Systems Age Maps 11 Body Clocks From One Blood Test

One biological age can hide a lot

Two people with the same overall biological-age score may have very different risk profiles: one may show more cardiovascular aging, another more metabolic or immune aging. Systems Age, published in Nature Aging in 2025, tries to expose that difference. A single blood DNA-methylation assay generates estimates for 11 physiological systems: Heart, Lung, Kidney, Liver, Brain, Immune, Inflammatory, Blood, Musculoskeletal, Hormone, and Metabolic.

That is a useful scientific shift. Aging does not arrive as a synchronized marching band. Blood vessels, kidneys, muscle, and immune function can decline at different rates within the same person. But the test is a risk-measurement framework, not an organ scan, diagnosis, or personalized treatment algorithm. It does not show that lowering a system-age score improves health, or that a high score tells a clinician which supplement, drug, or screening schedule will help.

How the researchers built the clocks

The researchers combined blood DNA methylation with clinical biomarkers, functional measures, and mortality-related information to train separate system-level scores. Training data came from the Health and Retirement Study and the Framingham Heart Study. Most prospective outcome validation came from several subsets of the Women's Health Initiative (WHI), totaling 5,129 women. The team also tested the framework in the Baltimore Longitudinal Study of Aging, the Swedish Adoption/Twin Study of Aging, and the dementia-enriched Alzheimer's Disease Neuroimaging Initiative.

The method differs from older epigenetic clocks that compress methylation patterns into one age estimate or one pace-of-aging measure. DunedinPACE, for example, estimates the pace of biological aging and predicted later morbidity, disability, and mortality in its validation work. Systems Age extends that basic idea by asking whether the blood signal can be separated into several physiologic domains.

That separation is plausible, but it is not literal localization. A blood sample contains signals from circulating cells and from exposures that affect the whole body. A high Kidney score may reflect biology associated with kidney aging; it does not prove that the kidney is the source of the methylation pattern or that kidney disease is present.

Where the system scores performed best

In the WHI outcome analyses, several scores predicted outcomes that matched their assigned systems. The Heart score predicted myocardial infarction with a pooled hazard ratio of 1.36. The Kidney score predicted myocardial infarction with a hazard ratio of 1.20 and mortality with a beta of 1.89. The Brain, Blood, and Liver scores predicted mortality, with reported betas of 1.82, 1.58, and 1.64, respectively. The Hormone score predicted thyroid disease (beta 1.10), the Metabolic score predicted diabetes (odds ratio 1.08), and the Musculoskeletal score predicted physical-function decline (hazard ratio 1.13).

These are prospective associations, which gives them more practical weight than a one-time correlation with a lab value. The paper also reported that some matched system-outcome associations outperformed whole-body comparators such as Horvath's clock and DunedinPACE, particularly the Heart score for myocardial infarction. Still, the publication details summarized here do not provide confidence intervals for those estimates, absolute event rates, or enough information to convert them into an individual's expected risk. A hazard ratio or odds ratio can be statistically interesting while changing personal decision-making very little.

Independent research supports the broader premise that blood methylation tracks health-relevant aging biology. A 2022 study of kidney-related aging linked blood-based epigenetic-age and mortality signatures to low renal function in samples totaling as many as 9,688 participants. That finding supports the presence of renal-aging signal in blood, though it does not validate any particular Kidney score.

Likewise, a nested case-control study of 132 incident type 2 diabetes cases and 132 matched controls found methylation changes and accelerated epigenetic aging before diabetes onset. Systems Age's metabolic result therefore fits an existing pattern: blood methylation can associate with future metabolic disease. It does not establish that methylation changes cause diabetes or provide a tested target for prevention.

Why a blood test can reflect more than blood

Epigenetic clocks respond to many exposures that affect long-term health. In a 2024 analysis of 3,990 postmenopausal women in WHI, Biemans and colleagues examined associations between food groups and PhenoAge acceleration. In older U.S. adults, Mi and colleagues linked higher blood cadmium concentrations with faster epigenetic age acceleration. A longitudinal cohort study of 359 people also associated lower childhood socioeconomic status with higher GrimAge acceleration and DunedinPACE at about age 50.

These studies help explain why methylation clocks can capture accumulated biology. Diet, toxic exposures, chronic stress, and social conditions influence health over decades. But every one of these findings is observational. People with higher cadmium exposure, different diets, or socioeconomic disadvantage also differ in many other ways that can affect methylation and disease risk. The clock may record part of that history without identifying the causal route.

Brain-related findings require the same restraint. A Framingham Heart Study analysis associated DNA-methylation age acceleration with digital clock-drawing performance, while a meta-analysis of schizophrenia found that blood-based epigenetic aging differed by age, sex, illness duration, and polygenic risk. These results show that methylation aging measures can relate to neurologic and cognitive phenotypes. They do not establish that a blood test can cleanly measure the biological age of the brain.

The weak spots matter as much as the wins

Systems Age did not deliver equally useful results across all 11 domains. In the WHI analyses, the Brain score did not significantly predict cognitive-function decline, and the Inflammation score did not significantly predict arthritis. Those null matched results matter because they test the central promise of the framework: whether a score assigned to a system predicts a relevant outcome in that system. The answer is mixed, not uniform.

Generalizability also remains unsettled. The main prospective evidence came largely from older women in WHI. The SATSA sample consisted of twins, and ADNI included older adults enriched for dementia-related research. Those cohorts are valuable, but they are not a broad mixed-sex, mixed-ancestry clinical population.

Technical reliability is another open question. The study checked duplicate-sample reliability in 36 whole-blood samples from a GEO dataset. The available publication details do not clearly report the intraclass correlation coefficients, cross-platform performance, batch sensitivity, or how stable each score remains when the same person gives blood at different times. That is a practical problem. A score cannot guide longitudinal self-tracking if ordinary laboratory variation can mimic biological change.

A predictive biomarker can be scientifically informative without being ready to steer treatment.

What biohackers should do with this information

Systems Age is promising for research and cautious hypothesis generation. A repeated pattern—such as a persistently elevated metabolic score alongside rising glucose, triglycerides, waist circumference, or blood pressure—could justify a closer look at established risk markers and standard medical care. The methylation score should not replace those measures.

Do not use a high Heart, Kidney, Brain, or Hormone score as a reason to start medication, buy a supplement stack, or intensify screening without conventional clinical evidence. The study did not test interventions based on score results, compare score-guided care with ordinary care, or show fewer illnesses or deaths after a score changed.

The next useful studies are straightforward: repeated-draw reliability testing, cross-platform validation, external validation in broader populations, comparisons with standard clinical risk models, and trials testing whether acting on a high system-age score improves decisions or patient outcomes. Until then, Systems Age offers a more detailed map of aging-related methylation patterns from one blood draw. It does not yet tell you which road to take.