Work

Abstract Number: Esoc2026lt82 Lt82 - Optimal Detection of Atrial Fibrillation in Tia Patients (odea-tia) Trial

Year: 2026
Type: conference-abstract
Abstract: Background and aims Effective stroke prevention in TIA patients depends on identifying underlying mechanisms including atrial fibrillation (AF). Currently, the appropriate extent and modality of cardiac monitoring for AF detection in TIA patients are unknown. We conducted an investigator-initiated trial comparing 3 methods of cardiac monitoring. Methods We randomly assigned, in a 1:1:1 ratio, TIA patients without known AF to receive either 24h Holter-ECG, 28d continuous non-invasive ECG monitoring or an implantable cardiac monitor (ICM; REVEAL-LINQ). Patients were enrolled within 30d of index TIA and followed-up for 24 months. The primary endpoint was AF detection rate (episodes >30sec) 6 months after enrollment. Pairwise comparisons between groups were sequentially tested at interim analysis (200 patients) and adapted for final analysis. Secondary endpoints included AF detection at 12 and 24 months. Clinical endpoints were adjudicated for exploratory analysis. Results 516 patients were enrolled by 27 German and Spanish sites. 504 patients started ECG monitoring. For the modified intention-to-treat analysis, 158 underwent ICM, 176 24h and 170 28d monitoring. Mean age was 69+/-10 years, 57% were male. At interim analysis and in the entire study population, AF was detected significantly more frequently in the ICM (15.6%; 14.6%) compared to the 24h (1.5%; 3.4%) and 28d (1.4%; 3.5%) groups. Secondary/exploratory outcomes are being analyzed. Conclusions In ODEA-TIA, the first RCT examining AF detection specifically in TIA patients, monitoring with ICM was superior to short-term and prolonged non-invasive ECG recording. The impact of more AF detection by ICM on clinical endpoints will be presented at ESOC. Conflict of interest
Language: English
Field: Medicine
Open Access status: hybrid
Estimated APC paid: $4,400
OpenAlex ID: W7160428407
Cites:
Cited by:
Related to: