Introduction
Osteoporosis and associated fractures present a major public health challenge [
1]. Body mass is an important determinant of bone mineral density (BMD) [
2], due to the associated mechanical loading to the bone [
3], and low body weight is a well-recognised risk factor for fracture [
4]. On the other hand, obesity is associated with increased cardiometabolic risk [
5] and in middle-aged adults we reported that the positive relationships between body mass index (BMI) and BMD were weaker at high BMI (i.e. when BMI above 30-35 kg/m
2) [
6]. In addition, a longitudinal cohort study showed that older adults with baseline BMI in the obesity category had greater annual loss of femoral neck BMD compared with participants with normal BMI, indicating that high body mass is not necessarily protective against bone loss [
7].
Randomised controlled trials (RCTs) of caloric restriction have shown that intervention induced weight loss is related to a decrease in BMD [
8], but weight regain after the trial does not increase BMD [
9]. In the Women's Health Initiative study, compared with stable body weight, weight gain, weight loss, and intentional weight loss were associated with increased incidence of fracture at different locations [
10]. Regarding the components of body mass, cross-sectional studies showed that in men and pre-menopausal women, lean mass is a stronger predictor than fat mass for bone mass, whereas in postmenopausal women, the influence of lean mass and fat mass on BMD were comparable [
11‐
13]. Few studies have evaluated the longitudinal relationships between change in body composition and BMD during middle to older age. In a study with 172 men and 165 women aged over 65 years, rate of loss of lean mass in the lower limbs was positively associated with the rate of total hip BMD loss in women, although the associations were attenuated when rate of loss in leg muscle strength was included in the model (
P =0.07) [
14]. Associations between change in fat mass and BMD were not evaluated in that study. In 539 participants of the Study of Women’s Health Across the Nation, during menopause transition (defined as the four years surrounding the final menstrual period) greater loss of lean mass was associated lower femoral neck BMD and greater gain in fat mass was associated with higher femoral neck and lumbar spine BMD at the end of the transition period [
15].
In a cross-sectional analysis of participants in the Busselton Healthy Ageing Study (BHAS) at baseline, we previously reported positive relationships between lean and fat mass with BMD, which were attenuated or absent in individuals with higher BMI [
6]. With the collection of further dual-energy x-ray absorptiometry (DXA) body composition and BMD data after ~6 years follow-up, the aim of the present study was to perform a longitudinal analysis of associations between changes in total body mass, lean mass and fat mass with changes in BMD of total hip, femoral neck and lumbar spine in this representative cohort of middle-to-older aged Australians.
Discussion
In this study of a representative cohort of middle to older aged Australians, in females we observed a positive, non-linear association of ∆TM and ∆LM with ∆BMD of total hip, femoral neck and lumber spine, with a threshold level of ~+5kg for ∆TM and ~+1kg for ∆LM. In males a positive, near-linear association of ∆TM and ∆LM with ∆BMD of total hip and femoral neck, but not lumbar spine, was observed. ∆FM was positively associated with ∆total hip BMD in both sexes, with a threshold level of ~3 kg gain in fat mass, but was not significantly associated with ∆BMD of femoral neck or lumbar spine.
Our study is the first longitudinal study to show that maintenance or increase of lean mass is associated with reduction in age-related bone loss in middle to older aged adults, whereas maintenance or increase in fat mass plays a lesser role for total hip BMD and was not associated with changes in femoral neck and lumber spine BMD. Furthermore, in body mass stable participants we showed that those who had increased lean mass but decreased fat mass over time (Group 3) had better maintenance of total hip BMD in both sexes and better maintenance of femoral neck BMD in males compared with those who had decreased lean mass but increased fat mass over time (Group 2). Previous observational studies [
22,
23] and weight loss RCTs [
24] showed that a decline in body mass was associated with bone loss, but there have been limited data on the relative contributions of change in lean and fat mass. To our knowledge, the only other longitudinal study that evaluated changes in body composition and BMD is The Study of Women’s Health Across the Nation, which focused on the changes in lean and fat mass during the menopause transition period and evaluated lumbar spine and femoral neck BMD [
15]. By including both male and female participants, having a larger sample size and wider age range (baseline age 46-70 years), and evaluating total hip BMD which has better precision compared with femoral neck BMD and is the preferred site for monitoring changes in older individuals who tend to have lumbar spine artifact [
25], our study advances knowledge on the complicated association between body composition and bone health, and shows the importance of maintaining lean mass for preventing age-related bone loss.
Cross-sectional studies in young, middle-age and older adults have shown positive associations between lean mass and BMD [
11,
12,
26]. The important role of lean mass in preserving bone mass is also supported by meta-analysis of RCTs, which showed that exercise could attenuate femoral neck bone loss resulted from diet-induced weight loss [
27]. Several mechanisms could link lean body mass with bone mass, including direct mechanical loading (muscle contraction and gravitational loading) [
28], genetic factors, hormonal factors, physical inactivity and nutrition factors that related to both muscle and bone loss [
29,
30]. In addition, muscle-derived myokines can affect bone remodelling, for example irisin could improve osteoblastgenesis whereas follistatin could inhibit myostatin induced bone resorption [
31]. Our study showed that women and men in the highest quartile of ∆LM had 2.8% and 1.4% less loss of total hip BMD over 6 years, respectively, compared with those in the lowest quartile. A 2% improvement in total hip BMD is related to 28% reduction in vertebral fracture risk and 16% reduction in hip fracture risk [
32]. Thus, the reduced bone loss observed in our study could be clinically relevant and of importance for fracture prevention.
In cross-sectional studies, associations between fat mass and BMD had been reported to be positive, neutral or negative, depending on participants’ age and sex, as well as analytical methods and covariates adjusted [
11‐
13,
26,
33]. In an analysis of 10,814 participants aged 20-59 from NHANES 2011-2018, after accounting for lean mass, fat mass had a moderate, negative association with total body BMD, especially in men [
33]. Adipose tissue influences BMD directly through gravitational loading and has indirect, positive effects via endocrine mechanisms, including aromatization of androgens to estrogens, secretion of leptin by adipocytes and increased pancreatic insulin secretion in individuals with higher fat mass [
34]. However, adipose tissue can also negatively impact bone through production of inflammatory cytokines which increase bone resorption [
35]. Furthermore higher fat mass could negatively affect bone metabolism by reduced circulating 25OHD levels, increased production of parathyroid hormone [
36], and dysregulation of the growth hormone /insulin-like growth factor-1 axis [
37]. These together could account for the weak or absence of associations with ∆BMD for ∆FM in our study. In the context of the global epidemic of obesity, our findings are of importance in that gain in fat mass provides little benefits in preventing bone loss in older age.
We observed stronger associations between changes in TM, LM and FM with changes in BMD in females compared with males. Partly it could be due to that females had greater bone loss compared with males during the follow up period. The annual rate of change was -0.7% to -0.8% in women and -0.2% in men for total hip/femoral neck BMD, and -0.5% in women and 0.4% in men for lumbar spine BMD, which are comparable to those reported in other longitudinal cohort studies [
38,
39]. Another possible explanation is that estrogen receptors have been reported to have different actions on male and female skeleton during loading. In animal studies, estrogen receptor alpha (ERα) plays a critical role in the adaptive response of bone to loading in female mice [
40], but not in male mice [
41]. Interestingly, an association of change in body mass and bone loss of lumbar spine was only observed in women, which was mainly related to change in lean mass. In males neither change in total mass, nor changes in lean or fat mass associated with change in lumbar spine BMD. The increase in lumbar spine BMD over time observed in males might indicate the high prevalence of spinal osteoarthritis in older men [
42], which could influence the accuracy of lumbar spine DXA BMD assessment and obscure any potential association with change in body composition. In addition, previous studies have shown that lumbar spine BMD is less influenced by weight loss compared with total hip BMD [
43].
Our study is the first large population-based longitudinal study to evaluate the relationships between change in body composition and BMD during middle to older age that included both male and female participants. In both sexes we showed the importance of maintaining lean mass for preventing age-related bone loss. Assessment at baseline and 6 years allowed us to examine the longitudinal associations of body composition and bone density over sufficient interval to assess age-related bone loss. Our study does have several limitations. Firstly, it is observational in nature and although in the restricted cubic spline analyses we controlled for important confounding variables including lifestyle factors, it remains possible that the greater bone loss observed with greater decline body mass/lean mass was due to uncontrolled or residual confounding, as we were unable to take into account the reason for weight loss and whether it was voluntary or involuntary (i.e. due to medical reasons). Secondly, only those who were able to attend the year 6 survey were included in this study, and therefore our study population might be healthier than people of similar age. Thirdly, DXA technology uses a three-compartment model and assumes that fat free mass has the same hydration. However, it has high reproducibility and has been shown to have strong agreement for body fat % with the reference 4C Lohman model (R
2=0.93) [
44]. Lastly, we used IPAQ to assess physical activity, which although has been validated for adults [
45], does not provide information on activity type. The impact of physical activity type and change in physical activity on changes in body composition and BMD deserves further study. Furthermore, the majority of participants were white, therefore further research in other ethnic groups is needed.
In conclusion, in this longitudinal study of middle-aged and older Australians, loss of body mass is associated with increased bone loss in both males and females. Changes in lean body mass are a stronger determinant than changes in fat mass of changes in bone density. Interventions which preserve or increase lean mass in ageing adults may have long term skeletal benefits.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.