Sleep apnea and Fitbit studies so far tested how well the trackers measure sleep in people who already have the diagnosis, not whether they can find apnea. In 127 patients with an average apnea-hypopnea index of 38.3, a Fitbit Charge 2 underestimated total sleep time by 27.5 minutes against a lab sleep study. Older Fitbit models without sleep stages recognized wake poorly (specificity 0.10 to 0.52); models that also use heart rate did better (0.58 to 0.69), and none of them measures breathing. Wearable AI in research can flag apnea with a pooled sensitivity of 0.938 and specificity of 0.752, which the reviewers called suboptimal for routine clinical use. The American Academy of Sleep Medicine said in 2018 that consumer sleep technology cannot be used to diagnose sleep disorders.
People who wear a Fitbit to bed see a sleep score each morning, and some wonder whether it can tell them about breathing pauses. The studies do not show that it can. The validation studies of Fitbit in sleep apnea ask a different, smaller question: how well does the tracker measure sleep in someone who already has the diagnosis? The research answers in three parts: validation studies of Fitbit in patients, work on wearable apnea detection in general, and the position of sleep medicine.
Sleep apnea and Fitbit in patients with a confirmed diagnosis
One validation study recruited 65 adults who presented with symptoms of sleep apnea at an outpatient sleep clinic. Each wore a Fitbit Charge 2 or Alta HR during a full sleep study, so the tracker and the lab could be compared on the same night. Sleep apnea was confirmed in 55 of them (84.6%). The trackers overestimated total sleep time and underestimated both time awake after falling asleep and time to fall asleep. Their sensitivity for sleep was acceptable and their specificity was poor, and the authors concluded that consumer trackers do not yet have enough accuracy for clinical settings, particularly in clinical populations.
Moreno-Pino F, Porras-Segovia A, López-Esteban P, Artés A, Baca-García E. Validation of Fitbit Charge 2 and Fitbit Alta HR Against Polysomnography for Assessing Sleep in Adults With Obstructive Sleep Apnea. J Clin Sleep Med 2019. PMID 31739855
Another study took 127 consecutive patients with sleep apnea (average age 50, average apnea-hypopnea index 38.3 events per hour) and put a Fitbit Charge 2 and a Galaxy Watch 2 on their wrist during the sleep study. Both devices underestimated total sleep time, by 27.5 minutes for the Fitbit and 24.9 minutes for the Galaxy Watch, and the size of the error did not follow the severity of the apnea. Recording failed on 15.7% of nights with the Galaxy Watch and 8.7% with the Fitbit, a difference that was not statistically significant.
Byun JI, Noh KC, Shin WC. Performance of the Fitbit Charge 2 and Galaxy Watch 2 compared with polysomnography in assessing patients with obstructive sleep apnoea. Chronobiol Int 2023. PMID 36971253
The two validation studies disagree on the direction of the error, over in one and under in the other. That is a reminder of how much the result depends on the device, the setting, the way the tracker decides when sleep begins and the people tested. They agree on the bigger point, which is that the question being asked was how long people slept. Neither asked the tracker to count breathing pauses.
Why wake is the hard part
A systematic review and meta-analysis of wristband Fitbit models explains where the errors come from. It included 22 articles, 8 of them with data that could be pooled. Compared with lab sleep studies, models that did not stage sleep tended to overestimate total sleep time and sleep efficiency and underestimate time awake after sleep onset. They labelled sleep epochs correctly with accuracy between 0.81 and 0.91 and sensitivity between 0.87 and 0.99, but specificity, the ability to recognize wake, was only 0.10 to 0.52. One reading of that: a still wrist and a calm pulse are easily taken for sleep, so quiet wakefulness can get counted as sleep. Recent models that combine heart rate variability and movement to stage sleep did better: specificity 0.58 to 0.69, and no significant difference from the sleep lab for total sleep time, time awake after falling asleep or sleep efficiency, although they underestimated the time to fall asleep. The authors called this promising, especially for telling wake from sleep, and also limited and no substitute for a sleep study.
Haghayegh S, Khoshnevis S, Smolensky MH, Diller KR, Castriotta RJ. Accuracy of Wristband Fitbit Models in Assessing Sleep: Systematic Review and Meta-Analysis. J Med Internet Res 2019. PMID 31778122
This matters for sleep apnea in a specific way. Many breathing pauses end in brief arousals, and a person with apnea can have many short awakenings that never become a remembered awake period. The meta-analysis scored sleep and wake, and it did not test whether the trackers see arousals of that kind, so this stays a reason for caution and not a proven fault.
What wearable AI can do for apnea in research
Separate from Fitbit's own sleep scores, there is a research field that applies machine learning to wearable signals to detect apnea. A 2024 systematic review and meta-analysis found 38 eligible studies. The pooled accuracy for detecting sleep apnea was 0.869, with sensitivity of 0.938 and specificity of 0.752. Classifying severity (normal, mild, moderate, severe) had a pooled accuracy of only 0.651. The reviewers concluded that wearable AI shows potential but is suboptimal for routine clinical use, and they recommended using it alongside traditional assessments until the evidence improves. They also called for certified commercial wearables, which suggests that much of the studied technology is not yet an off-the-shelf product.
Abd-Alrazaq A, Aslam H, AlSaad R, Alsahli M, Ahmed A, Damseh R, Aziz S, Sheikh J. Detection of Sleep Apnea Using Wearable AI: Systematic Review and Meta-Analysis. J Med Internet Res 2024. PMID 39255014
One example shows why the idea is not far-fetched. In 41 patients having a sleep study, researchers recorded the pulse with a wrist watch device and ran an algorithm that looks for the cyclical speeding and slowing of the heart that goes with apnea episodes. The hourly frequency of that pattern correlated with the apnea-hypopnea index (r=0.81) and separated people with an index of 15 or more from the rest with 82% sensitivity, 89% specificity and 85% accuracy. It was a research algorithm tested in a small group, not a consumer feature.
Hayano J, Yamamoto H, Nonaka I, Komazawa M, Itao K, Ueda N, Tanaka H, Yuda E. Quantitative detection of sleep apnea with wearable watch device. PLoS One 2020. PMID 33166293
These studies tested the Charge 2 and Alta HR for sleep measurement. They say nothing about newer models or about features a manufacturer has added since, so any claim about Fitbit sleep apnea detection or oxygen on a given device should be checked against what the maker states and whether a regulator has cleared it for that purpose.
What the sleep medicine societies say
The American Academy of Sleep Medicine put the issue plainly in a 2018 position statement on consumer sleep technology. Because of the lack of validation against polysomnography and of clearance by the US Food and Drug Administration, consumer sleep technologies cannot be used for the diagnosis or treatment of sleep disorders. The statement adds that the data can still be useful in a clinical conversation, as long as it is presented as part of a proper clinical evaluation. It also asked for future validation and for access to raw data and algorithms.
Khosla S, Deak MC, Gault D, Goldstein CA, Hwang D, Kwon Y, O'Hearn D, Schutte-Rodin S, Yurcheshen M, Rosen IM, Kirsch DB, Chervin RD, Carden KA, Ramar K, Aurora RN, Kristo DA, Malhotra RK, Martin JL, Olson EJ, Rosen CL, Rowley JA, American Academy of Sleep Medicine Board of Directors. Consumer Sleep Technology: An American Academy of Sleep Medicine Position Statement. J Clin Sleep Med 2018. PMID 29734997
The statement is eight years old, and eight years is long in wearable technology, so read it as a baseline: the argument is about validation, and any newer validation study deserves to be weighed on its own. For comparison, the Apple Watch's breathing disturbance feature is covered in How accurate is Apple Watch sleep apnea detection?, and Oximetry and sleep apnea covers the oxygen measurement that many wearables lean on.
How to use tracker data if you suspect apnea
Start with the symptoms, not the score. The signs doctors ask about are loud snoring, gasping or choking at night, pauses someone has seen, and unrefreshing sleep with sleepiness in the day. A tracker can add context: how many nights are broken, how long the night really is, whether a bad stretch lines up with a cold or a late meal. Write those down, bring the data as supporting detail, and ask about a sleep test. Best sleep tracking apps compares what different apps record.
SleepTrace listens through the phone's microphone and estimates sleep phases from sound; it can also be paired with an Apple Watch. It recognizes snoring, talking and coughing and keeps short clips for those (the free plan plays back 3 clips a month, Premium has no limit), and it shows a sleep apnea indicator as a screening, not a diagnosis. Keep the app open overnight with the phone charging; the screen dims on its own. It is not a medical device, and a clip of loud snoring with gaps is something to bring to an appointment, not a conclusion.
Frequently asked questions
The studies do not show that. Two validation studies tested Fitbit trackers on people with sleep apnea, and both compared sleep time and sleep stages with a lab sleep study, not breathing pauses. The Fitbit models tested were the Charge 2 and the Alta HR. A wearable AI meta-analysis of 38 studies found promising but not clinically ready performance for apnea detection across many different devices. If you suspect apnea, a sleep test ordered by a doctor is the way to find out.
In 127 patients with sleep apnea (average apnea-hypopnea index 38.3), the Fitbit Charge 2 underestimated total sleep time by 27.5 minutes compared with a sleep study, and the error did not depend on how severe the apnea was. In a separate study of 65 adults, Fitbit Charge 2 and Alta HR overestimated total sleep time and underestimated time awake after falling asleep, with acceptable sensitivity and poor specificity.
Trackers estimate sleep from movement and, in newer models, from heart rate variability. In a meta-analysis of Fitbit models, older models that did not stage sleep identified sleep periods with accuracy between 0.81 and 0.91 but recognized wake with a specificity of only 0.10 to 0.52. Models that also use heart rate variability did better (specificity 0.58 to 0.69), and none of them measures breathing.
They measure different things. A wrist tracker estimates sleep and wake from movement and pulse. A microphone hears snoring and the silences around it. Neither is a diagnosis. The American Academy of Sleep Medicine position statement says consumer sleep technology cannot be used for diagnosing or treating sleep disorders, and both approaches fall under that.
It can help the conversation. The position statement from the American Academy of Sleep Medicine says consumer data may be used to enhance the patient-clinician interaction in the context of a proper clinical evaluation. Describe what you notice (snoring, gasping, sleepiness), and show the data as supporting detail rather than as a finding.
SleepTrace is a wellness app, not a medical device. This article is general information, not medical advice. If your symptoms are frequent, severe or worrying, please talk to a doctor.
Trackers on the wrist measure sleep and wake, not breathing. SleepTrace is built on sound: it saves clips of snoring, marks possible breathing pauses with a time and estimates sleep phases. It does not diagnose sleep apnea, so use it to decide whether to ask for a sleep study, and use the Fitbit for what it does well.
Hear your own night. SleepTrace turns a night of audio into your sleep phases, the sounds you made, and how it all trends. Works without a wearable, better with one. Your phone on the nightstand is enough to start. Get it for iPhone or Android →