Abstract
Dual Energy X-ray Absorptiometry (DXA) is effective in measuring bone mineral density (BMD) in mice for early detection of osteoporosis. However, scanners designed for use with small animals (i.e. PIXImus) are very expensive. Used human DXA machines are cheaper to obtain, but analysis of scans from these instruments is operator-dependent. Obtaining reliable data depends on having a single operator analyze the scans in a blinded fashion. Scan quality is improved by excising the bone prior to scanning, which does not allow serial measurements. This study describes a novel method of analyzing lumbar spine BMD in mice using whole body DXA. This non-invasive technique has a high degree of precision and reproducibility, with good correlation between multiple observers. Inter-observer variability (0.063 +/- 0.00317 g/cm(2) [mean +/- SD], 5.05 [% coefficient of variation (CV)], repeat scan variability (0.063 +/- 0.00364 g/cm(2) [mean +/- SD], 5.94 [%CV]) were very low compared to variability between different animals (0.063 +/- 0.00588 g/cm(2) [mean +/- SD], 9.64 [%CV]) and variability seen in same animal over time (0.011 +/- 0.00885 g/cm(2) [mean +/- SD], 80.68 [%CV]). The measurement error is thus smaller than the biological variation. Accuracy was determined by comparing average peak BMD from two scans per mouse in-vivo (0.066 g/cm(2)) versus excised spine (0.065 g/cm(2)). Furthermore, correlation between bone ash weights and whole body lumbar spine BMD measurements (p < 0.0001) was highly significant. This technique thus shows a high degree of precision and accuracy, even with multiple observers, for measuring BMD in mice using a DXA machine designed for clinical use.