0954 CORRELATION OF PULSE RATE VARIABILITY (PRV) AND HEART RATE VARIABILITY (HRV) METRICS DURING SLEEP IN SUBJECTS SUSPECTED OF OSA

Abstract

Abstract Introduction Whether photoplethysmography (PPG)-derived pulse rate variability (PRV) metrics can be surrogates for ECG-derived heart rate variability (HRV) is a decade-long debate. Very few studies have assessed the performance of PRV metrics in wearables. We aimed to evaluate the performance of time- and frequency-domain PRV metrics derived from the Belun Ring (BR) in subjects suspected of sleep apnea. Methods Using Pearson’s correlation coefficient (PCC), we retrospectively analyzed the relationship between concurrent BR-PRV and PSG ECG-HRV metrics (standard deviation of NN intervals [SDNN], root mean square of successive RR interval differences [RMSSD], percentage of successive RR intervals that differ by more than 50 ms [PNN50], HF [high frequency] power, LF [low frequency] power, VLF [very low frequency] power, and LF/HF ratio) in 73 subjects suspected of OSA. Subgroup analyses were performed based on demographic information, OSA, co-morbidities, and HR-affecting medication status. Results Age (SD) 50.1 (13.4); female 68%; black 52%; BMI (SD) 36.1 (9.7); 56% OSA; mean AHI (SD) 14.7 (17.3). Nine (12%) subjects had significant arrhythmias (defined as having any arrhythmias in ≥25% of all 12-second segments.) Key findings: There was a unitary linear relationship between BR PPG-pulse rate and ECG-heart rate in the overall group. In those without significant arrhythmias (N=64), PRV-SDNN, PNN50, HF, LF, and VLF had excellent correlations (PCC ≥0.90) with corresponding HRV metrics in all subgroups. PRV-RMSSD and PRV-LF/HF also had a strong correlation (PCC ≥0.80) with ECG HRV-RMSSD and HRV-LF/HF in all subgroups except for LF/HF in the diabetes group. In those with significant arrhythmias (N=9), PRV performed less well, with most metrics having PCC < 0.90 or wide CI, except for PNN50. All PRV metrics had PCCs ≥0.90 in subjects on HR-affecting medications except for LF/HF. Conclusion BR-PRV metrics can serve as surrogates for ECG-HRV in patients suspected of OSA who do not have significant arrhythmias. PRV-RMSSD and LF/HF may be less robust than the rest of the PRV metrics but still showed strong correlations in most subgroups. BR-PRV can be a useful clinical and research tool in assessing autonomic nervous system function. Support (if any) This study is supported by a Grant from Belun Technology Company