Impact of testosterone therapy on bone turnover markers in obese males with type 2 diabetes and functional hypogonadism

Abstract

Methods

Fifty-five obese males with type 2 diabetes mellitus and functional hypogonadism participated in a 2-year, double-blind, placebo-controlled study of testosterone undecanoate (TU). Bone turnover markers C-telopeptide of type I collagen (CTX) and procollagen I N-terminal propeptide (PINP) were assessed at baseline, 12 and 24 months. Bone mineral density (BMD) changes were evaluated after 24 months using dual-energy X-ray absorptiometry. Group T (n = 28) received TU both years. Group P (n = 27) received placebo first year and TU second year.

Results

CTX decreased in group P from 1055 (676–1344) to 453 (365–665) pmol/L (p < 0.001) and from 897 (679–1506) to 523 (364–835) pmol/L (p < 0.001) in T. PINP decreased by 4.30 ± 8.05 μg/L in group P (p = 0.030) and 4.64 ± 8.86 μg/L in T (p < 0.023) after first year of therapy. No femoral neck BMD changes were observed in 32 patients from both groups (n = 16 per group). Lumbar spine BMD increased (by 0.075 ± 0.114 g/cm²; p = 0.019) in group T following two years of treatment.

Conclusions

We observed decreased CTX, decreased PINP and increased lumbar spine BMD after two years of testosterone treatment.

Clinical trials

NCT03792321; retrospectively registered trial on 4 January 2019.

Keywords:

• Type 2 diabetes
• hypogonadism
• testosterone
• fracture
• osteoporosis
• bone mineral density

Introduction

Both functional hypogonadism (FH) and type 2 diabetes mellitus (T2D) in men negatively affect bone mineral density (BMD), which increases the risk of fragility fractures, resulting in disabilities and increased mortality [1,2]. Diabetic osteopathy represents an often overlooked complication of T2D [3,4], which is characterized by microarchitectural changes that decrease bone quality, thus leading to an increased risk of bone fracture [5]. Pathophysiological mechanisms of diabetic bone disease include risks directly due to T2D and/or complications of T2D such as nephropathy, neuropathy and diabetic diarrhea [6]. In addition, microvascular complications of T2D lead to reduced blood flow to bone and may contribute to bone loss and its fragility [7,8]. Main clinical consequences of diabetic osteopathy are osteoporosis-associated bone fractures, especially at the hip and spine, which may be associated with serious complications such as progressive pain, disability, and death [9]. Diagnosis of diabetic osteopathy is challenging as current methods for fracture prediction such as BMD T-score and the Fracture Risk Assessment (FRAX) tool underestimate fracture risk for patients with T2D [10,11], particularly when disease duration is longer than 10 years [12].

FH is the most common cause of secondary osteoporosis in men and is present in up to 20% of men with symptomatic vertebral fractures and 50% of elderly men with hip fractures [13,14]. Low total testosterone (TT) levels were reported to be present in 40–50% of men with T2D [15]. Consequently, ADA guidelines recommend assessment of TT level in T2D patients showing symptoms of hypogonadism and included testosterone therapy (TTh) among the treatment modalities which may prevent T2D [16]. It is sufficiently established that patients with FH and metabolic syndrome benefit from TTh in cardio-metabolic and glycemic control and body composition [17–24].

Also, patients with FH treated with TTh show statistically significant improvements in BMD at spine and hip [25]. Testosterone modulates bone remodeling and increases muscle mass which has anabolic effects on bone [14].

Bone turnover can be assessed by measuring biochemical markers, which reflect bone turnover process and hence mirror bone resorption and formation processes and are known to be predictors of fracture in individuals without T2D [26]. Bone turnover markers (BTMs) include markers of both bone formation—procollagen type 1 amino terminal propeptide (PINP), osteocalcin, and bone-specific alkaline phosphatase (BAP), and bone resorption—C-telopeptide of type I collagen (CTX), N-terminal cross-linked telopeptide of type 1 collagen (NTX), and tartrate-resistant acid phosphatase (TRAP). Men with T2D and FH have smaller bone size and lower bone turnover rate [27].

Dual-energy X-ray absorptiometry (DXA) scan is considered to be the “gold standard” or most accurate test for measuring BMD for various reasons. DXA is simpler and rapid compared to micro-computed tomography for quantitative analysis of change in trabecular bone of test subjects [28]. Quantitative-computed tomography (QCT) is the alternative methodology to DXA for accurate BMD assessment but exposure to radiation with QCT is higher compared to DXA [29].

We have recently shown that TTh in obese men with T2D and obesity improved glycemia, insulin resistance, vascular function and morphology, and improved symptoms of hypogonadism [30,31]. However, whether TTh exerts any noteworthy effects on bone in population of obese males with both T2D and FH has not been sufficiently established to date. One of the aims of our study was therefore to evaluate changes due to TTh to BTMs and lumbar spine and hip BMD in obese males with T2D and FH.

Methods

Study design

Study on Effects of TTh in Hypogonadal obese type 2 diabetic male patients (SETH2 study) was a two-part single-center prospective clinical study (first year double-blind, randomized, placebo-controlled trial; second year open-label follow-up) conducted from January 2014 to March 2018 at General Hospital Celje (Slovenia). Study has been approved by the National Medical Ethic Committee (54/04/12) of Republic of Slovenia and was conducted in accordance with the Declaration of Helsinki and with all local applicable laws and regulations. Written informed consent was obtained from all study subjects prior to their participation in the study. The study has been registered at ClinicalTrials.gov (identifier: NCT03792321).

Study population

Fifty-five obese Caucasian male patients with confirmed symptomatic FH, aged 40–70 years and with T2D participated in SETH2 trial. FH was diagnosed as a biochemical deficiency of circulating testosterone levels (TT below 11 nmol/L and free testosterone below 220 pmol/L) on at least two separate morning measurements after an overnight fast in addition to exhibiting at least two symptoms of sexual dysfunction (less frequent morning erections, erectile dysfunction and decreased libido) [32–36]. Inclusion criteria for the SETH2 trial were: male, confirmed FH, age >35 years, BMI ≥ 30 kg/m², T2D treated exclusively with non-insulin anti-diabetic medications (metformin, sulfonylureas). Participants did not use concomitant medications that can reduce weight, BMI and waist circumference.

Exclusion criteria were: previously treated hypogonadism, history of current prostate or breast cancer, severe benign prostatic hyperplasia or prostate-specific antigen (PSA) level >4.0 μg/L, severe heart failure, acute coronary event or procedure during the six months prior to the study, chronic obstructive lung disease, severe obstructive sleep apnea with apnea–hypopnea index >30/h, and active infection.

Inclusion and exclusion criteria were reviewed at screening visit where patients were asked about symptoms and signs suggestive of hypogonadism in accordance with the clinical practice guidelines [32,34] and underwent clinical history and physical examination. We performed baseline clinical examination, anthropometric measurements and took blood samples for biochemical and hormonal tests. Data were collected at the Diabetic outpatient clinic at General Hospital Celje (Slovenia).

Study protocol

Subjects were randomly assigned in a concealed 1:1 allocation to either group T or P. An independent statistician generated the randomization sequence. Trial participants and trial investigators were blinded to treatment allocation for the first phase of the study. Group P received placebo and group T received testosterone in the form of intramuscular testosterone undecanoate (TU; Nebido^® 1000 mg, Bayer AG) injections during the first year (first part) of the study employing the following protocol: first injection of TU/placebo was administered at the first visit (baseline), second injection 6 weeks later (second visit), and each subsequent injection 10 weeks after the previous injection.

All study participants received TU during the second year of the study (second part) since placebo control was limited to one year due to ethical and medical concerns over withholding TTh from hypogonadal patients.

Safety parameters (complete blood count, PSA, markers of hepatic and renal functions) were performed at 0, 3, 6, 12, 15, 18 and 24 months. All patients were instructed to report any side effects during the treatment. Patients were not given any advice or instructions regarding dietary or other lifestyle changes prior to and during the study.

Use of non-insulin anti-diabetic medications was permitted and patients continued on the same (previously prescribed) anti-diabetic medication throughout the course of the study without any dose adjustments.

Methods

Biochemical and anthropometric parameters

All patients underwent clinical, anthropometric, biochemical and hormonal assessment at baseline, after first year and after second year of the study. Fasting blood samples were taken between 7.00 and 11.00am to measure serum TT, sex hormone-binding globulin (SHBG), albumin, CTX and PINP, estradiol, 25-hydroxyvitamin D, luteinizing hormone, follicle-stimulating hormone, fasting plasma glucose, HbA₁c, lipids (total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides), PSA, routine blood tests (complete blood count, electrolytes, urea, creatinine, liver tests). We calculated the following parameters: BMI, homeostasis model assessment of insulin resistance (HOMA-IR) index, calculated free testosterone (cFT) and bioavailable testosterone (BT). Assessments were made at the baseline visit, at 12 months and at 24 months.

Assessment of BMD

BMD of the femoral neck and lumbar spine L1–L4 was measured by DXA at the beginning and at the end of the trial (after two years). All DXA scans were taken on a Hologic QDR-1000 Plus (Waltham, MA, USA). BMD results are reported as areal density (in g/cm²). Osteoporosis was defined by a T-score more than 2.5 SDs below the mean value for young adult reference data on DXA, whereas osteopenia was defined by a T-score between 1.0 and 2.5 SDs below the mean value for young adult reference data. Fracture history was also retrieved from patient records.

Assessment of testosterone

TT was assessed using IMMULITE 2000 chemiluminescent enzyme immunoassay (Siemens Healthcare GmbH). Intra-assay coefficients of variation (CVs) in the relevant result range are 11.7% (at 2.99 nmol/L), 10.0% (5.27 nmol/L), 8.3% (9.70 nmol/L), 7.2% (14.35 nmol/L), and 5.1% (34.36 nmol/L). Inter-assay CVs at those same respective means are 13.0%, 10.3%, 9.1%, 8.2% and 7.2% as reported by the manufacturer. Vermeulen’s formulae were used to calculate cFT and BT values using SHBG, serum albumin and TT level data [37].

Statistical analysis

Data were compared within both study groups to determine CTX and PINP deviations from the baseline at statistically significant level (α = 0.05). Normality of distribution of residuals was assessed with Shapiro–Wilk test. CTX data were log-transformed before performing the analysis. Repeated measures analysis of variance (RM-ANOVA) was used to examine log-transformed CTX and PINP results. Mauchly’s test was used to test the assumption of sphericity. Adjusted p values are reported for post hoc tests with Bonferroni correction. BMD data were compared using paired samples t-test. SPSS Statistics 22.0 software (IBM Corporation, Armonk, NY, USA) was used to perform statistical analysis.

Outcomes

Main outcome measures reported in this text are: 1) changes in serum BTMs (CTX, PINP), 2) changes in lumbar spine and hip BMD and 3) changes in testosterone levels (TT, cFT, BT).

Results

Participants

Baseline demographic, anthropometrical and biochemical parameters of the study population are outlined in Table 1. Participants exhibited decreased levels of 25-OH-vitamin D, TT, cFT, BT. With the exception of one study participant, none were receiving any anti-osteoporotic therapy, vitamin D or calcium supplements or any other drug known to influence bone and calcium metabolism. Above-mentioned exception, who was already receiving anti-osteoporotic therapy (bisphosphonates) prior to his enrollment in our trial, was excluded from statistical analysis of effects of TTh on bone, bringing group T participant count to a total of 27 subjects for BTM assessments (as reflected by P group participant count in Table 2). This participant was also excluded from selection of 32 patients (16 from each study group) who underwent BMD assessment. Of these 32 patients, 8 had normal BMD, 7 were diagnosed with osteopenia and 1 with osteoporosis in group P, while 12 had normal BMD, 2 had osteopenia and 2 osteoporosis in group T.

Table 1. Study population baseline characteristics: key anthropometrical, biochemical and clinical values.

	Study population	Group P	Group T
Subjects recruited (n)	55	27	28
Age (years)	60.16 ± 7.234	62.19 ± 5.903	58.21 ± 7.941
Type 2 diabetes duration (years)	9.75 ± 6.662	10.67 ± 7.248	8.86 ± 6.041
Smoking status
Current smoker	13 (23.6%)	7 (25.9%)	6 (21.4%)
Never smoked	21 (38.2%)	13 (48.2%)	8 (28.6%)
Quit smoking	21 (38.2%)	7 (25.9%)	14 (50.0%)
Arterial hypertension	49 (89.1%)	24 (88.9%)	25 (89.3%)
Hyperlipidemia	55 (100.0%)	27 (100.0%)	28 (100.0%)
Anti-hypertensive therapy	49 (89.1%)	24 (88.9%)	25 (89.3%)
Statin therapy	29 (52.7%)	15 (55.6%)	14 (50%)
Thiazide therapy	8 (14.5%)	2 (7.4%)	6 (21.4%)
Anti-resorptive therapy	1 (1.8%)	0 (0.0%)	1 (3.6%)
Body mass index (kg/m^2)	31.9 (31–36.3)	31.4 (30.8–32.9)	33.6 (31–37.05)
Waist circumference (cm)	116.56 ± 4.969	116.64 ± 4.962	116.48 ± 5.066
Systolic blood pressure (mmHg)	136.36 ± 12.036	138.15 ± 13.240	134.64 ± 10.709
Diastolic blood pressure (mmHg)	78.18 ± 5.558	78.89 ± 5.250	77.5 ± 5.853
Fasting plasma glucose (mmol/L)	9.84 ± 1.446	9.60 ± 1.442	10.06 ± 1.439
Glycated hemoglobin A1c (%)	8.00 ± 0.914	7.89 ± 0.769	8.12 ± 1.037
Serum insulin (mIU/L)	25.47 ± 14.805	24.89 ± 13.899	26.03 ± 15.864
Homeostatic Model Assessment of Insulin Resistance index	11.08 ± 6.896	10.70 ± 6.522	11.45 ± 7.339
Total testosterone (nmol/L)	7.595 ± 1.7120	7.959 ± 1.3391	7.243 ± 1.9682
Calculated bioavailable testosterone (nmol/L)	4.167 ± 1.0505	4.611 ± 0.9542	3.739 ± 0.9711
Calculated free testosterone (pmol/L)	173.5 ± 46.07	192.1 ± 44.13	155.5 ± 41.11
Estradiol (nmol/L)	0.0904 ± 0.05192	0.0904 ± 0.0467	0.0904 ± 0.05738
25(OH)D (nmol/L)	69.78 ± 24.931	68.63 ± 23.931	70.88 ± 26.249
Total cholesterol (mmol/L)	5.311 ± 0.9327	5.307 ± 0.9719	5.314 ± 0.9111
LDL-cholesterol (mmol/L)	2.795 ± 0.8577	2.796 ± 0.9545	2.793 ± 0.7707
HDL-cholesterol (mmol/L)	1.027 ± 0.2718	1.048 ± 0.3167	1.007 ± 0.2243
Triglycerides (mmol/L)	3.180 ± 2.4503	3.515 ± 3.1508	2.857 ± 1.4926
Aging Male Symptoms Score	39 (35–43)	37 (35–41)	40 (35.5–47.75)
Serum creatinine (μmol/L)	80.71 ± 13.162	80.07 ± 12.661	81.32 ± 13.832
Lumbar bone mineral density (g/cm^2)	0.9179 ± 0.15180	0.9267 ± 0.17693	0.9093 ± 0.12564
Femoral neck bone mineral density (g/cm^2)	1.0334 ± 0.18740	0.9955 ± 0.16072	1.0699 ± 0.20626

Values are reported as mean ± SD when normally distributed, as median (interquartile range) when distribution was skewed, and as count (percentage within population/group) for clinical parameters.

Table 2. Changes in bone turnover markers in obese males with functional hypogonadism and type 2 diabetes on testosterone therapy (TTh) over the 2-year course of the study.

	Group P (n = 27)	p value	Group T (n = 27)	p value
CTX (pmol/L)
Baseline	1055 (676–1344)		897 (679–1506)
After 1 year	911 (556–1152)	0.053	523 (364–835)	<0.001*
After 2 years	453 (365–665)	0.001*	388 (265–619)	<0.001*
PINP (μg/L)
Baseline	32.7 ± 8.6		28.7 ± 7.9
After 1 year	32.1 ± 9.4	1.000	25.3 ± 6.8	0.023*
After 2 years	28.4 ± 7.0	0.030*	20.8 ± 6.1	<0.001*

Group P participants were receiving placebo during first year of the study and were then switched to testosterone undecanoate. Group T participants were receiving testosterone undecanoate throughout the entire 2-year course of the study. CTX levels are reported as median (inter-quartile range) and PINP levels as mean ± SD. p Values (with Bonferroni correction applied) marked with an asterisk (*) represent statistically significant change from the baseline.

Changes in outcome measures

CTX values are reported as median (interquartile range) and PINP as mean ± SD. CTX and PINP results are shown in Table 2.

With assumption of sphericity not met for log-transformed CTX data for either group P (p < 0.001) or T (p = 0.013), Greenhouse-Geisser correction was applied to RM-ANOVA degrees of freedom. Log-transformed CTX levels changed significantly in both group P (F(1.344, 34.957)=15.510, p < 0.001) and T (F(1.196, 40.205)=18.483, p < 0.001). Post hoc pairwise test employing Bonferroni correction showed that log-transformed CTX levels decreased from the baseline in group P after second year of our study—after one year of placebo, followed by one year of TTh (p = 0.001), while the change in T group occurred after the first year of the trial—after one year of TTh (p < 0.001). CTX decreased from 1055 (676–1344) to 453 (365–665) pmol/L in group P and from 897 (679–1506) to 523 (364–835) pmol/L in group T.

Assumption of sphericity was not violated for PINP data for either group P (p = 0.449) or T (p = 0.193). RM-ANOVA confirmed that PINP levels changed between time points in group P (F(2, 52)=4.081, p = 0.023) and T (F(2, 52)=15.055, p < 0.001). Post hoc pairwise test with Bonferroni correction confirmed that PINP levels decreased from the baseline of 32.7 ± 8.6 μg/L after second year of the trial in group P to 28.4 ± 7.0 μg/L (p = 0.030), and from 28.7 ± 7.9 to 25.3 ± 6.8 μg/L after first year in group T (p = 0.023), followed by further change to 20.8 ± 6.1 μg/L (p < 0.001) after two years of TTh.

Sixteen randomly selected individuals from each study group were examined for changes in femoral hip and lumbar spine BMD after two years of the study (Table 3). The only statistically significant change (increase) of BMD was lumbar spine BMD in 16 group T subjects (p = 0.019); change in femoral hip BMD was not significant (p = 0.194). There were no statistically significant changes in 16 group P subjects after two years of the study in either femoral hip BMD (p = 0.619) or lumbar spine BMD (p = 0.390).

Table 3. Changes in bone mineral density in obese males with functional hypogonadism and type 2 diabetes on testosterone therapy (TTh) over the 2-year course of the study.

	Group P (n = 16)	p value	Group T (n = 16)	p value
DXA femoral hip (g/cm^2)
Baseline	0.9765 ± 0.18417		0.9270 ± 0.12338
After 2 years	1.0023 ± 0.12486	0.619	0.9940 ± 0.21152	0.194
DXA lumbar spine (g/cm^2)
Baseline	1.0328 ± 0.14955		1.1231 ± 0.22574
After 2 years	1.0697 ± 0.19205	0.390	1.1982 ± 0.23891	0.019*

Group P participants were receiving placebo during first year of the study and were then switched to testosterone undecanoate. Group T participants were receiving testosterone undecanoate throughout the entire 2-year course of the study. Values are reported as mean ± SD. p Values marked with an asterisk (*) represent statistically significant change from the baseline.

Changes in testosterone levels are detailed in Table 4. Group T mean TT increased from 7.24 ± 1.97 to 17.04 ± 3.07 nmol/L after one year of TTh and to 23.50 ± 4.91 nmol/L after two years. All group T participants but one reached serum TT concentration above the 11 nmol/L reference value after first year of TTh while the sole exception went from baseline TT of 7.5–10.6 nmol/L. Statistically significant increase of mean TT in group P after 12 months (from 7.96 ± 1.34 to 9.83 ± 2.21 nmol/L) is of little clinical significance as it is still below 11 nmol/L reference value. Mean TT of group P patients increased considerably more (to 17.92 ± 2.21 nmol/L) after one year of TTh (after two years of the study). All group P participants reached TT above 11 nmol/L at this point. The calculated values of BT and cFT correlated highly with values of the TT, with clinically relevant increase manifesting in both groups after the first year of TTh (after one year of the study for group T and after two years for group P).

Table 4. Changes in testosterone levels in obese males with functional hypogonadism and type 2 diabetes on testosterone therapy (TTh) over the 2-year course of the study.

	Group P (n = 27)	p value	Group T (n = 28)	p value
Total testosterone (nmol/L)
Baseline	7.96 ± 1.34		7.24 ± 1.97
12 months	9.83 ± 1.51	<0.001*	17.04 ± 3.07	<0.001*
24 months	17.92 ± 2.21	<0.001*	23.50 ± 4.91	<0.001*
Calculated bioavailable testosterone (nmol/L)
Baseline	4.61 ± 0.95		3.74 ± 0.97
12 months	5.73 ± 0.89	<0.001*	9.50 ± 1.89	<0.001*
24 months	10.72 ± 1.84	<0.001*	13.93 ± 3.39	<0.001*
Calculated free testosterone (pmol/L)
Baseline	192.07 ± 44.13		155.54 ± 41.11
12 months	242.00 ± 42.23	<0.001*	403.84 ± 90.34	<0.001*
24 months	454.56 ± 87.78	<0.001*	594.04 ± 147.09	<0.001*

Group P participants were receiving placebo during first year of the study and were then switched to testosterone undecanoate. Group T participants were receiving testosterone undecanoate throughout the entire 2-year course of the study. Values are reported as mean ± SD. p Values marked with an asterisk (*) represent statistically significant change from the baseline.

Safety parameters

TTh led to expected increases in hemoglobin and hematocrit (Hct) concentrations, which remained within normal reference range during the entire 2-year observation period.

Adverse events

No adverse events or side effects of TTh have been observed over the 2-year course of this trial. An elaborate overview of safety aspect of our trial has been published separately [38].

Discussion

In our prospective clinical study of obese males with T2D and FH, we report changes on BTMs and lumbar spine and hip BMD due to TTh. Our results show decrease of CTX and PINP levels in both groups, after the second year (one year of placebo, followed by one year of TTh) in group P and after first year of the trial in group T (after one year of TTh), which implies that TTh modulates BTMs. Results on impact of TTh on BTMs in the literature are inconsistent. While some authors found significant changes in both osteocalcin and urinary N‐telopeptide [39], some only found changes in urine N‐telopeptide [40], CTX [41,42] and others found no statistically significant changes in either marker [43].

One important reason for the mixed results of TTh on bone in the literature is the duration of treatment. Behre et al. [44] did not observe any changes following six month placebo-controlled phase in their study with transdermal testosterone. In our study CTX decreased at statistically significant level after one year of introduction of TTh in both study groups, indicating a slowing down of bone destruction. Patients from T group who underwent DXA examination showed increased lumbar BMD at statistically significant level, while no such change was observed in patients from P group, showing apparent differences in BMD between patients treated with testosterone for two years and those with placebo followed by only one year of testosterone. This is in line with a study of effects of TTh on BMD in obese patients over 50 years of age without T2D where transdermal testosterone was employed during first year, followed by intramuscular testosterone during second year of the trial [45]. Results of the T Trial study have shown that one year of TTh increased volumetric BMD (vBMD) of trabecular bone in the spine by 6.8% more than placebo, and increased the estimated bone strength of trabecular bone in the spine by 8.5% more than placebo. Testosterone also substantially increased whole bone vBMD, the strength of the spine and trabecular and whole bone vBMD, and the strength of the hip [46]. As previously reported TTh normalized testosterone levels in both groups within 12 months along with improvements of components of metabolic syndrome (glycemia, insulin resistance, total cholesterol level) [30,31]. Taken together, these findings suggest that long-term TTh can increase BMD in obese males with T2D and FH. This increase in BMD would be expected to lower fracture risk in hypogonadal men. Long-term observational study by Haider demonstrated that T-scores improved over 6 years with statistical significance compared to the previous year in patients with osteoporosis [47]. RCT lasting 36 months reported increase in lumbar spine BMD, suggesting that longer treatment with TTh may be associated with a more robust response, however—in contrast to our study—testosterone enanthate was used in that trial and patients did not achieve concomitant weight loss, nor improved insulin sensitivity [48]. Participants of this study were not obese but only overweight. Clinically meaningful weight loss as a result of TTh has only been reported in overweight and obese men being treated for four years and longer [49].

Another important issue of TTh studies is also administration route of the therapy and medication adherence. Intramuscular injections are more effective than transdermal preparations in this regard [50,51]. Adherence to therapy was 100% in our study.

Testosterone acts on bone metabolism directly via binding to the androgen receptor (AR) or indirectly via aromatization to estradiol to activate estrogen receptor-a (ER-a) and/or estrogen receptor-b (ER-b) [52]. By its effect on periosteal apposition, testosterone contributes to bone mass, which explains the wider or bigger bones in men compared to women [53]. Through its conversion to estradiol by the enzyme aromatase in adipose tissue testosterone plays an important role in mediating age-related bone loss in men [54]. In vivo study by Movérare et al. investigated differential effects of AR and ER-a activation on bone in orchidectomized (ORX) adult male mice. Results indicate that activation of ER-a preserved the trabecular number, trabecular thickness, cortical thickness, and cortical density. In contrast, activation of AR only preserved trabecular number, which suggests that the bone-sparing effect of ER-a activation was distinct from the effects of AR activation in adult ORX male mice [55]. In human observational studies, estrogen was found to have an independent and stronger correlation with BMD compared to testosterone [56].

Testosterone also has an anabolic effect because of its ability to stimulate osteoblastic differentiation and proliferation [57]. Both androgen and estrogen receptors have been located in several bone cells including osteoblasts, osteoclasts, and mesenchymal stromal cells, which differentiate into osteoblasts [58]. By decreasing apoptosis, testosterone increases the lifespan of osteoblasts through its action on interleukin-6 production and is able to stimulate the proliferation of osteoblast progenitors and the differentiation of mature osteoblasts [58]. Also, androgens stimulate osteoblast proliferation and differentiation and an increase of osteoprotegerin, a product of osteoblasts, all of which result in decrease of osteoclast formation and bone resorption [59]. Osteoprotegerin and TTh (testosterone cypionate) recovered lumbar and femoral BMD in ORX rats after 8 weeks of treatment [60].

Findings from human studies suggest that selective estrogen receptor modulators (SERMs) are beneficial in increasing BMD—particularly at lumbar spine, total hip and femoral neck—by suppressing bone resorption and reduction of BTMs. However, studies on effects of SERMs on fracture risk are limited [61].

Increased red blood cell mass (erythrocytosis) is the most common adverse event associated with TTh in clinical practice and in testosterone trials. TTh-induced erythrocytosis is associated with stimulation of erythropoietin and reduced ferritin and hepcidin concentrations [62,63]. As previously published, no adverse events or side effects of TTh have been observed over the 2-year course of this trial [38]. We showed that Hct increased gradually from the baseline in both groups within 3–6 months after the introduction of TTh in each study group. This is in line with effects that TTh is reported to exert on Hct [64].

Main limitations of our study were relatively small sample size and the short-term observation period as any changes in BMD due to TTh have been shown to only manifest after prolonged (>1 year) administration of TTh. Its strengths are an exclusive focus on population of men with T2D and FH, who are at an elevated risk for osteopenia and osteoporosis, the placebo-controlled first phase of the trial where effects of TTh on BTMs can be observed, and its second year extension where changes on BTM can again be observed in group P, as can be changes in BMD, which only began manifesting after prolonged (>1 year) administration of TTh in group T.

Larger, longer term randomized trials are required to demonstrate whether the increase in BMD observed following TTh indeed reduces fracture risk in hypogonadal men with T2D. Future studies should also be directed at investigating the underlying mechanism of this association and to use this knowledge for improvement of personalized treatment strategies to prevent diabetic osteopathy. Whether supplementation of calcium and vitamin D in combination with TTh would result in even greater increase in BMD in comparison to TTh alone is another potential subject of future research.

Acknowledgements

Bayer AG (Berlin, Germany) provided testosterone and placebo for this study, but had no role in design of the study protocol, data collection, data analysis or writing of this manuscript.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.