Testosterone Levels by Age Chart

Start with the thing most charts get backwards. The reference range printed on a US laboratory report — 264 to 916 ng/dL — is not an age band. It was derived from healthy nonobese men aged 19 to 39 and is then applied unchanged to a man of 25 and a man of 75. That is a deliberate decision, argued explicitly in the paper it comes from: "Should the reference range be based on a sample of healthy young men (the so-called T-score approach) or should the reference range be age adjusted (the Z-score approach)? We have provided reference ranges in a young healthy reference sample as well as by decades of age." The first two tables below are those decade centiles. They tell you where you sit among men your age. They are not a range your result is "normal" inside.

The decade tables say something a staircase chart cannot. Between 19–39 and 80–99, the median in nonobese men falls by 55 ng/dL — 10% — while the 95th centile falls by 11 ng/dL, barely 1%. The 2.5th centile is where the collapse happens: 267 down to 157 ng/dL, a fall of 41%. Ageing does not slide the whole distribution downward. It stretches the bottom of it.

Every figure here carries its assay, because testosterone results from different assays are not interchangeable. The harmonised centiles were calibrated to the CDC reference method; the NHANES series was measured by standardised LC-MS/MS at the CDC. A laboratory whose assay is certified by the CDC Hormone Standardization Program has met "the performance criterion of ±6.4% mean bias when compared to the CDC reference measurement procedure", over 2.50–1,000 ng/dL. An uncertified direct immunoassay has met nothing, and Quest states plainly that immunoassay "lacks sensitivity, specificity, and precision at lower concentrations" — which is precisely the part of the range where a hypogonadism decision gets made.

The last table is the one that makes this sheet different from the others. It lists the decade figures we could not source and declines to print them. If you have seen "40–49: 252–916 ng/dL" on a dozen sites, that is why it is not on this one.

Total testosterone centiles by decade — nonobese men

Model-based estimates from 6,933 nonobese men across four cohorts, harmonised to the CDC reference method. This is a description of where men your age sit, not a range you are inside or outside.

Travison TG, et al. J Clin Endocrinol Metab. 2017;102(4):1161–1173, Table 4, "All nonobese men". Nonobese means body mass index at or below 30 kg/m². Ages 19–99 are the paper's own limits; there is no row below 19 because the paper has no data there and we will not extrapolate one. Note that the 19–39 row above is a model estimate and is NOT the same as the paper's headline range: the observed centiles in its 19–39 reference sample, which is where 264–916 comes from, are 264, 303, 531, 852 and 916 ng/dL. The two differ by up to 13 ng/dL at the top. Both are in the same paper; we print both rather than blending them.

The same table for all men — obese and comorbid included

The 9,054 men of the same four cohorts with nobody excluded. This is closer to the population a clinic actually sees, and every figure in it is lower.

Travison 2017 Table 4, "All men". The gap between the two tables is what excluding obesity and comorbidity is worth: at 19–39 the 2.5th centile moves from 229 ng/dL here to 267 in the table above. Which table applies to you is a clinical judgement, not an arithmetic one.

What ageing actually does to the distribution

Each published centile in nonobese men at 19–39, the same centile at 80–99, and the difference. Read the last column before you accept any chart that shifts a whole range down per decade.

The upper half of the distribution is close to flat across six decades. Almost the whole age effect sits in the lower tail — which is why a chart that lowers the top of the "normal range" for a 60-year-old is describing something the data does not show.

US men, measured: NHANES 2011–2012

Estimated geometric mean total testosterone with 95% confidence interval, from 3,419 male participants in a nationally representative survey, measured by standardised LC-MS/MS at the CDC. Nobody is excluded — this is the population, not a reference sample.

This curve is not a staircase and no decade table can represent it. It falls to 354 ng/dL at 35, rises again to 378 ng/dL at 55–60, and only then declines. The paper's own words: "CGM totalT concentrations in men peaked at ages 20 and 55–60 years. At age 80 years, concentrations were 30% lower than at age 20." Its authors attribute the midlife peak to treatment, concluding it "appeared to be consistent with the increased use of exogenous testosterone" — so these are means of a population that includes men on testosterone, and they are not a reference range.

Thresholds — decisions, not percentiles

Five different numbers are used to call a testosterone level "low". None of them is age-adjusted, and the widest pair disagree by 67 ng/dL.

The EMAS figure is published as 11 nmol/L; the paper prints its own US conversion as 3.2 ng/mL, or 320 ng/dL, where the molar mass gives 317. We print the derived figure and name the discrepancy rather than choosing silently. Note also that none of these five numbers changes with age, while the 2.5th centile of the population falls 110 ng/dL between the twenties and the eighties — which is the whole reason a man of 75 can be simultaneously typical for his age and below every threshold in this table.

Your result against the published centiles

A round-number result in both units, and where it falls among nonobese men of two very different ages. The same number is a different percentile at 30 and at 65 — that, and not a shifting "normal range", is what age does.

Centiles between the published points are linearly interpolated, which is an approximation and not a figure from the paper. "Below 264?" is the harmonised reference floor, which does not move with age. 9.2 nmol/L is the same threshold on a European report.

Assay — which is not a footnote

The centiles above apply to a fasting morning sample measured by an assay calibrated to the CDC reference method. Results from an uncalibrated assay are not interchangeable with them.

Certification is for MEAN bias, and it is annual. The CDC's own list reports an "individual samples pass rate" — the share of the 40 samples that met ±6.4% — and for currently certified participants it runs from 20% to the low nineties. A certified laboratory is one whose average is right, not one whose every result is. There is no CDC standardisation programme for calculated free testosterone and no harmonised free-testosterone reference range, which is why this sheet has no free-testosterone-by-age table.

Figures this sheet declines to print

The decade table that appears on most "testosterone by age" pages. We traced each row as far as it goes and could not reach a cohort, an assay or a sample size, so the numbers are named here and printed nowhere else on this page.

One consumer article attributes its decade table to a two-page editorial comment in Clinical Endocrinology (2005;62(3):263–4), which is a commentary on another paper rather than a reference-interval study. A laboratory's own consumer blog prints the same rows and attributes them to a Harvard Health newsletter article. Neither names an assay, a cohort or a percentile definition, and we could not locate a primary dataset containing those intervals. If you can point us to one, the page will be updated.

Frequently asked questions

What is a normal testosterone level for my age?

There is no age-specific normal range in routine clinical use, and that is not an omission. The harmonised reference range every major guideline works from is 264–916 ng/dL (9.2 nmol/L to 31.8 nmol/L), derived from healthy nonobese men aged 19–39 and applied unchanged at every adult age. Age-specific centiles do exist — the tables above — but they tell you where you rank among men your age, not whether your result is normal. A 70-year-old at the median for 70-year-olds is at 477 ng/dL, comfortably inside the reference range; a 70-year-old at the 2.5th centile for his age is at 218 ng/dL, which is below it.

Where does the "40–49: 252–916, 50–59: 215–878" decade chart come from?

We could not find out, which is why this page does not reproduce it. The version that circulates most — 40–49: 252–916, 50–59: 215–878, 60–69: 196–859, 70–79: 156–819 — is attributed by one consumer health site to a two-page editorial comment in Clinical Endocrinology from 2005, and by a laboratory's own consumer blog to a Harvard Health newsletter article. Neither attribution names a cohort, an assay, a sample size or a percentile definition. A genuine decade table does exist (Travison 2017, reproduced above) and it looks nothing like these rows.

Does testosterone really drop 1% a year after 30?

The best longitudinal estimate comes from the Baltimore Longitudinal Study of Aging: 890 men, mixed-effects model, "an average change of -0.124 nmol/L.yr" — about 3.6 ng/dL a year, or 36 ng/dL a decade. Two caveats matter. Testosterone was measured by radioimmunoassay, not mass spectrometry, so the absolute values are not interchangeable with the tables above. And the cross-sectional picture is different again: in NHANES 2011–2012 the geometric mean fell from 393 ng/dL at 20 to 354 at 35, then rose to 378 at 55.

Why does my lab report the same reference range at 65 as at 25?

Because the people who built the range chose to. Travison and colleagues put the question in print — whether to use a healthy-young-men reference (the "T-score approach") or an age-adjusted one (the "Z-score approach") — and published both, arguing: "For analytes that exhibit substantial age-related change, such as testosterone and estradiol, it might arguably be more appropriate to derive the reference ranges in a healthy young population." The Endocrine Society and the AUA both work from a single adult threshold. Quest states its adult lower limit of 250 ng/dL "is based on the 2.5th percentile of a distribution of results" for men aged 18 and over.

How much does testosterone actually fall between 30 and 80?

Less at the top than most charts imply, and far more at the bottom. In nonobese men the median falls from 531 to 476 ng/dL between 19–39 and 80–99 — 55 ng/dL, or 10%. The 95th centile falls only 11 ng/dL. But the 2.5th centile falls from 267 to 157 ng/dL, 41%. Ageing widens the bottom of the distribution rather than sliding the whole thing down.

Does it matter whether my testosterone was measured by LC-MS/MS or immunoassay?

Yes, and it is the single most under-stated thing on a testosterone report. The CDC runs a Hormone Standardization Program precisely because results were not comparable between laboratories; certified assays have met "the performance criterion of ±6.4% mean bias when compared to the CDC reference measurement procedure", over 2.50–1,000 ng/dL. Both LC-MS/MS methods and some chemiluminescence immunoassays hold that certification. An uncertified immunoassay holds nothing, and Quest advises that immunoassay "lacks sensitivity, specificity, and precision at lower concentrations. Thus, IA is not recommended for testing women, children, transgender women, and hypogonadal men". Every centile on this page assumes a CDC-calibrated result.

What testosterone level counts as low?

Four different answers are in current use and they disagree. The AUA guideline statement reads: "Clinicians should use a total testosterone level below 300 ng/dL as a reasonable cut-off in support of the diagnosis of low testosterone." The harmonised reference range puts its floor at 264. The Testosterone Trials enrolled men below 275 ng/dL. EMAS defined late-onset hypogonadism as total testosterone under 11 nmol/L (317 ng/dL) together with at least three sexual symptoms. All four are decisions about where to draw a line, not measurements of anything, and none of them is age-adjusted.

Do I need two tests?

Yes. The AUA guideline states: "The diagnosis of low testosterone should be made only after two total testosterone measurements are taken on separate occasions with both conducted in an early morning fashion." And the Endocrine Society likewise recommends measuring fasting morning total testosterone and confirming it by repeating the measurement. Every centile on this page was derived from single morning samples, which the authors flag as a limitation: the ranges "were derived from single morning samples and discounted the intraindividual variation in testosterone levels due to pulsatile, diurnal, and circannual secretory rhythms".

Is the population average the same as the reference range?

No, and conflating the two is how "average testosterone by age" charts go wrong. In NHANES 2011–2012 the 10th to 90th percentile for US men aged 20 and over was 150–698 ng/dL — a span whose lower end sits well below every diagnostic threshold, because it includes obese men, diabetic men and men with untreated illness. The harmonised 264–916 range excludes them by design. The population figure describes who is out there; the reference range describes who is healthy. (A small honesty note on that figure: the paper's abstract prints the 90th percentile as 698 ng/dL and its Results section prints 699 for the same statistic. We use the abstract's.)

Sources

1. [1]Travison TG, Vesper HW, Orwoll E, et al. Harmonized Reference Ranges for Circulating Testosterone Levels in Men of Four Cohort Studies in the United States and Europe. J Clin Endocrinol Metab. 2017;102(4):1161–1173 (PMID 28324103) — Tables 2 and 4, and the T-score/Z-score discussion
2. [2]Vesper HW, Wang Y, Vidal M, Botelho JC, Caudill SP. Serum Total Testosterone Concentrations in the US Household Population from the NHANES 2011–2012 Study Population. Clin Chem. 2015;61(12):1495–1504 (PMID 26510959) — Table 3, estimated geometric mean by age
3. [3]CDC Hormone Standardization Program (CDC HoSt) — Certified Total Testosterone Procedures: the ±6.4% mean bias criterion, the 2.50–1,000 ng/dL certification range and the individual-sample pass rates
4. [4]American Urological Association — Testosterone Deficiency Guideline: the 300 ng/dL cut-off and the two-early-morning-measurements requirement
5. [5]Quest Diagnostics Clinical Education Center FAQ 165 — Testosterone Testing: the adult male lower reference limit of 250 ng/dL and the limits of immunoassay at low concentrations
6. [6]Harman SM, Metter EJ, Tobin JD, Pearson J, Blackman MR. Longitudinal effects of aging on serum total and free testosterone levels in healthy men. Baltimore Longitudinal Study of Aging. J Clin Endocrinol Metab. 2001;86(2):724–731 (PMID 11158037) — the −0.124 nmol/L per year longitudinal slope, measured by RIA
7. [7]Wu FC, Tajar A, Beynon JM, et al. Identification of late-onset hypogonadism in middle-aged and elderly men. N Engl J Med. 2010;363(2):123–135 (PMID 20554979) — the 11 nmol/L plus three sexual symptoms definition, European Male Ageing Study
8. [8]Snyder PJ, Bhasin S, Cunningham GR, et al. Effects of Testosterone Treatment in Older Men. N Engl J Med. 2016;374(7):611–624 (PMID 26886521) — the Testosterone Trials enrolment criterion of less than 275 ng/dL
9. [9]Bhasin S, Brito JP, Cunningham GR, et al. Testosterone Therapy in Men With Hypogonadism: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2018;103(5):1715–1744 (PMID 29562364) — fasting morning measurement, confirmed by repeat
10. [10]Labcorp OnDemand — "Testing for testosterone by age": the widely copied 40–49 / 50–59 / 60–69 / 70–79 decade rows, attributed there to a Harvard Health article with no assay or cohort named
11. [11]Hone Health — "Testosterone Levels By Age in Men": the same decade rows, cited there to Barrett-Connor 2005
12. [12]Barrett-Connor E. Male testosterone: what is normal? Clin Endocrinol (Oxf). 2005;62(3):263–264 (PMID 15730406) — the two-page editorial comment cited elsewhere as the source of the decade table