Citation-backed associations between biomarkers and aging-related outcomes, with an intervention-modifiability layer.
Each row is a biomarker, grouped by physiological domain. The five outcome columns show how strongly that biomarker predicts each aging-related outcome — all-cause mortality (the highlighted anchor column), cardiovascular disease, cancer, dementia, and frailty. Every cell is a hazard ratio standardized to a common scale, HR per +1 standard deviation of the biomarker, so markers measured in different units can be compared directly. Cell colour encodes effect size and direction — blue = protective, red = risk — and the cell border encodes evidence quality (A > B > C). Exposures that are inherently categorical (genotypes, calcium scores) cannot be standardized; they are shown with their native HR and visually flagged, never ranked against the per-SD scale. The right-hand Standardization column tags those biomarkers with a short reason — genotype, U-shaped, non-normal, clinical cutpoint — hover it for the full explanation. Hover any cell for the source study; click a biomarker name for its full profile.
Prediction alone does not tell you what to do. The rightmost column rates how much each biomarker moves in response to the best-available intervention — high, moderate, low, or fixed (genetically or structurally set). A marker that is both highly predictive and highly modifiable is a priority target (marked ★): worth measuring and worth acting on. High predictive weight with low modifiability is for risk stratification; high modifiability with little predictive weight is largely noise. Sort by Priority to bring the priority targets to the top of each domain.