Sage Journals: Discover world-class research

Abstract

Objective:

Type 2 diabetes mellitus complicated with microvascular diseases can be used as a model to study the relationship between bone health and the microvascular situation.

Methods:

A total of 2,170 patients with type 2 diabetes mellitus (1,188 postmenopausal females and 982 males aged ⩾50 years) were included in our cross-sectional study. These patients were grouped according to 24-hour urine protein level: Group I (<30 mg), Group II (30-299 mg) and Group III (≥300 mg). Bone mineral density of the lumbar spine, hip and femoral neck was evaluated by dual-energy X-ray absorptiometry. Fundus oculi photography for diabetic retinopathy and 24-h urine protein for diabetic nephropathy were used as markers of microangiopathy in type 2 diabetes mellitus. Characteristics of the patients and bone mineral density were compared. Multivariate analysis was used to study the association between bone mineral density and microangiopathy. Statistical analysis was performed using SPSS 20.0. p < 0.05 was considered statistically significant.

Results:

Group III had the lowest bone mineral density level in both genders. Multivariate analysis revealed that microangiopathy was negatively correlated with bone mineral density in females (lumbar: r = –0.522, p < 0.001; hip: r = –0.301, p = 0.010; femoral neck: r = –0.314, p = 0.009), but not in males, after adjustment for age, body mass index, hypertension, hyperlipidemia, diabetic status, hepatic function, kidney function, sex hormones and 25(OH) vitamin D.

Conclusion:

These results demonstrate an independent negative correlation between microangiopathy and bone mineral density in postmenopausal female type 2 diabetes mellitus patients.

Keywords

24-h urine protein bone mineral density microangiopathy type 2 diabetes mellitus

Introduction

In the past decades, the pathophysiological study of osteoporosis has focused on bone cells including osteoblasts and osteoclasts. Some studies showed that vasculature degeneration may affect bone quality. A study¹ conducted in Hong Kong, China, that included 103 subjects revealed that vertebral marrow perfusion indexes had a significantly decreased trend in the osteoporotic group as compared with the osteopenic and the normal bone density group. This provided useful evidence that vasculature degeneration may affect bone health. In addition, many previous laboratory studies have suggested that angiogenesis promoted osteogenesis.^2,3 Choi et al.⁴ reported that the urine albumin-creatinine ratio (ACR) was negatively correlated with bone mineral density (BMD) at the lumbar spine and femoral neck in females. Another survey⁵ with follow-up lasing 8.4 years from Norway demonstrated that ACR was associated with non-vertebral fractures in females without diabetes. The prospective study found that the relative risk of fracture was 1.01 in males (p = 0.94) and 1.15 in females (p = 0.005) for a 1-standard deviation (SD) higher value for log-transformed ACR.⁵ These findings suggest that vasculature could affect BMD.

However, clinical evidence of the association between vasculature and bone disease is limited. Type 2 diabetes mellitus (T2DM) patients were vulnerable to vasculopathy and were especially complicated with microangiopathy. This provided a proper clinical model to investigate the relationship between the microvascular situation and bone health. Diabetic microvascular disease includes diabetic retinopathy and nephropathy. Fundus oculi photographs and proteinuria are traditional signs of early diabetic microvascular disease, and 24-h urine protein is an accurate, convenient, stable and reliable screening method to estimate urine protein excretion.⁶ The ocular fundus photographic changes and 24-h urine protein in T2DM patients might provide a clue of microangiopathy.^7–10

In this study, we sought to investigate the correlation between microangiopathy and BMD in postmenopausal women with T2DM and age-matched male patients using ocular fundus photographic changes and 24-h urine protein as surrogate microangiopathy markers.

Patients and methods

Patients

The study protocol was approved by the local ethics review committee at the authors’ affiliated institution, and the research was conducted according to the Declaration of Helsinki. Patient consent was not required due to the retrospective nature of the study.

In this retrospective study, we reviewed the medical records of patients with T2DM who visited the Department of Endocrinology and Metabolism, Shanghai Tenth People’s Hospital (Shanghai, P.R. China) between July 2011 and January 2016. Diabetes was defined according to the ADA standards.⁶ Male patients who were at least 50 years of age and postmenopausal women were included. Major exclusion criteria were kidney disease like nephritis and nephropathy, infections, fever or other conditions that could influence urine protein content, concurrent diseases that may affect bone metabolism like rheumatic arthritis and hyperparathyroidism, diabetic ketoacidosis and hyperglycemic hyperosmotic state, and concurrent medications that may influence bone metabolism such as bisphosphonates. Patients with incomplete data on BMD or 24-h urine protein or unknown menopausal status were also excluded.

Information was collected on age, gender, diabetic duration, menstrual history, previous history and current medications. Body mass index (BMI, kg/m²) was calculated as mass divided by height in metre squared. Women were defined as postmenopausal if they were reported having a natural menopause.¹¹ Hypertension was defined as systolic blood pressure (SBP) greater than 140 mmHg and/or diastolic pressure higher than 90 mmHg. Hyperlipidemia was considered present if total cholesterol (TC) was >5.2 mmol/L, triglyceride (TG) >1.7 mmol/L or low-density lipoprotein (LDL) >3.4 mmol/L. Hepatic function was considered impaired if alanine transaminase (ALT) was >40U/L or aspartate transaminase (AST) was >40 U/L or total bilirubin was >17.1 μmol/L or conjugated bilirubin >5.0 μmol/L. Kidney function was considered impaired if serum creatinine >106 μmol/L.

Biochemical analysis

Overnight fasting blood was collected. The fasting plasma glucose was determined by the glucose oxidase method. Haemoglobin A1 C (HbA1 C) was measured by high-performance liquid chromatography (HPLC) with the coefficient of variation (CV) of 1.11%.¹² Plasma insulin (Biosource, Nivelles, Belgium) was determined by immunoradiometric assays with the intra- and inter-assay CVs of 1.6%–2.2% and 6.1%–6.5% and 2.3%–3.0% and 3.5%–5.1%, respectively.¹³ The homeostasis model assessment–insulin resistance (HOMA-IR) index was calculated as fasting glucose (mmol/L) multiplied by fasting insulin (U/mL) divided by 2.25. TC, TG, high-density lipoprotein (HDL), and LDL were all measured by the electrochemical luminescence [enzyme-linked immunosorbent assay (ELISA)] method.¹⁴ Serum osteocalcin, C-terminal telopeptide fragments of type-I collagen (β-CTX) and 25 (OH) vitamin D were measured using radioimmunoassay (RIA).¹⁵ Serum calcium and phosphorus, ALT, AST, total bilirubin, conjugated bilirubin, serum creatinine (Cr), estradiol and testosterone were analysed with a routine autoanalyzer (Modular DP analyzer, Roche Diagnostics, Mannheim, Germany).

Microangiopathy

Ocular fundus photography

Patients with available ocular fundus photographs were evaluated for diabetic retinopathy. They were diagnosed as diabetic retinopathy microaneurysm formation, intraretinal haemorrhages, retinal capillary nonperfusion, cotton wool spots, increased numbers of haemorrhages, venous abnormalities and intraretinal microvascular abnormalities (IRMA). Photographs were tested by Digital Retinal CR-2.3 operated by experienced ophthalmologists.

2. 24-h urine protein

After a 24-h timed urine sample collection, 24-h proteinuria (mg) was measured by an automated nephelometric immunochemical method using reagents and a calibrator from Beckman Coulter (Fullerton, CA, USA). Internal reference samples were used in each analytical run.

3. Severity of microangiopathy

To examine the relationship between BMD and microangiopathy, we evaluated each patient for microangiopathy. So we collected the data of ocular fundus photography of these patients. Since 24-h urine protein was more objective than the grade of ocular fundus photographic changes, we used it to represent the degree of microangiopathy in multivariate analysis. Thus, severity of microangiopathy was graded as follows: (1) grade 0: no diabetic microangiopathy (no diabetic retinopathy); (2) grade 1: with diabetic retinopathy and 24-h urine protein < 30 mg; (3) grade 2: with diabetic retinopathy and 24-h urine protein 30–299 mg; and (4) grade 3: with diabetic retinopathy and 24-h urine protein ≥ 300 mg.

Dual-energy X-ray absorptiometry

BMD in the lumbar spine (L1 L4), hip and femoral neck was analysed using a Hologic QDR 4500 W densitometer with software version 9.03 (Hologic, USA). All BMD scans were conducted on the same device operated by a well-trained technologist under the same parameter settings following standard procedures. The dual-energy x-ray absorptiometry (DXA) device provided BMD (g/cm²), and T-score or Z-score. Osteoporosis was defined as a T-score value of ⩾2.5 SDs below the young adult mean (T-score <2.5 SD) according to the diagnostic criteria of the World Health Organization for osteoporosis in postmenopausal women and men ⩾50 years old.

Statistical analysis

Normally distributed data were presented as mean ± SD, non-normally distributed data were expressed as median (interquartile range, IQR), and categorical variables were expressed as percentage. All analyses were performed using SPSS 20.0 (SPSS Inc., Chicago, IL, USA) and stratified by gender owing to some physiological difference. Comparisons of three groups were made using analysis of variance (ANOVA) for continuous and homogeneous variables, nonparametric test (Kruskal–Wallis Test) for inhomogeneous variables and X² for categorical variables. Multivariate analysis was used to analyse the association between BMD at the lumbar spine, hip and femoral neck and other variables. In the model, the following variables were entered, including hypertension, hyperlipidemia, diabetic status, hepatic and kidney function, sex hormones, 25(OH) vitamin D, and microangiopathy. p < 0.05 was considered statistically significant.

Results

Demographic and baseline characteristics of the study population

The study flow chart is shown in Figure 1. A total of 3480 patients were enrolled. Forty-four patients were excluded due to kidney disease, 210 patients were excluded because of infection and fever or other conditions that might influence urine protein content and 513 patients were excluded due to medications or concurrent diseases that may impact bone metabolism. Fifty-two patients were excluded due to diabetic ketoacidosis and hyperglycemic hyperosmotic state. Furthermore, 491 patients were excluded due to lack of data on BMD, 24-h urine protein, or unknown menopausal status. Finally, 2,170 patients including 982 men and 1188 women were included. Patient demographic and baseline characteristics are shown in Table 1.

Figure 1.

The flow chart of study. A total of 3,480 T2DM patients were enrolled. According to the exclusion criteria, 1,310 patients were excluded. Finally, 2,170 T2DM patients (1,188 postmenopausal females and 982 age-matched males) were included.

Table 1.

Demographic and baseline characteristics of the study population.

	Males (982)				Females (1188)
	I	II	III	p	I	II	III	p
No.	171	687	124		201	809	178
Age (year)	61 (16)	63 (16)	62 (14)	0.006	66 (20)	66 (18)	72 (17)	<0.001
BMI (kg/m²)	24.28 ± 3.53	24.22 (4.78)	24.42 ± 3.04	0.420	24.85 ± 3.91	24.49 (4.76)	24.24 (4.46)	0.237
SBP (mmHg)	133.0 (30.0)	130.0 (21.0)	145.3 ± 20.8	0.002	133.0 ± 14.2	134.5 ± 19.57	140.0 (36.0)	0.018
DBP (mmHg)	80.0 (7.2)	80.0 (13.0)	80.0 (14.0)	0.067	77.5 (12.0)	75.1 ± 10.94	76.0 ± 11.5	0.867
Diabetic duration (year)	10.0 ± 7.25	9.25 (9.79)	11.0 (10.0)	0.001	10.5 ± 7.7	10.0 (10.0)	12.0 (12.0)	0.010
HbA1c (%)	8.91 ± 1.95	8.80 (3.37)	9.10 (2.70)	0.650	7.86 ± 1.68	8.60 (3.30)	9.72 ± 1.90	0.128
HOMA-IR	2.94 (4.31)	3.66 (4.47)	3.84 (5.86)	0.104	3.56 (2.58)	4.00 (5.67)	5.23 (9.21)	0.007
ALT(U/L)	16.90 (7.73)	17.30 (15.20)	18.30 (13.30)	0.943	16.05 (9.77)	15.90 (10.52)	12.40 (8.35)	0.011
Cr (μmol/L)	76.75 (19.15)	73.95 (22.65)	84.70 (40.30)	< 0.001	56.90 (15.72)	56.25 (18.30)	76.39 (46.85)	< 0.001
TC (mmol/L)	4.58 ± 1.06	4.28 (1.25)	4.30 ± 1.15	0.013	4.90 ± 1.05	4.89 (1.41)	4.81 (1.36)	0.083
TG (mmol/L)	1.79 ± 1.02	1.26 (0.93)	1.35 (0.61)	0.069	1.47 (1.23)	1.46 (0.98)	1.80 (1.53)	0.480
Insulin secretagogues, n (%)	26 (15.2)	104 (15.1)	17 (13.7)	0.915	35 (17.4)	135 (16.7)	26 (14.6)	0.739
Metformin, n (%)	22 (12.9)	87 (12.7)	17 (13.7)	0.950	24 (11.9)	118 (14.6)	20 (11.2)	0.371
α glucosidase inhibitors, n (%)	25 (14.6)	95 (13.8)	6 (3.0)	0.962	25 (12.4)	142 (17.6)	30 (16.9)	0.217
Thiazolidinediones, n (%)	4 (2.3)	28 (4.1)	5 (4.0)	0.558	6 (3.0)	55 (6.8)	7 (3.9)	0.061
Osteoporosis, n (%)	11 (6.6)	91 (13.2)	27 (22.0)	0.006	54 (26.9)	256 (31.6)	95 (53.1)	<0.001

BMI: body mass index; SBP: systolic blood pressure; DBP: diastolic blood pressure; TC: total cholesterol; TG: triglyceride.

24-h urine protein content: group I: < 30 mg, group II: 30–299 mg and group III: ⩾300 mg.

In men, there were 171 patients in group I, 687 patients in group II and 124 patients in group III. Statistically significant differences were observed in age, SBP, duration of diabetes, serum creatinine and TC among the three groups. In women, there were 201 patients in group I, 809 patients in group II and 178 patients in group III. Statistically significant differences were observed in age, SBP, duration of diabetes, serum creatinine, ALT and HOMA-IR among the three groups. We observed no statistically significant difference in the prescription of anti-diabetic medications (insulin secretagogues, metformin, α-glucosidase inhibitors and thiazolidinediones) in the three groups in both genders.

BMD and bone metabolism

The proportion of patients with osteoporosis was significantly different among the three groups in both men and women (male: p = 0.006, female: p < 0.001). Higher 24-h urine protein content was associated with a greater proportion of patients with osteoporosis. Both men and women in group III had significantly lower BMD at the lumbar spine, the hip and femoral neck than men and women in group II or I, respectively (p < 0.05 in all) (Figure 2). We found no statistically significant difference in calcium, phosphorus, testosterone, and osteocalcin among the three groups in both men and women (Table 2). Both men and women in group III had significantly lower 25(OH) VD levels than their counterparts in group II or I, respectively (p < 0.05 in all). Furthermore, women in group III had significantly higher estradiol or β-CTX than women in group II or I (p < 0.05 in all).

Figure 2.

BMD among different proteinuria groups in males and females. The circle (•) represented the male and the box (■) represented the females. The punctuation ‘*’ represented p-value < 0.05, when compared with group with 24-h proteinuria level less than 30. The punctuation ‘#’ represented p-value < 0.05, when compared with group with 24-h proteinuria level between 30 and 299: (a) The difference of BMD (g/cm²) at lumbar spine in three groups, for males: Group 1: 1.018 ± 0.173; Group 2: 1.002 ± 0.166; Group 3: 0.967 ± 0.160, p < 0.001; Group 1 vs Group 2, p < 0.001; Group 1 vs Group 3, p < 0.001; Group 2 vs Group 3, p = 0.003; for females: Group 1: 0.908 ± 0.185; Group 2: 0.863 ± 0.149; Group 3: 0.823 ± 0.178, p < 0.001; Group 1 vs Group 2, p < 0.001; Group 1 vs Group 3, p < 0.001; Group 2 vs Group 3, p < 0.001. (b) The difference of BMD at hip in three groups, for males: Group 1: 0.938 ± 0.169, Group 2: 0.908 ± 0.140, Group 3: 0.872 ± 0.148, p < 0.001; Group 1 vs Group 2, p < 0.001; Group 1 vs Group 3, p < 0.001; Group 2 vs Group 3, p = 0.001; for females: Group 1: 0.815 ± 0.155; Group 2: 0.784 ± 0.151; Group 3: 0.738 ± 0.131, p < 0.001; Group 1 vs Group 2, p < 0.001; Group 1 vs Group 3, p < 0.001; Group 2 vs Group 3; p < 0.001. (c) The difference of BMD at femoral neck in three groups, for males: Group 1: 0.823 ± 0.165; Group 2: 0.787 ± 0.144; Group 3: 0.753 ± 0.132, p < 0.001; Group 1 vs Group 2, p < 0.001; Group 1 vs Group 3, p < 0.001; Group 2 vs Group 3, p = 0.001; for females Group 1: 0.706 ± 0.150; Group 2: 0.674 ± 0.141; Group 3: 0.640 ± 0.119, p < 0.001; Group 1 vs Group 2, p < 0.001; Group 1 vs Group 3, p < 0.001; Group 2 vs Group 3, p < 0.001.

Table 2.

Biochemical profile of the study patients stratified by 24-h urinary protein content.

	Males (n = 982)				Females (n = 1,188)
	I	II	III	p	I	II	III	p
Calcium (mmol/L)	2.22 ± 0.09	2.22 (0.15)	2.19 (0.14)	0.479	2.24 (0.15)	2.27 ± 0.12	2.25 ± 0.14	0.159
Phosphorus (mmol/L)	1.22 ± 0.18	1.19 (0.24)	1.16 (0.29)	0.617	1.27 ± 0.24	1.25 (0.23)	1.21 (0.25)	0.284
Estradiol (pmol/L)	106.46 ± 39.26	103.10 (56.70)	103.30 (59.1)	0.372	33.91 (33.83)	29.27 (34.77)	43.88 (38.84)	0.040
Testosterone (nmol/L)	14.43 ± 6.13	13.80 (7.45)	14.12 ± 7.12	0.415	0.61 (0.68)	0.64 (0.54)	0.75 (0.43)	0.482
Osteocalcin (ng/mL)	11.53 (5.76)	11.77 (6.26)	11.58 (7.57)	0.958	14.49 ± 4.23	14.22 (7.21)	14.67 (9.20)	0.595
β-CTX (ng/mL)	0.33 (0.27)	0.33 (0.29)	0.32 (0.35)	0.794	0.38 ± 0.14	0.39 (0.32)	0.44 (0.23)	0.045
25(OH) vitamin D (nmol/L)	49.00 (28.50)	44.00 (24.00)	39.0 (25.0)	0.005	49.00 (25.00)	45.00 (24.00)	42.00 (17.00)	0.006

24-h urine protein content: group I: <30 mg; group II: 30–299 mg; and group III: ⩾300 mg.

Microangiopathy negatively correlates with BMD of the lumbar spine, hip and femoral neck in women

Diabetic retinopathy which is considered more specific should be assessed. So we collected the data of ocular fundus photography of these patients. However, since the patients were from July 2011 to January 2016, most of their ocular fundus photographic data could not be found. At last, only 1,244 patients were collected, and 661 of them had complete data. Four hundred and eight of them were diagnosed with diabetic retinopathy, 253 patients were without diabetic retinopathy. Severity of microangiopathy was graded as follows: (1) grade 0: no diabetic microangiopathy (no diabetic retinopathy, 151 males and 102 females); (2) grade 1: with diabetic retinopathy and 24-h urine protein <30 mg (35 males and 37 females); (3) grade 2: with diabetic retinopathy and 24-h urine protein 30–300 mg (133 males and 160 females); and (4) grade 3: with diabetic retinopathy and 24-h urine protein >300 mg (23 males and 20 females).

Other independent variables such as age, BMI, hypertension, hyperlipidemia, diabetic status, hepatic function, kidney function, sex hormones and vitamin D were also included in multivariate analysis. The result showed that among these patients (661 patients), microangiopathy was negatively correlated with BMD in females.

Our multivariate analysis revealed that age negatively correlated with BMD of the hip (r = −0.239; p = 0.018) and the femoral neck (r = –0.307; p = 0.003) in men and only with BMD of the hip in women (r = –0.313, p = 0.016) (Table 3). BMI only positively correlated with BMD of the lumbar spine in men (r = 0.266; p = 0.021) while hypertension only negatively correlated with BMD of the lumbar spine in women (r = –0.233; p = 0.027). Furthermore, total testosterone was negatively correlated with BMD of the hip (r = –0.352, p = 0.002) and the femoral neck in men (r = –0.372, p = 0.002). Moreover, we found that, after adjustment for age, BMI, hypertension, hyperlipidemia, diabetic status, sex hormones and 25(OH) vitamin D, microangiopathy were independently and negatively correlated with BMD of the lumbar spine (r = –0.522, p < 0.001), the hip (r = –0.301, p = 0.010) and the femoral neck (r = –0.314, p = 0.009) in postmenopausal women with diabetes while no such correlation was observed in men.

Table 3.

Multivariate analysis of correlation between BMD and other patient variables.

	Males						Females
	L1-L4 BMD		Hip BMD		Femoral neck BMD		L1-L4BMD		Hip BMD		Femoral neck BMD
	r	p	r	p	r	p	r	p	r	p	r	p
Age	−0.043	0.664	−0.239	0.018	−0.307	0.003	−0.217	0.052	−0.313	0.016	−0.259	0.052
Body mass index	0.266	0.021	0.158	0.163	0.069	0.547	−0.012	0.903	0.090	0.429	−0.035	0.766
Hypertension	0.154	0.113	0.114	0.237	0.129	0.187	−0.233	0.027	−0.024	0.838	−0.010	0.936
Hyperlipidemia	0.023	0.813	−0.064	0.508	−0.085	0.384	−0.006	0.955	0.036	0.757	−0.067	0.573
Diabetic status	0.152	0.110	0.071	0.452	0.009	0.924	0.121	0.216	0.187	0.100	0.073	0.529
Hepatic function	−0.048	0.593	−0.140	0.889	−0.014	0.875	−0.048	0.625	−0.094	0.408	−0.095	0.414
Kidney function	−0.080	0.409	−0.085	0.379	−0.142	0.146	−0.021	0.841	0.005	0.967	−0.045	0.708
Sex hormone	−0.225	0.050	−0.352	0.002	−0.372	0.002	−0.120	0.245	−0.114	0.342	−0.084	0.492
25(OH)VitD	0.046	0.630	0.026	0.779	0.059	0.538	0.047	0.643	0.151	0.201	0.149	0.218
Microangiopathy	−0.035	0.717	−0.124	0.195	−0.108	0.261	−0.522	<0.001	−0.301	0.010	−0.314	0.009

Hypertension: with or without hypertension (blood pressure higher than 140/90 mmHg); hyperlipidemia: with or without hyperlipidemia (TC > 5.2 mmol/L, TG > 1.7 mmol/L or LDL > 3.4 mmol/L); diabetic status: level of HbA1c; hepatic function: with or without impaired hepatic function (ALT > 40 U/L or AST > 40 U/L or TCB > 17.1μmol/L or CB > 5.0 μmol/L); kidney function: with or without impaired kidney function (serum creatinine > 106 μmol/L); microangiopathy: whether with microangiopathy or not and its severity.

Discussion

We found that BMD was the lowest at the lumbar spine, hip and femoral neck, and that the rate of osteoporosis was the highest in group Ⅲ of T2DM patients (both males and postmenopausal females) who had the highest 24-h urine protein concentration. It remains possible that BMD in the three groups may be affected by medication with glucose-lowering agents. It has been shown that metformin could increase the BMD of diabetic patients,¹⁶ while thiazolidinediones may lower the BMD of diabetic patients and increase their risk of fracture.¹⁷ However, we observed no statistically significant difference in the prescription of anti-diabetic medications in the three groups in both genders, suggesting that differences in BMD across the three groups of our study population were unlikely due to anti-diabetic medications. BMD is determined by different factors. Age has been negatively associated with BMD in previous studies.^18,19 Our study confirmed these results (Table 3). The slightly protective effect of high BMI on BMD in our finding is also similar to that found in other studies.¹¹ In addition, hypertension and diabetes are known risk factors for bones. Sex hormone is a well-known protective factor for BMD. However, our multivariate analysis revealed the total testosterone was negatively correlated with BMD in men, while estradiol was not correlated with BMD in women. Our results are in conflict with the fact that sex hormones are well-established factors for bone health. We speculate that it may be attributed to the following factors: (1) in this study, the sex hormones we tested were total testosterone and estradiol which included free forms and those bound to sex hormone-binding globulin (SHBG). In males, free testosterone, SHBG and total testosterone were different with regard to correlation with BMD in most studies.^20–22 Similar results were found in females.²³ We speculated that total testosterone and estradiol used in the current study may partially explain the differences in our findings from other studies. (2) The population of the current study consisted of diabetic patients, while most studies demonstrated a clear correlation between sex hormones and BMD were in non-diabetic patients. Abnormal fasting plasma glucose level in our study population could blunt such a correlation. Similar results were also found in other studies.²⁴ Even most of the independent variables such as age, BMI, hypertension, hyperlipidemia, diabetic status, hepatic function, kidney function, sex hormones, vitamin D and microangiopathy were included in the multivariate analysis, our results showed that among these patients (661 patients), microangiopathy was independently and negatively correlated with BMD in females. Interestingly, BMD was negatively associated with microangiopathy in postmenopausal female T2DM patients, whereas there was no such association in male T2DM patients. Furthermore, β-CTX, as the bone resorption marker, was the highest in Group III, which represented the highest bone resorption level. Thus, we assumed that microangiopathy was associated with bone loss in postmenopausal women. However, the underlying mechanism for sexual difference is unclear. It can be speculated that different sex hormones directly lead to difference in BMD itself. Bone loss in postmenopausal women was 1.7 times faster than that in men in the same age.²⁵ It may be speculated that females are more sensitive to the detrimental effect of vasculature on BMD than males.

Diabetic microangiopathy mainly manifests as diabetic nephropathy and diabetic retinopathy. Classification based on both 24-h urine protein and ocular fundus photography takes into consideration both diabetic nephropathy and diabetic retinopathy and may offer a useful scheme to better understand and stratify patients with diabetic microangiopathy. Our study demonstrated that the 24-h urine protein as a surrogate marker of microangiopathy independently and negatively correlated with BMD; however, it remains to be investigated whether diabetic microangiopathy correlated with lower bone density. We found that the bone resorption marker β-CTX increased with reducing BMD suggesting that the mechanism of bone loss was through increased bone resorption. A prospective randomized controlled study showed that calcitriol reduced proteinuria and improved BMD in patients with diabetic nephropathy.²⁵ Currently, we have no direct evidence for association of low bone mass and microangiopathy in diabetic patients. Burkhardt et al.²⁶ found that the number of arterial capillaries and sinuses per unit area decreased in osteoporotic bones. In turn, the number of capillaries next to remodelling sites in adult human bone increased. Kusumbe et al.²⁷ found that type H vessels promoted bone formation. Furthermore, a considerable body of laboratory studies demonstrated that molecular signals between bone and blood vessels are essential to bone growth and bone reconstruction. For example, pro-angiogenic factors including vascular endothelial growth factor (VEGF),^2,3 fibroblast growth factors (FGF),^28,29 platelet-derived growth factor (PDGF)³⁰ promoted blood vessel development and contributed to bone formation. Ovariectomized mice, as in mouse osteoporosis model, revealed a significant decrease in angiogenic stimulator VEGF³¹ or PDGF³⁰ expression in bones. After injecting angiogenic stimulator PDGF-BB to ovariectomized mice, the bone quality of the mice significantly improved.³⁰ These findings suggest that microvascular changes in the body may be associated or concurrent with bone loss in diabetic postmenopausal women, but clinical and laboratory evidence is required to establish whether such an association is present between microangiopathy and low bone mass in diabetic postmenopausal women.

There are several potential limitations in this study. The biomarkers in our study were all clinical parameters which were affected by many factors and histopathological studies should be undertaken to support the conclusion of the study findings. Moreover, this is only a cross-sectional study highlighting a phenomenon of higher bone loss in postmenopausal women with diabetic microangiopathy and could not establish a cause-and-effect relationship between microangiopathy and BMD. The causal relationship and the underlying mechanism remain to be elucidated. Furthermore, 24-h urine protein is an indirect marker for evaluating vasculopathy. A more direct tool for evaluating vascular function such as the functional perfusion magnetic resonance imaging (MRI) technique should be used to assess the association between vascular function and BMD in the future study.

In summary, these results demonstrate a negative correlation between BMD and microangiopathy in postmenopausal women with T2DM. However, more prospective studies are needed to confirm this.

Footnotes

Acknowledgements

N.Z., Y.Z., and X.P. contributed equally to this work.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This study was supported by the National Natural Science Foundation of China (NSFC 81500650), the Fund for Distinguished Young Scholars in Tongji University (2015KJ053), the Shanghai Municipal Natural Science Foundation (13ZR1432100) and the National High Technology Research and Development Programme 863 (2013AA032203).

References

Griffith

J F

Yeung

D K

Antonio

G E

et al . Vertebral marrow fat content and diffusion and perfusion indexes in women with varying bone density: MR evaluation.[J]. Radiology, 2006, 241(3):831–838.

Beamer

Hettrich

Lane

Vascular endothelial growth factor: an essential component of angiogenesis and fracture healing. HSS J 2010; 6: 85–94.

Clarkin

Gerstenfeld

LC.

VEGF and bone cell signalling: an essential vessel for communication?

Cell Biochem Funct 2013; 31: 1–11.

Choi

S W

Kim

H Y

Ahn

H R

et al . Association of bone mineral density with albuminuria and estimated glomerular filtration rate: the Dong-gu Study.[J]. Kidney & Blood Pressure Research, 2013, 37(2-3):132–141.

Jørgensen

Jenssen

Ahmed

et al . Albuminuria and risk of nonvertebral fractures.[J]. Archives of Internal Medicine, 2007, 167(13):1379–85.

American Diabetes

A. Standards of medical care in diabetes–2014.

Diabetes Care 2014; 37: S14–80.

Yudkin

Forrest

Jackson

AC.

Microalbuminuria as predictor of vascular disease in non-diabetic subjects. Islington diabetes survey. Lancet 1988; 2: 530–153.

Herrerapombo

J L

Aguilardiosdado

Hawkins

et al . Is Increasing Urinary Albumin a Better Marker for Microvascular than for Macrovascular Complication of Type 2 Diabetes mellitus?[J]. Nephron Clinical Practice, 2005, 101(3):116–21.

Strickland

A L

Rossetti

H C

Peshock

R M

et al . Urinary Albumin to Creatinine Ratio as Potential Biomarker for Cerebral Microvascular Disease[J]. Current Neurovascular Research, 2015, 11(3):242–247(6).

10.

Caramori

Fioretto

Mauer

Enhancing the predictive value of urinary albumin for diabetic nephropathy. J Am Soc Nephrol 2006; 17: 339–152.

11.

Shan

P F

X P

Zhang

et al . Bone mineral density and its relationship with body mass index in postmenopausal women with type 2 diabetes mellitus in mainland China[J]. Journal of Bone & Mineral Metabolism, 2009, 27(2):190.

12.

Rui

et al . Differential pattern for regulating insulin secretion, insulin resistance, and lipid metabolism by osteocalcin in male and female T2DM patients[J]. Medical Science Monitor International Medical Journal of Experimental & Clinical Research, 2014, 20(5):711–719.

13.

Hwang

Y C

Jeong

I K

Ahn

K J

et al . Circulating osteocalcin level is associated with improved glucose tolerance, insulin secretion and sensitivity independent of the plasma adiponectin level[J]. Osteoporosis International, 2012, 23(4):133.

14.

Trojak

Waluś-Miarka

Woźniakiewicz

et al . Nonalcoholic fatty liver disease is associated with low HDL cholesterol and coronary angioplasty in patients with type 2 diabetes.[J]. Medical Science Monitor International Medical Journal of Experimental & Clinical Research, 2013, 19(13):1167.

15.

Kanazawa

Yamaguchi

Sugimoto

Relationship between bone biochemical markers versus glucose/lipid metabolism and atherosclerosis; a longitudinal study in type 2 diabetes mellitus. Diabetes Res Clin Pract 2011; 92: 393–139.

16.

Awoniyi

Conner

Solomon

Effects of metformin alone or in combination with insulin on bone mineral density in osteopenic patients with type 2 diabetes. J Endocrinol Metab 2014; 4: 89–92.

17.

Billington

Grey

Bolland

MJ.

The effect of thiazolidinediones on bone mineral density and bone turnover: systematic review and meta-analysis. Diabetologia 2015; 58: 2238–2246.

18.

Scd

L M M

Lang

T F

Lambert

L C

et al . Dimensions and Volumetric BMD of the Proximal Femur and Their Relation to Age Among Older U.S. Men[J]. Journal of Bone & Mineral Research the Official Journal of the American Society for Bone & Mineral Research, 2006, 21(8):1197–206.

19.

Havill

L M

Mahaney

M C, T L B

et al . Effects of genes, sex, age, and activity on BMC, bone size, and areal and volumetric BMD[J]. Journal of Bone & Mineral Research, 2007, 22(5):737–746.

20.

Kim

Jung

Kim

et al . The relationship between testosterone and bone mineral density in middle aged men. J Korean Acad Farm Med 2003; 24: 365–374.

21.

Araujo

Travison

Leder

et al . Correlations between serum testosterone, estradiol, and sex hormone-binding globulin and bone mineral density in a diverse sample of men. J Clin Endocrinol Metab 2008; 93: 2135.

22.

Vasilkova

Mokhort

Sanec

et al . Testosterone is an independent determinant of bone mineral density in men with type 2 diabetes mellitus. Clin Chem Lab Med 2011; 49: 99–103.

23.

Napoli

Varadharajan

Rini

et al . Effects of polymorphisms of the sex hormone-binding globulin (SHBG) gene on free estradiol and bone mineral density. Bone 2009; 45: 1169.

24.

Corona

Monami

Rastrelli

et al . Type 2 diabetes mellitus and testosterone: a meta-analysis study. Int J Androl 2011; 34: 528.

25.

Jiang

Zhang

JA.

Calcitriol reduces proteinuria and improves bone mineral density in patients with diabetic nephropathy: a prospective randomized controlled study. Int J Clin Exp Med 2017; 10: 13194–13200.

26.

Burkhardt

Kettner

Bã¶Hm

et al . Changes in trabecular bone, hematopoiesis and bone marrow vessels in aplastic anemia, primary osteoporosis, and old age: a comparative histomorphometric study[J]. Bone, 1987, 8(3):157–164.

27.

Kusumbe

Ramasamy

Adams

RH.

Coupling of angiogenesis and osteogenesis by a specific vessel subtype in bone. Nature 2014; 507: 323–138.

28.

Behr

Leucht

Longaker

M T

et al . Fgf-9 is required for angiogenesis and osteogenesis in long bone repair[J]. Proc Natl Acad Sci U S A, 2010, 107(26):11853–11858.

29.

Kigami

Sato

Tsuchiya

et al . FGF-2 angiogenesis in bone regeneration within critical-sized bone defects in rat calvaria[J]. Implant Dentistry, 2013, 22(4):422–427.

30.

Xie

Cui

Wang

et al . PDGF-BB secreted by preosteoclasts induces angiogenesis during coupling with osteogenesis[J]. Nature Medicine, 2014, 20(11):1270–1278.

31.

Ding

Wei

Liu

JB.

Reduced local blood supply to the tibial metaphysis is associated with ovariectomy-induced osteoporosis in mice. Connect Tissue Res 2011; 52: 25–29.

Microangiopathy is associated with bone loss in female type 2 diabetes mellitus patients