Noninvasive Cardiac Power Output Estimation on Healthy Elderly Subjects during Dynamic Exercise

Abstract

Cardiac power output (CPO), the product of cardiac output (CO) and mean blood pressure (MBP) measured during cardiopulmonary exercise testing, has been well recognized as a powerful prognostic indicator in chronic heart failure. The accurate measurement of CPO requires synchronously beat to beat measurement of CO and BP during dynamic exercise, which is a challenge for current measuring techniques. In this paper we proposed a novel CPO estimation approach, which is based on the waveform analysis of two signals measurable from wearable devices, electrocardiogram (ECG) and photoplethysmogram (PPG), and hence this approach was potential to meet the strict measurement requirements of CPO under exercise. BP was estimated from pulse arrival time (PAT), the time delay from ECG R peak to the onset of PPG signal. And CO was estimated from pulse transit reflection ratio (PTRR), a CO index proposed in our previous study. We evaluated the success of this technique by testing on fifteen healthy elderly subjects before and after bicycle exercise. The results showed a CPO estimation accuracy of 0.06W ±0.46 W against reference CPO calculated as a product of BP measured by auscultatory method and CO measured by impedance cardiographic method. The estimated peak CPO can trace the variations of the reference peak CPO on different subjects. With further investigations on the calibration procedure, the accuracy of the proposed method may achieve the requirement of a medical CPO measurement.