Sleep Apnea Assessment by ECG Pattern

Sleep apnea (SA) is accompanied by a characteristic cyclic variation in heart rate in the waveform of the electrocardiogram (ECG). If sleep apnea were able to be diagnosed with ECG only, it could revolutionalise the diagnosis of sleep apnea by allowing patient's to be diagnosed within their own homes. The most informative feature of ECG recordings is a low frequency oscillation at about 0.02 Hz in heart rate variability (HRV) signal. In this study we calculate the HRV spectrum of the consecutive one minute ECG segments of a patient and label each minute as "yes" or "no" according whether the 0.02 Hz oscillation dominate spectrum. By voting strategy, the ratio of "yes" to total minutes can be calculated. Compared with traditional clinical SA diagnosis parameter of apnea and hyponea index (AHI), the ratio shows good relationship with AHI. The result suggests that ECG pattern can be a good evaluation tool of SA.