SNIPPET: DunedinPACE, a third-generation epigenetic clock measuring the pace of DNA methylation change, emerged as the strongest predictor of all-cause mortality among fourteen expert-consented biomarkers of aging in the BASE-II cohort (n=1,083, average follow-up 7.4 years). It outperformed inflammatory, functional, and physiological markers — and a minimal set of just three biomarkers including DunedinPACE predicted mortality nearly as well as all fourteen combined.

DunedinPACE Beats 13 Other Aging Biomarkers at Predicting Death: What the BASE-II Data Tells Us

THE PROTOHUMAN PERSPECTIVE#

The gap between lifespan and healthspan is the central problem of modern aging. We are living longer, but not necessarily better — and without a reliable way to measure biological aging, every longevity intervention is essentially flying blind. This study matters because it directly compares fourteen biomarkers that an expert panel already endorsed for use in intervention trials, and it tells us which ones actually predict the thing we care about most: staying alive.

DunedinPACE didn't just win. It won decisively. For anyone tracking their own biological age or designing interventions to slow it, this is the clearest signal yet that epigenetic pace — not grip strength, not CRP, not IGF-1 — is where your measurement dollars should go. The data also quietly killed several sacred cows. Blood pressure, muscle mass, gait speed, and the frailty phenotype failed to independently predict mortality in this cohort. That should make us uncomfortable, and it should change protocols.

The long game in longevity has always been about finding the right speedometer. This study suggests we may finally have one.

THE SCIENCE#

What DunedinPACE Actually Measures#

DunedinPACE is not an epigenetic clock in the traditional sense. First-generation clocks like the Horvath clock estimate your biological age at a single point in time — a snapshot. DunedinPACE, developed by Belsky, Caspi, and colleagues, measures something different: the rate at which you're aging right now ^[1]. It does this by analyzing DNA methylation patterns at specific CpG sites across the genome, calibrated against longitudinal data from the Dunedin birth cohort. A DunedinPACE score of 1.0 means you're aging at the expected rate. Below 1.0, you're aging slower. Above 1.0, faster.

This distinction matters enormously. Two people can have the same biological age estimate but very different trajectories. One is decelerating. The other is accelerating into decline. DunedinPACE captures that velocity.

The BASE-II Head-to-Head Comparison#

Vetter, Junge, Drevon, and colleagues used data from the Berlin Aging Study II — 1,083 adults aged 60–80 at baseline, followed for an average of 7.4 years (range 3.9–10.4 years) ^[2]. They tested fourteen biomarkers that had been specifically endorsed by an expert consensus panel for use in longevity intervention studies. These spanned four categories:

• Physiological: IGF-1, DNAmGDF15
• Inflammatory: High-sensitivity CRP, IL-6
• Functional: Muscle mass, muscle strength, hand grip strength, Timed-Up-and-Go, gait speed, standing balance, frailty phenotype, cognitive health, blood pressure
• Epigenetic: DunedinPACE

Cox proportional hazard regression, adjusted for age, sex, lifestyle factors, and genetic ancestry, identified five biomarkers that significantly predicted all-cause mortality: hand grip strength, IL-6, standing balance, cognitive health, and DunedinPACE. Of these, DunedinPACE emerged as the strongest and most consistent predictor.

But here's where it gets interesting — and where I push back slightly on the celebration. Nine of the fourteen biomarkers failed to predict mortality at all. CRP, gait speed, IGF-1, blood pressure, muscle mass, DNAmGDF15, the frailty phenotype, and Timed-Up-and-Go showed no statistically significant association. That's a lot of biomarkers that the longevity community treats as gospel, and the data told Vetter et al. they don't independently move the needle in this age group.

The Minimal Biomarker Set#

Feature selection analysis revealed that you don't need all fourteen markers. A minimal set of just three — muscle mass, standing balance, and DunedinPACE — predicted mortality with nearly the same accuracy as the full panel. This is a practical finding. It means a streamlined aging assessment could be built around one epigenetic test and two simple physical measurements. That's clinic-ready. That's scalable.

I'll note the irony: muscle mass on its own didn't predict mortality, but it contributed to the minimal set. The data tells me this is about combinatorial information — muscle mass adds signal when paired with DunedinPACE and balance, even if it can't carry the prediction alone.

Longitudinal Confirmation from InCHIANTI#

The BASE-II findings land alongside corroborating evidence from the InCHIANTI cohort. Kuo, Moore, Tanaka, and colleagues followed 699 adults for up to 24 years and found that longitudinal changes in epigenetic clocks — not just baseline measurements — independently predicted mortality ^[3]. The adjusted hazard ratio for the Hannum clock's rate of change was 1.12 (95% CI: 1.01–1.23). Faster acceleration across multiple clock generations predicted higher death risk, independent of where you started.

This is the piece that clinches it for me. Cross-sectional snapshots are useful. But the trajectory — how fast your epigenetic age is moving — carries independent, additional information. The InCHIANTI data suggests that even if your biological age looks acceptable today, a rapidly accelerating clock should raise alarms.

The Systems Age Approach#

Adding another dimension, Sehgal and colleagues developed a blood-based methylation test that estimates biological aging across eleven physiological systems simultaneously ^[4]. Published in Nature Aging, their Systems Age measure predicts mortality and health outcomes more precisely than existing single-output epigenetic clocks and reveals distinct organ-specific aging patterns. The key insight: your liver, your immune system, and your cardiovascular system may all be aging at different rates. A single composite score masks that heterogeneity.

This doesn't contradict DunedinPACE's performance — it extends it. The future likely involves both: DunedinPACE as the headline velocity metric, and system-specific clocks to tell you where the acceleration is happening.

COMPARISON TABLE#

THE PROTOCOL#

How to build an evidence-based biological age monitoring strategy based on the current data:

1. Get a DunedinPACE test as your baseline. Order through a provider offering Illumina EPIC methylation array analysis (TruDiagnostic, Elysium Health, or academic research programs). One blood draw. Your score gives you a pace-of-aging velocity — the single most predictive metric identified in the BASE-II study.

2. Measure your functional triad. Based on the minimal biomarker set, add three low-cost assessments: standing balance (single-leg stance, eyes closed, timed), hand grip strength (calibrated dynamometer, three attempts per hand, take the best), and body composition for muscle mass (DEXA scan preferred, bioimpedance acceptable). These are the physical markers that, combined with DunedinPACE, approach full-panel predictive power.

3. Test IL-6 alongside standard bloodwork. Add interleukin-6 to your next comprehensive metabolic panel. IL-6 was one of five significant mortality predictors and reflects systemic inflammatory burden — a driver of epigenetic aging acceleration through NF-κB pathway activation and its downstream effects on autophagy pathways and NAD+ synthesis.

4. Establish your re-testing cadence. The InCHIANTI data from Kuo et al. shows that changes in epigenetic age over time carry independent prognostic information ^[3]. A single measurement is valuable. Two measurements separated by 12–18 months tell you whether your interventions are working. Based on current evidence, annual DunedinPACE testing and quarterly functional assessments represent a reasonable monitoring protocol.

5. Intervene on modifiable inputs. DunedinPACE has been shown to respond to interventions in preliminary data. Caloric restriction, exercise protocols emphasizing both resistance training (for the muscle mass and grip strength components) and balance work (tai chi, single-leg exercises), and inflammation management through omega-3 supplementation (2–3g EPA/DHA daily) and sleep optimization (targeting 7–8 hours with consistent circadian timing) represent the highest-evidence interventions for slowing epigenetic pace. Monitor HRV as a low-cost daily proxy for autonomic and inflammatory status between formal testing windows.

6. Ignore the noise. The data is clear that CRP, IGF-1, blood pressure, gait speed, the frailty phenotype, and Timed-Up-and-Go did not independently predict mortality in this cohort. I'm not saying stop measuring blood pressure — it matters for cardiovascular management. But if you're trying to quantify your rate of aging, these markers aren't earning their place in the panel.

What is DunedinPACE and how does it differ from other epigenetic clocks?#

DunedinPACE is a third-generation DNA methylation biomarker that measures your current pace of aging rather than estimating a static biological age. Developed from longitudinal data in the Dunedin birth cohort by Belsky et al., it captures velocity — how fast you're accumulating age-related methylation changes right now ^[1]. First-generation clocks like Horvath give you a snapshot; DunedinPACE gives you a speedometer reading.

How was this study designed and who was included?#

The BASE-II study included 1,083 adults aged 60–80 at baseline, followed for an average of 7.4 years with a range of 3.9 to 10.4 years ^[2]. Cox proportional hazard regression models were adjusted for age, sex, lifestyle factors, and genetic ancestry. The fourteen biomarkers tested were selected from an expert consensus panel specifically for their potential use in longevity intervention studies.

Why did common biomarkers like CRP and blood pressure fail to predict mortality?#

The honest answer is we don't fully know, and I'd want to see this replicated in larger, more diverse cohorts. One possibility: in a 60–80-year-old cohort where many participants may already be treated for hypertension and inflammation, these markers lose their independent predictive signal. The adjustment for lifestyle factors and other covariates may also absorb variance that CRP and blood pressure would capture in a simpler model. The data doesn't mean these markers are useless — it means they didn't add independent mortality prediction beyond the other thirteen markers and covariates.

When should someone get their first epigenetic age test?#

Based on current evidence, testing becomes most informative from age 40 onward, when age-related methylation drift accelerates measurably. However, a baseline in your 30s provides a valuable reference point for future comparisons. The InCHIANTI longitudinal data emphasizes that serial measurements — not single tests — carry the strongest prognostic signal ^[3]. If you can only afford one test, make it DunedinPACE.

How can you slow down your epigenetic pace of aging?#

Preliminary data from intervention studies suggests that caloric restriction, structured exercise combining resistance and aerobic training, improved sleep quality, and anti-inflammatory dietary patterns may reduce DunedinPACE scores. Sehgal et al. noted in a preprint that Systems Age clocks respond to geroprotective interventions across 51 human studies ^[4]. Optimal dosing and protocols for epigenetic age reduction in humans are not yet fully established — but the signal is there. Start with what has the strongest evidence: consistent exercise, adequate sleep, and inflammation control.

VERDICT#

8.5/10. This is a well-designed head-to-head comparison that actually answers a question the longevity field has needed answered: among the biomarkers we've agreed to use, which ones predict death? The design was solid — adjusted models, cause-specific subgroup analyses, feature selection — and the sample size of 1,083 with 7.4 years of follow-up is respectable, though not enormous. The InCHIANTI longitudinal data from Nature Aging strengthens the overall case for epigenetic clocks as the leading aging biomarker class.

Where I dock points: this is a predominantly European cohort aged 60–80. I'd want to see DunedinPACE's dominance confirmed across diverse populations and younger age ranges before declaring total victory. The nine biomarkers that failed to predict mortality need replication — if CRP and gait speed truly don't predict death independently in older adults, that's a finding that should shake standard geriatric assessment. And frankly, the practical accessibility of methylation testing is still limited. DunedinPACE won the science. It hasn't yet won the clinic.

Still. This one actually moved me. The minimal three-biomarker set — DunedinPACE plus muscle mass plus standing balance — is elegant and actionable. That's where the field should focus.#