Pulse oximeters are generally accurate within about 2 to 4 percentage points of your true blood oxygen level. If your device reads 95%, your actual saturation could be anywhere from 91% to 99%. That margin matters more than most people realize, especially at lower readings. The National Institutes of Health notes that a reading of 90% means your true oxygen level could be as low as 86% or as high as 94%, a range that spans from “get to a hospital” to “perfectly fine.”
Medical-Grade vs. Consumer Devices
Not all pulse oximeters are held to the same standard. Devices intended for medical use go through FDA review, receive clearance, and must undergo clinical testing that confirms their accuracy across a range of oxygen levels. That testing includes desaturation studies with at least 200 paired data points comparing the device’s reading against arterial blood samples, the gold standard for measuring oxygen.
Consumer pulse oximeters, the kind you buy at a pharmacy or online, are a different category entirely. The FDA classifies most of these as “general wellness” or sporting and aviation products. They do not undergo FDA premarket review, and to date, the FDA has not cleared any over-the-counter pulse oximeter for medical purposes. These devices can give you a rough estimate of your oxygen saturation, but they haven’t been tested with the same rigor as the clip-on monitors used in hospitals. If you’re relying on one to track a health condition, that distinction matters.
What Throws Off a Reading
Pulse oximeters work by shining light through your fingertip and measuring how much is absorbed by oxygenated versus deoxygenated blood. Anything that interferes with that light signal or the blood flow underneath it can skew the result.
Cold hands are one of the most common culprits. Cold fingers reduce blood flow to the skin, which degrades the signal the sensor depends on. Research shows that keeping the sensor site near 33°C (about 91°F) produces the most reliable readings. If your hands are cold, warming them up or rubbing them together before clipping on the device can help. Poor circulation from low blood pressure, dehydration, or conditions that constrict blood vessels causes the same problem.
Nail polish, particularly darker colors, can also interfere. Black and blue polishes have been shown to lower readings by 2 to 6 percentage points. One study found blue nail polish dropped a reading from 97% to 87%, a clinically significant error. Brown and black polishes consistently lowered readings by about 2 percentage points. If you’re monitoring your oxygen regularly, bare nails on at least one finger give you the cleanest signal.
Movement is another major source of error. Shivering, trembling, or even fidgeting during a reading can create signal noise that the device misinterprets. FDA testing guidelines for medical-grade devices specifically include motion performance testing, but consumer devices may not handle movement well at all. For the best reading, sit still with your hand resting at heart level and wait for the number to stabilize for at least 10 to 15 seconds.
Skin Tone Affects Accuracy
This is one of the most important limitations of pulse oximetry, and one that has received growing attention since the COVID-19 pandemic. Pulse oximeters tend to overestimate oxygen levels in people with darker skin, sometimes by a meaningful amount.
A large study published in The BMJ found that readings ran 0.6 to 1.5 percentage points higher in patients with darker skin compared to those with lighter skin. That gap sounds small, but its clinical consequences are not. The study measured how often a pulse oximeter showed a normal reading (above 94%) when the patient’s true oxygen level was actually dangerously low (at or below 92%). For patients with darker skin tones, this false reassurance happened at rates 5.3 to 35.3 percentage points higher than for lighter-skinned patients. In the worst cases, more than 60% of readings in darker-skinned patients missed a truly low oxygen level, compared to under 27% in lighter-skinned patients.
The FDA has acknowledged this problem. In 2021, it issued a safety communication warning that skin pigmentation can affect accuracy. In February 2024, an FDA advisory committee discussed the issue specifically, and in January 2025, the agency released draft guidance that would update how pulse oximeters are tested and labeled before they reach the market. The updated recommendations aim to ensure devices are evaluated across a broader range of skin tones.
Carbon Monoxide Creates False Readings
Pulse oximeters cannot distinguish between oxygen attached to your red blood cells and carbon monoxide attached to them. Both absorb light in a similar way, so the device counts carbon monoxide as if it were oxygen. In cases of carbon monoxide poisoning, the oximeter overestimates your true oxygen level by roughly the same percentage as the carbon monoxide level in your blood. Someone with dangerously high carbon monoxide exposure could see a reading of 98% on the oximeter while their actual oxygen saturation is far lower. The average gap in one study was 17.5 percentage points. This is why emergency departments use blood draws rather than finger clips to assess suspected carbon monoxide poisoning.
When Readings Are Most Reliable
Pulse oximeters perform best when your oxygen saturation is in the normal range, roughly 94% to 100%. Clinical studies show a strong correlation between pulse oximeter readings and actual arterial blood oxygen in patients who are not hypoxemic, with a correlation coefficient of 0.95. In patients with low oxygen levels, the correlation drops to about 0.87. In practical terms, the device is more trustworthy when your levels are normal and less trustworthy precisely when accuracy matters most.
To get the most reliable reading from a home device, use it on a warm finger with no nail polish, sit quietly, and let the reading stabilize before recording it. Track your numbers over time rather than reacting to any single reading. A consistent downward trend is more informative than one low number. And if your reading is below 95% or dropping, especially if you also feel short of breath, lightheaded, or confused, that combination of symptoms and a low reading is more meaningful than either one alone.