DunedinPACE vs. GrimAge: Best Biological Age Test?

Two epigenetic age tests are dominating the longevity conversation right now: GrimAge and DunedinPACE. Both measure DNA methylation. Both come from serious academic research. Both are available from the same labs.

But they answer fundamentally different questions — and if you're trying to figure out whether your current protocol is actually working, the distinction matters more than most review articles let on.

What Epigenetic Clocks Actually Measure

Your DNA sequence stays the same throughout your life. But the chemical tags attached to your DNA — methyl groups bonded to cytosine bases at locations called CpG sites — change continuously in response to age, lifestyle, and environment. These patterns are called the methylome.

In 2013, Steve Horvath at UCLA made a landmark discovery: certain methylation patterns across hundreds of CpG sites correlate so reliably with chronological age that an algorithm could estimate how old a person is from a blood sample alone. This became the first widely validated epigenetic clock.

The insight that followed was even more interesting. If methylation ages predictably, then deviations from that expected pattern — an epigenome that looks older or younger than the calendar says it should — might tell you something meaningful about health and longevity.

That hypothesis has held up. The subsequent decade of research produced increasingly sophisticated clocks that go beyond "how old does your epigenome look?" and start answering "what does that mean for your health?" GrimAge and DunedinPACE represent two distinct approaches to that harder question.

Related: Try our Biological Age Calculator to test this yourself. Also worth reading: VO2 Max by Age: What's Excellent at 30, 40, and 50 and our The Complete Guide to Supplement Tracking.

GrimAge: Your Epigenetic Snapshot

GrimAge was developed by Steve Horvath's group at UCLA and published in 2019 in Aging (Albany NY). It represents a significant advance over the first-generation Horvath and Hannum clocks because it was explicitly designed to predict lifespan and healthspan — not just to estimate chronological age.

Where earlier clocks asked "how old does your epigenome look?", GrimAge asked "which methylation patterns best predict who is going to get sick or die first?"

To build it, researchers trained the algorithm on methylation surrogates for several plasma proteins linked to mortality and smoking history, combined with DNA methylation data from long-term cohort studies with decades of follow-up. The result is a composite score expressed as an "epigenetic age" in years.

What GrimAge predicts

GrimAge has been validated across multiple independent cohorts for:

• All-cause mortality
• Time to cancer diagnosis
• Cardiovascular disease risk
• Coronary heart disease
• Physical disability

The "GrimAge acceleration" metric — how many years older or younger your GrimAge is relative to your chronological age — is the number most people focus on. A GrimAge acceleration of +5 means your epigenome looks five years older than your passport. A GrimAge acceleration of -3 means it looks three years younger.

What GrimAge is sensitive to

GrimAge responds to chronic exposures and long-term lifestyle patterns. Factors with documented associations include:

• Smoking history — incorporated directly into the algorithm; smoking is one of the strongest accelerators
• Body mass index and adiposity
• Chronic stress and inflammation
• Physical activity levels over time
• Diet quality (Mediterranean-pattern diets are associated with lower GrimAge)

The limitation: GrimAge changes slowly. A three-month intervention — even a well-designed one — may produce little to no detectable change in GrimAge. It's measuring accumulated biology, not recent trends.

DunedinPACE: Your Aging Rate Right Now

DunedinPACE was developed by Daniel Belsky and colleagues at Duke University and Columbia, based on the Dunedin cohort — a longitudinal study tracking a group of New Zealanders from birth. It was published in eLife in 2022.

The conceptual shift from GrimAge is significant. DunedinPACE doesn't estimate how old your epigenome looks. It measures the rate at which you're currently aging — your pace of aging normalized against a reference population.

The score is expressed as a decimal around 1.0:

• 1.0 = aging at the average pace for your age group
• 0.8 = aging about 20% slower than average
• 1.2 = aging about 20% faster than average

The researchers derived this by identifying methylation patterns that tracked with concurrent change across 19 independent biomarkers of organ-system aging — lung function, kidney function, metabolic markers, immune markers, cardiovascular measures — in the same individuals measured repeatedly over decades.

The key advantage: responsiveness

Because DunedinPACE reflects the current pace of biological change rather than accumulated history, it is more sensitive to recent lifestyle interventions. Research has already shown DunedinPACE responding to:

• Caloric restriction — the CALERIE trial found significant DunedinPACE reduction with two years of sustained caloric restriction
• Exercise protocols — higher cardiorespiratory fitness is associated with lower DunedinPACE
• Sleep quality — short sleep and sleep disruption correlate with higher pace scores
• Psychological stress — chronic stress and adversity are associated with faster pace

This makes DunedinPACE more useful as an intervention tracking tool. If you change your protocol and retest six to twelve months later, DunedinPACE is more likely to reflect that change than GrimAge.

The Core Distinction: Snapshot vs. Speedometer

The clearest way to understand the difference is a car analogy.

GrimAge is the odometer. It tells you how many miles are already on the engine. High GrimAge acceleration means you've accumulated more biological wear than the calendar would predict. That's important, meaningful information about where you stand.

DunedinPACE is the speedometer. It tells you how fast you're traveling right now. A high DunedinPACE score means you're burning through biological resources quickly at this moment — regardless of what the odometer reads. A low DunedinPACE means you've slowed down.

You need both readings to understand your situation fully. A 42-year-old with a GrimAge of 47 and a DunedinPACE of 0.75 has accumulated meaningful biological aging, but is currently decelerating. A 38-year-old with a GrimAge of 36 and a DunedinPACE of 1.35 looks young today but is heading in the wrong direction fast.

Side-by-Side Comparison

How to Get Tested

Both GrimAge and DunedinPACE are available through direct-to-consumer labs that use dried blood spot or saliva collection kits — no clinic visit required.

TruDiagnostic (TruAge) is the most widely used option and offers both GrimAge and DunedinPACE in a single panel called TruAge Complete. The collection kit is mailed to you, you collect a small dried blood spot sample at home, and results come back through their portal with interpretation guidance. They're a legitimate clinical lab with peer-reviewed validation work.

Elysium Health (Index) offers an epigenetic age test based on their proprietary algorithm, which is derived from the Horvath clock family. This does not include DunedinPACE, making it less useful for intervention tracking.

myDNAge offers a Horvath-based clock test at a lower price point, without GrimAge or DunedinPACE.

If tracking interventions is your primary goal, TruDiagnostic's panel covering both GrimAge and DunedinPACE is the most information-dense option currently available to consumers.

Testing protocol recommendations

• Fasting state: Most labs recommend collecting samples in a fasted state (8-12 hours) for consistency
• Baseline first: Establish a baseline before starting any new intervention you plan to measure
• Minimum interval: Six months between tests is the practical minimum for detecting meaningful change; twelve months is more reliable
• Control variables: Avoid major illness, alcohol in the preceding few days, and extreme physical stress around test collection — these create noise in single readings
• Test twice, then decide: A single test is a snapshot. Two tests bracketing a protocol change tell you whether the protocol worked