QT interval adaptation to changes in autonomic balance

Variability in the RR interval has long been used to assess the state of the autonomic nervous system (ANS), with rapid changes being mediated by the action of the parasympathetic branch and slower variations through a combination of the sympathetic and parasympathetic branches. By contrast, the action of the ANS on QT variability may primarily be affected through the sympathetic branch. This motivates us here to assess QT interval variation during changes in autonomic balance. To detect the QT interval, we implemented a pair of detection algorithms based on the curve-length transform. These were tested against cardiologists' annotations of 102 records in the PhysioNet QT database. Our QT interval measurements achieved a mean error (± standard deviation) of -3.0 (±39.8) ms, comparable with the variability between the annotations of two cardiologists, and better than the -16.1 (±52.9) ms of the archived PhysioNet annotations associated with these records. Our algorithms were applied to archived ECG records of nine subjects undergoing tilt-table experiments. The interventions included passive rapid tilt-up, passive slow tilt-up and active stand-ups. The RR intervals responded quickly to each change in posture, while the QT intervals showed a distinctly slower response, with characteristics of a first-order exponential, exhibiting time constants in the range of 40 - 140 s across subjects.