The Usefulness of Heart Rate Variability Spectrum Analysis in Detecting Covert Atrial Fibrillation among Patients with Cryptogenic Stroke

We examined the usefulness of heart rate variability spectrum analysis as a predictor of covert paroxysmal atrial fibrillation (PAF), an essential risk factor for cryptogenic stroke. Between July 2021 and January 2022, 77 patients (mean age, 76.3 ± 10.9 years) hospitalized with cryptogenic stroke within 3 days of symptom onset were monitored by performing an electrocardiogram (ECG) using telemetry ECG transmitter Duranta®. Autonomic nervous activity up to maximum 30 days after mounting Duranta® was analyzed with and without PAF. PAF was detected in 20 episodes among 8 patients within 1 month. During 24 h post-mounting, 8 patients with PAF had significantly higher levels of low-frequency components (LF), high-frequency components (HF), and HF normalization units and significantly lower levels of LF/HF ratio than those without PAF. Conversely, LF and HF were significantly lower < 1 h before the onset of PAF, especially 15 min before onset. In patients with PAF, parasympathetic activity is predominant in the acute phase of brain infarction; however, parasympathetic activity is reduced just before the PAF occurrence. These findings suggest that autonomic changes may trigger PAF occurrence.