Automatic diagnostics of extrasystoles during cardiointervalogram preliminary assessment (filtration) in systems for heart rate variability analysis

Abstract

Heart rate variability analysis requires the following: registered rhythm must be sinus one, regular or arrhythmic; cardiac arrhythmias should be diagnosed automatically and interpoled without disturbing RR interval dynamic row. The study included 560 patients: 217 clinically healthy volunteers, 121 patients with acute myocardial infarction, and 222 individuals with chronic coronary heart disease (CHD). Basal rhythms were sinus rhythm (73,7 %), or sinus arrhythmia (26,3 %). Patients with atrial fibrillation, ventricular and supraventricular tachycardia, sinoatrial and AV blockade were excluded. Differential diagnostics of sinus and arrhythmic RR intervals was based on searching the cut-off point, discriminating sinus or non-sinus RR intervals. Similar approach could be applied to identifying shortened intervals between sinus impulse and «noise» artifacts. This gives a chance to cluster sinus and pathologically shortened RR intervals. In automatic assessment of heart rate variability, on cardiointervalograms with extrasystoles and artifacts, effective signal filtration could be performed. Clustering sinus and non-sinus RR intervals can be achieved by analyzing information on adjacent RR intervals. Secondary modification, like RRi/RRi+1, with cut-off point no less than 118, provides optimal balance of sensitivity (94,8 %) and specificity (94,9 %), with utility coefficient 0,89, during diagnosing pathologically shortened RR intervals in sinus rhythm or sinus arrhythmia.