GigaScience

Blood protein score improved heart attack risk prediction

A 23-protein score identified a 3.3-fold difference in 10-year heart attack incidence between the highest and lowest 20% risk groups in Chinese adults.

Paper by Neil Wright, Mohsen Mazidi, Alfred Pozarickij, +24 more authorsZhengming Chen · University of Oxford
Heart cross-section with intact muscle fibers beside fragmented gray tissue, above blood cells and folded proteins with one dark protein emphasized.

Our read

Biomarkers
Studied:
Cohort study of 4,000 Chinese adults, 10-year, against standard risk factors alone (Independent replication)
What changed:
Adding blood proteins improved heart attack prediction over standard risk factors alone, raising prediction accuracy from 0.846 to 0.864
The catch:
The analysis measured prediction, not heart attacks prevented by score-guided care

Editor of record: Boris Djordjevic · AgingResearch.News

Follow @agingbrief

Among 4,000 Chinese adults, an observational risk-prediction analysis measured more than 9,500 blood proteins. The group included 1,937 heart attack cases. Researchers used machine learning to select 23 proteins for a risk score and compared it with standard risk factors and scores based on inherited genetic variants.

Ten-year heart attack incidence was 3.6 per 100 in the protein score’s highest 20% risk group and 1.1 per 100 in its lowest 20%. Adding the protein score to standard risk factors improved prediction beyond standard factors alone. The team also tested the selected proteins in 37,187 Europeans in UK Biobank, including 630 heart attack cases. Adding the proteins to standard risk factors improved prediction there too.

Why it matters

Heart attacks threaten health in later life. The authors see potential for developing small blood-protein tests to refine disease-risk assessment.

Caveats

The analysis assessed prediction, not whether using the score to guide care would prevent heart attacks. Validation in Chinese adults and Europeans does not establish how well it would perform in other populations.

Comments

Be the first to comment.