Carbon monoxide poisoning

INTRODUCTION — Carbon monoxide (CO) is an odorless, tasteless, colorless, nonirritating gas formed by hydrocarbon combustion. The atmospheric concentration of CO is generally below 0.001 percent, but it may be higher in urban areas or enclosed environments. CO binds to hemoglobin with much greater affinity than oxygen, forming carboxyhemoglobin (COHb) and resulting in impaired oxygen transport and utilization. CO can also precipitate an inflammatory cascade that results in central nervous system (CNS) lipid peroxidation and delayed neurologic sequelae.

CO poisoning will be reviewed here. A summary table to facilitate emergency management is provided (table 1). Related topics include:

●(See "Inhalation injury from heat, smoke, or chemical irritants".)

●(See "Hyperbaric oxygen therapy".)

●(See "Cyanide poisoning".)

●(See "Overview of occupational and environmental health".)

EPIDEMIOLOGY — CO poisoning is estimated to occur in 50,000 people annually in the United States, and non-fire-related smoke inhalation is responsible for most cases [1]. Analysis of aggregated national data from the United States supports an overall mortality of 1 to 3 percent, with a mortality rate that is higher for intentional poisoning than for inadvertent exposure [2-5]. There are approximately 1000 to 1300 deaths from CO poisoning annually [1,6,7]. Intentional CO poisoning accounts for two-thirds of deaths and inadvertent, non-fire related CO poisoning cause the rest [6,8].

In contrast with intentional poisoning, unintended poisoning demonstrates both seasonal and regional variation, and occurs more frequently during the winter months in cold climates, most commonly from faulty furnaces [9]. An increase in CO poisoning has been reported to occur in the immediate aftermath of electrical power outages leading to increased use of portable, gasoline-powered generators. Generator misplacement indoors or proximate to home ventilation intake systems causes CO exposure and occasional fatalities [10-12].

SOURCES OF EXPOSURE — CO is produced from incomplete combustion of fuels. Potential sources of CO include:

●Smoke inhalation from fires.

●Poorly functioning heating systems.

●Improperly vented fuel-burning devices (eg, kerosene heaters, charcoal grills, camping stoves [13,14], gasoline-powered electrical generators [10,15]).

●Motor vehicles operating in poorly ventilated areas (eg, ice rinks, warehouses, parking garages, keyless cars left running, tailpipe blocked by snow drift).

●Open air exposure to motorboat exhaust [16].

●Underground electrical cable fires that produce large amounts of CO, which can seep into adjacent buildings and homes [17].

●Hookah (waterpipe) use [18,19]. (See "Patterns of tobacco use", section on 'Waterpipes or hookahs'.)

●Inhaled or ingested methylene chloride (dichloromethane), an industrial solvent and a component of paint remover, is hepatically metabolized to CO and can cause CO toxicity in the absence of ambient CO [20,21]. In 2012, the European Union banned paint strippers containing methylene chloride, and in 2019, it was banned from all paint and coating removers for consumer use in the United States. However, methylene chloride may still be available in some products for industrial applications. (See "Overview of occupational and environmental health".)

PATHOPHYSIOLOGY — CO poisoning causes impaired oxygen delivery and utilization as well as generation of reactive oxygen species. CO diffuses rapidly across the pulmonary capillary membrane. Elimination is dependent upon the degree of oxygenation and, to a lesser extent, minute ventilation.

●Impaired oxygen delivery – CO binds to the iron moiety of heme (and other porphyrins) with approximately 240 times the affinity of oxygen forming carboxyhemoglobin (COHb). This induces an allosteric change that greatly diminishes the ability of the other three oxygen binding sites in hemoglobin to off-load oxygen to peripheral tissues (figure 1). This results in a deformation and leftward shift of the oxyhemoglobin dissociation curve, and compounds the impairment in tissue oxygen delivery (figure 2). The degree of carboxyhemoglobinemia is a function of the relative amounts of CO and oxygen in the environment, duration of exposure, and minute ventilation. (See "Oxygen delivery and consumption".)

●Impaired oxygen utilization – Approximately 10 to 15 percent of CO is extravascular and bound to molecules such as myoglobin, cytochromes, and NADPH reductase, resulting in impairment of oxidative phosphorylation at the mitochondrial level (figure 3) [3,22]. The half-life of CO bound to these molecules is longer than that of COHb. The importance of these nonhemoglobin-mediated effects has been best documented in the heart, where mitochondrial dysfunction due to CO can produce myocardial stunning despite adequate oxygen delivery [23].

CO also inactivates cytochrome oxidase in a manner similar to cyanide. CO and cyanide poisoning can occur simultaneously in patients following smoke inhalation, and their combined effects on oxygen transport and utilization appear to be synergistic [24,25]. (See "Inhalation injury from heat, smoke, or chemical irritants".)

●Reactive oxygen species – CO results in superoxide generation and oxidative stress, which likely contributes to lipid peroxidation and neurologic injury. (See 'Delayed neuropsychiatric syndrome' below.)

CLINICAL FINDINGS

Common symptoms — The clinical findings of CO poisoning are highly variable and largely nonspecific [26,27]. Mildly or moderately CO-intoxicated patients often present with constitutional symptoms and may be misdiagnosed with acute viral syndromes [28]. In addition to current symptoms, the clinician should specifically inquire (of the patient and/or witnesses) about transient loss of consciousness, as the presence or absence of this finding is of critical importance in determining the need for hyperbaric oxygen (HBO).

In exposure calls to United States poison centers, patients who developed symptoms reported the following [1]:

●Headache: 58 percent

●Nausea: 33 percent

●Dizziness: 29 percent

●Drowsiness: 14 percent

●Vomiting: 14 percent

●Cough/choking: 6 percent

●Confusion: 5 percent

●Shortness of breath: 5 percent

●Syncope: 5 percent

●Throat or eye irritation, chest pain, weakness: <5 percent

Examination findings — In the absence of concurrent trauma or burns, physical findings in CO poisoning are usually confined to alterations in mental status, tachycardia, and tachypnea. Patients may manifest symptoms ranging from mild confusion to coma. There is no role for specific cognitive testing, such as the Carbon Monoxide Neuropsychological Screening Battery, in the acute setting [20,29].

The classically described finding of a "cherry red" appearance of the lips is neither a sensitive nor specific sign [26].

Severe toxicity — CO poisoning is considered severe in patients with any of the following findings:

●Neurologic – Seizures, syncope, transient loss of consciousness, or coma.

●Metabolic – Lactic acidosis (which may be profound) from cellular hypoxia and cytochrome oxidase inhibition.

●Cardiovascular – Acute myocardial ischemia, myocardial injury, ventricular arrhythmias, and pulmonary edema.

Myocardial ischemia is common among moderately to severe CO-poisoned patients and is associated with increased long-term mortality. A retrospective study of 230 patients with moderate or severe CO poisoning referred to a specialized center found evidence of myocardial ischemia (characteristic electrocardiographic changes or elevated serum cardiac biomarkers) in one-third of all cases [30]. (See 'Long-term outcomes' below.)

Delayed neuropsychiatric syndrome — The syndrome of delayed neurologic sequelae (DNS) includes variable degrees of cognitive deficits, personality changes, movement disorders, and focal neurologic deficits. DNS occurs in 15 to 40 percent of patients with significant CO exposure [6]. DNS is reported to arise 3 to 240 days after apparent recovery, generally occurring within 20 days of CO poisoning. Deficits may persist for a year or longer [31-34].

Although DNS is associated with abnormalities in the globus pallidus and deep white matter on computed tomography (CT), magnetic resonance imaging (MRI), and positron-emission tomography (PET), these findings are typically not found at acute presentation [32,35-39]. Hemorrhagic infarction of the globus pallidus and, less frequently, the deep white matter have been reported rarely following acute intoxication [40].

The development of DNS correlates poorly with carboxyhemoglobin (COHb) levels, although the majority of cases are associated with loss of consciousness during acute intoxication [3,20,41]. The mechanism of DNS is incompletely understood, but it probably involves lipid peroxidation by reactive oxygen species generated by xanthine oxidase. Xanthine oxidase is produced in situ from xanthine dehydrogenase via enzymes released by white blood cells that adhere to damaged endothelial cells [29,42-45]. During recovery from CO exposure, events analogous to ischemia-reperfusion injury and exposure to hyperoxia may exacerbate the initial oxidative damage [28,46].

DIAGNOSIS

The diagnosis of CO poisoning should be suspected in fire victims, in patients with flu-like symptoms in cold climates, especially if any other cohabitants or pets are also feeling ill, and patients with unexplained altered mental status or lactic acidosis. History obtained by emergency management personnel or ambient CO levels measured by the fire department may suggest CO poisoning. However, the ambient CO reading obtained by fire departments is often lower than the patient's exposure because emergency dispatchers (ie, 9-1-1 operators) often advise patients to open their windows.

The diagnosis of CO poisoning is made in patients with known or suspected CO exposure in conjunction with an elevated carboxyhemoglobin (COHb) level measured by cooximetry of a blood gas sample. In hemodynamically stable patients, venous samples are accurate and commonly used [47,48]. Nonsmokers may have up to 3 percent COHb at baseline; smokers may have levels of 10 percent [20]. Levels above these respective values are consistent with CO poisoning.

●COHb levels confirm exposure but may not tell extent – A COHb measurement is essential for determining exposure, but levels correlate imprecisely with the degree of poisoning. This is especially true if significant time has passed between exposure and when the level was obtained, owing to physiologic clearance of CO (see 'Pathophysiology' above). Any abnormally elevated COHb level associated with major symptoms (eg, loss of consciousness, altered mental status, cardiac ischemia) should be considered severe poisoning. Patient symptoms and signs guide management, not COHb level.

●Standard pulse oximetry is not useful – Standard pulse oximetry (SpO₂) cannot screen for CO exposure, as it does not differentiate COHb from oxyhemoglobin (figure 4) [49,50]. Eight-wavelength pulse oximeters measuring COHb and methemoglobin are available but are not considered accurate enough to substitute for blood co-oximetry, although they may have a role as a screening test [51-54]. (See "Pulse oximetry", section on 'Carboxyhemoglobin'.)

●Hydroxocobalamin interferes with COHb measurement – Treatment with hydroxocobalamin (which could be administered prehospital for presumed cyanide exposure) can interfere with the measurement of COHb leading to inaccurate results, and both falsely low and elevated levels are reported [55-59]. In patients with possible cyanide exposure (eg, brought to the hospital from a fire), clinicians assuming care should ask whether hydroxocobalamin was given. However, regardless of such treatment, clinicians should make a presumptive diagnosis of CO poisoning on the basis of the exposure history and should err on the side of treatment with oxygen and hyperbaric oxygen (HBO) if there is any doubt. If available, blood samples obtained before treatment with hydroxocobalamin would provide more accurate measurements.

POSTDIAGNOSTIC TESTING — The following tests are obtained selectively after diagnosis of CO poisoning:

●Electrocardiogram (ECG) – Obtain in all symptomatic patients.

●Cardiac biomarkers

•Patients with ECG evidence of ischemia.

•Patients with cardiac risk factors or a history of cardiac disease [30].

●Blood gas – Patients who are critically ill should undergo blood gas testing. Arterial sample is preferred if there are signs of severe hypoperfusion; otherwise, a venous sample is adequate.

Blood PO₂ measurements tend to be normal because PO₂ reflects O₂ dissolved in blood, and this process is not affected by CO. In contrast, hemoglobin-bound O₂ (which normally comprises 98 percent of arterial O₂ content) is profoundly reduced in the presence of carboxyhemoglobin (COHb). (See "Oxygen delivery and consumption".)

●Complete blood count (CBC), serum chemistries

●Urine or serum pregnancy test – In females of child-bearing age.

●Chest radiograph – Patients with dyspnea or hypoxia to evaluate for pulmonary edema.

MANAGEMENT

General principles

●The key interventions in the management of the CO-poisoned patient are prompt removal from the source of CO and provision of high-flow oxygen. (See 'High-flow oxygen' below.)

●In cases of unintentional CO poisoning, the clinician must consider that other people could have ongoing CO exposure if the source is not identified and corrected. Local fire departments or other emergency management personnel can assist with an assessment of ambient CO levels in the suspected environment and removal of victims [20].

●Small, critical access hospitals may lack the ability to measure carboxyhemoglobin (COHb) concentrations [60]. Patients suspected of CO poisoning who are treated in hospitals without on-site cooximetry should be started on 100 percent oxygen via nonrebreathing face mask. After initial stabilization, clinicians should discuss transfer with a medical toxicologist or a hyperbaric oxygen (HBO) specialist at a receiving hospital, especially for patients at high risk for adverse outcomes (eg, acidemia, ischemic electrocardiogram [ECG] changes, syncope, altered mental status, or persistent chest pain). The receiving facility should be capable of providing accurate measurement of COHb levels, ongoing intensive care, and HBO therapy.

High-flow oxygen — In a patient with suspected or confirmed CO poisoning, we recommend initial treatment with high-flow (100 percent) normobaric oxygen via nonrebreathing face mask, regardless of pulse oximetry or arterial PO₂. Elimination of CO starts once the patient is removed from the exposure and is almost exclusively via the pulmonary circulation through competitive binding of hemoglobin by oxygen. The half-life of COHb in a patient breathing room air is approximately 250 to 320 minutes; this decreases to 75 to 90 minutes with high-flow oxygen (>15 L/minute) provided via a nonrebreathing mask [61]. Normobaric high-flow oxygen therapy is relatively safe and available, and it hastens the elimination of COHb even though evidence does not exist that, compared with breathing room air, supplemental oxygen actually limits or prevents delayed neuropsychiatric syndrome (DNS). In a trial of CO-poisoned patients, six-week cognitive sequelae were lower in patients who were treated with normobaric oxygen compared with patients not receiving supplemental oxygen (41 [60/146] versus 53 [9/17] percent), but this finding did not achieve statistical significance [62].

Approach to patients with severe toxicity or critical illness

ABC's — Initial management starts with assessing and stabilizing airway, breathing, and circulation. Patients who are comatose, have severely impaired mental status, or who do not have sufficient respiratory effort should be intubated without delay and mechanically ventilated using 100 percent oxygen. A summary table to facilitate emergent management is provided (table 1).

Patients with smoke inhalation — In patients suffering from CO poisoning from a fire or smoke inhalation who are critically ill (ie, those with coma, seizure, or hemodynamic compromise associated with a metabolic acidosis and increased blood lactate concentration), we recommend empiric administration of hydroxocobalamin to treat possible cyanide toxicity. Since cyanide exposure can occur from fires and smoke inhalation and there is overlap in clinical presentation with CO, it is difficult to definitively exclude concomitant cyanide toxicity, which can further impair tissue oxygen utilization and exacerbate cellular hypoxia [22,63]. The hydroxocobalamin dose is 70 mg/kg intravenous (IV; 5 g is the standard adult dose) and can be repeated after 10 to 15 minutes if the patient does not have rapid clinical improvement. Emergency medicine services may have administered hydroxocobalamin en route to the hospital. (See "Cyanide poisoning", section on 'Empiric treatment for smoke inhalation' and "Inhalation injury from heat, smoke, or chemical irritants".)

Hyperbaric oxygen therapy — In patients with symptoms consistent with severe CO toxicity, we suggest treatment with HBO (algorithm 1). (See 'Severe toxicity' above.)

We use HBO in the following settings:

●COHb level >25 percent

●COHb level >15 percent in a pregnant patient (see 'Pregnant patients' below)

●Loss of consciousness

●Severe metabolic acidosis (pH <7.25)

●Evidence of end-organ ischemia (eg, ECG changes, elevated cardiac biomarkers, respiratory failure, focal neurologic deficit, or altered mental status)

HBO provides the greatest benefit if treatment begins as early as possible, ideally within six hours, to increase elimination of the carboxyhemoglobinemia and improve tissue oxygenation [64,65]. Importantly, even if COHb has decreased to below the recommended thresholds noted above, delayed HBO treatment may still provide benefit; however, it is unproven if more than 24 hours has passed after CO exposure. All patients selected to receive HBO should have at least one treatment at 2.5 to 3 atmospheres absolute (ATA) as soon as possible to reverse the acute effects of CO intoxication, with possible additional therapy directed toward limitation or prevention of DNS [20,24,30,66].

HBO exposes patients to 100 percent oxygen under supra-atmospheric conditions, which decreases the half-life of COHb to 30 minutes from approximately 75 to 90 minutes on 100 percent normobaric oxygen. The amount of oxygen dissolved in the blood also rises from approximately 0.3 to 6 mL per dL, which substantially increases the delivery of nonhemoglobin-bound oxygen to the tissues. HBO inhibits neutrophil oxidative burst, xanthine oxidase production, and lipid peroxidation, which is the purported mechanism of preventing DNS [67,68]. The mechanisms of action of HBO are discussed in more detail elsewhere. (See "Hyperbaric oxygen therapy", section on 'Mechanisms of action'.)

Although we routinely advocate for HBO therapy in patients with a COHb level greater than 25 percent, we recognize that other clinicians may disagree based on their interpretation of existing studies. Some societies use 40 percent as the appropriate threshold. There is no clear basis in the medical literature for choosing one level over the other. Experts have created a broad set of indications for HBO therapy in CO-poisoned patients, despite the uncertainty in identifying which patients will benefit from HBO treatment [20,27,46,69,70]. Consultation with a medical toxicologist, regional poison control center, or HBO specialist can be helpful with this decision. (See 'Additional resources' below.)

Efficacy — Evidence supporting the use of HBO is based on trials with significant limitations, observational studies, and mechanistic data.

•Mortality – HBO has been associated with decreased mortality after CO poisoning. In a retrospective national poison database study including over 25,000 patients with CO poisoning, patients who received HBO had lower mortality up to four years after treatment than those who did not (adjusted hazard ratio [aHR] 0.74, 95% CI 0.67-0.81) [70]. The benefit of HBO was most pronounced for patients with acute respiratory failure (aHR 0.45) and patients younger than 20 years of age (aHR 0.43). Among all patients undergoing HBO, those who received two or more HBO treatments had lower mortality than patients who received only one.

However, this study has significant limitations in data granularity; the severity of poisoning and timing of HBO were not available and were not factored in the analysis. Thus, there may have been a selection bias in favor of HBO therapy if mildly injured patients were more likely to receive HBO while severely injured patients were deemed futile or were too unstable for transfer. Despite these limitations, it provides indirect support for HBO therapy in patients with altered mental status or other severe manifestations of poisoning such as severe metabolic acidosis or end-organ ischemia.

•Delayed neuropsychiatric syndrome (DNS) – HBO may be beneficial in preventing the late neurocognitive deficits associated with severe CO poisoning. The quality and results of clinical trials designed to assess the efficacy of HBO in reducing the severity of DNS have varied widely [2,31,53,71-77].

-A metanalysis of six published trials (one published only in abstract form) included 1335 patients randomized to either HBO or normobaric oxygen therapy (NBO) and who had outcomes recorded. A pooled analysis using a random effects model showed a trend towards fewer neurologic sequalae in the HBO group (29 versus 34 percent, odds ratio [OR] 0.78, 95% CI 0.54-1.12) [77]. However, these six trials employed heterogeneous HBO protocols, inclusion criteria, and outcome definitions, thus limiting the ability to draw conclusions. Some studies excluded patients without syncope, while others excluded those with syncope. The number of HBO treatments and protocols varied, with several trials using lower than the recommended 2.5 ATA. The definition and measurement of neurologic sequalae and duration of follow-up varied between trials. Only two of the trials were double-blinded with "sham dives" in a chamber, which is critical in reducing bias since main outcomes are patient-reported symptoms [71,73].

-The trial with arguably the best methodology randomized 152 patients within 24 hours of presentation to hyperbaric or NBO [71]. Both hyperbaric and control groups received treatment in a hyperbaric chamber, effectively blinding the therapy that a patient was receiving. Six weeks after presentation, cognitive sequelae were less common in the group treated with HBO (25 versus 46 percent, OR 0.39, 95% CI 0.2-0.78). This advantage of hyperbaric therapy in terms of neurologic performance was maintained at six months and one year following initial presentation. A major limitation was the disparity in severity of toxicity and baseline characteristics between groups. Patients in the NBO group had a higher rate of cerebellar dysfunction before treatment (15 versus 4 percent), which was associated with a higher incidence of cognitive sequelae. Also, patients randomized to NBO had longer duration to CO than those in the HBO group (mean CO exposure 22 versus 13 hours).

-Another randomized, double-blinded trial of 191 patients referred to a tertiary center with CO poisoning during a two-year period failed to document benefit for patients who received HBO [73]. Rather, DNS and poor performance on neuropsychiatric tests after one month were more common among HBO-treated patients. However, one month follow-up was only 46 percent, making the results of this trial uninterpretable.

Challenges

•Limited availability of hyperbaric chambers is a major impediment to the wider application of HBO in the management of CO poisoning. In the United States, approximately 250 hyperbaric facilities offer either single occupant ("monoplace") or multiple occupant ("multiplace") chambers. Information regarding the location of hyperbaric facilities can be accessed through the Undersea and Hyperbaric Medical Society website or via the Divers Alert Network Emergency Hotline (1-919-684-9111) [20]. Outside of the United States, international poison control centers can be contacted to help locate the nearest HBO chamber. Contact information for poison centers around the world is provided separately. (See 'Additional resources' below.)

•Since most hospitals do not have an HBO chamber, the benefit of HBO must be weighed against the risks and logistical challenges of transferring a patient, especially a critically ill patient who may require close bedside attention.

•Victims of fires and smoke inhalation often have fluid requirements, hemodynamic instability, and airway issues necessitating frequent titration of therapies. Patients placed in an HBO chamber should be relatively stable since they cannot be removed quickly, as it takes time to depressurize. There should be no concern for dysrhythmias since electrical cardioversion or defibrillation cannot be performed in a chamber.

•In patients who are in extremis and are too unstable to be moved to an HBO chamber, extracorporeal membrane oxygenation (ECMO) may be employed for rescue therapy if it is available at the treating hospital. Successful outcomes with both venovenous and venoarterial ECMO for severe CO poisoning have been reported [78-81]. (See "Extracorporeal life support in adults in the intensive care unit: Overview".)

The complications of HBO are discussed in more detail elsewhere. (See "Hyperbaric oxygen therapy", section on 'Complications'.)

Approach to patients without severe/critical illness

Pitfalls — Some patients with CO exposure who appear clinically well should be considered to have severe toxicity if at any time during their intoxication they met criteria for HBO. (See 'Hyperbaric oxygen therapy' above.)

COHb measurement must be interpreted in the context of time since exposure ended because the COHb level is expected to decrease through pulmonary elimination over time. Patients who had transient loss of consciousness with a confirmed CO exposure could still benefit from HBO even if their COHb level does not reflect the extent of exposure when it occurred.

High-flow oxygen — In a patient with suspected or confirmed CO poisoning not requiring intubation or meeting the indications for HBO, we recommend treating with high-flow (100 percent) oxygen via nonrebreathing face mask. Oxygen is continued until the symptoms attributed to CO poisoning have resolved and the COHb level predicted based on the elimination half-life of 90 minutes with high-flow oxygen would be less than 10 percent. (See 'High-flow oxygen' above.)

Disposition — Patients who are not clinically ill and do not meet indications for HBO can typically be managed in the emergency department since most symptoms will resolve with high-flow oxygen. Serial COHb levels are generally unnecessary once the diagnosis of CO poisoning is established since symptoms as well as the initial COHb level are usually sufficient to guide management. The unlikely scenario of an oral ingestion of methylene chloride would be an exception to this rule.

Patients whose symptoms do not resolve, who demonstrate ECG or laboratory evidence of severe poisoning, or who have other medical or social cause for concern should be hospitalized. Obtain psychiatric assessment and determination of suicidality in patients with intentional CO poisoning. (See "Suicidal ideation and behavior in adults".)

SPECIAL POPULATIONS

Pediatrics

Symptoms — In young children signs of CO poisoning may be more subtle and nonspecific than those in adults. Infants and toddlers may present with complaints such as fussiness and feeding difficulty as the sole manifestation of CO poisoning [82]. Young children may develop signs and symptoms of CO poisoning faster than older children and adults who experience the same exposure due to their higher oxygen utilization and higher minute ventilation. In contrast, older children have similar symptoms to adults, as they are able to verbalize when they have headache or nausea.

The incidence of delayed neurologic sequelae (DNS) in the pediatric population is reported between 3 and 17 percent (lower than that of adults) [83-85].

Diagnosis and management — Young age does not alter the diagnosis and management of CO poisoning, and hyperbaric oxygen (HBO) treatment centers generally do not modify their approach to therapy based upon age. Authorities make theoretical arguments for both more or less aggressive treatment in young children, but there are no human studies to support changes in current recommendations.

When HBO is administered to young children, the following specific concerns must be addressed:

●Myringotomy should be performed in children with active otitis media who are younger than five years or who are unable to equalize their middle ear pressure [86].

●It may be helpful to allow a family member to accompany a frightened child into the chamber.

●When HBO is administered to an infant, keep the infant warm since hypothermia can easily develop.

●Congenital abnormalities must also be taken into account:

•A chest radiograph should be obtained to detect congenital anomalies such as lobar emphysema, which could lead to a pneumothorax.

•Patients with unpalliated ductal dependent cardiac lesions should undergo HBO with caution as oxygen may precipitate duct closure. In most cases, such patients are poor candidates for HBO, and a pediatric cardiologist should be involved.

Pregnant patients — The threshold to use HBO in pregnant patients is lower because of the greater affinity and longer half-life of CO bound to fetal hemoglobin, the inability to substantially increase placental perfusion, and the direct effects of hypoxemia and acidosis on the fetus. A prospective, multicenter study of fetal outcome following accidental CO poisoning found no physical or neurobehavioral deficits in 31 infants who were exposed to CO in utero when their mothers suffered mild to moderate CO poisoning [87]. Severe maternal poisoning resulted in adverse outcomes in three of five patients treated with normobaric oxygen alone; HBO was used in two other cases, and those children did not demonstrate evidence of prenatal injury. Exposure to HBO does not seem to adversely affect the fetus, but the published experience is limited [88,89].

There is limited information on the characteristics of the fetal heart rate tracing of pregnant patients with CO poisoning in the third trimester. In the few cases with detailed reports, the initial tracing showed baseline fetal tachycardia of 160 to 190 beats per minute in three of four fetuses, and all four had minimal variability with no accelerations or decelerations [90,91]. After 60 to 90 minutes of maternal HBO therapy, all of the tracings became normal.

LONG-TERM OUTCOMES — In patients who survive the initial poisoning, morbidity is primarily related to late neurocognitive impairment and persists in up to 40 percent of victims [2,3]. Significant CO poisoning may be associated with an increased risk of developing a subsequent seizure disorder [92]. The prognosis of patients who sustain hypoxic-ischemic brain injury from any cause, including severe CO poisoning, is reviewed separately. (See "Hypoxic-ischemic brain injury in adults: Evaluation and prognosis", section on 'Prognosis based on clinical findings'.)

Long-term mortality is increased if CO poisoning causes myocardial injury. Long-term follow-up (median 7.6 years) of a cohort of patients with moderate to severe CO poisoning who sustained acute myocardial injury noted a mortality rate of 24 percent [93]. Mortality among patients with myocardial injury was more than twice that of poisoned patients without evidence of such injury and was estimated to be triple the expected rate for a comparable non-poisoned cohort. This study population was young (mean age 47 years) and had a low incidence of established cardiac disease or cardiac risk factors, other than smoking.

PREVENTION — CO monitors equipped with alarms are relatively inexpensive, widely available, and potentially life-saving. The United States Consumer Product Safety Commission (CPSC) recommends that every home have a CO monitor equipped with an alarm; further information is available online www.cpsc.gov or via their hotline (1-800-638-2772). Given the evidence that CO is capable of diffusing rapidly through standard wallboard and floorboard materials, we recommend that every home (even those without an obvious source of CO) be equipped with a CO monitor [94].

ADDITIONAL RESOURCES

Regional poison control centers — Regional poison control centers in the United States are available at all times for consultation on patients with known or suspected poisoning, and who may be critically ill, require admission, or have clinical pictures that are unclear (1-800-222-1222). In addition, some hospitals have medical toxicologists available for bedside consultation. Whenever available, these are invaluable resources to help in the diagnosis and management of ingestions or overdoses. Contact information for poison centers around the world is provided separately. (See "Society guideline links: Regional poison control centers".)

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: Treatment of acute poisoning caused by specific agents other than drugs of abuse".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topic (see "Patient education: Carbon monoxide (CO) poisoning (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Introduction – Carbon monoxide (CO), an odorless, tasteless, colorless, nonirritating gas, is formed by incomplete hydrocarbon combustion. CO poisoning is common, potentially fatal, can lead to permanent neurologic effects, and is probably underdiagnosed because of its nonspecific flu-like clinical presentation. An algorithm and brief table to facilitate emergent management are provided (algorithm 1 and table 1).

●Sources of exposure – CO poisoning is most common during winter in cold climates; in aftermath of natural disasters when portable, gasoline-powered generators are increasingly used; and in smoke inhalation and fire victims. Other potential sources of CO include poorly functioning heating systems, improperly vented fuel-burning devices (eg, kerosene heaters, charcoal grills, camping stoves, gasoline-powered electrical generators), motor vehicles operating in poorly ventilated areas, hookah (water pipe) smoking, and exposure to methylene chloride. (See 'Epidemiology' above and 'Sources of exposure' above.)

●Pathophysiology – CO diffuses rapidly across the pulmonary capillary membrane and binds to the iron moiety of heme forming carboxyhemoglobin (COHb), which is less effective at tissue oxygen delivery. CO also interferes with mitochondrial oxygen utilization and increases lipid peroxidation in the central nervous system (CNS). (See 'Pathophysiology' above.)

●Clinical findings – Common symptoms are nonspecific and include headache, malaise, nausea, and dizziness. Infants and toddlers may present with only fussiness and feeding difficulty. Symptoms of more severe poisoning include chest pain, dyspnea, syncope, and confusion. In the absence of concurrent trauma or burns, physical findings are usually confined to tachycardia, tachypnea, and alterations in mental status, ranging from mild confusion to seizures and coma. (See 'Clinical findings' above.)

The syndrome of delayed neurologic sequelae (DNS) includes variable degrees of cognitive deficits, personality changes, movement disorders, and focal neurologic deficits and occurs in 15 to 40 percent of patients with significant CO exposure. (See 'Delayed neuropsychiatric syndrome' above.)

●Diagnosis – The diagnosis of CO poisoning is made in patients with known or suspected exposure in conjunction with an elevated COHb level measured by cooximetry of a venous blood gas sample. COHb levels greater than three in nonsmokers and greater than 10 to 15 in smokers confirms the diagnosis. The diagnosis of CO poisoning should be suspected in fire victims, in patients with flu-like symptoms in cold climates especially if any other cohabitants or pets are also feeling ill, and patients with unexplained altered mental status or lactic acidosis. Importantly, CO levels do not necessarily correlate with the severity of poisoning. (See 'Diagnosis' above.)

●Post-diagnostic testing – Obtain electrocardiogram (ECG), complete blood count (CBC), chemistries, blood gas, cardiac biomarkers as in (algorithm 1).

●Management – Assess and stabilize airway, breathing, and circulation. Patients who are comatose, have severely impaired mental status, or who do not have sufficient respiratory effort should be intubated without delay and mechanically ventilated using 100 percent oxygen. (See 'Management' above.)

•In a patient with suspected or confirmed CO poisoning, we recommend initial treatment with 100 percent normobaric oxygen via nonrebreathing face mask, regardless of pulse oximetry or arterial PO₂ (Grade 1C). High-flow oxygen decreases the half-life of COHb. (See 'High-flow oxygen' above.)

•In patients with the following indications, we suggest treating with hyperbaric oxygen therapy (HBO) in addition to normobaric 100 percent oxygen (Grade 2C):

-COHb level >25 percent

-COHb level >15 percent in pregnant patient

-Loss of consciousness

-Severe metabolic acidosis (pH <7.25)

-Evidence of end-organ ischemia (eg, ECG changes, elevated cardiac biomarkers, respiratory failure, focal neurologic deficit or altered mental status) (see 'Hyperbaric oxygen therapy' above)

HBO increases elimination of CO in the presence of elevated COHb and potentially prevents DNS.

•In patients suffering from CO poisoning from a fire or smoke inhalation who are critically ill (ie, those with coma, seizure, or hemodynamic compromise associated with a metabolic acidosis and increased blood lactate concentration), we empirically treat possible cyanide toxicity with hydroxocobalamin. The dose is 70 mg/kg intravenous (IV; 5 g is the standard adult dose), and can be repeated after 10 to 15 minutes if the patient does not have rapid clinical improvement. (See 'Patients with smoke inhalation' above and "Cyanide poisoning", section on 'Empiric treatment for smoke inhalation'.)

•In cases of unintentional CO poisoning, the clinician must consider that other people could have ongoing CO exposure if the source is not identified and corrected. Local fire departments or other emergency management personnel can assist with an assessment of ambient CO levels in the suspected environment and removal of victims. (See 'General principles' above.)

●Disposition – Many patients with mild symptoms from an unintentional poisoning can be managed in an emergency department and safely discharged. Patients whose symptoms do not resolve, who demonstrate ECG or laboratory evidence of severe poisoning, or who have other medical or social cause for concern should be hospitalized. (See 'Disposition' above.)