Pulse Oximeter Accuracy in Dark Skin: FDA Guidance and What Clinicians Should Know in 2025 and 2026

Pulse oximeter bias across skin tones is not a solved issue. On January 7, 2025, the FDA issued draft guidance proposing updated testing and labeling recommendations to improve pulse oximeter performance across the range of skin pigmentation. As of April 21, 2026, the FDA’s public pulse oximeter pages still described that document as draft guidance, not final guidance. Clinically, that means the core message has not changed: pulse oximeters remain useful, but clinicians and patients should be cautious about normal-looking readings that do not fit the patient, especially near decision thresholds and especially in people with darker skin pigmentation.

This topic deserves precision because two opposite mistakes keep happening. One is to act as though pulse oximeters are untrustworthy for everyone all the time. The other is to act as though the problem was fully addressed once the FDA issued updated recommendations. Neither is correct.

What the FDA actually said in 2025

The FDA’s January 2025 action was important, but it was a draft guidance, not an immediate performance fix.

The document proposed updated recommendations for:

• non-clinical testing
• clinical performance testing
• labeling
• premarket submission expectations

The FDA’s own public explanation connected the guidance directly to concerns that pulse oximeter accuracy can vary across skin pigmentation. It also emphasized that labeling should better communicate benefits, risks, and limitations.

That matters because the regulatory issue is not simply whether a device can pass a historical accuracy threshold under controlled conditions. It is whether manufacturers are adequately testing and describing device performance across the range of people who use the device.

What remained true on April 21, 2026

Because this is an actively evolving regulatory area, the date matters.

As of April 21, 2026:

• the FDA pulse oximeter page still listed the January 2025 document as draft guidance
• the agency continued to describe pulse oximetry as useful but limited
• the FDA still warned that readings should not be used in isolation
• the agency still highlighted skin pigmentation as one of multiple factors that can affect accuracy

That is the current practical regulatory status, based on the FDA pages available on that date. In other words, there was no public FDA signal that the issue had been "closed." The agency was still in a guidance-development and evidence-generation posture.

Why skin-tone bias matters clinically

Pulse oximeters estimate oxygen saturation noninvasively. They do not measure arterial oxygen saturation directly. The estimate can be biased, and a small bias matters most when the patient is near a treatment threshold.

The clinical danger is occult hypoxemia:

• the device reading looks acceptable
• the patient’s true arterial saturation is lower
• escalation of care or oxygen treatment may be delayed

This is the main reason the issue became so visible during the COVID-19 era, but the concern is broader than COVID. It applies to pneumonia, heart failure, COPD, postoperative care, emergency triage, and home monitoring pathways.

What the evidence showed before the 2025 FDA draft guidance

The 2020 letter by Sjoding and colleagues in the New England Journal of Medicine became a landmark because it translated a long-known technical issue into a clinical consequence. In paired pulse oximeter and arterial blood gas data, Black patients were more likely than White patients to have occult hypoxemia at the same displayed SpO2 range.

Subsequent systematic reviews did not make the concern disappear. The 2022 BMC Medicine review by Shi and colleagues concluded that pulse oximetry may overestimate oxygen saturation in people with high levels of skin pigmentation and in people described as Black or African American, while also showing that overall pooled device-level metrics can still appear acceptable in aggregate. That is exactly the regulatory and clinical tension.

Why? Because average device accuracy can look "acceptable" even when subgroup-specific error around important thresholds is still clinically meaningful.

Why this is not just a technicality

Small average bias can still produce large real-world consequences.

Threshold-based care is common

Clinicians often respond to ranges, not just trends. If a patient is hovering near a threshold for escalation, supplemental oxygen, or further evaluation, a few points of positive bias matter.

Home users do not have arterial blood gases

In the hospital, discordance can sometimes be resolved with confirmatory testing. At home, a patient may rely heavily on the device display.

The sickest patients are not the only ones affected

Bias matters in moderate hypoxemia too, because that is where the device may appear reassuring enough to delay action.

Workflow can amplify the problem

If protocols rely too rigidly on pulse ox numbers rather than symptoms and exam, device error becomes care-pathway error.

What the FDA guidance is trying to improve

Even without quoting the draft line by line, the direction is clear.

The FDA is pushing toward:

• better study design for premarket testing
• clearer evaluation across the range of skin pigmentation
• improved labeling about limitations and safe use
• more consistency in submissions and review

That is important progress, but it is still a forward-looking regulatory framework. It does not instantly upgrade older devices already in use, and it does not erase the need for clinical judgment.

What clinicians should do now

The most practical response is conservative and workflow-based.

Do not interpret the pulse ox in isolation

This is the FDA’s message and it remains the correct bedside rule.

Pay attention when the number does not fit the patient

If work of breathing, cyanosis, mental status, chest findings, or overall trajectory suggest trouble, a reassuring SpO2 should not end the evaluation.

Be cautious near action thresholds

Bias becomes most consequential when management hinges on a narrow numeric range.

Recognize subgroup vulnerability without stereotyping

The point is not that every reading in a darker-skinned patient is inaccurate. The point is that systematic overestimation risk is real enough that overconfidence is unsafe.

Know what kind of device the patient is using

FDA-cleared medical-purpose pulse oximeters and general wellness products are not the same category. Many home users do not realize the difference.

What patients should do now

Patients need advice that is specific enough to be usable.

Use the device as one piece of information

Keep track of symptoms, not just the number.

Follow the instructions carefully

Warm hand, still finger, good contact, and repeated checks matter.

Do not wait for a low number if symptoms are clearly worsening

Shortness of breath, chest discomfort, confusion, or bluish discoloration still warrant clinical attention even if the pulse ox does not look alarming.

Bring the device context to the visit

Patients should tell clinicians what type of pulse oximeter they use and how the readings were obtained.

The FDA’s consumer page makes essentially the same point: home users should not rely only on the pulse oximeter.

Why "dark skin" is not the whole story, but still the right clinical focus

Skin pigmentation is one factor among several. Perfusion, temperature, movement, nail products, tobacco exposure, skin thickness, and device quality all matter too.

But the skin-pigmentation issue deserves special attention because:

• it has been repeatedly documented
• it intersects with equity and delayed recognition of illness
• it can be hidden inside aggregate accuracy claims
• it remains under active FDA review

That combination makes it more than a routine measurement caveat.

What not to conclude from the evidence

Several overreactions are worth avoiding.

Not every normal pulse ox in a darker-skinned patient is wrong

The concern is increased risk of clinically important overestimation, not universal failure.

Not every device performs the same way

Device design, sensor quality, labeling, and testing matter.

Regulatory attention does not mean patients should abandon pulse oximetry entirely

Pulse oximeters still provide useful information. The issue is how that information is interpreted.

A future final FDA guidance will not eliminate clinical judgment

Even better devices will remain estimates rather than perfect truth machines.

Where this fits with broader wearable and PPG accuracy

Pulse oximeter bias sits inside the broader PPG accuracy conversation. Skin tone can affect more than SpO2. It can also influence signal quality and performance of other optical wearables.

For related context, see PPG skin tone bias accuracy, PPG SpO2 accuracy limitations, and wearable pulse oximeter guide.

Bottom line

The FDA took an important step on January 7, 2025 by issuing draft guidance aimed at improving pulse oximeter testing and labeling across the range of skin pigmentation. As of April 21, 2026, that guidance still appeared on FDA’s site as draft guidance, not final guidance.

For clinicians and patients, the operational message remains the same: pulse oximeters are useful, but they should not be treated as infallible, especially near clinical thresholds and especially when the reading does not fit the patient. Skin-tone bias remains a live patient-safety issue, not just a historical footnote.