The use of different dual X-ray absorptiometry brands in a multicenter clinical trial: consequences and limits.

Accurate cross-calibration (CC) and quality control (QC) programs for dual X-ray absorptiometry (DXA) instruments are necessary in order to guarantee appropriate measurements of bone mineral density (BMD) during longitudinal studies. This article details the CC-QC program established for the STRATOS study, a multicenter clinical trial investigating the effects of strontium ranelate on osteoporotic women with vertebral fractures. Forty-five DXA instruments of different brands (namely, 27 Hologic, 9 Lunar, 5 Norland, and 4 Sopha) were cross-calibrated at the beginning of the study. Twenty-seven of these were still in use by the end of the study. The CC was performed at the beginning and at the end of the study by measuring a unique spine phantom 20 times. The in vitro reproducibility of measurements. (coefficient of variation [CV]) was calculated from the results of the phantom measurements. The in vivo CV was obtained from pairs of measurements of the lumbar spine and the hip of the patients at the time of inclusion in the study. Initial in vitro CV averaged 0.5%. At the end of the study, the CC performed for the 27 apparatus in use at the end of the trial provided long-term intrabrand in vitro CV of 0.7% for the Hologic (n = 18), 1% for the Lunar (n = 5), and 0.3% for the Norland (n = 4) DXA instruments. The in vivo short-term CV for the lumbar spine BMD measurements was suboptimal, as opposed to the hip measurements, and was most likely due to the age of the population investigated. The results of measurements of multibrand DXA apparatus in this multicenter study suggest several practical conclusions: (1) the CC should be performed by using a single phantom independent of the DXA brand tested; (2) duplicate measurements should be performed at the time of patient inclusion; (3) the most efficient QC program should include CC, central reading of in vivo scans, and central review of daily QC.