerations and potential applications".)

●HFNC compared with low-flow oxygen – HFNC has been compared with low-flow oxygen in medical patients with severe nonhypercapnic hypoxemic respiratory failure. Data have demonstrated that while HFNC improves oxygenation in patients with acute hypoxemic respiratory failure compared with low-flow oxygen systems, there is no impact of HFNC on mortality, and the impact on intubation rates, length of stay, dyspnea, and comfort are uncertain [4-26]. However, data have been hampered by imprecision and heterogeneity in study design, patient population characteristics, and type of respiratory failure. As examples:

•A meta-analysis of six trials including 2769 patients reported that while HFNC did not reduce 28-day mortality compared with conventional oxygen therapy (COT), it did significantly reduce the rate of reintubation (relative risk [RR] 0.89, 95% CI 0.81-0.97) [26]. Most trials were consistent in this effect, and subgroup differences were reported.

•In a meta-analysis of eight randomized trials that compared HFNC with COT (ie, low-flow systems) in patients with acute hypoxemic respiratory failure due to several etiologies (eg, pulmonary edema, AECOPD, pneumonia in immunocompromised and immunocompetent patients), HFNC had little or no impact on intubation rates (26 percent each; odds ratio 0.98, 95% CI 0.34-2.82; low-certainty evidence) or mortality (26 versus 27 percent; RR 0.97, 95% CI 0.82-1.14; low-certainty evidence) [21]. The same meta-analysis also reported improved dyspnea and comfort, a possible reduction in hospital-acquired pneumonia (one trial), but uncertain effect on ICU admissions and length of stay.

•In a meta-analysis of nine trials that compared HFNC with low-flow oxygen in patients with hypoxemic respiratory failure, HFNC decreased the need for both intubation (RR 0.85, 95% CI 0.74-0.99; moderate-certainty evidence) and escalation of respiratory support (RR 0.71, 95% CI 0.51-0.98; moderate-certainty evidence) [22]. No improvements were reported on mortality, length of stay, or patient dyspnea and comfort.

•In a network meta-analysis, HFNC was shown to reduce rates of intubation in patients with acute hypoxemic respiratory failure compared with COT but had no impact on mortality (RR 0.76, 95% CI 0.55-0.99; absolute risk difference, -0.11; moderate-certainty evidence) [25]; this study is discussed in detail separately. (See "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications", section on 'Hypoxemic nonhypercapnic respiratory failure not due to ACPE'.)

•Data are limited in specific subgroups:

-Immunocompromised patients with acute nonhypercapnic hypoxemic respiratory failure – While one prospective study of 1611 immunocompromised patients with acute hypoxemic respiratory failure reported that HFNC reduced intubation rates [27], two post-hoc analyses and one unblinded randomized trial reported no difference in intubation rates or mortality with HFNC when compared with standard oxygen therapy in the same population [9,10,28].

-Select etiologies – Only case series describe successful use of HFNC in acute hypoxemic respiratory failure due to severe H1N2 influenza [11], hematologic malignancies [12], organ transplant [13], congestive heart failure [14], and acute respiratory distress syndrome [15]. The impact of HFNC on COVID-19 is discussed separately. (See "COVID-19: Respiratory care of the nonintubated hypoxemic adult (supplemental oxygen, noninvasive ventilation, and intubation)", section on 'Noninvasive modalities'.)

●HFNC compared with NIV – Data comparing HFNC with NIV in patients with hypoxemic respiratory failure are limited and have not shown consistent benefit in terms of intubation rates or other outcomes. These data are discussed separately:

•Medical patients with hypoxemic nonhypercapnic acute respiratory failure – (see "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications", section on 'Hypoxemic nonhypercapnic respiratory failure not due to ACPE')

•Patients who need postextubation support – (see "Extubation management in the adult intensive care unit", section on 'Low-flow versus high-flow oxygen' and "Extubation management in the adult intensive care unit", section on 'High-flow oxygen via nasal cannulae')

•Patients with postoperative respiratory failure – (see "Overview of the management of postoperative pulmonary complications", section on 'Postoperative respiratory failure')

•Patients who need intubation support – (see "Airway management for induction of general anesthesia", section on 'Preoxygenation' and "Rapid sequence induction and intubation (RSII) for anesthesia", section on 'Preoxygenation' and "Heated and humidified high-flow nasal oxygen in adults: Practical considerations and potential applications", section on 'Intubation support')

•Patients who undergo tracheostomy weaning – (see "Heated and humidified high-flow nasal oxygen in adults: Practical considerations and potential applications", section on 'Miscellaneous')

•Patients who undergo fiberoptic bronchoscopy – (see "Heated and humidified high-flow nasal oxygen in adults: Practical considerations and potential applications", section on 'Miscellaneous')

Oxygen saturation goals — For patients with hypoxemic respiratory failure who are treated with supplemental oxygen, there is no single optimal target range for oxygen saturation. Nonetheless, experts agree that oxygenation goals should be individualized based upon the specific condition being treated, that hyperoxia should be avoided, and that the lowest possible FiO₂ necessary to meet oxygenation goals should be used.

General population — In most patients with acute hypoxemic respiratory failure who do not have a condition requiring a select oxygenation goal (see 'Special populations' below), we typically target an SpO₂ between 90 and 96 percent, if feasible [29]. An alternative option is to target a minimum arterial oxygen tension goal of 60 mmHg. These limits decrease the likelihood that adverse consequences of supplemental oxygen will develop (eg, absorption atelectasis, accentuation of hypercapnia, parenchymal injury) (see "Adverse effects of supplemental oxygen"), while simultaneously avoiding dangerously low saturations. Guidelines and data that support this approach are discussed separately. (See "Overview of initiating invasive mechanical ventilation in adults in the intensive care unit", section on 'Fraction of inspired oxygen'.)

Special populations — Saturation goals unique to special populations are discussed separately:

●Acute respiratory distress syndrome – (see "Acute respiratory distress syndrome: Ventilator management strategies for adults", section on 'Positive end-expiratory pressure (PEEP), fraction of inspired oxygen, oxygenation target')

●COVID-19 pneumonia – (see "COVID-19: Respiratory care of the nonintubated hypoxemic adult (supplemental oxygen, noninvasive ventilation, and intubation)", section on 'Oxygenation targets')

●Carbon monoxide poisoning – (see "Carbon monoxide poisoning", section on 'Hyperbaric oxygen therapy')

●Air embolism – (see "Air embolism", section on 'Supportive therapy')

●Cluster headaches – (see "Cluster headache: Treatment and prognosis", section on 'Oxygen')

●Acute chest syndrome in sickle cell disease – (see "Acute chest syndrome (ACS) in sickle cell disease (adults and children)", section on 'Respiratory support')

●Stroke – (see "Initial assessment and management of acute stroke", section on 'Airway, breathing, and circulation')

●Myocardial infarction – (see "Overview of the acute management of ST-elevation myocardial infarction", section on 'Therapies of unclear benefit')

●Pregnancy – (see "Critical illness during pregnancy and the peripartum period", section on 'Mechanical ventilation')

●Oxygen-sensitive hypercapnic hypoxemic respiratory failure (eg, due to chronic obstructive pulmonary disease) – (see "The evaluation, diagnosis, and treatment of the adult patient with acute hypercapnic respiratory failure", section on 'Titration of oxygen' and "Management of refractory chronic obstructive pulmonary disease", section on 'Oxygen' and "Management of refractory chronic obstructive pulmonary disease", section on 'Nocturnal noninvasive ventilation')

●Chronic pulmonary hypertension in congenital heart conditions (in some cases, oxygen saturation levels can be below SpO₂ <90 percent) – (see "Pulmonary hypertension in adults with congenital heart disease: General management and prognosis", section on 'General approach')

Oxygen monitoring — In most patients, we typically use SpO₂ to follow oxygenation. (See "Pulse oximetry".)

If oximetry is unreliable or concomitant follow-up of carbon dioxide tension is required, intermittent ABG analysis is appropriate. If patients are unstable, or multiple ABGs required, arterial access may be required. Importantly, we advise that correlation between ABG and oximetry findings be performed at least once during the course of a patient's illness, particularly in patients with impaired circulation (such as Raynaud syndrome), dark nail polish, or dark skin tone, when oximetry can underestimate the degree of hypoxemia. Further details on interpreting pulse oximetry and influence of skin color and other factors are provided separately. (See "Pulse oximetry", section on 'Interpreting the results' and "Pulse oximetry", section on 'Troubleshooting sources of error' and "Arterial blood gases" and "Intra-arterial catheterization for invasive monitoring: Indications, insertion techniques, and interpretation".)

Venous blood gases are unacceptable for measuring oxygen levels but are often considered for monitoring the pH and carbon dioxide levels in combination with SpO₂ when oximetry is accurate. (See "Venous blood gases and other alternatives to arterial blood gases".)

Address presumed cause for hypoxemia — In patients with hypoxemia, we empirically treat the suspected underlying cause while oxygen therapy is ongoing. Since the etiology of hypoxemic respiratory failure is so broad, the spectrum of treatment is also broad. Narrowing the differential is important and is discussed below. (See 'Narrowing the differential' below.)

A few examples of common life-threatening conditions that can be easily treated at the bedside are listed and discussed in the linked topics below:

●Diuresis for suspected cardiogenic pulmonary edema – (see "Treatment of acute decompensated heart failure: Specific therapies")

●Bronchodilators and corticosteroids for suspected acute exacerbation chronic obstructive pulmonary disease and asthma – (see "COPD exacerbations: Management" and "Acute exacerbations of asthma in adults: Emergency department and inpatient management")

●Fluids and antibiotics for suspected sepsis – (see "Evaluation and management of suspected sepsis and septic shock in adults")

●Incentive spirometry for suspected atelectasis – (see "Respiratory problems in the post-anesthesia care unit (PACU)" and "Overview of the management of postoperative pulmonary complications")

●Chest tube thoracotomy for suspected pneumothorax – (see "Pneumothorax: Definitive management and prevention of recurrence")

●Anticoagulation or thrombolysis for suspected pulmonary embolism – (see "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults")

●Antiarrhythmics for cardiac rate and rhythm disturbances – (see "Advanced cardiac life support (ACLS) in adults")

●Pulmonary toilet and antibiotics for suspected aspiration or pneumonia – (see "Aspiration pneumonia in adults" and "Treatment of hospital-acquired and ventilator-associated pneumonia in adults" and "Overview of community-acquired pneumonia in adults", section on 'Treatment')

●Chest tube thoracotomy for pleural effusion – (see "Thoracostomy tubes and catheters: Indications and tube selection in adults and children" and "Thoracostomy tubes and catheters: Placement techniques and complications")

●Decubitus positioning and high FiO₂ for suspected air embolism – (see "Air embolism", section on 'Treatment')

●Reversal agents for suspected hypoventilation due to sedative or toxin overdose (eg, naloxone for opioid toxicity) – (see "Benzodiazepine poisoning", section on 'Role of antidote (flumazenil)' and "Acute opioid intoxication in adults", section on 'Basic measures and antidotal therapy')

Less common bedside treatments that may need to be considered include, but are not limited to, the following:

●High FiO₂ for suspected carbon monoxide poisoning; methylene blue and discontinuation of offending agent for suspected methemoglobinemia – (see "Carbon monoxide poisoning", section on 'Management' and "Methemoglobinemia", section on 'Management (acquired/toxic)')

●Discontinuation of the transfusion with or without diuresis for suspected transfusion-related reaction – (see "Approach to the patient with a suspected acute transfusion reaction")

●Intravenous glucocorticoids for suspected adrenal crisis – (see "Treatment of adrenal insufficiency in adults")

●Intramuscular epinephrine for suspected anaphylaxis – (see "Anaphylaxis: Emergency treatment")

●Antidotes for suspected poisoning and animal or insect bites – (see "Animal bites (dogs, cats, and other mammals): Evaluation and management" and "Insect and other arthropod bites" and "General approach to drug poisoning in adults")

●Pericardiocentesis for suspected pericardial tamponade – (see "Emergency pericardiocentesis")

DIAGNOSTIC EVALUATION — If the patient is stable on initial assessment or becomes stable after initial bedside treatment, our diagnostic approach is as follows (algorithm 1):

●Perform a thorough clinical assessment. (See 'History and examination' below.)

●Draw an arterial blood gas (ABG) and basic laboratory tests. (See 'Arterial blood gas' below and 'Laboratory testing' below.)

●Obtain chest imaging and electrocardiography (ECG). (See 'Chest imaging (radiography and/or ultrasonography)' below and 'Electrocardiography' below.)

Following initial testing, the etiology can usually be determined or narrowed to a few possibilities. If needed, subsequent testing and therapies can be appropriately tailored. (See 'Narrowing the differential' below.)

History and examination — We focus on identifying any underlying comorbidities and recent changes in the patient's health.

We ask targeted questions and perform a focused examination to elucidate whether symptoms are due to a particular organ system:

●Cardiac disease (eg, known cardiac diagnosis; palpitations; dizziness; syncope; chest pain or pressure; orthopnea; altered rate or rhythm; murmurs or gallops; elevated jugular venous pressure; lower extremity, sacral, or dependent edema; cold extremities).

●Pulmonary disease (eg, known pulmonary diagnosis, cough, dyspnea, change in sputum production, wheezing, baseline level of dyspnea, any home oxygen usage, crackles, decreased breath sounds).

●Anemia (eg, hemoptysis, hematemesis, melena, hematochezia, trauma or recent accidents or falls, recent procedures, pallor, active bleeding from wounds, evidence of oropharyngeal blood, evidence of rectal bleeding).

●Infectious disease (eg, chills, malaise, sick contacts, recent travel fever, warm extremities, uncontrolled sources of infection such as indwelling lines or decubitus wounds).

●Malignancy (eg, recent pleural effusion, extensive metastatic disease impairing respiratory status, cachexia, enlarged lymph nodes, distant heart or lung sounds to suggest a pericardial or pleural effusion, respectively).

●New medications or recent medication changes (eg, medicines that decrease respiratory drive such as narcotics or cause methemoglobinemia such as lidocaine), or recreational or other drug abuse.

Laboratory testing — In most patients with acute hypoxemia, we obtain routine laboratory studies including the following:

●Basic chemistries

●Complete blood count

●Liver function studies

●Coagulation studies

●Troponin level

●N-terminal pro-brain natriuretic peptide level

Additional testing may need to be performed depending on the preliminary differential and clinical condition for suspected conditions. (See 'Additional testing based upon suspected etiologies' below.)

Arterial blood gas — In patients with hypoxemic respiratory failure, we almost always obtain an ABG, especially if inaccurate pulse oximetry is suspected, the patient has hemodynamic instability, altered mental status, severely reduced peripheral saturation, a new requirement for supplemental oxygen, or suspected concomitant hypercapnia, or there is a predicted need for non-invasive or invasive ventilation (eg, respiratory distress, labored breathing).

An ABG establishes the diagnosis of hypoxemia and can also help narrow the differential by assessment of the alveolar-arterial gradient (calculator 1) and the arterial carbon dioxide tension (table 1). (See 'Narrowing the differential' below.)

We sometimes obtain a mixed venous oxygen saturation or central venous oxygen saturation to help identify whether low cardiac output or anemia is potentially contributing to hypoxemia.

Chest imaging (radiography and/or ultrasonography) — In patients with acute hypoxemia, we obtain chest imaging, which, in many cases, will provide an explanation for the underlying etiology and is crucial for narrowing the differential. (See 'Narrowing the differential' below.)

In most patients, we obtain a chest radiograph unless emergent therapy is needed for an assumed diagnosis such as tension pneumothorax. (See "Evaluation of and initial approach to the adult patient with undifferentiated hypotension and shock", section on 'Tension pneumothorax'.)

Bedside cardiothoracic ultrasonography is being increasingly used instead of or as a supplement to chest radiography (eg, additional bedside cardiac or lower extremity ultrasonography). (See "Indications for bedside ultrasonography in the critically ill adult patient".)

Additional imaging may need to be performed depending on the initial differential and suspected etiologies (eg, chest computed pulmonary angiography for suspected pulmonary embolism). (See 'Additional testing based upon suspected etiologies' below.)

Electrocardiography — We typically obtain an ECG, especially in those with rate and rhythm disturbances, dyspnea, and/or chest pain. ECG together with troponin and brain natriuretic peptide levels can help diagnose or exclude a cardiac disorder as an etiology for hypoxemia.

NARROWING THE DIFFERENTIAL — Etiologies associated with acute hypoxemia are numerous, ranging from an erroneous pulse oximeter reading to acute respiratory distress syndrome requiring mechanical ventilation. Narrowing the differential for the underlying cause is challenging and there is no ideal approach.

●We typically rely on initial clinical assessment and initial testing including arterial blood gas (ABG) analysis to make a preliminary diagnosis for an underlying etiology. We target both therapy and any other additional testing towards that differential.

●As an alternative, some clinicians use a multi-layered "rule out approach":

•As a first step, concomitant hypercapnia is confirmed or excluded on ABG, ruling in or ruling out potential etiologies for hypoventilation (table 2), respectively. (See 'Arterial blood gas analysis' below.)

•Common life-threatening medical causes or "can't miss diagnoses" are assessed using clinical findings, initial imaging, laboratory, and electrocardiography (eg, pulmonary edema, acute myocardial infarction, pneumothorax, pneumonia, sepsis, shock). Additional testing may be needed depending on the suspicion for select conditions (eg, chest computed pulmonary angiography, COVID-19 testing, methemoglobin level). (See 'Additional testing based upon suspected etiologies' below.)

•If serious and common etiologies have been excluded and the cause remains unknown, testing for less common or rare causes depending upon the suspicion is appropriate (eg, testing for mechanical shunt, hemoglobinopathies).

Arterial blood gas analysis — The ABG can be used to distinguish shunt from ventilation/perfusion (V/Q) mismatch and hypoventilation, which may have separate etiologies (table 1). However, in practice, this distinction between mechanisms is not always that clear, since many patients can have two or more mechanisms to explain hypoxemia.

Nonetheless, the following principles may be useful:

●An elevated arterial carbon dioxide tension (PaCO₂) without an elevated alveolar-arterial (A-a) gradient may suggest etiologies associated with hypoventilation (eg, narcotic overdose). If hypoventilation is suspected, a search for etiologies associated with hypoventilation is prudent (table 2). An elevated PaCO₂ may also be found in conditions that increase dead space (eg, acute exacerbations of chronic obstructive lung disease), but the A-a gradient is usually elevated. (See "The evaluation, diagnosis, and treatment of the adult patient with acute hypercapnic respiratory failure" and "Mechanisms, causes, and effects of hypercapnia".)

●A normal PaCO₂ and a widened A-a gradient may suggest etiologies associated with V/Q mismatch (eg, acute respiratory distress syndrome) or shunt (severe multilobar pneumonia or venoarterial mechanical shunt). V/Q mismatch and shunt can in turn be distinguished by a good or poor response to oxygen supplementation, respectively. If shunt is suspected, a search for associated etiologies is prudent (eg, computed tomography of the chest, contrast echocardiography, liver ultrasonography).

●Identifying a disparity between the arterial oxygen tension (PaO₂) and the peripheral saturation (SpO₂) may also be helpful. For example, a low PaO₂ and normal SpO₂ in a patient with extremely high white cell or platelet count may suggest leukocyte or platelet larceny, while a low SpO₂, normal PaO₂, and failure of the SpO₂ to improve with supplemental oxygen may suggest methemoglobinemia. (See "Methemoglobinemia", section on 'Initial evaluation' and "Measures of oxygenation and mechanisms of hypoxemia", section on 'Spurious hypoxemia'.)

Additional testing based upon suspected etiologies

Patients with findings suggestive of specific etiologies — In many patients, suspected underlying etiologies for acute hypoxemia may be apparent on initial testing (see 'Diagnostic evaluation' above). However, some patients may require additional confirmatory diagnostic testing. Common examples include the following:

●For patients with suspected pulmonary edema, a bedside or formal echocardiogram, empiric diuresis, follow-up troponins, and electrocardiography may be warranted. (See "Approach to diagnosis and evaluation of acute decompensated heart failure in adults".)

●For patients with suspected respiratory sepsis, additional respiratory and blood cultures, respiratory viral panel, COVID-19 or influenza testing, and procalcitonin and lactate levels may be required. (See "Overview of community-acquired pneumonia in adults", section on 'Clinical presentation' and "Sepsis syndromes in adults: Epidemiology, definitions, clinical presentation, diagnosis, and prognosis", section on 'Clinical presentation' and "Seasonal influenza in adults: Clinical manifestations and diagnosis" and "Seasonal influenza in adults: Clinical manifestations and diagnosis", section on 'Diagnosis'.)

●For patients with suspected acute exacerbation of chronic obstructive pulmonary disease or asthma, peak expiratory flow values and documented response to bronchodilators may be warranted. (See "COPD exacerbations: Clinical manifestations and evaluation" and "Acute exacerbations of asthma in adults: Emergency department and inpatient management", section on 'Clinical findings'.)

●In patients with suspected tongue and airway swelling due to angioedema following angiotensin-converting enzyme inhibitor use, complement levels may be appropriate. (See "An overview of angioedema: Clinical features, diagnosis, and management".)

●In patients with hypoventilation from suspected drug intoxication, serum and urine toxin screening and alcohol level is appropriate. (See "The evaluation, diagnosis, and treatment of the adult patient with acute hypercapnic respiratory failure".)

●For patients with risk factors for methemoglobinemia (table 5) or carboxyhemoglobinemia (eg, smoke exposure), co-oximetry and carboxyhemoglobin levels, respectively, are appropriate. (See "Carbon monoxide poisoning" and "Methemoglobinemia".)

Patients with unclear etiology after initial testing — In some patients, the etiology of hypoxemia may be more subtle or multifactorial in nature and therefore, not apparent after initial testing (see 'Diagnostic evaluation' above). The following are important considerations that we keep in mind and may require additional testing:

●Pulmonary embolism may be suspected in a patient with hypoxemia, a clear chest radiograph, and risk factors for thrombosis; in such cases, chest computed pulmonary angiography or ultrasonography of the lower extremities may be warranted. (See "Clinical presentation and diagnosis of the nonpregnant adult with suspected deep vein thrombosis of the lower extremity" and "Clinical presentation, evaluation, and diagnosis of the nonpregnant adult with suspected acute pulmonary embolism".)

●Nonrespiratory infections may be suspected in patients with risk factors including venous catheters or ischemic bowel disease and may prompt obtaining cultures (including blood cultures) and additional imaging for a source. (See "Evaluation and management of suspected sepsis and septic shock in adults", section on 'Septic focus identification and source control'.)

●A mechanical shunt may be suspected or need to be excluded in patients with unexplained or severe hypoxemia that is minimally responsive to supplemental oxygen or in patients with platypnea. In such cases, a bubble echocardiogram and/or contrast-enhanced computed tomography of the chest and/or liver may be warranted. (See "Pulmonary arteriovenous malformations: Clinical features and diagnostic evaluation in adults", section on 'Transthoracic contrast echocardiography' and "Hepatopulmonary syndrome in adults: Prevalence, causes, clinical manifestations, and diagnosis", section on 'Shunt assessment' and "Clinical manifestations and diagnosis of atrial septal defects in adults", section on 'Agitated saline contrast'.)

●Hypo-or hyperthyroidism may be suspected in patients with unexplained hypoxemia who have mental status changes or arrythmias and may prompt evaluation with thyroid studies. (See "Diagnosis of and screening for hypothyroidism in nonpregnant adults" and "Diagnosis of hyperthyroidism".)

Rarely, the following may need to be considered:

●The diagnosis of pulmonary hypertension may need to be considered or excluded by right heart catheterization. (See "Clinical features and diagnosis of pulmonary hypertension of unclear etiology in adults".)

●In patients with suspected interstitial, malignant, or alveolar diseases of unclear etiology, high-resolution chest computed tomography and bronchoscopy may be needed (eg, pulmonary alveolar proteinosis, alveolar hemorrhage, acute eosinophilic pneumonia, fungal pneumonia). (See "Flexible bronchoscopy in adults: Indications and contraindications", section on 'Diagnostic indications' and "Acute respiratory distress syndrome: Clinical features, diagnosis, and complications in adults", section on 'Diagnosis and etiology is unclear'.)

SUPPORTIVE THERAPIES

●General – For patients with acute hypoxemic respiratory failure, especially those who are on advanced therapies, we adopt similar principles to those in critically ill patients including venous thromboembolism and stress ulcer prophylaxis, nutrition, and glycemic control. These issues are discussed separately:

•(See "Prevention of venous thromboembolic disease in acutely ill hospitalized medical adults".)

•(See "Stress ulcers in the intensive care unit: Diagnosis, management, and prevention" and "Acute respiratory distress syndrome: Fluid management, pharmacotherapy, and supportive care in adults", section on 'Stress ulcer prophylaxis'.)

•(See "Glycemic control in critically ill adult and pediatric patients" and "Acute respiratory distress syndrome: Fluid management, pharmacotherapy, and supportive care in adults", section on 'Glucose control'.)

•(See "Nutrition support in intubated critically ill adult patients: Initial evaluation and prescription" and "Acute respiratory distress syndrome: Fluid management, pharmacotherapy, and supportive care in adults".)

●Self-proning – We do not routinely perform awake pronation in patients with acute hypoxemic respiratory failure unless the patient has COVID-19, where data support its practice. Whether self-proning is of value in patients with early non-COVID-19-related ARDS is less clear, although weak data suggest a possible oxygenation benefit in some patients [30-32].

FOLLOW-UP

Monitoring — For patients with acute hypoxemia, particularly those who require advanced respiratory support, monitoring of the patient's respiratory status is warranted for progression or improvement.

●For patients receiving minimal oxygen requirements and who are comfortable, intermittent monitoring by clinical evaluation every four to six hours is reasonable depending on the underlying etiology.

●For patients who require advanced support (eg, high-flow oxygen via nasal cannulae [HFNC], noninvasive ventilation [NIV]), we evaluate clinically every one to two hours. We also obtain an arterial blood gas after the first one to two hours to ensure effective and safe oxygenation and ventilation. We advocate a low threshold for intubation in this population.

There is no set duration for a trial period of advanced respiratory therapy. In our experience, some patients deteriorate quickly (hours to a few days), while others tolerate advanced therapy for prolonged periods (eg, one week to 10 days) and may even switch between therapies intermittently (eg, nocturnal NIV and NIV while asleep with "HFNC breaks" during the day or while eating).

There are no factors that reliably predict which course a patient will experience. Invariably, patients who require advanced therapies for prolonged periods or who progress despite such therapy are at high risk of requiring intubation and mechanical ventilation. (See 'Invasive mechanical ventilation' below.)

Patients who improve — For patients who improve, we de-escalate oxygen supplementation to the lowest possible level or to the patient's baseline oxygen requirement. There is no specific protocol for weaning oxygen. Oxygen therapy can be discontinued once oxygen saturations are at goal at rest on room air.

For some patients, de-escalation can take hours (eg, pneumothorax), while for others it may take a few days or longer (eg, pneumonia) and ultimately depends upon the rate of resolution of the underlying disease.

We do not routinely assess ambulatory or nocturnal oximetry but do evaluate select patients for oxygen needs during ambulation and sleep. Examples of suitable patients include frail or weak patients, patients with tenuous respiratory status, and patients with suspected pulmonary hypertension or with residual or slowly resolving lung disease. In such cases, further de-escalation may be feasible before or after discharge when resolution of the underlying disorder is complete. For ambulatory needs, a formal six-minute walk is not necessary and we typically assess ambulatory oxygenation during walking for as long as the patient can tolerate it (generally a few minutes). (See "Long-term supplemental oxygen therapy" and "Portable oxygen delivery and oxygen conserving devices".)

Patients with escalating oxygen needs — Some patients will continue to have worsening hypoxemia despite optimal management of the underlying disorder. Options include increasing the amount of delivered oxygen or ventilating the patient (noninvasively). Choosing among the options depends upon the patient's clinical characteristics, and in many cases, the patient may transition between both higher flows of oxygen and NIV. It is prudent at this juncture to simultaneously determine the reason for clinical deterioration and treat it (eg, progression or complication of the underlying disorder or second etiology such as fluid overload, pulmonary embolism, aspiration). Should patients fail these noninvasive options, intubation is typically indicated.

Increase oxygen supplementation — For patients with worsening hypoxemia who do not require immediate ventilation, we steadily increase the flow or fraction of inspired oxygen (FiO₂) to target the original set goal.

●For patients on low-flow oxygen, this may require increasing the flow rate or transitioning to humidified HFNC. The maximal amount of oxygen that can be delivered in low-flow systems is only 6 to 10 L/minute, at which point patients may need to be transitioned to a high-flow system.

●For patients on HFNC, this involves increasing the flow or FiO₂, usually both. The maximal amounts that can be delivered in high-flow systems is 1.0 FiO₂ and 60 liters/minute (L/min) of flow (sometimes higher). Once patients are close to maximal flow rates and FiO₂ on HFNC (eg, 60 L/min and 0.6 FiO₂), we begin evaluation for intubation. (See 'Humidified, high-flow oxygen delivered via nasal cannulae (HFNC)' above and "Heated and humidified high-flow nasal oxygen in adults: Practical considerations and potential applications".)

Noninvasive ventilation — Patients with severe or rapidly escalating oxygen requirements may benefit from NIV, the indications for which are discussed separately. (See 'Patients with indications for noninvasive ventilation' above and "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications", section on 'Patients likely to benefit' and "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications", section on 'Patients less likely to benefit' and "Respiratory muscle weakness due to neuromuscular disease: Management", section on 'Acute ventilatory support'.)

Depending on the NIV device, the FiO₂ can be titrated (in many newer devices up to 1.0) to provide enough oxygen to reach the target goal. Most clinicians agree that trials of NIV should be short (eg, two hours) and evidence of further deterioration or failure of NIV should prompt early intubation.

The application, indications, and contraindications (table 4) to NIV are discussed separately. (See "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications" and "Noninvasive ventilation in adults with acute respiratory failure: Practical aspects of initiation".)

Invasive mechanical ventilation — For patients who fail HFNC and/or NIV, patients with severe respiratory distress, or patients who are hemodynamically unstable, intubation and mechanical ventilation is appropriate, provided it is compliant with the patient's wishes.

The optimal timing of intubation is individualized and is discussed separately. (See "The decision to intubate" and "COVID-19: Respiratory care of the nonintubated hypoxemic adult (supplemental oxygen, noninvasive ventilation, and intubation)", section on 'The decision to intubate'.)

The management of refractory hypoxemia on mechanical ventilation for patients with acute respiratory distress syndrome is provided separately. (See "Acute respiratory distress syndrome: Ventilator management strategies for adults", section on 'Refractory patients'.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Supplemental oxygen".)

SUMMARY AND RECOMMENDATIONS

●Initial bedside evaluation and treatment – When patients present with suspected hypoxemia, we simultaneously perform the following (algorithm 1) (see 'Initial triage and treatment' above):

•Assess and stabilize the airway, breathing, and circulation – We evaluate the patient's airway for patency, their breathing pattern, pulse rate and rhythm, temperature, and blood pressure, all of which should be stabilized, if compromised. (See 'Assess and stabilize airway, breathing, and circulation' above.)

•Immediately administer supplemental oxygen (figure 1). (See 'Oxygen and respiratory support' above.)

-Patients with minimal oxygen requirements – In most patients with acute hypoxemic respiratory failure who have minimal oxygen requirements, low-flow oxygen via nasal cannulae is usually sufficient. (See 'Patients with minimal oxygen requirements (eg, ≤6 L/min)' above.)

-Patients who need advanced respiratory support – For patients who have a need for advanced respiratory support, options include humidified high-flow oxygen via nasal cannulae (HFNC) and noninvasive ventilation (NIV). High-flow oxygen via a low-flow system is an alternative if HFNC or NIV are not available or feasible.

For patients in whom there is a known benefit and no contraindications (table 4), we administer a trial of NIV. Data to support NIV are discussed separately. (See 'Patients with requirements for advanced respiratory support' above and 'Patients with indications for noninvasive ventilation' above and "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications", section on 'Patients likely to benefit' and "Respiratory muscle weakness due to neuromuscular disease: Management", section on 'Noninvasive ventilation'.)

For all other patients who need advanced respiratory support, we suggest HFNC (Grade 2C). HFNC is a comfortable modality that effectively improves oxygenation. However, HFNC does not impact mortality, and its effect on intubation rates, length of stay, dyspnea, and comfort are uncertain. (See 'Humidified, high-flow oxygen delivered via nasal cannulae (HFNC)' above.)

-Oxygen monitoring and targets – We typically use peripheral oxygen saturation (SpO₂) to follow oxygenation. Intermittent arterial blood gas (ABG) analysis may be required (eg, unreliable oximetry). In most patients, we typically target an SpO₂ between 90 and 96 percent but individualize targets for specific populations. (See 'Oxygen monitoring' above and 'Oxygen saturation goals' above and "Pulse oximetry".)

•Initiate empiric treatment for the suspected underlying cause – For all patients with hypoxemia, we target the suspected etiology (eg, diuresis for heart failure, bronchodilators for wheeze, antibiotics for infection). (See 'Address presumed cause for hypoxemia' above.)

●Diagnostic evaluation – Once the patient is stable, we perform the following:

•A thorough clinical assessment. (See 'History and examination' above.)

•An ABG and basic laboratory tests (eg, chemistries, complete blood count, liver function tests, coagulation studies, troponin level, N-terminal pro-brain natriuretic peptide level). (See 'Arterial blood gas' above and 'Laboratory testing' above.)

•Chest imaging (radiography and/or ultrasonography) and electrocardiography. (See 'Chest imaging (radiography and/or ultrasonography)' above and 'Electrocardiography' above.)

●Narrowing the differential – Following initial testing, the etiology can usually be determined or narrowed to a few possibilities (table 1 and algorithm 1). If needed, subsequent testing can be appropriately tailored (eg, echocardiogram, computed tomography of the chest and abdomen, specialized cultures, COVID-19 testing, toxin screening). (See 'Narrowing the differential' above.)

In some patients, the etiology of hypoxemia may be more subtle or multifactorial and require additional testing. Important considerations include pulmonary embolism, nonrespiratory infections, mechanical shunt, hypo-or hyperthyroidism, pulmonary hypertension (PH), or interstitial or alveolar lung diseases. (See 'Patients with unclear etiology after initial testing' above.)

●Monitoring – For patients who are comfortable and receiving minimal oxygen requirements, intermittent monitoring by clinical evaluation every four to six hours is reasonable. For patients who require advanced support, we evaluate clinically every one to two hours and also obtain an ABG after the first one to two hours; we advocate a low threshold for intubation in this population. (See 'Monitoring' above.)

•Patients who improve – For patients who improve, we de-escalate oxygen supplementation to the lowest possible level or to the patient's baseline oxygen requirement. We do not routinely perform ambulatory or nocturnal oximetry unless indicated (eg, frail patients, tenuous respiratory status, suspected PH, slowly resolving lung disease). (See 'Patients who improve' above.)

•Patients with increasing oxygen needs – For patients with escalating oxygen needs, we either increase oxygen supplementation and/or trial NIV and determine the reason for clinical deterioration (eg, progression or complication of the underlying disorder or second etiology such as fluid overload, pulmonary embolism, aspiration). Should patients fail these noninvasive options, intubation is typically indicated. (See 'Patients with escalating oxygen needs' above.)