Lean mass predicts the size-dependent and size-adjusted BMD better than fat mass in premenopausal females : a cross-sectional analysis

Methods: A group of 128 healthy premenopausal females, aged 25-50 years was selected randomly from the local community, stratified according to their BMI. Lean mass (LM), fat mass (FM), total body bone mineral density (TBBMD) and total body bone mineral content (TBBMC) were measured using a DXA scan. TBBMD was divided by the standing height to estimate height-adjusted total body BMD (TBBMD-ht). Furthermore, TBBMD was divided by the bone area and height to estimate bone mineral apparent density (BMAD).


Introduction
2][3] While an isolated abnormality in a given body compartment has defined clinical implications, co-existence of an abnormality in another body compartment may lead to an interaction between them and could possibly modify the expected clinical outcome.While reduced BMC/BMD increases the fracture risk, co-existent sarcopenia may enhance the tendency for recurrent falls resulting in greater fracture risk. 4Similarly, if low BMD/BMC is co-existent with high fat content, fat tissue can, to some extent, offset the fracture risk following a fall as fat tissue would absorb the excess energy generated at the site of impact. 5fferent methods are used to measure body composition and Dual-Energy X-ray Absorptiometry (DXA) is a safe and non-invasive method with a high measurement precision. 6,7Other methods available to measure body composition include CT and MRI technologies.
There is an uncertainity regarding the relative contribution made by lean and fat tissues to the variation of BMD.Previous studies have shown conflicting results where some studies found lean mass to be the main determinent of BMD 8,10 while others found fat mass to be more predictive of BMD than lean mass. 11,12The exact mechanisms that explain the association between different body compartments FM.Magnitude and directions of the correlations between bone mass and LM were different when BMC was adjusted for bone size and height of subjects.These results support the idea that lean mass has dual effects on bone tissue where it influences both linear measurements and mineral content to different extents.

Key words: premenopausal, bone mineral density, bone mineral content, bone mineral apparent density
Vol. 45, Nos. 1 & 2, 2014   Lean mass predicts the size-dependent and size-adjusted BMD are not well known.While physical activity is the most plausible explanation, 13 other possibilities such as vitamin D 14 and simultaneous programming of different body compartments during the early period of life 15 have also been postulated.
Most of the previous studies on this subject have been done in Western countries and are based on white Caucasian populations.The direct applicability of such results to Asians is questionable as the factors that are relevant to body composition; genetic background, dietary habits and physical activity, have an ethnic variation. 16Furthermore, most of the previous studies have used either a real BMD or BMC as surrogate markers of bone mass.BMC and a real BMD are not adjusted for body size (body sizedependent variables) hence body measurements such as height and weight could have influenced the previous data and may have led to spurious associations.Recent studies have made an attempt to adjust BMD and BMC to accommodate body size to make these indices more representative of volumetric BMD. 17 However, not many studies have used these body sizeadjusted bone measurements in analyzing the associations between bone mass and other body tissue compartments.
This study examined the association of different body compartments in order to find out the major determinant of bone mass and used both sizedependent and size-adjusted bone mass measures in the analysis.

Methods
This cross-sectional study included 128 healthy pre-menopausal females aged 25 to 50 years selected randomly from four PHM areas of the Bope Poddala MOH division (32 subjects from each area).Consenting subjects were stratified into four groups (32 in each) according to their body mass index (BMI) (BMI=18-22.5, 22.6-25, 25.1-30, >30kg/m 2 ).Subjects who were postmenopausal (period of amenorrhoea >6 months without being pregnant or other valid reason), pregnant or breast feeding were not included.Futhermore, subjects who were on long-term medications such as oral hypoglycaemics, anti hypertensives, lipid lowering medications, gluco-corticoids, thyroxine, hormonal contraceptives or hormone-replacement therapy were excluded.Subjects with the history of hypertension, diabetes, hyperlipidaema, ischemic heart disease, heart failure, liver diseases, renal diseases, endocrine diseases, metabolic bone diseases or epilepsy were not included in the study.
Body weight and height were measured while wearing light indoor clothes and without shoes (Weigh Master International; Japan, model BW-110 H).The waist and hip circumferences were measured according to the standard protocols.
Lean mass (LM), fat mass (FM), total body bone mineral density (TBBMD) and total body bone mineral content (TBBMC) were measured using DXA (Hologic Discovery QDR 4500 A, Hologic, Bedford,MA).The area of the bone image was also obtained from the Region of Interest in DXA results.All analyses were performed adhering to the instructions given by the manufacturer and the same technician performed all scans.The in-vitro precision of the machine was checked on each scanning day by scanning the two phantoms provided by the manufacturer.In-vivo precision of the machine has been published previously. 18atistical analysis; TBBMD was divided by the standing height to estimate height-adjusted total body BMD (TBBMD-ht).Furthermore, TBBMD was divided by the area and height to estimate bone mineral apparent density (BMAD). 17Absolute FM and LM measured in kg were divided by body weight to estimate percentage fat and lean masses.Correlations of lean and fat contents with TBBMC, TBBMD, TBBMD-ht and BMAD were examined by Pearson correlations.Linear regression analyses were done for TBBMD and TBBMC (as dependent variables) to detect the predictive value of lean and fat contents.Bone mineral content; BMC (g) = bone mineral content Bone mineral density; BMD (g/cm 2 ) = BMC/ area Height adjusted bone mineral density; BMD-ht (g/cm 3 ) = BMD/height Bone mineral apparent density; BMAD (g/cm 3 ) = BMD/ ht/area

Results
Descriptive data related to body composition and anthropometry of the selected 128 women are given in Table 1.None of them had ever smoked or taken alcohol.
Height, weight and BMI showed positive correlations with BMD/BMC, fat and lean masses (Table 2).Both fat and lean masses showed significant positive correlations with TBBMC and bone area.LM, but not the FM showed a significant correlation with TBBMD.When TBBMD was adjusted for height, the correlation with LM became less strong but still remained statistically significant.When TBBMD was further adjusted to BMAD, the correlation with LM became negative (Table 3).
Regression analysis supported our observations described above.A unit (1kg) change in LM was associated with greater regression coefficients than a unit (1kg) change in FM for all size-dependent and size-adjusted bone mass measurements.Similar to the observations made in correlation analysis, adjusting BMD for height nullified the association with LM and

Table 2. Correlations between anthropometric measures and body composition of 128 premenopausal women
further adjusting reversed the direction of association (Table 4).
When the regression models were fitted with LM, FM and percentages of lean and fat masses as independent variables and either TBBMC or TBBMD as the dependent variable and weaker variables were excluded in forward stepwise manner, LM remained the strongest predictor of both BMD and BMC (data not shown).

Discussion
This study shows the strengths and directions of associations between lean, fat and bone tissue among healthy premenopausal women in our study sample.While both FM and LM showed positive correlations TBBMC, LM showed stronger associations with TBBMD, TBBMD-ht and BMAD indicating that LM has more influence than FM on bone tissue of these young healthy women.LM showed stronger correlations with both bone area and TBBMC but the correlation between LM and TBBMD was comparatively less.The correlation reduced further when TBBMD was adjusted for height indicating that as the association between LM on TBBMD is partly explained by its effect on Vol.45, Nos. 1 & 2, 2014   Lean mass predicts the size-dependent and size-adjusted BMD height.In support of this, we observed a correlation of 0.46 (p<0.001) between LM and height among our subjects (Table 2).Interestingly, when TBBMD was adjusted for height and bone area, correlation became inversed but remained significant.
Our observations are keeping with the previous analysis involving a group of healthy young female volunteers from Southern part of Sri Lanka in whom LM showed stronger associations than FM with both BMD and BMC in different skeletal sites. 10However, only the size-dependent BMD measurements were used in that study, hence the mechanism of the associations witnessed was not clear.Our data are comparable with those observed by Zagarins et al in a similar analysis where they used both size-dependent and size-adjusted BMD variables. 17In this analysis, LM showed positive correlations, stronger than FM with BMC, BMD and BMD-ht.Similar to our data LM showed a negative correlation with BMAD in this study.
Our results possibly explain the influence of LM on different aspects of the bone tissue.It is well known that LM has a positive influence on BMC due to its tensile strains on the skeleton.Blain et al in 2001 found that 30% variation of femoral neck BMD is related to the strength of quadriceps muscles in healthy elderly women. 19Similarly, Calbet et al in 2001 found higher BMC, BMD and muscle mass among pre-pubertal football players when compared with matching controls. 20Furthermore, skeletal tissue increases cross-sectional bone area both in the growing and adult skeleton. 21,22Janz et al in 2007 reported a positive association between femoral neck cross-sectional area and moderate/vigorous physical activities in children and part of this association was attributable to the variation in lean mass content. 21Similar observations were made by Daly and others who observed most measurements of muscle functions to correlate, positively, with BMD, BMC and bone geometry of prepubertal subjects. 22These studies indicate that LM influences many facets of human bone tissue such as bone length, bone width and the amount of bone material.
The diverse associations witnessed between LM and different expressions of bone tissue need an explanation.This could partly be a technical phenomenon as described in the following model which incorporates amount of bone material and bone dimensions; two measurements we found related to LM (Table 5).As illustrated in model 1 in column 2, if a bone cube of 1cmx1cmx1cm contains 1g of bone mineral, all measurements such as BMC, BMD, BMDht and BMAD will be equal to 1.If LM has an effect only confined to bone area, we would consider the example illustrated in model 2 where bone linear measurements have increased but the amount of material has remained the same.Here, BMC would increase by 8 fold, but BMD would only double.However, BMD-ht would remain same as in model 1 but BMAD would decline.Our data would not fit this model as we observed a positive, relatively small but statistically significant correlation between LM and BMD-ht.If LM has a dual effect both on linear measurements and material content of bone tissue, we could consider the illustration shown in the model 3 where in addition to linear measurements, the amount of bone material also has increased.Here, BMC and bone area have the greatest increase followed by BMD with a relatively less increase.BMD-ht will also increase marginally but BMAD will decline from the original value.This appears the most likely explanation for the associations observed between LM and different expressions of the bone mass in this analysis.
The knowledge that LM has a dual action on bone tissue has clinical relevance, as higher lean mass will associate with more bone material as well as wider and longer bones.While BMD and BMC are proven indices of bone strength, cross-sectional area and bone width are also known to influence bone strength. 23Wider bones have a wider width and a higher cross-sectional area (cortical bone equivalent of the cross-section of As a limitation of the study, we did not measure the volumetric BMD directly but estimated BMAD as a surrogate.However we selected our subjects from the community on random basis and all DXA analyses were done by a single technician.These would enhance the validity of our findings.
In summary, we demonstrate significant relationships between the three major body components; lean, fat and bone mass in this group of healthy premenopausal women.Although both FM and LM showed positive correlations with measurements of bone mass, the association between LM and bone mass was stronger than that between FM and bone mass.The strength and directions of associations were different when BMC was adjusted for bone size and the height of subjects.We also illustrated the possibility that LM has dual effects on bone tissue where it influences both linear measurements and bone mineral content to different extents.

Table 1 . Descriptive data of 128 premenopausal women included in the analysis
Given values in cells are Pearson correlation coefficients and p values (within brackets)TBBMC; total body bone mineral content, TMBMD; total body bone mineral density, BMD-ht; height adjusted bone mineral density, BMAD; bone mineral apparent density

Table 3 . Correlatons of lean and fat measurements with size-dependent and size-adjusted bone mineral density measurements among 128 women included in the analysis
Given values in cells are Pearson correlation coefficients and p values (within brackets), TBBMD= total body bone mineral density, TBBMC= total body bone mineral content.

Table 5 . The hypothetical model based on linear measurements and material properties of bone tissue used to explain the association between lean mass and different measurements of bone tissue
, therefore, less vulnerable to buckling.Furthermore, this could explain the lower incidence of fractures seen among women of Afro-Caribbean descent and anti-fracture benefits seen in interventions based on physical activities.