Which Aging Biomarkers Best Predict Who Dies Early? A 16-Year Study Answers
A 16-year mortality study pits proteomic organ clocks against epigenetic, brain, and physical biomarkers — revealing what actually predicts death.
Summary
Scientists tracked 861 older adults for 16 years to compare how well different aging biomarkers predicted death. Proteomic clocks estimating the biological age of organs like the liver, immune system, and heart were meaningfully linked to mortality risk. However, the epigenetic clock GrimAge2, brain volume measures, respiratory function, and cognitive scores performed equally well or better. Scanning nearly 10,000 blood proteins, the researchers also identified 368 proteins tied to mortality risk — with GDF15, WFDC2, and TIMP1 as the strongest individual signals. High-risk proteins were clustered around immune dysfunction, while protective proteins related to genomic stability. The study offers a rigorous head-to-head comparison of aging measurement tools, helping clarify which biomarkers deserve priority in longevity research and clinical practice.
Detailed Summary
Aging biomarkers promise to tell us how old our bodies truly are — and whether we are heading toward an early death. But dozens of competing tools now exist, from DNA methylation clocks to organ-specific protein panels, and no systematic comparison has established which are most useful for predicting mortality.
This study used the Lothian Birth Cohort 1936 (LBC1936), a well-characterized Scottish cohort of older adults born in 1936, following 861 participants over 16 years during which 444 died. Cox regression models benchmarked proteomic organ clocks — derived from plasma proteomics and designed to estimate the biological age of 11 specific organs — against a broad multimodal panel of established aging biomarkers including the epigenetic clock GrimAge2, telomere length, grip strength, walking speed, respiratory function, brain MRI volumes, and cognitive test scores.
Among the proteomic organ clocks, accelerated aging of the liver, immune system, and heart showed the strongest mortality associations, with hazard ratios per standard deviation of 1.43, 1.42, and 1.38, respectively. Yet GrimAge2, total brain volume, gray matter volume, respiratory function, and cognition performed equally well or better, with hazard ratios reaching 1.44 to 1.62 per standard deviation. This suggests that established biomarkers remain highly competitive and should not be displaced by organ clocks prematurely.
In a separate proteome-wide survival analysis across 9,703 plasma protein targets, 368 proteins were significantly associated with mortality. GDF15, WFDC2, and TIMP1 emerged as the top individual predictors, all reflecting inflammatory and tissue-remodeling pathways. Protective proteins were enriched for roles in DNA repair and cellular maintenance.
The findings underscore that no single biomarker class dominates. Brain health, lung function, epigenetic age, and immune-related proteins each capture distinct mortality-relevant biology, making multimodal assessment the strongest approach for predicting longevity outcomes.
Key Findings
- Epigenetic clock GrimAge2 and brain volume measures outperformed most proteomic organ clocks in predicting 16-year mortality.
- Accelerated liver, immune, and heart aging (proteomic clocks) were the strongest organ-specific protein predictors of early death.
- Of 9,703 plasma proteins analyzed, GDF15, WFDC2, and TIMP1 were the top individual mortality-risk proteins.
- High-risk proteins clustered around immune dysfunction; protective proteins were linked to genomic stability and cellular maintenance.
- Respiratory function and cognitive scores rivaled the best molecular biomarkers, reinforcing their value as mortality predictors.
Methodology
Cox regression survival analysis was applied to the Lothian Birth Cohort 1936 (n=861; 444 deaths over 16 years), comparing proteomic organ clocks against multimodal aging biomarkers. A separate proteome-wide survival scan examined 9,703 plasma protein targets for mortality associations. All analyses adjusted for relevant covariates in this well-characterized longitudinal cohort.
Study Limitations
Summary is based on the abstract only, as the full text is not open access. The cohort is composed entirely of Scottish adults born in 1936, limiting generalizability across ethnicities, ages, and geographies. Proteomic organ clocks and some biomarkers may carry measurement-specific noise that affects the hazard ratio comparisons.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: