Device-Detected Atrial Fibrillation and Risk for Stroke: An Analysis of >10,000 Patients From the SOS AF Project (Stroke Prevention Strategies Based On Atrial Fibrillation Information From Implanted Devices)

Study Questions:

In patients with cardiac implanted electronic devices (CIEDs), what is the association between the maximum daily atrial fibrillation (AF) burden and the risk of ischemic stroke?


This was a pooled analysis of patient data from three large prospective observational studies (TRENDS, PANORAMA, and the Italian ClinicalService Registry Project). Eligible patients had been previously implanted with devices capable of continuous AF detection and had at least 3 months of continuous device monitoring; participants with permanent AF were excluded. The risk of ischemic stroke associated with prespecified cut-off points of AF burden (5 minutes; 1, 6, 12, and 23 hours, respectively) was assessed.


The median follow-up was 24 months, during which 43% experienced at least 1 day with a burden of at least 5 minutes of AF. The annual rate of ischemic stroke or transient ischemic attack was 0.39%. The 1-hour burden of AF was associated with the highest hazard ratio (HR) for ischemic stroke (2.11, 95% confidence interval [CI], 1.22-3.64; p = 0.008). Thresholds of ≥6, ≥12, and ≥3 hours did not reach statistical significance. In analyses adjusted for CHADS2 score and use of anticoagulants at baseline, device-detected AF burden ≥1 hour remained an independent predictor of stroke.


The authors concluded that daily AF burden in a population of patients with an arrhythmia detecting CIED in place is associated with an increased risk of ischemic stroke.


The limitations of this study aside (including the heterogeneity of the three studies used for pooled analysis and the low absolute event rate of stroke), the authors provide useful information describing the association between AF burden and stroke. It is interesting that the 1-hour threshold of AF burden was significantly associated with the outcome, whereas other longer thresholds did not reach statistical significance. The authors suggest that ‘discriminatory capability is lost [at these higher burdens], probably because the risk is consistent below the proposed threshold.’ Future studies should clarify how continuous monitoring of AF burden may inform the risk of stroke in this population and the role of remote transmission of data on AF burden.

Keywords: Prostheses and Implants, Stroke, Ischemic Attack, Transient

< Back to Listings