A New Approach to Detecting Atrial Fibrillation Using Count Statistics of Relative Changes between Consecutive RR...

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• The ratio of the difference between neighboring RR intervals to the preceding RR interval (x%) quantifies relative cardiac cycle duration changes.
• Investigating the percentage of relative RR interval differences exceeding x% (pRRx%) for distinguishing atrial fibrillation (AF) from sinus rhythm (SR).

Purpose of the Study:

• To evaluate the diagnostic capabilities of pRRx% parameters for differentiating AF from SR.
• To identify the optimal x% value within the 0.25% to 25% range for AF detection.

Main Methods:

• Analysis of 1-minute ECG segments from the Physionet Long-Term Atrial Fibrillation Database (LTAFDB) containing AF and SR.
• Statistical evaluation of pRRx% properties across various x% values using standard diagnostic metrics.

Main Results:

• pRRx% distributions for AF and SR differed significantly across the 0.25% to 25% range, with optimal performance between 1.5% and 6%.
• pRR3.25% demonstrated superior diagnostic accuracy (95.44%), sensitivity (97.16%), and specificity (93.76%), with an AUC of 0.971.
• pRR3.25% outperformed the established pRR31 parameter in distinguishing AF from SR.

Conclusions:

• The pRRx% parameter family exhibits excellent diagnostic properties for AF detection, particularly within the 1.5% to 6% range.
• pRR3.25% emerges as a highly effective single parameter for AF diagnosis, surpassing pRR31.
• The simplicity and efficacy of pRRx% make it suitable for long-term AF monitoring using modern wearable ECG technologies.