Ecg derived respiratory rhythms for improved diagnosis of sleep apnea

Abstract

Respiratory rhythms of a subject are derived from measured ECG signals utilizing leads placed for significant influence of chest movement on the ECG signals. The QRS pulses within ECG signals measured in two substantially orthogonal planes are located by applying a sequence of filters. The knot and period of the QRS pulses are then determined, with adjustments made during a learning phase of data sampling. The QRS pulse areas in both planes is then calculated. These pulse areas are employed to determine the angle of orientation of the depolarization wave's mean electrical axis (MEA) at the QRS pulse locations. Cubic spline interpolation of the data points for the MEA angle provides a smooth breathing curve, which may be scored for sleep disordered breathing events. The ECG-derived respiratory (EDR) signal may be employed in lieu of airflow measurements where such measurements are not available, or may be employed in conjunction with airflow measurements and/or measured cardiac activity data to discriminate between arrhythmias associated with disordered breathing versus those associated with cardiac malfunction, reducing misdiagnosis. The additional processing of ECG data required for derivation of respiratory rhythms may be easily automated and implemented at nominal incremental cost per unit.