Four Biomarkers Predict Mortality Better Than All Others Combined

Most people think of aging as a single process, but the biology tells a more fragmented story. A new study using the Lothian Birth Cohort 1936 — 861 individuals tracked from age 70 to 89 — set out to determine which biomarkers most reliably predict who will die sooner, and how much those biomarkers overlap or add unique information.

The headline finding is that GrimAge2, a second-generation epigenetic clock, outperformed nearly every other measure as a composite predictor of all-cause mortality. Organ-specific proteomic age gaps — particularly for the liver, immune system, heart, pancreas, and brain — were also strongly associated with mortality. Crucially, however, organ age gaps were only loosely correlated with each other within the same person, confirming that organs age at different rates independently.

When researchers stripped away redundancy and asked which biomarkers carry truly independent predictive signal, four rose to the top: white matter volume, total brain volume, walking time, and general cognitive function (g). These four alone explained 19% of mortality variance; all 17 remaining biomarkers combined added just 4% more. This underscores the outsized value of simple, measurable functional tests — particularly walking speed — relative to expensive molecular assays.

On the protein side, GDF15 — a well-established marker of cellular senescence and mitochondrial stress — was the single protein most predictive of mortality after lifestyle adjustment, surpassed only by GrimAge2 overall. Telomere attrition in white blood cells, once a cornerstone of aging research, showed no significant predictive relationship in this cohort.

Practically, these findings suggest that brain health metrics and walking performance deserve priority in any longevity monitoring panel, alongside epigenetic clocks. The study is observational, and the cohort is largely homogeneous, so generalizability needs further validation across diverse populations.