39
Short time window analysis of heart rate variability during tilt table testing provides insights in to the timing of changes in the autonomic nervous system Free

Introduction: The autonomic nervous system (ANS) responds within seconds to homeostatic perturbations. Heart rate variability (HRV), though a widely accepted ANS measurement, fails to detect sudden changes as sampling windows of 5 minutes or more are typically used.

Aim: We sought to determine whether 1 minute sampling windows for frequency-domain and non-linear HRV parameters could be used to describe rapid changes in the ANS accurately during tilt table testing (TTT).

Methods: Frequency-domain parameters (LFnu, HFnu) and non-linear short term measures of variability (SD1 from the Poincaré plot and α1 slope from detrended fluctuation analysis) were derived from patients undergoing clinically indicated TTTs. To determine reproducibility of 1 minute windows, a 5 minute rest period was analysed then divided into 1 minute intervals. Using 1minute windows we analysed the TTTs: 1minute before and 5 minutes after tilt and GTN then 5 minutes pre-syncope. In those without syncope an equivalent pre-syncope 5 minute time frame was estimated.

Results: TTT assessments in 295 consecutive patients were reviewed. Subjects were excluded if they had unusable traces (e.g. atrial fibrillation, paced rhythms) leaving 84 patients without syncope (tilt negatives) and 27 tilt-induced syncope patients (tilt positives).

All four HRV parameters were reproducible at rest using 1 minute windows. LFnu and α1 increased significantly while HFnu and SD1 fell during the 5 minutes after tilt-up and GTN administration. The changes post tilt-up were first detectable at 1-2 minutes (LFnu, HFnu) but there were no differences in any parameter with respect to tilt outcome negative/positive. However, in the pre-syncope period all parameters were significantly different between negative and positive tilts. The earliest difference recognised was at 3 minutes pre-syncope when, compared with tilt negatives, syncope patients displayed reduced LFnu (65.2±20.9 vs 77.3±19.0, p=0.04), increased HFnu (34.5±33.8 vs 22.6±18.8, p=0.04) and a reduced α1 (1.0±0.4 vs 1.4±0.3, p=0.004).

Conclusions: One minute time windows may be used to analyse HRV with improved temporal sensitivity. Using this technique, we demonstrated changes consistent with sympathetic activation in patients at tilt-up and nitrate administration followed by a shift to vagal predominance pre-syncope in positive tests. Our work uniquely identifies changes in the frequency-domain parameters to occur early, between minute 1 and 2 of tilt-up, and that vagal activation could be detected as early as 3 minutes before syncope. This technique may be useful for detecting acute changes in the ANS in a variety of situations.