Withings funded the study and paid clinicians for their time. The first author is a Withings employee, and the corresponding author is its medical consultant. The sleep study check covered 14 people for one night. One CPAP model only, the ResMed AirSense 10. The authors say the result may not hold for other brands. Nobody saw the sensor data during the study, so it doesn't show whether having it changes anything.

Bed Sensor vs CPAP Data: A 2026 Study

Q: What did the two devices report?

Breathing events per hour: sensor 11.3, CPAP machine 2.0. Time: the sensor logged 437.4 minutes of sleep a night. The machine logged 395.6 minutes of use. Night to night: the sensor's count moved by 5.4 an hour within the same person. The machine's moved by 1.1 to 1.2 (the paper prints both). The authors give two reasons for the 42-minute gap. Some sleep happened with the mask off, and earlier studies found the sensor overestimates sleep by 20 to 26 minutes.

A 2026 study put an under-mattress sensor beside CPAP machines for 448 nights. The sensor counted 11.3 breathing events an hour, the machines 2.0. The sensor's maker funded it.

PulseOxTracker Team9 October 2026

Infographic. Bed sensor vs CPAP machine, breathing events per hour on the same nights: under-mattress sensor 11.3, CPAP machine 2.0. Who was studied: 75 adults on CPAP at home, 448 nights, in the Paris region, on CPAP for at least a month, and 14 had a sleep study. What the study found. Sleep time and mask time: sensor sleep 437.4 minutes a night, CPAP use 395.6 minutes. Night to night: the sensor count varied by 5.4 an hour, the CPAP count by 1.1 to 1.2. Sleep study check on 14 people: sleep study 10.9, sensor 8.8, CPAP 2.9 events an hour. Sensor against the sleep study: 2.1 an hour lower on average, with a wide spread, standard deviation 8.3. Mask on while awake: that night CPAP use was 483.9 minutes and sleep on the sleep study 408.6 minutes. What it cannot show: Withings makes the sensor and funded the study, the first author works there and another is its medical consultant, one CPAP model only, all patients used CPAP a lot, and and earlier studies found the sensor overestimates sleep time. — A summary of the 2026 study: what an under-mattress sensor and a CPAP machine each reported on the same nights, and who paid for the study.

Your CPAP machine reports how many breathing events it still sees each hour. A 2026 study put a second device in the same bed and got a higher count. The second device's maker paid for the study.

Who was in the study?

75 adults with obstructive sleep apnea from six sleep centres around Paris. 115 enrolled and 40 were excluded for protocol deviations.
All had been on CPAP for at least a month and used it a lot.
Each had a Withings Sleep Analyzer, a pressure pad under the mattress, for about a week.
448 nights were analysed, all with over 5 hours of sensor-logged sleep.

What did the two devices report?

Breathing events per hour: sensor 11.3, CPAP machine 2.0.
Time: the sensor logged 437.4 minutes of sleep a night. The machine logged 395.6 minutes of use.
Night to night: the sensor's count moved by 5.4 an hour within the same person. The machine's moved by 1.1 to 1.2 (the paper prints both).

The authors give two reasons for the 42-minute gap. Some sleep happened with the mask off, and earlier studies found the sensor overestimates sleep by 20 to 26 minutes.

Which one was closer to a sleep study?

14 patients had a one-night home sleep study with both devices running.

Sleep study: 10.9 events an hour.
Sensor: 8.8.
CPAP machine: 2.9.

The sensor was closer on average, but its count was still statistically lower than the sleep study's. The differences were spread wide (standard deviation 8.3 an hour).

That night the patients kept their masks on until a technician arrived. So CPAP use (483.9 minutes) was longer than sleep on the sleep study (408.6 minutes). The authors say time awake with the mask on can make the machine's hourly count look lower.

What can't it show?

Withings funded the study and paid clinicians for their time.
The first author is a Withings employee, and the corresponding author is its medical consultant.
The sleep study check covered 14 people for one night.
One CPAP model only, the ResMed AirSense 10. The authors say the result may not hold for other brands.
Nobody saw the sensor data during the study, so it doesn't show whether having it changes anything.

What can you do with it?

A CPAP report covers only the time the mask is on. For a record of the whole night, you can track your oxygen alongside CPAP. Take questions about your numbers to your sleep clinic.

Not a medical device. Does not diagnose, treat or prevent any condition. Discuss results and changes with a clinician.

References

Yang RY, Seailles T, Aisenberg N, Felix O, Sabil A, Escourrou P. Comparison of under-mattress sensor and CPAP data for home monitoring in obstructive sleep apnea. Sleep and Breathing. 2026;30(4):194. doi:10.1007/s11325-026-03741-9