Although men do not experience a phase of accelerated bone loss similar to the menopause in women, their bone health status declines gradually with age (1). According to the World Health Organization (WHO) diagnostic standard using bone mineral density (BMD) measured by dual-energy X-ray absorptiometry (DXA), approximately 3-6% of men have osteoporosis and 28-47% have osteopenia, and the prevalence of osteoporosis rises to 19% in men who are aged 50 and older (2, 3). While it is established that the primary cause of menopausal-bone loss in women is driven by ovarian failure, in men, the etiology of age-related bone loss is less clear.

In men, sex steroids have an important role in the maintenance of bone strength, as demonstrated by the rapid reductions in bone density following androgen deprivation therapy (4). There is evidence of a small and progressive age-related decline in free testosterone, estradiol and to a lesser extent, progesterone (5, 6) and increases in luteinizing hormone, follicle-stimulating hormone (FSH) and sex hormone-binding globulin (SHBG) (7). However, research into the role of age-related sex hormone changes in the development of male osteoporosis has primarily focused on testosterone and estradiol. Endogenous testosterone levels have been related to bone turnover markers, BMD, and hip structural geometry in normal aging men (8–11). The actions of testosterone on the male skeleton are partly mediated by the aromatization of testosterone to estradiol, with estrogen deficiency also contributing to age-related bone loss in men, and there is cross-sectional and longitudinal evidence to suggest that bioavailable estradiol is more strongly associated with male BMD and fracture (10, 12, 13). Testosterone has been suggested to also play an indirect role in male skeletal health with aging by allowing for relative maintenance of balance and muscle strength in men compared to women (14). SHBG levels in men increase with age and influence the amount of hormone that is available to enter cells, and are predictive of vertebral fracture (15, 16) and/or lower bone density (17).

Relatively less attention has been directed toward the role of other sex steroids in male bone health, particularly of FSH. In women, a sharp rise in serum FSH during late perimenopause coincides with the most rapid rate of bone loss (18) and epidemiological data suggest that bone turnover markers and BMD in perimenopausal and early postmenopausal women are independent of serum estrogen, but negatively associated with FSH (19–21). However, the relationship between FSH and bone mass in men has not yet been fully explored.

Musculoskeletal fragility associated with sarcopenia (loss of muscle mass) and osteopenia (loss of bone mass) can result in fall and fracture. People are diagnosed with osteoporosis often combining with muscles weakness, and increased spine kyphosis leading vertebral fractures (22). Partial androgen deficiency may contribute to the age-related decrease in muscle mass and strength and to the higher risk of fall and fractures in elderly men (23, 24).

In the present cross-sectional study, we observed the variations of endogenous sex hormone concentrations with increasing of age, and investigated their relations to bone mass. Considering the potential roles of marrow adiposity and paravertebral muscle fragility in bone loss and vertebral fractures in old men which may be correlated with sex hormones, we furthermore explored the relationships between endogenous sex hormones concentrations and vertebral marrow fat content as well as paravertebral muscle fat content.

The general characteristics of the study participants are shown in Table 1 . Participants were stratified into four sub-groups according to age, and the characteristics of each group and inter-group comparisons are also presented in Table 1 . FSH concentrations were higher with increasing age (p < 0.001), but there were no age related differences in total estradiol concentration (p = 0.90). With aging, lumbar spine vertebral vBMD decreased (p < 0.001) and erector muscle fat content increased (p = 0.034) gradually. There were no significant age-related differences in marrow fat content (p = 0.574).

Further sub-group analysis was completed according to presence of osteopenia and osteoporosis (normal bone mass: vBMD ≥ 120 mg/mm³ and osteopenia/osteoporosis: vBMD < 120 mg/mm³) and of obesity (non-obesity: BMI < 25 kg/m² and obesity: BMI ≥ 25 kg/m²), respectively, and inter-group comparisons are shown in Table 2 . Participants who were osteopenic or osteoporotic had significantly higher FSH concentrations (p = 0.005), marrow fat content (p = 0.002), and erector muscle fat content (p = 0.003). Obese participants had significantly higher total testosterone concentrations (p = 0.004), but there were no differences in vBMD or fat content measurements.

Associations between sex hormone concentrations, vBMD, and fat content measurements are shown in Table 3 . Lower vBMD was associated with higher FSH concentration (β = -0.216, p < 0.001), but not with total estradiol or total testosterone concentration ( Table 3 ), after adjustment for age, BMI, and other sex hormones. Neither bone marrow fat content nor erector muscle fat content was associated with sex hormone concentrations after adjustment for age, BMI, and other sex hormones.

Adjusted and non-adjusted logistic regression analysis found that for each SD increase in FSH there was a 50% higher risk of an individual having osteopenia or osteoporosis ( Table 4 ). All the three sex hormones showed no significant contributions to the risk of higher marrow fat content or higher erector fat content after multivariate adjustment.

This study is the first to explore associations between sex hormones, bone density, marrow adiposity and muscle adiposity in community-dwelling middle and older aged men. In doing so, we found that high concentration of FSH, but not total testosterone or estradiol, was a risk factor for lower vBMD, osteopenia and osteoporosis. While marrow fat content and erector muscle fat content were greater in osteopenic and osteoporotic men, there were no associations with sex hormones concentrations.

Our primary finding was of significant associations between high FSH and low vertebral trabecular vBMD in men. There is accumulating evidence that FSH acts on bone through both direct and indirect mechanisms. First, FSH acts directly through the FSHR on osteoclast precursors to increase osteoclastogenesis by sensitizing MAP kinase, NF-κB, and Akt pathways, and blocking FSHRs or the β-subunit of FSH, can prevent bone loss in mice independent of estrogen (28). Second, FSH increases expression of the receptor activator of NF-κB (29) and indirectly stimulates osteoclastogenesis by releasing cytokines, namely IL-1β, TNF-α, and IL-6 in proportion to the surface expression of FSHR (21, 30). Elsewhere, in-vivo evidence regarding the relationship between FSH and bone mass in men is limited. Hsu B, et al. reported that lower levels of serum FSH were protective against bone loss and were negatively associated with incident fractures in a cohort of elderly men (31). In addition, Jing, et al. have demonstrated associations between FSH, lumbar spine BMD and osteoporosis in men with type 2 diabetes mellitus (32).

In this study, we found no associations between total estradiol, total testosterone and vertebral trabecular vBMD. Elsewhere, studies have supported estradiol as the crucial sex steroid for bone health in men, and others, provide evidence for testosterone and SHBG (8–11, 32–37). In elderly men, bioavailable estradiol has been found to be positively associated with BMD at multiple sites (8, 10, 11, 33–35), and higher bioavailable estradiol is suggested to be a protective factor for osteopenia (9). Total estradiol has also been reported to be a significant determinant of bone mass in elderly men (10, 24, 34, 36). In a study involving a diverse sample of men (Black, Hispanic, and White, aged 30 to 79 years), total and free estradiol showed larger positive correlations with BMD outcomes and correlations between estradiol levels and hip BMD were robust after multivariate-adjustment, whereas no correlation between total or free estradiol and lumbar spine BMD after multivariate-adjustment (8). In aging Chinese men, both bioavailable and total estradiol have been demonstrated with a positive association with both total hip and lumbar spine BMD after multivariate- or age-adjustment (32, 37). The bioavailable sex steroids comprise the fractions that are free or associated with albumin in the circulation (38), and it is these fractions that have rapid access to target tissues (39). The free fraction constitutes only 1-3% of the total circulating sex steroids (38). In contrast, the fraction bound to SHBG does not have free access to target tissue. As SHBG levels increase with age in men, measurement of total testosterone or estradiol levels does not accurately reflect the actual levels of these steroids available to tissue.

The evidence around the role of testosterone for bone health in elderly men is more conflicting, than the evidence for estradiol (9–11, 24, 34, 36). A study on American men aged 20–90 years found that men with free testosterone level at the lowest quartile were more likely to be osteopenic (9). A study by Van Den Beld et al. also found a positive and significant association between testosterone levels (total, free, and bioavailable fraction and androgen index) and proximal femoral BMD in older men (10). The Swedish Osteoporotic Fractures in Men (MrOS) Study discovered that free testosterone was an independent positive predictor of BMD in total body, total hip, femur trochanter, and arm but not in the lumbar spine (11). The Hong Kong MrOS study also found a positive and significant relationship between femoral and vertebral BMD and free testosterone in Chinese men aged 65 years and above (37). However, discrepancies exist among different studies, and some have reported no associations between bone health status in men and total testosterone (8, 9, 34), and studies have found that the relationship between testosterone and BMD is not consistent at different testing sites (35). The effects of testosterone can be exerted either directly through the androgen receptor or indirectly through aromatization to estrogens and further through estrogen receptor-α and/or –β (40). These sex steroid receptors are expressed in bone, and experimental animal studies, using sex steroid receptor-inactivated transgenic mouse model, have indicated that each of these three receptors mediate the site-specific skeletal effects of sex steroids (41, 42).

To date, most studies have been conducted using DXA measured BMD, which cannot separate trabecular from cortical bone. Trabecular bone is known to have a more rapid rate of age-related loss than cortical bone. This may diminish the sensitivity of DXA for assessing osteoporosis. QCT is a truly three dimensional technique for quantifying volumetric trabecular bone density that is not affected by spine degeneration and abdominal aortic calcification (43). Khosla et al, measured vBMD using QCT and showed that vertebral trabecular vBMD was significantly associated with bioavailable estradiol and bioavailable testosterone in the entire group of men (22 to 91 years) (33). After adjusting for age, bioavailable testosterone was more strongly associated with vertebral trabecular vBMD than bioavailable estradiol in the middle-aged men (40-59 years), while bioavailable estradiol was the best predictor of vBMD at most sites in the elderly men (≥ 60 years) (33).

In the current study, there were no associations between sex hormone concentrations and marrow fat content, despite evidence that marrow fat increases with age (44, 45). Mistry et al, have reported that marrow adiposity is negatively associated with both total estradiol and total testosterone in aging men (74 to 95 years of age) after adjusting for age and total percent fat (46). The possible reasons for the discrepancies with Mistry’s results may be firstly, the number size of this study (n = 62) was small and secondly, the participants in this study are much younger than those in Mistry’s study (61 years vs. 82.4 years). FSH receptor cDNA and protein have been identified on osteoclasts and mesenchymal stem cells (28). An animal study showed that both sham-operated and ovariectomized mice presented a decrease in marrow adipose volume after FSH antibody treatments (47). Our study did not find a significant association between marrow fat content and FSH concentration in this middle-aged to older male population, and furthermore, the increase of FSH was not a risk factor for higher marrow fat content.

In the context of muscle aging, it is important to remember that it is not just a decline in muscle mass which contributes to the deterioration of muscle function. Factors underpinning muscle quality come into play, including muscle composition, aerobic capacity and metabolism, fatty infiltration, insulin resistance, fibrosis and neural activation (48). Observational studies have shown that lean mass and strength are reduced in men with low testosterone levels (10, 23), and testosterone replacement increases lean mass in men with hypogonadism (49). In the present cross-sectional study, total testosterone concentration was detected with a weakly negative correlation with erector muscle fat content by unadjusted linear regression analysis, however, the correlation was insignificant after adjusting for age, BMI, total estradiol concentration, and FSH concentration. The increase of total testosterone for 1 SD was not a significant protective factor for higher erector muscle fat content (≥ 10%). Considering lean mass, muscle area, and muscle strength were preserved until gonadal steroid deficiency was more mark according to Finkelstein et al. (a testosterone level ≤ 200 ng per deciliter) (50), our results are not unexpected.

The main limitation of this study was the relative small sample size (n = 199), especially only 62 participants had MR measurements of marrow fat content and muscle fat content, and as a result, the relationships between sex hormones and MR-quantified marrow fat and muscle fat content may not be fully revealed. In addition, luteinizing hormone and SHBG were not measured, and therefore, free estradiol and free testosterone could not be calculated and investigated in this study. The smoking and drinking history were not obtained in this study, and therefore, neither multivariate linear regression analyses for the contributions of sex hormones to vBMD nor ORs of osteopenia/osteoporosis were adjusted for smoking and drinking.

FSH but not total estradiol or total testosterone is related to vertebral trabecular vBMD in middle-aged and older Chinese men. Neither marrow adiposity nor muscle adiposity is associated with total estradiol, total testosterone, or FSH.

The original contributions presented in the study are included in the article/supplementary files. Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Beijing Jishuitan Hospital Ethics Committee. The patients/participants provided their written informed consent to participate in this study.

The first authorship of LX and QZ is of an equal rank. LX and QZ designed the study and prepared the first draft of the paper. KL, YZ, and YL contributed the experimental work and data collection. KH and LW edited the draft. CW was responsible of the statistical analysis of the data. XC supervised the study and paper organization. All authors revised the paper critically for intellectual content and approved the final version. All authors agree to be accountable for the work and to ensure that any questions relating to the accuracy and integrity of the paper are investigated and properly resolved.

This work was funded by the grants from the National Key Research and Development Program of China [No. 2020YFC2004902], the Beijing Hospitals Authority Clinical Medicine Development of Special Funding Support [No. ZYLX202107], the National Natural Science Foundation of China [No. 81771831, 2018], and the High Level Talents “Discipline backbone” Project of Beijing Jishuitan Hospital [No. XKGG2021123].

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.