Arrhythmias & electrophysiology
New-onset atrial fibrillation prediction: the HARMS2-AF risk score
Authors' conclusion
The HARMS2-AF score is a novel lifestyle risk score which may help identify individuals at risk of AF in the general community and assist population screening.
Study context
- Population
- General population without prior atrial fibrillation (AF) (UK Biobank derivation; Framingham Heart Study validation)UK Biobank (UKB)
- Participants
- 314,280 participants
- Exposure
- HARMS2-AF score vs unweighted model, Framingham-AF, ARIC and CHARGE-AF risk models
- Outcome
- Incident AF (ICD-10 coding)
Methodological detailsrisk score development and external validation, derivation in UK Biobank, external validation in Framingham Heart Study
- Risk score development and external validation
- Derivation in UK Biobank
- External validation in Framingham Heart Study
Main finding
As reported in the abstract
Predictive performance for incident AF
| Group | AUC (UKB) | AUC (Framingham Heart Study (FHS) external validation) | AUC |
|---|---|---|---|
| HARMS2-AF | 0.782 | 0.757 (95% CI 0.735-0.779) | — not stated in the abstract |
| Unweighted model | 0.802 | — not stated in the abstract | — not stated in the abstract |
| Framingham-AF | — not stated in the abstract | — not stated in the abstract | 0.568 |
| ARIC | — not stated in the abstract | — not stated in the abstract | 0.713 |
| CHARGE-AF | — not stated in the abstract | — not stated in the abstract | 0.754 |
AUC (UKB)
HARMS2-AF0.782
Unweighted model0.802
Framingham-AF— not stated in the abstract
ARIC— not stated in the abstract
CHARGE-AF— not stated in the abstract
AUC (Framingham Heart Study (FHS) external validation)
HARMS2-AF0.757 (95% CI 0.735-0.779)
Unweighted model— not stated in the abstract
Framingham-AF— not stated in the abstract
ARIC— not stated in the abstract
CHARGE-AF— not stated in the abstract
AUC
HARMS2-AF— not stated in the abstract
Unweighted model— not stated in the abstract
Framingham-AF0.568
ARIC0.713
CHARGE-AF0.754
— not stated in the abstract
- HARMS2-AF vs Framingham-AF: P < 0.001 — HARMS2-AF outperformed
- HARMS2-AF vs ARIC: P < 0.001 — HARMS2-AF outperformed
- HARMS2-AF vs CHARGE-AF: P = 0.73 — comparable
Observational data: associations, not effects.
Further findings
AF incidence and time to AF (UKB)
- AF incidence (UKB)
- 5.7%
- Median time to AF (IQR 4.5-10.2)
- 7.6 years
- AF incidence (FHS)
- 6.0%
AF risk by HARMS2-AF score (≥5 points)
- Score 5-9, HR
- 12.79
- Score 10-14, HR
- 38.70
Individual predictors
- Hypertension, age, BMI, male sex, sleep apnoea, smoking and alcohol predictive
- all P < 0.001
- Physical inactivity, HR 1.01 (95% CI 0.96-1.05)
- P = 0.80
- Diabetes, HR 1.03 (95% CI 0.97-1.09)
- P = 0.38
Original publication
New-onset atrial fibrillation prediction: the HARMS2-AF risk score. Eur Heart J. Published 23 June 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →