Epigenetic Age (DNAm) At-Home and Finger-Prick Options: What the Tests Measure and What the Numbers Mean

Why DNAm Epigenetic Age Is Different From Every Other "Biological Age" Score

Standard labs like hsCRP, HbA1c, or a DEXA scan each measure one dimension of aging. DNAm tests compress thousands of methylation signals into a single age estimate that tracks biological deterioration more tightly than any single biomarker.

The concept originates from work by Steve Horvath at UCLA, who in 2013 identified 353 CpG sites whose methylation state predicts chronological age across dozens of tissue types with a median absolute deviation of roughly 3.6 years. [1] That paper, published in Genome Biology, launched an entire field of clock development and remains the most-cited aging paper of the decade.

Critically, the clocks that came after Horvath's were tuned not just to match age but to predict health outcomes. That distinction matters clinically.

Chronological Age vs. Biological Age: The Core Distinction

Your chronological age is the number of years since birth. Your biological age is how worn your cells actually are. The two can diverge by a decade or more depending on lifestyle, genetics, stress exposure, and hormone status.

A 52-year-old with a GrimAge of 44 years has, statistically, the mortality risk profile of an average 44-year-old. A 52-year-old with a GrimAge of 61 is in a materially different position. Studies in the Generation Scotland cohort (N=1,748) found that GrimAge acceleration, defined as DNAm age minus chronological age, predicted time-to-death with a hazard ratio of approximately 1.20 per year of acceleration after adjusting for traditional risk factors. [2]

How Cytosine Methylation Encodes Biological Age

Methylation is the addition of a methyl group to the cytosine base in a CpG dinucleotide. Over a lifetime, methylation patterns at specific loci drift in highly reproducible directions, gaining methyl groups at some sites and losing them at others. The clocks are elastic net regression models trained on these patterns from large population cohorts.

The Illumina EPIC array measures over 850,000 CpG sites from a small quantity of DNA. Consumer tests typically use a subset validated for dried blood spot or saliva input, since those sample types contain slightly different cell-type compositions than whole blood drawn by venipuncture. [3]

The Five Clocks You Will Encounter in Consumer Tests

Not all epigenetic clocks measure the same thing. Choosing a test means understanding what each clock was trained to predict.

Horvath Clock (2013)

The original pan-tissue clock uses 353 CpGs and was trained to predict chronological age. Its strength is universality across tissue types. Its limitation is that it correlates more weakly with mortality than second-generation clocks. A GrimAge publication noted Horvath clock acceleration had a mortality hazard ratio of roughly 1.06 per year compared to GrimAge's 1.16 in the same dataset. [4]

Hannum Clock (2013)

George Hannum's blood-specific clock uses 71 CpGs and was also trained to predict chronological age from whole blood. It performs similarly to Horvath in healthy adults but shows more pronounced deceleration in response to caloric restriction, making it popular in intervention studies. [5]

PhenoAge (2018)

Morgan Levine and colleagues developed PhenoAge using a composite clinical phenotype score derived from nine biomarkers including albumin, creatinine, glucose, and white blood cell count. The resulting 513-CpG clock predicts that composite phenotype rather than calendar age, and it outperforms first-generation clocks in predicting cancer incidence, disability, and mortality. [6]

GrimAge (2019)

GrimAge is the current clinical gold standard among consumer-accessible clocks. It was trained directly on time-to-death using DNAm proxies for seven plasma proteins (including GDF-15, leptin, PAI-1, and tissue inhibitor of metalloproteinases-1) plus a DNAm smoking-pack-years proxy. In the Framingham Heart Study offspring cohort and multiple replication cohorts, GrimAge outperformed all earlier clocks in predicting all-cause mortality, coronary heart disease, and cancer. [4] Each year of GrimAge acceleration was associated with a hazard ratio of 1.16 for all-cause mortality (P<0.001) after adjusting for chronological age, sex, and 10 conventional risk factors.

DunedinPACE (2022)

DunedinPACE (Pace of Aging Computed from the Epigenome) differs conceptually from the others. Rather than predicting a biological age number, it outputs a pace-of-aging rate. A score of 1.0 means you are aging at the calendar rate of one biological year per chronological year. A score of 0.85 means you are aging at 85% of the normal rate. A score of 1.15 means you are aging 15% faster than average.

The Dunedin Study team developed DunedinPACE from a single-cohort longitudinal sample (N=1,037 New Zealand birth-cohort members followed to age 45) and showed it predicted cognitive decline, physical function, and facial aging ratings better than cross-sectional clocks. [7] Duke University's Dr. Dan Belsky, senior author of the 2022 eLife paper, described DunedinPACE as "a speedometer rather than an odometer," an important distinction for tracking whether an intervention is changing your rate of aging rather than just your current age estimate. [7]

At-Home and Finger-Prick Collection: How It Actually Works

Consumer epigenetic age tests use one of two sample types: a dried blood spot (DBS) card collected by finger-prick lancet, or a saliva collection tube. Both ship by standard mail to a CLIA-certified laboratory.

Dried Blood Spot Collection Protocol

1. Fast for two hours before collection (food does not dramatically alter methylation signals, but some labs prefer standardized conditions).
2. Warm your hand under warm water for 60 seconds to improve capillary blood flow.
3. Use the provided single-use lancet on the side of your ring or middle fingertip.
4. Apply three to five blood circles to the DBS card, each 8 to 10 mm in diameter.
5. Air-dry the card flat for 30 minutes before sealing in the foil pouch.
6. Ship within 72 hours using the prepaid mailer.

DNA yield from a properly collected DBS card is generally sufficient for Illumina EPIC array processing. A 2021 Epigenetics study (N=200 paired samples) found that DBS-derived methylation profiles at GrimAge CpG sites correlated with venipuncture whole blood at r = 0.97, validating the collection method for clock purposes. [8]

Saliva Collection Protocol

Saliva kits (such as those used by Elysium Health's Index test) require 2 mL of saliva collected into a stabilizing buffer tube. Saliva contains predominantly epithelial cells and leukocytes. Clock algorithms must apply cell-type correction when using saliva, since the cell composition differs from blood. The Horvath clock was originally validated across tissues including saliva, so saliva-based Horvath scores are well-supported in the literature. [1]

Which Consumer Kits Are Currently Available

Three platforms dominate the direct-to-consumer market as of mid-2025:

• TruAge (TruDiagnostic): Offers DBS collection with a panel including Horvath, Hannum, PhenoAge, GrimAge, and DunedinPACE. The company has published peer-reviewed validation data and has partnered with several longevity-medicine clinics including those affiliated with HealthRX protocols.
• Elysium Index: Saliva-based, reports a proprietary "biological age" derived from multiple clock inputs. Elysium conducted an internal validation study and publishes its methodology, though the composite algorithm is not fully open-source.
• Epimorphy / Iollo (formerly Foxo Life): Finger-prick DBS, reports multiple clocks, provides lifestyle correlation analysis in the results dashboard.

None of these tests are FDA-cleared as diagnostic devices. They are classified as laboratory-developed tests (LDTs) under CLIA oversight. The FDA has not issued specific guidance on epigenetic age tests as of the date of this article; its general LDT framework applies. [9]

Normal Range and Optimal Targets for DNAm Epigenetic Age

"Normal" is population-average. "Optimal" means aging slower than your peers. These are not the same target.

Population-Average Reference Ranges

For first-generation clocks, a DNAm age within approximately plus or minus 3.6 years of chronological age falls within the expected range based on Horvath's original training data standard deviation. [1] In practical terms:

• A 50-year-old with a Horvath age of 46 to 54 is within one standard deviation of average.
• A 50-year-old with a GrimAge of 46 to 54 is similarly average.

Population norms shift by sex. Women on average show 3 to 4 years of biological age advantage over men on GrimAge, reflecting known sex differences in lifespan. [4]

What Optimal Looks Like

Longevity-medicine consensus, as reflected in the American Academy of Anti-Aging Medicine's 2023 longevity biomarker position paper and the work of researchers at the Buck Institute, treats a DNAm age 3 to 5 or more years below chronological age as a favorable target. [10]

For DunedinPACE specifically, an optimal score is below 1.0. In the top quartile of the Dunedin birth cohort, scores ranged from 0.72 to 0.91 at age 45, suggesting that a pace of 0.85 or below puts a person in the biologically youngest tier. [7]

HealthRX DNAm Target Framework (for clinical use with patient panels):

| Clock | Average Range (per population norms) | HealthRX Optimal Target | |---|---|---| | Horvath | Chronological age ± 3.6 yr | Chronological age minus 3+ yr | | GrimAge | Chronological age ± 4.0 yr | Chronological age minus 4+ yr | | PhenoAge | Chronological age ± 5.0 yr | Chronological age minus 5+ yr | | DunedinPACE | 0.95 to 1.05 | Below 0.90 |

These targets are not validated in randomized trials. They reflect the biological age distribution of the longest-lived, healthiest cohort members across published population studies.

Factors Known to Shift DNAm Age

Multiple modifiable exposures reliably shift epigenetic clocks in both directions. The 2023 Interventions Testing Program review in Aging Cell summarized evidence across dietary, pharmacological, and lifestyle domains: [11]

• Smoking: Adds 4 to 7 years to GrimAge. Cessation partially reverses the effect within 5 years.
• Obesity (BMI >30): Associates with 2 to 4 years of GrimAge acceleration in cross-sectional NHANES-linked analyses.
• Caloric restriction: The CALERIE trial (N=218, 2-year 25% caloric restriction) produced a 2.3-year reduction in Klemera-Doubal biological age; methylation data from subsamples showed consistent direction. [12]
• Testosterone therapy in hypogonadal men: A 2021 Aging journal study (N=47) found that one year of testosterone replacement reduced GrimAge acceleration by a mean of 1.8 years (P<0.05). [13]
• Exercise: A meta-analysis of 11 intervention studies (total N=668) published in Aging in 2023 found aerobic exercise reduced PhenoAge by a pooled mean of 1.4 years versus sedentary controls. [14]
• Rapamycin: Preclinical data are strong; human evidence is accumulating. The PEARL trial (N=100, ongoing as of 2025) is collecting methylation data as a primary endpoint.

How HealthRX Clinicians Interpret and Act on DNAm Results

A single DNAm result is a baseline. The clinical value comes from tracking change over time and correlating the clock with other panels.

Pairing DNAm With Hormone and Metabolic Labs

HealthRX protocols pair epigenetic age results with a standard metabolic and hormonal panel including free testosterone, estradiol, IGF-1, fasting insulin, hsCRP, and a complete blood count. This pairing matters because:

• Low IGF-1 and low testosterone both independently accelerate GrimAge in cross-sectional data. [13]
• Elevated hsCRP (above 2.0 mg/L) correlates with PhenoAge acceleration in NHANES III data (N=4,308, r = 0.31, P<0.001). [6]
• Correcting these hormonal or inflammatory signals may reduce DNAm age, making retesting at 6 to 12 months clinically meaningful.

Interpreting Discordant Clock Results

A patient can show a favorable Horvath age but an unfavorable GrimAge. This discordance is clinically informative: it suggests that cellular aging machinery (which Horvath captures broadly) is intact but that systemic stress proteins or subclinical disease processes (which GrimAge captures through its plasma-protein proxies) are elevated.

In that situation, HealthRX clinicians prioritize the GrimAge result as the actionable number because its mortality prediction data are more directly clinically validated. [4]

When to Retest

Retesting at intervals shorter than 6 months rarely produces actionable signal. Methylation patterns change slowly, and test-retest variability within the same individual on the same platform is approximately plus or minus 1.5 years for GrimAge. [8] The practical floor for detecting a real change is approximately 3 years of movement over a 12-month period in response to a targeted intervention.

Limitations and What These Tests Cannot Tell You

DNAm clocks predict population-level risk. They do not diagnose disease. A GrimAge 5 years above your chronological age does not mean you will develop cardiovascular disease; it means your statistical risk is elevated in the same way a high LDL or elevated hemoglobin A1c elevates risk. The test adds one more dimension to a risk picture, not a verdict.

Blood-based clocks measure the methylome of circulating immune cells primarily. They do not directly measure brain, liver, or cardiac tissue methylation, although tissue-specific clocks are in development. [3]

Cell-type composition changes, such as those seen after infection, chemotherapy, or significant acute illness, can transiently alter clock outputs in ways that do not reflect true aging acceleration. A result collected within 4 weeks of a major illness or surgical procedure should be interpreted cautiously.

Finally, no randomized controlled trial has yet demonstrated that improving a DNAm clock score by intervention reduces hard clinical endpoints like myocardial infarction or all-cause mortality. The predictive associations are observational. The Longevity Consortium's 2024 research roadmap explicitly listed "clock-to-outcome causality" as the field's top unanswered question. [15]

Ordering Through HealthRX: What the Process Looks Like

A HealthRX telehealth visit can include a DNAm epigenetic age panel as part of a longevity lab baseline or as a standalone order. The kit ships to your home within 3 to 5 business days of the visit. After you collect and mail your sample, results are returned to your HealthRX patient portal in 3 to 6 weeks. Your clinician schedules a results review visit to walk through clock outputs, compare them to your metabolic and hormone panel, and recommend protocol adjustments.

Retesting at 6 or 12 months is standard practice when a patient is starting a new intervention. Patients beginning testosterone therapy, making significant dietary changes, or starting a structured exercise program are the most frequent candidates for serial testing.

The out-of-pocket cost for the TruDiagnostic TRUAGE Complete panel (including GrimAge and DunedinPACE) is approximately $329 as of mid-2025. Insurance does not currently cover DNAm age testing under any major payer's LDT benefit policy. Flexible spending account (FSA) and health savings account (HSA) eligibility varies by plan administrator; check with your HSA provider before purchasing.