Bone mineral and soft-tissue changes in AIDS-associated lipoatrophy

It is well established that body weight influences bone mass, but there is considerable controversy in the literature as to whether the principal determinant is lean mass, fat mass, or combinations of both, with variable degrees of relative influence as a function of age and sex. Total body and regional tissue composition studies were performed with DXA in a cohort of 102 HIV+ males who were on highly active antiretroviral therapy. These medications may be associated with a unique peripheral lipoatrophy without commensurate loss of lean tissue mass, and thereby provide an opportunity to assess the relative influence of fat mass on BMC levels in the absence of lean mass change. A comparison of the first quartile of peripheral fat (the lowest fat mass) with the remainder of the cohort showed no significant difference in lean mass in the total body, trunk, lower limbs, and upper limbs. In contrast, a significantly lower BMC was registered in the first peripheral fat quartile for all regions, implying an association of fat mass with BMC. By quartile stratification of the regional BMC, a comparison of the first and fourth quartiles demonstrated that the percentage fat mass decrease exceeded the lean mass decrease by a factor of 1.8 to 4.5. Regional BMCs were also stratified into two groups, ≤50 percentile and >50 percentile, and analyzed by Spearman correlation and robust multiple regression. It was found that lean mass was a determinant of BMC in both groups, whereas fat mass was an independent predictor in the >50 percentile BMC group only. The BMD t-score for the total hip was significantly lower in the first quartile of lower limb fat mass than the t-score of the remaining cohort, but this difference was not significant for the BMD lumbar spine t-score. This is a reflection of the influence of preferential local peripheral lipoatrophy on the adjacent mineral content and provides further evidence of fat mass as a determinant of BMC in addition to lean mass.