Prediction of left ventricle ejection fraction from HRV in diabetics and cardiac diseased subjects: A mathematical model approach

Abstract

Heart Rate variability (HRV) is the variation between beat-to-beat intervals in a consecutive heartbeat. HRV provides enormous information about cardiac health. Electrocardiogram and echocardiography (2D echo) are the two major techniques used by medical practitioners to diagnose cardiac health. The 2D echo provides more accurate results as compared to electrocardiography but costs ten times more. This research provides low-cost solution for cardiac health diagnosis to the level of 2D echo derived from the electrocardiogram. ECG and the 2D echo data of the same cohort are collected from 15 diabetics and 15 cardiac-diseased subjects. In this research, we found a correlation between significant HRV parameters and 2D Echo parameters. Maximum correlation is obtained between LF/HF (low-frequency power to high-frequency power ratio), the HRV parameter, and LVEF (left ventricle ejection fraction), the 2D echo parameter. We derived various mathematical models, to predict the LVEF, from LF/HF for both cardiac diseased and diabetics separately. The combined best-fit model is obtained using nonlinear regression with the goodness of fit, R2 ​= ​94.54 %.

Keywords

• Left ventricle ejection fraction
• Heart rate variability
• Autonomous nervous system
• Cardiac health
• Diabetes