Blood-based age estimate predicted death risk in human cohorts
In UK Biobank, its model-fit score was 910749.6 versus 912914.9 for an existing age clock when assessed alone; lower scores indicate better fit.
Limit: Model-fit scores compared clocks but do not directly measure mortality prediction accuracy in new patients
Edited by Boris Djordjevic · AgingResearch.News
People in the UK Biobank and US National Health and Nutrition Examination Survey cohorts formed the basis of this observational analysis. Researchers built an age estimate from 11 routine blood measurements. Each measurement contributed years added or subtracted according to its association with death from any cause.
In UK Biobank, the estimate predicted deaths and the onset of eight common chronic diseases better than PhenoAge, another age clock, when assessed alone. After adjustment for other factors, PhenoAge predicted deaths better. An optimized version of the new clock outperformed PhenoAge for death prediction both alone and after adjustment. Comparisons for disease prediction after adjustment were mixed. In the US cohort, the new clock and PhenoAge performed similarly, while two other tested clocks predicted deaths best.
Why it matters
For biological-aging research, an interpretable clock could help clarify how individual blood measurements contribute to estimated age and mortality predictions.
Caveats
These observational associations do not show that changing the blood measurements would change lifespan. The abstract does not report cohort sizes, participants’ ages or follow-up duration.
The paper
University Clinic of Pulmonary and Allergic Diseases Golnik; NU B.V.
Mechanisms of Ageing and Development, Jul 12, 2026
