The evaluation and management of the nonventilated adult with acute hypercapnic respiratory failure

INTRODUCTION — 

Acute hypercapnic respiratory failure may become life-threatening if untreated, potentially resulting in respiratory arrest, seizures, coma, and/or death.

This topic discusses the approach to the spontaneously breathing adult patient with acute hypercapnic respiratory failure. The etiologies and end-organ effects of hypercapnia and the approach to patients with acute hypoxemic respiratory failure are discussed separately. (See "Mechanisms, causes, and effects of hypercapnia" and "Evaluation and management of the nonventilated, hospitalized adult patient with acute hypoxemia".)

SUSPECTING AND DIAGNOSING HYPERCAPNIA

Features suggestive of hypercapnia — Acute hypercapnic respiratory failure should be suspected in patients with risk factors (eg, sedative use, obstructive sleep apnea, chronic obstructive pulmonary disease exacerbation) who have dyspnea and/or altered sensorium (eg, hypersomnolence). Hypercapnia should also be suspected in patients with a markedly elevated bicarbonate. The etiologies and mechanisms of hypercapnia are presented in the table (table 1) and discussed elsewhere. (See "Mechanisms, causes, and effects of hypercapnia".)

Features vary depending upon the level and rate of carbon dioxide (CO₂) accumulation in arterial blood:

●Patients with mild to moderate acute hypercapnia may be anxious and/or complain of mild dyspnea, daytime sluggishness or hypersomnolence, or a headache.

●Patients with severe and rapidly developing hypercapnia develop delirium, paranoia, depression, and confusion. Signs include asterixis, myoclonus, and seizures, as well as papilledema and dilated superficial veins. As levels rise, somnolence and then coma (CO₂ narcosis) ensue.

The development of symptoms also differs between patients who are normocapnic or hypercapnic at baseline. Normocapnic individuals exhibit a depressed level of consciousness when the arterial CO₂ tension (PaCO₂) is >75 to 80 mmHg (9.9 to 10.9 kPa). Patients with chronic hypercapnia may not develop symptoms until the PaCO₂ rises acutely to >90 to 100 mmHg (11.9 to 13.3 kPa). The mechanisms that underlie the cerebral and cardiorespiratory effects of hypercapnia are discussed separately. (See "Mechanisms, causes, and effects of hypercapnia", section on 'Effects of hypercapnia'.)

Patients with hypercapnia due to hypoventilation have hypopnea (ie, low respiratory rate and/or shallow breathing) (table 1). A common misperception is that the presence of tachypnea implies that hypercapnia is not present. However, patients with hypercapnia due to increased dead space and thoracic cage abnormalities often have tachypnea and accessory muscle use. Likewise, patients with neuromuscular weakness may develop rapid shallow breathing, leading to an increase in dead space. (See "Mechanisms, causes, and effects of hypercapnia", section on 'Formula for arterial carbon dioxide tension'.)

A commonly missed presentation of oxygen-induced acute hypercapnic respiratory failure is the onset of hypoxemia due to hypoventilation in the postoperative patient. Supplemental oxygen may transiently improve peripheral oxygen saturation but worsens the hypercapnia, which in turn leads to worsening hypoxemia and further increases in supplemental oxygen. If not detected, this cycle invariably results in hypercapnic respiratory failure and CO₂ narcosis. Mechanisms and management of oxygen-induced hypercapnia are discussed separately. (See "Mechanisms, causes, and effects of hypercapnia", section on 'Oxygen-induced hypercapnia' and 'Managing oxygen and ventilatory needs' below.)

Diagnostic blood gas — Every patient with suspected hypercapnia should have an arterial blood gas (ABG) or venous blood gas (VBG) analysis. The identification of an elevated CO₂ tension (PCO₂) >45 mmHg is diagnostic of hypercapnia.

Due to greater accuracy, we prefer that an ABG rather than a VBG be performed. Peripheral VBGs tend to have a slightly higher PCO₂ and bicarbonate level and a lower pH (table 2). VBGs can be used if ABGs cannot be obtained or if frequent samples are necessary to follow therapy. Detailed discussion on the use and interpretation of VBGs is provided separately. (See "Venous blood gases and alternatives to arterial carbon dioxide measurement in adults".)

Further blood gas analysis determines whether hypercapnia is acute or chronic, whether it is a primary or secondary phenomenon, and can assist the clinician when formulating the differential diagnosis. Further details regarding that analysis are discussed below. (See 'Additional blood gas analysis' below.)

INITIAL EMERGENCY MANAGEMENT — 

If needed, we stabilize the airway and support breathing with oxygen and/or ventilatory assistance. We also secure intravenous access while laboratory tests and imaging are being performed. (See 'Brief clinical assessment' below.)

●Airway – We assess patients for upper airway obstruction, which could be the result of carbon dioxide narcosis or a contributing factor in the development of hypercapnia (eg, edema, foreign body, oropharyngeal collapse due to sleep apnea). A chin lift/jaw thrust can be used to achieve airway patency and allow airway inspection. Etiologies and assessment of the upper airway are described in detail separately. (See "Postoperative airway and pulmonary complications in adults: Etiologies and initial assessment and stabilization", section on 'Acute upper airway obstruction' and "Postoperative airway and pulmonary complications in adults: Etiologies and initial assessment and stabilization", section on 'Ensure a patent upper airway and adequate ventilation'.)

●Oxygen – Oxygen should be administered to patients with evidence of clinically significant hypoxemia (eg, peripheral oxygen saturation [SpO₂] <90 percent) to maintain an SpO₂ between 90 and 93 percent. Managing oxygen titration to avoid oxygen-induced hypercapnia and ideal oxygen targets are provided below. (See 'Managing oxygen and ventilatory needs' below.)

When oxygen-induced hypercapnic respiratory failure is suspected (see 'Features suggestive of hypercapnia' above) and the SpO₂ is in the high normal range (eg, ≥95 percent), we gradually reduce supplemental oxygen to maintain SpO₂ between 90 and 93 percent. This strategy (with or without noninvasive ventilation [NIV]) may avoid mechanical ventilation and identify at-risk patients during future admissions. (See 'Low-flow oxygen titration' below and 'Patients at risk for oxygen-induced hypercapnia' below.)

●Ventilatory support – Some patients may require ventilatory support including NIV, bag-mask ventilation, or intubation. For example, bag-mask ventilation or NIV may be used to temporarily ventilate patients in whom reversal of hypercapnic respiratory failure is expected (eg, rapid reversal within minutes [bag-mask]; slower reversal within hours [NIV]). Before initiating NIV, one must be sure the patient is alert enough to protect their airway to minimize the chances of aspiration. Patients with severe respiratory distress are usually intubated. Further details on providing ventilatory support are provided separately. (See 'Patients with advanced respiratory support or oxygen needs' below and "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications" and "Noninvasive ventilation in adults with acute respiratory failure: Practical aspects of initiation" and "Rapid sequence intubation in adults for emergency medicine and critical care" and "Direct laryngoscopy and endotracheal intubation in adults" and "Basic airway management in adults", section on 'Bag-mask ventilation' and "Emergency cricothyrotomy (cricothyroidotomy)".)

●Intravenous access – Peripheral venous access is usually sufficient for initial management.

NARROWING THE DIFFERENTIAL

Brief clinical assessment — This assessment is targeted at identifying the underlying cause (table 1) so it can be treated promptly. A general approach is outlined in the table (table 3).

●History and examination – We obtain a brief history from prehospital or hospital providers, the patient (if feasible), their relatives or caregivers, and/or the medical record. This includes a review of prescribed medications (eg, sedatives, neuromuscular blockers, procainamide); illicit drug ingestion; potential toxins (eg, tetanus, botulism, shellfish, organophosphates); a history of chronic lung disease, obstructive sleep apnea (OSA), or a neuromuscular disorder; a history of smoking; and a history of recent trauma. For patients on oxygen, we ask about recent changes in the flow settings.

On examination, we assess the following:

•Temperature (hyper- or hypothermia) and respiratory rate.

•The patient's depth of breathing (eg, deep or shallow).

•Mental status.

•Body habitus (for obesity or evidence of malnutrition).

•Oropharynx and neck circumference (for OSA or obesity hypoventilation syndrome).

•Lungs for evidence of chronic lung disease (eg, wheeze and hyperinflation to suggest asthma or chronic obstructive pulmonary disease [COPD], inspiratory crackles to suggest interstitial lung disease [ILD] or pneumonia).

•Thoracic cage for kyphosis or other structural abnormalities (eg, pectus excavatum, thoracoplasty).

•Brief neurologic examination to look for any potential central or peripheral neurologic etiology responsible for acute and/or chronic hypercapnia (eg, evidence of neuromuscular weakness or diminished air entry in one lung [may suggest unilateral diaphragmatic paralysis]).

•Paradoxical movement of the abdomen, which may indicate diaphragmatic weakness or paralysis; weakness of the diaphragm is also frequently associated with use of accessory muscles of ventilation in the upper chest (eg, sternocleidomastoid) as the patient attempts to compensate.

●Laboratory assessment – We order the following laboratory studies:

•A complete blood count and differential – Chronic hypoxemia from underlying lung disease may be associated with polycythemia. Although rare, an elevated eosinophil fraction on a complete blood count may be consistent with eosinophilic myalgia. Elevated eosinophils may also prompt suspicion of reactive airways disease as a contributor to hypercapnia.

•Serum chemistries including bicarbonate and electrolytes (eg, potassium, sodium, phosphate, calcium, magnesium) – Altered chemistries may contribute to hypercapnia (especially hypophosphatemia and hypomagnesemia, which lead to muscle weakness). An elevated serum bicarbonate may suggest chronic hypercapnia or evidence of a primary metabolic alkalosis.

•Others – Additional laboratory tests may be obtained on a case-by-case basis (eg, toxicology screen if overdose is suspected, thyroid function tests if hypothyroidism is suspected or the patient is on thyroid hormone replacement).

●Chest radiograph – We obtain a chest radiograph in most patients to look for intrapulmonary pathology that may be associated with the development of hypercapnia (eg, new pneumonia, atelectasis, pulmonary edema, oligemia to suggest pulmonary embolism, COPD, ILD, thoracic cage disorder).

●Others – Additional, less commonly performed imaging, laboratory, and physiologic testing targeted at a suspected underlying etiology is discussed below. (See 'Additional testing for underlying etiology' below.)

Additional blood gas analysis — The original diagnostic blood gas can be used to distinguish acute, chronic, or acute-on-chronic hypercapnia and help narrow the differential.

●Distinguish acute and chronic hypercapnia – Acute hypercapnia needs more immediate attention than chronic hypercapnia. While arterial blood gas (ABG) interpretation is influenced by the presence of mixed and complex acid-base disturbances (see "Simple and mixed acid-base disorders"), we use the following general principles:

•Acute hypercapnia – Acute hypercapnia occurs when an abrupt failure of ventilation occurs (eg, central nervous system disease or drug-induced respiratory depression, severe acute airways disease) or when there is an acute process that worsens alveolar ventilation in a patient with limited pulmonary reserve (eg, pneumonia in a patient with COPD).

Acute hypercapnia is always accompanied by a respiratory acidosis (pH <7.35). In patients with previous normocapnia, the measured level of hypercapnia and bicarbonate accurately predicts the pH (figure 1 and figure 2). In mild acute respiratory acidosis, the serum bicarbonate typically rises by 2 to 5 mEq/L as CO₂ and water combine to form carbonic acid, which dissociates into a bicarbonate molecule and a proton. Buffering of carbonic acid in red blood cells by hemoglobin as CO₂ diffuses into the red cell also contributes to the acute rise in serum bicarbonate. A quick rule of thumb is that an acute change in the arterial CO₂ tension (PaCO₂) of 10 mmHg leads to a 0.08 unit change in pH; in contrast, a chronic change in PaCO₂, with appropriate metabolic compensation, leads to a 0.03 unit change in pH.

•Chronic hypercapnia – Chronic hypercapnia develops over days to weeks such that bicarbonate rises as an indication that renal compensation (ie, secretion of acid) has occurred (the latter usually takes >3 to 5 days for full compensation to occur). Unlike acute hypercapnia, chronic hypercapnia is associated with a low-normal or near-normal pH (assuming the patient has normal kidney function and the ability to excrete acid by the kidney). In mild to moderate chronic compensated respiratory acidosis, hypercapnia is associated with a pH at the lower limit of normal or near-normal pH (eg, pH 7.33 to 7.35) (figure 1 and figure 3) depending upon the magnitude of the elevated PaCO₂.

It is important to note that the metabolic compensation for chronic respiratory acidosis is the secretion of acid by the kidney; the elevated serum bicarbonate in this scenario is a byproduct of acid secretion and does not play a role in buffering the respiratory acid.

While an elevated bicarbonate level suggests underlying chronic hypercapnia at baseline, it is nonspecific since other etiologies associated with a primary metabolic alkalosis can increase bicarbonate concentration (eg, volume contraction, diuretics, hypochloremia).

Comparison of the ABG with previous ABGs drawn on prior admissions or in the chronic stable state may help identify the presence of normocapnia or chronic hypercapnia at baseline.

•Acute-on-chronic – Acute-on-chronic respiratory acidosis has components of both acute and chronic hypercapnia such that the measured pH will be higher than predicted for acute hypercapnia but lower than that predicted for chronic hypercapnia.

•Mixed – When hypercapnia is associated with an alkaline pH, we consider a mixed acid-based disorder (eg, a primary metabolic alkalosis with a secondary, compensatory respiratory acidosis); with correction of the metabolic alkalosis, the hypercapnia should resolve.

●Determine the alveolar-arterial gradient — Hypoxemia frequently coexists with hypercapnic respiratory failure. When feasible, we calculate the alveolar-arterial gradient (alveolar oxygen tension [PAO₂] – arterial oxygen tension [PaO₂]; also known as the A-a gradient) from a room air blood gas (calculator 1). A normal A-a gradient in the setting of an elevated PaCO₂ is indicative of total (also known as global) hypoventilation (ie, alveolar plus dead space ventilation) whereas a widened gradient suggests that increased dead space (ie, ventilation/perfusion [V/Q] mismatch) from underlying lung disease may be contributing to the measured hypercapnia.

However, accurately measuring the A-a gradient is difficult in patients with acute hypercapnic respiratory failure since they are often receiving supplemental oxygen through nasal cannulae or masks (figure 4 and figure 5) at the time of evaluation; consequently assessing the delivered fraction of inspired oxygen (FiO₂) is often inaccurate. Nonetheless, it can be reliably determined in situations where the FiO₂ is assured (eg, mechanical ventilation).

Further details on measuring the A-a gradient are provided separately. (See "Measures of oxygenation and mechanisms of hypoxemia", section on 'Alveolar-arterial (A-a) oxygen gradient'.)

Formulating a preliminary diagnosis — A preliminary diagnosis (table 1) is usually formulated using clinical assessment (table 3) and ABG analysis. (See 'Brief clinical assessment' above and 'Additional blood gas analysis' above.)

We encourage an individualized approach to determining the etiology. The approach below is a suggested one.

Patients with symptomatic hypercapnia and low respiratory rate and/or shallow breathing (hypoventilation) — Patients with hypercapnia with a low respiratory rate and/or shallow tidal breathing (ie, "can't breathe" [central nervous system issues] or "won't breathe" [peripheral neuromuscular issues]) may have a dominant component of hypoventilation. In such cases, ventilatory assistance is often required and causes of hypoventilation should be investigated and reversed, if feasible. As examples:

●Global hypoventilation from sedatives is likely the primary cause of acute hypercapnia in an otherwise healthy young individual with narcotic drug use who has a normal chest radiograph, positive toxicology screen, and a normal A-a gradient (eg, <20 mmHg in young individuals).

●Hypoventilation may also be the likely cause in a patient with normal chest radiography and normal A-a gradient who has trauma of the spinal cord, large stroke, diaphragmatic paresis, Guillain-Barré syndrome, myasthenia gravis crisis, or toxin poisoning.

●Oxygen-induced hypercapnia is suggested in a patient (especially postoperative setting) with worsening hypersomnolence and desaturation following an increase in supplemental oxygen resulting in a peripheral oxygen saturation between 98 and 100 percent. (See 'Patients at risk for oxygen-induced hypercapnia' below and 'Features suggestive of hypercapnia' above.)

Patients with symptomatic hypercapnia and excessive respiratory effort (increased dead space) — Patients with symptomatic hypercapnia and excessive respiratory effort (eg, tachypnea, respiratory distress) often have a dominant component of increased dead space ventilation due to underlying parenchymal lung disease or thoracic cage disease (ie, V/Q mismatch; "can't breathe enough"). They have reduced ventilatory reserve; at baseline, they may be able to sustain a normal PaCO₂, but with an acute problem that worsens V/Q mismatch or increases CO₂ production, they are no longer able to sustain a normal PaCO₂.

This includes patients with an acute COPD exacerbation or end-stage ILD and is typically associated with the usual clinical signs and symptoms of COPD or ILD, a chest radiograph with abnormal findings of COPD or ILD, and an ABG that reveals hypercapnia and a widened A-a gradient. In such cases, ventilatory assistance may or may not be needed and investigations should focus on parenchymal lung and thoracic cage diseases.

Patients with mixed findings — Some patients may have mixed findings. For example:

●Hypercapnia due to hypoventilation and increased dead space may be seen in patients with neuromuscular disorders who develop pneumonia and mucus plugging due to dehydration. Often these patients complain of dyspnea but breathe shallowly.

●It may also be seen in patients with hypoventilation from sedative use who have underlying lung disease.

●Upper airway disorders are rare causes of hypercapnia. They either diminish total ventilation or lead to dynamic hyperinflation and reduced tidal volume while simultaneously causing increased work of breathing and CO₂ production.

Asymptomatic hypercapnia — Some patients who have an ABG or venous blood gas drawn have hypercapnia (even significant hypercapnia, such as >65 mmHg), but do not appear to be very symptomatic. In such cases, hypercapnia is likely chronic in nature or due to significant alkalemia and acute ventilatory assistance is unlikely to be indicated or beneficial. However, chronic nocturnal ventilatory assistance may be indicated. Further details are provided separately. (See "Nocturnal ventilatory support in COPD" and "Treatment and prognosis of the obesity hypoventilation syndrome".)

Additional testing for underlying etiology — For most patients, the underlying etiology (table 1) is apparent after the initial clinical assessment (table 3). However, in some patients, the etiology is not apparent and may require additional attention. For such cases, we perform a more detailed history and examination with a specific etiology in mind (eg, obesity hypoventilation, neuromuscular weakness) and consider targeted testing to confirm a suspected etiology.

Additional tests that can be obtained on an individual basis to identify suspected etiological risk factors include the following:

●Chest imaging – Chest computed tomography (CT; without contrast) may reveal underlying COPD (eg, emphysema, hyperinflation, flattened diaphragms) and ILD (reticular nodular shadows), as well as thoracic cage abnormalities (kyphoscoliosis, pectus excavatum, ankylosing spondylitis, fractured ribs) or elevated diaphragm (to suggest diaphragmatic paralysis). The latter may also be readily diagnosed with point-of-care diaphragmatic ultrasound.

●Brain and spinal cord imaging – In patients with suspected central or spinal hypoventilation (eg, history of trauma), CT or magnetic resonance imaging of the neck or brain may reveal additional central (especially brainstem) or peripheral nervous system etiologies for hypercapnia (eg, stroke, tumor, traumatic transection of the spinal cord).

●Creatine phosphokinase – If a primary muscle pathology is suspected, we obtain creatine phosphokinase levels. If elevated, this may suggest infectious or autoimmune polymyositis, hypothyroidism, rhabdomyolysis (eg, secondary to colchicine or chloroquine toxicity), or procainamide myopathy. Further investigations, such as muscle biopsy, can be obtained on an individual basis.

●Pulmonary function testing – In patients with myasthenic crisis, serial bedside measurement of vital capacity (VC) and negative inspiratory force is of value in predicting respiratory failure progression or resolution. A VC <1 L in a patient with normal baseline lung function warrants observation in an intensive care unit because of the increased risk of acute respiratory failure. Notably, the fall in VC may occur prior to the rise in PaCO₂. Further details are provided separately. (See "Myasthenic crisis", section on 'Assessment of respiratory function' and "Myasthenic crisis", section on 'Weaning from ventilatory support'.)

Interrogating the medical record for prior pulmonary function tests (PFTs) may be valuable in detecting subtle signs of underlying obstructive or restrictive lung disease. Following recovery of acute hypercapnia, follow-up PFTs at a later date are also reasonable.

Maximal inspiratory or expiratory pressures and maximum voluntary ventilation may be needed to detect an underlying respiratory neuromuscular disorder. (See "Respiratory muscle weakness due to neuromuscular disease: Clinical manifestations and evaluation", section on 'Diagnostic evaluation'.)

●Others – A sleep study is needed when underlying sleep apnea is suspected. (See "Clinical presentation and diagnosis of obstructive sleep apnea in adults", section on 'Diagnostic tests'.)

DEFINITIVE MANAGEMENT — 

Definitive management involves treating the presumed cause, which may be confirmed with testing or a response to empiric treatment. Supportive therapy for hypoxemia with oxygen and ventilatory assistance is often (but not always) needed. When oxygen is administered, we ensure careful titration to avoid oxygen-induced hypercapnia. Ideal oxygen and carbon dioxide (CO₂) targets should be adhered to.

Treat presumed cause of hypercapnia — Definitive management covers a wide array of etiologies. As examples:

●In patients with hypoventilation due to sedative or neuromuscular blockade overdose, the underlying cause is reversible. We may allow the sedative to wear off while supporting the patient with noninvasive ventilation (NIV). Alternatively, we may reverse sedatives or neuromuscular blocking agents if the indication is present (eg, significant somnolence, pending respiratory arrest). Further details on agent reversal are provided separately. (See "Benzodiazepine poisoning", section on 'Role of antidote (flumazenil)' and "Acute opioid intoxication in adults", section on 'Management' and "Clinical use of neuromuscular blocking agents in anesthesia", section on 'Reversal of neuromuscular block'.)

●For patients with hypoventilation in whom quick reversal is less likely (eg, trauma of the spinal cord, large stroke, diaphragmatic paresis, Guillain-Barré syndrome, myasthenic crisis), we initiate ventilatory support (invasive or noninvasive) as recovery is ongoing. Details regarding ventilatory support in patients with neuromuscular disease are provided separately. (See "Respiratory muscle weakness due to neuromuscular disorder: Management", section on 'Acute ventilatory support'.)

●For patients with oxygen-induced hypercapnia, we gradually reduce the fraction of inspired oxygen (FiO₂) to achieve a reasonable target saturation of 90 to 93 percent. NIV support may be needed. Reasonable oxygen targets are discussed below. (See 'Oxygen and carbon dioxide targets' below and 'Low-flow oxygen titration' below and 'Patients at risk for oxygen-induced hypercapnia' below.)

●For patients with increased dead space-related hypercapnia due to intrinsic lung disease, such as a chronic obstructive pulmonary disease (COPD) exacerbation, hypercapnia should respond to treatment with bronchodilators and corticosteroids, usually with NIV support. A similar approach is reasonable in patients with hypercapnia due to heart failure or interstitial lung disease exacerbations. (See "COPD exacerbations: Management" and "Treatment of acute decompensated heart failure: Specific therapies" and "Acute interstitial pneumonia (Hamman-Rich syndrome)".)

●In patients with neuromuscular disorders who have pneumonia and mucus plugging, we would treat them with antibiotics and fluids while supporting ventilation with NIV. (See "Treatment of community-acquired pneumonia in adults who require hospitalization".)

General measures also include the following:

●We reverse metabolic disturbances that may be contributing to hypoventilation (eg, hypophosphatemia, hypomagnesemia). (See "Hypophosphatemia: Evaluation and treatment".)

●We avoid sedatives in this population. However, in rare cases, such as hypercapnia associated with anxiety and rapid shallow breathing in patients with significant obstructive lung disease and flow limitation, small doses of an anxiolytic or morphine may be beneficial by reducing dynamic hyperinflation associated with tachypnea [1].

Managing oxygen and ventilatory needs

Patients at risk for oxygen-induced hypercapnia — While most patients do not retain CO₂ when supplemental oxygen is administered (nonretainers), a small proportion do retain CO₂ ("CO₂ retainers;" approximately 10 percent). In our experience, the following patients are at increased risk:

●COPD – The risk of oxygen-induced hypercapnia is best studied and most commonly seen in patients with COPD. There are three possible outcomes when administering uncontrolled oxygen therapy to a patient with COPD and respiratory insufficiency [2]:

•The patient's clinical state and arterial CO₂ tension (PaCO₂) may improve or not change.

•The patient may become drowsy but can be roused to cooperate with therapy; in these cases, the PaCO₂ generally rises slowly by up to 20 mmHg.

•The patient rapidly becomes unconscious and the PaCO₂ rises at a rate of ≥30 mmHg/hour.

Although risk factors are poorly described, the following patients with COPD are at risk for developing oxygen-induced hypercapnia and require vigilant oversight [2-4]:

•Patients with a history of CO₂ retention at baseline (sometimes suggested by an elevated serum bicarbonate).

•Patients with a history of oxygen-induced CO₂ retention.

•Patients with a low initial pH (<7.33) and/or arterial oxygen tension (PaO₂).

●Neuromuscular disease – The risk of oxygen-induced hypercapnia in patients with neuromuscular disease (eg, amyotrophic lateral sclerosis) is less well-studied. One retrospective series found that among eight patients identified with neuromuscular disease, seven had baseline hypercapnia and six worsened their hypercapnia by a mean of 28 mmHg after receiving low-flow supplemental oxygen [5]. There were no measurements of breathing pattern, respiratory drive, or dead space, and it is unclear if these findings are due to the same mechanisms that are operative in COPD.

●Postoperative patients – Oxygen-induced hypercapnia is commonly missed in this population, although we are unaware of data describing this risk.

The mechanism underlying oxygen-induced hypercapnia is described separately. (See "Mechanisms, causes, and effects of hypercapnia", section on 'Oxygen-induced hypercapnia'.)

Patients with minimal oxygen and ventilatory needs

Low-flow oxygen titration — For most patients with acute hypoxemic hypercapnic respiratory failure with minimal oxygen (eg, <6 L/minute) and ventilatory needs (eg, mild or no acute respiratory acidosis), we initiate or increase oxygen therapy with an FiO₂ to reach our target (eg, peripheral oxygen saturation [SpO₂] 90 to 93 percent) (see 'Oxygen and carbon dioxide targets' below) [6,7]. These patients do not typically need ventilatory support.

●Our oxygen titration approach – We start with low-flow oxygen devices (eg, 1 to 2 L/minute nasal cannulae or 24 to 28 percent FiO₂ via Venturi mask; the latter achieves greater FiO₂ control than a simple nasal cannula (figure 4)). If the target goal is not achieved, we gradually increase the FiO₂ in increments of 1 L/minute (via nasal cannula) or 4 to 7 percent (via Venturi mask). Once the target goal is achieved, we obtain an arterial blood gas (ABG; or venous blood gas) and, if acceptable, we maintain supplemental oxygen until the underlying disorder improves and oxygen can be weaned. In contrast, a fall in the SpO₂ below 88 percent may suggest oxygen-induced hypercapnia and should prompt another ABG. Should hypercapnia be worsened by the administration of oxygen, NIV may be indicated rather than an increase in oxygen supplementation. (See 'Patients with advanced respiratory support or oxygen needs' below.)

●Avoiding prolonged hypoxemia – Oxygen should not be removed entirely from the patient with hypercapnia and hypoxemia, and prolonged episodes of severe hypoxemia should be avoided during titration as this may aggravate tissue hypoxemia and worsen acidosis. Thus, this approach works best for patients with minimal oxygen needs who respond quickly (within 15 minutes) or patients who have no complications of hypoxemia (eg, chest pain, altered mental status, lactic acidosis).

●Frequency of incremental increases – The optimal rate of incremental increases in supplemental oxygen is unknown. We increase the FiO₂ every 5 to 15 minutes. Shorter periods of 5 minutes are usually adequate for patients with no underlying ventilation/perfusion mismatch and normal airways. In contrast, patients with significant mismatch and airways disease require longer periods for the change in FiO₂ to be fully reflected at the alveolar level. This principle is explained by the time constant of the lung (time constant = resistance x compliance; ie, the time it takes for equilibration of the FiO₂ and the alveolus). For example, a patient with normal lungs might equilibrate within three to five minutes while a patient with severe COPD may not equilibrate for 15 minutes.

●Low-flow oxygen device preference – We prefer low-flow devices, typically nasal cannulae (figure 4), for initial titration since they are more convenient, better tolerated, and readily available. However, it is difficult to accurately predict the FiO₂ when using nasal cannulae or simple face masks since it depends on the patient's total ventilation (higher ventilation dilutes the oxygen being supplied via nasal cannula/masks) and the patient's breathing pattern (many patients in respiratory distress primarily breathe through their mouth rather than nose).

In contrast, Venturi masks permit tighter regulation of the maximum FiO₂ administered to the patient (figure 5), but are generally employed when tighter FiO₂ control is required (eg, difficulty regulating the PaCO₂ with nasal cannulae, altered mental status due to hypercapnia).

Patients with advanced respiratory support or oxygen needs — For patients with more advanced needs (eg, severe hypoxemia, such as an SpO₂ of 75 percent, respiratory distress, significant acute respiratory acidosis [pH <7.3]) or for patients in whom the above strategy fails, supplemental oxygen should be administered at an FiO₂ that achieves adequate oxygenation (ie, SpO₂ 90 to 93 percent) and ventilatory support with NIV initiated. We use the same principles of oxygen titration since the risk of oxygen-induced hypercapnia still exists in this population. (See 'Patients with minimal oxygen and ventilatory needs' above.)

In most cases, we typically choose NIV since it can oxygenate and ventilate the patient simultaneously. HFNC may be an option if NIV is not tolerated, hypoxemia is severe, or intubation is not yet indicated or is outside the patient's goals of care. Management should be individualized and may need a pulmonologist or intensivist.

●NIV – For patients who fail to improve on low-flow oxygen or need advanced respiratory support (eg, patients with severe hypoxemia, such as an SpO₂ of 75 percent, respiratory distress, significant acute respiratory acidosis [pH <7.3]), we generally start NIV, typically bilevel positive airway pressure, provided there is no contraindication (table 4). Patients most likely to benefit are those with acute hypercapnia due to an acute exacerbation of COPD, acute cardiogenic pulmonary edema, obesity hypoventilation, and hypoventilation associated with central or peripheral neuromuscular or thoracic cage disorders. The purpose of NIV is the reduction of PaCO₂ (via improved ventilation) and avoidance of invasive mechanical ventilation. Oxygenation also improves for most patients if the treatment is effective in reducing the PaCO₂. The indications for and contraindications to NIV and data supporting its use in these populations are provided separately. (See "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications" and "Noninvasive positive airway pressure therapy for the obesity hypoventilation syndrome", section on 'Patients with acute hypercapnic respiratory failure and OHS' and "Respiratory muscle weakness due to neuromuscular disorder: Management", section on 'Noninvasive ventilation'.)

Initial settings, practical application, and follow-up of patients on NIV are described in the table (table 5) and discussed separately. (See "Noninvasive ventilation in adults with acute respiratory failure: Practical aspects of initiation".)

●HFNC – HFNC (figure 4) has a proven role in the treatment of hypoxemic respiratory failure (see "Evaluation and management of the nonventilated, hospitalized adult patient with acute hypoxemia", section on 'Humidified, high-flow oxygen delivered via nasal cannulae (HFNC)'). Its role in the treatment of hypercapnic respiratory failure is increasing, but optimal patient selection is unclear. HFNC may be an option if NIV is not tolerated, hypoxemia is severe, but intubation is not yet indicated or prohibited (eg, do-not-intubate order).

Data are limited [8-10]. A meta-analysis of seven randomized trials in patients with acute hypercapnic respiratory failure reported that HFNC had similar improvements in gas exchange and intubation rates as NIV [8]. In another meta-analysis of eight studies in patients with hypercapnic COPD (five in patients with acute hypercapnia and three in patients with chronic hypercapnia), HFNC reduced PaCO₂ in those with acute hypercapnia compared with conventional low-flow nasal oxygen [9].

Patients who fail these options, develop severe respiratory distress or arrest, or have a marked depression in their level of consciousness are typically intubated. (See "Direct laryngoscopy and endotracheal intubation in adults" and "Overview of initiating invasive mechanical ventilation in adults in the intensive care unit".)

Extracorporeal CO₂ removal is a newer technique that is rarely used [11]. The indications, applications, and complications of this technique are discussed separately. (See "Extracorporeal life support in adults: Extracorporeal carbon dioxide removal (ECCO2R)".)

Oxygen and carbon dioxide targets

●Oxygen target – Although the absolute lower limit of acceptable oxygenation in this population is unknown, we generally target an SpO₂ of 90 to 93 percent or a PaO₂ of 60 to 70 mmHg (8 to 9.3 kPa).

Rarely, in patients with severe chronic hypoxemic hypercapnic respiratory failure, levels lower than that (eg, SpO₂ ≥88 to 90 percent or PaO₂ >55 to 60 mmHg) may be maintained especially if it is close to their baseline and the patient is clinically stable (eg, normal mentation). This strategy provides adequate oxygenation by keeping the SpO₂ at the intersection between the ascending limb and the plateau of the hemoglobin-oxygen saturation curve (figure 6).

●CO₂ targets – We aim to return the PaCO₂ to premorbid baseline levels by treating the underlying cause with or without ventilation support.

The baseline for many patients is normocapnia. However, in patients who are chronically hypercapnic prior to admission, a PaCO₂ that approximates their premorbid state should be targeted. Overcorrection of chronic hypercapnia can result in alkalemia, which in turn can reduce respiratory drive and, if severe, induce seizures. Obtaining a previous ABG in the stable state is useful (eg, outpatient ABG or predischarge ABG in the past year or two).

Criteria for intensive care unit admission — There are no agreed-upon criteria for intensive care unit (ICU) admission other than the need for NIV or mechanical ventilation.

Common reasons for admission may include hypercapnic respiratory failure due to narcotic overdose, COPD exacerbation, or obesity hypoventilation syndrome [12]; the need for frequent monitoring (especially during fine titration of supplemental oxygen); and the prolonged application of NIV. Other examples include patients with severely impaired consciousness, patients with respiratory muscle fatigue (eg, neuromuscular disorders), patients not improving on NIV or HFNC, patients with severe acidosis (eg, <7.25), and patients with indications for admission other than hypercapnic respiratory failure (eg, hemodynamic instability). In patients with neuromuscular disease, vital capacity <1 L is also an indication for monitoring in an ICU setting.

In contrast, other patients with an etiology for acute hypercapnia that is easily reversible may require minor additional supervision and may not even require hospital admission once it has been established that it will not recur (eg, benzodiazepine-induced hypercapnia following conscious sedation for a minor procedure).

In complex cases, consultation with intensivists, pulmonologists, neurologists, and other subspecialists is reasonable for management.

THERAPIES OF UNPROVEN BENEFIT — 

We do not use respiratory stimulants (eg, medroxyprogesterone, acetazolamide, theophylline) since they have no proven benefit in the treatment of acute (or chronic) hypercapnic states.

Similarly, sodium bicarbonate infusions should only be considered in patients with a true indication for its use (eg, severe symptomatic metabolic acidosis, tricyclic antidepressant overdose) because bicarbonate is not an effective buffer for acute respiratory acidosis.

SUMMARY AND RECOMMENDATIONS

●Suspecting hypercapnia – Acute hypercapnic respiratory failure should be suspected in patients with risk factors (eg, sedative use, obstructive sleep apnea [OSA], chronic obstructive pulmonary disease [COPD] exacerbation) who have dyspnea and/or altered sensorium (eg, hypersomnolence). Hypercapnia should also be suspected in patients with a markedly elevated bicarbonate. The identification of an elevated arterial carbon dioxide tension (PaCO₂) >45 mmHg is diagnostic of hypercapnia. (See 'Suspecting and diagnosing hypercapnia' above.)

●Initial emergency management – Once diagnosed, we assess patients for upper airway obstruction and, if indicated, administer oxygen or ventilatory support. We also ensure that intravenous access is secured. (See 'Initial emergency management' above.)

●Narrowing the differential – Our approach is the following:

•Obtain a brief clinical assessment at the bedside – We obtain a brief history and perform an examination targeted at identifying the underlying cause (table 1 and table 3). We obtain a complete blood count with differential, serum chemistries including bicarbonate and electrolytes (including calcium, magnesium, and phosphate), and a chest radiograph. Additional laboratories may be obtained when specific etiologies are suspected (eg, toxicology screen, thyroid function tests). (See 'Brief clinical assessment' above.)

•Additional blood gas analysis – Hypercapnia should then be classified as acute (accompanying appropriate respiratory acidosis), chronic (a low-normal or near-normal pH), or acute-on-chronic (figure 1 and figure 2 and figure 3). When feasible, we calculate the alveolar-arterial (A-a) gradient from a room air blood gas (calculator 1). A normal A-a gradient is indicative of total hypoventilation whereas a widened gradient suggests that dead space ventilation from underlying lung disease may be contributing to the measured hypercapnia. (See 'Additional blood gas analysis' above.)

•Formulate and treat presumed underlying cause – Our approach is the following (see 'Formulating a preliminary diagnosis' above and 'Treat presumed cause of hypercapnia' above):

-Patients with symptomatic hypercapnia with a low respiratory rate and/or shallow tidal breathing ("can't breathe" [central nervous system issues] or "won't breathe" [peripheral neuromuscular issues]) may have a dominant component of hypoventilation. If there is no intrinsic lung disease, the A-a gradient and chest radiograph are typically normal. In such cases, ventilatory assistance is often needed and causes of hypoventilation should be investigated if feasible. (See 'Patients with symptomatic hypercapnia and low respiratory rate and/or shallow breathing (hypoventilation)' above.)

-Patients with symptomatic hypercapnia and excessive respiratory effort (tachypnea, respiratory distress) often have a dominant component of increased dead space ventilation due to underlying parenchymal lung disease or thoracic cage disease (ie, ventilation/perfusion mismatch; "can't breathe enough"). Such patients need reversal of the underlying etiology (eg, acute exacerbation of COPD) with or without ventilatory support. (See 'Patients with symptomatic hypercapnia and excessive respiratory effort (increased dead space)' above.)

-Some patients have mixed findings (eg, pneumonia and mucus plugging in a patient with neuromuscular disease). These patients may require both ventilatory assistance and reversal of the underlying disorder. (See 'Patients with mixed findings' above.)

-Patients with asymptomatic hypercapnia likely have chronic hypercapnia or hypercapnia due to significant alkalemia. In such cases, ventilatory assistance is unlikely to be indicated or beneficial. (See 'Asymptomatic hypercapnia' above.)

•Additional testing for underlying etiology – With this approach, the etiology can usually be determined or narrowed to a few possibilities (table 1). When the etiology is unclear, subsequent testing can be appropriately tailored with a specific etiology in mind. This may involve chest imaging (eg, parenchymal or thoracic cage pathology), brain or spinal cord imaging (central reasons for hypoventilation), pulmonary function testing (eg, respiratory neuromuscular weakness), or sleep study (OSA). (See 'Additional testing for underlying etiology' above.)

●Managing oxygen and ventilatory needs – For most patients with acute hypoxemic hypercapnic respiratory failure, we initiate or increase existing oxygen therapy with a fraction of inspired oxygen (FiO₂) to reach a set target. A reasonable target is a peripheral oxygen saturation (SpO₂) of 90 to 93 percent or an arterial oxygen tension of 60 to 70 mmHg (8 to 9.3 kPa). We aim to return the PaCO₂ back to premorbid baseline levels. (See 'Managing oxygen and ventilatory needs' above and 'Oxygen and carbon dioxide targets' above.)

•Patients with minimal oxygen and ventilatory needs – For patients who are estimated to require small amounts of supplemental FiO₂ (eg, <6 L/minute), we use low-flow oxygen devices to achieve target oxygen saturations. Ventilatory assistance is not typically needed. Oxygen options include 1 to 2 L/minute via nasal cannulae (figure 4) or 24 to 28 percent FiO₂ via Venturi mask (figure 5); the latter achieves greater FiO₂ control. We gradually increase the delivered FiO₂ (eg, every 5 to 15 minutes) in increments of 1 L/minute (via nasal cannula) or 4 to 7 percent (via Venturi mask). This approach avoids the risk of oxygen-induced hypercapnia. Once the target goal is achieved, we obtain an arterial blood gas. (See 'Patients with minimal oxygen and ventilatory needs' above.)

•Patients with advanced respiratory needs – For patients with more advanced needs or those in whom the above strategy fails, options are noninvasive ventilation (NIV) or high-flow oxygen via nasal cannulae (HFNC). (See 'Patients with advanced respiratory support or oxygen needs' above.)

-Most patients with advanced respiratory needs (eg, SpO₂ of 75 percent, respiratory distress, significant acute respiratory acidosis [pH <7.3]) benefit from NIV if they do not have contraindications (table 4). This includes patients with acute hypercapnia due to an acute exacerbation of COPD, acute cardiogenic pulmonary edema, obesity hypoventilation, and hypoventilation associated with central or peripheral neuromuscular or thoracic cage disorders. Data supporting NIV use in these populations are provided separately. (See "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications" and "Noninvasive positive airway pressure therapy for the obesity hypoventilation syndrome", section on 'Patients with acute hypercapnic respiratory failure and OHS' and "Respiratory muscle weakness due to neuromuscular disorder: Management", section on 'Noninvasive ventilation'.)

-HFNC may be an option if NIV is not tolerated, the hypoxemia is severe, and intubation is not yet indicated or is outside the patient's goals of care.

-Patients who fail these options, develop severe respiratory distress or arrest, or have a marked depression in their level of consciousness are typically intubated.