Fast, free, expert help
24 hours a day, every day

Carbon Monoxide Poisoning



Carbon monoxide (CO) poisoning remains an important cause of illness and death. A colorless, odorless gas produced by the combustion of any organic material, it has been implicated in approximately 5,000 deaths per year in the United States (based on a review of death certificates). The American Association of Poison Control Centers (AAPCC) received reports of over `17,000 CO exposures in 2001, resulting in over 5,000 hospitalizations and 35 deaths. Intoxication can range from mild headache to coma and death, and neurological and neuropsychiatric sequelae are not uncommon in survivors. Controversy remains about the role of hyperbaric oxygen in the management of CO-poisoned patients.

Case presentation

A 60 year old woman was found unconscious in her car. She was treated by paramedics with high-flow oxygen and she was waking by the time they reached the emergency department. The oxygen saturation by pulse oximetry was 100%. The initial carboxyhemoglobin (COHgb) level was 29%. Three hours later she was fully awake, and ambulating without difficulty. Her blood pressure was 110/70 and her heart rate was 76/min. The hospital did not have a hyperbaric chamber, and called the poison control center to obtain information about the location of chambers in the area and advice about indications for HBO treatment.


  1. What is the pathophysiology of CO poisoning?
  2. What are the common sources of CO?
  3. What are the typical presenting symptoms and signs?
  4. Why was the patient’s pulse oximetry normal if she had a 29% saturation of her hemoglobin by CO?
  5. What is the treatment for CO poisoning? What are the indications for HBO?


Carbon monoxide is an odorless, nonirritating gas that readily binds to hemoglobin, reducing the ability of the blood to carry oxygen. The affinity of CO for hemoglobin in approximately 250 times that of oxygen, so a CO concentration of as little as 0.1% is capable of producing a COHgb of 50%. Besides the reduction in oxygen-carrying capacity (a “functional anemia”), binding of CO to hemoglobin alters its ability to release oxygen to tissues, further exacerbating the hypoxic injury. Animal evidence also points to ischemia due to hypotension and shunting of blood from poorly vascularized cerebral tissue as an important contributor to brain damage. Finally, animal models suggest an important role for lipid peroxidation and other post-anoxic injury mechanisms in the genesis of CNS injury.

Clinical presentation

Patients with CO poisoning are easily recognized if they are brought comatose from a burning building or found in an automobile unconscious with a suicide note. On the other hand, the diagnosis may be elusive if the patient presents with no known history of exposure and nonspecific symptoms such as headache, nausea or dizziness. One study found a surprising number of patients presenting to emergency departments in the winter with nonspecific flu-like complaints who had evidence of CO poisoning on screening breath testing. Because of the ubiquitous sources of CO, exposure can occur almost anywhere; some important sources include: riding in the back of covered pickup trucks; using small gas-powered tools such as pressure-washers, compressors and concrete cutters; indoor ice arenas using the Zamboni ice-resurfacing machine; indoor use of unvented gas-powered heaters or cooking equipment; and many others .

With mild poisoning, patients often complain of headache, dizziness, nausea and sometimes vomiting. Shortness of breath is not uncommon. With more serious intoxication syncope, seizures and coma may occur. Findings on examination may include unusually pink skin and mucous membranes, although this is actually uncommon in clinical case reports. Tachycardia is common, and hypotension may occur. Neurological findings may range from mild confusion to coma. Cerebellar ataxia has been associated with a greater risk of neuropsychiatric sequelae .


Pulse oximetry may give falsely normal readings because it is not capable of distinguishing between oxyhemoglobin and carboxyhemoglobin, which are both red in color. Arterial blood gases usually reveal a normal pO2, because the carbon monoxide concentration required to cause poisoning is low enough that it does not alter the amount of oxygen physically dissolved in the plasma. However, ABGs will usually reveal a metabolic acidosis in patients with serious poisoning. Specific testing for carboxyhemoglobin (COHgb) is needed to determine the degree of hemoglobin saturation. Although COHgb levels do not always correlate well with the severity of intoxication, levels >25% are considered significant. In patients with smoke inhalation (eg, in a fire), other gases may also have been inhaled, causing irritant injury to the lung as well as systemic toxicity such as cyanide poisoning or methemoglobinemia . CT scanning may reveal damage to the central white matter, although this is not usually apparent on initial presentation.


Initial treatment includes establishing a patent airway and administration of the highest available concentration and flow of oxygen. In ambient air (21% oxygen) the half-life of the COHgb complex is approximately 180 minutes, compared with approximately 75 minutes using 100% oxygen at sea level pressure (1 atmosphere), also referred to as “normobaric oxygen” (NBO). Hyperbaric oxygen (HBO), which is 100% oxygen delivered in a chamber pressurized to pressures above 1 atmosphere (usually 2.5-3 atmospheres) can further speed the removal of carbon monoxide (half-life approximately 20 min) and can also provide sufficient oxygen dissolved in plasma to meet basic metabolic oxygen demands. Animal studies have also demonstrated that hyperbaric oxygen reduces lipid peroxidation and perivascular injury associated with hypoxia-ischemia. Human case reports and clinical series suggest a benefit of HBO in reducing post-exposure neuropsychiatric complications compared with normobaric oxygen or ambient air, but there are few controlled studies and they report conflicting results. A prospective, randomized, double-blind study using sham HBO found no benefit from HBO (Scheinkestel CD et al: Med J Australia 1999; 170:203-10). A more recent double-blind randomized study (Weaver LK et al: New Eng J Med 2002; 347:1057-67) found a small but statistically significant benefit of HBO based on one subgroup score on neuropsychiatric testing (Trailmaking Part A) and self-reported symptoms of memory impairment at 6-week follow-up after treatment. Other neuropsychiatric scores and self-reported symptoms did not differ between the groups. Thorough information on results of testing at 6 months or longer was not reported. Previous studies (Thom SR et al: Ann Emerg Med 1995; 25: 474-80) have suggested that differences between patients treated with HBO and NBO seen early after treatment often resolve after several months.

Based on these studies, it remains unclear whether HBO is the treatment of choice for CO poisoning. Many variables need to be evaluated, including the severity of the intoxication, the interval between removal from exposure and the initiation of HBO, distance to a hyperbaric chamber, and the presence of complications such as hypotension or seizures. Based on the most recent study, Weaver recommends HBO be considered if the following are present: history of loss of consciousness; metabolic acidosis; COHgb >25%; age >50 years; or the presence of cerebellar abnormalities on neurologic examination.

Answers to Qs: Key take-home points

  1. Carbon monoxide causes cellular hypoxia by reducing oxygen carrying capacity and oxygen delivery to tissues, and it may also affect intracellular oxygen utilization.
  2. CO can be produced by burning any organic material, and is a ubiquitous poison.
  3. Syndromes and signs can be nonspecific, making diagnosis difficult in patients with mild or moderate poisoning.
  4. Pulse oximetry is usually normal because the device is unable to distinguish between the color of carboxyhemoglobin and oxyhemoglobin. A specific test for carboxyhemoglobin saturation is needed (using a co-oximeter type of blood gas machine) is required.
  5. Delivery of 100% oxygen is the initial treatment of choice. Hyperbaric oxygen should be considered for patients with severe poisoning (eg, loss of consciousness, metabolic acidosis, cerebellar findings) although clinical studies offer conflicting evidence for its value. Contact the California Poison Control Center for assistance in determining the need for HBO and location of regional hyperbaric chambers.

Table I: Symptoms and Signs of CO Poisoning

  • Headache
  • Dizziness
  • Nausea and vomiting (occasionally, diarrhea)
  • Confusion, agitation, anxiety
  • Seizure
  • Syncope
  • Stupor or coma
  • Tachycardia
  • Hypotension
  • Chest pain, cardiac ischemia (in patients with coronary artery disease)

Table II: Treatment of CO Poisoning

  • Remove the victim from exposure
  • Administer oxygen by highest available concentration
  • Endotracheal intubation if appropriate
  • Normobaric oxygen: 100% oxygen by tight-fitting mask or ET tube for several hours, until COHgb level is <5%
  • Hyperbaric oxygen: 100% oxygen delivered in a hyperbaric chamber (2.5 ATM). Controversial but may reduce neuropsychiatric sequelae.

Table III: Suggested indications for HBO

  • History of loss of consciousness
  • Metabolic acidosis
  • COHgb > 25%
  • Age > 50 years
  • Cerebellar abnormalities on neurologic exam