Regularity of Heart Rate Fluctuations Analysis in Congestive Heart Failure Patients Using Information-Based Similarity

Abstract

Congestive Heart Failure (CHF) is a chronic progressive condition that affects the pumping power of your heart muscle. There are a number of works investigating Obstructive Sleep Apnea (OSA) detection based on heart rate variability and obtain outstanding results. Therefore, using HRV analysis for the screening of CHF patients has great potential. This study included 30 electrocardiogram (ECG) recordings (15 CHF recordings and 15 normal recordings) from the PhysioNet database. These recordings included 24h RR interval data and were divided into 5-minnute segments. Comparing with traditional time-domain analysis and frequency-domain analysis, information-based similarity (IBS) has a better performance on showing significant differences between normal group and CHF group (p < 0.001). The accuracies of time-domain analysis and frequency-domain analysis in the CHF detection are 86.7% and 83.3%, respectively, while IBS performed an accuracy of 86.7% with a better balance between sensitivity and specificity. This research find that the similarity of heart rate decreased in CHF group because of the low-level similarity of adjacent RR segments. This finding is probably the reason that CHF patients have arrhythmia. Consequently, this IBS method has certain clinical significance, and could be used to detect CHF.