Abstract
BACKGROUND: Atrial fibrillation (AF) is currently diagnosed by ECG, creating a binary, lifelong diagnosis. AF burden, estimated as the proportion of time spent in AF, quantifies AF severity dynamically. AF burden can modulate the risk of AF-related outcomes. Whether AF burden modulates cardiovascular outcomes with rhythm-control therapy is unknown. METHODS: AF burden on early rhythm-control was estimated using supervised artificial-intelligence-based rhythm classification of patient-operated telemetric short-term ECGs in patients randomised to early rhythm-control in the EAST-AFNET 4 trial (NCT01288352, ISRCTN04708680, conducted between 2011 and 2020). ECGs were transmitted 1-2 times per week and during symptoms. A landmark was set at 12 months and efficacy and safety outcomes occurring during the subsequent 4.1 years of follow-up were compared by estimated AF burden quartiles (Q1-Q4). FINDINGS: In 1178 patients (70 years, 47% women, CHA(2)DS(2)-VA 2.8 ± 1.2) transmitting 303,308 ECGs over 5.1 years, (median 1/week, IQR 1; 2) estimated AF burden was 6% [0%; 22%] in the first year of follow-up. Estimated AF burden below the median was associated with low rates of cardiovascular death, stroke, or unplanned hospitalisation for heart failure or acute coronary syndrome (Q1: 2.0 events/100 patient-years; Q2: 2.6 events/100 patient-years). A higher estimated AF burden was associated with higher event rates (Q3: 4.8 events/100 patient-years; Q4: 4.2 events/100 patient-years), comparable to events with usual care (4.5 events/100 patient-years). Sensitivity analyses confirmed these findings. INTERPRETATION: These hypothesis-generating findings suggest that AF burden estimated by weekly short-term patient-operated ECGs modulates AF-related events on rhythm-control therapy. Pending validation and evaluation of residual confounding, estimation of AF burden can refine AF diagnosis. FUNDING: EAST-AFNET4 was supported by a grant from the German Ministry of Education and Research (01 GI0204) via the German Center for Cardiovascular Research (DZHK), the Atrial Fibrillation NETwork (AFNET), the European Heart Rhythm Association, St. Jude Medical/Abbott, Sanofi, and the German Heart Foundation. These analyses received additional support from the European Union (grant agreement 965286 [MAESTRIA]), British Heart Foundation (AA/18/2/34218), German Center for Cardiovascular Research supported by the German Ministry of Education and Research (DZHK, grant numbers DZHK FKZ 81X2800182, 81Z0710116, and 81Z0710110), German Research Foundation (Ki 509167694) and the Else Kröner-Fresenius Foundation, Dutch Heart Foundation (Grant number 01-002-2022-0118, EmbRACE), and the Leducq Foundation (2024, Immune Targets for Atrial Fibrillation).