REVIEW URRENT C OPINION
Male hypogonadism and skeletal health Michael S. Irwig
Purpose of review To examine the role of testosterone in skeletal health in men. Recent findings Evidence from recent studies shows that the contributing role of testosterone to osteoporosis is modest and likely trumped by other factors such as estradiol levels. A few studies have documented an association between low testosterone levels and lower bone mineral density (BMD), increased prevalence of osteoporosis of the hip and low bone mass-related fractures. Other studies, however, have found that testosterone levels are not independent predictors of bone resorption or formation markers, BMD at the hip or incident fractures. Curiously, hypogonadism does not account for the increased osteoporosis seen in men with Klinefelter Syndrome. Regardless of hypogonadism status, two recent clinical trials have found fewer new morphometric vertebral fractures in men treated with zoledronic acid and increased BMD in men treated with denosumab. Denosumab was also shown to modestly increase bone-metastasis-free survival in men with castration-resistant prostate cancer. Summary Although male hypogonadism is associated with osteoporosis, estradiol is likely to be the more important hormone for bone health. Although a few large randomized controlled trials have been conducted in men with low bone density (a subset of whom have hypogonadism), more trials are needed, particularly with fractures as the main outcome. Keywords estradiol, fractures, male hypogonadism, osteoporosis, testosterone
INTRODUCTION Although male hypogonadism is a well established cause of secondary osteoporosis, the relationship between sex steroids and skeletal health is still an area of active investigation. The prevalence and incidence of osteoporosis in men differs from women as a result of several factors including timing of puberty, larger bone size and strength in men, fewer falls in men, shorter life expectancy in men, and different levels of sex steroids including estradiol and testosterone. Although both androgens and estrogens regulate bone turnover in men, teasing apart the precise role of each hormone can be difficult as testosterone is converted into estradiol via the enzyme aromatase. Men with higher levels of total testosterone and bioavailable testosterone also have higher levels of total estradiol and bioavailable estradiol [1]. Specifically, total testosterone and calculated bioavailable testosterone had correlations with calculated bioavailable estradiol with r values of 0.22 and 0.39, respectively [1]. To further complicate matters, sex-hormone binding globulin (SHBG), the major carrier protein of testosterone and estradiol, is also related to skeletal health. It
is unclear how much of the association between higher SHBG levels and lower bone mineral density (BMD) is due to reduced levels of free estradiol and free testosterone versus direct independent influences of SHBG. This review examines the role of sex hormones in relation to bone remodeling, predictors of BMD and predictors of fracture. Two special populations of male hypogonadism are considered: men with Klinefelter Syndrome and men receiving androgen deprivation therapy (ADT) for the treatment of prostate cancer. The review also summarizes some of the major recent randomized controlled trials (RCTs) in the field and highlights some of the recommendations from the Endocrine Society’s Center for Andrology and Division of Endocrinology, The George Washington University, Washington, District of Columbia, USA Correspondence to Michael S. Irwig, MD, Division of Endocrinology, Medical Faculty Associates, 2150 Pennsylvania Ave NW, Washington, DC 20037, USA. Tel: +1 202 741 2498; fax: +1 202 741 2490; e-mail: [email protected] Curr Opin Endocrinol Diabetes Obes 2013, 20:517–522 DOI:10.1097/01.med.0000436185.36717.76
1752-296X ß 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-endocrinology.com
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Parathyroids, bone and mineral metabolism
KEY POINTS In men, levels of estradiol are more closely linked to skeletal health than the levels of testosterone. In high-risk men with and without hypogonadism, zoledronic acid decreases the incidence of new morphometric vertebral fractures, but does not decrease clinical vertebral and nonvertebral fractures. Denosumab increases BMD regardless of baseline testosterone level and also modestly increases bonemetastasis-free survival in men with castration-resistant prostate cancer.
recent clinical practice guideline on ‘osteoporosis in men’ [2 ]. Much of the epidemiological data presented comes from two important cross-sectional and longitudinal studies: the European Male Aging Study (EMAS) and the Osteoporotic Fractures in Men Study (MrOS). &
BONE REMODELING In younger men, cortical modeling of long bones (radius and tibia) appears to extend beyond adolescence. Sex hormones and parathyroid hormone were not associated with cortical perimeter and cortical thickness at either site [3]. At the radius site only, insulin-like growth factor 1 (IGF-1) was associated with these cortical properties [3]. In older men, increased bone turnover markers have been associated with lower BMD as well as poor bone microarchitecture. One of the objectives of the EMAS was to assess the determinants of bone turnover rate. The bone marker measurements consisted of the serum N-terminal propeptide of type 1 collagen for bone formation and the serum b C-terminal cross-linked telopeptide for bone resorption. Using multivariable linear regression that adjusted for potential confounders, this study found that total and free testosterone were not independently associated with either bone marker [4]. In contrast, estradiol, SHBG, IGF-1 and parathyroid hormone (PTH) were independent predictors of bone remodeling. The study found no evidence of threshold effects for any of the hormones when analyzed by tertiles or quintiles, although the number of frankly hypogonadal men was small. The study also found that higher SHBG may have an adverse effect on BMD. It was estimated that age, lifestyle and hormones accounted for 8–20% of the variability in the bone turnover markers. In men with late-onset hypogonadism (LOH), serum levels of osteoprotegerin and Receptor Activator of Nuclear Factor kappa B (RANK) ligand were 518
www.co-endocrinology.com
higher than those in nonhypogonadal controls [5]. An inverse relationship was seen between osteoprotegerin and bioavailable testosterone. There were no associations between serum sex steroids and RANK ligand. Older studies have shown inconsistent results.
PREDICTORS OF BONE MINERAL DENSITY In the EMAS, men were divided into three groups based upon their total testosterone levels and presence of certain symptoms [6 ]. Moderate LOH was defined as the presence of three sexual symptoms and a total testosterone between 8 and 11 nmol/l (230–320 ng/dl). Severe LOH was based upon the same three symptoms but with a total testosterone of less than 8 nmol/l (230 ng/dl). As compared with eugonadal men, men with both moderate and severe LOH had lower mean estimated heel BMD after adjusting for age, BMI, smoking status and comorbidity. Two of the sites of the EMAS (Belgium and UK) participated in a substudy of hormonal determinants of bone density and geometry using peripheral quantitative computed tomography (CT) [7]. This technology can overcome some of the limitations of dual-energy X-ray absorptiometry and assess bone geometry. This study found that in men over 60 years in Belgium, lower bioavailable estradiol was associated with lower cortical BMD (50% radius site) and trabecular BMD (4% radius site). The findings were not evident in the UK participants or in men under age 60. Of note, differences between the two sites included slightly different scan locations in the distal radial site, as well as higher testosterone concentrations and lower estradiol concentrations in the Belgian cohort. The MrOS Study also looked at sex steroid hormones and BMD. BMD of the total hip was analyzed both cross-sectionally (n ¼ 1238) and longitudinally (n ¼ 969) with adjustment for age, race, clinic and weight [1]. Results from the cross-sectional data revealed positive associations between hip BMD and quintiles of total estradiol, bioavailable estradiol and estrone, and a negative association with SHBG. There was no relationship between hip BMD and either total testosterone or bioavailable testosterone. In the longitudinal analysis over 4.6 years, only bioavailable estradiol and SHBG were associated with the loss of BMD at the total hip. Specifically, men in the lowest quintile of bioavailable estradiol (66.0 pmol/l or 18.0 pg/ml). In addition, a threshold effect was seen for BMD loss when SHBG levels were above 49–60 nmol/l. &
Volume 20 Number 6 December 2013
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Male hypogonadism and skeletal health Irwig
Smaller studies have also found a minimal role of androgens and androgen receptor polymorphisms on BMD. In a study of college-aged athletes chosen from three distinct sports teams (running, wrestling and golf), total and free testosterone were not significant determinants of BMD [8]. Instead, free estradiol accounted for 5% of posteroanterior spine BMD variability and 19% of radius BMD variability. Although reduced activity has been associated with long CAG (polyglutamine) repeats in the androgen receptor, the length of CAG and GGN (polyglycine) repeats was not found to be associated with BMD or bone mineral content after age adjustment [9].
PREDICTORS OF FRACTURE According to a systematic review and meta-analysis, a history of male hypogonadism was identified as one of the risk factors associated with low bone mass-related fractures [10 ]. The odds ratios were 2.77 for natural hypogonadism (P ¼ 0.01) and 1.53 for drug-induced hypogonadism in men treated with ADT for prostate cancer (P ¼ 0.01). Other statistically significant associations were seen with age, low bone mass index, excessive alcohol use, current smoking, chronic glucocorticoid use, history of prior fractures, history of falls, history of stroke and history of diabetes. In the cross-sectional analysis of the MrOS Study in the USA, men with a total testosterone less than 200 ng/dl (6.9 nmol/l) had a three-fold risk for osteoporosis at the hip as compared to men with the reference group (total testosterone levels > 500 ng/dl or 17.4 nmol/l) [11]. It is important to note that men with borderline testosterone levels between 200 and 300 ng/dl (6.9–10.4 nmol/l) did not have an increased prevalence of osteoporosis at the hip. The MrOS Study also included a cohort of 1489 community-dwelling men 65 years and older from Hong Kong, China that were prospectively followed [12]. Baseline sex steroids were measured by gas chromatography mass-spectroscopy (GC-MS) and BMD was assessed at baseline and at 4 years. In this study, testosterone was not a major determinant of fractures. However, men in the lowest quartile of estradiol and bioavailable estradiol had a roughly 50% increased risk of incident fractures as compared with the other quartiles. The fracture threshold for bioavailable estradiol was 12.51 pg/ml (45.9 pmol/l), which was similar to other studies in white populations. The Framingham Study also prospectively followed 793 elderly men (mean age 71) for 16–18 years with baseline measurements of total &
estradiol and testosterone. The adjusted hazard ratio for incident hip fractures was 3.1 (95% confidence interval, CI 1.4–6.9) in men in the low estradiol group [34.3 pg/ml (126 pmol/l)] [13]. Similar adjusted analyses were performed for testosterone but they found no significant associations. Nevertheless, the combination of a low estradiol and low testosterone conferred the highest hazard risk of incident hip fracture of 6.5 (95% CI 2.9–14.3) as compared with the reference group. Using a dose response relation, this study suggested an increased risk of hip fracture for men with serum estradiol levels under 20 pg/ml (74 pmol/l). The WHO, Monitoring of Trends and Determinants in Cardiovascular Disease Project in Gothenburg, Sweden calculated the incidence of X-ray verified fractures in three groups of men aged 35–64 in 1995 (n ¼ 590) and in 2008 (n ¼ 525) [14]. Despite the lower testosterone levels in men in 2008 as compared with the levels in aged-matched men in 1995, there was no difference in fracture incidence. This lends further support that testosterone levels have a minimal role in fracture incidence among the general population. Limitations in this study include higher rates of use of medications that can affect bone (calcium, vitamin D, antiosteoporotic agents) among men in 2008 than in 1995, and a small number of men with hypogonadism.
KLINEFELTER SYNDROME It is well established that men with Klinefelter Syndrome have lower BMD than control populations. Although many men with Klinefelter Syndrome have hypogonadism, they also are different from healthy men in other regards that may affect skeletal health: increased height, weight, gonadotropins and estradiol levels, and decreased 25 hydroxy vitamin D levels [15]. Among men with Klinefelter Syndrome, those with hypogonadism (total testosterone
Male hypogonadism and skeletal health Michael S. Irwig
Purpose of review To examine the role of testosterone in skeletal health in men. Recent findings Evidence from recent studies shows that the contributing role of testosterone to osteoporosis is modest and likely trumped by other factors such as estradiol levels. A few studies have documented an association between low testosterone levels and lower bone mineral density (BMD), increased prevalence of osteoporosis of the hip and low bone mass-related fractures. Other studies, however, have found that testosterone levels are not independent predictors of bone resorption or formation markers, BMD at the hip or incident fractures. Curiously, hypogonadism does not account for the increased osteoporosis seen in men with Klinefelter Syndrome. Regardless of hypogonadism status, two recent clinical trials have found fewer new morphometric vertebral fractures in men treated with zoledronic acid and increased BMD in men treated with denosumab. Denosumab was also shown to modestly increase bone-metastasis-free survival in men with castration-resistant prostate cancer. Summary Although male hypogonadism is associated with osteoporosis, estradiol is likely to be the more important hormone for bone health. Although a few large randomized controlled trials have been conducted in men with low bone density (a subset of whom have hypogonadism), more trials are needed, particularly with fractures as the main outcome. Keywords estradiol, fractures, male hypogonadism, osteoporosis, testosterone
INTRODUCTION Although male hypogonadism is a well established cause of secondary osteoporosis, the relationship between sex steroids and skeletal health is still an area of active investigation. The prevalence and incidence of osteoporosis in men differs from women as a result of several factors including timing of puberty, larger bone size and strength in men, fewer falls in men, shorter life expectancy in men, and different levels of sex steroids including estradiol and testosterone. Although both androgens and estrogens regulate bone turnover in men, teasing apart the precise role of each hormone can be difficult as testosterone is converted into estradiol via the enzyme aromatase. Men with higher levels of total testosterone and bioavailable testosterone also have higher levels of total estradiol and bioavailable estradiol [1]. Specifically, total testosterone and calculated bioavailable testosterone had correlations with calculated bioavailable estradiol with r values of 0.22 and 0.39, respectively [1]. To further complicate matters, sex-hormone binding globulin (SHBG), the major carrier protein of testosterone and estradiol, is also related to skeletal health. It
is unclear how much of the association between higher SHBG levels and lower bone mineral density (BMD) is due to reduced levels of free estradiol and free testosterone versus direct independent influences of SHBG. This review examines the role of sex hormones in relation to bone remodeling, predictors of BMD and predictors of fracture. Two special populations of male hypogonadism are considered: men with Klinefelter Syndrome and men receiving androgen deprivation therapy (ADT) for the treatment of prostate cancer. The review also summarizes some of the major recent randomized controlled trials (RCTs) in the field and highlights some of the recommendations from the Endocrine Society’s Center for Andrology and Division of Endocrinology, The George Washington University, Washington, District of Columbia, USA Correspondence to Michael S. Irwig, MD, Division of Endocrinology, Medical Faculty Associates, 2150 Pennsylvania Ave NW, Washington, DC 20037, USA. Tel: +1 202 741 2498; fax: +1 202 741 2490; e-mail: [email protected] Curr Opin Endocrinol Diabetes Obes 2013, 20:517–522 DOI:10.1097/01.med.0000436185.36717.76
1752-296X ß 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-endocrinology.com
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Parathyroids, bone and mineral metabolism
KEY POINTS In men, levels of estradiol are more closely linked to skeletal health than the levels of testosterone. In high-risk men with and without hypogonadism, zoledronic acid decreases the incidence of new morphometric vertebral fractures, but does not decrease clinical vertebral and nonvertebral fractures. Denosumab increases BMD regardless of baseline testosterone level and also modestly increases bonemetastasis-free survival in men with castration-resistant prostate cancer.
recent clinical practice guideline on ‘osteoporosis in men’ [2 ]. Much of the epidemiological data presented comes from two important cross-sectional and longitudinal studies: the European Male Aging Study (EMAS) and the Osteoporotic Fractures in Men Study (MrOS). &
BONE REMODELING In younger men, cortical modeling of long bones (radius and tibia) appears to extend beyond adolescence. Sex hormones and parathyroid hormone were not associated with cortical perimeter and cortical thickness at either site [3]. At the radius site only, insulin-like growth factor 1 (IGF-1) was associated with these cortical properties [3]. In older men, increased bone turnover markers have been associated with lower BMD as well as poor bone microarchitecture. One of the objectives of the EMAS was to assess the determinants of bone turnover rate. The bone marker measurements consisted of the serum N-terminal propeptide of type 1 collagen for bone formation and the serum b C-terminal cross-linked telopeptide for bone resorption. Using multivariable linear regression that adjusted for potential confounders, this study found that total and free testosterone were not independently associated with either bone marker [4]. In contrast, estradiol, SHBG, IGF-1 and parathyroid hormone (PTH) were independent predictors of bone remodeling. The study found no evidence of threshold effects for any of the hormones when analyzed by tertiles or quintiles, although the number of frankly hypogonadal men was small. The study also found that higher SHBG may have an adverse effect on BMD. It was estimated that age, lifestyle and hormones accounted for 8–20% of the variability in the bone turnover markers. In men with late-onset hypogonadism (LOH), serum levels of osteoprotegerin and Receptor Activator of Nuclear Factor kappa B (RANK) ligand were 518
www.co-endocrinology.com
higher than those in nonhypogonadal controls [5]. An inverse relationship was seen between osteoprotegerin and bioavailable testosterone. There were no associations between serum sex steroids and RANK ligand. Older studies have shown inconsistent results.
PREDICTORS OF BONE MINERAL DENSITY In the EMAS, men were divided into three groups based upon their total testosterone levels and presence of certain symptoms [6 ]. Moderate LOH was defined as the presence of three sexual symptoms and a total testosterone between 8 and 11 nmol/l (230–320 ng/dl). Severe LOH was based upon the same three symptoms but with a total testosterone of less than 8 nmol/l (230 ng/dl). As compared with eugonadal men, men with both moderate and severe LOH had lower mean estimated heel BMD after adjusting for age, BMI, smoking status and comorbidity. Two of the sites of the EMAS (Belgium and UK) participated in a substudy of hormonal determinants of bone density and geometry using peripheral quantitative computed tomography (CT) [7]. This technology can overcome some of the limitations of dual-energy X-ray absorptiometry and assess bone geometry. This study found that in men over 60 years in Belgium, lower bioavailable estradiol was associated with lower cortical BMD (50% radius site) and trabecular BMD (4% radius site). The findings were not evident in the UK participants or in men under age 60. Of note, differences between the two sites included slightly different scan locations in the distal radial site, as well as higher testosterone concentrations and lower estradiol concentrations in the Belgian cohort. The MrOS Study also looked at sex steroid hormones and BMD. BMD of the total hip was analyzed both cross-sectionally (n ¼ 1238) and longitudinally (n ¼ 969) with adjustment for age, race, clinic and weight [1]. Results from the cross-sectional data revealed positive associations between hip BMD and quintiles of total estradiol, bioavailable estradiol and estrone, and a negative association with SHBG. There was no relationship between hip BMD and either total testosterone or bioavailable testosterone. In the longitudinal analysis over 4.6 years, only bioavailable estradiol and SHBG were associated with the loss of BMD at the total hip. Specifically, men in the lowest quintile of bioavailable estradiol (66.0 pmol/l or 18.0 pg/ml). In addition, a threshold effect was seen for BMD loss when SHBG levels were above 49–60 nmol/l. &
Volume 20 Number 6 December 2013
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Male hypogonadism and skeletal health Irwig
Smaller studies have also found a minimal role of androgens and androgen receptor polymorphisms on BMD. In a study of college-aged athletes chosen from three distinct sports teams (running, wrestling and golf), total and free testosterone were not significant determinants of BMD [8]. Instead, free estradiol accounted for 5% of posteroanterior spine BMD variability and 19% of radius BMD variability. Although reduced activity has been associated with long CAG (polyglutamine) repeats in the androgen receptor, the length of CAG and GGN (polyglycine) repeats was not found to be associated with BMD or bone mineral content after age adjustment [9].
PREDICTORS OF FRACTURE According to a systematic review and meta-analysis, a history of male hypogonadism was identified as one of the risk factors associated with low bone mass-related fractures [10 ]. The odds ratios were 2.77 for natural hypogonadism (P ¼ 0.01) and 1.53 for drug-induced hypogonadism in men treated with ADT for prostate cancer (P ¼ 0.01). Other statistically significant associations were seen with age, low bone mass index, excessive alcohol use, current smoking, chronic glucocorticoid use, history of prior fractures, history of falls, history of stroke and history of diabetes. In the cross-sectional analysis of the MrOS Study in the USA, men with a total testosterone less than 200 ng/dl (6.9 nmol/l) had a three-fold risk for osteoporosis at the hip as compared to men with the reference group (total testosterone levels > 500 ng/dl or 17.4 nmol/l) [11]. It is important to note that men with borderline testosterone levels between 200 and 300 ng/dl (6.9–10.4 nmol/l) did not have an increased prevalence of osteoporosis at the hip. The MrOS Study also included a cohort of 1489 community-dwelling men 65 years and older from Hong Kong, China that were prospectively followed [12]. Baseline sex steroids were measured by gas chromatography mass-spectroscopy (GC-MS) and BMD was assessed at baseline and at 4 years. In this study, testosterone was not a major determinant of fractures. However, men in the lowest quartile of estradiol and bioavailable estradiol had a roughly 50% increased risk of incident fractures as compared with the other quartiles. The fracture threshold for bioavailable estradiol was 12.51 pg/ml (45.9 pmol/l), which was similar to other studies in white populations. The Framingham Study also prospectively followed 793 elderly men (mean age 71) for 16–18 years with baseline measurements of total &
estradiol and testosterone. The adjusted hazard ratio for incident hip fractures was 3.1 (95% confidence interval, CI 1.4–6.9) in men in the low estradiol group [34.3 pg/ml (126 pmol/l)] [13]. Similar adjusted analyses were performed for testosterone but they found no significant associations. Nevertheless, the combination of a low estradiol and low testosterone conferred the highest hazard risk of incident hip fracture of 6.5 (95% CI 2.9–14.3) as compared with the reference group. Using a dose response relation, this study suggested an increased risk of hip fracture for men with serum estradiol levels under 20 pg/ml (74 pmol/l). The WHO, Monitoring of Trends and Determinants in Cardiovascular Disease Project in Gothenburg, Sweden calculated the incidence of X-ray verified fractures in three groups of men aged 35–64 in 1995 (n ¼ 590) and in 2008 (n ¼ 525) [14]. Despite the lower testosterone levels in men in 2008 as compared with the levels in aged-matched men in 1995, there was no difference in fracture incidence. This lends further support that testosterone levels have a minimal role in fracture incidence among the general population. Limitations in this study include higher rates of use of medications that can affect bone (calcium, vitamin D, antiosteoporotic agents) among men in 2008 than in 1995, and a small number of men with hypogonadism.
KLINEFELTER SYNDROME It is well established that men with Klinefelter Syndrome have lower BMD than control populations. Although many men with Klinefelter Syndrome have hypogonadism, they also are different from healthy men in other regards that may affect skeletal health: increased height, weight, gonadotropins and estradiol levels, and decreased 25 hydroxy vitamin D levels [15]. Among men with Klinefelter Syndrome, those with hypogonadism (total testosterone
