Arrhythmias & electrophysiologyCohort study
External validation and comparison of published stroke risk scores in a Chinese cohort of patients with atrial fibrillation
- retrospective cohort
- external validation
- Chinese tertiary-care center
Authors' conclusion
Candidate models did not consistently outperform CHA2DS2-VASc in this external validation; CHA2DS2-VASc remains practical, whereas other models may require population-specific validation and recalibration.
Study context
- Population
- Adults hospitalized with nonvalvular atrial fibrillation (AF) at a Chinese tertiary-care centeradults hospitalized with nonvalvular AF, 17 published models validated
- Participants
- 1,283 adults
- Exposure
- 17 published stroke risk scores externally validated and compared with CHA2DS2-VASc
Main finding
As reported in the abstract
Primary outcomes:
- Ischemic stroke
- Systemic embolism
Primary outcomes
103
Events per 100 person-years
2.85
Observational data: associations, not effects.
Further findings
| Outcome | CHA2DS2-VASc | GARFIELD-AF 2017 | GARFIELD-AF 2021 | ATRIA |
|---|---|---|---|---|
| Discrimination (4-year Uno C-index) | Uno C-index 0.690 | Not reported in the abstract | Not reported in the abstract | Not reported in the abstract |
| 1-year calibration | ||||
| Predicted risk | 3.39% | Not reported in the abstract | Not reported in the abstract | Not reported in the abstract |
| Observed risk | 3.54% | Not reported in the abstract | Not reported in the abstract | Not reported in the abstract |
| Observed-to-expected ratio | 1.04 | Not reported in the abstract | Not reported in the abstract | Not reported in the abstract |
- Discrimination (4-year Uno C-index) — Range across models
- 0.572 to 0.767
Discrimination (4-year Uno C-index)
- GARFIELD-AF 2017, GARFIELD-AF 2021, and ATRIA had higher Uno C-indices than CHA2DS2-VASc
- No model had a significantly higher time-dependent areas under the curve at any time point
1-year calibration
- Several other models underestimated the risk
Brier scores and reclassification
- No consistent advantage
Competing-risk sensitivity analysis
- Lowered the cumulative incidence without changing comparative patterns
- death as a competing event
Patients not receiving oral anticoagulant at baseline
- No model outperformed CHA2DS2-VASc
Original publication
External validation and comparison of published stroke risk scores in a Chinese cohort of patients with atrial fibrillation. Heart Rhythm. Published 3 September 2026. doi:10.1016/j.hrthm.2026.08.048
Summarised from the published abstract; every number is checked against it