AHA Scientific Sessions 2026 Abstract

Blood proteins improved models of cardiovascular disease risk

Adding 30 proteins to young-adult cardiovascular health models raised risk-ranking scores from 0.735 to 0.900 at ages 40–49 and 0.725 to 0.889 at 50–59.

Paper by Jackson Song, Xincheng Li, Yinan Zheng, +14 more authorsHongmei Jiang · Northwestern University
Blood cells and sparse proteins sit beneath an intact heart section, while clustered proteins sit beneath a heart section with gaps in its muscle.

Our read

Biomarkers
Studied:
Cohort study of CARDIA participants
What changed:
Adding 30 proteins increased cardiovascular disease risk discrimination C-index from 0.735 to 0.900
The catch:
The risk prediction models were evaluated only in participants from the CARDIA study

Editor of record: Boris Djordjevic · AgingResearch.News

Follow @agingbrief

CARDIA participants assessed at ages 40–49 (2,227 people) and 50–59 (2,707 people) were analyzed in an observational cohort study linking cardiovascular health at ages 18–35 to midlife blood proteins and later cardiovascular disease. Researchers combined health measures from six examinations.

Better young-adult cardiovascular health was associated with differences in 1,223 proteins in the younger group and 2,259 in the older group. Adding 30 selected proteins to models based on young-adult cardiovascular health improved how well the models ranked people by future disease risk. The groups had 154 and 132 cardiovascular events, respectively. Further analyses linked better early health to lower levels of certain proteins, which were themselves associated with lower disease risk. The authors proposed that some proteins partly connect early cardiovascular health with later disease.

Why it matters

Understanding the biological links between early-adult cardiovascular health and later disease could help clarify how cardiovascular risk develops with age.

Caveats

The protein pathway analysis relied on observational links that other health differences could have influenced. The conference abstract does not report whether the prediction models were tested in an independent population.

Comments

Be the first to comment.