Short-term in vivo precision of BMD and parameters of trabecular architecture at the distal forearm and tibia

Summary In vivo hr-pQCT precision was determined in 42 postmenopausal women using double baseline measurements from a multicenter trial of odanacatib. Errors, e.g., at the radius below 1.3% for BMD and below 6.3% for trabecular structure, were comparable to single-center results. Motion artifacts remain a challenge, particularly at the forearm. Introduction The short-term in vivo precision of BMD, trabecular bone structure, cortical thickness and porosity of the forearm and tibia was measured by hr-pQCT. Also the effect of image quality on precision was evaluated. Methods In 42 postmenopausal women (age 64.4 ± 6.8 years) out of 214 subjects enrolled in a multi center advanced imaging phase III study of odanacatib (DXA spine or hip T-scores between −1.5 and −3.5), double baseline hr-pQCT (XtremeCT) measurements with repositioning were performed. The standard ultradistal location and a second, more proximally located VOI were measured at the radius and tibia to better assess cortical thickness and porosity. Image analysis and quality grading (grades: perfect, slight artifacts, pronounced artifacts, unacceptable) were performed centrally. Results At the radius RMS%CV values varied from 0.7% to 1.3% for BMD and BV/TV and from 5.6% to 6.3% for Tb.Sp, Tb.Th, Tb.N, and cortical porosity. Numerically at the tibia, precision errors were approx. 0.5% lower for BMD and 1% to 2% lower for structural parameters although most differences were insignificant. In the radius but not in the tibia, precision errors for cortical thickness were smaller at the distal compared to the ultradistal location (1% versus 2%). Conclusions BMD precision errors were lower than those for trabecular architecture and cortical porosity. Motion artifacts remain a challenge, particularly at the forearm. Quality grading remains subjective, and more objective evaluation methods are needed. Precision in the context of a multicenter clinical trial, with centralized training and scan analysis, was comparable to single-center results previously reported.