Carbon monoxide poisoning

Contents

Overview

Carbon monoxide (CO) is an odourless, colourless gas produced by incomplete combustion of carbon-containing fuels. It is the most common cause of fatal poisoning in the UK. CO poisoning can be accidental (faulty appliances, house fires, inadequate ventilation) or deliberate. Early and accurate diagnosis depends on clinical suspicion and CO-oximetry — standard pulse oximetry is unreliable.

Pathophysiology

CO exerts toxicity through several mechanisms:

Haemoglobin binding: CO binds haemoglobin with approximately 240 times the affinity of oxygen, forming carboxyhaemoglobin (COHb). COHb is incapable of carrying or releasing oxygen. The oxygen dissociation curve of remaining oxyhaemoglobin shifts to the left (the Haldane effect), impairing oxygen delivery to tissues even at lower COHb levels than would be predicted from simple displacement alone.

Cytochrome oxidase inhibition: CO binds cytochrome c oxidase (complex IV of the mitochondrial electron transport chain), directly impairing cellular respiration and causing intracellular hypoxia independent of its effects on haemoglobin. This accounts for toxicity at tissue level, and explains why severe poisoning can produce profound cellular injury even when some oxygen-carrying capacity remains.

Myoglobin binding: CO binds myoglobin in cardiac and skeletal muscle, impairing oxygen storage and contractile function. This contributes to myocardial dysfunction and rhabdomyolysis.

Oxidative stress: CO generates reactive oxygen species during reperfusion, causing lipid peroxidation and delayed neurological injury.

Sources and Epidemiology

Common sources of CO:

  • House fires and confined space exposure (smoke inhalation — CO is the leading cause of death in house fires)
  • Faulty gas boilers, cookers, and central heating systems
  • Poorly ventilated use of generators, petrol-powered tools, and BBQs indoors
  • Car exhaust in enclosed spaces

Seasonal variation is pronounced — poisoning peaks in winter months when heating systems are in use and buildings are poorly ventilated.

Clinical Features

Severity correlates broadly with COHb level and duration of exposure, but there is significant individual variability.

Mild to moderate (COHb 10–30%): headache (classically frontal), nausea, dizziness, malaise, and reduced exercise tolerance. These symptoms are non-specific and frequently misattributed to viral illness.

Severe (COHb >30–40%): confusion, syncope, chest pain, cardiac arrhythmias, seizures, and coma. In house fires, victims may be found unconscious.

Signs: The classic 'cherry-red' skin colour occurs from COHb and is unreliable — it is often absent and may mislead clinicians into excluding the diagnosis. Cyanosis is absent (COHb is red). Tachycardia is common; hypotension indicates severe poisoning with cardiac involvement.

Cardiac involvement: CO causes myocardial ischaemia and arrhythmias. Troponin elevation is a marker of severity and predicts poor outcome. New ECG changes (ST changes, arrhythmias) should prompt urgent cardiology review.

Neurological: Confusion, reduced GCS, and focal deficits may occur acutely. Seizures indicate severe poisoning.

Diagnosis

Pulse oximetry is unreliable in CO poisoning. Standard pulse oximeters measure the differential absorption of light at two wavelengths (660 nm and 940 nm) to distinguish oxyhaemoglobin from deoxyhaemoglobin. They cannot distinguish COHb from oxyhaemoglobin — both absorb similarly at 660 nm. As a result, SpO2 reads falsely normal (often 98–100%) despite severe COHb. A normal SpO2 does not exclude CO poisoning.

CO-oximetry (multi-wavelength spectrophotometry on an ABG analyser) measures COHb directly. An ABG must be taken and analysed on a CO-oximeter-equipped blood gas machine. COHb levels:

COHb Interpretation
<3% Normal (non-smokers); up to 10% in heavy smokers
10–25% Mild-moderate poisoning
25–40% Severe poisoning
>40% Life-threatening; high mortality without treatment
>50% Often fatal

Note: COHb levels fall with time and supplemental oxygen treatment. A low level at presentation does not exclude prior severe exposure if treatment has already started or if time has elapsed.

Additional investigations: ECG (ischaemia, arrhythmia), troponin, CK (rhabdomyolysis), lactate (cellular hypoxia), renal function, urine myoglobin.

Management

Immediate

Remove the patient from the source of CO. Resuscitate as needed.

High-flow oxygen via non-rebreather mask at 15 L/min: The most important initial intervention. Breathing 100% oxygen reduces the elimination half-life of COHb from approximately 5 hours (on room air) to 60–90 minutes by competitively displacing CO from haemoglobin and increasing dissolved oxygen in plasma. Continue until COHb is below 5% and the patient is asymptomatic.

Intubation: Required for patients with GCS ≤8, respiratory failure, or refractory seizures. In these patients, 100% FiO2 via the ventilator accelerates COHb clearance.

Seizures: Benzodiazepines are first-line treatment. Standard AED protocols apply.

Cardiac monitoring: Continuous ECG monitoring for at least 12 hours in patients with cardiac symptoms or high COHb. Arrhythmias are treated per standard protocols.

Cyanide co-poisoning: Consider in victims of house fires, particularly if there is a combined high anion gap metabolic acidosis disproportionate to the COHb level. Hydroxocobalamin (Cyanokit) 5 g IV is the preferred cyanide antidote and does not interfere with oxygen delivery (cf. sodium thiosulphate, which may be harmful in severe CO poisoning).

Admission

All symptomatic patients should be admitted. Home carbon monoxide detector advice and environmental assessment (gas safety check) should be arranged on discharge. Psychiatric assessment is required if the exposure may be deliberate.

Hyperbaric Oxygen

Hyperbaric oxygen (HBO) at 2.5–3 atmospheres reduces the COHb half-life to approximately 20 minutes, and is thought to reduce late neurological injury by several mechanisms beyond accelerating CO elimination: it increases dissolved plasma oxygen, may mitigate CO-mediated oxidative injury, and may reduce the delayed neuropsychiatric syndrome.

UK indications for HBO (UHMS guidelines; also referenced in Toxbase and national guidance):

  • Loss of consciousness at any point
  • Neurological symptoms (confusion, seizure, focal deficit)
  • COHb >25% (some centres use >20%)
  • Cardiac involvement (ischaemia, arrhythmia, elevated troponin)
  • Pregnancy (any significant CO exposure — fetal haemoglobin has higher CO affinity; fetal susceptibility is greater)

HBO is delivered in a pressurised chamber; not all centres have access. The National Poisons Information Service (NPIS) and toxicology advice should guide decisions, and transfer to a HBO centre should be arranged where indicated and safe.

Evidence for HBO: The Weaver trial (NEJM 2002) showed that three sessions of HBO within 24 hours of poisoning significantly reduced cognitive sequelae at 6 weeks compared with normobaric oxygen alone. This is the strongest evidence base for HBO benefit.

Delayed Neuropsychiatric Syndrome

A clinically important complication of severe CO poisoning is the delayed neuropsychiatric syndrome (DNS), occurring 2–40 days after apparent recovery. Features include:

  • Cognitive impairment and memory loss
  • Personality change, emotional lability, irritability
  • Parkinsonism (bradykinesia, rigidity, tremor)
  • Urinary incontinence
  • Psychosis

DNS occurs in up to 30% of patients with severe poisoning. The mechanism involves delayed neuronal injury from oxidative stress and lipid peroxidation, with characteristic white matter changes on MRI (particularly in the globus pallidus and periventricular regions). HBO therapy reduces the incidence of DNS, which contributed to its adoption in severe poisoning.

Patients should be warned of DNS risk, and GPs and families should be informed to watch for neuropsychiatric changes in the weeks after discharge.

Viva Questions

Why is standard pulse oximetry unreliable in carbon monoxide poisoning?

Standard pulse oximeters use two wavelengths of light — 660 nm (red) and 940 nm (infrared) — to calculate the ratio of oxyhaemoglobin to deoxyhaemoglobin. The algorithm is calibrated on the assumption that haemoglobin exists in only two states. Carboxyhaemoglobin has a similar optical absorption profile to oxyhaemoglobin at 660 nm, and unlike deoxyhaemoglobin, it does not absorb strongly at 940 nm. As a result, the pulse oximeter misidentifies COHb as oxyhaemoglobin and reports a falsely elevated (or normal) SpO2 — often 98–100% — even in severe poisoning. This is clinically dangerous because it provides false reassurance. CO-oximetry, which uses four or more wavelengths of light simultaneously, can differentiate COHb, oxyhaemoglobin, and deoxyhaemoglobin accurately, and must be used to diagnose and monitor CO poisoning. A point-of-care finger clip that claims to measure SpCO (carboxyhaemoglobin saturation) may be available in some units, though its accuracy is inferior to formal co-oximetry on an ABG analyser.

What are the indications for hyperbaric oxygen in CO poisoning?

Hyperbaric oxygen is indicated when there is evidence of significant systemic toxicity from CO. The main indications are: loss of consciousness at any point, even if the patient has recovered on arrival; neurological symptoms including confusion, seizures, or focal deficits; COHb above 25%; cardiac involvement in the form of ischaemia, arrhythmias, or troponin elevation; and any significant exposure in pregnancy, because fetal haemoglobin has a higher affinity for CO and the fetus is more vulnerable to CO-mediated hypoxia. The evidence base for HBO is strongest for reducing the delayed neuropsychiatric syndrome, as demonstrated in the Weaver trial (NEJM 2002), which showed significantly lower cognitive sequelae at 6 weeks with three HBO sessions compared with normobaric oxygen. In practice, access to an HBO chamber is limited, and transfer decisions must weigh the potential benefit against the risks of transport in a patient who may be haemodynamically unstable. The National Poisons Information Service provides real-time advice on indications and transfer logistics.

What is the mechanism of delayed neuropsychiatric syndrome after CO poisoning?

The delayed neuropsychiatric syndrome occurs days to weeks after apparently successful treatment of acute CO poisoning, and is thought to result from ongoing neuronal injury driven by oxidative stress. During reperfusion and recovery, reactive oxygen species are generated — a consequence of CO's effects on the mitochondrial electron transport chain and subsequent lipid peroxidation. White matter is particularly vulnerable, and characteristic changes are seen in the globus pallidus and periventricular regions on MRI. Inflammatory mechanisms, including neutrophil-mediated oxidative injury and complement activation, also contribute to delayed injury. Clinically, the syndrome manifests as cognitive impairment, memory loss, personality change, Parkinsonism, and in severe cases psychosis. It affects up to 30% of patients with severe poisoning. The mechanism provides the rationale for hyperbaric oxygen therapy — HBO reduces lipid peroxidation and oxidative injury, and the Weaver trial demonstrated a significant reduction in cognitive sequelae with three sessions of HBO compared with normobaric oxygen, supporting a direct effect on the delayed injury process.