A carbon monoxide blood test measures carboxyhemoglobin, the percentage of Hemoglobin in your blood that is bound to carbon monoxide instead of oxygen. It is run on a blood sample with a CO-oximeter, it confirms recent exposure, and it is read alongside symptoms, because the number alone does not grade how unwell someone is.
Hemoglobin is the iron-containing protein inside red blood cells that picks up oxygen in the lungs and releases it to tissue. Each hemoglobin molecule has four binding sites, and carbon monoxide competes for exactly those sites. That is the whole reason a gas you cannot see or smell shows up on a blood test at all.
How much hemoglobin you carry also sets the size of the tank. Someone with a lower hemoglobin level has less total oxygen-carrying capacity to start with, so the same share of sites occupied by carbon monoxide leaves less oxygen moving.
What hemoglobin is, how it is tested, what the standard and optimal ranges mean and what a high or low value can reflect are covered in full on the hemoglobin biomarker page. The rest of this guide stays on the carbon monoxide question.
It measures carboxyhemoglobin, usually written COHb, and reports it as a percentage of your total hemoglobin that is currently bound to carbon monoxide rather than oxygen. A result of 5% means roughly five in every hundred binding sites are occupied by carbon monoxide at the moment the blood was drawn.
The analysis is done by co-oximetry. A co-oximeter passes several wavelengths of light through a blood sample and separates the different hemoglobin species, so oxyhemoglobin, deoxyhemoglobin, carboxyhemoglobin and methemoglobin are each reported separately 1. Because the method reads hemoglobin species rather than dissolved gas, the sample does not have to be arterial in most situations, and a venous sample is commonly acceptable for measuring carboxyhemoglobin 2.
Two things the percentage is not. It is not the same measurement a household alarm makes; an alarm reads carbon monoxide concentration in air in parts per million, while the blood test reads what has already crossed into the bloodstream. It is also not a lifetime record. Carbon monoxide leaves the blood once a person is away from the source, so the result describes recent exposure, not exposure last month.
A small amount of carbon monoxide is made inside the body during normal breakdown of heme, which is why the baseline percentage in a healthy person is low but not zero 1.
Carbon monoxide binds hemoglobin far more tightly than oxygen does, so even small amounts in the air can occupy a meaningful share of binding sites and reduce how much oxygen the blood can carry 1. That displacement is the part most people know about, and on its own it would already explain fatigue, headache and breathlessness.
The second effect is less obvious and matters just as much. When carbon monoxide occupies some sites, the remaining oxygen is held more tightly and released less readily to tissue, so delivery suffers more than the raw percentage would suggest 3. Blood can look adequately loaded while muscle and brain receive less than they need.
There is also an effect inside the cell. Research suggests that carbon monoxide binds proteins beyond hemoglobin, including myoglobin in muscle and components of the mitochondrial respiratory chain, which may impair energy production independently of oxygen delivery 1. Studies indicate that oxidative stress and inflammatory injury follow significant exposure and contribute to effects on the heart and brain 3.
This combination explains a pattern clinicians describe. The organs with the highest oxygen demand, brain and heart, tend to show symptoms first, which is why confusion, chest discomfort, poor balance and loss of consciousness are the signs that most urgently need emergency care 1. Observation, meaning and next step stay separate here. A raised percentage is an observation. What it may reflect depends on the exposure story. What happens next is a clinical decision.
Typical laboratory reference ranges for carboxyhemoglobin are low in people who do not smoke and noticeably higher in people who do, which is why most labs report the two separately.
| Marker | Typical lab reference range | FitScript optimal range |
|---|---|---|
| Carboxyhemoglobin, adult who does not smoke | roughly 0% to 3% of total hemoglobin | no optimal range published for this marker |
| Carboxyhemoglobin, adult who smokes | commonly higher, often reported up to around 10% of total hemoglobin | no optimal range published for this marker |
| Hemoglobin, men | 13 to 17.5 g/dL | 14 to 16 g/dL |
| Hemoglobin, women | 11.5 to 16 g/dL | 12 to 15 g/dL |
These are typical lab reference ranges, and they vary by laboratory, by analyser, by age and by sex, so always read your result against the range printed on your own report. The hemoglobin optimal ranges shown are FitScript's optimization targets for a healthy adult, not diagnostic cut-offs, and they are narrower than the standard range by design.
Carboxyhemoglobin behaves differently from most biomarkers in one important way. There is no optimization target for it, because the goal is simply as little exogenous carbon monoxide as possible. It is a point-in-time exposure measurement ordered in urgent settings, not a marker you track quarterly for performance. A hemoglobin and hematocrit result, by contrast, is a stable part of routine bloodwork and is worth knowing in advance, because it tells you how much oxygen-carrying capacity you had before anything went wrong.
Tobacco smoke contains carbon monoxide, so people who smoke carry a continuously topped-up carboxyhemoglobin level that sits above the non-smoking baseline. Smoking is associated with a persistently elevated carboxyhemoglobin level, which is the reason labs report a separate reference range for smokers 4.
This creates a genuine interpretation problem in two directions. A smoker with a mildly raised result may simply be showing their own baseline rather than an environmental exposure. Equally, a smoker exposed to a faulty appliance starts from a higher point, so the same additional exposure lands them at a higher total. Clinicians handle this by asking when the last cigarette was, how heavy the habit is, and whether anyone else in the household feels unwell.
Secondhand smoke, hookah use, and working in an enclosed space with combustion engines can all nudge the baseline too. So can occupational exposure near furnaces, boilers, forklifts, boats and ice resurfacing machines.
The practical takeaway. A number read without the story around it is close to meaningless. The story includes smoking status, timing, who else is affected, what appliances are in the building, and whether symptoms improve when the person leaves that building.
An elevated result reflects recent absorption of carbon monoxide from somewhere, and the job after the test is finding that somewhere. The usual sources are incomplete combustion indoors or in a poorly ventilated space, which includes faulty or unvented furnaces and water heaters, blocked flues, gas or kerosene heaters, wood and charcoal burning, portable generators, and vehicle exhaust in attached garages 1.
Two less familiar routes are worth knowing. Smoke inhalation from a house fire delivers carbon monoxide along with everything else in the smoke 5. And methylene chloride, a solvent found in some paint strippers, is absorbed and then converted to carbon monoxide inside the body, so exposure can keep producing carbon monoxide after the person has left the room 4.
Here is the honest limit of the test. A result within the typical range does not rule out an exposure that happened hours earlier, because the level falls once the person is breathing cleaner air 2. This is why clinicians weigh the exposure history and the symptom pattern at least as heavily as the number itself.
If a result comes back above the range on your own report, that is worth discussing with a clinician promptly rather than filing away, and the source needs identifying before anyone returns to the space. Any next step, including further testing, belongs to the clinician reviewing the whole picture.
Because a standard fingertip pulse oximeter cannot tell carbon monoxide-bound hemoglobin apart from oxygen-bound hemoglobin, so it counts carboxyhemoglobin as if it were carrying oxygen. A conventional two-wavelength pulse oximeter may therefore display a reassuring saturation in a person with significant carbon monoxide exposure 1.
The physics explains it. Carboxyhemoglobin and oxyhemoglobin absorb red light similarly enough that a two-wavelength device cannot separate them. A laboratory co-oximeter uses more wavelengths and reports each hemoglobin species individually, which is exactly why the blood measurement exists 1. Dedicated non-invasive CO-oximeters used in some emergency settings work on the same multi-wavelength principle.
There is a parallel trap with blood gas reports. If the oxygen saturation on a blood gas is calculated from the measured oxygen tension rather than directly measured by co-oximetry, it can also look normal, because dissolved oxygen is not the problem here 2. The saturation has to be measured, not computed, for the result to mean anything in this setting.
The same applies to skin colour. The cherry red appearance described in older teaching is uncommon and unreliable, and its absence says nothing useful 6. The practical consequence for a reader at home is simple. A normal pulse oximeter reading is not a safety check for carbon monoxide, and a working alarm is.
The early symptoms are headache, dizziness, nausea, fatigue, poor concentration and general malaise, which overlap almost completely with viral illness, migraine, food poisoning and simple exhaustion. Carbon monoxide poisoning is described as a clinical chameleon precisely because this non-specific presentation leads to it being mistaken for other conditions 6.
A few features shift the odds and are worth holding onto.
More than one person in the same building feeling unwell at the same time is the single most useful clue, and it includes pets, who are often affected first because of their size and their time spent indoors. Symptoms that fade within hours of leaving the house and return on coming back point at the building rather than the person. Onset that tracks the heating season, a newly serviced or newly installed appliance, a generator running during a power cut, or a blocked chimney all matter.
At higher exposure the picture changes to confusion, poor coordination, chest pain, shortness of breath, fainting, seizure or loss of consciousness 1. Any of those, in anyone, is an emergency and needs 911, not a wait-and-see approach.
In pregnancy there is an added consideration, because carbon monoxide crosses the placenta and the fetus is more vulnerable than the pregnant person, so exposure during pregnancy warrants urgent assessment 3.
No, and this is the most consistently misunderstood part of the test. Studies indicate that the measured carboxyhemoglobin level correlates poorly with symptom severity and with eventual outcome, so a modest percentage does not promise a mild course and a high one does not reliably predict a severe one 5.
There are solid reasons for the mismatch. The level falls continuously from the moment a person leaves the source, so a sample drawn after an ambulance ride, or after oxygen has been given, reads lower than the peak the body actually experienced 2. The test captures a single moment, while the harm relates to how high the level went and how long it stayed there. And part of the injury happens inside cells rather than in the blood, so a blood percentage cannot fully represent it 1.
This is why guidance treats the result as one input among several. Clinical assessment, the exposure history, neurological examination and other investigations carry weight alongside the number, and the diagnosis is not made on the percentage alone 2.
For a reader, the usable version is this. Do not let a low-ish number talk you out of seeking care when the story and the symptoms point at an exposure, and do not panic at a number in isolation. The person in front of the clinician matters more than the decimal on the report.
A venous blood draw is usually sufficient, because co-oximetry measures hemoglobin species rather than gas tension, and arterial sampling is reserved for situations where the clinician also needs the other information a blood gas provides 2. The draw itself is an ordinary blood collection into a tube suitable for the analyser.
Timing is the critical variable. Carbon monoxide is cleared through the lungs, and the blood level drops once the person is breathing air without it, so the sooner the sample is taken after removal from the source, the closer it sits to the true peak 2. Research suggests that elimination is considerably faster when supplemental oxygen is given, which is useful clinically but also means a sample drawn after oxygen has started will read lower 4. Treatment decisions belong to a licensed provider.
Two practical points follow. First, whoever interprets the result needs to know the clock, how long since the person left the environment and whether oxygen was given before the draw, or the number cannot be placed in context. Second, samples are analysed promptly and handled according to the lab's protocol, because delays and poor handling degrade accuracy.
This is also why carboxyhemoglobin is not a routine wellness marker. It answers a question about a specific moment, usually in an urgent care setting, rather than describing a stable feature of your physiology.
The most directly relevant companions are the ones that describe oxygen-carrying capacity, because carbon monoxide reduces the usable fraction of whatever capacity you have. A hemoglobin level and hematocrit tell you the size of that capacity, and lower values mean less margin for the same proportional loss.
In an acute assessment, clinicians commonly look wider than the carboxyhemoglobin percentage alone. A blood gas with measured co-oximetry, lactate as a marker of tissue oxygen debt, cardiac markers where chest pain or an abnormal electrocardiogram is present, and basic chemistry and muscle markers after prolonged unconsciousness are all part of the standard assessment described in clinical reviews 1. Carbon monoxide exposure is associated with cardiac injury in a meaningful share of people with significant poisoning, which is why the heart is specifically assessed rather than assumed to be fine 1.
For someone whose exposure has been dealt with and who is asking what to monitor afterwards, the answer is usually neurological and cognitive follow-up arranged by their clinician rather than a specific blood marker.
For general health tracking outside an emergency, knowing your baseline hemoglobin and hematocrit, iron status and overall red cell picture is simply useful context. It tells you, and a clinician, how much oxygen-carrying reserve you normally run with. That context is decided with your clinician, not from a single value read in isolation.
Some people develop new neurological or cognitive symptoms days to weeks after an apparent recovery, a pattern described as delayed neurological sequelae. Research suggests that these delayed effects can include memory and concentration difficulties, mood changes, personality change, and movement or balance problems appearing after an initial improvement 7.
The possibility exists across a range of exposure severities, and studies indicate that the initial carboxyhemoglobin level does not reliably identify who will go on to develop delayed symptoms 5. That is one more reason the single percentage is a weak predictor and the overall clinical assessment carries the weight.
This is why follow-up matters after any significant exposure, even when the person feels well on discharge. Reviews of acute carbon monoxide poisoning describe persistent cognitive and neurological symptoms as an important outcome to look for rather than an unusual curiosity 7. Carbon monoxide poisoning is associated with effects on both the brain and the heart, so cardiac symptoms appearing afterwards are worth reporting too 3.
Practically, if someone in your household had a confirmed exposure and weeks later seems slower, more forgetful, unusually low in mood or unsteady, that observation is worth raising with their clinician. It should not be filed under stress or a bad patch without a conversation. The clinician decides what assessment, if any, follows.
Because an alarm detects the gas before anyone absorbs it, while a blood test can only confirm what has already been absorbed. Carbon monoxide has no colour, no smell and no taste, so human senses provide no warning at all, and detection has to be mechanical 1.
A few durable points of practice. Place alarms on every level of the home and near sleeping areas, since symptoms during sleep progress without anyone noticing. Test the alarms on the schedule the manufacturer specifies, replace batteries when prompted, and replace the units themselves at the end of their stated service life, because the sensors degrade. Have fuel-burning appliances, chimneys and flues checked by a qualified technician at the intervals the manufacturer or installer recommends.
Never run a portable generator, a grill or a camping stove inside a home, garage, tent or boat cabin, and keep generators well away from doors, windows and vents. Do not warm a vehicle in an attached garage, even with the door open.
If an alarm sounds, get everyone and every pet outside into fresh air first and call emergency services from outside. Do not reopen windows and go hunting for the source, and do not go back in until the responders say it is safe. The blood test is what happens after that, and it is better never to need it.
Any suspected exposure deserves medical assessment on the same day, and anyone who is confused, faint, short of breath, has chest pain, is pregnant, or has lost consciousness needs emergency care immediately. Call 911 rather than driving yourself.
There is also a slower version of this question. People who repeatedly feel unwell at home and better when away, households where more than one person has unexplained headaches through the heating season, and anyone working around combustion engines or furnaces who notices recurring symptoms all have a reason to raise carbon monoxide specifically with a clinician, because the presentation is easy to attribute to something else.
Bring the context with you. Timing of symptoms, who else and which pets are affected, what fuel-burning appliances are in the building, whether an alarm is fitted and when it was last tested, smoking status, and any solvent or paint stripper use. That information changes how a carboxyhemoglobin result is read.
This article is educational and is not medical advice. Discuss your results, your symptoms and any treatment with a clinician or licensed provider, who evaluates what is appropriate for you; treatment decisions belong to a licensed provider.
For readers tracking oxygen-carrying capacity as part of general health rather than in an emergency, the related biomarker guides cover the standard markers in depth, including hematocrit, the proportion of blood volume made up of red cells. FitScript members have 150+ biomarkers tested per year, with results interpreted by Atlas, FitScript's AI health advisor, and any treatment decision made by a licensed provider after reviewing the member.
No. Carboxyhemoglobin has to be requested specifically and is measured on a co-oximeter, which is standard equipment in emergency and hospital laboratories rather than part of a general wellness panel. Routine panels include hemoglobin and hematocrit, which describe oxygen-carrying capacity but say nothing about carbon monoxide.
As soon as possible after the person is away from the source. The level falls steadily once someone is breathing cleaner air, so a sample taken hours later reads lower than the peak. A result inside the usual range does not rule out an earlier exposure, which is why the clinician also weighs symptoms and the exposure history.
Usually not. A venous sample is commonly acceptable for measuring carboxyhemoglobin, because the analyser reads hemoglobin species rather than dissolved gas. An arterial sample may be taken when the clinician also needs the other information a blood gas provides, and that choice is theirs to make.
A standard fingertip pulse oximeter cannot distinguish hemoglobin bound to carbon monoxide from hemoglobin bound to oxygen, so it can display a reassuring number during an exposure. Only a laboratory co-oximeter, or a dedicated multi-wavelength CO-oximeter used in some emergency settings, separates the two. Never use a pulse oximeter as a carbon monoxide safety check.
Smokers generally carry a higher baseline than non-smokers, which is why most labs print a separate reference range for them. That baseline makes interpretation harder in both directions, because a mildly raised value may be the person's normal, or may be an environmental exposure stacked on top of it. Smoking status, timing of the last cigarette and whether others in the building feel unwell all help sort it out.
Get every person and pet outside into fresh air first, then call emergency services from outside the building. Do not stay inside to look for the source or to open windows, and do not go back in until responders confirm it is safe. Medical assessment follows for anyone with symptoms.
No. Home alarms measure carbon monoxide concentration in the air, not in your blood, and there is no consumer blood test for this. The two measurements answer different questions, and the alarm is the one that warns you before absorption happens.
A clinician reviews the result alongside symptoms, the exposure history, the time since leaving the source and the examination, and decides what assessment or monitoring follows. The source also needs identifying and fixing before anyone returns to the space. Treatment decisions belong to a licensed provider.