Carbon Monoxide Toxicity

Updated: Jan 26, 2023
Author: Guy N Shochat, MD; Chief Editor: Gil Z Shlamovitz, MD, FACEP 


Practice Essentials

Carbon monoxide (CO) is a colorless, odorless gas produced by incomplete combustion of carbonaceous material. Clinical presentation in patients with CO poisoning ranges from headache and dizziness to coma and death. Hyperbaric oxygen therapy (see the image below) can significantly reduce the morbidity of CO poisoning, but a portion of survivors still suffer significant long-term neurologic and affective sequelae.[1]

Monoplace hyperbaric chamber. Courtesy JG Benitez, Monoplace hyperbaric chamber. Courtesy JG Benitez, MD, MPH.

CO is formed as a by-product of burning organic compounds. Many cases of CO exposure occur in private residences.[2]  CO toxicity is especially common during power outages due to storms, as a result of the improper use of gasoline-powered portable generators to provide electricity and indoor use of charcoal briquettes for cooking and heating.[3, 4]  Exhaust from generators and propulsion engines on houseboats has also been linked to CO poisoning.[5]

Most fatalities from CO toxicity result from fires, but stoves, portable heaters, and automobile exhaust cause approximately one third of deaths. These often are associated with malfunctioning or obstructed exhaust systems and suicide attempts. Cigarette smoke is a significant source of CO. Natural gas contains no CO, but improperly vented gas water heaters, kerosene space heaters, charcoal grills, hibachis, and Sterno stoves all emit CO. Other sources of CO exposure include the following[6, 7] :

  • Propane-fueled forklifts
  • Gas-powered concrete saws
  • Inhaling spray paint
  • Indoor tractor pulls
  • Swimming behind a motorboat

CO intoxication also occurs by inhalation of methylene chloride vapors, a volatile liquid found in degreasers, solvents, and paint removers. Dermal methylene chloride exposure may not result in significant systemic effects but can cause significant dermal burns. Rarely, methylene chloride is ingested, and can result in delayed CO toxicity. The liver metabolizes as much as one third of inhaled methylene chloride to CO. A significant percentage of methylene chloride is stored in the tissues, and continued release results in elevated CO levels for at least twice as long as with direct CO inhalation.

Children riding in the back of enclosed pickup trucks seem to be at particularly high risk. Industrial workers at pulp mills, steel foundries, and plants producing formaldehyde or coke are at risk for exposure, as are personnel at fire scenes and individuals working indoors with combustion engines or combustible gases.


CO toxicity causes impaired oxygen delivery and utilization at the cellular level. CO affects several different sites within the body but has its most profound impact on the organs (eg, brain, heart) with the highest oxygen requirement.

Cellular hypoxia from CO toxicity is caused by impedance of oxygen delivery. CO reversibly binds hemoglobin, resulting in relative functional anemia. Because it binds hemoglobin 230-270 times more avidly than oxygen, even small concentrations can result in significant levels of carboxyhemoglobin (HbCO).

An ambient CO level of 100 ppm produces an HbCO of 16% at equilibration, which is enough to produce clinical symptoms. Binding of CO to hemoglobin causes an increased binding of oxygen molecules at the three other oxygen-binding sites, resulting in a leftward shift in the oxyhemoglobin dissociation curve and decreasing the availability of oxygen to the already hypoxic tissues.

CO binds to cardiac myoglobin with an even greater affinity than to hemoglobin; the resulting myocardial depression and hypotension exacerbates the tissue hypoxia. Decrease in oxygen delivery is insufficient, however, to explain the extent of the CO toxicity. Clinical status often does not correlate well with HbCO level, leading some to postulate an additional impairment of cellular respiration.

CO can produce direct cellular changes involving immunological or inflammatory damage by a variety of mechanisms, including the following[3] :

  • Binding to intracellular proteins (myoglobin, cytochrome a,a 3)
  • Nitric oxide generation leading to peroxynitrite production
  • Lipid peroxidation by neutrophils
  • Mitochondrial oxidative stress
  • Apoptosis
  • Immune-mediated injury
  • Delayed inflammation

Studies have indicated that CO may cause brain lipid peroxidation and leukocyte-mediated inflammatory changes in the brain, a process that may be inhibited by hyperbaric oxygen therapy. Following severe intoxication, patients display central nervous system (CNS) pathology, including white matter demyelination. This leads to edema and focal areas of necrosis, typically of the bilateral globus pallidus. Interestingly, the pallidus lesions, as well as the other lesions, are watershed area tissues with relatively low oxygen demand, suggesting elements of hypoperfusion and hypoxia.[8]

Studies have demonstrated release of nitric oxide free radicals (implicated in the pathophysiology of atherosclerosis) from platelet and vascular endothelium, following exposure to CO concentrations of 100 ppm. One study suggests a direct toxicity of CO on myocardium that is separate from the effect of hypoxia.[9]

HbCO levels often do not reflect the clinical picture, yet symptoms typically begin with headaches at levels around 10%. Levels of 50-70% may result in seizure, coma, and fatality.

CO is eliminated through the lungs. The half-life of CO at room air temperature is 3-4 hours. One hundred percent oxygen reduces the half-life to 30-90 minutes; hyperbaric oxygen at 2.5 atm with 100% oxygen reduces it to 15-23 minutes.


Most unintentional CO fatalities occur in stationary vehicles from preventable causes such as malfunctioning exhaust systems, inadequately ventilated passenger compartments, operation in an enclosed space, and utilization of auxiliary fuel-burning heaters inside a car or camper.

Most unintentional automobile-related CO deaths in garages have occurred despite open garage doors or windows, demonstrating the inadequacy of passive ventilation in such situations.

Colorado state data from 1986-1991 revealed that leading sources of 1149 unintentional nonfatal CO poisonings were residential furnaces (40%), automobile exhaust (24%), and fires (12%); however, furnaces were responsible for onlly 10% of fatal poisonings[10]

In the setting of structure fires, CO presents greater risk than thermal injury or oxygen deprivation, both for firefighters and victims.[11]

In most developing countries, cooking or heating is often done with unvented cookstoves, wood, charcoal, animal dung, or agricultural waste,  which has been linked with elevated HbCO levels.

Boats and houseboats represent a significant and underappreciated source of exposure, with multiple case reports and studies.[5]



United States

Unintentional, non–fire-related CO poisoning is responsible for approximately 15,000 emergency department visits annually in the United States. In 2000-2009, the exposure site was reported as residence in 77.6% of cases and workplace in 12%.[12] The most common source of CO exposure in the home is furnaces (18.5%), followed by motor vehicles, stoves, gas lines, water heaters, and generators.[13] During 1999–2012, deaths from unintentional non–fire-related CO poisoning in the US totaled 6136, an average of 438 deaths per year.[14]

In 2021, the American Association of Poison Control Centers reported 12,728 single exposures to CO, 366 of which were intentional. Major outcomes occurred in 349 cases, and 124 deaths were reported.[15]


Quantifying the global incidence of CO poisoning is impossible because of the transient duration of symptoms in mild intoxication, the ubiquitous and occult nature of exposure, and the tendency of misdiagnosis. Nevertheless, the worldwide cumulative incidence of CO poisoning can be currently estimated at 137 cases per million population. The worldwide incidence has remained stable during the last 25 years, while the mortality rate has declined; annual mortality is currently estimated at 4.6 deaths per million population.[16]

In contrast to findings in the United States, one Australian study of suicidal poisonings indicated no decrease following significantly lowered CO emissions from 1970-1996 and revealed no difference between the HbCO levels of occupants in cars with and without catalytic converters.[17]


All ages, ethnic populations, and social groups are affected, yet particular groups may be at higher risk.

  • Earlier data stated that, for unintentional fatalities, race-specific death rates were 20% higher for Blacks. More recent data reveal that non-Hispanic Whites and non-Hispanic Blacks to have equally high death rates, significantly above that of Hispanics and those classified as Other.[18]

  • Conversely, intentional fatalities demonstrate that race-specific rates for Blacks and other minority racial groups are 87% lower than for Whites, revealing a cultural partiality to this form of suicide.

Two North American studies, from the 1990s and 2005, examined the incidence of CO toxicity from indoor heating devices used during severe winter storms. Both studies identified a strong association between CO toxicity and US immigrants who were non–English speaking.[19] However, a study of acute, severe CO poisoning from portable electric generators in the US from August 1, 2008 to July 31, 2011 found that 96% of patients spoke English.[4]


During 1999–2010, the average annual death rate from unintentional non–fire-related CO poisoning was more than three times higher for males than for females (0.22 versus 0.07 per 100,000 population, respectively). Males represented an overwhelming 74% of unintentional non–fire-related deaths.[18]


Age-specific fatality rates increase with age and are highest in those older than 65 years. However, nonfatal exposures are more common in older teens and young adults (aged 15-34 y) than in older adults and are most common in young children (aged 0-4 y).[18, 13]

Individuals with pulmonary and cardiovascular disease tolerate CO intoxication poorly; this is particularly evident in those with chronic obstructive pulmonary disease (COPD) who have the additional concern of ventilation-perfusion abnormalities and possible respiratory depressive response to 100% oxygen therapy.

Neonates and the in utero fetus are more vulnerable to CO toxicity because of the natural leftward shift of the dissociation curve of fetal hemoglobin, a lower baseline PaO2, and levels of HbCO at equilibration that are 10-15% higher than maternal levels.

Climate and weather

Age-adjusted fatality rates are higher in cold and mountainous Midwestern and Western states and peak in the winter months. In addition, hurricanes and other natural disasters that result in power outages can lead to a spike in CO poisonings, as those affected turn to alternative sources of fuel or electricity.[20] For example, multiple incidents of CO poisoning were reported in Southern states following the Katrina and Rita hurricanes of 2005, in Northeastern states following Hurricane Sandy in 2012, and in Florida following Hurricane Irma in 2017.[21, 22, 23] The Centers for Disease Control and Prevention has established a Web page providing clinical guidance for carbon monoxide (CO) poisoning after a disaster.


Considerations regarding prognosis include the following:

  • Variability of clinical severity, laboratory values, and outcome limits prognostic accuracy
  • Cardiac arrest, coma, metabolic acidosis, and high HbCO levels are associated with poor outcome
  • Abnormal CT or MRI findings are associated with persistent neurologic impairment
  • Neuropsychiatric testing may have prognostic efficacy in determining delayed sequelae

An analysis of 331 pediatric patients with CO poisoning seen at a single-site emergency department found risk factors associated with severe disease course were a low Glasgow Coma Scale score, high leukocyte count, and high troponin T levels at presentation.[24]

Survivors of intentional CO poisoning are at extreme risk for subsequent completion of suicide.[3]




Carbon monoxide (CO) toxicity is often misdiagnosed because of the vagueness and broad spectrum of complaints; symptoms often are attributed to a viral illness. Specifically inquiring about possible exposures when considering the diagnosis is important (see Overview/Practice Essentials and Overview/Etiology).

For nonfatal nonintentional non–fire-related exposures, the most common symptom was headache (37%) followed by dizziness (18%) and nausea (17%).[13] However, any of the following symptoms should alert suspicion in the winter months, especially when the patient has a history compatible with CO exposure and when more than one patient in a group or household presents with similar complaints:

  • Malaise, flulike symptoms, fatigue
  • Dyspnea on exertion
  • Chest pain, palpitations
  • Lethargy
  • Confusion
  • Depression
  • Impulsiveness
  • Distractibility
  • Hallucination, confabulation
  • Agitation
  • Nausea, vomiting, diarrhea
  • Abdominal pain
  • Headache, drowsiness
  • Dizziness, weakness, confusion
  • Visual disturbance, syncope, seizure
  • Fecal and urinary incontinence
  • Memory and gait disturbances
  • Bizarre neurologic symptoms, coma

Chronic exposure also produces the above symptoms; however, patients with chronic CO exposure may present with loss of dentition, gradual-onset neuropsychiatric symptoms, or, simply, recent impairment of cognitive ability.

Physical Examination

Physical examination is of limited value. Inhalation injury or burns should always alert the clinician to the possibility of CO exposure.

Vital signs may include the following:

  • Tachycardia
  • Hypertension or hypotension
  • Hyperthermia
  • Marked tachypnea (rare; severe intoxication often associated with mild or no tachypnea)

Although so-called cherry-red skin has traditionally been considered a sign of CO poisoning (ie, "When you're cherry red, you're dead"), it is in fact rare.[3]  Pallor is present more often.

Ophthalmologic findings include the following:

  • Flame-shaped retinal hemorrhages
  • Bright red retinal veins (a sensitive early sign)
  • Papilledema
  • Homonymous hemianopsia
  • Noncardiogenic pulmonary edema

Neurologic and/or neuropsychiatric findings may include the following;

  • Memory disturbance (most common), including retrograde and anterograde amnesia with amnestic confabulatory states
  • Emotional lability, impaired judgment, and decreased cognitive ability
  • Stupor, coma, gait disturbance, movement disorders, and rigidity
  • Brisk reflexes, apraxia, agnosia, tic disorders, hearing and vestibular dysfunction, blindness, psychosis

Long-term exposures or severe acute exposures frequently result in long-term neuropsychiatric sequelae. Additionally, some individuals develop delayed neuropsychiatric symptoms, often after severe intoxications associated with coma.

After recovery from the initial incident, patients present several days to weeks later with neuropsychiatric symptoms such as those just described. Two thirds of patients eventually recover completely.





Laboratory Studies

The clinical diagnosis of acute carbon monoxide (CO) poisoning should be confirmed by demonstrating an elevated level of carboxyhemoglobin (HbCO). Either arterial or venous blood can be used for testing.[3]

Analysis of HbCO requires direct spectrophotometric measurement in specific blood gas analyzers. Bedside pulse carbon monoxide (CO)-oximetry is now available but requires a special unit and is not a component of routine pulse oximetry. A 2012 study showed that noninvasive pulse CO-oximetry correlates with more rapid diagnosis and initiation of hyperbaric oxygen therapy than laboratory CO-oximetry. However, the impact on clinical outcome is still not proven.[25]  A 2017 clinical policy statement from the American College of Emergency Physicians (ACEP) recommends against using pulse CO-oximetry to diagnose CO toxicity in patients with suspected acute CO poisoning (level B recommendation).[26]  

Elevated CO levels of at least 3–4% in nonsmokers and at least 10% in smokers are significant.[3] However, low levels do not rule out exposure, especially if the patient already has received 100% oxygen or if significant time has elapsed since exposure. HbCO levels in cigarette smokers typically range from 3-5%, but may be as high as 10% in some heavy smokers.[3] Presence of fetal hemoglobin, as high as 30% at 3 months, may be read as an elevation of HbCO level to 7%.[27] Symptoms may not correlate well with HbCO levels.

Findings on arterial blood gas measurement include the following:

  • Partial pressure of oxygen (PaO2) levels should remain normal; oxygen saturation is accurate only if directly measured but not if calculated from PaO2, which is common in many blood gas analyzers.

  • As with pulse oximetry, estimate PCO2 levels by subtracting the HbCO level from the calculated saturation. PCO2 level may be normal or slightly decreased. Metabolic acidosis occurs secondary to lactic acidosis from ischemia.

The ACEP recommends obtaining an electrocardiogram and cardiac biomarker levels in emergency department patients with moderate to severe CO poisoning (level B recommendation).[26] Cardiac marker results include the following:

  • Elevated high-sensitive troponin I levels often indicate cardiomyopathy, including reversible global dysfunction and a Takotsubo-like pattern[28]

  • Myocardial ischemia is common in patients hospitalized for moderate-to-severe CO exposure and is a predictor of mortality.[29]

  • Patients with preexisting cardiovascular disease can experience increased exertional angina with HbCO levels of just 5-10%; at high HbCO levels, even young healthy patients develop myocardial depression

Other test results include the following:

  • Creatinine kinase, urine myoglobin - Nontraumatic rhabdomyolysis can result from severe CO toxicity and can lead to acute renal failure.

  • Complete blood count - Mild leukocytosis may be present; disseminated intravascular coagulation (DIC) and thrombotic thrombocytopenic purpura (TTP) require further hematologic studies.

  • Electrolytes and glucose level - Hypokalemia and hyperglycemia occur with severe intoxication.

  • Blood lactate level - Elevation is an indication of severity,[3, 30, 31]  and may correlate with neurologic outcomes.[32]  If the source of the CO was a house fire and the lactate level is 10 mmol/L or higher, the patient may have concomitant cyanide poisoning.[3]

  • Blood urea nitrogen (BUN) and creatinine levels - Acute kidney failure may result from myoglobinuria.

  • Liver function tests - Mild elevation in fulminant hepatic failure

  • Urinalysis - Positive for albumin and glucose in chronic intoxication

  • Methemoglobin level - Included in the differential diagnosis of cyanosis with low oxygen saturation but normal PaO2

  • Toxicology screen - For instances of suicide attempt

  • Ethanol level - A confounding factor of both intentional and unintentional poisonings

  • Cyanide level - If cyanide toxicity also is suspected (eg, industrial fire); cyanide exposure is suggested by an unexplained metabolic acidosis; rapid determinations rarely are available. Smoke inhalation is the most common cause of acute cyanide poisoning.

Imaging Studies

Obtain a chest radiograph in patients with significant intoxications, pulmonary symptoms, evidence of hypoxia, or if hyperbaric oxygen is to be used. Findings usually are normal. Changes such as the following imply a worse prognosis than normal findings:

  • Ground-glass appearance
  • Perihilar haze
  • Peribronchial cuffing
  • Intra-alveolar edema

Computed tomography

Obtain a CT scan of the head with severe intoxication or change in mental status that does not resolve rapidly. Assess cerebral edema and focal lesions; most are typically low-density lesions of the basal ganglia.[33]

Positive CT scan findings generally predict neurologic complications. In one study, 53% of patients hospitalized for acute CO intoxication had abnormal CT scan findings; all of these patients had neurologic sequelae. Of those patients with negative scan results, only 11% had neurologic sequelae.[33]

Serial CT scans may be necessary, especially with mental status deterioration. One report describes the evolution of acute hydrocephalus in a child poisoned with CO, documented by serial CT scans.[34]

Magnetic resonance imaging

MRI is more accurate than CT scans for detection of focal lesions and white matter demyelination and is often used for follow-up care.[35] The progression from conventional MRI to diffusion-weighted imaging (DWI) and then diffusion tensor imaging (DTI) has enabled increasingly sensive evaluation of damage from CO poisoning. DTI can visualize progressive pathologic changes in the early stage of CO toxicity, allowing prediction of chronic conditions.[8]

Other Tests

On electrocardiography, sinus tachycardia is the most common abnormality. Arrhythmias may be secondary to hypoxia, ischemia, or infarction. Even low HbCO levels can have a severe impact on patients with cardiovascular disease.

Neuropsychologic testing

Formal neuropsychologic testing of concentration, fine motor function, and problem solving consistently reveal subtle deficits in even mildly poisoned patients.

Abridged versions of these tests are available that can be performed in about 30 minutes by a well-trained examiner. These are more applicable to the emergency department (ED) setting. These tests have been developed as possible means to assess the risk of delayed neurologic sequelae, to assess the need for hyperbaric oxygen therapy, and to determine the success of hyperbaric therapy in preventing delayed sequelae. The tests are used in some institutions, but studies prospectively confirming the conclusions are lacking.

Research indicates a specific link to deficits in context-aided memory in CO poisoing.Use of such specific testing in the ED has been proposed, as a tool for measuring the severity of neurologic involvement.



Prehospital Care

Prehospital care for patients with carbon monoxide (CO) toxicity includes the following:

  • Promptly remove the patient from continued exposure and immediately institute oxygen therapy with a nonrebreather mask.

  • Perform intubation for the comatose patient or, if necessary for airway protection, and provide 100% oxygen therapy.

  • Institute cardiac monitoring. Pulse oximetry, although not useful in detecting carboxyhemoglobin (HbCO), is still important because a low saturation causes even greater apprehension in this setting.

  • Give notification to the emergency department for comatose or unstable patients because rapid or direct transfer to a hyperbaric center may be indicated.

  • If possible, obtain ambient CO measurements from fire department or utility company personnel, when present.

  • Early blood samples may provide much more accurate correlation between HbCO and clinical status; however, do not delay oxygen administration to acquire them.

  • Obtain an estimate of exposure time, if possible.

  • Avoid exertion to limit tissue oxygen demand.

Emergency Department Care

Considerations in emergency department (ED) care include the following:

  • Cardiac monitor: Sudden death has occurred in patients with severe atherosclerotic disease at HbCO levels of only 20%.

  • Pulse oximetry: HbCO absorbs light almost identically to that of oxyhemoglobin. Although a linear drop in oxyhemoglobin occurs as HbCO level rises, pulse oximetry will not reflect it. Pulse oximetry gap, the difference between the saturation as measured by pulse oximetry and one measured directly, is equal to the HbCO level.[36] However, pulse CO-oximetry units are available that can screen for CO toxicity at the bedside.

Oxygen therapy is usually provided via a non-rebreather mask. However, Roth et al describe effective use of noninvasive continuous positive airway pressure (CPAP) ventilation using a tight mask and an inspired fraction of oxygen (FiO2) of 100%. These authors provide case reports of simultaneous CO toxicity in a couple, in which HbCO levels fell from 21% at admission to 6% within 1 hour and 3% after 90 minutes in the patient treated with CPAP. In the spouse, who was treated with conventional oxygen therapy, reduction of HbCO from the admission level of 21% to 3% took 6 hours.[37]

Continue 100% oxygen therapy until the patient is asymptomatic and HbCO levels are below 10%. In patients with cardiovascular or pulmonary compromise, lower thresholds of 2% have been suggested. Calculate a gross estimate of the necessary duration of therapy using the initial level and half-life of 30-90 minutes at 100% oxygen.

In uncomplicated intoxications, venous HbCO levels and oxygen therapy are likely sufficient. Evaluate patients with significant cardiovascular disease and initial HbCO levels above 15% for myocardial ischemia and infarction.

Consider immediate transfer of patients with levels above 40% or cardiovascular or neurologic impairment to a hyperbaric facility, if feasible. Persistent impairment after 4 hours of normobaric oxygen therapy necessitates transfer to a hyperbaric center. Pregnant patients should be considered for hyperbaric treatment at lower HbCO levels (above 15%). Because fetal hemoglobin has a greater affinity for CO than hemoglobin in the mother's red blood cells, the fetus acts as a sink for the CO, so HbCO levels will be higher in the fetus than in the mother.[38]

Serial neurologic examinations, including funduscopy, CT scans, and, possibly, MRI, are important in detecting the development of cerebral edema. Cerebral edema requires intracranial pressure (ICP) and invasive blood pressure monitoring to further guide therapy. Head elevation, mannitol, and moderate hyperventilation to 28-30 mm Hg PCO2 are indicated in the initial absence of ICP monitoring. Glucocorticoids have not been proven efficacious, yet the negative aspects of their use in severe cases are limited.

Do not aggressively treat acidosis with a pH above 7.15 because it results in a rightward shift in the oxyhemoglobin dissociation curve, increasing tissue oxygen availability. Acidosis generally improves with oxygen therapy.

In patients who fail to improve clinically, consider other toxic inhalants or thermal inhalation injury as contributing factors. Be aware that the nitrites used in cyanide kits cause methemoglobinemia, shifting the dissociation curve leftward and further inhibiting oxygen delivery at the tissue level. Combined intoxications of cyanide and CO may be treated with sodium thiosulfate 12.5 g intravenously to prevent the leftward shift.

Admit patients to a monitored setting and evaluate acid-base status if HbCO levels are 30-40% or above 25% with associated symptoms.  Admitted patients generally require monitored settings, telemetry beds, or cardiac care unit/medical intensive care unit (CCU/MICU) beds for more severe cases.

Patients with cerebral edema may be most appropriately treated in a neurosurgical ICU setting; this may dictate transfer to another facility. Admission or consult by toxicology service is helpful in these cases.

Hyperbaric Oxygen Therapy

Locate the nearest hyperbaric oxygen center by contacting the Divers Alert Network (DAN) at Duke University at (919) 684-2948. However, note that a survey of hyperbaric programs in the United States found that only 43 of 361 centers (11.9%) had equipment, intravenous infusion pumps and ventilators, and staff necessary to treat high-acuity patients.[39]

Hyperbaric oxygen (HBO) therapy currently rests at the center of controversy surrounding management of CO poisoning. Increased elimination of HbCO clearly occurs. Certain studies proclaim major reductions in delayed neurologic sequelae, cerebral edema, pathologic central nervous system (CNS) changes, and reduced cytochrome oxidase impairment.

Despite these individual claims, systematic reviews have not revealed a clear reduction in neurologic sequelae with HBO.[40, 41] A 2017 clinical policy statement from the American College of Emergency Physicians (ACEP) concluded that it remains unclear whether HBO therapy is superior to normobaric oxygen therapy for improving long-term neurocognitive outcomes.[26]  

A study by Han et al, in which 224 patients with acute CO poisoning were followed for up to 6 months, found no difference in the incidence of delayed neuropsychiatric sequelae between patients receiving HBO (n=198) and those receiving normobaric oxygen (n=26).[42]  In contrast, a nationwide observational study from Japan reported that patients who received HBO therapy had significantly lower rates of depressed mental status and reduced activities of daily living at discharge compared with a control group. Similar benefits were seen in patients with non-severe CO poisoning. The study used one-to-one propensity score matching to pair 2034 patients who received HBO therapy within 1 day of admission with an equal number of patients who did not receive HBO.[43]

However, evidence of a mortality benefit with HBO therapy has emerged. A retrospective study by Rose et al that reviewed 1099 cases of CO poisoning in adults concluded that HBO therapy was associated with an absolute risk reduction of 2.1% in both inpatient and 1-year mortality.[44]  

Lower mortality with HBO therapy was also reported in a retrospective nationwide population-based cohort study from Taiwan that included 7,278 patients who received HBO and 18,459 patients who did not. Overall, the adjusted hazard ratio [AHR] for death in HBO-treated patients was 0.74 (95% CI, 0.67-0.81). In patients younger than 20 years, the AHR was 0.45 (95% CI, 0.26-0.80) and for those with acute respiratory failure, the AHR was 0.43 (95% CI, 0.35-0.53). The lower mortality rate was noted for a period of 4 years.[45]  

Presently, universal treatment criteria do not exist; however, a survey of directors of North American HBO facilities with 85% responding demonstrates some consensus. The most common selection criteria (regardless of HbCO level) include the following:

  • Coma (98%)
  • Transient loss of consciousness (77%)
  • Ischemic ECG changes (91%)
  • Focal neurologic deficits (94%)
  • Abnormal neuropsychiatric test results (91%).

Ninety-two percent of HBO facility directors use HBO for headache, nausea, and HbCO levels above 40%; yet only 62% have a specific minimum HbCO level in asymptomatic patients. One half of the centers place a time limit on delay of treatment in patients with transient loss of consciousness alone.

Untreated pneumothorax is the only major contraindication for HBO therapy.[38]

HBO at 3 atm raises the amount of oxygen dissolved in the serum to 6.8%, enough to sustain cerebral metabolism. Elimination half-life is reduced to 15-23 minutes. Elimination half-life of CO from methylene chloride intoxication of 13 hours at room air temperature is reduced to 5.8 hours.

Chambers are small monoplace hulls, allowing space for a single patient in a supine position who can be viewed through a window at the head, or they are acrylic walled and allow full visualization. Many of these monoplace chambers allow for care of critically ill patients, including intravenous lines, arterial lines, and ventilator. Others are large multiplace chambers that permit ventilation equipment and allow medical teams to accompany the patient. A monoplace chamber is shown below.

Monoplace hyperbaric chamber. Courtesy JG Benitez, Monoplace hyperbaric chamber. Courtesy JG Benitez, MD, MPH.

Treatment regimens usually involve 100% oxygen at 2.4-3 atm for 90-120 minutes. Re-treatment, although controversial, may be performed for acutely and chronically persistent symptoms. One study suggests that degree of metabolic acidosis can predict the need for re-treatment.[46]

Complications of therapy include decompression sickness, sinus and middle ear barotrauma, seizure, progression of pneumothorax to tension pneumothorax, gas embolism, reversible visual refractive changes, and complications related to transport of unstable patients.

For treatment of complications from therapy, decongestants are useful, prophylactic myringotomy is common and a requirement for intubated patients, and chest tube placement is mandatory with pneumothorax. Exercise caution in patients who have experienced chest compressions, central venous catheterization, intubation, and positive pressure ventilation. Seizures are most often secondary to oxygen toxicity and do not mandate anticonvulsant therapy or discontinuation of HBO therapy.

In multiplace chambers, seizure therapy consists of removing the oxygen mask. In monoplace chambers, decompression lowers oxygen concentration. It is crucial not to do this during the tonic phase of the seizure because it may cause pulmonary barotrauma secondary to gas expansion in the lungs.

A 10-year retrospective study found that transfer to an HBO facility did not need to be delayed for concern of cardiac arrest, respiratory arrest, myocardial infarction, or worsening mental status if they had not occurred during initial resuscitation; however, hypotension, dysrhythmia, seizure, emesis, and agitation were of concern in transit as well as in initial resuscitation.[47]


Survivors of CO poisoning are at risk for a range of neurologic and psychiatric complications, including the following[48] :

  • Impaired intellectual function
  • Short-term memory loss
  • Dementia
  • Amnesia
  • Psychosis
  • Irritability
  • Dysfunctional gait
  • Speech disorders
  • Parkinson disease [49]
  • Cortical blindness
  • Depression

Discuss the possibility of delayed neurologic complications, although they are much more common in patients with toxicity severe enough to require hospital admission.


Carbon monoxide (CO) detectors: Home CO detectors with audible alarms are available, and can limit CO toxicity.[50]  One study of 911 calls for suspected CO poisoning showed in 80% of calls for detector alarms, verifiable ambient CO levels were present in the home; the mean concentration of CO was 18.6 ppm in homes tested because of detector alarms but was 96.6 ppm in homes without alarms where calls were prompted by suspicious symptoms.[51]


For patient education information, see Carbon Monoxide Poisoning.

Long-Term Monitoring

Asymptomatic patients with HbCO levels below 10% may be discharged. In cases of accidental CO poisoning, patients should be followed up in 4-6 weeks to screen for cognitive sequelae. With intentional poisoning, psychiatric follow-up is mandatory, given the high rate of subsequent completed suicide.[3]

A nationwide population-based study from Korea found that the risk of venous thromboembolism was significantly elevated in the first 90 days after CO poisoning (odds ratio [OR] 3.96; 95% confidence interval [CI] 2.50 to 6.25). Risk was especially high in the first 30 days for pulmonary embolism (OR 22.00; 95% CI 5.33 to 90.75) and deep venous thrombosis (OR 10.33; 95% CI 3.16 to 33.80). These researchers recommend monitoring patients for venous thromboembolism risk in the 3 months following CO poisoning.[52]



Questions & Answers


What is carbon monoxide (CO) toxicity?

What are the most common causes of carbon monoxide (CO) toxicity?

What is the role of methylene chloride vapors in carbon monoxide (CO) toxicity?

Who are at high risk for carbon monoxide (CO) toxicity?

What is the pathophysiology of carbon monoxide (CO) toxicity?

What are the physiologic mechanisms of carbon monoxide (CO) toxicity?

How does carbon monoxide (CO) cause harm to the brain?

How is carbon monoxide (CO) eliminated from the body?

How does the mortality rate from carbon monoxide (CO) toxicity differ between males and females?

What is the incidence of carbon monoxide (CO) toxicity in the US?

What is the global incidence of carbon monoxide (CO) toxicity?

What are the racial predilections of carbon monoxide (CO) toxicity?

How do fatality rates from carbon monoxide (CO) toxicity vary by age?

Which risk factors for carbon monoxide (CO) toxicity are related to weather?


Why is misdiagnosis of carbon monoxide (CO) toxicity common?

What are symptoms of carbon monoxide (CO) toxicity?

Which physical findings suggest carbon monoxide (CO) toxicity?

Which ophthalmologic findings suggest carbon monoxide (CO) toxicity?

Which neurologic findings suggest carbon monoxide (CO) toxicity?

What are the causes of carbon monoxide (CO) toxicity?


What are the differential diagnoses for Carbon Monoxide Toxicity?


How is the clinical diagnosis of acute carbon monoxide (CO) toxicity confirmed?

Which findings on arterial blood gas measurements suggest carbon monoxide (CO) toxicity?

What are cardiac markers for carbon monoxide (CO) toxicity?

What is the role of lab studies in the evaluation of carbon monoxide (CO) toxicity?

What is the role of chest radiograph in the evaluation of carbon monoxide (CO) toxicity?

What is the role of CT scanning in the evaluation of carbon monoxide (CO) toxicity?

What is the role of MRI in the evaluation of carbon monoxide (CO) toxicity?

What is the role of electrocardiography in the evaluation of carbon monoxide (CO) toxicity?

What is the role of neuropsychologic testing in the evaluation of carbon monoxide (CO) toxicity?


What is included in prehospital care for carbon monoxide (CO) toxicity?

What is included in emergency department (ED) care for carbon monoxide (CO) toxicity?

How can clinicians locate a hyperbaric oxygen center for treatment of a patient with carbon monoxide (CO) toxicity?

What is the efficacy of hyperbaric oxygen therapy for treating carbon monoxide (CO) toxicity?

When is hyperbaric oxygen therapy indicated for the treatment of carbon monoxide (CO) toxicity?

How is hyperbaric oxygen therapy administered for carbon monoxide (CO) toxicity?

What are possible complications of hyperbaric oxygen therapy for carbon monoxide (CO) toxicity?

What are possible complications of carbon monoxide (CO) toxicity?