Slow sinus heart rate as a potential predictive factor of paroxysmal atrial fibrillation in stroke patients

Background: Paroxysmal atrial fibrillation (PAF) is a strong independent risk factor for ischemic stroke and is also associated with stroke severity and mortality. However, the detection rate of PAF in patients with ischemic stroke is limited because they are often asymptomatic or present with sinus rhythm on electrocardiograms (ECGs). In the present study, we aimed to identify predictors of PAF in stroke patients by continuous ECG monitoring of the heart rate in sinus rhythm.

Methods: We enrolled 741 consecutive patients with acute symptomatic ischemic stroke who were admitted to our hospital. Exclusion criteria were the following: (1) patients with persistent AF, (2) cardiac pacemaker users and (3) incomplete clinical investigations. Each patient was subject to 24-hour Holter ECG, cardiac monitoring by inpatient telemetry and routine transthoracic echocardiography. The minimum and mean sinus heart rates (SHRs) on 24-hour Holter ECGs were recorded. The presence of PAF was judged on the basis of previous history and cardiac studies during hospitalization. Clinical characteristics of the patients with and without PAF were compared, and multiple logistic regression analysis was performed to identify the predictors of PAF.

Results: Of all enrolled patients, 606 (mean age 68.2 years, 63.4% male) were eligible for analysis, and the presence of PAF was confirmed in 116 subjects (19.1%). In the univariate analysis, the patients with and without PAF showed significant differences in age (74.3 vs. 66.7 years, p < 0.001), dyslipidemia (32.8 vs. 50.4%, p = 0.001), chronic heart failure (16.4 vs. 4.7%, p < 0.001), admission National Institute of Health Stroke Scale score (8.5 vs. 6, p < 0.001) and absence of vascular etiology (80.2 vs. 54.1%, p < 0.001). Furthermore, the minimum and mean SHRs were slower in the patients with than in those without PAF (46.5 vs. 54.0 bpm, p < 0.001 and 70.7 vs. 73.8 bpm, p = 0.009, respectively). In 501 (82.7%) out of 606 patients, minimum SHRs were detected at night (from 9:00 p.m. to 7:00 a.m.). A minimum SHR was an independent predictive factor of PAF in the multivariate analysis (odds ratio 1.08, 95% confidence interval 1.05 - 1.12, p < 0.001).

Conclusions: A slow SHR on monitoring ECG is a potential predictive factor of PAF in patients with ischemic stroke.