Correlations among whole-body fat, bone, and biomarkers in boys and girls with obesity: a cross-sectional study
Article information
Abstract
Purpose
This study aimed to investigate correlations among body composition, bone parameters, and biomarkers in boys and girls with excess body fat percentage (%fat).
Methods
Healthy boys and girls aged 13–14 years with >95th percentile %fat for age and sex were included. Body composition and bone parameters of the whole body (WB) were measured using dual-energy x-ray absorptiometry. Serum biomarkers were measured via enzyme-linked immunosorbent assay. Comparisons of these parameters between sexes were analyzed using bivariate and multivariate correlation analyses.
Results
Boys and girls had no differences in %fat or body fat mass (BFM), but boys had more lean body mass (LBM) than girls. %Fat and BFM were key negative predictors of %bone in both sexes, while serum parathyroid hormone (PTH) and C-terminal cross-linking telopeptide (CTX) were predictors of %bone in girls. Both PTH and CTX were correlated with %bone in boys. Serum leptin was a predictive factor of %bone in both sexes. In addition, %bone was strongly correlated with bone mineral density (BMD) z-score and BMD z-score of participants was negatively correlated with %fat and BFM. In girls, %fat, PTH, and leptin were predictors of BMD z-score. Furthermore, BFM in girls and both BFM and LBM in boys were positively correlated with WB bone mineral content.
Conclusions
Excess %fat has a deleterious effect on WB bone in both boys and girls, potentially due to bone resorption. BFM may have a protective effect on bone through a mechanical loading mechanism.
Highlights
· An increase in %fat and body fat mass (BFM) in boys and girls can predict a decrease in %bone.
· The decreased %bone may be modulated by serum parathyroid hormone, C-terminal cross-linking telopeptide, and leptin.
· Increase in BFM in adolescents is positively correlated with the whole-body bone mineral content.
Introduction
Pediatric obesity is a major public health concern, and excessive adiposity influences growth patterns and pubertal development [1]. Body composition during puberty is a marker of metabolic changes during this intense period of growth, and body composition, including whole-body (WB) fat and lean body mass (LBM), is sex-specific [2]. Dual-energy x-ray absorptiometry (DXA) has been recommended as a tool for quantifying total and regional body fat mass (BFM) because of its low radiation, high precision, and strong clinical feasibility [3-5]. There are age- and sexbased differences in body composition and in the effect of body composition variables on bone mass. In boys, body fat percentage (%fat) is stable from age 6 to 18 years, with a slight decrease from 13 to 14 years, whereas %fat in girls increases steadily from 6 to 18 years [3]. At an age of 13–14 years, the >95th percentile of %fat for boys and girls is >30–35 and >35–40, respectively [3,5]. Boys have a lower proportion of BFM and a higher proportion of fat-free mass, which has a better protective effect on WB bone mineral density (BMD), compared with girls [6]. Previous studies have shown a positive effect of LBM in boys and BFM in girls but a deleterious effect of %fat on bone mineral content (BMC) and BMD in children and adolescents [7-11].
Biomarkers of bone metabolism reflect dynamic changes in formation and resorption processes during different periods of growth that vary with age, sex, and pubertal stage [12-14]. Bone formation markers include bone alkaline phosphatase (BALP), osteocalcin (OC), and pro-collagen 1 peptides. Bone resorption markers are products of type 1 collagen degradation and N-terminal and C-terminal cross-linking telopeptides (NTX and CTX). BALP and CTX levels are associated with sex and correlations among body fat, bone parameters, and bone biomarkers are sex-specific [15,16]. All bone markers (except OC) decrease during late puberty [13]. OC has a negative influence on the WB BMD in girls and on the lumbar spine BMD in both sexes [17]. Furthermore, 25-hydroxyvitamin D (25(OH)D) and parathyroid hormone (PTH) regulate calcium and phosphate homeostasis, which are required for bone mineralization. A previous study concluded that vitamin D status, age, and sex were important for establishing reference intervals for bone markers in healthy children [15]. However, a recent study reported no significant differences in 25(OH) D levels between boys and girls at different pubertal stages [12]. PTH is secreted from the parathyroid glands under hypocalcemic conditions, and low 25(OH)D levels augment 25(OH)D synthesis, which has a protective effect on bone health [18,19]. Moreover, PTH is associated with increased bone turnover and bone loss, exerting a catabolic effect on the bone.
Growth hormone (GH) promotes osteoblast differentiation and muscle development during puberty [20], and it also affects body composition, bone mineralization, and lipid metabolism [21]. Children with obesity are frequently tall for their age despite low GH levels, levels that can be affected by several regulatory hormones; high levels of insulin and leptin are often observed in obese children [22]. Leptin acts as a negative regulator of osteoblasts, osteocytes, and muscles and is a positive regulator of osteoclasts [23]. Previous studies have reported a strong relationship between leptin levels and BFM, and serum leptin was inversely associated with BMD [24,25]. Moreover, leptin interacts with the hypothalamic–pituitary–adrenal and GH axes [26].
Age- and sex-specific relationships between body composition and bone have been studied continuously. This study focused on the correlations among body composition, WB bone parameters, and biomarkers (including bone formation and resorption markers and related hormones) in adolescent boys and girls with a >95th percentile %fat for age (excess %fat).
Materials and methods
1. Study participants
Healthy boys and girls aged 13–14 years with a >95th percentile body mass index (BMI) for age [27] were selected from 4 secondary schools in southern Thailand. The exclusion criteria were any chronic or bone diseases; history of bone injury or bone fracture; other conditions, such as asthma, allergies, or gastritis; and use of steroids or anticonvulsant drugs [16]. Boys (n=68) and girls (n=58) were assessed for %fat by bioelectrical impedance analysis using a TANITA SC-330ST series body composition analyzer (Tanita Corporation, Japan) [16] (Fig. 1). Considering the >95th percentile %fat for age [3], 62 boys and 58 girls were enrolled in this study. The participants and their parents received information regarding the purpose and methods of the study prior to enrollment. There were 21 boys and 16 girls who declined bone and biomarker measurements (Fig. 1). Participants (41 boys and 42 girls) were assessed for WB bone density, body composition, and biomarkers, and were included in this study based on a >95th percentile %fat on DXA measurement [5].
Flowchart of the study.
2. Anthropometric measurements
Anthropometric, body composition, and bone parameter data were collected with the participants wearing light clothes without shoes or jewelry. BMI was calculated as body weight (BW) (kg) divided by height squared (m2). BW was measured to the nearest 0.1 kg, and height (cm) was measured to the nearest 0.5 cm using a stationary vertical height board with the participants in the standing position. The z-scores for BW, height, and BMI were calculated using reference data (https://thaipedendo.org/growth-and-bmi-charts/). Waist circumference (WC) (cm) was measured with the participants in the standing position using a plastic tape in the horizontal plane at the midpoint between the costal margin and iliac crest and to the nearest 0.1 cm at the end of a normal expiration [9].
3. Body composition and bone measurements
Body composition, including LBM (g/cm2), BFM (g/cm2), %fat, body lean percentage (%lean), percentage of bone or BMC (%bone), and bone parameters, including BMD (g/cm2), BMC (g), and BMD z-score of the WB (subregions of interest: trunk, extremities, and head), were measured with DXA using the pediatric mode of Stratos and a pencil-beam densitometer (Diagnostic Medical Systems, France) [4,9,28]. Quality control was maintained throughout DXA measurements, including phantom calibration, every morning before scanning. The BMD z-score is expressed as a percentile or standard deviation (SD) score that compares the participant’s BMD with the average BMD for the participant’s age, sex, and race. A z-score of zero is equivalent to the mean, and a z-score of -1 is equivalent to a value of 1 SD below the mean value [29].
4. Measurement of biomarkers and hormonal levels by enzyme-linked immunosorbent assay
Five milliliters of blood was obtained from each participant by venipuncture into a clotted blood tube. Blood samples were kept at 37°C for 1 hour, allowing retraction of the clot and subsequent centrifugation at 2,000 rpm for 10 minutes. Serum samples were harvested and stored in an aliquot state at -70°C until use for measurement. Serum concentrations of bone formation markers were determined using a specific OC enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, Inc., USA; sensitivity to 0.898 ng/mL and intra-assay coefficient of variability <10%) and BALP ELISA kit (Cloud-Clone-Corp., USA; sensitivity to 0.064 ng/mL and intra-assay coefficient of variability <10%). Serum concentrations of CTX were determined using a specific ELISA kit (Cloud-Clone-Corp.; sensitivity to 50.5 pg/mL and intra-assay coefficient of variability <10%). Serum levels of leptin, GH, PTH, and 25(OH)D were measured using a commercial leptin ELISA kit (R&D Systems, Inc., sensitivity to 7.80 pg/mL and intra-assay coefficient of variability <10 %), GH ELISA kit (R&D Systems, Inc., sensitivity to 7.18 pg/mL and intra-assay coefficient of variability <10%), PTH EIA kit (Sigma, USA, sensitivity to 1.27 pg/mL and intra-assay coefficient of variability <10 %), and 25(OH)D ELISA kit (DRG Diagnostics, Germany; sensitivity to 2.89 ng/mL and intra-assay coefficient of variability <10%), according to the manufacturers’ instructions. All assays were performed in duplicate.
5. Statistical analyses
This was a cross-sectional study. The sample size was estimated based on previously published data [30] using G*Power software. The minimal significance (a) and statistical power (1 - b) were set at 0.05 and 0.80, respectively. Calculations were performed for 2 groups (independent-samples t-test). The total sample size was 83 participants in 2 groups (group 1: 41 boys and group 2: 42 girls).
Descriptive statistics were presented as the mean and SD or the median and interquartile range for normally and nonnormally distributed data, respectively. The differences between boys and girls were analyzed using the independent-samples t-test or the nonparametric independent-samples test. Correlations between body composition parameters, bone parameters, and biomarkers were analyzed using bivariate (Pearson correlation) and multivariate (multiple regression) analyses separately for boys and girls. Pearson correlation and simple linear regression analyses were performed to examine the relationships among body composition, bone parameters, and bone biomarkers. Multiple linear regression analysis was applied to determine the relationships of each bone parameter (dependent variables) with body composition parameters and biomarkers (independent variables) that were significantly correlated with bone on bivariate analysis. Pearson correlation coefficient (r) and the standardized regression coefficient (β) were used to determine the strength of these correlations. Additionally, R2 values were the coefficients of determination.
Data analysis was performed using IBM SPSS Statistics ver. 22.0 (IBM Co., USA). P-values <0.05, 0.01, or 0.001 were considered statistically significant.
6. Ethical statement
This study was performed according to the Helsinki Declaration and approved by the Human Research Ethics Committee of Walailak University, Thailand (protocol number: 58/079). Written informed consent was obtained from all the participants or their legal representative before enrollment.
Results
1. Characteristics of participants
Boys and girls had mean ages of 13.46±0.51 and 13.21±0.51 years, respectively. Mean BMI and BMI z-scores of both sexes were not significantly different, but mean BW, BW z-score, height, height z-score, and WC of boys were higher than those of girls (P<0.001) as shown in Table 1. Boys and girls had no significant difference in mean BFM, %fat, or %lean, while the mean LBM of boys was more than that of girls (P<0.001). With regard to WB bone, there was no significant difference in BMD; however, boys had significantly higher BMC and lower BMD z-score than girls (P<0.001). Furthermore, serum BALP, OC, and CTX levels were significantly higher and serum leptin levels were significantly lower in boys than in girls (P<0.05). Serum GH, PTH, and 25(OH)D levels were not significantly different between boys and girls.
Descriptive data and comparisons of variables between boys and girls
2. Bivariate correlation analyses among body composition, bone parameters, and bone biomarkers in boys and girls
Pearson correlations and simple linear regression among body composition, bone parameters, and biomarkers in boys and girls are shown in Table 2 and Fig. 2, respectively. %Bone in boys and girls was strongly negatively correlated with %fat (β=-0.625 and -0.754, P<0.001) (Fig. 2A and B) and BFM (β=-0.670 and -0.773, P<0.001) (Fig. 2C and D), while BFM was strongly positively correlated with BMC in both sexes (boys, β=0.472, P=0.002; girls, β=0.707, P<0.001) (Fig. 2E and F) (Table 2). Table 2 shows that BFM and %fat in girls and %fat in boys were negatively correlated with BMD z-score. In both sexes, %lean was strongly correlated with %bone, LBM was slightly correlated with BMD and BMC, and %bone was strongly correlated with BMD z-score. Moreover, BFM and %fat were positively correlated with leptin and GH levels in boys and negatively correlated with PTH levels in girls (Table 2).
Pearson correlation coefficients (r) among body composition, BMI z-score, bone parameters, and biomarkers of boys and girls
Correlations between %Fat and %Bone, BFM and %Bone, and BFM and BMC of boy (A, C, E) and girl (B, D, F). %Fat, body fat percentage; %Bone, body bone percentage; BFM, body fat mass; BMC, bone mineral content.
In addition, Table 2 shows that BMI z-scores of boys and girls were positively correlated with BMD and BMC but negatively correlated with %bone. In boys, BMI z-score was negatively correlated with CTX levels. Furthermore, BMI z-scores were strongly correlated with BFM and %fat in boys (r=0.854 and 0.710, P<0.001) and girls (r=0.911 and 0.835, P<0.001) (data not shown).
Table 3 shows Pearson correlations between bone parameters and biomarkers in boys and girls. %Bone in both sexes were correlated with CTX and PTH levels. Moreover, %bone in boys was negatively correlated with GH levels. In addition, BMD z-score in boys was positively correlated with PTH and BALP levels but negatively correlated with 25(OH)D levels. In girls, BMD z-score was positively correlated with CTX, PTH, and leptin levels.
Pearson correlation coefficients (r) between whole-body bone parameters and biomarkers of boys and girls
Furthermore, correlations between biomarkers in boys and girls were analyzed (data not shown). In boys, PTH levels were positively correlated with CTX levels (r=0.463, P<0.01), but negatively correlated with 25(OH)D (r=-0.366, P<0.05). In girls, leptin levels were positively correlated with CTX (r=0.469, P<0.01), PTH (r=0.574, P<0.001), and 25(OH)D (r=0.309, P<0.05), but negatively correlated with OC (r=-0.488, P<0.01).
3. Multiple linear regression analysis among bone parameters, body composition, and biomarkers
Predictive factors for %bone and WB BMD z-score of boys and girls are shown in Table 4. In Models 1 and 2, %fat and BFM were key negative predictors of %bone in boys (β=-0.554 and -0.604, P<0.001) and girls (β=-0.666 and -0.683, P<0.001), while serum CTX and PTH were positive predictors of %bone in girls (β=0.208, P=0.042 and β=0.207, P=0.038). In Model 3, BFM in boys was a negative predictor and serum leptin was a positive predictor of %bone (β=-0.682, P=0.042 and β=0.310, P=0.014). Thus, increased %fat and BFM predicted decreased %bone in both sexes, while decreased %bone in girls involved serum CTX and PTH levels. In addition, in Model 4, BMI z-score of boys was a negative predictor of %bone, indicating that an increase in BMI z-score could predict a decrease in %bone.
Predictors for %bone and BMD z-score of boys and girls
Furthermore, with regard to BMD z-score of girls (model 1), %fat (not BFM) was a negative predictor (β=-0.329, P=0.021) and serum PTH was a positive predictor (β=0.319, P=0.024). In model 3, in girls, BFM and serum leptin were positive predictors of BMD z-score (β=0.867, P=0.026 and β=0.342, P=0.009), while %fat was a negative predictor (β=-1.257, P=0.002). In boys, %fat was a negative predictor (β=-1.011, P=0.017), and serum leptin was a positive predictor of the BMD z-score (β=0.391, P=0.012) (model 3). Therefore, increased %fat in boys and girls predicted a decrease in BMD z-score related to serum leptin levels in both sexes and serum PTH levels in girls.
Discussion
This study investigated correlations between body composition, bone parameters of the WB, and bone biomarkers in boys and girls aged 13–14 years with >95th percentile mean %fat for age [5,27]. The body composition of participants differed between sexes, with boys having a higher LBM than girls [6,7]. The increased BW of boys was due to an increase in LBM, whereas that of girls was related to increased BFM. This study found a negative effect of excess %fat and BFM on WB %bone and BMD z-scores of boys and girls. Compared with previous results [31,32], mean WB BMD of participants was slightly below age-specific mean WB BMD, and it did not adapt to increased BW [33], which might be related to the decreased the WB BMD z-score. Moreover, previous studies have reported that the deleterious effect of %fat on bone in children and adolescents [8,34] and the role of adipose tissue in bone are probably related to the secretion of inflammatory cytokines such as interleukin-6 and tumor necrosis factor-β, which stimulate osteoclast differentiation and promote bone resorption [35]. In addition, BMI z-score, %fat, and BFM of participants in this study were significantly correlated and BMI z-score was negatively correlated with %bone and CTX levels. Mean serum CTX and PTH levels in participants were above the 50th percentile for age and sex [12] and PTH and CTX correlated with WB %bone and BMD z-score in both sexes. Moreover, PTH and CTX were predictive factors of WB %bone and BMD z-score in girls with excess %fat and BFM. PTH exerts a catabolic effect on bone by stimulating bone resorption and the release of Ca2+ from bone [18], whereas CTX is a product of type 1 collagen degradation that reflects bone resorption [12]. Thus, serum CTX and PTH levels indicate bone resorption, which may decrease WB BMD z-score and %bone in adolescents with excess %fat. BMD z-score has been widely used to assess young individuals with a previous fracture or other major risk factors for osteoporosis, fracture, or accelerated bone loss, while whole-body composition on DXA is not a routine clinical application but can be useful in a selected population, such as those with sarcopenia (low LBM) and obesity with diseases or conditions associated with accelerated bone loss [36,37]. Due to the strong correlation between excess %fat and decreased %bone, which is associated with accelerated bone loss, %bone, which represents the absolute amount of minerals in the bone, should be evaluated in obese individuals in addition to BMC measurement, as it is useful for assessing skeletal health in the context of overall body composition.
This study found correlations among serum leptin, CTX, and PTH levels in girls, and among serum leptin, %fat, and BFM in boys. Moreover, serum leptin was a predictor of %bone in boys and BMD z-score in boys and girls with excess %fat. Therefore, decreased %bone and WB BMD z-score in participants with increased %fat may be modulated by serum PTH, CTX, and leptin. According to a previous report, leptin is a potential biomarker for childhood obesity because it promotes fat accumulation and %fat [38]. This study found sex differences in the increase in leptin concentration with excess %fat; serum leptin and %fat were higher in girls than in boys, which is consistent with a previous finding [39]. The mechanisms of satiety and appetite control are regulated by leptin, and obesity can result from an imbalance in energy expenditure and energy intake caused by changes in leptin sensitivity and serum leptin levels (called leptin resistance) [40]; thus, in this study, leptin resistance was suspected in girls, but a relationship between leptin resistance and obesity was not clear in boys.
In addition, serum GH levels of the participants were within normal range [41], and this study found correlations among GH levels, BFM, and %fat in boys, and a negative correlation between GH levels and %bone. Therefore, increased BFM and %fat related to GH are advantageous or detrimental to bone depending on sex. However, evidence reported that obesity in children has advantages in terms of growth during the prepubertal years, but this advantage gradually decreases during puberty, possibly due to age-related decreases GH secretion [22]. Moreover, obesity-related growth spurts are further influenced by leptin. Leptin plays a role in modulating growth and metabolic processes through the molecular mechanisms of the hypothalamic-pituitary-GH axis, directly stimulating GH secretion from the pituitary and reducing somatostatin secretion from the hypothalamus, which in turn increases GH release [42]. Childhood obesity is characterized by normal or even accelerated growth in spite of abnormalities in the GH-insulin-like growth factor-1 (GH-IGF-1) axis, which are mainly characterized by reduced GH secretion with normal IGF-1 levels. During puberty, obese children often show less of a growth spurt compared to lean subjects, with early signs of puberty in girls and either early or delayed puberty in boys [43]. In addition, hyperleptinemia can interact with the GH-IGF-1 axis; IGF-1 levels in obese adolescents can be normal or even elevated due to factors like elevated insulin [42-44].
This study found that LBM of boys and BFM of both sexes had a positive effect on WB-BMC. The mechanical load of LBM added to the skeleton increases BMD as a result of the potential of the mechanosensitive cells in bone tissue to adapt to mechanical stress [45]. Similarly, BFM is considered a protective factor for bone health due to the increased loading of high total body mass on bone. Thus, both LBM and BFM are positive determinants of bone [34].
Finally, this study found that serum BALP and OC levels in participants with excess %fat were lower compared to healthy adolescents in previous studies [14]. BALP and OC contribute to bone formation and skeletal mineralization. Serum BALP levels increase with age and peak at ages 12–13 years for boys and 9–10 years for girls, whereas a peak in serum OC levels is observed at ages 12–13 years in both sexes [14]. This study found a positive correlation between BALP levels and WB z-score; thus, low serum BALP levels may be related to low BMD z-score caused by excess %fat in boys.
This study limited the ages of the participants to a narrow range of 13–14 years old and comprised a small number of participants with in the >95th percentile of %fat for age and sex, which may have been insufficient to allow comparisons between sexes. In addition, the participants were drawn from a specific geographic and demographic group, which may limit the applicability of the findings to other populations with different ethnicities, lifestyles, or socioeconomic backgrounds. Furthermore, body composition and serum levels of bone biomarkers are age- and sex-specific, and the average ages at which boys and girls begin and complete puberty are different, but this study did not report pubertal stages of participants. Pubertal growth data in Thai children reported a mean age of pubertal onset of 10.2±1.2 years for girls and 12.2±1.0 years for boys [46]. Therefore, the age of participants should be extended according to puberty and sex, and participants with different lifestyles and socioeconomic backgrounds should be included in a longitudinal study. Further studies should focus on bone biomarkers and related hormones, including sex hormones that may affect bone growth in adolescents with obesity. However, the results of this study may be useful as primary data to examine the risk of bone loss based on body composition in a larger number of adolescents.
In conclusion, excess %fat has a deleterious effect on WB bone in adolescents of both sexes, and this may be a result of bone resorption. However, increased BFM has advantages in terms of WB-BMC in both boys and girls, and has a positive effect on GH levels in boys.
Notes
Conflicts of interest
No potential conflict of interest relevant to this article was reported.
Funding
This work was supported by the Research and Innovation Institute of Excellence, Walailak University, Thailand, under contact no. WU59105.
Data availability
The data that support the findings of this study can be provided by the corresponding author upon reasonable request.
Acknowledgments
The authors would like to thank all the participants for their cooperation, teachers at high schools in Nakhonsithammarat for their help and collaboration, and a research assistant for their help with DXA measurement.
Author contribution
Conceptualization: RK; Data curation: RK; Formal analysis: RK; Funding acquisition: RK, CP; Methodology: RK, CP; Project administration: RK; Visualization: RK; Writing - original draft: RK, CP; Writing - review & editing: RK, CP