Field note
Blood oxygen on a watch is a wrist estimate of how much oxygen your blood is carrying, usually taken while you sleep. For most healthy people that number sits high and steady. It is still a rough check, not a hospital finger clip.
Your wrist is not a clinic. The useful move is to read your own multi-night trend, not to panic at one low spot.
The sensor never touches your blood. It shines several colors of light, including red and infrared, into your wrist and reads how much of each gets absorbed. Oxygen-rich blood and oxygen-poor blood soak up those colors differently. From that difference the watch makes an estimate. Jubran (2015) reviewed pulse oximetry as an indirect optical method: clever, widely used, and still an estimate rather than a lab blood-gas value.
The watch is guessing well from light bouncing through skin and tissue. It is not measuring oxygen molecules directly.
| What you see | Honest read |
|---|---|
| One low night | Often cold wrist, loose band, or motion |
| Steady high nights | Common; the expected quiet baseline for many people |
| A shift that holds across nights | Worth a glance against your own usual |
| Any single wrist value as a medical call | Not what this sensor is for |
Reading from the wrist is harder than reading from a fingertip. The pulse of blood the sensor needs is much fainter at the wrist than at the finger. So a wrist read starts out rougher than a clinic clip before anything else gets in the way.
Cold hands genuinely degrade the reading. When you are cold, the small vessels near your skin tighten, and that shrinks the signal the sensor depends on. Movement does the same. If your arm shifts while it reads, the number can wander. One low spot value, caught on a cold or restless night, tells you very little.
Skin tone, how warm your hands are, and how still you lie all nudge the number. Sjoding and colleagues (2020) documented racial bias in pulse oximetry measurement: the same optical method can quietly fail different people differently. That is a fairness and accuracy limit of the technique, not a reason to treat one wrist night as a verdict. Under perfect conditions, still and warm, a good wrist sensor can land in the right ballpark. Perfect conditions are rare at 3am.
It pairs with What is overnight breathing rate? and What is sleeping heart rate?: all three are quieter overnight floors best read against your own nights, not against a stranger's chart.
Because each reading wobbles, the single number is not the signal. The signal is your trend against your own baseline, gathered over many nights. Once you know where you usually sit, a real shift stands out from the nightly noise. A one-off dip almost never means your oxygen changed. It usually means a cold wrist, a loose band, or a moved arm.
A cold wrist, a loose band, a shifted arm, and a real change can all produce the same low number. The watch cannot tell them apart. That is exactly why the useful read leans on your trend, not any one value. See also What does recovery actually measure? for how overnight heart and breath sit together against your usual.
A watch SpO2 number is a wrist-based estimate, not a diagnosis. It cannot reliably catch a true low-oxygen event, and it cannot rule one out. It should never be the thing that decides whether you need care.
If you have breathing symptoms that stick around, that is a conversation for a doctor and a real medical oximeter. Not the watch on your wrist. The overnight chart is a quiet background check so you can notice patterns over time. That is all it is meant to be.
Trust the trend, not the spot. And for anything that worries you, trust a clinician over a wrist.
I write these because I build Aera, an iPhone app that reads Apple Health against your own baseline and puts the research and the error bars underneath every number it shows you, including this one. You do not need it to use anything above.