Arrhythmias & electrophysiology
The predictive power of left atrial reservoir strain for long-term atrial fibrillation in the Atherosclerosis Risk in Communities study
Authors' conclusion
The standalone predictive performance of LA reservoir strain for long-term AF risk is comparable with the CHARGE-AF score (11 variables) and slightly lower than that of a machine learning-based model with 12 variables, highlighting the importance of LA reservoir in AF risk prediction.
Study context
- Population
- Older adults without previous atrial fibrillation (AF), Atherosclerosis Risk in Communities study (median age 74)
- Participants
- 5,442 adults
- Comparison
- Left atrial (LA) strain measures vs CHARGE-AF score vs machine learning-based model for AF prediction
- Outcome
- Standalone predictive performance (Harrell's C-index) of LA function measures for long-term AF
- Follow-up
- 5 and 10 years
Methodological detailssecondary analysis of the ARIC cohort, training and testing sets (7:3 ratio), machine learning-based model
- Secondary analysis of the ARIC cohort
- Training and testing sets (7:3 ratio)
- Machine learning-based model
Performance
As reported in the abstract
Harrell's C-index for AF prediction
| Test or model | C-index, 5-year | C-index, 10-year |
|---|---|---|
| LA reservoir strain | 0.664 | 0.646 |
| LA reservoir strain + age | 0.677 | 0.663 |
| CHARGE-AF score | 0.667 | 0.655 |
| Machine learning-based model | 0.732 (0.686-0.777) | 0.725 (0.693-0.757) |
C-index, 5-year
LA reservoir strain0.664
LA reservoir strain + age0.677
CHARGE-AF score0.667
Machine learning-based model0.732 (0.686-0.777)
C-index, 10-year
LA reservoir strain0.646
LA reservoir strain + age0.663
CHARGE-AF score0.655
Machine learning-based model0.725 (0.693-0.757)
Original publication
The predictive power of left atrial reservoir strain for long-term atrial fibrillation in the Atherosclerosis Risk in Communities study. Heart Rhythm. Published 17 October 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →