Improving 10-year cardiovascular risk prediction in apparently healthy people: flexible addition of risk modifiers on top of SCORE2
An erratum to this article was published (Eur J Prev Cardiol, 20 Sep 2024 ↗)
Authors' conclusion
This flexible methodology improves the accuracy of predicted risks and increases applicability of prediction models for individuals with additional risk known modifiers.
Study context
- Population
- Individuals without previous cardiovascular disease (CVD) or DM (UK Biobank, Atherosclerosis Risk in Communities (ARIC), Multi-Ethnic Study of Atherosclerosis (MESA), EPIC-NL, Heinz Nixdorf Recall (HNR)); external validation in Clinical Practice Research Datalink (CPRD) (UK)external validation n = 518,015
- Participants
- 409,757 individuals
- Comparison
- SCORE2 adjusted with risk modifiers vs unimproved SCORE2
Methodological detailsmulticohort, competing risk-adjusted Fine and Gray models
- Multicohort
- Competing risk-adjusted Fine and Gray models
Performance
As reported in the abstract
C-index (discrimination), external validation (CPRD)
SCORE2 + risk modifiers
0.740
C-index · 95% CI 0.736–0.745
SCORE2
0.737
C-index · 95% CI 0.732–0.741
- Increase in discrimination, coronary artery calcium (CAC) percentile
- 0.0198 [95% CI 0.0115; 0.0281]
- Increase in discrimination, hs-Troponin-T
- 0.0100 (95% CI 0.0063; 0.0137)
Further findings
| Outcome | Increase in discrimination (95% CI) | |
|---|---|---|
| Increase in discrimination by risk modifier (derivation cohorts) — CAC percentile vs Unimproved SCORE2 | 0.0198 (0.0115–0.0281) | |
| Increase in discrimination by risk modifier (derivation cohorts) — hs-Troponin-T vs Unimproved SCORE2 | 0.0100 (0.0063–0.0137) | |
| −0.0200.020.04Difference (linear scale) |
- Increase in discrimination by risk modifier (derivation cohorts): 16 166 CVD events, 19 149 non-CV deaths over 10.0 years
Original publication
Improving 10-year cardiovascular risk prediction in apparently healthy people: flexible addition of risk modifiers on top of SCORE2. Eur J Prev Cardiol. Published 2 June 2023. doi:10.
Summarised from the published abstract; every number is checked against it