Relationship between Clinical and Polysomnography Measures Corrected for CPAP Use

Abstract

Objectives: The changes in patient-reported measures of obstructive sleep apnea (OSA) burden are largely discordant with the change in apnea-hypopnea index (AHI) and other polysomnography measures before and after treatment. For patients treated with continuous positive airway pressure (CPAP), some investigators have theorized that this discordance is due in part to the variability in CPAP use. We aim to test the hypothesis that patient-reported outcomes of CPAP treatment have stronger correlations with AHI when it is corrected for mean nightly CPAP use.

Methods: This was a cross-sectional study of 459 adults treated with CPAP for OSA. Five patient-reported measures of OSA burden were collected at baseline and after 6 months of CPAP therapy. The correlations between the change in each patient-reported measure and the change in AHI as well as mean nightly AHI (corrected for CPAP use with a weighted average formula) were measured after 6 months of treatment. The same analysis was repeated for 4 additional polysomnography measures, including apnea index, arousal index, lowest oxyhemoglobin saturation, and desaturation index.

Results: The change in AHI was weakly but significantly correlated with change in 2 of the 5 clinical measures. The change in mean nightly AHI demonstrated statistically significant correlations with 4 out of 5 clinical measures, though each with coefficients less than 0.3. Similar results were seen for apnea index, arousal index, lowest oxyhemoglobin saturation, and desaturation index.

Conclusions: Correction for CPAP use yielded overall small but significant improvements in the correlations between patient-reported measures of sleep apnea burden and polysomnography measures after 6 months of treatment.

Keywords: CPAP; apnea-hypopnea index; outcome; quality of life; sleep apnea; sleepiness; symptom.