Accuracy, precision, and utility of spine and whole-skeleton mineral measurements by dxa in rats

Open Access
Publisher Website
Google Scholar
Abstract
We evaluated the precision and accuracy of dual-energy x-ray bone densitometry (DXA) in 38 male and female rats aged 1–10 months. The coefficients of variation (CV) estimated from same-day paired measurements of bone mineral content (BMC) were 1.26% at the lumbar spine and 0.69% at the whole skeleton, and the corresponding CV for BMC corrected for projected bone area (i.e., bone mineral density, BMD) were 0.57 and 0.66%. BMC, measured in vivo, correlated closely with the subsequently determined ash weights (spine r2 = 0.94, whole-skeleton r2 = 0.97). The long-term CV for BMC measurements, assessed by measuring a frozen animal daily for 4 weeks, were 1.28% for the spine and 1.03% for the whole skeleton; for BMD the corresponding CV were 0.88 and 1.15%. To examine the utility of serial DXA measurements we followed female rats subjected to ovariectomy (OVX) or sham operation at 10 months of age and male rats given daily subcutaneous injections of hPTH-(1–34) or vehicle starting at 10 months of age every 3 weeks for 15 weeks. In the OVX rats a progressive decrease in spine BMC was observed that was most rapid during the first 6 weeks. By 15 weeks the mean spine BMC decreased by 17% in the OVX rats (p < 0.007 versus sham operation). OVX did not affect the accuracy of DXA measurements as assessed by comparison with the ash weight at the end of the 15 week study. PTH treatment increased spine BMC by a mean of 32% and increased whole-skeleton BMC by a mean of 19% within 15 weeks. The long-term in vivo variation was estimated by fitting a model for BMC and BMD in the vehicle-treated male rats and sham-operated female rats, assuming a linear rate of change over the 15 weeks and a different slope and intercept for each animal; the residual root mean squared error was taken as an estimate of variability. Defined in this manner, the long-term variation for BMC measurements in vivo is 5% for the spine and 3% for the whole skeleton; for BMD the corresponding figures are 3 and 2%. This variability is significantly greater than that attributable to instrumental variation and/or same-day repositioning and must be considered when designing experiments requiring serial measurements.XS
Funding Information
  • Hologic, Inc.
  • Chugai Pharmaceutical Co., Inc.