OBJECTIVE: To evaluate the ability of the heart rate power spectral density of the very low frequency component expressed as a percentage of the total power spectral band (%VLFI) to diagnose obstructive sleep apnoea (OSA) in the elderly. PATIENTS AND METHODS: Eight hundred and thirty-two subjects aged 68.5±0.8 years were selected from the participants of a 7-year cross-sectional study on the association between OSA and vascular morbidity-mortality. All subjects underwent electrocardiogram Holter monitoring and an at-home polysomnogram to measure the apnoea/hypopnoea index (AHI). OSA was assessed as an AHI>15. RESULTS: The very low, the low frequency powers, and the %VLFI increased in subjects with OSA and were correlated with the autonomic arousal index and AHI (p<0.0001). A %VLFI of 2.4% and 4% demonstrated a sensitivity of 62% and 69% and a specificity of 45% and 44% to predict OSA in the population. CONCLUSION: The %VLFI did not provide a high level of diagnostic sensitivity and specificity at different degrees of OSA severity in an elderly healthy population. The ability for prediction was lower in mild cases and in women. These findings suggest that the ageing-related changes in the autonomic nervous system reduce the applicability of HRV to diagnose OSA in older populations.
OBJECTIVE: To evaluate the ability of the heart rate power spectral density of the very low frequency component expressed as a percentage of the total power spectral band (%VLFI) to diagnose obstructive sleep apnoea (OSA) in the elderly. PATIENTS AND METHODS: Eight hundred and thirty-two subjects aged 68.5±0.8 years were selected from the participants of a 7-year cross-sectional study on the association between OSA and vascular morbidity-mortality. All subjects underwent electrocardiogram Holter monitoring and an at-home polysomnogram to measure the apnoea/hypopnoea index (AHI). OSA was assessed as an AHI>15. RESULTS: The very low, the low frequency powers, and the %VLFI increased in subjects with OSA and were correlated with the autonomic arousal index and AHI (p<0.0001). A %VLFI of 2.4% and 4% demonstrated a sensitivity of 62% and 69% and a specificity of 45% and 44% to predict OSA in the population. CONCLUSION: The %VLFI did not provide a high level of diagnostic sensitivity and specificity at different degrees of OSA severity in an elderly healthy population. The ability for prediction was lower in mild cases and in women. These findings suggest that the ageing-related changes in the autonomic nervous system reduce the applicability of HRV to diagnose OSA in older populations.