Carbon Monoxide Misconceptions: Tolerance, Canaries, and Why There's No Air Hunger

Corrections around carbon monoxide, several of which invert the intuitive answer. ## About tolerance **You cannot build useful resistance, and acclimatisation makes things worse.** The binding chemistry is fixed. The one documented human acclimatisation — Esther Killick's 1930s self-experiments — produced adaptation by raising red cell mass, which increases baseline carboxyhaemoglobin and does nothing against an acute dose. See Esther Killick and the CO Self-Experiments. **The headache is the alarm, so losing it is a loss.** Someone acclimatised to a chronically leaking appliance is the person least likely to notice the night it turns lethal. This is the inverse of how tolerance is usually imagined — compare Mithridatism: Building Poison Tolerance Through Repeated Exposure, where the same intuition is examined for a different poison. **Being a smoker does not help.** A smoker runs at 10–15% carboxyhaemoglobin, so an acute exposure starts from a worse baseline, and any symptomatic dulling actively removes the warning. Carboxyhaemoglobin also normalises within about a day of quitting, and the extra red cell mass reverses over the following weeks as red cells turn over — so a recent ex-smoker has no residual "tolerance" in either direction. ## About the survival case The widely-shared account of a woman who identified a CO leak while her family slept is instructive, and the causal factors are usually ranked wrongly. **The decisive factor was being awake.** CO kills sleeping people because it deepens CNS depression without triggering the arousal reflex that would wake you to a threat. She was upright when the heating started; the others took their dose unconscious. One family member's collapse on standing is textbook CO ataxia and reflects a far deeper impairment produced by exactly that difference. **Quitting smoking helped through attribution clarity, not tolerance.** CO poisoning feels like being stoned or hungover. Having been clean for months, that sensation had no innocent explanation available and registered as an anomaly demanding a cause. A current smoker would have had a ready explanation and used it. The mechanism is epistemic, not physiological — and the physiological version would have hurt. **Prior priming is what produced the hypothesis.** Someone had drilled the possibility into her for years. Under CO impairment, recalling an available hypothesis is far easier than generating a novel one. ## About sensing it **Humans are not uniquely blind — no animal senses CO.** There was no ancestral exposure to select for a receptor, CO is chemically inert, and it is invisible to the oxygen alarm we do have. **There is no air hunger, and that is the point.** The carotid body measures oxygen *tension*, not *content*. In CO poisoning dissolved oxygen pressure is normal, so the sensor never fires. A suffocating person fights for air; a CO victim feels nothing of the kind. See The Carotid Body Measures Oxygen Tension, Not Oxygen Content. **Pulse oximetry can read normal in severe poisoning**, because standard oximeters cannot distinguish oxyhaemoglobin from carboxyhaemoglobin. CO-oximetry is required. **The canary was not a detector.** It was a faster casualty — high metabolic rate and highly efficient flow-through lungs meant it collapsed sooner. The same ordering applies to pets and children in a house. **Even nature's best fire-finder doesn't use CO.** *Melanophila* fire beetles locate forest fires from tens of kilometres away using infrared receptors; in tests, smoke did not attract them. ## About the biology **Your body makes CO deliberately.** Heme oxygenase releases it during normal heme recycling, which is why every non-smoker carries 1–3% carboxyhaemoglobin. It is one of three gasotransmitters alongside NO and H₂S. See Gasotransmitters: The Lethal Gases Your Body Manufactures as Signals. **Cells detect CO everywhere — for signalling, never for danger.** Bacterial CooA and RcoM, and mammalian NPAS2 and soluble guanylate cyclase, all detect it. None of them warns you about anything. ## About detectors and response **The alarm delays are deliberate, not defects.** UL 2034 alarms integrate concentration × time because that models carboxyhaemoglobin accumulation. A brief spike is harmless; sustained exposure is not. **Standard alarms are silent below roughly 30 ppm by design**, while 20–40 ppm chronically causes real symptoms in vulnerable people. A low-level monitor covers that gap. See Electrochemical CO Detectors and the UL 2034 Time-Delay Curve. **Detector sensors expire in about 5–7 years** regardless of battery status. **Airing out the room does not clear your blood.** Carboxyhaemoglobin has a 4–5 hour half-life on room air, versus 40–80 minutes on 100% oxygen. "Open the windows and go back to bed" is how people relapse. **Feeling fine afterwards is not reassurance.** Significant exposures carry a 10–30% risk of delayed neurological sequelae appearing days to weeks later.