The authors state that they have no conflicts of interest.
We studied the relation of leptin to bone, bone loss, and bone turnover in community-dwelling men and women. Leptin predicted higher BMD and lower bone turnover only in women. Leptin was not associated with 4-year bone loss in either sex.
Introduction: Leptin, the protein product of the obesity (OB) gene produced in fat tissue, was originally thought to be involved only in the regulation of food intake and energy balance. Recent evidence suggests that leptin may play a role in the pathophysiology of several chronic diseases. Studies of the association between leptin and bone have been numerous yet inconclusive. Only one previous longitudinal study has been reported, which showed no association of leptin with BMD after adjusting for body size.
Materials and Methods: We report the association of serum leptin with BMD at the hip, spine, and midshaft radius in community-dwelling men (n = 498) and nonestrogen-using postmenopausal women (n = 411) 45–92 years of age. Serum leptin was measured in blood obtained between 1984 and 1987. Between 1988 and 1991, BMD was measured at the midshaft radius by single photon absorptiometry and at the femoral neck, total hip, and lumbar spine by DXA; at the same visit, height, weight, and body fat (by bioelectrical impedance analysis) were measured, and bone resorption was assessed in a subset of men (n = 286) and women (n = 241) using urine N-telopeptide (NTX). Four years later, axial BMD was remeasured in 536 participants. Sex-specific associations of leptin with BMD, NTX, and bone loss were tested using regression analysis.
Results: In unadjusted analyses, leptin was associated with BMD at the femoral neck, total hip, lumbar spine, and midshaft radius in both sexes (p < 0.01). In multiple regression analyses, adjusted for age, BMI, and other bone-related factors, only women showed a graded stepwise positive association between serum leptin and BMD at all sites and a negative stepwise association with NTX (all p for trend < 0.01). Baseline leptin levels did not predict 4-year bone loss in either sex.
Conclusions: A favorable dose-dependent leptin–BMD association unexplained by obesity was observed only in women. The reason for the sex difference is unknown.
Leptin is a hormone produced by adipocytes. It was originally thought that leptin's target was the central nervous system where it regulates energy balance. However, leptin receptors have been identified in many peripheral tissues including the skeleton.(1)
Several lines of evidence suggest a relationship between leptin and bone health,(2) but published results are contradictory. Some studies found that leptin was associated with greater BMD,(3–5) reduced risk of fracture,(4,6) and decreased bone resorption,(5–8) whereas others reported a negative association with BMD(9–14) and bone formation,(14) or no association with BMD,(7,8,15–23) bone loss,(24) or bone turnover.(15,17,18,24)
Body size is positively associated with both leptin(25) and BMD.(24,26) Obese people have relatively low bone loss and a decreased risk of osteoporosis.(27) The bone-sparing effects of weight may be caused by increased loading on the skeleton,(28) increased production of estrogen in adipose tissue,(29) increased bone formation caused by the anabolic effects of high levels of insulin,(30) and/or high levels of bone-related hormones, such as leptin.(20)
Differences in the study populations, measures of body size, and bone sites assessed may contribute to the inconsistency among these studies. In addition, small sample sizes and lack of adjustment for bone-related covariates, including calcium intake and hormone therapy, may also contribute to the discrepancies between studies of women. Many studies did not adjust for body size; most studies that found no association between leptin and BMD adjusted for body mass index (BMI),(8,16,17,19,21,22,24,31) fat mass,(7,15,18) or weight.(32) The majority of studies were in women.
We report here the association between serum leptin and BMD, bone resorption, and bone loss in 909 community-dwelling men and postmenopausal women not using estrogen therapy. Current hormone users were excluded to avoid the confounding favorable effects of estrogen on bone.
MATERIALS AND METHODS
The cohort consists of older, ambulatory, community-dwelling white men and women who were participants in the Rancho Bernardo Study, established in 1972. In 1984–1987, fasting blood samples were obtained and frozen for later measurement of biomarkers, including leptin. Four years later, between 1988 and 1991, BMD was first measured, and a morning urine sample was obtained to measure urine N-telopeptide (NTX) from a subset of fasting subjects. Approximately 4 years after the initial BMD measurement, a subset of men and women returned for a second BMD measurement.
During the 1988–1991 visit, information on smoking habits, alcohol intake, exercise frequency, reproductive history, and use of vitamins, thiazides, thyroid hormones, and estrogen therapy was obtained using standard questionnaires. Participants brought their pills and prescriptions to the study center, where a specially trained nurse confirmed current supplement and medication use. Dietary calcium intake was obtained using the self-administered, semiquantitative Harvard-Willett Diet Assessment Questionnaire.(33) Reported foods were converted into nutrients using the food composition database of the U.S. Department of Agriculture. Total calcium intake was calculated by adding dietary calcium intake and calcium from supplements.
Measurement of serum leptin
Serum samples obtained in 1984–1987 were stored frozen at −70°C. Leptin was measured by radioimmunoassay (RIA) at Linco Diagnostics Laboratory in 2004. The leptin assay sensitivity was 0.5 ng/ml; the interassay and intra-assay CVs were 4.7% and 3.9%, respectively.
Measurement of insulin
Fasting insulin levels from plasma samples obtained in 1984–1987 and stored at −70°C were determined for 354 men and 254 women by double-antibody RIA method in the diabetes laboratory of JM Olefsky(34) in 1996. The CV for insulin was 8.5%.
Measurement of bone resorption
Urine samples were collected in 1988–1991 and stored at −70°C until analysis. NTX levels were measured in 286 men and 241 women who provided morning fasting urine samples using an ELISA (Ostex International) in 1996. For normalization of urinary NTX, urine creatinine was measured by an enzymatic assay (Ostex International). The assay was calibrated using standard amounts of human bone collagen digested with bacterial collagenase, and the results were expressed as nanomoles of bone collagen equivalents per millimoles of creatinine (nmol BCE/mmol CR).(35)
In 1988–1991, BMD was measured at the midshaft radius of the nondominant arm using single photon absorptiometry (SPA; Lunar model SP2B; Lunar Corp.) and at the hip (femoral neck, total hip) and lumbar spine (mean L1–L4) using DXA (Hologic QDR model 1000; Hologic). Between 1992 and 1996, axial BMD measurements were repeated using the same DXA machine. The machines were calibrated daily, with measurement precisions of ≤1% for the spine, ≤1.5% for the hip, and ≤5.0% for the radius. The BMD of the total hip was calculated as the total BMD of the greater trochanter, femoral neck, and intertrochanteric regions. The midshaft radius was the mean of four scanned lines.
Height and weight were measured in 1984–1987 and again in 1988–1991 with participants wearing light clothing and no shoes. BMI was calculated as weight (kg) divided by height (m)2. Total body fat was assessed by bioelectrical impedance analysis (BIA) using a body composition analyzer (Model 1990B; Valhalla Scientific, San Diego, CA, USA).(36)
Statistical analysis was performed using SPSS (SPSS for Windows 11.5; SPSS) and SAS statistical software packages (version 8.2; SAS Institute). In preliminary analyses, we tested for interactions between leptin and body size (BMI, fat mass, weight) and leptin and sex related to BMD. There were substantial sex differences in the association between leptin and BMD at the midshaft radius and with NTX. Therefore, data from men and women were analyzed separately in the final models. Independent t-tests were performed to test for differences between men and women. Pearson correlations were used to assess correlations between BMI and weight at the 1984–1987 visit (measurement of leptin) and 1988–1991 visit (measurement of BMD). Univariate, age-adjusted, age- and BMI-adjusted, and multiple regression analyses, adjusted for standard osteoporosis covariates including age, BMI, alcohol intake, total calcium intake, exercise status, smoking history, and use of common medications that impact bone (thiazides, thyroid hormones), were used to test the association of BMD at the femoral neck, total hip, lumbar spine, and midshaft radius and of NTX with serum leptin. We also tested for linearity by including leptin quintiles in the regression model. In additional analyses, insulin was included in the regression models. In the analyses of bone loss, the same regression models were used to test the relationship of annual percentage change in BMD over an average of 4 years at the femoral neck, total hip, and lumbar spine with serum leptin. Paired t-tests were used to test for significant bone loss over an average of 4 years. Linear trends for BMD and NTX across leptin quartiles were tested using analysis of covariance and linear contrasts. Regression models were repeated including fat mass or weight instead of BMI, with no change in the observed associations. There was no evidence of collinearity between leptin and measures of body size on bone. Serum leptin, NTX, insulin, and total calcium intake were log-transformed to meet assumptions of statistical tests. All tests were considered significant at the 0.05 level.
Men (n = 498) and postmenopausal women (n = 411) who had measures of serum leptin and BMD were included in the study of leptin and BMD: the 299 men and 237 women who returned 4 years later for a follow-up visit when BMD was measured again were included in the analysis of BMD change.
Sex-specific cohort characteristics are shown in Table 1. Compared with women, men were younger, taller, heavier, had significantly lower leptin levels, higher fasting insulin levels, lower levels of NTX, and higher BMD at the hip, spine, and radius (all p < 0.05). Log leptin levels were not correlated with age but were strongly correlated with BMI, fat mass, and weight in men and women, respectively (BMI: r = 0.58 and 0.62; fat mass: r = 0.58 and 0.60; weight: r = 0.51 and 0.59; all p < 0.01). BMD at two or more bone sites was negatively correlated with age and positively correlated with BMI, fat mass, weight, alcohol intake, and total calcium intake in both sexes (all p < 0.01). BMD at the femoral neck and total hip was higher among men who exercised and never smoked. In both sexes, weight and BMI from the 1984–1987 visit (measurement of leptin) and the 1988–1991 visit (BMD measurement) were significantly correlated (men: weight, r = 0.924, BMI, r = 0.906; women: weight, r = 0.918, BMI, r = 0.903; all p < 0.0001).
Table Table 1.. Characteristics of Study Participants, 1988–1991
In univariate regression analysis, leptin was associated with BMD at all sites in both sexes (p < 0.01). Adjustment for age did not change the associations; however, after further adjustment for BMI, leptin remained positively associated with BMD at all four bone sites in women only (all p ≤ 0.01; Table 2), where leptin explained 4.0% of the hip BMD variance in the full model. Leptin was not associated with BMD in men after adjustment for BMI. Leptin was inversely associated with NTX, again in women only (p < 0.01; Table 3), where it contributed 8.2% of the variance. Leptin was not associated with NTX in men. Addition of other bone-related covariates to the regression models did not materially change the results. When insulin was included as a covariate in the regression models, the association between leptin and BMD did not materially change. When leptin quintiles were included in the regression models, the term was significant at all BMD sites and with NTX in women only. No significant linear associations were found in men.
Table Table 2.. Regression Models of Leptin on BMD in Men and Postmenopausal Women
Table Table 3.. Regression Models of Leptin on NTX in Men and Postmenopausal Women
Age- and BMI-adjusted site-specific BMD (Fig. 1) and NTX levels (Fig. 2) by serum leptin quartile in men and postmenopausal women are shown. In women, there was a significant stepwise increase in BMD at all bone sites (femoral neck, p for trend = 0.005; total hip, p for trend < 0.0001; lumbar spine, p for trend = 0.002; midshaft radius, p for trend < 0.0001) by increasing quartile of leptin and a significant stepwise decrease in NTX (p for trend < 0.001). No significant trends were found in men (femoral neck, p for trend = 0.39; total hip, p for trend = 0.66; lumbar spine, p for trend = 0.31; midshaft radius, p for trend = 0.65; NTX, p for trend = 0.15).
We tested for interactions between leptin and body size and found no statistically significant interactions. When we tested for a leptin–sex interaction in a model that combined both sexes, we found significant interactions for BMD at the radius (p = 0.04) and for NTX (p = 0.01). Sex–leptin interactions were not statistically significant at other bone sites.
After a mean interval of 4.04 ± 0.49 (SD) years, 299 men and 237 women had repeat DXA measurements of the hip and spine. In men and women, the annual average bone loss was statistically significant at all sites (all p ≤ 0.001) except the spine in women, and there were no significant sex differences in bone loss rates. Leptin was not a significant predictor of annual percent change in BMD at the femoral neck, total hip, or lumbar spine in either sex.
In this study, leptin (measured an average of 4 years before BMD) predicted higher BMD levels and lower NTX in a graded dose-dependent manner, independent of age and BMI, in women but not in men. These findings support the thesis that leptin plays a fundamental role in both short-term (antiresorption) and long-term (bone mass) female bone health.(1) Some other studies have supported the current findings in women.(3–5) Pasco et al.(3) found that leptin was associated with BMD at the spine and (borderline) hip independent of fat mass and weight in 214 pre- and postmenopausal women. Two studies in postmenopausal women found that leptin was associated with BMD at the femoral neck and total body in models adjusted for percent body fat(4) or fat mass.(5) Others reported that leptin was associated with decreased bone resorption in postmenopausal women after adjustment for body fat(5,8) or BMI.(8)
However, several other studies in women found that leptin was not associated with BMD or markers of bone turnover after adjusting for BMI(8,17,18,22) and/or fat mass.(7,15,18) Two studies found that leptin was inversely associated with BMD in women after controlling for insulin(13) and weight.(10) In the largest reported study, Ruhl and Everhart(19) found that increasing levels of leptin were associated with higher BMD levels in 3054 pre- and postmenopausal women from NHANES III but the association was no longer significant after adjusting for BMI. In the only previous bone loss study, Dennison et al.(24) also found that leptin, adjusted for BMI, was not correlated with 4-year BMD change in 132 older women.(24)
Fewer studies have been conducted in men. Two studies in men found an inverse association(7,11) and four reported no association.(12,14,19,24) Another study found that leptin was positively associated with total body BMD in 92 older men, but the association disappeared after BMI was added to the regression model,(22) similar to our results. In another study, leptin was negatively correlated with BMD at the lumbar spine in 80 Korean men 42–70 years of age after adjusting for BMI.(37)
Because women have 2- to 3-fold higher leptin levels than men,(20) higher levels may be necessary for beneficial leptin effects or the association in women may be more evident because of their broader range of leptin levels compared with men. Men and women had similar rates of annual bone loss over the 4 years so this did not explain baseline sex differences. Analyses combining men and women showed a statistically significant leptin–sex interaction for BMD at the radius and for NTX, supporting the observed sex differences. Previous studies that have investigated the association between leptin and bone in men and women have performed sex-specific analyses.(6,19,24) We believe the sex differences are real and unexplained.
Although leptin has been shown to exert both anti-osteogenic and anabolic effects on bone formation,(38) its exact role in bone metabolism has yet to be elucidated. It has been suggested that leptin inhibits bone formation by acting centrally on the hypothalamus.(39) Alternatively, leptin has been shown to promote osteoblastic cell growth, bone formation, and bone mineralization by acting peripherally on osteoblasts(40) and human stem cells undergoing osteogenic differentiation.(41) Leptin deficient obese mice (ob/ob) exhibit either increased(39) or reduced bone mass.(42) In ob/ob mice, central administration of leptin decreased bone mass,(39) but peripheral administration increased bone growth.(42) In vitro studies also suggest that leptin has osteogenic properties mediated by receptors in osteoblasts, osteoclasts, and chondrocytes, with resulting increased bone mineralization.(40,43,44) It has been suggested that because leptin is produced and secreted systemically, the direct peripheral positive effects may override negative central effects of leptin on bone.(41,44)
It has been hypothesized that skeletal loading is the dominant mass-related factor associated with bone(1,28) and that leptin mediates the bone-sparing effects of fat mass on the skeleton.(39) To gain a better understanding of the independent contributions of body size and leptin to bone, we included the non–weight-bearing radius as an outcome. If leptin plays a critical role in BMD independent of the loading effect of weight, leptin should contribute to BMD at a site that is minimally affected by loading.(26) We found that leptin explained 4% of the variance in hip BMD and 2.5% at the radius, independent of BMI, suggesting that leptin plays a role in bone health, independent of body fat in postmenopausal women. Smaller studies that reported no independent association between leptin and radial BMD after adjustment for BMI or fat in men(7) or women(3,4,7,16) may have been underpowered to detect these weak associations.
Serum leptin has been found to covary with several measures of body composition including BMI and fat mass.(25) It has been suggested that controlling for fat mass is more useful than controlling for BMI(1) in assessing the association between leptin and bone. We obtained similar results when using BMI, fat mass, or weight, suggesting that these different measures of body size are equivalent covariates when studying the association between leptin and bone in women. However, because leptin is associated with body size, it is possible that adjusting for anthropometric measures is, in fact, overadjusting for some other attribute in the causal pathway, and leptin could be one mechanism whereby weight preserves BMD in some older women.
One potential limitation of this study is the single leptin assay. However, it has been shown that a single morning fasting leptin, as used here, can characterize usual leptin levels for an individual within a population.(45) Also, the long duration of freezing may have affected leptin levels, but the levels observed here were similar to those reported in other studies for men and women of comparable age and body size.(3,5,8,15,17,19,21,22,24) We cannot exclude the possibility that leptin levels may have changed between the measurement of leptin and 4 years later at the time that BMD was assessed, but we can think of no reason for a differential change over time by sex. In addition, change in body size was unlikely to have confounded the association given the high correlation of BMI between visits.
The absence of an association with the 4-year change in BMD may reflect small numbers, short follow-up, or the very large effect of baseline BMD, reflecting lifetime determinants of BMD. Longer prospective studies are necessary.
This is the largest reported epidemiological study of the association between leptin, BMD, bone resorption, and bone change in older men and women. Most studies investigating the association between leptin and bone have been small, cross-sectional studies limited to women. The strength of this study is the large sample, inclusion of both sexes, and the ability to adjust for bone-related covariates including calcium intake, serum insulin, several measures of body size, and physical activity, which few other studies have done.
In conclusion, leptin had favorable associations with BMD and bone resorption (NTX) independent of body size in postmenopausal women, with no independent association in men. Understanding the mechanism for the sex difference role of leptin may provide new targets for the prevention of osteoporosis.
This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK 31801 and National Institute on Aging Grant AG 07181. LAW received an unrestricted research grant from the collaboration of Procter & Gamble Pharmaceuticals and Sanofi-Aventis Pharmaceuticals.