A study of heart rate variability as sleep apnoea predictor over two different databases

Abstract

The main aim of this work is to study the consistency of ECG-based approaches for the screening of apnoeic patients, which are based on estimation of the Heart Rate Variability (HRV) by computing the so called High Frequency to Low Frequency (HF/LF) ratio. Different intervals and thresholds proposed in the literature for the previous ratio are compared with those found by the authors, and tested on two different databases. While more studies need to be launched on this area, our results suggest that patient classification according to this approach suffers from high variability and is widely dependent of the set of patients used for the tests.