Oxygen Saturation Readings: Meaning and Limits

Urgent first

Severe trouble breathing or any breathing emergency needs emergency care, whatever a reading says. In the United States call 911. The FDA advises tracking symptoms as well as the number, and lists shortness of breath, difficulty breathing or a cough that gets worse among the symptoms to watch. [1] It also says some people with low oxygen may not show any of the symptoms. [1]

The short version

A pulse oximeter clips on a finger (or sits on a wrist or ring) and shines light through or onto the skin to estimate the percentage of blood carrying oxygen. That estimate is called SpO2. It is useful in the right setting and easy to over-read at home. Evidence labels used here: established, mixed, insufficient evidence, marketing claim.

What is measured and what is calculated

• Direct: the device works with light and does not take a blood sample. [1]
• Estimated: the FDA says a pulse oximeter "uses light beams to estimate" oxygen saturation and pulse rate. [1] The figure on screen is therefore a calculation from light, usually labeled SpO2 and shown as a percentage. [1]
• Reference: a blood sample from an artery, whose result is called SaO2. Studies compare SpO2 with it. [2]
• Not calculated here: the "wellness" extras (sleep-time oxygen charts, "breathing disturbance" scores) are covered in consumer sleep-apnea features.

What is "normal"?

The FDA says values are between 95% and 100% for most healthy individuals, but can be lower in people with lung and heart problems, and are generally slightly lower at higher altitudes. [1] There is no single cut-off that suits everyone, and a number alone does not say why it is low.

Conditions that change the reading

The FDA lists poor circulation, skin pigmentation, skin thickness, skin temperature, current tobacco use and fingernail polish as factors that can affect accuracy. [1] For home use it advises a warm, relaxed hand held below heart level, removing nail polish, sitting still and waiting for a steady number. [1] Movement matters: in a test of four wearable pulse oximeters in 33 people, tasks such as tapping, rubbing and turning pages made at least one device exceed 4% root-mean-square error, although all could detect low oxygen. [2]

Skin pigmentation

The FDA says current evidence suggests accuracy differences between people with lighter and darker skin pigmentation. [1] In 30,039 paired readings from Veterans Health Administration inpatients, the chance that a reading of 92% or higher hid a true arterial saturation below 88% ("occult hypoxemia") was 15.6% in white patients and 19.6% in Black patients. [3] That result comes from hospital patients with medical-grade devices, not from consumer wearables. In January 2025 the FDA proposed draft recommendations to improve testing across skin tones; the draft applies to oximeters for medical purposes and does not apply to general wellness products. [4] A draft is not a final rule.

Medical versus wellness oximeters

The FDA explains that oximeters intended for medical purposes are cleared by the agency, while those sold as general wellness or sporting products are not reviewed or evaluated by it before sale. [1] It says wellness products are not evaluated for clinical decisions or for deciding whether to seek medical care. [1] So a watch's oxygen feature is not automatically a medical test. Check the maker's stated intended use and the jurisdiction. See also how validation works.

Wearable validation: one device, one study

A study of 97 adults with sleep disturbances compared one wrist-worn watch model (Galaxy Watch 4) with polysomnography oximetry. It reported an overall RMSE of 2.3% and negligible bias of -0.2%. [5] But 26.5% of data were rejected. [5] A device-derived oxygen desaturation index of 5 or more per hour predicted an apnea-hypopnea index of 15 or more with sensitivity 89.7% and specificity 64.1%. [5] With specificity of 64.1%, about 36% of people below that apnea level would still be flagged by that cut-off. Do not extend this to other models or versions, and it does not show that tracking improves health. See wearable validation in different populations.

Appropriate uses, and what a reading cannot do

• Reasonable: a spot check a clinician asked you to do, written down with the time and your symptoms. [1]
• Cannot do: diagnose asthma, lung disease, heart disease or sleep apnea, or rule them out. The FDA says only a health care provider can diagnose a condition such as hypoxia. [1]
• Not for chasing numbers: oxygen "boosting" practices are discussed in oxygen-based wellness claims.

When professional assessment may be appropriate

The FDA says to contact a provider if you are concerned about a reading or your symptoms are serious or getting worse. [1] Concerns about breathing during sleep belong with a clinician; see sleep apnea.