Short-Term HRV Analysis Using Nonparametric Sample Entropy for Obstructive Sleep Apnea

Duan Liang; Shan Wu; Lan Tang; Kaicheng Feng; Guanzheng Liu

doi:10.3390/e23030267

Short-Term HRV Analysis Using Nonparametric Sample Entropy for Obstructive Sleep Apnea

Entropy (Basel). 2021 Feb 24;23(3):267. doi: 10.3390/e23030267.

Authors

Duan Liang^{1

2

3}, Shan Wu^{1

2

3}, Lan Tang^{1

2

3}, Kaicheng Feng^{1

2

3}, Guanzheng Liu^{1

2

3}

Affiliations

¹ School of Biomedical Engineering, Sun Yat-Sen University, Guangzhou 510275, China.
² Key Laboratory of Sensing Technology and Biomedical Instruments of Guangdong Province, School of Engineering, Sun Yat-Sen University, Guangzhou 510275, China.
³ Guangdong Provincial Engineering and Technology Centre of Advanced and Portable Medical Device, Guangzhou 510006, China.

Abstract

Obstructive sleep apnea (OSA) is associated with reduced heart rate variability (HRV) and autonomic nervous system dysfunction. Sample entropy (SampEn) is commonly used for regularity analysis. However, it has limitations in processing short-term segments of HRV signals due to the extreme dependence of its functional parameters. We used the nonparametric sample entropy (NPSampEn) as a novel index for short-term HRV analysis in the case of OSA. The manuscript included 60 6-h electrocardiogram recordings (20 healthy, 14 mild-moderate OSA, and 26 severe OSA) from the PhysioNet database. The NPSampEn value was compared with the SampEn value and frequency domain indices. The empirical results showed that NPSampEn could better differentiate the three groups (p < 0.01) than the ratio of low frequency power to high frequency power (LF/HF) and SampEn. Moreover, NPSampEn (83.3%) approached a higher OSA screening accuracy than the LF/HF (73.3%) and SampEn (68.3%). Compared with SampEn (|r| = 0.602, p < 0.05), NPSampEn (|r| = 0.756, p < 0.05) had a significantly stronger association with the apnea-hypopnea index (AHI). Hence, NPSampEn can fully overcome the influence of individual differences that are prevalent in biomedical signal processing, and might be useful in processing short-term segments of HRV signal.

Keywords: heart rate variability (HRV); nonparametric sample entropy (NPSampEn); obstructive sleep apnea (OSA); short-term HRV analysis.