Screening of heart diseases with multivariate short-term heart rate variability analysis

Abstract

Heart diseases are the leading cause of mortality in the western world. The aim of this pilot study was to examine the possibility of screening heart diseases for the general practitioner by short-term analysis (five-minute ECG segments) of the heart rate variability (HRV) applying linear (time domain and frequency domain) and nonlinear (detrended fluctuation analysis (DFA), compression entropy (CE), symbolic dynamics (SD), short-term symbolic dynamics (STSD), Poincaré Plot analysis (PPA)) methods. 78 healthy men and 378 male patients with several heart diseases (myocardial infarction, heart failure with and without peripheral arterial disease) were investigated considering two age groups: young men (30-50 years) and elder men (51-70 years). The identification of heart disease patients was performed using a scoring system that based on parameters calculated from the healthy subject groups. The combination of parameters from TD, FD, SD, STSD and CE achieved sensitivities of 70.37% for the group of young men and of 76.09% for the elder male group and a specificity in both age groups of 100%. In conclusion, multivariate short-term analyses of HRV applying both linear and nonlinear methods are suitable for screening heart diseases in men.