Who writes this page?
Between 1999 and 2005, I worked at the CHUV in Lausanne, particularly in the ER and the hyperbaric chamber, where carbon monoxide poisonings are treated, among other things. I then started flying, on Robin DR400 and Cessna, before corporate aviation. Here I describe what I've seen from both sides. But I am not a doctor: this page is neither a diagnosis nor medical advice.
In short
Hypoxia and carbon monoxide both deprive your tissues of oxygen, through two different paths, with the same consequence: the first function affected is the one that would allow you to realize it. CO adds a layer: it is odorless, colorless, and tasteless. No external sign is reliable, and, this is the most dangerous trap on the subject, a pulse oximeter does not detect CO: it can display perfectly normal saturation on heavily poisoned blood. The only means of warning you have in flight is a CO detector—the only one, not a foolproof one. In case of doubt: turn off the heating, ventilate, land.
A common point: they first attack the judge
Hypoxia and carbon monoxide are often put in the same box, and for once the shortcut is useful: in both cases, what happens to the pilot is that their cells no longer receive enough oxygen. The path differs, altitude hypoxia because there is less oxygen available to breathe, CO because the blood can no longer transport it, but the most oxygen-hungry organ is the same: the brain.
Hence the real difficulty. Hemorrhaging hurts. An engine failure makes noise. A lack of oxygen, no: it gradually degrades attention, working memory, the ability to question oneself. In other words, it early affects the instrument that should detect the failure. This explains those stories of pilots who "felt very well" while stringing together incoherent decisions.
The practical consequence is easy to remember: on these two subjects, never rely on your feelings as an alarm system. You don't compensate for a faulty sensor with willpower. You replace it with a device (a CO detector) and a rule decided on the ground, in a calm state, applied without discussion the day doubt arrives.
Altitude hypoxia: the lack of oxygen that doesn't hurt
First misunderstanding to clear up: at altitude, the air does not contain "less oxygen". The proportion of oxygen in the air remains largely the same, around 21%, from sea level to very high up. What changes is the pressure. Less pressure means fewer molecules in the same volume of air, hence a lower partial pressure of oxygen in your lungs, and less oxygen actually passing into the blood. You breathe the same, you get less out of it.
This is the mechanism of so-called hypoxic hypoxia, the one that concerns the pilot of an unpressurized aircraft climbing, and the one that strikes the crew of a pressurized aircraft in the event of a depressurization. Two very different situations in terms of timeframe, but the same underlying phenomenon.
The pilot's perspective: the ear as the first sensor
In a pressurized aircraft, I have occasionally experienced the sensation of pressure in my ears preceding my reading of the instruments. I say this because it is true, not because it is a means of detection: a sensation replaces no instrumental monitoring and must never serve as an alarm feeling nothing proves nothing. What is systematic, however, on the jets I fly: the reaction to a depressurization is part of the memory items, oxygen mask first, the checklist following.
What it does, and why everyone has their own signature
The manifestations described are broad and unspecific: headaches, sensation of heat or tingling, shortness of breath, narrowing of vision or loss of colors, clumsy fingers, and above all a change in mood and judgment, euphoria, false confidence, or conversely irritability. The truly important point: these signs do not present themselves in the same order or with the same intensity depending on the person, and they vary in the same person depending on fatigue, hydration, smoking or flu-like symptoms.
This is the purpose of altitude chamber courses, offered in several aeromedical centers: they allow you to discover your own sequence of symptoms in a supervised medical setting, rather than discovering it in flight. This is not an experiment to try on your own, by climbing "just to see": it is precisely the type of test whose result you cannot judge yourself.
Two things this page will not do, for lack of having been able to collate them against the official text: give you the altitude from which supplemental oxygen is required, and give you a quantified time of useful consciousness by flight level. The former is not a matter of medical fitness (Part-MED) but of operational rules (EU Regulation 965/2012 and its annexes depending on the type of operation); the latter circulate in the form of tables whose values differ from one source to another. Refer to your aircraft's flight manual and the text applicable to your operation. What can safely be said: this time is counted in minutes, and it collapses as you climb.
Carbon monoxide: no odor, no color, no taste
Carbon monoxide is a product of incomplete combustion. In light aviation, it enters the cabin through two main routes: the heating system, which recovers heat from the exhaust, and leaks (firewall, control passages, holes around the exhaust). The risk increases in winter, when the heating is on, but it is not seasonal: cases also occur outside the heating season.
What makes it formidable is its chemistry. CO binds to hemoglobin, the molecule that transports oxygen in the blood, with an affinity on the order of 200 to 250 times that of oxygen, forming carboxyhemoglobin. In concrete terms: the seats are taken, oxygen no longer gets on board, and it is not enough to breathe better to compensate. This is a classic order of magnitude in medicine, cited here as such.
The pilot's and nurse's perspective, what I saw in the ER
I worked for several years in the ER of the CHUV, in Lausanne, where I was also assigned to the hyperbaric chamber dedicated mainly to diving accidents in Lake Geneva, but also to CO poisonings. We would see people pulled from fires, sometimes without the slightest burn, who were hospitalized anyway for a check-up. The signs that tipped us off: violent headaches, nausea, sometimes confusion.
And in winter, it wasn't just fires: oil heaters, wood fireplaces, especially during high-pressure weather, when the smoke draws poorly and CO stagnates in the room. Every time, the same trap: the gas is odorless and invisible, it is absorbed without being felt, and without knowing the signs, you don't realize anything. Treatment is the responsibility of the doctor and relies on oxygen, according to modalities they decide.
The "cherry red skin": a real sign, but one you absolutely shouldn't wait for
This is the image everyone remembers: the victim of CO supposedly has an abnormally pink, "cherry red" complexion. The sign exists, carboxyhemoglobin is bright red, but it is rare: medical literature describes it as infrequent in living patients, where pallor rather dominates, and it is mostly reported in the most severe poisonings. An equally important detail: CO does not cause cyanosis, those blueish lips associated with a lack of oxygen, and yet the opposite idea circulates.
Remember the conclusion rather than the detail: no external sign is reliable to spot CO poisoning. Neither on others, nor on oneself. This is exactly why a detector is not a gimmick.
Why it is a problem for piston aircraft, and almost not for jets
The typical scenario primarily affects piston-engine aircraft, both single and twin-engine, where the engine is right in front of the pilot. All it takes is a slightly punctured exhaust pipe, or a seal that is no longer completely tight, for exhaust gases to end up in the cabin.
The pilot's perspective, piston vs. jet
In a business jet or an airliner, cabin air is indeed drawn from the engines, but at the compressor level, upstream of the combustion chamber. No combustion, no CO: this makes carbon monoxide poisoning practically impossible through this route on these aircraft.
On a piston aircraft, two causes dominate: a crack in the exhaust manifold, or (much more frequently) a leak in the cabin heating. This heating is disarmingly simple: an air shroud wrapped around the burning exhaust pipe recovers its heat and sends it into the cabin. If the seal degrades between the two, exhaust gases, hence CO, pass directly into the cabin.
Not to be confused with the fume events talked about on transport aircraft: those concern contamination of the bleed air, most often by engine oils, sometimes by hydraulic or de-icing fluids. It's largely another subject, with its own substances and its own debates, even if some studies discuss the presence of CO among the degradation products of these oils.
The double penalty: when CO joins in at altitude
This is why the two subjects fit on the same page. In an unpressurized aircraft, the higher you climb, the less oxygen is available: that's already altitude hypoxia, and it can be managed. If CO is added on top, the two effects combine: less oxygen to breathe on one side, less capacity to transport it on the other. Vigilance can then drop very quickly, with decisions that have little to do with the reality of the flight, leading up to unconsciousness, without having ever "felt" anything.
It is also this addition that explains why a modest exposure, without consequence on the ground, can count in flight: your margin has already been dented by altitude. And several studies go further than mere addition (pressure drop, shift in the hemoglobin dissociation curve, increased ventilation) and describe an aggravation of the clinical effects of CO at altitude rather than a simple combination. All the more reason not to reason "small dose, small effect".
What the investigations show
The interest here is being able to refer to European and British safety investigations rather than generalities. Three documented situations, each with its lesson.
| Situation | What it teaches |
|---|---|
| A training aircraft, spot detector changed color and exhaust smell after takeoff (BEA investigation) | The identified cause was not a broken part but a lack of sealing around the muffler attachment passages. A discrete defect, invisible during the circuit, but enough to let gases into the cabin. |
| A female instructor poisoned over several flights and several days (BEA investigation: the exposure figure noted has not been collated with the original report for this page, so it is not cited) | Hospitalized after dizziness, nausea, and fatigue. The lesson is accumulation: the danger is not only the sudden peak, it is also the repeated small dose that is blamed on end-of-week fatigue. |
| A fatal piston single-engine accident (AAIB, report AAR 1/2020, Piper PA-46, accident of January 21, 2019) | A high level of carboxyhemoglobin was found. The case has been widely commented on in the UK; the CAA published a safety directive there several years later. Two facts, which I refrain from connecting with an arrow. |
These three cases are cited for what they illustrate, and not as statistics: there is no reliable public figure, to my knowledge, on the actual frequency of CO poisonings in light aviation in Europe. A portion of mild cases are probably never linked to CO; that is the nature of a poison that resembles fatigue.
The symptoms, and why they deceive
Headaches, dizziness, nausea, drowsiness, sudden fatigue. Nothing specific: it feels like a bad night, the onset of airsickness, or dehydration. And while you are looking for a reassuring explanation, the function that should decide, judgment, is among those that are already impaired.
The intensity of the signs depends on the concentration and duration of exposure, but no symptom allows you to estimate for yourself the severity of what is happening. This is precisely what makes CO dangerous: the time taken to link a headache to carbon monoxide is additional exposure time.
The pilot's perspective, my only episode, and it was on the ground
About twenty years ago, as a young First Officer on a Citation CJ1, I was waiting for a client near the terminal while the Captain, already on board, started an engine to prepare the systems and the flight plan. The client took a good half hour longer than expected, and during all this time, the wind blew the jet's exhaust gases towards me. I could smell the kerosene getting stronger and stronger; I didn't pay attention to it. I loaded the luggage, seated the passenger, closed the door.
On takeoff, I felt terribly ill: violent nausea, lightheadedness, a headache coming on. Five minutes after takeoff, I had to vomit in a bag: the first and only time in my career. Looking back, I am convinced it was exhaust gas poisoning. Which of its components was the cause, I have no idea, there was never a test, and the signs I had do not allow any particular gas to be blamed. An important detail: what I smelled was the unburnt kerosene; the CO, however, was the invisible and odorless guest.
That day, there were two of us, and the Captain was flying the plane; I was quite simply incapable of thinking or flying. Imagine the same thing alone, in the middle of a cross-country flight on a light aircraft. That's why I repeat it: at the slightest doubt, headache, nausea, confusion, turn off the heating, ventilate, and land. What the detector says does not change this decision.
Honest clarification: it was an exposure on the ground, in the direct plume of a jet engine, not cabin air in flight, which on this type of aircraft is drawn before combustion. And twenty years later, without a test, it remains a memory, not a diagnosis.
This list of symptoms is a list of warning signals, not a self-diagnosis chart. It allows neither to affirm nor to rule out poisoning: too many ordinary things give the same signs. What it should trigger is an action, see below, then a medical opinion.
Reflexes: prevent, detect, react
EASA popularized a three-step sequence, Prevent, Detect, React, in its Safety Information Bulletin SIB 2020-01R1. It has the merit of fitting into three words and being easy to remember under stress.
| Step | What it means in practice |
|---|---|
| Prevent | This is maintenance: exhaust, muffler, heating system. The pilot has no control over it in flight, but can ask to see the records, and report the slightest exhaust smell noticed on the ground rather than blaming it on the hangar atmosphere. |
| Detect | An on-board detector, visible from the pilot's seat. A spot detector at a minimum, an active detector preferably: details below. It is the only one of these three steps that gives you a sensor your body doesn't have, it being understood that a silent detector does not prove the absence of CO. |
| React | Exhaust smell, detector alarm, or unexplained symptoms: turn off the heating, ventilate (air vents, crack open the canopy if the flight manual allows), land as soon as possible at the nearest airfield, notify air traffic control, ask for medical assistance if you feel unwell. |
A remark on the order: "land as soon as possible" comes after "turn off the heating" and "ventilate" because these two actions take three seconds and immediately reduce exposure. They do not replace landing, they give you what you need to achieve it. And for the exact maneuvers, your aircraft's flight manual is the authority, not a web page.
The pilot's perspective, smoke on board, no arguing
CO is just one facet of a broader principle: as soon as there is a suspicion of fire or smoke on board, electrical, mechanical, whatever the origin, it is a vital emergency, period. You don't try to "finish" the flight: the only objective is to land the plane as quickly as possible. A fire in a plane is like a fire on a boat, there is no room for discussion.
On fast jets flying high, the "cockpit/cabin smoke" procedure begins with memory items: oxygen mask, goggles, oxygen for passengers, descent; the checklist comes next. These actions are specific to pressurized aircraft, and they are those of my flight manual: yours are in yours. What is general is the principle. In a light aircraft, the CO reflex remains the same: turn off the heating, ventilate, land.
Detectors: spot, active unit, headset
This is the only section on this page over which a pilot has direct control. Here are the commonly available means, with what really distinguishes them: the type of alert.
| Means | Alert type | What you need to know |
|---|---|---|
| Chemical spot detector a few euros | Passive it changes color, but you still have to look at it | Limited and highly variable lifespan: from a few dozen days for some common spots to 12 or 18 months for the most durable ones. Must be dated with a marker on the day of opening, and replaced according to the manufacturer's instructions, not according to a general rule. An expired spot is a false sense of security. |
| Active electronic detector a few dozen euros | Active audible and/or visual alarm | It alerts you even if you are not looking at it: that makes all the difference. This is the type of device that British regulations have required since 2025 in certain cases (see below). Remember the battery, and know that most commercial units are not certified for aviation, with alarm thresholds and delays that vary from one model to another. |
| Headset with integrated CO sensor | Active directly in the ear | Some high-end headsets incorporate a sensor, notably the Lightspeed Delta Zulu. To be considered as an uncertified aid to vigilance: an additional argument for a headset that is primarily bought for its other qualities, not approved safety equipment. |
| ⚠️ Pulse oximeter | - | Does not detect CO. Useful for monitoring oxygenation at altitude, useless, and deceptive, for carbon monoxide. See just below. |
The pilot's perspective, it hasn't always been a reflex
When I started out in light aircraft, as a pilot then as an instructor, especially on Robin DR400, and on Cessna 152, 172 and 177, CO detectors had not yet become commonplace. They probably existed, but I never heard of them once during my training: no classes, no prevention on the subject. Awareness came later, following accidents. Today, taking one on board has become much more common in flying clubs and schools, and that's a good thing.
Prices and lifespans are given as an order of magnitude: they vary depending on the model and the country, and the manufacturer's instructions are authoritative. No model is recommended here. The principle, however, does not vary: a detector you don't look at is useless.
The dangerous trap: the pulse oximeter does not "see" CO
Many light aircraft pilots carry a pulse oximeter, that little device you clip to your fingertip, to monitor their saturation on the climb. This is a legitimate use. But it is accompanied by a belief that is false, and dangerous: that it would also serve as a carbon monoxide alert.
The standard pulse oximeter does not differentiate between hemoglobin loaded with oxygen and hemoglobin loaded with carbon monoxide. As a result: in case of CO poisoning, it can display a perfectly normal, even reassuring saturation, while the blood is heavily poisoned. This is an established point in medical literature, and it is a classic false friend, even for healthcare professionals.
Only a dedicated device, a CO-oximeter, or a blood test actually measures carboxyhemoglobin. In flight, the consequence is direct: a pulse oximeter does not replace a CO detector. Two different devices, for two different dangers.
Corollary not to be missed: a low saturation recorded in flight is not a warning sign of CO. It can indicate altitude hypoxia, which is already useful, but it says nothing about carbon monoxide, one way or the other.
The regulatory perspective: a British obligation, not a European one
This is the kind of subject where exact information becomes false when crossing a border. So let's be precise.
| Where | What applies |
|---|---|
| United Kingdom | The CAA published Safety Directive SD-2024/001 (version 3 of February 3, 2025). Since January 1, 2025, operators of piston-engine aircraft registered in the UK, or registered elsewhere and flying in UK airspace, must have a functioning active CO detector on board, alerting by an audible and/or visual signal, when flying with passengers who do not hold a recognized pilot qualification. The stated objective is the protection of uninformed passengers. Exemptions exist (single-seat, open cockpit, aerobatics, engine located above or behind the cabin, engines only on the wings, with some caveats): they are listed in the directive itself, to be consulted directly. If you cross the Channel, this rule concerns you. |
| EASA Airspace France, Belgium, Luxembourg, and Switzerland, which participates in the EASA system without belonging to the European Union (point to be confirmed with the FOCA for any Swiss question) | No equivalent obligation to my knowledge to date. EASA addressed the CO risk in light aviation through the Safety Information Bulletin SIB 2020-01R1, Carbon Monoxide Risk in Small Aeroplanes and Helicopters (issued January 27, 2020, revised October 19, 2021), which itself bears the statement "This is information only. Recommendations are not mandatory.": it is a non-binding recommendation, unlike an airworthiness directive. In other words: strongly advised, not imposed. |
Never present the British obligation as European, it is a common confusion. And note the date: this subject evolves, and it is national. If your operation is concerned, it is the text of your authority that decides.
A final remark of common sense: the absence of an obligation is not an argument. An active detector represents a small expense compared to the cost of operating an aircraft.
Frequently Asked Questions
Can you smell carbon monoxide in flight?
Where does CO in an aircraft cabin come from?
Are jets and airliners concerned?
Can a pulse oximeter detect CO poisoning?
What are the first symptoms of CO poisoning?
Does the skin turn red in case of CO poisoning?
What to do if the detector goes off or I feel unwell in flight?
Is a spot detector enough, or is an active detector needed?
Are hypoxia and carbon monoxide the same problem?
From what altitude is oxygen required in flight?
Pour aller plus loin
Sur les autres effets de la pression et de l'altitude sur l'organisme : plongée et vol, quel délai respecter le même mécanisme de pression, pris par l'autre bout. Sur l'oreille et le bruit du cockpit : l'audition du pilote. Pour l'ensemble de la visite médicale : la page d'ensemble et préparer sa visite. Le hub : santé du pilote.
Les pages du Labo restent en accès libre, celle-ci comprise. À côté, il existe désormais un guide payant consacré à la visite médicale : le rétroplanning, les questions précises à poser à l'AME, les annexes par pays. Rien de ce qui est gratuit aujourd'hui ne passera derrière un paiement. Voir le guide.