Continuous positive airway pressure device-based automated detection of obstructive sleep apnea compared to standard laboratory polysomnography

Sleep Breath. 2010 Jun;14(2):101-7. doi: 10.1007/s11325-009-0285-z. Epub 2009 Oct 14.


Purpose: Obstructive sleep apnea (OSA) is a common health problem that affects more than 2-4% of the US population. Polysomnography (PSG) is the gold standard for diagnosing OSA. PSG is, however, expensive, time-consuming, and not always readily accessible. Hence, alternative diagnostic methods such as home-based testing have been evaluated. We studied the ability of the REMstar Pro (RSP2, a brand of continuous positive airway pressure (CPAP) device) to identify abnormal breathing events in subjects with OSA and compared this with breathing events simultaneously determined by laboratory-based PSG.

Methods: We evaluated 10 subjects previously diagnosed with OSA (apnea hypopnea index (AHI) > 15, known therapeutic level of CPAP). Subjects underwent attended PSG using the REMstar Pro M series machine and their prescribed interface/mask type. The first 3 h of the study were conducted using a subtherapeutic CPAP (4 cm H2O). The last 3 h or remaining portion of the PSG was completed using the previously determined therapeutic CPAP. Comparison of respiratory events detected by PSG vs the RSP2 was performed.

Results: Subjects included four men and six women, aged 32 to 57 years and with a body mass index ranging from 29.5-66.4. The baseline AHI ranged from 18.3-93.1, with the AHI at therapeutic CPAP ranging from 0-3. Apnea counts at baseline and at therapeutic CPAP by manually scored PSG and REMstar were not significantly different (mean at subtherapeutic 11.7 vs 12.5, p = 0.76; median at therapeutic CPAP 2.0 vs 4.5, p = 0.15). Hypopnea counts at baseline and at effective CPAP by PSG and REMstar were not significantly different (mean at subtherapeutic 38.1 vs. 40.9, p = 0.72; median at therapeutic CPAP 5.0 vs. 2.5, p = 0.34). The correlation coefficient of REMstar and PSG for apnea and hypopnea was significant in subtherapeutic phase only (apnea r = 0.78, p = 0.007; hypopnea r = 0.76, p = 0.01). Agreement between the two methods declined for hypopnea detection at therapeutic CPAP.

Conclusions: The monitoring of residual sleep-disordered breathing on treatment, in addition to adherence, is an important objective therapeutic target in OSA. The REMstar Pro detects sleep-disordered breathing events similar to that of a manually scored PSG-for apnea but not for hypopnea-and merits further investigation as a device to determine disease severity and treatment efficacy.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Algorithms
  • Body Mass Index
  • Continuous Positive Airway Pressure / instrumentation*
  • Cross-Over Studies
  • Diagnosis, Computer-Assisted / instrumentation*
  • Equipment Design
  • Female
  • Humans
  • Male
  • Middle Aged
  • Monitoring, Ambulatory / instrumentation*
  • Pilot Projects
  • Polysomnography / instrumentation*
  • Predictive Value of Tests
  • Prospective Studies
  • Recurrence
  • Signal Processing, Computer-Assisted / instrumentation*
  • Sleep Apnea, Obstructive / diagnosis*
  • Sleep Apnea, Obstructive / therapy*