PhenoAge (Phenotypic Age)
A research score that combines nine routine blood tests and your age into the age your body's chemistry most resembles.
AI-verified against cited sources. Awaiting expert reviewAlso called PhenoAge
Check your result
The usual range for PhenoAge depends on age. Add your age above to see where you stand.
Nothing you type here leaves your browser.
What it is
PhenoAge, also called Phenotypic Age, was published in 2018 by Levine and colleagues. It combines your age with nine routine blood results: albumin, creatinine, glucose, C-reactive protein (CRP), lymphocyte percent, mean cell volume (MCV), red cell distribution width (RDW), alkaline phosphatase (ALP) and white blood cell count. The equation was built to estimate the 10-year risk of dying in a large US health survey. That risk is then turned into an age: the age at which an average person in that survey had the same risk. Labs don't report PhenoAge. It is calculated from your results.
Why it matters
In US health-survey data, a person whose PhenoAge was one year higher than expected for their age had about 9% higher risk of dying during follow-up. People with more chronic conditions also tended to have a higher PhenoAge. These are patterns across large groups, not a forecast for one person. PhenoAge is a research tool. No guideline sets a normal range or a target for it, and it can't tell you why the number is what it is. Many people use it to follow a trend and to see which of their routine results are pulling the number up or down.
If it’s low
A PhenoAge below your calendar age means your results together look more like those of a younger person in the reference survey. In group studies this pattern went with lower risk on average. It doesn't prove your body is younger, and it is not a reason to skip routine checkups.
Common causes
- Results in the lower part of the usual range for the heavily weighted inputs, such as RDW, glucose and CRP, together with higher albumin
- A very low CRP from a high-sensitivity test, which the original equation wasn't built to handle and which can lower the score by a year or more compared with a CRP of 3 mg/L
- Low creatinine from low muscle mass, which lowers the score without reflecting better health
- A unit or rounding error when the score was calculated
What can help
- Keep up regular checkups and the screening your clinician recommends, since a low number doesn't replace them
- If one input looks unusually low, such as CRP or creatinine, ask your clinician whether it reflects the test or your body
If it’s high
A PhenoAge above your calendar age means your nine results, taken together, look more like those of an older person in the survey the equation was built on. This is not a diagnosis. One or two inputs, often the heavily weighted ones such as RDW or glucose, can account for much of it, and some causes are temporary.
Common causes
- A recent infection, injury or hard workout, which can raise CRP and white blood cell count for a few days
- Higher blood sugar, including a draw taken after eating or a longer-term pattern of high glucose
- A higher red cell distribution width (RDW), which can be linked to low iron, vitamin B12 or folate, among other causes
- Lower albumin, which can follow illness, inflammation or low protein intake
- Higher creatinine, from lower kidney filtering, dehydration, a recent meat meal, creatine supplements or high muscle mass
- Unit mix-ups or a wrong birth date when the score was calculated
What can help
- Look at which of the nine results contribute most, and ask your clinician whether any of them deserve follow-up
- Recheck after an illness has passed, and use a fasting draw if glucose is one of the drivers
- Make sure all nine values came from the same draw and were entered in the right units
- Not smoking, regular physical activity, a balanced diet and enough sleep support health in general; your clinician can tailor advice to you
What can affect your result
- Age: each calendar year adds about 0.9 years to the score when your results stay the same
- Sex: women averaged about 1.3 years lower than men in US survey data
- Illness: infections and injuries can raise CRP and white blood cell count for days or weeks
- Fasting: glucose is part of the equation, and a draw after eating can raise it
- Which CRP test you had: the equation was built on an older test that couldn't report below 0.3 mg/dL (3 mg/L)
- Which RDW you use: the percent version (RDW-CV), not RDW-SD in fL
- Kidney function, muscle mass, diet and supplements, which change creatinine
- Anemia and low iron, B12 or folate, which change MCV and RDW
- The equation was built on adults. It isn't designed for children, and pregnancy changes several of the inputs
- Unit mix-ups. The published equation expects albumin in g/L, creatinine in µmol/L, glucose in mmol/L and CRP in mg/dL (as a natural log). US reports use g/dL, mg/dL, mg/dL and usually mg/L, so an unconverted value gives a wrong score.
- Which CRP test was used. The equation was built on an older CRP test used in NHANES III (1988 to 1994) that reported anything under 0.3 mg/dL (3 mg/L) as '<0.3'. Modern high-sensitivity results go much lower, and the equation has no rule for a CRP reported as 'less than'. Using 0.5 mg/L instead of 3 mg/L lowers the score by about 1.9 years.
- A recent infection, injury or hard workout can raise CRP and white blood cell count for days, and the score rises with them.
- Glucose enters the equation directly. A draw after eating can read higher than a fasting one; each 10 mg/dL (0.55 mmol/L) adds about 1.2 years.
- RDW must be the percent version (RDW-CV). RDW-SD, reported in fL, is a different measure.
- Lymphocyte percent must be the percent of white cells from the differential, not the absolute lymphocyte count.
- Conditions and medicines that change an input also change the score: steroids (glucose, white cells), anemia or low iron, B12 or folate (MCV, RDW), kidney disease, muscle mass or creatine supplements (creatinine).
No time of day is required. For comparisons over time, draw blood at a similar time of day and with the same fasting status each time. Each input moves at its own speed. Glucose, white blood cell count and CRP can change within days, while albumin and the red-cell measures (MCV and RDW) move over weeks to months. A single illness can shift the score for a short time.
When to talk to a clinician
PhenoAge has no urgent cutoff. If the lab flagged any of the nine results as urgent, call the clinician who ordered the tests. This score should not stand in for talking with your clinician about symptoms that worry you.
Questions you could ask
- Is it useful for me to track this over time, or are the individual results enough?
- Which of these nine results matter most for my own health?
- Which of my nine results are pulling this number up, and are any worth rechecking?
- Could a recent illness, a non-fasting draw or a medicine be affecting some of these results?
- Does a PhenoAge below my age change anything about how often I should have checkups?
- Is any single result, such as CRP or creatinine, unusually low in a way that should be looked at?
Questions people ask
Why does my PhenoAge differ from another calculator?
Usually it is units or CRP. The equation expects albumin in g/L, creatinine in µmol/L, glucose in mmol/L and CRP in mg/dL, and most US reports use other units. Tools also differ in how they handle high-sensitivity CRP and 'less than' values, and some round the constants. Gaps of a fraction of a year come from rounding. Gaps of several years usually mean a unit or input mix-up.
Which of my results move PhenoAge the most?
The published weights make this exact. Each 1-point rise in RDW adds about 3.6 years. Each 1 g/dL drop in albumin adds about 3.7 years. Each 18 mg/dL (1 mmol/L) rise in glucose adds about 2.1 years. Each 0.1 mg/dL rise in creatinine adds about 0.9 years, and each extra 1,000 white blood cells per µL adds about 0.6 years. CRP counts for less: each doubling adds about 0.7 years. This describes how the equation works. It isn't advice to change any single result.
Is a PhenoAge lower than my age good news?
It is a pattern that went with lower risk across groups in studies, but it isn't a promise about you. A single input, such as a recent cold, a non-fasting draw or a very low high-sensitivity CRP, can move the number by a year or more. Researchers also still lack agreed standards for using aging scores in clinical care, so it is best used as one more way to look at your routine results.
Is this the same as the PhenoAge from a DNA methylation test?
No. DNAm PhenoAge is a separate number that predicts this blood-chemistry score from chemical tags at 513 sites on DNA. The two are related but not interchangeable, so keep them apart.
Show the science and sources
For clinicians and researchers
PhenoAge (Levine 2018; Liu 2018) was built in NHANES III. A penalized Cox model picked nine biomarkers from 42 candidates in 9,926 adults. A Gompertz proportional hazards model then gave a 10-year (120-month) mortality score, which was converted to years using a second Gompertz model fit on chronological age alone. The linear predictor is xb = −19.907 − 0.0336 × albumin + 0.0095 × creatinine + 0.1953 × glucose + 0.0954 × ln(CRP) − 0.0120 × lymphocyte% + 0.0268 × MCV + 0.3306 × RDW + 0.00188 × ALP + 0.0554 × WBC + 0.0804 × age, with albumin in g/L, creatinine in µmol/L, glucose in mmol/L, CRP in mg/dL, lymphocytes in %, MCV in fL, RDW in %, ALP in U/L, WBC in 10^3 cells/µL and age in years. The mortality score is M = 1 − exp(−1.51714 × exp(xb) / 0.0076927), and PhenoAge = 141.50 + ln[−0.00553 × ln(1 − M)] / 0.09165. The two logarithms cancel, so PhenoAge equals about 142.45 + 10.91 × xb. It is linear in the inputs, and computing it this way avoids rounding M to 1 for very high-risk profiles. Each unit of an input adds its weight × 10.91 years: RDW 3.6 years per percentage point, glucose 2.1 per mmol/L, albumin −3.7 per g/dL, creatinine 0.9 per 0.1 mg/dL, WBC 0.6 per 10^3/µL, MCV 0.3 per fL, lymphocytes −0.13 per percentage point, ALP 2.1 per 100 U/L, ln(CRP) 1.04 per natural-log unit (0.7 per doubling) and chronological age 0.88 per year. The equation as first printed in Liu 2018 had errors (glucose 0.0195, RDW 0.3356 and a missing step) that a 2019 correction fixed; the Levine 2018 Table 1 weights match the correction, apart from ALP, which Table 1 rounds to 0.0019. Some implementations use an intercept of 141.50225; the difference is 0.002 years. The NHANES III CRP method was latex-enhanced nephelometry that reported results under 0.3 mg/dL as '<0.3', and about 65% of participants in one analysis were at or below its detection limit. RDW in NHANES III is the coefficient of variation of red cell volume. In NHANES IV (Liu 2018, 11,432 adults aged 20 to 84), each year of higher PhenoAge at a given chronological age carried a hazard ratio of 1.09 (95% CI 1.08 to 1.10) for all-cause mortality, about 13%, 10% and 8% per year in ages 20 to 39, 40 to 64 and 65 to 84. In a separate NHANES IV validation sample of 6,209 adults (Levine 2018), PhenoAge correlated with chronological age at r = 0.94. PhenoAge is not the same as DNAm PhenoAge, a DNA methylation predictor (513 CpGs) of this score, or the KDM biological age. No clinical reference interval or decision threshold exists, and the field lacks agreed standards for clinical use of aging biomarkers.
Sources
- An epigenetic biomarker of aging for lifespan and healthspan, Aging (Albany NY) (2018)
- A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV: A cohort study, PLOS Medicine (2018)
- Correction: A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV: A cohort study, PLOS Medicine (2019)
- NHANES III Laboratory Procedures Manual (C-reactive protein in serum; hematology methods), CDC National Center for Health Statistics
- Genetic Variation Is Associated With C-Reactive Protein Levels in the Third National Health and Nutrition Examination Survey, Circulation (2006)
- Biomarkers of aging for the identification and evaluation of longevity interventions, Cell (2023)
AI-verified against cited sources. Awaiting expert review. Last reviewed 2026-10-03. How we build and check these pages.