Despite both densitometers being highly concordant in a single BMD measurement, discordance in the assessment of BMD changes between the Prodigy and DPX densitometers was observed.

stratified patients based on the baseline weight and compared RMS: For femoral neck BMD the root mean square (RMS) residual percentage change was statistically significantly larger in the high BMI group (P < 0.05) but all other bone density variables showed no significant difference. With patients allocated to two groups according to their absolute percentage change in weight (< 5% and > 5%, respectively) the RMS residual percentage changes in the bone density variables were statistically significantly larger in the large weight change group for femoral neck BMD (P < 0.05) and for standard and enhanced total body BMC (P < 0.01 and P < 0.05, respectively). With the exception of the standard total body algorithm, weight change in a longitudinal study of postmenopausal women was not found to cause systematic errors in the results of DXA studies but may adversely affect precision.

Short-term BMD variability is constant in a wide range of BMD values. Consequently, to determine cut-off values for the smallest detectable differences in BMD at the individual level, precision errors should be based on SD (expressed in absolute units) rather than on CV (expressed in percentage).

Should be checked in more detail. Although it repeatedly uses the word precision, it seems it 's actually analyzed factors affecting BMD change (e.g. OLS regression) not SD

BMD (change?) affected by baseline BMI and change in BMI

SD of change was not correlated with baseline BMD