Early Decline of Androgen Levels in Healthy Adult Men

An Effect of Aging Per Se? A Prospective Cohort Study

Thiberiu Banica; Charlotte Verroken; Tim Reyns; Ahmed Mahmoud; Guy T'Sjoen; Tom Fiers; Jean-Marc Kaufman; Bruno Lapauw


J Clin Endocrinol Metab. 2021;106(4):1074-1083. 

In This Article

Abstract and Introduction


Context: Androgen levels have been shown to decline in aging men. However, there is no consensus on the effect of aging, (changes in) body mass index (BMI), lifestyle factors, and intercurrent disease.

Objective: Investigating longitudinal changes in serum androgen levels in healthy men in relation to body composition, lifestyle factors, and intercurrent disease.

Design, Setting, and Participants: Longitudinal, population-based sibling pair study at a university research center. 999 healthy men aged 24 to 46 years of whom 691 were reevaluated after a mean period of 12 years.

Main Outcome Measures: Serum SHBG, LH, and FSH levels measured using immuno-assays. Testosterone (T), estradiol (E2), dihydro-testosterone (DHT), and androstenedione (Adione) measured using liquid chromatography-tandem mass spectometry, free T calculated (cFT).

Results: Baseline age was 34 ± 6 years. Mean BMI increased by 1.19 kg/m2, T levels decreased by 14.2% (20.8 nmol/L vs. 17.8 nmol/L), cFT by 19.1% (392 pmol/L vs. 317 pmol/L), DHT by 15.6% (1.5 nmol/L vs.1.3 nmol/L), and Adione by 10.7% (3.7 nmol/L vs. 3.3 nmol/L; all P < 0.001). E2 did not change over time. SHBG increased by 3.0% (39.8 nmol/L vs. 41.0 nmol/L), LH by 5.8% (4.6 U/L vs. 4.9 U/L) and FSH by 14.7% (4.3 U/L vs. 5.1 U/L) (all P < 0.001). For T, cFT, DHT, Adione, and SHBG, these longitudinal changes persisted after adjustment for confounders (all P < 0.001).

Conclusion: Serum androgen levels start declining early during adult life and independently from changes in BMI and other lifestyle factors, suggesting that aging per se leads to an altered sex steroid status. Given the concurrent rise in gonadotropin levels, the decline in androgen status most likely arises from primary decrease in testicular function.


The decline of serum androgen levels with aging in men is well described in many cross-sectional studies and also in some longitudinal studies.[1–6] This decline is most apparent in populations of elderly men, though some cross-sectional studies reported it starting as early as the fourth decade of life.[7,8] It remains unclear whether this decrease in androgen exposure results from the aging process per se or is driven by intercurrent disease and treatments or changes in body composition and lifestyle factors.[1,2,7] For instance, the European Male Aging Study found that weight gain accentuated the age-related decline in total testosterone (T) in elderly men.[9] Both Shi et al. and Kim et al. even suggested that the age-related decline in T levels in middle-aged men is mainly related to changes in smoking behavior and intercurrent health conditions such as obesity and depression, and not to be considered an inevitable part of aging.[6,10] Indeed, in at least 1 cross-sectional study in middle-aged men, there was no association between age and serum T levels.[11] Because androgen exposure is of importance for many bodily systems in men, better knowledge of the natural history and determinants of changes in androgen exposure during aging is warranted.

Importantly, T is to be considered a pro-hormone with many of its effects mediated by tissue-specific conversion into di-hydro-testosterone (DHT) and estradiol (E2). Moreover, part of androgen exposure results from intracellular conversion of androstenedione (Adione) to T and DHT. Similar to T, circulating Adione and DHT concentrations are reported to decrease with aging,[7,12,13] whereas data on changes in E2 levels in aging men are largely inconsistent.[7] Further, in the circulation, sex steroids are variably bound to the SHBG and to a lesser extent to albumin, with only a minor fraction circulating freely. Because aging in men is accompanied by a rise in SHBG levels, age-related decreases in calculated free T levels (cFT) are even more pronounced than those in total T.[7]

To date, no longitudinal data exist on age-related changes in sex steroids in healthy young adult men. Moreover, many of the available studies measured serum T levels using immunoassays, instead of state-of-the-art liquid chromatography tandem mass spectrometry (LC-MS/MS), and might not have accurately addressed all confounding factors. Therefore, this study evaluated longitudinal changes in serum T, DHT, Adione, E2 (using LC-MS/MS), cFT, SHBG, and gonadotropin levels in a cohort of healthy young and middle-aged men and investigated effects of changes in body composition, lifestyle factors, and intercurrent disease as compared to the effects of aging per se.