← Longevity & Supplement Guides
Bedroom Air Quality and Sleep: The Free Variable Nobody Measures
Somewhere in the last decade we learned to measure sleep in exquisite detail and quietly stopped looking at the room it happens in. There are scores for the sleeper, stages for the sleeper, a readiness percentage for the sleeper. For the sealed box the sleeper spends eight hours inside, there is almost nothing.
It is an odd blind spot, because the room is the part you can change tonight, for free, without changing a single thing about your behaviour.
A closed bedroom is a slow experiment
An adult asleep is a small, steady source of carbon dioxide. Shut the door and the window, and the concentration in the room climbs for as long as you lie there. Outdoor air runs a little over 400 parts per million. In the Danish dormitory rooms studied by Strøm-Tejsen and colleagues, a night with the window closed averaged 2,585 ppm — and with the window open, 660 ppm.
Here is the part that makes this interesting rather than alarming. In this literature, carbon dioxide is almost certainly not the villain. It is the odometer. It happens to be cheap to measure continuously, and it rises in lockstep with everything else that accumulates in a sealed room overnight — the compounds a warm body gives off, the humidity, whatever the mattress and the carpet are exhaling. CO₂ is the number researchers use because it stands in for the real question, which is simply: how much of this air have you already breathed?
Bedroom air
What happens to the air in a closed room overnight
800 ppm or below
The design target the review recommends bedrooms be ventilated to hit.
1,000 ppm
The lowest average concentration at which disturbed sleep showed up across the pooled datasets.
≈8 L/s per person
The airflow needed to hold that target — roughly double what many residential standards prescribe.
The measured effect sizes, in full
Kang and colleagues put 36 healthy young adults through three ventilation conditions averaging 750, 1,000 and 1,300 ppm. Against the 750 ppm night: sleep efficiency fell 1.3% at 1,000 ppm and 1.8% at 1,300 ppm; time awake rose by 5.0 and 7.8 minutes; deep sleep was shorter at 1,300 ppm; and morning salivary cortisol was higher at 1,300 ppm, which the authors read as raised sympathetic activity.
The two end-points are measured study averages from Strøm-Tejsen et al., Indoor Air 2016 (window open vs closed in single-occupancy dormitory rooms). The shape of the overnight rise between them is a schematic, not recorded data. The 800 and 1,000 ppm lines are the values proposed by Akimoto et al., Science and Technology for the Built Environment 2025 (ASHRAE 1837-RP).
The number a ventilation committee argued its way to
Last September a team from Waseda University, the Technical University of Denmark, Shanghai Jiao Tong and the National University of Singapore published the result of an ASHRAE-funded project with the unglamorous name 1837-RP. They pooled 17 studies containing 22 experimental datasets and asked two blunt questions: at what point does bad bedroom air start costing you sleep, and how much ventilation does it take to stay under that point.
Their answers: 1,000 ppm was the lowest average concentration at which sleep disturbance turned up. 850 ppm was the highest at which nothing was observed — and they declined to treat that as a safe threshold, on the honest grounds that the sensors aren't precise enough to defend it. So they proposed a design target of 800 ppm, which needs roughly 8 litres per second per person of outdoor air, about twice what many residential standards currently require.
There is something quietly bracing about watching a standards committee conclude that the standard is wrong.
How big is this, honestly?
Small. Let me not dress it up.
In the most carefully dosed experiment we have, 36 young adults slept under conditions averaging 750, 1,000 and 1,300 ppm. Moving from the cleanest air to the worst cost them 1.8% of sleep efficiency and 7.8 additional minutes awake. Deep sleep shrank at the highest concentration, and morning cortisol was higher.
Eight minutes is not a transformation, and anyone selling you a ventilation product on the strength of it is overselling. What makes the finding worth your attention is not the size of the effect but its denominator. It costs nothing. It requires no willpower, no schedule change, no new habit to fail at in week three. It applies to roughly a third of your life, every night, indefinitely. Very few things in this field are that cheap to run, and most of the ones people spend money on are supported by less.
The trial where nobody knew what was happening
The methodological problem with sleep-environment advice is that you always know which condition you're in. Open the window and you have also given yourself the expectation of sleeping better.
Fan and colleagues solved this elegantly. Twenty-nine bedrooms, four weeks: one baseline week, then a week each of low, medium and high ventilation in balanced order, produced by covertly changing the speed of the existing extract fan. Nothing else in the room changed. Participants were not told when the changes would happen, or whether they would happen at all.
In the twelve bedrooms where the CO₂ readings confirmed the three conditions had genuinely separated, people got significantly less deep sleep, more light sleep and more awakenings on the low-ventilation nights. In the twenty-three bedrooms where at least the high and low conditions separated cleanly, deep sleep was shorter under low ventilation. The evening and morning cognitive tests found nothing at all.
Two things are true here and the post is dishonest if it only tells you one. The effect survived blinding, which is more than most sleep-hygiene advice can say. And the analysis is conditional on the intervention having actually worked — which, in a real building, it often didn't. Turning up a fan is not the same as changing the air in a room. That is a limitation of the study and, at the same time, its most practical finding.
The morning after
The oldest of these experiments is still the most useful, because it followed people into the next day.
Strøm-Tejsen and colleagues ran two week-long crossovers in single-occupancy student rooms. In the pilot, ventilation was changed by opening a window: 660 ppm versus 2,585 ppm. In the second, a silent fan in the intake vent either ran whenever CO₂ passed 900 ppm or was disabled: 835 versus 2,395 ppm. Sleep was measured with wrist actigraphs; mornings brought questionnaires and two performance tests.
With the lower carbon dioxide, measured sleep quality improved, the air was perceived as fresher, reported sleepiness and ability to concentrate improved, and people did better on a test of logical thinking.
The detail I find most persuasive is a negative one. Bedroom temperature varied widely across the study — but it did not differ between the ventilation conditions. Which kills the obvious objection, that opening a window simply cools the room and temperature is doing all the work. Here it wasn't.
It was never only carbon dioxide
In February, Lin and colleagues published a field study of 183 young adults whose bedroom environment and sleep were monitored on the night before a standardised fitness test. Fine particulate matter — PM2.5 — was significantly associated with a smaller proportion of deep sleep and with worse performance on the long-distance run the following day. That second association got worse on nights when CO₂ was also high; the high-CO₂ nights in that sample averaged 3,961 ppm.
It is a field study using regression, so it shows association, not cause, in a young and fit sample on a single night. Take it as a pointer rather than a proof. But the pointer aims somewhere useful: the thing to keep out of the room matters as much as the thing to let in.
Which is why "just open a window" is not a universal answer. Outdoor air arrives with its own particulate load, its own pollen, its own traffic noise, and in February its own heating bill. Ventilating a bedroom on a still night beside a busy road is a trade, not a free win. This is precisely the kind of question a general rule handles badly and your own measurements handle well.
What to change this week
Tonight
Three levers, cheapest first
1 · Open something
The bedroom door is free, silent, doesn't cool the room much and doesn't let the street in. In most homes it is the single largest change you can make to overnight air without buying anything. A window on the latch does more, at the cost of noise and heat.
2 · Run a three-and-three
Three nights with the door open, three with it shut, holding bedtime, room temperature, alcohol and last coffee as steady as you can manage. Judge it on how the air smells when you wake and on your own overnight trend — not on a single good night, which every condition produces eventually.
3 · If you can't ventilate, shift the timing
Loud street, cold month, poor outdoor air: air the room out well before bed rather than during it, keep the door open once you're asleep, and deal with particulates during the day. A closed room that started clean is better than one that started stale.
Practical steps drawn from the interventions used in the cited trials (window, vent fan, extract fan). Effects in those studies were modest; treat this as a cheap experiment, not a fix.
Use your nose as the crude sensor. The perception these trials kept picking up, night after night, was the freshness of the air on waking — and unlike sleep efficiency, you get that reading for nothing. If the room smells slept-in in the morning, ventilation lost.
If you own a carbon dioxide meter, the number to aim at overnight is 800 ppm or below. If you don't, don't rush to buy one: an open door will get most people most of the way, and the meter is more useful for settling an argument than for changing a behaviour you'd have changed anyway.
Then read the trend rather than the night. Overnight resting heart rate, heart-rate variability and how long you take to fall asleep, compared with your own baseline over a fortnight rather than anyone else's numbers — that is what the Agen Band is for, and it is the only honest way to tell whether a change this small did anything for you specifically. One caution worth stating plainly: if you snore heavily, gasp or wake unrefreshed however well the room is ventilated, that is a conversation for a doctor, not a ventilation project. No consumer device diagnoses anything.
The bottom line
The evidence here is better than most sleep advice and the effect is smaller than most sleep advice claims. Bad bedroom air costs healthy young adults a percent or two of sleep efficiency and a handful of minutes awake; it shows up in deep sleep, it survives blinding, and it seems to carry into the next morning's concentration. The threshold the specialists converged on is 800 ppm, and most bedrooms with the door shut sail past it before midnight.
What I like about this finding is that it is the opposite of optimisation theatre. Nothing to buy, nothing to track, nothing to be disciplined about. Just the recognition that you spend a third of your life inside a small sealed volume, and that the number on your sleep app has never once asked what was in the air. Leave the door open. Notice what the room smells like in the morning. It's the cheapest experiment in this whole field, and it's the one nobody is selling.
If you want the wider frame, this fits into the longevity protocol hub alongside the parts of sleep you can actually steer — light timing, regularity, and temperature.
