Diagnostic evaluation of bone densitometric size adjustment techniques in children with and without low trauma fractures

Abstract

Several established methods are used to size adjust dual-energy X-ray absorptiometry (DXA) measurements in children. However, there is no consensus as to which method is most diagnostically accurate. All size-adjusted bone mineral density (BMD) values were more diagnostically accurate than non-size-adjusted values. The greatest odds ratio was estimated volumetric BMD for vertebral fracture. The size dependence of areal bone density (BMDa) complicates the use of DXA in children with abnormal stature. Despite several size adjustment techniques being proposed, there is no consensus as to the most appropriate size adjustment technique for estimating fracture risk in children. The aim of this study was to establish whether size adjustment techniques improve the diagnostic ability of DXA in a cohort of children with chronic diseases. DXA measurements were performed on 450 children, 181 of whom had sustained at least one low trauma fracture. Lumbar spine (L2-L4) and total body less head (TBLH) Z-scores were calculated using different size adjustment techniques, namely BMDa and volumetric BMD for age (bone mineral apparent density (BMAD)); bone mineral content (BMC) and bone area for height; BMC for bone area; BMC for lean mass (adjusted for height); and BMC for bone and body size. Unadjusted L2-L4 and TBLH BMDa were most sensitive but least specific at distinguishing children with fracture. All size adjustments reduced sensitivity but increased post-test probabilities, from a pre-test probability of 40 % to between 58 and 77 %. The greatest odds ratio for fracture was L2-L4 BMAD for a vertebral fracture and TBLH for lean body mass (LBM) (adjusted for height) for a long bone fracture with diagnostic odds ratios of 9.3 (5.8-14.9) and 6.5 (4.1-10.2), respectively. All size adjustment techniques improved the predictive ability of DXA. The most accurate method for assessing vertebral fracture was BMAD for age. The most accurate method for assessing long bone fracture was TBLH for LBM adjusted for height.