High altitude affects nocturnal non-linear heart rate variability: PATCH-HA study.

Abstract

© 2018 Boos, Bye, Sevier, Bakker-Dyos, Woods, Sullivan, Quinlan and Mellor. Background: High altitude (HA) exposure can lead to changes in resting heart rate variability (HRV), which may be linked to acute mountain sickness (AMS) development. Compared with traditional HRV measures, non-linear HRV appears to offer incremental and prognostic data, yet its utility and relationship to AMS have been barely examined at HA. This study sought to examine this relationship at terrestrial HA. Methods: Sixteen healthy British military servicemen were studied at baseline (800 m, first night) and over eight consecutive nights, at a sleeping altitude of up to 3600 m. A disposable cardiac patch monitor was used, to record the nocturnal cardiac inter-beat interval data, over 1 h (0200-0300 h), for offline HRV assessment. Non-linear HRV measures included Sample entropy (SampEn), the short (α1, 4-12 beats) and long-term (α2, 13-64 beats) detrend fluctuation analysis slope and the correlation dimension (D2). The maximal rating of perceived exertion (RPE), during daily exercise, was assessed using the Borg 6-20 RPE scale. Results: All subjects completed the HA exposure. The average age of included subjects was 31.4 ± 8.1 years. HA led to a significant fall in SpO 2 and increase in heart rate, LLS and RPE. There were no significant changes in the ECG-derived respiratory rate or in any of the time domain measures of HRV during sleep. The only notable changes in frequency domain measures of HRV were an increase in LF and fall in HFnu power at the highest altitude. Conversely, SampEn, SD1/SD2 and D2 all fell, whereas α1 and α2 increased (p < 0.05). RPE inversely correlated with SD1/SD2 (r = -0.31; p = 0.002), SampEn (r = -0.22; p = 0.03), HFnu (r = -0.27; p = 0.007) and positively correlated with LF (r = 0.24; p = 0.02), LF/HF (r = 0.24; p = 0.02), α1 (r = 0.32; p = 0.002) and α2 (r = 0.21; p = 0.04). AMS occurred in 7/16 subjects (43.8%) and was very mild in 85.7% of cases. HRV failed to predict AMS. Conclusion: Non-linear HRV is more sensitive to the effects of HA than time and frequency domain indices. HA leads to a compensatory decrease in nocturnal HRV and complexity, which is influenced by the RPE measured at the end of the previous day. HRV failed to predict AMS development.