Hypogonadal men with type 2 diabetes mellitus have smaller bone size and lower bone turnover

Abstract

Introduction:

Both hypogonadism and type 2 diabetes mellitus (T2D) are associated with increased fracture risk. Emerging data support the negative effect of low testosterone on glucose metabolism, however, there is little information on the bone health of hypogonadal men with diabetes. We evaluated the bone mineral density (BMD), bone geometry and bone turnover of hypogonadal men with T2D compared to hypogonadal men without diabetes.

Materials and Methods:

Cross-sectional study, men 40–74 years old, with average morning testosterone (done twice) of < 300 ng/dl. Areal BMD (aBMD) was measured by DXA; volumetric BMD (vBMD) and bone geometry by peripheral-quantitative-computed-tomography; serum C-telopeptide (CTX), osteocalcin, sclerostin and sex hormone-binding globulin (SHBG) by ELISA, testosterone and 25-hydroxyvitamin D (25OHD) by automated immunoassay and estradiol by liquid-chromatography/mass-spectrometry. Groups were compared by ANOVA adjusted for covariates.

Results:

One-hundred five men, 49 with and 56 without diabetes were enrolled. Adjusted vBMD at 38% tibia was higher in diabetic than non-diabetic men (857.3 ± 69.0 mg/cm³ vs. 828.7 ± 96.7 mg/cm³, p = 0.02). Endosteal (43.9 ± 5.8 mm vs. 47.1 ± 7.8 mm, p = 0.04) and periosteal (78.4 ± 5.0 mm vs. 81.3 ± 6.5 mm, p = 0.02) circumferences and total area (491.0 ± 61.0 mm² vs. 527.7 ± 87.2 mm², p = 0.02) at 38% tibia, were lower in diabetic men even after adjustments for covariates. CTX (0.25 ± 0.14 ng/ml vs. 0.40 ± 0.19 ng/ml, p < 0.001) and osteocalcin (4.8 ± 2.8 ng/ml vs. 6.8 ± 3.5 ng/ml, p = 0.006) were lower in diabetic men; there were no differences in sclerostin and 25OHD. Circulating gonadal hormones were comparable between the groups.

Conclusion:

Among hypogonadal men, those with T2D have higher BMD, poorer bone geometry and relatively suppressed bone turnover. Studies with larger sample size are needed to verify our findings and possible even greater risk for fractures among hypogonadal diabetic men.

1. Introduction

The interplay between testosterone and glucose metabolism has been suggested by several studies [1,2]; men with low testosterone had impaired glucose tolerance, while men with high testosterone levels had a 42% lower risk of developing type 2 diabetes mellitus (T2D) [2]. Moreover, testosterone therapy improves glycometabolic parameters in hypogonadal individuals [3,4]. Studies have also shown that T2D and obesity exacerbate low testosterone production such that both are considered as risk factors for the development of hypogonadism [2,5–7].

A decrease in testosterone levels is associated with bone loss leading to osteoporosis and an increase in the incidence of fragility fractures [8,9]. Not surprisingly, testosterone deficiency has been reported in over half of elderly men with a history of hip fracture [10,11]. On the other hand, despite the normal to high bone mineral density (BMD), patients with T2D have a higher risk for fractures compared to patients without diabetes [12–14]. Because of the increased risk for fractures despite relatively higher BMD in individuals with T2D, there is intense interest on modalities other than BMD testing by dual energy x-ray absorptiometry (DXA) to assess the bone health of these patients [12]. Measures reflecting bone quality such as bone microarchitecture, bone material strength index, bone geometry and biochemical markers of bone turnover are considered useful in this context [12,14–18], although at present, only bone turnover markers are approved for clinical use. Interestingly, similar to what was previously reported in patients on long term treatment with bisphosphonates [19], patients with T2D have suppressed bone turnover markers compared to those without diabetes [20], which is considered as a risk factor for atypical fractures [21,22].

As both hypogonadism and T2D are detrimental to bone, it is possible that the combination of both may portend a worse skeletal health. Prior reports show that low testosterone can be found in 33% to as much as 64% [7,23,23,24] of patients with T2D suggesting that a significant number of men could be at a greater risk for fractures if indeed the presence of both conditions confers an even higher risk than either one alone.

In a previous study in a group of men with T2D, Dhindsa et al. reported no difference in areal BMD (aBMD) between eugonadal and hypogonadal subjects [25]. This limited study suggests that among patients with T2D, testosterone deficiency may not have an effect on BMD. However, no other parameter of bone health was evaluated. In this study we evaluated other parameters of bone health, such as bone geometry and bone turnover markers, among hypogonadal patients with T2D compared to those without diabetes. Given the high fracture risk among men with hypogonadism and the associated increased mortality from a hip fracture [26], the identification of comorbidities that may exacerbate this risk in subjects with low testosterone is crucial [27]. As T2D is associated with negative skeletal effects, it is possible that among hypogonadal men, the subgroup with T2D will have a worse skeletal profile compared to those without diabetes. Thus, the objective of this study is to evaluate the effect of T2D on the bone turnover and bone geometry in men with hypogonadism compared to hypogonadal men without diabetes.

2. Materials and methods

2.1. Study design and study population

This study is a cross-sectional analysis of baseline data from subjects who volunteered to participate in a study investigating the effect of genetics on testosterone therapy in male veterans with low total testosterone, defined as an average total testosterone of <300 ng/dl from 2 samples taken in the morning [28]. This study was conducted at the New Mexico VA Health Care System and at the Michael E. DeBakey VA Medical Center in accordance with the guidelines in the Declaration of Helsinki for the ethical treatment of human subjects. The protocol was approved by the University of New Mexico and the Baylor College of Medicine Institutional Review Board. Participants were recruited from patients attending the Endocrine, Urology and Primary Care Clinics of the New Mexico VA Health Care System and Michael E. DeBakey VA Medical Center. This was accomplished either through flyers or letters to physicians about patients who may qualify for the study. A written informed consent was obtained from each subject. Inclusion criteria included: males between 40 and 75 years of age with no medical problems that may prevent them from finishing the study. Exclusion criteria included: treatment with bone-acting drugs (e.g. bisphosphonates, glucocorticoids, sex-steroid compounds, selective estrogen receptor modulators, androgen deprivation therapy and anticonvulsants), and finasteride. Additional exclusion criteria included: osteoporosis and history of fragility fractures or diseases known to affect bone metabolism such as: hyperparathyroidism, chronic liver disease, uncontrolled or untreated hyperthyroidism and significant renal impairment (creatinine of >1.5 mg/dl). Those with a history of prostate cancer, breast cancer and untreated sleep apnea also met the criteria for exclusion.

The presence of T2D was ascertained from the medical history, intake of medications for diabetes, and for those who did not have the first 2 criteria, a hemoglobin A1c (A1c) at study entry of ≥6.5%. None of our subjects had type 1 diabetes mellitus.

A group of eugonadal patients who satisfied all the above inclusion criteria but with testosterone levels ≥300 ng/dl, was included as a normal comparison group in the sub-analysis of the bone turnover markers involving the entire population.

2.2. Body mass index (BMI)

Body weight was measured using a standard weighing scale and height was obtained using a stadiometer. BMI was calculated as body weight in kilograms (kg) divided by the square of the height in meters (m²) and expressed as kg/m².

2.3. Areal bone mineral density (aBMD) by dual energy x-ray absorptiometry (DXA)

Areal BMD was measured by DXA on the lumbar spine and the proximal femur using Hologic Discovery (Hologic Inc., Bedford, MA, USA). Regions of interest in the femur include the total hip and femoral neck. The coefficients of variation (CV) at our centers are ~1.1% for the lumbar spine and ~1.2% for the proximal femur [29].

2.4. Volumetric bone mineral density (vBMD) by peripheral quantitative computed tomography (pQCT)

Peripheral quantitative computed tomography was performed on the left tibia (except for those who had surgery or metal plates on the left tibia where the right tibia was used) using Stratec XCT-2000 (Stratec GmbH, Pforzheim, Germany) as previously described [30]. The voxel size was 500 μm, scan speed 30 mm/s and slice thickness 2.4 mm. The X-ray current was between 0.14 and 0.22 mA and the voltage was 58 kV. Scanning was performed by first obtaining a scout view on the lower leg to identify anatomical landmarks. From the scout view, the cortical endplates of the tibia were identified visually and used as anatomical landmarks. Then, a reference line was fixed to bisect the distal tibial endplate in a transverse plane set at right angles to the long axis of the tibia. Scans were then obtained at regions of interest equal to a percentage of the limb length measured from the distal to proximal direction. A slice was each taken at the 4% and 38% for assessment of bone volume and bone quality of the tibia.

For bone assessments, trabecular and total vBMD, bone mineral content (BMC) and bone area were measured distally (4%) while cortical and total vBMD, BMC, bone area and cortical thickness more proximally (38%). Short term CV for these measurements in our center is 0.96%. Different thresholds were applied to separate each of trabecular, cortical and soft tissues from each other. At 4% tibia, we applied the following thresholds: 180 mg/cm³ and 45% of the area for the evaluation of total and trabecular vBMD, BMC and total area; and thresholds 280 mg/cm³ and 400 mg/cm³ for the analysis of trabecular area. The threshold for separating trabecular and cortical bone at 38% was 710 mg/cm³. All images were reviewed for motion grade by the PI and the researcher acquiring the scan at the end of each study.

2.5. Biochemical measurements

Fasting blood samples were collected at baseline; serum samples were extracted and stored at −80 °C until analysis except for the screening testosterone levels. Baseline serum testosterone represents an average from 2 determinations taken 30 min apart between 8 and 11 AM, and measured using automated immunoassay, detection range 10–3200 ng/dl (Vitros®, Ortho Clinical Diagnostics, Rochester), NY. The coefficient of variability for testosterone assay is ≤20% for testosterone levels of <50 ng/dl, and ≤10% for testosterone levels between 200 and 1000 ng/dl. Twenty-five-hydroxyvitamin D (25OHD) was measured using automated immunoassay (IDS-ISYS, France). Fasting glucose was measured using Unicel DxC 800 Auto-analyzer (Beckman Coulter, Fullerton, CA, USA); A1c by high performance liquid chromatography (Tosoh G8, South San Francisco, CA, USA). The following were measured using enzyme-linked immunosorbent assay (ELISA) kits: sex hormone-binding globulin (SHBG) (Alpco Diagnostics, Salem, NH), C-terminal telopeptide of type I collagen (CTX, a marker of bone resorption) (Crosslaps; Immunodiagnostic System Inc., Gaithersburg, MD), osteocalcin, a marker of bone formation, (Metra OC; Quidel Corporation, San Diego, CA), and sclerostin (TECOmedical Sclerostin HS Enzyme Immunoassay Kit, Quidel Corp, San Diego, CA). The coefficients of variation (CVs) for the above assays in our laboratory are <10% and <3.5% for A1c. Estradiol was measured by liquid chromatography/mass spectrometry (Mayo Clinic Laboratories, Mayo Clinic, Rochester, MN) with assay sensitivity of 0.23 pg/ml to 405 pg/ml, intra-assay CV of 1.4% to 11.8% and inter-assay CV of 4.8% to 10.8%. The free estradiol index (FEI) was calculated as the molar ratio of total estradiol to SHBG (pmol/nmol) [31]. The Free Androgen Index (FAI) and the estradiol to testosterone ratio (E2/T) were calculated using the formula as described by Sowers et al. [32]. FAI = 100 × T (ng/dl)/28.84 × SHBG (nM) and E2/T = 10 × estradiol/testosterone, both of which are unit-free [32].

2.6. Statistical analysis

Results are expressed as means ± SD for the text and tables, and means ± SE for the figure. A p value of <0.05 was considered statistically significant. Group comparisons between those with T2D and those without diabetes were analyzed using analysis of variance (ANOVA) without and with adjustments for covariates; model 1: no adjustment; model 2: adjusted for age, BMI and estradiol levels; model 3: model 2 plus adjustment for diabetes medication use. Correlations among the different variables were analyzed by simple correlation analysis. The data were managed using Excel 2010 (Microsoft, Redmond, WA) and were analyzed using SAS version 9.3 (SAS Institute, Inc., Cary, NC, USA).

3. Results

Our study population consisted of 105 men with an average age of 59.6 ± 8.4 years, average testosterone levels from 2 samples taken in the morning 30 min apart of 210.4 ± 63.3 ng/dl (Reference in our laboratory: 262–1593 ng/dl), mean BMI of 32.3 ± 5.5 kg/m² (range: 21.8–48.6 kg/m²) and mean A1c of 6.7 ± 1.7%. Although the primary criteria for enrollment was an average morning testosterone level of <300 ng/dl, all of the subjects recruited in the study had at least one complaint that could be related to lack of testosterone, thus can be classified as hypogonadal [28]. The baseline characteristics of most of the participants have been reported previously [29]. Fifty percent of the participants had a history of smoking, while 24% were current smokers.

Forty-nine subjects had T2D, while fifty-six did not have diabetes. Among the 49 patients with T2D, 42 were treated with blood-glucose lowering agents: 2 were treated with sulfonylureas alone, 8 with metformin alone, 6 were on sulfonylurea + metformin, 7 were on insulin alone, 1 patient was on a combination of sulfonylurea + insulin, 11 were on metformin + insulin, 4 were on a combination of sulfonylurea + metformin + insulin, 2 were on metformin + incretin agonists, and 1 patient was on combination of metformin + insulin + incretin agonist.

The group with T2D was significantly older and showed a trend toward a higher BMI compared to those without diabetes. As expected, subjects with T2D had a significantly higher hemoglobin A1c (Table 1). There were no significant differences in estradiol, testosterone, SHBG, FEI, FAI and E2/T ratio between the groups (all p values are >0.05) (Table 1).

Table 1.

Clinical and hormonal profile of hypogonadal men according to the presence of type 2 diabetes. Significant p values are in bold print.

	Type 2 diabetes (N = 49)	No diabetes (N = 56)	P value
Age (yr)	63.0 ± 5.9	56.8 ± 9.5	<0.001
BMI (kg/m2)	33.2 ± 5.3	31.3 ± 5.2	0.07
A1c (%-)	7.9 ± 2.0	5.7 ± 0.4	<0.001
Hormones
Testosterone (ng/dl)	203.7 ± 61.5	215.5 ± 66.5	0.39
Estradiol (pg/ml)	17.2 ± 5.9	16.7 ± 6.7	0.68
E2/T	0.6 ± 0.8	0.4 ± 0.4	0.06
FAI	24.1 ± 20.7	21.0 ± 19.5	0.45
FEI (pmol/nmol)	2.3 ± 2.5	1.7 ± 1.9	0.22

Values shown as Mean ± SD. Hormones, A1c, fasting glucose P values adjusted for age. BMI: Body Mass Index; A1c: hemoglobin A1c; E2/T: Estradiol to Testosterone ratio; FAI: Free Androgen Index; FEI: Free Estrogen Index.

Unadjusted and adjusted aBMD of the lumbar spine and femoral neck were not significantly different between the groups. While total hip aBMD was significantly higher in the group with T2D, this disappeared after adjusting for the different variables in model 2 (BMI, age and estradiol levels) and model 3 (model 2 plus use of diabetes medication) (Table 2). Unadjusted and adjusted total and trabecular vBMD, as well as trabecular parameters of bone geometry measured at 4% tibia were similar between the groups. Unadjusted total vBMD at 38% tibia was not significantly different between those with and without T2D. However, after adjusting for the variables in models 2 and 3, total vBMD was higher among those with T2D compared to those without diabetes. Both unadjusted and adjusted endosteal and periosteal circumferences at the 38% tibia were significantly lower in the group with T2D (Table 2) while unadjusted total area at the 38% tibia was significantly lower in hypogonadal men with T2D which persisted with adjustment for BMI, age and estradiol. However, additional adjustment for the use of diabetes medication in the total area at the 38% tibia reduced the significance of the difference between the 2 groups to borderline.

Table 2.

Skeletal Profile of hypogonadal men according to the presence of type 2 diabetes. Significant p values are in bold print.

	Type 2 diabetes (N = 49)	No diabetes (N = 56)	P value model 1	P value model 2	P value model 3
Lumbar spine BMD (g/cm2)	1.132 ± 0.25	1.073 ± 0.16	0.17	0.47	0.32
Femoral neck BMD (g/cm2)	0.816 ± 0.12	0.821 ± 0.13	0.82	0.56	0.89
Total hip BMD (g/cm2)	1.114 ± 0.14	1.051 ± 0.13	0.02	0.37	0.44
4% Tibia
Trabecular density (mg/cm3)	235.2 ± 31.6	233.4 ± 42.6	0.84	0.78	0.48
Trabecular content (mg/mm)	144.5 ± 26.0	140.5 ± 27.9	0.51	0.88	0.75
Trabecular area (mm2)	1072.8 ± 195.5	1041.8 ± 206.7	0.49	0.63	0.61
Total density (mg/cm3)	299.0 ± 35.2	304.1 ± 48.6	0.59	0.49	0.83
Total content (mg/mm)	406.6 ± 57.1	404.7 ± 58.7	0.88	0.48	0.98
Total area (mm2)	1364.9 ± 149.4	1343.7 ± 154.5	0.53	0.41	0.61
38% Tibia
Cortical density (mg/cm3)	1154.4 ± 62.7	1142.7 ± 76.9	0.45	0.31	0.86
Cortical content (mg/mm)	386.6 ± 51.9	397.7 ± 59.0	0.37	0.71	0.71
Cortical area (mm2)	334.8 ± 39.2	348.1 ± 47.4	0.17	0.33	0.64
Total density (mg/cm3)	857.3 ± 69.0	828.7 ± 96.7	0.13	0.04	0.05
Total content (mg/mm)	414.7 ± 55.0	435.0 ± 61.1	0.23	0.54	0.54
Total area (mm2)	491.0 ± 61.0	527.7 ± 87.2	0.02	0.04	0.06
Cortical thickness (mm)	5.5 ± 0.5	5.4 ± 0.7	0.27	0.46	0.22
Endosteal circumference (mm)	43.9 ± 5.8	47.1 ± 7.8	0.04	0.03	0.02
Periosteal circumference (mm)	78.4 ± 5.0	81.3 ± 6.5	0.02	0.05	0.05

Values shown as Mean ± SD. BMD: Bone Mineral Density. Model 1: unadjusted data; model 2: p value adjusted for age, BMI and estradiol levels; model 3: p value adjusted for model 2 plus treatment of type 2 diabetes.

Biochemical markers of bone turnover (Fig. 1) showed that serum CTX (A) and osteocalcin (B) were significantly lower in those with T2D compared to those without diabetes CTX (0.25 ± 0.14 ng/ml vs.0.40 ± 0.19 ng/ml, p < 0.001) and osteocalcin (4.8 ± 2.8 ng/ml vs. 6.8 ± 3.5 ng/ml, p = 0.006) respectively, however the significance was lost for both variables after adjustment for medication use (both p > 0.05). No differences were found for serum sclerostin (0.84 ± 0.22 ng/ml vs. 0.81 ± 0.30 ng/ml, p = 0.54) and 25OHD (27.2 ± 8.8 ng/ml vs. 27.4 ± 10.7 ng/ml, p = 0.92) levels, for those with and without T2D, respectively. Further analysis showed that the group of patients with diabetes on insulin (n = 23), had significantly lower osteocalcin (3.65 ± 2.00 ng/ml vs 5.40 ± 2.79 ng/ml, p = 0.04) and CTX (0.19 ± 0.19 ng/ml vs 0.28 ± 0.14 ng/ml, p = 0.02) compared to untreated patients and those on medication other than insulin combined (n = 26). Analysis of glucose control showed that among patients with T2D, the group on insulin had a higher A1c compared to the untreated patients and those on medications other than insulin combined (8.8 ± 2.0% vs 7.3 ± 1.6%, p = 0.008). Simple regression analysis showed significant negative correlation between osteocalcin and A1c (r = −0.25, p = 0.01) and CTX and A1c (r = −0.41, p = 0.0001) in the whole population. However, separate analyses according to the presence of T2D showed that the negative correlation between bone markers and A1c was observed only between CTX and A1c and only among those with T2D (r = −0.30, p = 0.05). There were no significant correlations between parameters of bone geometry and vBMD with any of the markers of bone turnover (osteocalcin and CTX), sclerostin, vitamin D and A1c. Testosterone levels were not related to bone turnover markers or bone geometry and bone density by pQCT. However, simple regression analysis revealed a significant negative correlation between total hip aBMD by DXA and total testosterone (r = −0.25, p = 0.01). There were no correlations between testosterone and aBMD by DXA in the other skeletal sites.

Fig. 1.

Values are means ± SE. CTX (C-telopeptide) and osteocalcin levels in hypogonadal men with type 2 diabetes and hypogonadal men without diabetes. Hypogonadal men with type 2 diabetes have significantly lower CTX and osteocalcin compared to those without diabetes. Analyses were adjusted for age.

Among the eugonadal subjects (n = 50), 20 were non-diabetic, while 30 had T2D. In our eugonadal population, mean age and testosterone were 62.7 ± 9.0 years old and 458.6 ± 147.1 ng/dl, respectively. Comparing the four groups (eugonadal diabetic, eugonadal non-diabetic, hypogonadal diabetic and hypogonadal non-diabetic) showed that CTX levels were significantly lower in the diabetic groups (both hypogonadal and eugonadal) vs. the non-diabetic groups (hypogonadal and eugonadal) (p = 0.0001). Osteocalcin levels were significantly lower in the group of hypogonadal diabetic patients compared to all the other groups (p = 0.0006) (Table 3).

Table 3.

Bone turnover markers of eugonadal and hypogonadal men according the presence of type 2 diabetes mellitus.

	Diabetic Eugonadal N = 20	Diabetic Hypogonadal N = 49	Non-Diabetic Eugonadal N = 30	Non-Diabetic Hypogonadal N = 56	P value
CTX (ng/ml)	0.29 ± 0.1*	0.25 ± 0.1*	0.41 ± 0.1	0.40 ± 0.1	0.0001
Osteocalcin (ng/ml)	8.05 ± 5.0	4.80 ± 2.8δ	8.59 ± 4.0	6.81 ± 3.5	0.0006

Values shown as Mean ± SD.

^*Post-hoc analysis showed lower CTX in diabetic compared to non-diabetic groups, p value < 0.05;

^δPost-hoc analysis showed lower osteocalcin in hypogonadal diabetic compared to all groups, p value < 0.05.

4. Discussion

Our results suggest that hypogonadal men with T2D have relatively higher BMD, but smaller bone area and reduced periosteal and endosteal circumferences at the cortical site (38% tibia) compared to hypogonadal men without diabetes. Furthermore, among our hypogonadal men, subjects with T2D have relatively lower bone turnover compared to those without diabetes. No significant differences in the sex hormone levels were detected between the two groups. Thus, our results suggest that in men with hypogonadism, the subset with T2D have a higher BMD, but poorer bone geometry and relatively suppressed bone turnover compared to those who do not have T2D.

An existing mutual influence between glucose and androgen metabolism has been hypothesized [1], i.e. low testosterone is associated with glucose intolerance while hyperglycemia is in turn associated with low testosterone production [3,5]. According to a meta-analysis involving 37 studies, patients with T2D have lower testosterone levels by approximately 87 ng/dl than non-diabetic individuals [3]. In our population of hypogonadal men, we could not detect any difference in testosterone levels between those with and without T2D, most likely because of the relatively narrow range of testosterone values in our study (inclusion criteria testosterone <300 ng/dl) and the small sample size.

Despite having a normal to higher BMD, patients with T2D have an increased fracture risk compared to non-diabetics, especially when treated with certain antidiabetic medications [12,33,34]. Several mechanisms have been proposed for this observation and include the duration of T2D, poor glycemic control, accumulation of advanced glycation end products (AGEs), the onset of microvascular complications, reduced bone turnover, poor bone microarchitecture or altered bone geometry [15,35–38]. The impact of hypogonadism on the bone of patients with T2D remains unclear. In a group of men with T2D, Dhindsa et al. reported no differences in the lumbar spine, femoral neck and total hip aBMD between those with and without hypogonadism [25]. In our study in a population of hypogonadal men, there was no difference in adjusted aBMD between those with and without diabetes. However aBMD is a 2-dimentional measurement which does not take into account the actual bone volume as opposed to the 3-dimentional information provided by vBMD. Given the lack of correlation between aBMD by DXA and fracture risk in patients with T2D, results from their study and that of ours (when limited to DXA measurements), do not necessarily reflect the bone health of patients with T2D and hypogonadism. Yet, we also have data showing that hypogonadal patients with T2D have smaller bone size and relatively lower bone turnover than those without diabetes, information that were not available in prior studies. In addition, regardless of gonadal status, CTX levels were lower in patients with T2D relative to their non-diabetic counterparts while osteocalcin levels are much lower in the group who has both hypogonadism and T2D. To some extent, these findings suggest that the suppression of bone turnover markers associated with T2D prevails over effect of hypogonadism on markers of bone turnover.

The low bone turnover in T2D is associated with diminished quantity of osteoid and lower mineralization surface of bone [12,39]. Histomorphometric evaluation of transiliac crest bone biopsies of 5 patients with T2D and 4 non-diabetic subjects revealed no differences in trabecular parameters, but subjects with T2D had lower cortical width and osteoid surface and significantly lower bone turnover markers [39]. Although the effects of each of T2D and hypogonadism alone on bone health have been documented, the impact of T2D on the skeleton of hypogonadal men and vice-versa remains unestablished. Our data collected in men with low testosterone showed that while vBMD at the 38% tibia was higher among diabetics, these subjects have smaller bone size and lower CTX and osteocalcin relative to non-diabetic hypogonadal men. However, this difference in the bone turnover markers between the groups disappeared with adjustment for medication use. The role of medication in our findings is less clear as those who are on more potent agents, such as insulin, have lower bone turnover than those who are not. Medications, in particular insulin, likely have no role on the suppressed bone turnover, but these patients are likely on insulin because of uncontrolled diabetes, a conclusion supported by the higher mean A1c in this subgroup. Our analysis showed an inverse correlation between bone markers and A1c. Unfortunately, studies on the effect of glucose control on bone metabolism are lacking.

A big bone size indicates a greater bone capacity to adapt to mechanical load, elevated bone strength and resistance to compression forces which are considered the main trigger for fractures [40]. According to Haapasalo et al., the high bone strength in the playing arm of tennis players is due to an increased bone size rather than vBMD [40]. Subjects with higher cross-sectional area (CSA) are thus considered to have a better bone geometry independent of vBMD [40,41]. Although a causal relationship cannot be established, it is possible that the comparatively smaller bone size in men with hypogonadism and T2D is because of the relatively less active bone turnover in these subjects. The slower bone remodeling and the consequent accumulation of old bone of poor quality in conjunction with the deposition of AGEs perhaps foster bone brittleness in T2D [15]. Because of the testosterone lowering effect of hyperglycemia [3], patients with T2D could have been exposed to lower testosterone levels at a younger age [27] and likely for a longer period of time than those without diabetes. As a result, periosteal expansion, promoted by testosterone especially during male pubertal development [42], could have been suppressed earlier in diabetic subjects. Although there was a weak but significant negative correlation between the aBMD of the total hip and testosterone levels, this finding is consistent with previous finding by the group of Khosla et al. in a population of 97 middle age men [43]. Considering the well-accepted notion that estradiol, as opposed to testosterone, is the main regulator of the BMD in men, this finding may have little significance.

Our study has several limitations. First, being cross-sectional, it does not provide information on the changes in bone turnover and bone geometry that follow the onset of T2D and hypogonadism combined. Second, the differences in bone geometry and vBMD were mainly observed at the 38% tibia and do not necessarily apply to all skeletal sites, but differences in bone turnover suggest that the effect is likely systemic. Third, we do not have the BMD and bone geometry data of the eugonadal control groups (with and without diabetes) that would have provided a better picture of the effect of each condition alone on other parameters of bone health compared to a combination of T2D and hypogonadism. Fourth, patients with T2D in our study were significantly older compared to those without diabetes. However, our analyses were all adjusted for age, which showed that the differences between the two groups remained. Fifth, we did not consider other co-morbidities associated with T2D, the duration of both T2D and hypogonadism, all of which could have an impact on bone metabolism. Finally, we are limited by the small sample size of our study.

5. Conclusion

In conclusion, despite the higher vBMD, hypogonadal men with T2D have worse skeletal profile due to the smaller bone size and the relatively lower bone turnover which over time may result in overabundance of aged bones with reduced ability to resist fractures than those who have hypogonadism alone. Although antiresorptives, such as bisphosphonates, improve bone density and prevent fractures by suppressing bone turnover, prolonged use with persistent suppression of bone turnover for several years results in atypical femur fractures [19,44]. Similarly, it is conceivable that prolonged suppression of bone turnover in patients with T2D and hypogonadism will result in skeletal fragility. Hence, longitudinal studies with bigger sample size are necessary to determine this possibility. As morbidity and mortality from a hip fracture is worse in men than in women [26], if the co-existence of both conditions confers additional risk for fractures, perhaps efforts should be directed to control the blood glucose and replete testosterone to improve skeletal health. To our knowledge, our study is the first one to report the effect of T2D among hypogonadal men on bone turnover and bone geometry.