Precision of pQCT-measured total, trabecular and cortical bone area, content, density and estimated bone strength in children

OBJECTIVES To define pQCT precision errors, least-significant-changes, and identify associated factors for bone outcomes at the radius and tibia in children. METHODS We obtained duplicate radius and tibia pQCT scans from 35 children (8-14yrs). We report root-mean-squared coefficient of variation (CV%RMS) and 95% limits-of-agreement to characterize repeatability across scan quality and least-significant-changes for bone outcomes at distal (total and trabecular area, content and density; and compressive bone strength) and shaft sites (total area and content; cortical area content, density and thickness; and torsional bone strength). We used Spearman's rho to identify associations between CV% and time between measurements, child's age or anthropometrics. RESULTS After excluding unanalyzable scans (6-10% of scans per bone site), CV%RMS ranged from 4% (total density) to 19% (trabecular content) at the distal radius, 4% (cortical content) to 8% (cortical thickness) at the radius shaft, 2% (total density) to 14% (trabecular content) at the distal tibia and from 2% (cortical content) to 6% (bone strength) at the tibia shaft. Precision errors were within 95% limits-of-agreement across scan quality. Age was associated (rho -0.4 to -0.5, p⟨0.05) with CV% at the tibia. CONCLUSION Bone density outcomes and cortical bone properties appeared most precise (CV%RMS⟨5%) in children.