ABSTRACT: Recent evidence suggests that adipose tissue may negatively impact bone health, challenging the traditional paradigm that increased fat mass, through mechanical loading or endogenous estrogen production, is beneficial to the skeleton. We hypothesized that it is primarily the visceral compartment of body fat that is detrimental to bone metabolism, resulting in impaired bone density and architecture. In an age-stratified population sample of 218 women and 291 men (age 20-97 years), we assessed visceral (VAT) and subcutaneous (SAT) adipose tissue areas at the L2-L3 interspace level by single slice quantitative computed tomography (QCT) and measured total body fat mass (TBF) by dual-energy X-ray absorptiometry. We then correlated these findings with volumetric bone mineral density (vBMD) at the femoral neck (FN) and lumbar spine (LS) assessed by central QCT, and with vBMD and microstructural parameters at the ultradistal radius (UDR) by high resolution peripheral QCT (HRpQCT). In unadjusted analyses in postmenopausal women, TBF and SAT were positively correlated with total, trabecular, and cortical vBMD at the FN, LS, and UDR and with trabecular microstructure at the UDR. By contrast, VAT was not correlated with vBMD at the FN or LS but was positively correlated with UDR total and trabecular vBMD but not cortical vBMD. Adjustment for age or for bioavailable estradiol and testosterone levels reduced these correlations, while adjustment for body weight eliminated most positive associations. Assessment of the VAT/SAT ratio, however, demonstrated a negative relationship with vBMD at the FN and LS in postmenopausal women, a relationship eliminated when adjusted for age. Correlations between skeletal parameters and adipose measurements in pre-menopausal women and older men were weaker and mostly non-significant. In younger men, VAT was negatively associated with vBMD, cortical thickness, and trabecular microstructure at the UDR, and with LS vBMD and FN cortical vBMD. These associations generally remained after adjustment, with some negative associations (e.g., UDR cortical area) being accentuated. Similar results were found when the VAT/SAT ratio was correlated with FN vBMD in younger men; in contrast, VAT/SAT was positively correlated with FN vBMD in older men and this relationship was strengthened by age-adjustment. Together, our data suggest that adiposity has associations with bone that are age-, gender-, menopausal status-, adipose depot-, and bone compartment-specific. These novel observations warrant further investigations to establish any causal relationships.
