Atrial fibrillation density as a biomarker for ischaemic stroke risk prediction.

Abstract (English)

BACKGROUND AND AIMS: Atrial fibrillation (AF) is frequently classified by episode duration and cumulative burden, yet these methods fail to characterize temporal episode distribution (AF density), which may be more clinically relevant and prognostically important. This study evaluated AF density's association with ischaemic stroke and whether it improves risk stratification compared with AF burden. METHODS: Data from two US cohorts with cardiac implantable electronic devices remotely monitored by the Veterans Health Administration and University of North Carolina (January 2010-May 2025) were analysed. AF burden (percentage of time in AF) and density [range: 0 (dispersed episodes) to 1 (consolidated episodes)] were assessed in 30-day intervals and categorized as low (>0-0.3), medium (>0.3-0.6), medium-high (>0.6-0.9), and high (>0.9-1.0). Patients with permanent AF or no episodes ≥6 min were excluded. G-formula modelling estimated 1-year stroke risk ratios (RRs), adjusting for baseline and time-varying covariates. Results were pooled using random-effects meta-analysis. RESULTS: Of 41 780 patients, 12 868 met inclusion criteria (mean age 72.0 years; median CHA2DS2-VASc 4.0); 336 experienced ischaemic stroke over a median 4.0-year follow-up (6.3 per 1000 person-years). Atrial fibrillation density demonstrated a dose-response relationship with 1-year stroke risk (RR 1.75; 95% confidence interval 1.25-2.44) and findings were consistent across device types, comorbidities, age, and anticoagulation status. At each level of AF burden, patients with high density exhibited greater stroke risk. CONCLUSIONS: Atrial fibrillation density demonstrated robust dose-response relationships with ischaemic stroke independent of AF burden and enhanced risk stratification, suggesting that density may enable more precise stroke risk assessment and personalized prevention strategies.