Polysomnographic phenotypes and their cardiovascular implications in obstructive sleep apnoea

Area of Science:

• Sleep Medicine
• Cardiovascular Medicine
• Data Science

Background:

• Obstructive sleep apnoea (OSA) is a complex condition with varying physiological underpinnings.
• Understanding these physiological phenotypes is crucial for accurate clinical risk assessment.
• Current methods may not fully capture the heterogeneity of OSA and its cardiovascular implications.

Purpose of the Study:

• To determine if routine polysomnography can identify distinct OSA phenotypes.
• To assess the association between these identified phenotypes and cardiovascular outcomes.
• To evaluate if these phenotypes provide additional risk information beyond conventional OSA severity metrics.

Main Methods:

• Analysis of a large US Veteran cohort (n=1247) using cross-sectional and longitudinal data.
• Application of principal components-based clustering on polysomnographic features across four pathophysiological domains.
• Identification of OSA phenotypes using K-means cluster analysis and evaluation of cardiovascular event risk via Cox survival analysis.

Main Results:

• Seven distinct OSA phenotypes were identified, including 'mild', 'PLMS', 'NREM and arousal', 'REM and hypoxia', 'hypopnoea and hypoxia', 'arousal and poor sleep', and 'combined severe'.
• The 'PLMS', 'hypopnoea and hypoxia', and 'combined severe' phenotypes were significantly associated with increased risk of major adverse cardiovascular events.
• Conventional apnoea-hypopnoea index (AHI) severity categories did not show significant associations with increased cardiovascular risk.

Conclusions:

• Routine polysomnography data can effectively identify distinct physiological OSA phenotypes.
• These phenotypes are more effective than conventional AHI severity in identifying patients at high risk for adverse cardiovascular outcomes.
• Phenotype-based risk stratification may improve clinical management and outcomes for patients with OSA.