Detection of cardiac autonomic neuropathy using linear parametric modeling of QT dynamics

In this study a linear parametric modeling technique is applied to model ventricular repolarisation (QT) variability from heart rate variability (RR) and respiration to investigate the change in cardiovascular system complexity in diabetes patients with cardiac autonomic neuropathy (CAN+) and without (CAN-). ECGs were recorded from 32 participants (20 CAN- and 12 CAN+), whilst in a supine resting position for 20 minutes and linear parametric autoregressive models using the first 250 beats were analysed. Models developed from the ECG derived respiration information could explain the QT-RR dynamics better than the models designed without any respiration information and may improve classification of CAN. QT-RR interaction provides useful information about cardiovascular system dynamics and autonomic nervous system modulation and ventricular repolarization on heart rate and heart rate variability.