Biological age clocks vs pace-of-aging clocks: How DunedinPACE differs from PhenoAge and KDM

Q: I only have budget for one test per year. Which?

If you have never measured: start with PhenoAge from a blood panel — low cost, actionable. If you already have a PhenoAge baseline and are trialing a serious intervention: spend on DunedinPACE for follow-up.

A 42-year-old reader emailed us with two numbers that seemed to contradict each other. Her PhenoAge came back at 45 years — three years “older” than her birthday. But her DunedinPACE score was 1.15, meaning her biology was aging 1.15 years for every chronological year lived. “Am I aging fast, or am I three years behind?” she asked.

Both. They measure different things. The confusion is common in the longevity community, where readers assume a biological age algorithm and a pace-of-aging algorithm are two flavors of the same dish. They are not. One is a snapshot. The other is a velocity reading.

This article is a category-level map. It will not rewrite the single-clock explainers already on this site. Instead, it answers the meta-question: when do you want a point-in-time age number, and when do you want a rate?

What you’ll learn:

Why point-in-time biological age clocks and pace-of-aging clocks answer fundamentally different questions
How PhenoAge, KDM, Horvath 2013, and GrimAge fit into the first category — and DunedinPACE into the second
A decision framework for picking the right signal depending on your goal (benchmark, intervention tracking, medical risk, metabolic focus)
How to combine both signals for a more complete view of how you are aging

Quick answer

Point-in-time biological age clocks (PhenoAge, KDM, Horvath 2013, GrimAge) tell you how old you look biologically right now — a single number in years, compared to your chronological age.

Pace-of-aging clocks (DunedinPACE, DunedinPoAm) tell you how fast you are aging per chronological year — a rate, usually centered around 1.0, where 1.2 means you are aging 20% faster than an average adult and 0.8 means 20% slower.

Age clocks answer “where am I?”. Pace clocks answer “which direction and how quickly am I moving?”. They are complementary, not interchangeable.

Key facts

Point-in-time biological age clocks (PhenoAge, KDM, Horvath, GrimAge) answer “where am I biologically?” in years.
Pace-of-aging clocks (DunedinPACE) answer “how fast am I aging per chronological year?” as a rate.
The distinction is position vs velocity — both are needed for a complete picture, and neither substitutes for the other.
Choose a point-in-time clock for benchmarking and medical risk; choose a pace-of-aging clock for tracking interventions over months.
PhenoAge is the highest-value entry point because it runs on a standard blood panel. DunedinPACE is the highest-value addition once you have a protocol you want to evaluate.
Disagreement between an age clock and a pace clock is diagnostic, not contradictory — it tells you whether you are accumulating or shedding biological debt.

What a biological age clock actually measures

A biological age clock produces a point estimate in years. It compares your current biology — blood chemistry, DNA methylation patterns, or plasma proteins — to reference distributions, and returns the chronological age of a person whose biology looks like yours.

The four algorithms most longevity-literate readers encounter:

PhenoAge (Levine 2018). A blood-based clock trained on NHANES mortality data, integrating 9 clinical biomarkers — albumin, creatinine, glucose, hs-CRP, lymphocyte %, MCV, RDW, alkaline phosphatase, WBC — plus chronological age. See our deep dive on PhenoAge and KDM from blood tests for the formula.

KDM — Klemera-Doubal Method. A statistical approach that estimates biological age by minimizing residual variance between a panel of biomarkers and chronological age across a reference population. Both PhenoAge and KDM run on routine lab panels — no epigenetic sequencing.

Horvath 2013. The first widely adopted multi-tissue epigenetic clock. It reads DNA methylation at 353 CpG sites and returns a biological age correlated with chronological age at R ≥ 0.96. Trained to predict calendar age, not mortality.

GrimAge (Lu 2019). A mortality-trained second-generation epigenetic clock using DNA methylation surrogates of 7 plasma proteins plus smoking pack-years. Strongly associated with time-to-death, cardiovascular events, and cancer incidence. See our epigenetic clocks explained: Horvath, GrimAge, DunedinPACE guide.

What these four share: the output is a year number. A 42-year-old with PhenoAge 45 is biologically equivalent, by that algorithm, to the average 45-year-old in the training cohort — integrating everything that happened up to the blood draw.

What a pace-of-aging clock measures

A pace-of-aging clock does not return an age. It returns a rate — a unitless multiplier centered on 1.0 that describes how many years of biological aging are accumulating per chronological year right now.

The canonical example is DunedinPACE (Belsky 2022, eLife). Derived from the Dunedin Study — a New Zealand birth cohort followed since 1972 — with 19 organ-system biomarkers measured at ages 26, 32, 38, and 45. From those four longitudinal snapshots, researchers fit each participant’s personal rate of physiological decline, then trained a DNA methylation algorithm (173 CpG sites) to reproduce that rate from a single blood draw. The predecessor DunedinPoAm (Belsky 2020) used the first three waves; DunedinPACE added age 45 and improved test-retest reliability.

Output interpretation:

1.00 → aging at cohort average — one biological year per chronological year
1.20 → aging 20% faster than average
0.85 → aging 15% slower than average

A pace-of-aging clock is agnostic about where you started — it tells you what is happening now, which is why it responds more sensitively to recent interventions and is increasingly the preferred endpoint in geroscience trials. Our article on pace of aging and how fast your biological age is moving covers the mechanics in depth.

Why the distinction matters

Consider two 45-year-old men tested on the same day.

Marco — PhenoAge 50, DunedinPACE 0.88. Biology reads 5 years older than his birthday, but his pace is 12% below average. He accumulated damage in his 30s (poor sleep, untreated hypertension), but fixed those things over the last 18 months. The damage is still on the books; the acceleration has stopped. Behind, slowing down.

David — PhenoAge 43, DunedinPACE 1.22. Biology reads 2 years younger, but pace is 22% above average. His 30s were kind to him, but something shifted recently — weight gain, sleep fragmentation, rising inflammation. The snapshot still looks good because 18 months of acceleration has not yet reshaped the long-run average. Ahead, accelerating.

Who is in more trouble? Arguably David. A single point-in-time measurement would miss this entirely — that is why pace-of-aging clocks were developed. Conversely, relying only on DunedinPACE would miss Marco’s accumulated deficit.

These are position and velocity — you need both to know where the car is heading.

Head-to-head comparison

Dimension	Point-in-time age clocks (PhenoAge, KDM, Horvath, GrimAge)	Pace-of-aging clocks (DunedinPACE)
Unit	Years (biological age)	Unitless rate, centered on 1.0
Question answered	How old does your biology look right now?	How fast is your biology aging right now?
Best use case	Cross-sectional benchmarking, “where am I vs my birthday”	Tracking intervention effects over 6-24 months
Sensitivity to recent change	Low-to-moderate — integrates lifetime history	Higher — weighted toward current trajectory
Integrates historical damage	Yes — accumulated burden shows up	Partially — dominant signal is recent rate
Availability	PhenoAge/KDM from a standard blood panel; Horvath/GrimAge require DNA methylation array	Requires DNA methylation array (Illumina EPIC or 450K)
Typical cost	PhenoAge/KDM: free to ~$100 via blood work; epigenetic: $250-500	$250-500 (same assay as epigenetic age clocks)
Validation strength	Very strong for mortality (GrimAge, PhenoAge)	Strong for morbidity, mortality, and pace change
Used as clinical trial endpoint	Yes (mainly GrimAge, PhenoAge)	Yes (increasingly, due to sensitivity)

Notice what the table does not say: it does not say one clock is “better”. A velocity reading from a parked car is 0. An age reading of a new battery is low. Neither tells you the whole story.

When to use each type of clock

A decision framework based on what you are trying to learn:

One-off benchmark — “how am I doing for my age?” → Pick a point-in-time age clock. PhenoAge or KDM if you already get regular blood work; GrimAge if you can afford an epigenetic test and care most about mortality prediction.

Tracking whether a lifestyle or therapeutic intervention works → Pick a pace-of-aging clock (DunedinPACE). A PhenoAge delta of 1 year over 6 months is within noise; a DunedinPACE drop from 1.15 to 0.95 over the same window is a real signal.

Medical context — cardiovascular risk, cancer screening, primary prevention → Prefer GrimAge. Strongest mortality prediction among validated clocks. Pair with conventional risk scores (ASCVD, FRAX).

Metabolic-health focus — prediabetes, inflammation, cardiometabolic drift → Prefer PhenoAge. The 9 biomarkers are intrinsically metabolic and inflammatory, so you can see which component is driving your acceleration and target it. See how to lower your biological age.

Low-friction ongoing proxy without quarterly blood or saliva assays → You cannot replicate DunedinPACE from a wearable, but PhenoAge every 6 months plus continuous wearable signals produces a directionally correlated trend. See how SuperAge calculates biological age.

Can you combine them?

Yes — and for serious longevity tracking, you should. The two signals are complementary.

A principled combination:

Anchor your position — PhenoAge every 6-12 months (cost-effective entry point); annual GrimAge if budget allows.
Measure your velocity — DunedinPACE every 12-24 months; most useful when you have at least two readings to compare.
Use the pair to diagnose:
- Age above chronological, pace above 1.0 → behind and getting worse. Highest urgency.
- Age above chronological, pace below 1.0 → behind but recovering. Stay the course.
- Age below chronological, pace above 1.0 → ahead but decelerating. Investigate what changed.
- Age below chronological, pace below 1.0 → ahead and compounding. Protect the protocol.

An intervention can move the pace months before it moves the age snapshot. Conversely, a person with strong historical health can have a quietly deteriorating pace that the age clock does not yet see.

How SuperAge visualizes your trend over time

SuperAge computes biological age using PhenoAge and KDM — the two most validated blood-based algorithms — from labs you enter manually or import from a PDF. It does not run epigenetic assays, so it does not produce a DunedinPACE number.

What it does produce is a trend line: every PhenoAge and KDM calculation is stored and charted. The slope of that line is a blood-based pace-of-aging proxy — not identical to DunedinPACE, but directionally informative and far cheaper to run at high frequency. For a full guide on why tracking biological age as a trend beats any single score, see our dedicated explainer.

Combine the trend line with continuous wearable data — resting heart rate, HRV, sleep stages, VO₂max — and you get a personal health dashboard that stitches wearables and blood tests. Together they approximate what an epigenetic pace-of-aging clock tells you, affordable enough to actually repeat.

Download SuperAge on the App Store and start a biological-age trend line today.

Frequently asked questions

Is DunedinPACE more accurate than PhenoAge?

They measure different things, so “accuracy” is not the right frame. DunedinPACE is more reliable at detecting short-term changes in aging rate. PhenoAge is more accurate as a mortality-associated snapshot from blood work.

Can my PhenoAge go down if my DunedinPACE is above 1.0?

Short-term, yes — noise in a single blood panel can drop a PhenoAge number by a year while your true pace is above 1. Sustained, no — if DunedinPACE stays above 1 for several years, PhenoAge drifts upward to match.

Which one do longevity clinicians use in practice?

Mortality-focused clinics lean on GrimAge. Metabolic-focused practices lean on PhenoAge or KDM. Research-active longevity clinics increasingly use DunedinPACE as the primary intervention-response endpoint, with a point-in-time clock as anchor.

I only have budget for one test per year. Which?

If you have never measured: start with PhenoAge from a blood panel — low cost, actionable. If you already have a PhenoAge baseline and are trialing a serious intervention: spend on DunedinPACE for follow-up.

Do these clocks agree within a single person?

Loosely. GrimAge and PhenoAge correlate at r ≈ 0.5-0.7. Correlation between any single-age clock and DunedinPACE is lower because pace and position are conceptually distinct. Disagreement is information, not a bug.

Key takeaways

Point-in-time biological age clocks (PhenoAge, KDM, Horvath, GrimAge) answer “where am I biologically?” in years.
Pace-of-aging clocks (DunedinPACE) answer “how fast am I aging per chronological year?” as a rate.
The distinction is position vs velocity — both are needed for a complete picture, and neither substitutes for the other.
Choose a point-in-time clock for benchmarking and medical risk; choose a pace-of-aging clock for tracking interventions over months.
PhenoAge is the highest-value entry point because it runs on a standard blood panel. DunedinPACE is the highest-value addition once you have a protocol you want to evaluate.
Disagreement between an age clock and a pace clock is diagnostic, not contradictory — it tells you whether you are accumulating or shedding biological debt.

Start tracking the right signal

The most important decision in biological-age tracking is not which clock to buy — it is to start a trend line and stick to it. A single reading is a data point; ten readings over three years is a trajectory.

SuperAge gives you that trajectory: PhenoAge and KDM from your blood work, charted over time alongside your wearable data, for a directionally accurate pace-of-aging proxy without a $400 epigenetic assay.

Get SuperAge on the App Store — and the next time two numbers seem to contradict each other, you will know which question each is answering.

References

Horvath, S. (2013). DNA methylation age of human tissues and cell types. Genome Biology, 14(10), R115. https://doi.org/10.1186/gb-2013-14-10-r115
Levine, M. E., Lu, A. T., Quach, A., et al. (2018). An epigenetic biomarker of aging for lifespan and healthspan. Aging, 10(4), 573-591. https://doi.org/10.18632/aging.101414
Lu, A. T., Quach, A., Wilson, J. G., et al. (2019). DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging, 11(2), 303-327. https://doi.org/10.18632/aging.101684
Belsky, D. W., Caspi, A., Arseneault, L., et al. (2020). Quantification of the pace of biological aging in humans through a blood test, the DunedinPoAm DNA methylation algorithm. eLife, 9, e54870. https://doi.org/10.7554/eLife.54870
Belsky, D. W., Caspi, A., Corcoran, D. L., et al. (2022). DunedinPACE, a DNA methylation biomarker of the pace of aging. eLife, 11, e73420. https://doi.org/10.7554/eLife.73420
Klemera, P., & Doubal, S. (2006). A new approach to the concept and computation of biological age. Mechanisms of Ageing and Development, 127(3), 240-248. https://doi.org/10.1016/j.mad.2005.10.004

Last updated: 2026-04-15. This article is educational content and does not replace medical advice. Always consult a qualified professional before making decisions about testing, supplementation, or therapeutic interventions.