A Primer on Carbon Monoxide Poisoning: The Myths, the Evidence, and the Expert Opinion

By Eric C. Walter, MD, MSc, Pulmonary and Critical Care Medicine, Northwest Permanente and Kaiser Sunnyside Medical Center, Portland, is Associate Editor for Critical Care Alert.

Dr. Walter reports no financial relationships relevant to this field of study.

Synopsis: This is a generally well-written review of the (admittedly limited) literature regarding diagnosis and management of carbon monoxide poisoning from several experts in the field.

Source: Hampson NB, et al. Practice recommendations in the diagnosis, management, and prevention of carbon monoxide poisoning. Am J Respir Crit Care Med 2012;186:1095-1101.

Hampson and colleagues present a review of the literature and pathophysiology of carbon monoxide (CO) poisoning and provide treatment and prevention recommendations. CO poisoning accounts for an estimated 50,000 emergency department visits annually and is a leading cause of poisoning deaths. It is well recognized that CO leads to tissue hypoxia by binding with hemoglobin, displacing oxygen, and shifting the oxyhemoglobin dissociation curve to the left. It is now appreciated that direct cellular damage from immunologic and inflammatory mechanisms also is important.

The authors stress that CO poisoning is a clinical diagnosis. The diagnostic triad involves: 1) recent CO exposure, 2) symptoms of CO poisoning, and 3) an elevated carboxyhemoglobin (COHb) level (> 3-4% in nonsmokers or > 10% in smokers). Symptoms are nonspecific and do not correlate with COHb levels. The COHb level does not predict clinical course or outcome and only serves to document exposure. Symptoms may include headache, dizziness, nausea, vomiting, confusion, fatigue, chest pain, shortness of breath, or loss of consciousness. The “cherry red” skin color often described in association with CO poisoning only occurs at lethal COHb levels, and even then is rarely seen.

The presence of COHb in the blood can cause confusion when interpreting arterial oxygenation. Historically, blood gas analyzers calculated arterial oxygen saturation using the partial pressure of oxygen and pH rather than measuring saturation directly. This calculation was made irrespective of the amount of COHb in the blood and could lead to erroneously normal calculated oxygen saturations. For example, a person with 40% COHb could have a calculated oxygen saturation of 100% when, in fact, 40% of the hemoglobin was bound to CO. Fortunately, most current blood gas analyzers now measure the concentrations of oxy-, deoxy-, carboxy- and methemoglobin directly, avoiding this potential clinical pitfall. Arterial oxygenation estimates via pulse oximetry can also be erroneous in the setting of elevated COHb levels. Standard two-wavelength pulse oximeters cannot differentiate COHb and oxyhemoglobin. Therefore, in the presence of COHb, pulse oximetry may not accurately reflect the true percent of oxyhemoglobin saturation. In a study of 30 patients with at least 25% COHb (in whom true oxyhemoglobin levels could thus be no greater than 75%), the saturation indicated by pulse oximetry was > 90% in all of them.1

Oxygen is recommended as treatment despite the lack of clinical trials demonstrating benefit of 100% normobaric oxygen over air. Treatment should be started as soon as the diagnosis is considered. When hyperbaric oxygen is not available, the authors recommend 100% oxygen be given until the COHb level normalizes (< 3%) and the patient’s presenting symptoms have resolved. However, the authors go on to say that COHb levels can be expected to normalize within 6 hours, and if a patient has been compliant with oxygen therapy for 6 hours and is asymptomatic, a repeat COHb level is not needed.

A number of trials have attempted to compare normobaric and hyperbaric oxygen therapy for CO poisoning. Unfortunately, many have had significant methodological flaws — making interpretation of their results challenging. However, based on these studies and clinical experience, the authors recommend at least considering hyperbaric oxygen in all cases of acute CO poisoning. The goal of hyperbaric therapy is not to decrease mortality — only about 3% of CO-poisoned patients who present to medical care die — but to decrease short- and long-term neurocognitive dysfunction.

Finally, the authors discuss prevention of CO poisoning. Many accidental CO poisonings occur from the use of CO-producing materials indoors — such as the use of charcoal briquettes or gasoline-powered generators for heating and cooking following power outages. Public health programs and CO alarms are both advocated to help decrease CO poisonings.


My recollections of the teaching of CO poisoning from medical school and residency were to look for the “cherry red” skin color and that all patients must be rushed to the nearest hyperbaric chamber — turns out both were wrong (mostly). Given the lack of good science and historically dogmatic teaching about this common poisoning, this clinical review by four experts in the field is timely and needed. The authors freely admit this review should not be considered a “meta-analysis” of CO poisoning management but rather a consensus of expert opinion.

Overall, the review is well written and provides information that will be very useful to providers caring for patients with CO poisoning. The authors could have better clarified the need to check COHb levels after treatment with normobaric oxygen. However, on this topic dominated by expert opinion and lacking evidence-based medicine to guide treatment decisions, it is likely that no consensus exists. Since most patients are seen and treated in the emergency department, this review may be most useful to providers working in that setting. Nevertheless, critical care providers are likely to be asked for treatment recommendations and decisions regarding disposition.


  1. Hampson NB. Pulse oximetry in severe carbon monoxide poisoning. Chest 1998;114:1036-1041.