J Endocrinol Invest DOI 10.1007/s40618-013-0037-6
ORIGINAL ARTICLE
Relationship of body composition with bone mineral density in northern Chinese men by body mass index levels D. Kang • Z. Liu • Y. Wang • H. Zhang X. Feng • W. Cao • P. Wang
•
Received: 17 December 2012 / Accepted: 1 December 2013 Ó Italian Society of Endocrinology (SIE) 2014
Abstract Summary Osteoporosis and obesity are severe public health problems in an aging society, and as we all know, bone mineral density (BMD) is closely related to fat mass (FM) and fat distribution. However, studies have long focused on pre- or post-menopausal women, and its presence in men has been underestimated. To investigate the differential impact of fat on BMD, we characterized body composition of northern Chinese men and examined the relationship with BMD according to body mass index (BMI) levels. Methods A cross-sectional study was conducted on 502 healthy northern Chinese men aged 20–89 screened from the participants in a community-based osteoporosis prevention study conducted by the Research Center of Qianfoshan Hospital of Shandong University from 2009 to 2010. The qualified subjects were stratified according to BMI levels as normal weight (18.5 B BMI \ 24 kg/m2, n = 137), overweight (24 B BMI \ 28 kg/m2, n = 225), and obesity (BMI C 28 kg/m2, n = 140). Total body, left femur, lumbar spine BMD and lean mass (LM), FM, percent body fat (%BF) were measured by dual-energy X-ray absorptiometry. Pearson correlation and age-adjusted partial correlation analyses between body composition-related parameters and BMD were performed. Multiple regression analysis was performed to examine the relationship of BMD with LM, FM and %BF. Results Height and weight had positive associations with BMD at all sites, although age had negative associations.
D. Kang (&) Z. Liu Y. Wang H. Zhang X. Feng W. Cao P. Wang Department of Endocrinology, Qianfoshan Hospital of Shandong University, No. 16766 Jingshi Road, Jinan 250014, China e-mail: [email protected]
Of all subjects, LM and FM were positively correlated with BMD at almost sites (P \ 0.01). However, when the subjects were divided into normal weight, overweight and obesity, no relations were reflected between FM and BMD. %BF showed negative correlations with BMD at arm and leg (P \ 0.01) in overweight, and with BMD at total body, arm, leg, hip (P \ 0.01) in obesity. In regression models, both FM and LM showed statistically positively significant relations with total body and regional BMD in all subjects (all P \ 0.05). LM was positively correlated with BMD at almost site (all P \ 0.05) in groups, while FM had no association. Interestingly, percent body fat (%BF) had negative associations with BMD at total body, arm, leg and total femur in overweight and obesity. Conclusions The relationship between LM and BMD was certain in northern Chinese men while fat–bone relationship was complicated. %BF had a significantly negative association with total body and regional BMD in overweight and obese men. Keywords Lean mass Fat mass Percent body fat Bone mineral density Men Body mass index
Introduction Osteoporosis is a systemic disease of the skeleton, characterized by a low bone mass and a deterioration of the microarchitecture of the bones that lead to an increased risk of fracture. Osteoporosis and related fractures have become a world wide public health problem [1, 2]. There are many genetic and environmental factors influencing bone mineral density (BMD) [3–6]. It has been shown that body weight is a strong predictor of bone mass for different parts of the skeleton [7, 8]. Body
123
J Endocrinol Invest
weight is composed of three main components: fat mass (FM), lean mass (LM) and bone. Traditionally, body composition can be classed into FM and LM. There have been related researches between bone and body composition from the 1990s of twentieth century to this day [9, 10]. Previous studies showed that both FM and LM might help to determine bone mass. LM was found to have more impact on BMD and may against fracture [11–15]. The fat– bone connection is important clinically since thinness is a potentially preventable risk factor for fracture [16]. However, many recent studies showed contradict viewpoint by an inverse relationship between FM and BMD [17–20]. According to this, it may be caused by some effect factors such as age, sex, BMI, etc. It is shown that most of studies for fat–bone relationship are about pre- and postmenopausal women [21–23]. Recent studies have suggested that obese men are associated with increased fracture risk at some sites [24–26]. But there are differences between obese men and women [27, 28]. Further studies should elucidate the role of body composition in the occurrence of osteoporosis in men. Moreover, it is a critical area of our understanding of the regulation of bone mass and a potentially important pharmaceutical target for the development of osteoporosis therapies. To discover what this reported link between FM and BMD could be, we took this investigation in northern Chinese men.
Subjects and methods Subjects A total of 515 healthy men aged from 20 to 89 years old were screened from the participants in a community-based osteoporosis prevention study conducted by the Research Center of Qianfoshan Hospital of Shandong University from 2009 to 2010. The exclusion criteria were as follows: history of metabolic bone diseases such as chronic liver and renal failure, hyperparathyroidism and rheumatoid arthritis; history of diseases affecting body weight or composition such as thyrotoxicosis, hypothyroidism. None of the subjects had been treated with medicines capable of influencing BMD, weight and body composition such as thyroid hormones, glucocorticosteroids, bisphosphonates and antiobesity drugs. Besides, five men were excluded due to prostatectomy, and seven men excluded due to their bone mass density at any of the regions (CMean ± 3 SD). In the end, 502 men were included in the analysis. All subjects completed a questionnaire on demographic and lifestyle information about smoking, alcohol consumption, exercise, and medical history including past illness and current medication. In the questionnaire, there were two types of drinking: nondrinkers and drinkers. Drinkers were those
123
who drank an alcoholic beverage about twice a week and nondrinkers were those who drank no beer, wine, or hard liquor. Nobody was heavy drinkers. There were two types of smoking: nonsmokers and smokers. Written informed consent was obtained and the study was approved by the ethics committee of Qianfoshan Hospital, Shandong University. Anthropometry and body composition measurement Weight was measured to the nearest 0.1 kg using a calibrated standard balance beam scale. Height was measured to the nearest 0.1 cm using a wall-mounted stadiometer. When measured, all subjects should wear light clothing and no shoes. All values were recorded as the mean of three measures. Body mass index (BMI) was calculated as weight (kg)/height (m2). All subjects were divided into three groups according to BMI levels as normal weight (18.5B BMI \24 kg/m2, n = 137), overweight (24B BMI \28 kg/m2, n = 225), and obesity (BMI C 28 kg/m2, n = 140) [29–31]. Dual-energy X-ray absorptiometry (DXA); software version 11.40.004; GE-lunar, WI, USA) was used to measure LM, FM, %BF, total body BMD (TB) and regional BMD through whole-body scans. Regional BMD consisted of arm leg, femoral neck (FN), femoral shaft (FS), total femur (TF), and lumbar Spine BMD (L1–L4). For ethical reasons, we did not make any further assessments of the precision error of this equipment. According to the manufacturer’s brochure, this is a standardized commercial machine with an in vivo precision (%coefficient of variation) of \1 % for anterior–posterior spinal, femoral, total body BMD and body composition. DXA was calibrated using a standard phantom provided by the manufacturer, performed daily and demonstrated long-term ([2 years) CVs of B0.8 %. Statistical analysis All data are presented as means and SD of the mean for continuous variables and as frequencies for categorical variables. Generalized Kolmogorov–Smirnov tests were used to ascertain normality. If necessary, logarithmic transformation was performed to achieve a normal distribution. ANOVA for continuous variables and Chi square tests for categorical variables were used to compare baseline characteristics, body composition, and body mineral density in normal weight, overweight, and obesity. Pearson correlation and age-adjusted partial correlation coefficients were detected to assess linear relationships among body composition-related parameters and BMD at total body, arm, leg, femur and lumbar spine. Because of a significant interaction between the body composition and BMD, the
J Endocrinol Invest Table 1 Characteristics of the subjects by body mass index levels All subjects (n = 502)
Normal weight (n = 137)
Overweight (n = 225)
Obesity (n = 140)
Age (years)
62.2 ± 16.0
64.7 ± 17.1
61.4 ± 16.2
61.2 ± 14.5
Height (m)
1.685 ± 0.065
1.681 ± 0.065
1.692 ± 0.063
1.677 ± 0.068*
Weight (kg)
73.8 ± 11.1
62.0 ± 6.6
74.0 ± 6.3**
84.9 ± 8.6**,
mm
BMI (kg/m2)
25.9 ± 3.4
21.9 ± 1.6
25.9 ± 1.2**
30.1 ± 1.7**,
mm
Body composition measures FM (kg)
20.6 ± 6.6
13.7 ± 4.1
20.7 ± 4.2**
27.2 ± 4.8**,
mm
LM (kg)
48.5 ± 6.5
44.2 ± 5.7
48.7 ± 5.3**
52.5 ± 6.5**,
mm
%BF
29.3 ± 6.6
23.6 ± 6.1
29.8 ± 5.2**
34.1 ± 4.8**,
mm
Body mineral density measures (g/cm2) TB
1.158 ± 0.100
1.090 ± 0.091
1.173 ± 0.092**
1.198 ± 0.099**,
Arm Leg
0.869 ± 0.094 1.224 ± 0.124
0.823 ± 0.088 1.156 ± 0.121
0.880 ± 0.088** 1.242 ± 0.116**
0.898 ± 0.094** 1.261 ± 0.114**
FN
0.911 ± 0.138
0.838 ± 0.142
0.934 ± 0.131**
0.946 ± 0.118**
FS
1.184 ± 0.185
1.072 ± 0.189
1.201 ± 0.170**
1.243 ± 0.157**
TF
0.981 ± 0.145
0.892 ± 0.150
1.006 ± 0.131**
1.029 ± 0.121**
L1–L4
1.093 ± 0.164
1.031 ± 0.154
1.115 ± 0.168**
1.119 ± 0.151**
Drinkers (%)
91 (18.1)
24 (48.9)
42 (18.6)
66 (47.6)
Smokers (%)
225 (44.8)
61 (44.5)
101 (44.9)
63 (45.0)
m
Mean ± standard deviation or number (%). One-way ANOVA for continuous variables and Chi-square tests for categorical variables BMI body mass index, FM fat mass, LM lean mass, %BF percent body fat, TB total body, FN femoral neck, FS femoral shaft, TF total femur * P \ 0.05, ** P \ 0.01 compared to normal weight,
m
P \ 0.05,
mm
regression models were used to evaluate separately the associations of BMD with FM, LM and with %BF. Firstly, the associations of FM and LM with BMD were explored, and covariates such as age, height, smoking and regular alcohol consumption were included in the regression model one. Secondly, the associations of %BF with BMD were explored and covariates such as age, height, weight, smoking and regular alcohol consumption were included in the regression model two. SPSS (version 16.0 for Windows, SPSS Inc., Chicago, IL, USA) was used for analysis. All tests were two-sided, and P \ 0.05 was considered statistically significant.
Results Descriptive statistics The basic characteristics of the subjects are shown in Table 1. Comparisons of FM, LM, %BF among three groups were significantly different (all P \ 0.01). Compared to normal weight, total body and regional BMD of overweight and obesity were significantly higher (all P \ 0.01). When comparing overweight with obesity, it showed no significant difference except BMD at total body. The number of drinker and smoker appeared no significantly different in three groups.
P \ 0.01 compared to overweight
Correlation analysis Correlation analyses between anthropometry or body composition-related parameters and BMD are provided in Table 2. In Pearson’ and age-adjusted correlation analyses, LM was positively correlated with BMD at almost sites (r = 0.258–0.550 P \ 0.01; qr = 0.185–0.493 P \ 0.01). However, the relationship was not positive at lumbar spine in normal weight and overweight. Of all subjects, FM was positively correlated with total body and regional BMD (all r = 0.219–0.347 P \ 0.01; qr = 0.228–0.299 P \ 0.01). However, when the subjects were divided into three groups according to BMI, no relations were reflected between FM and BMD at almost sites. Relationship between %BF and BMD was different in groups. On the one hand, %BF was positively correlated with BMD at total body, TF and lumbar spine (r or qr = 0.126–0.186, P \ 0.01) in all subjects. On the other hand, %BF showed negative correlations with BMD at arm and leg (r or qr = -0.371 to -0.157, P \ 0.05) in overweight and obesity. Multiple regression analysis Table 3 shows determination coefficients for FM, LM, age and height to BMD in model one and determination coefficients for %BF, weight, height, age to BMD in model
123
123
0.494**
0.155**
LM
%BF
-0.203**
%BF
-0.105
0.148 0.411**
-0.083
0.325**
0.073
-0.022
0.185**
0.053
0.199**
0.443**
0.352**
-0.371**
-0.033 0.550**
-0.227**
0.436**
-0.034
-0.125
0.377**
0.031
0.008
0.527**
0.219**
-0.227**
-0.013 0.395**
-0.175**
0.359**
-0.019
-0.035
0.192*
0.052
0.008
0.424**
0.230**
Partial
-0.316**
0.031 0.550**
-0.157*
0.352**
0.013
-0.154
0.435**
0.014
0.047
0.511**
0.251**
Simple
Leg
-0.191*
0.073 0.436**
-0.133*
0.330**
0.020
-0.082
0.298**
0.029
0.096
0.443**
0.256**
Partial
-0.305**
-0.004 0.493**
-0.122
0.322**
0.023
-0.039
0.427**
0.133
0.078
0.472**
0.249**
Simple
FN
-0.132
0.050 0.301**
-0.087
0.277**
0.032
-0.097
0.217*
0.188*
0.150**
0.363**
0.261**
Partial
-0.196*
0.057 0.397**
-0.130
0.245**
-0.028
-0.044
0.338**
0.088
0.126**
0.424**
0.280**
Simple
FS
-0.059
0.098 0.253**
-0.121
0.245**
-0.025
-0.078
0.112
0.125
0.180**
0.341**
0.286**
Partial
-0.220**
0.046 0.423**
-0.120
0.258**
-0.004
-0.055
0.379**
0.094
0.129**
0.446**
0.292**
Simple
TF
* Statistical significance (P \ 0.05); ** Statistical significance (P \ 0.01)
BMD bone mineral density, TB total body, FN femoral neck, FS femoral shaft, TF total femur, FM fat mass, LM lean mass, %BF percent body fat
0.115 0.485**
FM LM
Obesity
-0.098
0.329**
LM
%BF
0.069
FM
Overweight
-0.099
0.399**
LM
%BF
0.036
FM
Normal weight
0.347**
FM
All subjects
Simple
Simple
Partial
Arm
TB
Table 2 Simple and age-adjusted partial correlation coefficients between body composition-related parameters and BMD
-0.087
0.087 0.286**
-0.106
0.249**
-0.004
-0.062
0.171*
0.131
0.186**
0.363**
0.299**
Partial
0.108
0.293** 0.203*
-0.001
0.107
0.058
-0.081
0.236*
0.020
0.127**
0.242**
0.228**
Simple
L1–L4
0.089
0.288** 0.284*
-0.054
0.230**
0.048
-0.025
0.117
0.030
0.125**
0.280**
0.228**
Partial
J Endocrinol Invest
J Endocrinol Invest Table 3 Determination coefficients for FM, LM and %BF to BMD All subjects 2
Normal weight
Overweight
R
R2
SE of estimate
0.125
0.087
0.493
0.243
0.077
0.156
0.085
0.494
0.244
0.077
0.458
0.210
0.078
0.597
0.356
0.077
0.493
0.243
0.077
0.606
0.367
0.076
0.104
0.368
0.135
0.109
0.565
0.320
0.095
0.269
0.105
0.401
0.161
0.108
0.572
0.327
0.095
0.649
0.421
0.109
0.345
0.119
0.103
0.576
0.332
0.098
0.648
0.419
0.109
0.382
0.146
0.123
0.576
0.332
0.098
0.163
0.582
0.339
0.156
0.252
0.063
0.166
0.444
0.197
0.142
0.162
0.580
0.336
0.156
0.295
0.087
0.163
0.439
0.192
0.142
0.245
0.126
0.587
0.345
0.124
0.268
0.072
0.127
0.459
0.211
0.109
0.255
0.126
0.582
0.339
0.125
0.313
0.098
0.126
0.456
0.208
0.109
0.316
0.100
0.156
0.327
0.107
0.148
0.333
0.111
0.095
0.387
0.150
0.142
0.321
0.103
0.156
0.322
0.104
0.148
0.362
0.131
0.158
0.379
0.144
0.142
R
R
SE of estimate
R
R
0.283
0.085
0.455
0.293
0.084
0.448
0.207
0.082
0.353
0.200
0.083
0.395
0.570
0.325
0.078
0.575
0.331
0.078
0.535
0.286
0.076
0.526
0.276
0.076
Leg M1
0.529
0.280
0.106
0.524
0.275
M2
0.537
0.289
0.105
0.519
M1 M2
0.533
0.284
0.117
0.541
0.293
0.117
M1
0.475
0.226
M2
0.483
0.233
M1
0.495
M2
0.505
M1 M2
R
R
M1
0.532
M2
0.542
M1 M2
2
Obesity SE of estimate
SE of estimate
2
TB
Arm
FN
FS
TF
L1–L4
Dependent variables as BMD; independent variables as FM, LM, age, height in model 1; independent variables as %BF, weight, age in model 2 BMD bone mineral density, TB total body, FN femoral neck, FS femoral shaft, TF total femur, FM fat mass, LM lean mass, %BF percent body fat, M1 model 1, M2 model 2
two. In model one, there were 22.2–32.5 % of BMD variability at almost sites in all subjects, 20.7–42.1 % in normal weight, 19.7–35.6 % in obesity. In model two, there were 23.3–33.1 % of BMD variability at almost sites in all subjects, 20.0–41.9 % in normal weight, 19.2–36.7 % in obesity. The determination coefficients of models in overweight were lower than other groups. Tables 4 and 5 show regression coefficients and P values for FM, LM to BMD in model one and regression coefficients and P values for %BF to BMD in model two. In model one, both FM and LM showed statistically positively significant relations with total body and regional BMD in all subjects (all P \ 0.05). But the change did not appear in groups. LM was positively correlated with BMD at all sites (all P \ 0.05) except for lumbar spine in normal weight, while FM had no association with BMD except for lumbar spine in obesity. Interestingly, %BF had negative associations with BMD at total body, arm, leg, FN, FS, and TF in overweight and at arm, leg in obesity, but not in normal weight in model two (Tables 4, 5).
Discussion As mentioned above, body weight is a significant predictor of osteoporosis. The present result of our research showed that weight is the significant determinant of BMD in northern Chinese men, which was consistent with previous reports. Robbins et al. [32] suggested that weight alone or weight and height is a better predictor of BMD. Body weight impacts both bone turnover and bone density. According to Wolff’s law [33, 34], if loading on a particular bone increases, the bone will remodel itself to become stronger to resist the sort of loading. In order to adapt to changes, the internal architecture of the trabeculae is certain to become thicker as a result. The converse is true as well: If the loading on a bone decreases, the bone become weaker due to turnover. The loading is required to maintain bone mass. The effect of body weight on BMD is probably contributed to FM and LM. However, the influence of its major components, FM and LM, on BMD remains unclear in men. So we aimed to explore the associations of body composition with BMD according to BMI in this study.
123
123
%BF
%BF
-0.0032 (-0.0063 to 0.0001)
0.424
0.201
-0.175
0.485
0.069
-0.173
0.329
0.099
-0.056
0.188
0.066
-0.142
b
0.051
ORIGINAL ARTICLE
Relationship of body composition with bone mineral density in northern Chinese men by body mass index levels D. Kang • Z. Liu • Y. Wang • H. Zhang X. Feng • W. Cao • P. Wang
•
Received: 17 December 2012 / Accepted: 1 December 2013 Ó Italian Society of Endocrinology (SIE) 2014
Abstract Summary Osteoporosis and obesity are severe public health problems in an aging society, and as we all know, bone mineral density (BMD) is closely related to fat mass (FM) and fat distribution. However, studies have long focused on pre- or post-menopausal women, and its presence in men has been underestimated. To investigate the differential impact of fat on BMD, we characterized body composition of northern Chinese men and examined the relationship with BMD according to body mass index (BMI) levels. Methods A cross-sectional study was conducted on 502 healthy northern Chinese men aged 20–89 screened from the participants in a community-based osteoporosis prevention study conducted by the Research Center of Qianfoshan Hospital of Shandong University from 2009 to 2010. The qualified subjects were stratified according to BMI levels as normal weight (18.5 B BMI \ 24 kg/m2, n = 137), overweight (24 B BMI \ 28 kg/m2, n = 225), and obesity (BMI C 28 kg/m2, n = 140). Total body, left femur, lumbar spine BMD and lean mass (LM), FM, percent body fat (%BF) were measured by dual-energy X-ray absorptiometry. Pearson correlation and age-adjusted partial correlation analyses between body composition-related parameters and BMD were performed. Multiple regression analysis was performed to examine the relationship of BMD with LM, FM and %BF. Results Height and weight had positive associations with BMD at all sites, although age had negative associations.
D. Kang (&) Z. Liu Y. Wang H. Zhang X. Feng W. Cao P. Wang Department of Endocrinology, Qianfoshan Hospital of Shandong University, No. 16766 Jingshi Road, Jinan 250014, China e-mail: [email protected]
Of all subjects, LM and FM were positively correlated with BMD at almost sites (P \ 0.01). However, when the subjects were divided into normal weight, overweight and obesity, no relations were reflected between FM and BMD. %BF showed negative correlations with BMD at arm and leg (P \ 0.01) in overweight, and with BMD at total body, arm, leg, hip (P \ 0.01) in obesity. In regression models, both FM and LM showed statistically positively significant relations with total body and regional BMD in all subjects (all P \ 0.05). LM was positively correlated with BMD at almost site (all P \ 0.05) in groups, while FM had no association. Interestingly, percent body fat (%BF) had negative associations with BMD at total body, arm, leg and total femur in overweight and obesity. Conclusions The relationship between LM and BMD was certain in northern Chinese men while fat–bone relationship was complicated. %BF had a significantly negative association with total body and regional BMD in overweight and obese men. Keywords Lean mass Fat mass Percent body fat Bone mineral density Men Body mass index
Introduction Osteoporosis is a systemic disease of the skeleton, characterized by a low bone mass and a deterioration of the microarchitecture of the bones that lead to an increased risk of fracture. Osteoporosis and related fractures have become a world wide public health problem [1, 2]. There are many genetic and environmental factors influencing bone mineral density (BMD) [3–6]. It has been shown that body weight is a strong predictor of bone mass for different parts of the skeleton [7, 8]. Body
123
J Endocrinol Invest
weight is composed of three main components: fat mass (FM), lean mass (LM) and bone. Traditionally, body composition can be classed into FM and LM. There have been related researches between bone and body composition from the 1990s of twentieth century to this day [9, 10]. Previous studies showed that both FM and LM might help to determine bone mass. LM was found to have more impact on BMD and may against fracture [11–15]. The fat– bone connection is important clinically since thinness is a potentially preventable risk factor for fracture [16]. However, many recent studies showed contradict viewpoint by an inverse relationship between FM and BMD [17–20]. According to this, it may be caused by some effect factors such as age, sex, BMI, etc. It is shown that most of studies for fat–bone relationship are about pre- and postmenopausal women [21–23]. Recent studies have suggested that obese men are associated with increased fracture risk at some sites [24–26]. But there are differences between obese men and women [27, 28]. Further studies should elucidate the role of body composition in the occurrence of osteoporosis in men. Moreover, it is a critical area of our understanding of the regulation of bone mass and a potentially important pharmaceutical target for the development of osteoporosis therapies. To discover what this reported link between FM and BMD could be, we took this investigation in northern Chinese men.
Subjects and methods Subjects A total of 515 healthy men aged from 20 to 89 years old were screened from the participants in a community-based osteoporosis prevention study conducted by the Research Center of Qianfoshan Hospital of Shandong University from 2009 to 2010. The exclusion criteria were as follows: history of metabolic bone diseases such as chronic liver and renal failure, hyperparathyroidism and rheumatoid arthritis; history of diseases affecting body weight or composition such as thyrotoxicosis, hypothyroidism. None of the subjects had been treated with medicines capable of influencing BMD, weight and body composition such as thyroid hormones, glucocorticosteroids, bisphosphonates and antiobesity drugs. Besides, five men were excluded due to prostatectomy, and seven men excluded due to their bone mass density at any of the regions (CMean ± 3 SD). In the end, 502 men were included in the analysis. All subjects completed a questionnaire on demographic and lifestyle information about smoking, alcohol consumption, exercise, and medical history including past illness and current medication. In the questionnaire, there were two types of drinking: nondrinkers and drinkers. Drinkers were those
123
who drank an alcoholic beverage about twice a week and nondrinkers were those who drank no beer, wine, or hard liquor. Nobody was heavy drinkers. There were two types of smoking: nonsmokers and smokers. Written informed consent was obtained and the study was approved by the ethics committee of Qianfoshan Hospital, Shandong University. Anthropometry and body composition measurement Weight was measured to the nearest 0.1 kg using a calibrated standard balance beam scale. Height was measured to the nearest 0.1 cm using a wall-mounted stadiometer. When measured, all subjects should wear light clothing and no shoes. All values were recorded as the mean of three measures. Body mass index (BMI) was calculated as weight (kg)/height (m2). All subjects were divided into three groups according to BMI levels as normal weight (18.5B BMI \24 kg/m2, n = 137), overweight (24B BMI \28 kg/m2, n = 225), and obesity (BMI C 28 kg/m2, n = 140) [29–31]. Dual-energy X-ray absorptiometry (DXA); software version 11.40.004; GE-lunar, WI, USA) was used to measure LM, FM, %BF, total body BMD (TB) and regional BMD through whole-body scans. Regional BMD consisted of arm leg, femoral neck (FN), femoral shaft (FS), total femur (TF), and lumbar Spine BMD (L1–L4). For ethical reasons, we did not make any further assessments of the precision error of this equipment. According to the manufacturer’s brochure, this is a standardized commercial machine with an in vivo precision (%coefficient of variation) of \1 % for anterior–posterior spinal, femoral, total body BMD and body composition. DXA was calibrated using a standard phantom provided by the manufacturer, performed daily and demonstrated long-term ([2 years) CVs of B0.8 %. Statistical analysis All data are presented as means and SD of the mean for continuous variables and as frequencies for categorical variables. Generalized Kolmogorov–Smirnov tests were used to ascertain normality. If necessary, logarithmic transformation was performed to achieve a normal distribution. ANOVA for continuous variables and Chi square tests for categorical variables were used to compare baseline characteristics, body composition, and body mineral density in normal weight, overweight, and obesity. Pearson correlation and age-adjusted partial correlation coefficients were detected to assess linear relationships among body composition-related parameters and BMD at total body, arm, leg, femur and lumbar spine. Because of a significant interaction between the body composition and BMD, the
J Endocrinol Invest Table 1 Characteristics of the subjects by body mass index levels All subjects (n = 502)
Normal weight (n = 137)
Overweight (n = 225)
Obesity (n = 140)
Age (years)
62.2 ± 16.0
64.7 ± 17.1
61.4 ± 16.2
61.2 ± 14.5
Height (m)
1.685 ± 0.065
1.681 ± 0.065
1.692 ± 0.063
1.677 ± 0.068*
Weight (kg)
73.8 ± 11.1
62.0 ± 6.6
74.0 ± 6.3**
84.9 ± 8.6**,
mm
BMI (kg/m2)
25.9 ± 3.4
21.9 ± 1.6
25.9 ± 1.2**
30.1 ± 1.7**,
mm
Body composition measures FM (kg)
20.6 ± 6.6
13.7 ± 4.1
20.7 ± 4.2**
27.2 ± 4.8**,
mm
LM (kg)
48.5 ± 6.5
44.2 ± 5.7
48.7 ± 5.3**
52.5 ± 6.5**,
mm
%BF
29.3 ± 6.6
23.6 ± 6.1
29.8 ± 5.2**
34.1 ± 4.8**,
mm
Body mineral density measures (g/cm2) TB
1.158 ± 0.100
1.090 ± 0.091
1.173 ± 0.092**
1.198 ± 0.099**,
Arm Leg
0.869 ± 0.094 1.224 ± 0.124
0.823 ± 0.088 1.156 ± 0.121
0.880 ± 0.088** 1.242 ± 0.116**
0.898 ± 0.094** 1.261 ± 0.114**
FN
0.911 ± 0.138
0.838 ± 0.142
0.934 ± 0.131**
0.946 ± 0.118**
FS
1.184 ± 0.185
1.072 ± 0.189
1.201 ± 0.170**
1.243 ± 0.157**
TF
0.981 ± 0.145
0.892 ± 0.150
1.006 ± 0.131**
1.029 ± 0.121**
L1–L4
1.093 ± 0.164
1.031 ± 0.154
1.115 ± 0.168**
1.119 ± 0.151**
Drinkers (%)
91 (18.1)
24 (48.9)
42 (18.6)
66 (47.6)
Smokers (%)
225 (44.8)
61 (44.5)
101 (44.9)
63 (45.0)
m
Mean ± standard deviation or number (%). One-way ANOVA for continuous variables and Chi-square tests for categorical variables BMI body mass index, FM fat mass, LM lean mass, %BF percent body fat, TB total body, FN femoral neck, FS femoral shaft, TF total femur * P \ 0.05, ** P \ 0.01 compared to normal weight,
m
P \ 0.05,
mm
regression models were used to evaluate separately the associations of BMD with FM, LM and with %BF. Firstly, the associations of FM and LM with BMD were explored, and covariates such as age, height, smoking and regular alcohol consumption were included in the regression model one. Secondly, the associations of %BF with BMD were explored and covariates such as age, height, weight, smoking and regular alcohol consumption were included in the regression model two. SPSS (version 16.0 for Windows, SPSS Inc., Chicago, IL, USA) was used for analysis. All tests were two-sided, and P \ 0.05 was considered statistically significant.
Results Descriptive statistics The basic characteristics of the subjects are shown in Table 1. Comparisons of FM, LM, %BF among three groups were significantly different (all P \ 0.01). Compared to normal weight, total body and regional BMD of overweight and obesity were significantly higher (all P \ 0.01). When comparing overweight with obesity, it showed no significant difference except BMD at total body. The number of drinker and smoker appeared no significantly different in three groups.
P \ 0.01 compared to overweight
Correlation analysis Correlation analyses between anthropometry or body composition-related parameters and BMD are provided in Table 2. In Pearson’ and age-adjusted correlation analyses, LM was positively correlated with BMD at almost sites (r = 0.258–0.550 P \ 0.01; qr = 0.185–0.493 P \ 0.01). However, the relationship was not positive at lumbar spine in normal weight and overweight. Of all subjects, FM was positively correlated with total body and regional BMD (all r = 0.219–0.347 P \ 0.01; qr = 0.228–0.299 P \ 0.01). However, when the subjects were divided into three groups according to BMI, no relations were reflected between FM and BMD at almost sites. Relationship between %BF and BMD was different in groups. On the one hand, %BF was positively correlated with BMD at total body, TF and lumbar spine (r or qr = 0.126–0.186, P \ 0.01) in all subjects. On the other hand, %BF showed negative correlations with BMD at arm and leg (r or qr = -0.371 to -0.157, P \ 0.05) in overweight and obesity. Multiple regression analysis Table 3 shows determination coefficients for FM, LM, age and height to BMD in model one and determination coefficients for %BF, weight, height, age to BMD in model
123
123
0.494**
0.155**
LM
%BF
-0.203**
%BF
-0.105
0.148 0.411**
-0.083
0.325**
0.073
-0.022
0.185**
0.053
0.199**
0.443**
0.352**
-0.371**
-0.033 0.550**
-0.227**
0.436**
-0.034
-0.125
0.377**
0.031
0.008
0.527**
0.219**
-0.227**
-0.013 0.395**
-0.175**
0.359**
-0.019
-0.035
0.192*
0.052
0.008
0.424**
0.230**
Partial
-0.316**
0.031 0.550**
-0.157*
0.352**
0.013
-0.154
0.435**
0.014
0.047
0.511**
0.251**
Simple
Leg
-0.191*
0.073 0.436**
-0.133*
0.330**
0.020
-0.082
0.298**
0.029
0.096
0.443**
0.256**
Partial
-0.305**
-0.004 0.493**
-0.122
0.322**
0.023
-0.039
0.427**
0.133
0.078
0.472**
0.249**
Simple
FN
-0.132
0.050 0.301**
-0.087
0.277**
0.032
-0.097
0.217*
0.188*
0.150**
0.363**
0.261**
Partial
-0.196*
0.057 0.397**
-0.130
0.245**
-0.028
-0.044
0.338**
0.088
0.126**
0.424**
0.280**
Simple
FS
-0.059
0.098 0.253**
-0.121
0.245**
-0.025
-0.078
0.112
0.125
0.180**
0.341**
0.286**
Partial
-0.220**
0.046 0.423**
-0.120
0.258**
-0.004
-0.055
0.379**
0.094
0.129**
0.446**
0.292**
Simple
TF
* Statistical significance (P \ 0.05); ** Statistical significance (P \ 0.01)
BMD bone mineral density, TB total body, FN femoral neck, FS femoral shaft, TF total femur, FM fat mass, LM lean mass, %BF percent body fat
0.115 0.485**
FM LM
Obesity
-0.098
0.329**
LM
%BF
0.069
FM
Overweight
-0.099
0.399**
LM
%BF
0.036
FM
Normal weight
0.347**
FM
All subjects
Simple
Simple
Partial
Arm
TB
Table 2 Simple and age-adjusted partial correlation coefficients between body composition-related parameters and BMD
-0.087
0.087 0.286**
-0.106
0.249**
-0.004
-0.062
0.171*
0.131
0.186**
0.363**
0.299**
Partial
0.108
0.293** 0.203*
-0.001
0.107
0.058
-0.081
0.236*
0.020
0.127**
0.242**
0.228**
Simple
L1–L4
0.089
0.288** 0.284*
-0.054
0.230**
0.048
-0.025
0.117
0.030
0.125**
0.280**
0.228**
Partial
J Endocrinol Invest
J Endocrinol Invest Table 3 Determination coefficients for FM, LM and %BF to BMD All subjects 2
Normal weight
Overweight
R
R2
SE of estimate
0.125
0.087
0.493
0.243
0.077
0.156
0.085
0.494
0.244
0.077
0.458
0.210
0.078
0.597
0.356
0.077
0.493
0.243
0.077
0.606
0.367
0.076
0.104
0.368
0.135
0.109
0.565
0.320
0.095
0.269
0.105
0.401
0.161
0.108
0.572
0.327
0.095
0.649
0.421
0.109
0.345
0.119
0.103
0.576
0.332
0.098
0.648
0.419
0.109
0.382
0.146
0.123
0.576
0.332
0.098
0.163
0.582
0.339
0.156
0.252
0.063
0.166
0.444
0.197
0.142
0.162
0.580
0.336
0.156
0.295
0.087
0.163
0.439
0.192
0.142
0.245
0.126
0.587
0.345
0.124
0.268
0.072
0.127
0.459
0.211
0.109
0.255
0.126
0.582
0.339
0.125
0.313
0.098
0.126
0.456
0.208
0.109
0.316
0.100
0.156
0.327
0.107
0.148
0.333
0.111
0.095
0.387
0.150
0.142
0.321
0.103
0.156
0.322
0.104
0.148
0.362
0.131
0.158
0.379
0.144
0.142
R
R
SE of estimate
R
R
0.283
0.085
0.455
0.293
0.084
0.448
0.207
0.082
0.353
0.200
0.083
0.395
0.570
0.325
0.078
0.575
0.331
0.078
0.535
0.286
0.076
0.526
0.276
0.076
Leg M1
0.529
0.280
0.106
0.524
0.275
M2
0.537
0.289
0.105
0.519
M1 M2
0.533
0.284
0.117
0.541
0.293
0.117
M1
0.475
0.226
M2
0.483
0.233
M1
0.495
M2
0.505
M1 M2
R
R
M1
0.532
M2
0.542
M1 M2
2
Obesity SE of estimate
SE of estimate
2
TB
Arm
FN
FS
TF
L1–L4
Dependent variables as BMD; independent variables as FM, LM, age, height in model 1; independent variables as %BF, weight, age in model 2 BMD bone mineral density, TB total body, FN femoral neck, FS femoral shaft, TF total femur, FM fat mass, LM lean mass, %BF percent body fat, M1 model 1, M2 model 2
two. In model one, there were 22.2–32.5 % of BMD variability at almost sites in all subjects, 20.7–42.1 % in normal weight, 19.7–35.6 % in obesity. In model two, there were 23.3–33.1 % of BMD variability at almost sites in all subjects, 20.0–41.9 % in normal weight, 19.2–36.7 % in obesity. The determination coefficients of models in overweight were lower than other groups. Tables 4 and 5 show regression coefficients and P values for FM, LM to BMD in model one and regression coefficients and P values for %BF to BMD in model two. In model one, both FM and LM showed statistically positively significant relations with total body and regional BMD in all subjects (all P \ 0.05). But the change did not appear in groups. LM was positively correlated with BMD at all sites (all P \ 0.05) except for lumbar spine in normal weight, while FM had no association with BMD except for lumbar spine in obesity. Interestingly, %BF had negative associations with BMD at total body, arm, leg, FN, FS, and TF in overweight and at arm, leg in obesity, but not in normal weight in model two (Tables 4, 5).
Discussion As mentioned above, body weight is a significant predictor of osteoporosis. The present result of our research showed that weight is the significant determinant of BMD in northern Chinese men, which was consistent with previous reports. Robbins et al. [32] suggested that weight alone or weight and height is a better predictor of BMD. Body weight impacts both bone turnover and bone density. According to Wolff’s law [33, 34], if loading on a particular bone increases, the bone will remodel itself to become stronger to resist the sort of loading. In order to adapt to changes, the internal architecture of the trabeculae is certain to become thicker as a result. The converse is true as well: If the loading on a bone decreases, the bone become weaker due to turnover. The loading is required to maintain bone mass. The effect of body weight on BMD is probably contributed to FM and LM. However, the influence of its major components, FM and LM, on BMD remains unclear in men. So we aimed to explore the associations of body composition with BMD according to BMI in this study.
123
123
%BF
%BF
-0.0032 (-0.0063 to 0.0001)
0.424
0.201
-0.175
0.485
0.069
-0.173
0.329
0.099
-0.056
0.188
0.066
-0.142
b
0.051
