Pulse Oximetry in Carbon Monoxide Poisoning
Standard pulse oximeters will show falsely elevated readings (typically >90%, often 96-100%) in carbon monoxide poisoning patients, even when actual oxyhemoglobin saturation is critically low.
The Critical Pitfall
Standard two-wavelength pulse oximeters (660 and 990 nm) cannot differentiate carboxyhemoglobin (COHb) from oxyhemoglobin (O2Hb) because they have similar light absorbance characteristics at 660 nm 1. This creates a dangerous clinical scenario where the pulse oximeter reads the COHb as if it were O2Hb, providing falsely reassuring values.
Clinical Evidence
The evidence consistently demonstrates this phenomenon:
In a series of 30 CO-poisoned patients with COHb levels ≥25%, all had SpO2 readings >90% 1. This means patients with only 75% or less of their hemoglobin actually carrying oxygen appeared "normal" on pulse oximetry.
The "pulse oximetry gap" (SpO2 minus actual O2Hb) approximates the COHb level 2, 3. For example, a patient with 40% COHb will show a pulse oximetry reading around 96-100%, while their actual oxyhemoglobin saturation is only 60% or less.
Pulse oximetry readings remained elevated (mean 97.7%) across COHb levels ranging from 2.2% to 44% 2. Even with COHb as high as 44%, pulse oximetry failed to drop below 96% 2.
In severe CO poisoning (COHb >25%), SpO2 consistently overestimated arterial oxygen saturation, with the error increasing as COHb levels rose 4.
Why This Happens
The physics are straightforward: COHb and O2Hb have nearly identical extinction coefficients at the wavelengths used by standard pulse oximeters 1. The pulse oximeter essentially adds COHb + O2Hb together and reports this sum as "oxygen saturation" 5.
Clinical Implications
Never rely on pulse oximetry to rule out or assess the severity of CO poisoning 3, 6, 7. The answer to the question is definitively: Falsely elevated.
What You Must Do Instead
Obtain laboratory CO-oximetry (spectrophotometry) on arterial or venous blood 1. This directly measures COHb, O2Hb, deoxyhemoglobin, and methemoglobin separately.
Start 100% oxygen immediately based on clinical suspicion alone, without waiting for confirmatory testing 1.
Be aware that calculated oxygen saturation from older blood gas machines (using PaO2 and pH algorithms) will also be falsely normal 1. Only direct spectrophotometric measurement is reliable.
Note on Pulse CO-Oximeters
Newer pulse CO-oximeters (available since 2005) that specifically measure SpCO exist, but their accuracy remains controversial 8, 1. They have poor sensitivity (48% in one study) for detecting elevated COHb 8 and should not be used to rule out CO poisoning 6, 7. If used, always confirm with laboratory spectrophotometry before making treatment decisions about hyperbaric oxygen 1.