Carbon monoxide poisoning is confirmed through a blood test that measures how much carbon monoxide is bound to your hemoglobin. This measurement, called carboxyhemoglobin (COHb), is the gold standard and the only reliable way to confirm exposure. The test is straightforward, but timing matters: carbon monoxide clears from your blood relatively quickly, especially if you’ve been breathing fresh air or receiving oxygen, so earlier testing gives more accurate results.
The Blood Test That Confirms It
The definitive test for carbon monoxide poisoning is a blood draw analyzed by a device called a CO-oximeter. A standard blood sample is collected in a heparinized tube (a green-top tube in most hospitals), and the lab measures what percentage of your hemoglobin is occupied by carbon monoxide instead of oxygen. The sample must be kept sealed and airtight, since exposing it to room air can alter the results.
Either arterial or venous blood can be used. Arterial draws (from the wrist) are slightly more involved but are often collected anyway as part of a blood gas panel in the emergency department. Venous blood from a standard arm draw works for the COHb measurement itself.
What the Numbers Mean
Normal carboxyhemoglobin in a nonsmoker is below 3%. Smokers can have baseline levels up to 12% due to chronic carbon monoxide inhalation from cigarettes. Newborns also tend to run higher, at 12% or above, because of differences in how their hemoglobin breaks down after birth.
Levels above 20% are considered potentially critical. Symptoms, however, don’t always match the number neatly. Some people feel severely ill at 15%, while others with higher levels may initially seem fine. That’s why the blood test is essential: you can’t gauge the severity of exposure by symptoms alone.
Why Standard Pulse Oximeters Miss It
One of the most dangerous pitfalls in carbon monoxide poisoning is that a regular pulse oximeter, the clip placed on your finger, will often show a normal or near-normal oxygen reading even when your blood is saturated with carbon monoxide. This is because standard pulse oximeters can’t distinguish between hemoglobin carrying oxygen and hemoglobin carrying carbon monoxide. Both absorb light in a similar way, so the device reads them as the same thing.
Research published in CHEST confirmed that pulse oximeter readings consistently overestimate true oxygen levels in carbon monoxide patients, and the error grows larger as carbon monoxide levels rise. In other words, the sicker you are, the more the device lies. A reading of 98% on the finger clip could mask a dangerously high carbon monoxide level. The only way to get an accurate picture is through laboratory CO-oximetry on an actual blood sample.
Non-Invasive Screening Devices
Some emergency departments and fire departments use a newer type of pulse oximeter designed specifically to detect carbon monoxide through the skin. These devices (the Masimo Rad-57 is the most widely studied) clip onto your finger like a standard oximeter but use additional wavelengths of light to estimate carboxyhemoglobin without a blood draw.
The concept is appealing, but accuracy remains a problem. A study in Annals of Emergency Medicine found that these devices had a sensitivity of only 48% for detecting carboxyhemoglobin levels at or above 15%, meaning they missed roughly half of significant poisoning cases. The readings could swing as much as 11 to 14 percentage points away from the true lab value in either direction. While a high reading on one of these devices is useful as a red flag, a normal reading cannot rule out poisoning. A blood test is still required for confirmation.
Timing Changes Everything
Carbon monoxide clears from your blood faster than many people realize, and this directly affects how accurate your test results will be. When you’re breathing 100% oxygen through a mask (the standard first treatment), the half-life of carboxyhemoglobin averages about 74 minutes. That means roughly half of the carbon monoxide bound to your hemoglobin is gone in just over an hour. On room air alone, the half-life is longer, commonly cited around 4 to 5 hours, though individual variation is significant.
This has a practical implication: if paramedics put you on high-flow oxygen during transport and you don’t get your blood drawn for another hour or two, your COHb level at the hospital may be dramatically lower than it was at the time of your worst exposure. A “mildly elevated” result doesn’t necessarily mean you had mild poisoning. Emergency physicians take the timing of oxygen administration into account when interpreting results, but the takeaway for you is that getting tested as quickly as possible gives the most accurate snapshot of your exposure.
Heart Damage Screening
Carbon monoxide doesn’t just starve your brain of oxygen. It also directly injures heart muscle, and this damage is far more common than many people expect. In one large study, 54% of carbon monoxide poisoning patients had elevated troponin levels, a protein released into the blood when heart cells are damaged. That’s more than half of all patients showing evidence of cardiac injury.
Among those with elevated troponin who underwent advanced cardiac imaging, 69% showed signs of injury within the heart muscle itself. Notably, standard ultrasound of the heart (echocardiography) missed many of these cases, with only 42% showing visible problems on that exam. The reassuring news is that the vast majority of patients, roughly 88%, recovered normal heart function over time.
Because of these findings, troponin blood tests and heart rhythm monitoring (ECG) are now considered important parts of the workup for anyone with significant carbon monoxide exposure. If your troponin comes back elevated, your medical team may recommend more detailed cardiac imaging to assess the extent of injury.
Testing for Delayed Neurological Effects
One of the more unsettling aspects of carbon monoxide poisoning is that neurological symptoms can appear days to weeks after the initial exposure, even in people who seemed to recover fully. These delayed effects can include memory problems, difficulty concentrating, personality changes, movement disorders, and vision loss.
There is no single blood test that predicts who will develop these delayed symptoms. Instead, doctors rely on neurological exams, cognitive screening tools, and brain imaging. MRI scans can reveal characteristic patterns of damage in the brain’s white matter and deep gray structures. In some cases, formal neuropsychological testing, essentially a structured series of memory, attention, and reasoning tasks, helps document the extent of cognitive impairment. Vision testing and nerve conduction studies may be added if visual symptoms are present.
The challenge is that most patients don’t have baseline neurological testing from before their poisoning, making it difficult to measure exactly how much function was lost. Because delayed symptoms are unpredictable, follow-up appointments in the weeks after a significant exposure are important for catching problems early. If you or someone around you notices new cognitive or behavioral changes after a carbon monoxide event, that warrants prompt evaluation even if the initial recovery seemed complete.
Home Detection vs. Medical Testing
It’s worth distinguishing between detecting carbon monoxide in your environment and detecting it in your body. A home carbon monoxide detector measures gas concentration in the air (in parts per million) and can alert you to a dangerous leak. This is prevention, not diagnosis. If your alarm goes off or you suspect exposure, the medical tests described above are what determine whether your body has actually been affected and how severely.
If you suspect chronic low-level exposure, such as recurring headaches that improve when you leave home, a carbon monoxide detector can help confirm an environmental source. But to document the physical impact on your body, you’ll need a blood COHb level drawn while you’re still in or recently removed from the exposure environment. Waiting a day to visit your doctor after airing out the house may result in a normal blood level that doesn’t reflect what was happening during the exposure.