Components of variance when assessing the reproducibility of body composition measurements using bio-impedance and the Hologic QDR-2000 DXA scanner

This study evaluated the reproducibility of measuring lean tissue mass (LTM), fat mass (FMDXA), bone mineral content (BMC) and density (BMD) with a Hologic QDR-2000 DXA scanner, and both fat free mass (FFMbio) and fat mass (FMbio) with an impedance meter. Furthermore, the study aimed to assess whether sheets and pillows significantly altered the DXA results. Fifty-one healthy volunteers aged 20-61 years were examined before and after repositioning (n = 20), removal of the sheets (n = 10), addition of two foam pillows (n = 11) or a regular pillow (n = 10). One to 7 days later a re-examination was performed (n = 35). Eight surgical patients were also scanned twice 1-7 days apart. The day to day coefficients of variation were about 1% for most measurements, though more than 2% for FMDXA and FMbio. For total tissue mass, LTM, and BMC there was a significant biological day to day variance whereas for the other measurements the day to day variation could be accounted for by just measurement variance. Pillows systematically changed the estimates of bone mineral and soft tissue, but sheets did not.