Heart risk models predicted co-occurring disease risk
Scores measuring how well models distinguished who developed co-occurring diseases ranged from 0.74 to 0.83, versus 0.71 to 0.74 for cardiovascular disease.
Our read
Biomarkers- Studied:
- Preprint: cohort study of UK Biobank volunteers and UK primary care patients, median 10 years (Independent replication)
- What changed:
- Models distinguished co-occurring disease cases with scores of 0.74 to 0.83, versus 0.71 to 0.74 for cardiovascular disease
- The catch:
- The analysis measured prediction accuracy, not improvements in prevention or screening
- Stake:
- Cancer Research UK funded the study.
- In the record:
- The 6th paper on QRISK3 with results in people; 43 on record in all.
- Why it’s aging research:
- It measures an outcome that worsens with age.
Editor of record: Boris Djordjevic · AgingResearch.News
UK Biobank volunteers and UK primary care patients were studied in an observational analysis of four heart and blood vessel disease risk models. Available data covered 502,126 volunteers, averaging 56.6 years old, and 4,810,089 patients, averaging 58.5. Women accounted for 54.4% and 50%, respectively. Median follow-up was 10 years in both groups.
The researchers adjusted the models using 20% of Biobank data and tested them in the remaining 80%. They then checked predictions against primary care records. The models ranked 10-year risk of heart failure and other conditions, plus heart and blood vessel disease occurring alongside type 2 diabetes, dementia or Parkinson’s disease. They slightly underestimated risk in the primary care group. After age and sex, smoking and systolic blood pressure—the higher blood pressure reading—contributed most to predictions.
Why it matters
For diseases common in later life, the authors suggest these models may support prevention and screening based on a person’s risk of developing several conditions.
Caveats
The analysis tested predictions, not whether using them improved prevention or screening; the primary analysis included only people with complete data. External validation covered primary care patients aged 40 to 84, and the work is a preprint, not yet peer-reviewed.
The paper
British Heart Foundation; Amsterdam UMC Location University of Amsterdam
medRxiv, Jun 15, 2026, Open access, CC BY, Preprint, not yet peer-reviewed
Comments
Be the first to comment.