Supervised Machine Learning to Examine Factors Associated with Respiratory Sinus Arrhythmias and Ectopic Heart Beats in Adults: A Pilot Study

Abstract

Background: There are many types of arrhythmias which may threaten health that are well-known or opaque. The purpose of this pilot study was to examine how different cardiac health risk factors rank together in association with arrhythmias in young, middle-aged, and older adults. Methods: The analytic sample included 101 adults aged 50.6 ± 22.6 years. Several prominent heart-health-related risk factors were self-reported. Mean arterial pressure and body mass index were collected using standard procedures. Hydraulic handgrip dynamometry measured strength capacity. A 6 min single-lead electrocardiogram evaluated arrhythmias. Respiratory sinus arrhythmias (RSAs) and ectopic heart beats were observed and specified for analyses. Classification and Regression Tree analyses were employed. Results: A mean arterial pressure ≥ 104 mmHg was the first level predictor for ectopic beats, while age ≥ 41 years was the first level predictor for RSAs. Age, heart rate, stress and anxiety, and physical activity emerged as important variables for ectopic beats (p < 0.05), whereas age, sodium, heart rate, and gender were important for RSAs (p < 0.05). Conclusions: RSAs and ectopic arrhythmias may have unique modifiable and non-modifiable factors that may help in understanding their etiology for prevention and treatment as appropriate across the lifespan.