Carbon monoxide · Health
A reassuring oxygen number is not a clearance test
The important distinction is between what a fingertip device displays and what suspected carbon monoxide exposure requires.
- Do not wait for SpO2 to fallA conventional pulse oximeter cannot exclude carbon monoxide poisoning.
- A blood result needs a timelineTime away from the source and oxygen already received affect the interpretation of carboxyhemoglobin.
A small glowing percentage feels like an answer. You clip an instrument onto a finger, wait for the display to settle, and see a number that looks reassuring. It is tempting to let that number settle a much larger question: am I safe?
With possible carbon monoxide exposure, that inference is unsafe. The problem is not simply that somebody moved their hand or bought a cheap device. It concerns what conventional pulse oximetry can distinguish. An instrument can produce a plausible result while lacking the information needed to answer the question being asked of it.
Carbon monoxide interferes with the distinction the sensor makes
Hemoglobin carries oxygen in blood. Carbon monoxide can bind to it, forming carboxyhemoglobin, abbreviated COHb. The American College of Emergency Physicians explains that a standard pulse oximeter can misidentify this form as oxygenated hemoglobin. Its estimate, SpO2, can therefore look reassuring despite impaired oxygen delivery.
This is also why “the oxygen test” is an imprecise phrase. Oxygen dissolved in blood, measured as a partial pressure called PaO2, is not the same quantity as oxygen carried by hemoglobin. ACEP notes that PaO2 can be normal in this setting. A different measurement does not become interchangeable merely because both results concern oxygen.
The practical consequence is narrower than “never trust medical devices.” It is that repeating the same kind of measurement does not necessarily resolve its blind spot. More confidence in the displayed percentage is not more evidence against CO exposure.
Sixteen patients made the false reassurance visible
In a 1994 prospective case series, Buckley and colleagues collected 25 pulse-oximetry readings alongside arterial blood sampling from 16 adults exposed to carbon monoxide. The contrast was striking: none of the reported SpO2 readings was below 96%, although measured COHb reached 44%.
Two reported ranges from the same case series
| Measurement | Reported range | What to retain |
|---|---|---|
| Pulse-oximetry SpO2 | 96–100% | Reassuring-looking readings occurred during CO exposure. |
| Blood COHb | 2.2–44% | This measures a different form of hemoglobin. |
This small, older series is not an accuracy ranking of today's devices. Nor does it estimate how often every exposed person will receive a misleading reading. It demonstrates the specific failure this article concerns. The CDC's clinical guidance still explicitly warns that conventional two-wavelength oximetry is inaccurate when COHb is present.
The blood sample arrives after the exposure has already changed
CDC identifies measurement of blood COHb as central to confirming the diagnosis. Hospital laboratory CO-oximetry can use venous or arterial blood. But even the appropriate test is not a recording of everything that happened before the sample was taken.
Imagine the sequence: somebody leaves the building, travels to care and then has blood taken. These are not three descriptions of the same moment. CDC warns that several hours breathing ordinary air can make COHb testing less useful. ACEP also emphasizes whether supplemental oxygen has already been given, because COHb falls with time, particularly during oxygen treatment.
That does not make the blood test pointless. It changes how the result must be read. CDC states that COHb does not correlate well with illness severity, outcome or treatment response. Clinical assessment and the exposure history remain important. A later number should not be treated as a complete account of an earlier event.
Tell the clinician what the display cannot know
For a handover, the useful information is concrete. Do not remain in an unsafe building to collect it. Once safe, describe:
- Where the possible exposure happened and any suspected fuel-burning source.
- When symptoms began, when you left and whether anyone else became unwell.
- Whether oxygen or other medical care has already been provided.
This checklist is an editorial synthesis of the history and timing questions in the cited guidance, not a diagnostic score. It has no total and no “safe” result.
The NHS notes that symptoms can worsen inside an affected building and improve outside. Headache, nausea and dizziness are among possible symptoms. Those patterns are clues, not proof of a particular cause. UKHSA advises seeking medical advice even when leaving the source makes somebody feel better. An alarm or suspected leak calls for evacuation; do not wait for a second instrument to agree.
The lesson is specific to the question, not hostile to measurement
A conventional fingertip oximeter, a specialist pulse CO-oximeter and a laboratory blood measurement are not interchangeable names. This article is not a buying guide for any of them. It explains why the ordinary SpO2 number must not be used to dismiss a suspected exposure.
There is a broader habit worth keeping: before accepting a result as reassurance, ask what it actually measured. Our feature on platelet storage and transfusion explores a different version of that distinction—preserving material in a bag is not identical to establishing its usefulness in a recipient. Here the stakes are immediate: an oxygen percentage is not permission to remain in a possibly contaminated room.
This is an educational explanation of published evidence, not individualized medical advice or a medically reviewed diagnostic protocol.
Questions about the apparently normal reading
Does a low SpO2 result prove carbon monoxide poisoning?
No. The limitation does not work in reverse: failure to exclude one cause is not a way to identify it. A concerning result or symptoms require appropriate medical assessment.
Can I calculate my COHb from the fingertip percentage?
No home conversion formula is supported here. The case-series ranges are not a calculator, and ordinary SpO2 is not a COHb measurement.
