Estimating cardiac autonomic activity during sleep: impedance cardiography, spectral analysis, and Poincaré plots

Objective: To compare noninvasive measures of cardiac autonomic activity during sleep. Methods: The absolute and normalized (n.u.) high and low frequency peaks from the spectral analysis of R-R intervals (HF, LF, HFn.u., LFn.u.), LF/HF ratio, pre-ejection period (PEP) from impedance cardiography, and the autocorrelation coefficient (rRR) as illustrated in Poincaré plots were measured during night-time sleep in 9 young healthy subjects. Heart rate and blood pressure were also recorded. Results: Heart rate was significantly associated with cardiac sympathetic activity (PEP, average r=−0.46), but not with cardiac parasympathetic activity (HF, average r=−0.17). rRR was significantly associated with heart rate (average r=0.41), and LF/HF (average r=0.69), but not with PEP or HF. From NREM to REM sleep, heart rate, LFn.u., LF and rRR significantly increased, HFn.u. significantly decreased, LF/HF showed an increasing trend (P=0.07) and PEP showed a decreasing trend (P=0.06). Blood pressure and HF were highly variable without significant changes from NREM to REM sleep. Conclusions: Cardiac parasympathetic activity (HF) does not vary greatly between sleep stages. Cardiac sympathetic activity (PEP) decreases linearly during sleep. rRR and LF/HF can track sympathovagal changes during sleep, but cannot differentiate between changes in cardiac parasympathetic and sympathetic activity. The relative advantages and disadvantages of the different measures are discussed.