Best Test for Carbon Monoxide Measurement
The gold standard test for measuring carbon monoxide is laboratory co-oximetry (spectrophotometry) of arterial or venous blood to measure carboxyhemoglobin (COHb) levels. 1
Recommended Testing Method
Laboratory blood co-oximetry should be performed on either arterial or venous blood samples, as COHb levels are similar between the two when CO body stores are in equilibrium with lung CO partial pressure. 1 The test works by transilluminating blood with multiple wavelengths of light, measuring differential absorbance, and calculating concentrations from the known absorption spectrum of each hemoglobin species. 1
Diagnostic Thresholds
- Nonsmokers: COHb ≥3-4% is considered elevated 1
- Smokers: COHb ≥10% is considered elevated 1
- Typical smoker baseline: 3-5% (approximately 2.5% increase per pack/day) 1
Noninvasive Pulse CO-Oximetry: NOT Recommended for Diagnosis
Noninvasive pulse CO-oximetry cannot be used to accurately diagnose CO toxicity in the emergency department and should not replace laboratory blood testing. 1 The 2017 ACEP guidelines found that noninvasive testing had only 48% sensitivity (95% CI 27-69%) for detecting COHb >15%, missing more than half of poisoned patients. 1
Why Pulse CO-Oximetry Fails
- Poor sensitivity (65% pooled across studies) makes it unreliable for ruling out poisoning 2
- Agreement with blood testing exceeded acceptable ranges in 33% of patients 1
- Device fails to report values when oxygen saturation drops below approximately 85% 3
- The American Journal of Respiratory and Critical Care Medicine recommends laboratory confirmation for any patient being considered for hyperbaric oxygen therapy 1
Limited Role for Pulse CO-Oximetry
If pulse CO-oximetry is used at all, it should only serve as a rapid screening tool, with all positive or clinically suspicious cases confirmed by laboratory blood co-oximetry. 1, 4 The high positive likelihood ratio (9.4) means a positive result is useful, but the poor negative likelihood ratio (0.38) means a negative result cannot exclude poisoning. 2
Critical Pitfalls to Avoid
Standard Pulse Oximetry is Useless
Standard two-wavelength pulse oximeters (measuring SpO2) cannot detect carboxyhemoglobin and will show falsely normal readings. 1 COHb and oxyhemoglobin have similar absorbances at 660 nm, causing standard pulse oximeters to read COHb as if it were oxyhemoglobin. 1 Patients with 25% COHb can have SpO2 readings >90%. 1
Older Blood Gas Machines
Verify that your blood gas analyzer has an integrated co-oximeter that directly measures COHb. 1 Older machines calculate oxygen saturation from PaO2 and pH using algorithms, reporting normal saturation (97-98%) regardless of COHb levels—a patient with 40% COHb and PaO2 100 mmHg would be incorrectly reported as 97-98% saturated. 1
Clinical Context
Initiate 100% oxygen immediately in any patient with suspected CO poisoning while awaiting laboratory confirmation—do not delay treatment for test results. 1 The COHb level confirms diagnosis but does not correlate with symptoms or predict outcome. 1 Levels may be low or normal if significant time has elapsed between exposure and testing, or if oxygen has already been administered. 1