Arrhythmias & electrophysiologyDiagnostic or prognostic study
Proteomic Signatures for Risk Prediction of Atrial Fibrillation
- Prognostic cohort study (risk score derivation and validation)
- derivation 70% / test 30% split
- external replication (ARIC)
- lasso-penalized Cox regression
Authors' conclusion
Protein risk score derived from a single plasma sample improved risk prediction of AF.
Study context
- Population
- Individuals without a history of atrial fibrillation (AF) with plasma proteomics, UK Biobank Pharma Proteomics Project (UKB-PPP)ARIC replication n = 11,012
- Participants
- 51,680 individuals
- Comparison
- CHARGE-AF + NT-proBNP + polygenic risk score with vs without protein risk score
- Outcome
- Incident AF
Performance
As reported in the abstract
Incident AF (UKB-PPP test set)
HR per 1-SD increase in protein risk score
2.20 (95% CI 2.05-2.41)
C index, CHARGE-AF + NT-proBNP + polygenic + protein risk score
0.816 (95% CI 0.802-0.829)
C index, CHARGE-AF + NT-proBNP + polygenic risk score
0.771 (95% CI 0.755-0.787)
C-index change
0.044 (95% CI 0.039-0.055)
Further findings
Risk reclassification with protein risk score added (5-year risk threshold 5%)
- Risk reclassification
- 5.4% (95% CI 2.9%-7.9%)
Decision curve net benefit (5-year risk threshold 5%)
- Net benefit before addition of protein risk score, per 1000 people
- 3.8
- Net benefit after addition of protein risk score, per 1000 people
- 5.4
External replication (ARIC)
- Consistent improvement in risk stratification of AF
Protein risk score composition
- Unique plasma proteins in score
- 165
- Proteins associated with atrial remodeling
- 15
Original publication
Proteomic Signatures for Risk Prediction of Atrial Fibrillation. Circulation. Published 22 May 2025. doi:10.1161/circulationaha.124.073457
Summarised from the published abstract; every number is checked against it