Precision and accuracy of in vivo bone mineral measurement in rats using dual-energy X-ray absorptiometry

The aim of this study was to evaluate the precision and accuracy of dual-energy X-ray absorptiometry (DXA) for measuring bone mineral content at different sites of the skeleton in rats. In vitro the reproducibility error was very small (<1%), but in vivo the intra-observer variability ranged from 0.9% to 6.0%. Several factors have been shown to affect in vivo reproducibility: the reproducibility was better when the results were expressed as bone mineral density (BMD) rather than bone mineral content (BMC), intra-observer variability was better than the inter-observer variability, and a higher error was observed for the tibia compared with that for vertebrae and femur. The accuracy of measurement at the femur and tibia was assessed by comparing the values with ash weight and with biochemically determined calcium content. The correlation coefficients (R) between the in vitro BMC and the dry weight or the calcium content were higher than 0.99 for both the femur and the tibia. SEE ranged between 0.0 g (ash weight) and 2.0 mg (Ca content). Using in vitro BMC, ash weight could be estimated with an accuracy error close to 0 and calcium content with an error ranging between 0.82% and 6.80%. TheR values obtained between the in vivo and in vitro BMC were 0.98 and 0.97 respectively for femur and tibia, with SEE of 0.04 and 0.02 g respectively. In conclusion, the in vivo precision of the technique was found to be too low. To be of practical use it is important in the design of experimentation to try to reduce the measurement error. This can be achieved by performing measurements in the same position, by repeating measurements several times and by using the mean values of several BMD calculations performed by the same observer on each BMD measurement. Furthermore, better reproducibility can be obtained on the vertebra or the femur than on the tibia.