Secular decline in male serum testosterone across five countries, 1972–2019
This pooled analysis combined 118,593 male testosterone observations from the USA, Denmark, Finland, Israel and Brazil spanning 1972–2019. It was presented at the 2026 annual meeting of the European Society of Human Reproduction and Embryology and reported the largest cross-country estimate of the secular decline discussed on this site.
Design
Pooled analysis
Conference presentation, July 2026
Observations
118,593
Male testosterone observations
Countries
Five
USA, Denmark, Finland, Israel, Brazil
Reported change
≈54% lower
1972 compared with 2019
Structured abstract
Method
118,593 male testosterone observations spanning 1972 to 2019 were pooled across the USA, Denmark, Finland, Israel and Brazil. The scale matters: single-country cohorts can always be dismissed as local, and a five-decade window covers the whole period in which diet, body composition, activity and environmental exposure changed most.
Result
Average male testosterone was reported to be approximately 54% lower in 2019 than in 1972. It is the largest cross-country estimate of the secular decline discussed on this site.
Conclusion
The smaller declines measured in the American, Finnish and Danish cohorts sit inside this longer pattern rather than contradicting it. Each of those measured a slice; this analysis lines the slices up.
Why it matters
118,593 observations across five countries: average male testosterone about 54% lower in 2019 than in 1972; conference data, the figure is provisional.
Study design
Five countries, five decades
Pooling is what turns a set of national observations into a statement about a shared direction.
Scale
118,593 observations
Male testosterone observations combined across five national datasets rather than analysed one country at a time.
Span
1972 to 2019
A window long enough to include the whole period in which body composition, diet, activity and exposure changed most.
Geography
USA, Denmark, Finland, Israel, Brazil
Five countries across three continents, which is what makes a purely local explanation harder to sustain.
Presented
Scientific meeting, July 2026
Peer review
Full paper pending at time of writing
Methods
Complete methods not yet published
Status on this site
Reported as conference evidence
The headline figure, and what surrounds it
The number is large and it is provisional. Both of those statements have to travel together.
Lower in 2019 than 1972
≈54%
The reported difference in average male testosterone across the pooled datasets.
Observations pooled
118,593
The dataset size is what distinguishes this estimate from the single-country cohorts.
Until publication
Provisional
Conference analysis: the exact percentage may move when the full paper appears.
Because complete methods were not published at the time of writing, this page does not reproduce sub-analyses, confidence intervals or per-country breakdowns. Only the figures stated in the presentation coverage are shown.
The measured cohorts underneath it
The pooled figure is new; the direction is not. These are the peer-reviewed studies that measured pieces of the same trend.
American men
Repeated sampling in the same community found lower testosterone at the same ages in later sampling periods, after adjustment.
Travison et al., JCEM, 2007
Finnish cohorts
Men aged 60–69 had 21.9 nmol/L when born 1913–1922 and 13.8 nmol/L when born 1942–1951 — the same age, a different generation.
Perheentupa et al., EJE, 2013
Young American men
Mean total testosterone fell across NHANES survey cycles in men aged 15–39, and the trend held even at normal BMI.
Lokeshwar et al., European Urology Focus, 2020
Clinical interpretation
What survives whatever the final number is
Whatever figure the published paper eventually carries, the direction has now been reproduced across countries, decades and independent research groups. That part no longer depends on any single study.
For an individual man, the size of the population decline changes nothing about his own result. It changes what the reference range around that result means.
We flag this study as conference evidence everywhere it appears on this site, and we will update the figure when the peer-reviewed paper is available.
Related studies in this library
Journal of Clinical Endocrinology & Metabolism · 2007
A population-level decline in serum testosterone levels in American men
Journal of Clinical Endocrinology & Metabolism 86: 724–731 · 2001
Longitudinal effects of aging on serum total and free testosterone levels in healthy men
European Journal of Endocrinology 168: 227–233 · 2013
A cohort effect on serum testosterone levels in Finnish men
Your own numbers
The population trend is not your result.
Whatever has happened to the average, the only figure that can guide a decision is your own measured level.
- Total testosterone and SHBG by LC-MS/MS
- Free testosterone calculated from both
- Symptom questionnaire alongside the result
- A physician to interpret the two together