Abstract presented as a poster at the annual meeting of the Society for Epidemiologic Research, Toronto, Canada, 2001
Article first published online: 1 OCT 2002
Copyright © 2002 ASBMR
Journal of Bone and Mineral Research
Volume 17, Issue 10, pages 1896–1903, October 2002
How to Cite
Ruhl, C. E. and Everhart, J. E. (2002), Relationship of Serum Leptin Concentration With Bone Mineral Density in the United States Population. J Bone Miner Res, 17: 1896–1903. doi: 10.1359/jbmr.2002.17.10.1896
The authors have no conflict of interest.
- Issue published online: 2 DEC 2009
- Article first published online: 1 OCT 2002
- Manuscript Revised: 26 APR 2002
- Manuscript Accepted: 26 APR 2002
- Manuscript Received: 11 JAN 2002
- bone mineral density;
- Third National Health and Nutrition Examination Survey
Overweight is associated with both higher bone mineral density (BMD) and higher serum leptin concentrations. In humans, little is known about the relationship of leptin concentration and bone density. We studied this relationship in a large, national population-based sample. Participants included 5815 adults in the Third U.S. National Health and Nutrition Examination Survey (NHANES III; 1988–1994) who underwent DXA of the proximal femur and measurement of fasting serum leptin. Mean ± SE BMD (gm/cm2) of the total hip was 1.01 ± 0.005 in men, 0.94 ± 0.004 in premenopausal women, and 0.78 ± 0.007 in postmenopausal women. Bone density increased with increasing leptin concentration in men (p = 0.003), premenopausal women (p < 0.001), and postmenopausal women (p < 0.001). However, after adjusting for body mass index (BMI) and other bone density-related factors, an inverse association emerged in men (p < 0.001), being most evident among men <60 years old. There was no association of leptin and BMD in premenopausal women (p = 0.66) or postmenopausal women (p = 0.69) in multivariate analysis. Controlling for leptin had no effect on the strong positive association of BMI and BMD in either men or women. Serum leptin concentration did not appear to affect directly BMD. If present, the association appeared to be limited to younger men who are at lower risk of osteoporosis.
Being overweight is associated with higher bone mineral density (BMD) and decreased risk of fracture.(1–3) The mechanisms responsible are incompletely understood, but those proposed include muscle-mediated mechanical effects of increased weight bearing,(4) increased aromatization of androgen to estrogen in adipose tissue,(5) decreased sex hormone-binding globulin with increased free sex steroids,(6) and hyperinsulinemia.(7)
Leptin, the protein product of the obesity (ob) gene, is synthesized and secreted by adipocytes, and serum concentrations are highly correlated with adipose tissue mass.(8–9) Leptin binds to receptors in the hypothalamus and influences expression of several neuropeptides that regulate energy intake and expenditure. Although rare individuals with extreme obesity are leptin deficient, most obese persons have hyperleptinemia proportionate to body fat and appear to be leptin resistant.(8) In addition to its role in food intake and energy balance, leptin has been found to exert an influence on a variety of endocrine axes, particularly on the hypothalamic-pituitary-gonadal axis and on insulin metabolism.(10)
Although both BMD and serum leptin concentration are increased with obesity, nonhuman models suggest that leptin may be related inversely to bone density.(11) Obese leptin-deficient (ob/ob) mice have high BMD, even when obesity is prevented by diet, while intracerebroventricular leptin administration to these animals decreases bone mass. In contrast, leptin reduced ovariectomy-induced bone loss in rats.(12) In humans, the relationship of leptin concentration and BMD is unclear. No association of leptin concentration with bone density, independent of overall adiposity, was found in four small nonpopulation studies of middle-aged or postmenopausal women using various definitions of bone mass.(13–16) There was an inconsistent positive relationship of leptin with BMD at regional sites in premenopausal but not postmenopausal women in one study(17) and at one or more but not all bone sites in studies of nonobese women(18) and postmenopausal women.(19) Bone density was inversely associated with leptin concentration in a study of men.(20) In a small, population-based study, leptin was unassociated with BMD independent of fat mass in women, and there was an inconsistent inverse relationship in men.(21)
To examine the relationship of leptin concentration with BMD in adult men and women, we measured serum leptin concentration in a large, representative U.S. population sample and studied its relationship with proximal femoral BMD.
MATERIALS AND METHODS
The Third National Health and Nutrition Examination Survey (NHANES III) was conducted in the United States from 1988 through 1994 by the National Center for Health Statistics of the Centers for Disease Control and Prevention (CDC). It consisted of an interview, an examination, and laboratory data collected from a complex multistage stratified clustered probability sample of the civilian noninstitutionalized population aged ≥2 months, with oversampling of the elderly, non-Hispanic blacks, and Mexican Americans.(22) The study was approved by the CDC Institutional Review Board and all participants provided written consent to participate.
The sample for this study consisted of 2761 men, 1906 premenopausal women, and 1148 postmenopausal women aged ≥20 years who were examined in the morning at a mobile examination center after an overnight fast. Excluded were persons who fasted <9 h (n = 536) or ≥24 h (n = 9) or whose fasting time was unknown (n = 129), those for whom no serum specimen was available (n = 732) or the quantity was inadequate to perform the leptin assay (n = 11), pregnant women (n = 105), and persons with missing data on BMD (n = 495). Women were considered to be postmenopausal if they were ≥55 years (n = 1052) or if they had a uterus and at least one ovary, were not breast-feeding, and reported not having had a menstrual period during the past 12 months (n = 96).
Total proximal femoral areal BMD (bone mass per unit of area scanned) was measured by DXA (Hologic, Inc. QDR-1000; Hologic, Inc., Waltham, MA, USA).(23–24)
A morning, fasting venous blood sample was collected. Serum specimens were stored at −70°C and went through at least one freeze-thaw cycle during an average of 8 years of storage before measurement of leptin concentrations. Leptin has been shown to remain stable through five freeze-thaw cycles(25) and after storage for as long as 29 years.(26) Serum leptin concentrations were measured by radioimmunoassay at Linco Research, Inc. (St. Charles, MO, USA).(25) The minimum detectable concentration of the assay is 0.5 μg/liter. Intra- and interassay coefficients of variation are both <5%.
Data were collected on factors known or thought to be related to BMD based on other studies, including age (years); ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, or other); education (less than high school, some high school, high school graduate, some college, college graduate, or graduate school); body mass index (BMI; kg/m2); cigarette smoking (never, former, less than one pack per day, or one or more packs per day); alcohol drinking (never, former, less than one drink per day, one to two drinks per day, or more than two drinks per day); a doctor diagnosis of diabetes mellitus, congestive heart failure, stroke, chronic bronchitis, goiter, or other thyroid diseases; use of adrenal corticosteroids or thiazide diuretics; and for women, number of live births, a bilateral oophorectomy, oral contraceptive use (never, former, or current), and postmenopausal oral estrogen use (never, former, or current).
A BMI change variable (kg/m2) was created by subtracting one of the following measures from the current BMI: maximum BMI (persons aged 20–25 years), BMI at the age of 25 years (persons aged 26–35 years), or BMI 10 years ago (persons aged 36 years and older). All BMI values used in the calculation of the BMI change variable were based on reported weights (reported weights were highly correlated with measured weights, ρ = 0.98). A physical activity intensity-level variable was created by summing the products of activity frequency in the previous month and an intensity rating (or rate of energy expenditure, calculated as the ratio of work metabolic rate to resting metabolic rate [METs](27)) for nine common activities (walking, jogging or running, bicycling, swimming, aerobics, dancing, calisthenics, gardening or yard work, and weight lifting). Total caffeine intake from beverages (mg/day) was calculated based on 136 mg/cup of regular coffee, 64 mg/cup of regular tea, and 46 mg/bottle or can of cola or soda.(28) Daily food energy (kcal), protein (g), total fat (g), and calcium (mg) intakes were calculated from a 24-h dietary recall, as was daily calcium intake from supplements. Serum vitamin D (ng/ml), thyroxine (μg/dl), thyroid-stimulating hormone (TSH; μU/ml), and fasting insulin (pmol/liter) concentrations were measured as previously described.(29)
Because of differences in serum leptin concentration and BMD, separate analyses were conducted for men and pre- and postmenopausal women. Relations of leptin with continuous bone density-related factors were examined by calculating weighted Pearson's correlation coefficients (ρ). To investigate the relation of leptin concentration and bone density, we controlled for the effects of BMI and other bone density-related factors using multiple linear regression analysis. We examined the relation of bone density with leptin concentration both by treating leptin concentration as a continuous variable and by calculating mean bone density estimates for leptin concentration quintiles. For the latter analysis, leptin quintiles were categorized as indicator variables and mean bone density was estimated by least squares means computed using SUDAAN PROC REGRESS (Research Triangle Institute, Research Triangle Park, NC, USA).(30) The p values for trend in mean bone density across leptin quintiles were computed by including leptin concentration in the models as an ordinal variable of five levels. Interaction terms for leptin and covariates were evaluated by adding them individually to multivariate models. Because leptin concentrations were skewed to the right, log10 transformation was performed to normalize the distribution for regression analyses with leptin as a continuous variable. Because of the large number of factors potentially related to bone density, only those factors that were associated with bone density in multivariate analysis (p < 0.1) were retained in the models to which leptin was added. Multivariate analyses excluded persons with missing values for any factor included in the model (numbers excluded for men and premenopausal and postmenopausal women, respectively, were 214, 227, and 210). A β-coefficient with an associated value of p < 0.05 was considered statistically significant. Analyses incorporated sample weights, stratification, and clustering using SUDAAN software.(30)
Characteristics of men and pre- and postmenopausal women in the study sample are shown in Table 1. Mean BMD of the total hip was higher in men than in women and higher in premenopausal women than postmenopausal women (p < 0.001). Conversely, mean fasting serum leptin concentration was higher in women than in men and higher in postmenopausal women than in premenopausal women (p < 0.001).
Mean bone density increased with increasing leptin concentration in men (p = 0.003; Table 2). However, after adjusting for BMI, an inverse association emerged between leptin concentration and bone density (p < 0.001; Table 2). The inverse association remained in multivariate analyses controlling for BMI and age (p < 0.001; Table 2) or BMI, age, and other bone density-related factors (p < 0.001; Table 2 and Fig. 1). This inverse relationship was examined further in subgroups based on BMI or age. The relationship of leptin and bone density varied with level of BMI (test for interaction, p = 0.011). When examined by BMI subgroup (normal weight, <25 kg/m2; overweight, 25-<30 kg/m2; obese, ≥30 kg/m2) in multivariate analysis, the inverse association of bone density with leptin concentration was limited to overweight men (p = 0.004), and there was no relationship in normal weight (p = 0.52) or obese (p = 0.58) men (Fig. 2A). Interaction was also found between leptin concentration and age (p < 0.001). In multivariate analysis by age group, the inverse association of bone density with leptin concentration was found in young (p = 0.002) and middle-aged (p = 0.003) men, and there was no relationship in older men (p = 0.27; Fig. 2B).
The effect of fasting serum insulin concentration on the relation between serum leptin concentration and BMD was examined in the subgroup of men without a doctor diagnosis of diabetes. Insulin concentration was correlated with leptin concentration (ρ = 0.64) and was strongly associated with bone density (β = 0.049, SE = 0.0066, and p < 0.001 for log-transformed insulin concentration [pmol/liter]). After adjusting for other bone density-related factors, insulin concentration was inversely associated with bone density (β = −0.018, SE = 0.0070, and p = 0.012). However, when insulin and leptin concentrations were included together in the multivariate model, insulin concentration was not independently associated with bone density (β = −0.0038, SE = 0.0088, and p = 0.67) and had little effect on the inverse association of leptin concentration with bone density.
As expected, BMD was strongly associated with BMI (β = 0.014 per kg/m2, SE = 0.0012, and p < 0.001 in unadjusted analysis) and adjusting for leptin concentration strengthened this relationship (β = 0.022 per kg/m2, SE = 0.0015, and p < 0.001). This positive relationship of bone density with BMI after controlling for leptin concentration is also seen in Fig. 2A where bone density increased with increasing BMI within each leptin concentration quintile.
Among premenopausal women, mean bone density increased with increasing leptin concentration (p < 0.001; Table 2). After adjusting for BMI, an inverse association was found (p = 0.019; Table 2). This relationship was diminished after controlling for age (p = 0.076; Table 2) and eliminated after adjusting for additional bone density-related factors (Table 2 and Fig. 1). Among premenopausal women without doctor-diagnosed diabetes, insulin concentration was correlated with leptin concentration (ρ = 0.64) and was strongly associated with bone density (β = 0.076, SE = 0.012, and p < 0.001 for log-transformed insulin concentration [pmol/liter]) but not independent of other bone density-related factors (β = −0.016, SE = 0.013, and p = 0.22). Insulin concentration had no effect on the lack of a relationship between leptin concentration and bone density in multivariate analysis. As in men, BMD was strongly associated with BMI (β = 0.010 per kg/m2, SE = 0.00072, and p < 0.001) and adjusting for leptin concentration did not diminish this relationship (β = 0.012 per kg/m2, SE = 0.0012, and p < 0.001).
Among postmenopausal women, mean bone density increased with increasing leptin concentration (p < 0.001; Table 2). After adjusting for BMI, no association was found (p = 0.16; Table 2). This relationship was unchanged by adding age (p = 0.24) (Table 2) or additional bone density-related factors (Table 2 and Fig. 1) to the model. Among postmenopausal women without doctor-diagnosed diabetes, insulin concentration was correlated with leptin concentration (ρ = 0.64) and strongly associated with bone density (β = 0.086, SE = 0.011, and p < 0.001 for log-transformed insulin concentration [pmol/liter]) but not independent of other bone density-related factors (β = 0.010, SE = 0.013, and p = 0.44) and had no effect on the lack of a relationship between leptin concentration and bone density. Among postmenopausal women, there was a relationship between leptin and bone density that depended on estrogen-replacement status (p value for interaction = 0.027). Thus, on stratifying by estrogen-replacement status in multivariate analysis, an inverse association of bone density with leptin concentration was found among current users (p = 0.013), whereas there was no relationship among never users (p = 0.95) and former users (p = 0.13; Fig. 3). BMD was strongly associated with BMI (β = 0.014 per kg/m2, SE = 0.0010, and p < 0.001) and this relationship remained after adjusting for leptin concentration (β = 0.012 per kg/m2, SE = 0.0014, and p < 0.001).
The results of this large, national, population-based, cross-sectional study of adult men and women do not provide evidence for a direct effect of serum leptin concentration on BMD. In addition, leptin concentration does not appear to mediate the positive relationship of BMD with BMI. Among men, this latter relationship is seen graphically in Fig. 2A. Small studies of selected samples of women(13–19) or men(20) and a smaller population study of men and women(21) have, likewise, found no strong associations of leptin concentration with bone density. In mice, leptin has been found to inhibit bone formation, acting through a central control mechanism involving the hypothalamus.(11) Obese, leptin-deficient (ob/ob) and leptin receptor-deficient (db/db) mice have increased bone formation leading to high bone mass. The high bone mass phenotype cannot be explained by obesity alone and occurs despite the negative influences of hypogonadism and hypercortisolism. Intracerebroventricular leptin administration completely reverses the high bone mass of ob/ob mice and induces bone loss in wild-type mice. However, leptin and leptin receptor-deficient mice represent extreme situations in which there is a total loss of leptin function, which may not be relevant to human populations. In obese humans, high leptin concentrations can be associated with leptin resistance because of impaired transport of leptin to the hypothalamus. The lack of a direct correlation of leptin concentration with bone density in our study and other human studies may be caused by, at least in part, decreased leptin function despite high levels that occur with leptin resistance.(31)
Serum leptin and insulin concentrations are highly correlated, and both have been considered part of the metabolic syndrome and its attendant disease-related complications.(32) Insulin levels have also been shown to be related to bone density.(7,33) However, in this analysis, insulin concentration was not associated independently with BMD and did not influence the relationship of leptin concentration to bone density. Nevertheless, because of the limitations of a cross-sectional study, we cannot exclude a relationship of leptin and insulin to bone density.
In multivariate analyses, we found a negative association of serum leptin concentrations with BMD in certain hormonally influenced situations such as younger men, who have higher testosterone levels,(34) and postmenopausal women on hormone-replacement therapy. Additionally, in premenopausal women who would be expected to have high estrogen levels, there was an inverse association independent of BMI but not of additional bone density-related factors (Table 2).
These intriguing findings that suggest a hormonally mediated relationship of leptin with BMD could not be evaluated further because sex steroid levels were not measured in NHANES III. Nevertheless, other studies have suggested a relationship that should be pursued. Among men, serum leptin levels have been found to correlate negatively with total and bioavailable testosterone levels.(35) Testosterone appears to increase bone density in men.(36) It is possible that the inverse association of bone density with leptin concentration that we found in young and middle-aged men was caused by confounding by serum testosterone concentration rather than to an effect of leptin on bone density. In addition, bioavailable estrogen has been shown to correlate with bone density in men.(21,37) Among postmenopausal women not on hormone-replacement therapy, serum leptin levels have been found to correlate positively with bioavailable estrogen.(35) As with men, bioavailable estrogen has been shown to correlate with bone density in postmenopausal women.(21,37) Our findings in postmenopausal users of estrogen-replacement therapy should be interpreted with caution because of the small number of current postmenopausal hormone users (n = 104). The reason for the inverse association of bone density with leptin concentration in such women is unclear because hormone-replacement therapy has been found to increase(38) or be unrelated to(39–40) leptin levels.
Our study had other limitations. In NHANES III, BMD was measured only at the proximal femur. We cannot exclude the possibility of an effect of leptin concentration on bone density at nonweight-bearing sites such as the spine or radius. As with all cross-sectional studies, leptin was measured at one point in time, and bone mass is accumulated and lost over a lifetime. Also, because of the cross-sectional design, we limited our study to BMD, rather than fractures that might have occurred many years before leptin measurement. In addition, approximately one-third of participants had to be excluded because they were not examined in the morning after an appropriate fast or had missing data. Finally, the correlation of leptin concentration and BMI makes their independent effects on bone density more difficult to determine.
In conclusion, this study did not find a clear-cut association of leptin with BMD. Importantly, leptin did not explain the association of BMI with bone density. We did find an inverse association of leptin with bone density in younger men. Because younger men have a lower risk of fracture, an effect of leptin on bone density does not appear to have public health significance. Whether there is any biological effect of leptin on bone density in adult humans requires additional study.
This work was supported by a contract from the National Institute of Diabetes and Digestive and Kidney Diseases (NO1-DK-6-2220).
- 11983 Risk factors for spinal osteoporosis in men. Am J Med 75:977–983., , ,
- 21986 Involutional osteoporosis. N Engl J Med 314:1676–1686.,
- 31989 Epidemiology of vertebral fractures in women. Am J Epidemiol 129:1000–1011., , , , ,
- 41985 Regulation of bone mass by mechanical strain magnitude. Calcif Tissue Int 37:411–417.,
- 51980 Relationship of fasting urinary calcium to circulating estrogen and body weight in postmenopausal women. J Clin Endocrinol Metab 50:70–75., , , , , ,
- 61996 Obesity as a protective factor for postmenopausal osteoporosis. Int J Obes Relat Metab Disord 20:1027–1032., , , , ,
- 71993 Circulating insulin levels are related to bone density in normal postmenopausal women. Am J Physiol 265:E655–E659., , , ,
- 81996 Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med 334:292–295., , , , , , , , , ,
- 92001 Leptin concentrations in the United States: Relations with demographic and anthropometric measures. Am J Clin Nutr 74:295–301.,
- 102000 Human leptin: From an adipocyte hormone to an endocrine mediator. Eur J Endocrinol 143:293–311., ,
- 112000 Leptin inhibits bone formation through a hypothalamic relay: A central control of bone mass. Cell 100:197–207., , , , , , , , ,
- 122001 Leptin reduces ovariectomy-induced bone loss in rats. Endocrinology 142:3546–3553., , , , , , ,
- 131998 Does leptin have an effect on bone in adult women? Calcif Tissue Int 63:453–455., , , ,
- 141998 Plasma leptin values in relation to bone mass and density and to dynamic biochemical markers of bone resorption and formation in postmenopausal women. Calcif Tissue Int 63:456–458.,
- 152000 Plasma leptin concentrations in postmenopausal women with osteoporosis. Eur J Endocrinol 142:170–173., , , , , , ,
- 162001 Influence of insulin-like growth factor-1 and leptin on bone mass in healthy postmenopausal women. Bone 28:113–117., , , , ,
- 172000 Relationships between serum leptin level and regional bone mineral density, bone metabolic markers in healthy women. Acta Obstet Gynecol Scand 79:1060–1064., , , , ,
- 182001 Serum leptin levels are associated with bone mass in nonobese women. J Clin Endocrinol Metab 86:1884–1887., , , , , ,
- 192001 Plasma leptin concentrations are associated with bone mineral density and the presence of vertebral fractures in postmenopausal women. Clin Endocrinol (Oxf) 55:341–347., , , , , , , ,
- 202001 Association between serum leptin concentrations and bone mineral density, and biochemical markers of bone turnover in adult men. J Clin Endocrinol Metab 86:5273–5276., , , , , , , , ,
- 212001 Role of serum leptin, insulin, and estrogen levels as potential mediators of the relationship between fat mass and bone mineral density in men versus women. Bone 29:114–120., , , , , ,
- 22National Center for Health Statistics 1994 Plan and operation of the Third National Health and Nutrition Examination Survey, 1988–94. Vital Health Stat 1 32:1–34.
- 23National Center for Health Statistics 1996 National Health and Nutrition Examination Survey (NHANES III) Densitometry Manual. Westat, Inc. [on CD-ROM] NCHS. NHANES III Reference Manuals and Reports.
- 241994 Quality control of bone densitometry in a national health survey (NHANES III) using three mobile examination centers. J Bone Miner Res 9:951–960., , , , ,
- 251996 Radioimmunoassay of leptin in human plasma. Clin Chem 42:942–946., , , , ,
- 261999 Birth weight, adulthood BMI, and subsequent weight gain in relation to leptin levels in Swedish women. Obes Res 7:150–154., , , , , ,
- 271993 Compendium of physical activities: Classification of energy costs of human physical activities. Med Sci Sports Exerc 25:71–80., , , , ,
- 281991 Caffeine, moderate alcohol intake, and risk of fractures of the hip and forearm in middle-aged women. Am J Clin Nutr 54:157–163., , , , ,
- 29GunterEW, LewisBG, KoncikowskiSM (eds.) 1996 NHANES III Reference Manuals and Reports: Laboratory Procedures Used for the Third National Health and Nutrition Examination Survey (NHANES III), 1988–1994 [CD-ROM]. National Center for Health Statistics, Hyattsville, MD, USA.
- 301996 SUDAAN User's Manual Release 7.0. Research Triangle Institute, Research Triangle, NC, USA, pp. 9-1–9-19., ,
- 312000 Leptin and bone: Does the brain control bone biology? Nutr Rev 58:209–211.
- 322001 Leptin and other components of the Metabolic Syndrome in Mauritius—a factor analysis. Int J Obes Relat Metab Disord 25:126–131., , , , , ,
- 331996 Does hyperinsulinemia preserve bone? Diabetes Care 19:1388–1392.,
- 341997 Longitudinal changes in testosterone, luteinizing hormone, and follicle-stimulating hormone in healthy older men. Metabolism 46:410–413., , , , , , , ,
- 352000 Relationship of serum leptin levels with body composition and sex steroid and insulin levels in men and women. Metabolism 49:1278–1284., , , , , ,
- 362000 Androgen deficiency in aging men: Role of testosterone replacement therapy. J Lab Clin Med 135:370–378.,
- 371998 Relationship of serum sex steroid levels and bone turnover markers with bone mineral density in men and women: A key role for bioavailable estrogen. J Clin Endocrinol Metab 83:2266–2274., , , , ,
- 381999 Effects of short-term hormone replacement on serum leptin levels in postmenopausal women. Clin Endocrinol (Oxf) 51:415–422., , , , ,
- 391999 Hormone replacement therapy affects body composition and leptin differently in obese and non-obese postmenopausal women. J Endocrinol 163:55–62., , , ,
- 402000 Serum leptin levels and body weight in postmenopausal women under transdermal hormone replacement therapy. Eur J Med Res 5:63–66., , , ,