ORIGINAL ARTICLE: HEPATOLOGY
AND
NUTRITION
High Bone Density in Adolescents With Obesity Is Related to Fat Mass and Serum Leptin Concentrations
Albane B.R. Maggio, yDominique C. Belli, zJulie Wacker Bou Puigdefabregas, §Rene´ Rizzoli, Nathalie J. Farpour-Lambert, zMaurice Beghetti, and yVale´rie A. McLin
ABSTRACT Objectives: Obesity has been associated with increased bone mass, but the mechanisms involved are still poorly understood. We aimed to explore the relation between bone mineral density and factors known to influence bone formation in obese and lean adolescents. Methods: We recruited 24 obese and 25 lean adolescents in a case-control study. Total body bone mineral density (TB-BMD) z scores and body composition were determined using dual-energy x-ray absorptiometry. We measured 25-hydroxyvitamin D (25-OH-D), glucose, insulin, and leptin concentrations. Physical activity (PA) level was quantified using accelerometer. Results: TB-BMD z score was higher, whereas 25-OH-D and PA levels were lower in obese compared with lean subjects (TB-BMD z score 1.06 0.96 vs 0.26 0.91, P ¼ 0.004; 25-OH-D 9.9 6.4 vs 18.5 7.4 ng mL1, P < 0.001; PA level 308.3 22.1 vs 406.8 29.2 count min1, P ¼ 0.01). TB-BMD z score was not related to 25-OH-D or PA levels, but was positively correlated with leptin concentration and fat mass (P < 0.05). Vitamin D concentration was negatively correlated with fat mass (P < 0.001). Conclusions: Despite lower serum vitamin D and PA levels, BMD was higher in adolescents with obesity and associated with higher serum leptin concentrations. Furthermore, adolescents with obesity have lower vitamin D serum concentrations than lean controls, probably owing to its distribution in adipose tissue. Key Words: bone mineral density, leptin, obesity, physical activity, vitamin D
(JPGN 2014;58: 723–728)
O
besity has been associated with increased bone mass in some studies (1,2), but conflicting results exist (3–5). In fact, the mechanisms involved are still poorly understood. Studies Received and accepted January 1, 2014. From the Pediatric Sport Medicine and Obesity Care Program, Service of Pediatric Specialties, the yPediatric Gastroenterology Unit, the zPediatric Cardiology Unit, Service of Pediatric Specialties, Department of Child and Adolescent, University Hospitals of Geneva and University of Geneva, and the §Service of Bone Diseases, Department of Rehabilitation and Geriatrics, University Hospitals of Geneva, Geneva, Switzerland. Address correspondence and reprint requests to Dr Albane Maggio, MD, Pediatric Sport Medicine and Obesity Care Program, Department of Child and Adolescent, 6 rue Willy-Donze´, 1211 Geneva 14, Switzerland (e-mail: [email protected]). This study was supported by the Prim’Enfance Foundation and the Geneva University Hospital Research and Development Fund. The work was independent of the funding. The authors report no conflicts of interest. Copyright # 2014 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition DOI: 10.1097/MPG.0000000000000297
JPGN
Volume 58, Number 6, June 2014
in obese subjects have demonstrated higher bone mineral density (BMD) with increased fat mass (FM), but the respective effect of weight-induced mechanical loading and biochemical effects of adipose tissue is still debated. Vitamin D concentration is known to be decreased in obese subjects (6,7), and may contribute to lower BMD, as suggested in some studies (3,4). Vitamin D has an important role in bone growth before birth, in early life, and during adolescence (8,9). Conversely, the role of vitamin D on bone remodeling has been questioned lately because a few studies failed to show an association between BMD and vitamin D concentration in healthy and obese adolescents (10,11). In fact, BMD studies in obese subjects are scarce and differences may exist because of different bone sites studied. Adolescence is a critical period for bone growth during which bone mass accrual and remodeling determine the peak bone mass (12). Many factors are implicated in adequate peak bone mass acquisition (12): weight, physical activity (PA), genetic predisposition, and other environmental and hormonal factors. A high-impact PA has been shown to increase areal bone mineral density (aBMD or bone mineral mass) when performed during childhood and adolescence. aBMD integrates the size of the bone, its thickness, and the true volumetric density. It is assumed to account for up to 17% of BMD variance between individuals in their late 20s (13). Furthermore, polymorphisms in several genes, including vitamin D and leptin receptors, have also been suggested to influence bone mass (14). The effect of hormones such as insulin-like growth factor 1 (IGF-1), leptin, and insulin on various aspects of bone metabolism are, however, not well understood in healthy or obese individuals (15–17). Because serum concentrations of these hormones are altered in obese subjects, examining their relation to BMD may offer insight into indicators influencing bone formation in adolescents. The aim of this study was to investigate vitamin D, hormones, body composition (fat and lean mass), and PA in obese adolescents compared with normal weight-matched healthy controls, and to assess their possible relation to bone mineral mass.
METHODS Study Design and Subjects This case-control study was performed in 50 obese and lean adolescents, ages between 10 and 16 years. The main aim of the protocol was to compare cardiovascular risk factors between obese and lean adolescents. The present study deals with the secondary aims, which were to assess their bone mineral status, metabolism, and PA level. Obese subjects were recruited between April and August 2009 from the Pediatric Obesity Care Clinic of the Geneva University Hospital. Obesity was defined as body mass index (BMI) above the 97th percentile for age and sex (European standard) (18). Lean children were recruited according to the same age range
723
Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
Maggio et al through advertising during the same period. Their height and weight were within the normal range 2 standard deviations (SDs), and their BMI was under the 90th percentile for age and sex. The inclusion criteria were onset of puberty, no previous diagnosis of hypertension, no previous antihypertensive or antidiabetic treatment, no previous use of medications affecting glucose and lipid metabolism, no history of familial hypertension or familial dyslipidemia, absence of diabetes or other chronic disease, and no participation in competition sports. One obese adolescent had to be excluded after initial laboratory testing because of previously unknown familial hypercholesterolemia. A total of 24 obese and 25 lean adolescents completed the study. Informed written consent was obtained from both parents and child. The mother and child ethics committee of the University Hospitals of Geneva approved the study (02.2009).
Measures Anthropometrics and Body Composition We assessed body weight in kilograms and height in centimeters, and calculated body mass index (BMI) as weight/height squared (kg m2). BMI z scores were derived using the World Health Organization references (19). FM (kg), lean body mass (LBM, kg), percentage of trunk fat, and percentage of total body (TB) fat were measured using dual-energy x-ray absorptiometry (DXA; GE Lunar Prodigy, Lunar Corp, Madison, WI). Pubertal development was assessed using Tanner stages by means of a validated self-assessment questionnaire (20). Data were complete for all of the obese subjects but were missing in 14 lean children (56%); however, all of them had entered puberty.
Biomarkers Blood samples were collected by phlebotomy following a 10-hour overnight fast. Vitamin D status was evaluated using serum 25(OH)D3 concentration (ng mL1). Its concentration was measured using enzyme-linked immunosorbent assay method (Immunodiagnostik AG, Bensheim, Germany). Vitamin D concentration (25-hydroxyvitamin D [25-OH-D]) was separated in 3 categories: deficiency, 30 ng mL1 (21). Serum leptin and IGF-1 were measured by colorimetric enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN), following manufacturer’s instructions. The limit of detection was 2.56 ng mL1 for 25-OH-D, 62.50 pg mL1 for leptin, and 31.25 pg mL1 for IGF-1. Mean intra- and interassay variation coefficients were 0.05 for both). Despite higher insulin concentration and insulin resistance indices (Table 1), no obese adolescents presented glucose intolerance or type 2 diabetes mellitus upon oral glucose tolerance test (results published (27)). Insulin and leptin concentrations were higher in girls (insulin, P ¼ 0.037; leptin, P ¼ 0.032) compared with boys. The sex differences for serum leptin concentrations persisted when separating lean and obese subjects (lean, P ¼ 0.019; obese, P ¼ 0.016), but insulin concentrations were significantly higher only in lean girls (P ¼ 0.022). We did not find any difference in IGF-1 concentration between groups or sexes.
Relation Between Bone Mineral Density and Other Variables We investigated the relation between TB aBMD z score and factors known to exert an effect on bone growth. Results of the linear regression are presented in Table 2. TB aBMD z score was highly positively related to BMI z score (Fig. 1A), lean and FM, insulin, and leptin concentrations (Fig. 1B), but negatively to leisure-time PA. Furthermore, TB aBMD z score was neither related to total daily PA level or time spent in moderate-to-vigorous PA (MVPA), nor related to 25-OH-D concentration (Fig. 1C), even when separating lean and obese subjects (lean, F ¼ 1.98, P ¼ 0.172; obese, F ¼ 0.325, P ¼ 0.575). Results were similar for BMC (data not shown). Next, we looked for sex differences in BMD and possible associations. Insulin and LBM were related to TB aBMD z score in girls only (insulin: girls, P ¼ 0.027; boys, P ¼ 0.160, and LBM: girls, P < 0.001, boys, P ¼ 0.522). There were no other sex differences. Standard multiple regressions were performed to assess which body weight constituent and which biological factor were most associated to BMD (Table 2) and BMC. FM was the only variable related to TB aBMD z score, whereas both LBM and FM were related to BMC (FM: b, 0.306, P ¼ 0.006; LBM: b, 0.575, P < 0.001). Next, leptin correlated better than insulin to TB aBMD z score; and none of these hormones were statistically related to BMC (leptin: b, 0.169, P ¼ 0.346; insulin: b, 0.137, P ¼ 0.445).
DISCUSSION The present study shows that adolescents with obesity had higher BMD than lean subjects, despite lower vitamin D and PA levels. The difference in BMD was independent of LBM. Instead, it was related to high serum leptin and insulin concentrations. The relation to leptin concentration is likely secondary to increased biologically active FM relative to lean subjects. The relation to serum insulin is probably linked to a direct anabolic effect through its receptor in bone. In addition, the present study confirms that adolescents with obesity have lower vitamin D serum concentrations than lean controls, probably because of its distribution in
725
Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
Maggio et al
JPGN
Volume 58, Number 6, June 2014
TABLE 2. Results of linear and multiple regression analysis for total BMD z score Univariate analysis
Age BMI z score Body composition LBM FM Trunk fat, % Biomarkers Insulin IGF-1 Leptin 25-OH-D PA Total daily PA Vigorous PA MVPA School PA Leisure time PA
Multivariate analysis
r2 change, %
b
t
P
r2 change, %
2.1 37.3 30.1 11.4 29.8 20.7 20.4 15.3 2.0 19.0 1.8
0.012 0.621 11.3 0.364 0.559 0.472 7.1 0.413 0.028 0.455 0.057
0.1 5.4
AND
NUTRITION
High Bone Density in Adolescents With Obesity Is Related to Fat Mass and Serum Leptin Concentrations
Albane B.R. Maggio, yDominique C. Belli, zJulie Wacker Bou Puigdefabregas, §Rene´ Rizzoli, Nathalie J. Farpour-Lambert, zMaurice Beghetti, and yVale´rie A. McLin
ABSTRACT Objectives: Obesity has been associated with increased bone mass, but the mechanisms involved are still poorly understood. We aimed to explore the relation between bone mineral density and factors known to influence bone formation in obese and lean adolescents. Methods: We recruited 24 obese and 25 lean adolescents in a case-control study. Total body bone mineral density (TB-BMD) z scores and body composition were determined using dual-energy x-ray absorptiometry. We measured 25-hydroxyvitamin D (25-OH-D), glucose, insulin, and leptin concentrations. Physical activity (PA) level was quantified using accelerometer. Results: TB-BMD z score was higher, whereas 25-OH-D and PA levels were lower in obese compared with lean subjects (TB-BMD z score 1.06 0.96 vs 0.26 0.91, P ¼ 0.004; 25-OH-D 9.9 6.4 vs 18.5 7.4 ng mL1, P < 0.001; PA level 308.3 22.1 vs 406.8 29.2 count min1, P ¼ 0.01). TB-BMD z score was not related to 25-OH-D or PA levels, but was positively correlated with leptin concentration and fat mass (P < 0.05). Vitamin D concentration was negatively correlated with fat mass (P < 0.001). Conclusions: Despite lower serum vitamin D and PA levels, BMD was higher in adolescents with obesity and associated with higher serum leptin concentrations. Furthermore, adolescents with obesity have lower vitamin D serum concentrations than lean controls, probably owing to its distribution in adipose tissue. Key Words: bone mineral density, leptin, obesity, physical activity, vitamin D
(JPGN 2014;58: 723–728)
O
besity has been associated with increased bone mass in some studies (1,2), but conflicting results exist (3–5). In fact, the mechanisms involved are still poorly understood. Studies Received and accepted January 1, 2014. From the Pediatric Sport Medicine and Obesity Care Program, Service of Pediatric Specialties, the yPediatric Gastroenterology Unit, the zPediatric Cardiology Unit, Service of Pediatric Specialties, Department of Child and Adolescent, University Hospitals of Geneva and University of Geneva, and the §Service of Bone Diseases, Department of Rehabilitation and Geriatrics, University Hospitals of Geneva, Geneva, Switzerland. Address correspondence and reprint requests to Dr Albane Maggio, MD, Pediatric Sport Medicine and Obesity Care Program, Department of Child and Adolescent, 6 rue Willy-Donze´, 1211 Geneva 14, Switzerland (e-mail: [email protected]). This study was supported by the Prim’Enfance Foundation and the Geneva University Hospital Research and Development Fund. The work was independent of the funding. The authors report no conflicts of interest. Copyright # 2014 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition DOI: 10.1097/MPG.0000000000000297
JPGN
Volume 58, Number 6, June 2014
in obese subjects have demonstrated higher bone mineral density (BMD) with increased fat mass (FM), but the respective effect of weight-induced mechanical loading and biochemical effects of adipose tissue is still debated. Vitamin D concentration is known to be decreased in obese subjects (6,7), and may contribute to lower BMD, as suggested in some studies (3,4). Vitamin D has an important role in bone growth before birth, in early life, and during adolescence (8,9). Conversely, the role of vitamin D on bone remodeling has been questioned lately because a few studies failed to show an association between BMD and vitamin D concentration in healthy and obese adolescents (10,11). In fact, BMD studies in obese subjects are scarce and differences may exist because of different bone sites studied. Adolescence is a critical period for bone growth during which bone mass accrual and remodeling determine the peak bone mass (12). Many factors are implicated in adequate peak bone mass acquisition (12): weight, physical activity (PA), genetic predisposition, and other environmental and hormonal factors. A high-impact PA has been shown to increase areal bone mineral density (aBMD or bone mineral mass) when performed during childhood and adolescence. aBMD integrates the size of the bone, its thickness, and the true volumetric density. It is assumed to account for up to 17% of BMD variance between individuals in their late 20s (13). Furthermore, polymorphisms in several genes, including vitamin D and leptin receptors, have also been suggested to influence bone mass (14). The effect of hormones such as insulin-like growth factor 1 (IGF-1), leptin, and insulin on various aspects of bone metabolism are, however, not well understood in healthy or obese individuals (15–17). Because serum concentrations of these hormones are altered in obese subjects, examining their relation to BMD may offer insight into indicators influencing bone formation in adolescents. The aim of this study was to investigate vitamin D, hormones, body composition (fat and lean mass), and PA in obese adolescents compared with normal weight-matched healthy controls, and to assess their possible relation to bone mineral mass.
METHODS Study Design and Subjects This case-control study was performed in 50 obese and lean adolescents, ages between 10 and 16 years. The main aim of the protocol was to compare cardiovascular risk factors between obese and lean adolescents. The present study deals with the secondary aims, which were to assess their bone mineral status, metabolism, and PA level. Obese subjects were recruited between April and August 2009 from the Pediatric Obesity Care Clinic of the Geneva University Hospital. Obesity was defined as body mass index (BMI) above the 97th percentile for age and sex (European standard) (18). Lean children were recruited according to the same age range
723
Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
Maggio et al through advertising during the same period. Their height and weight were within the normal range 2 standard deviations (SDs), and their BMI was under the 90th percentile for age and sex. The inclusion criteria were onset of puberty, no previous diagnosis of hypertension, no previous antihypertensive or antidiabetic treatment, no previous use of medications affecting glucose and lipid metabolism, no history of familial hypertension or familial dyslipidemia, absence of diabetes or other chronic disease, and no participation in competition sports. One obese adolescent had to be excluded after initial laboratory testing because of previously unknown familial hypercholesterolemia. A total of 24 obese and 25 lean adolescents completed the study. Informed written consent was obtained from both parents and child. The mother and child ethics committee of the University Hospitals of Geneva approved the study (02.2009).
Measures Anthropometrics and Body Composition We assessed body weight in kilograms and height in centimeters, and calculated body mass index (BMI) as weight/height squared (kg m2). BMI z scores were derived using the World Health Organization references (19). FM (kg), lean body mass (LBM, kg), percentage of trunk fat, and percentage of total body (TB) fat were measured using dual-energy x-ray absorptiometry (DXA; GE Lunar Prodigy, Lunar Corp, Madison, WI). Pubertal development was assessed using Tanner stages by means of a validated self-assessment questionnaire (20). Data were complete for all of the obese subjects but were missing in 14 lean children (56%); however, all of them had entered puberty.
Biomarkers Blood samples were collected by phlebotomy following a 10-hour overnight fast. Vitamin D status was evaluated using serum 25(OH)D3 concentration (ng mL1). Its concentration was measured using enzyme-linked immunosorbent assay method (Immunodiagnostik AG, Bensheim, Germany). Vitamin D concentration (25-hydroxyvitamin D [25-OH-D]) was separated in 3 categories: deficiency, 30 ng mL1 (21). Serum leptin and IGF-1 were measured by colorimetric enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN), following manufacturer’s instructions. The limit of detection was 2.56 ng mL1 for 25-OH-D, 62.50 pg mL1 for leptin, and 31.25 pg mL1 for IGF-1. Mean intra- and interassay variation coefficients were 0.05 for both). Despite higher insulin concentration and insulin resistance indices (Table 1), no obese adolescents presented glucose intolerance or type 2 diabetes mellitus upon oral glucose tolerance test (results published (27)). Insulin and leptin concentrations were higher in girls (insulin, P ¼ 0.037; leptin, P ¼ 0.032) compared with boys. The sex differences for serum leptin concentrations persisted when separating lean and obese subjects (lean, P ¼ 0.019; obese, P ¼ 0.016), but insulin concentrations were significantly higher only in lean girls (P ¼ 0.022). We did not find any difference in IGF-1 concentration between groups or sexes.
Relation Between Bone Mineral Density and Other Variables We investigated the relation between TB aBMD z score and factors known to exert an effect on bone growth. Results of the linear regression are presented in Table 2. TB aBMD z score was highly positively related to BMI z score (Fig. 1A), lean and FM, insulin, and leptin concentrations (Fig. 1B), but negatively to leisure-time PA. Furthermore, TB aBMD z score was neither related to total daily PA level or time spent in moderate-to-vigorous PA (MVPA), nor related to 25-OH-D concentration (Fig. 1C), even when separating lean and obese subjects (lean, F ¼ 1.98, P ¼ 0.172; obese, F ¼ 0.325, P ¼ 0.575). Results were similar for BMC (data not shown). Next, we looked for sex differences in BMD and possible associations. Insulin and LBM were related to TB aBMD z score in girls only (insulin: girls, P ¼ 0.027; boys, P ¼ 0.160, and LBM: girls, P < 0.001, boys, P ¼ 0.522). There were no other sex differences. Standard multiple regressions were performed to assess which body weight constituent and which biological factor were most associated to BMD (Table 2) and BMC. FM was the only variable related to TB aBMD z score, whereas both LBM and FM were related to BMC (FM: b, 0.306, P ¼ 0.006; LBM: b, 0.575, P < 0.001). Next, leptin correlated better than insulin to TB aBMD z score; and none of these hormones were statistically related to BMC (leptin: b, 0.169, P ¼ 0.346; insulin: b, 0.137, P ¼ 0.445).
DISCUSSION The present study shows that adolescents with obesity had higher BMD than lean subjects, despite lower vitamin D and PA levels. The difference in BMD was independent of LBM. Instead, it was related to high serum leptin and insulin concentrations. The relation to leptin concentration is likely secondary to increased biologically active FM relative to lean subjects. The relation to serum insulin is probably linked to a direct anabolic effect through its receptor in bone. In addition, the present study confirms that adolescents with obesity have lower vitamin D serum concentrations than lean controls, probably because of its distribution in
725
Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
Maggio et al
JPGN
Volume 58, Number 6, June 2014
TABLE 2. Results of linear and multiple regression analysis for total BMD z score Univariate analysis
Age BMI z score Body composition LBM FM Trunk fat, % Biomarkers Insulin IGF-1 Leptin 25-OH-D PA Total daily PA Vigorous PA MVPA School PA Leisure time PA
Multivariate analysis
r2 change, %
b
t
P
r2 change, %
2.1 37.3 30.1 11.4 29.8 20.7 20.4 15.3 2.0 19.0 1.8
0.012 0.621 11.3 0.364 0.559 0.472 7.1 0.413 0.028 0.455 0.057
0.1 5.4
