Precision and sensitivity of dual-energy x-ray absorptiometry in spinal osteoporosis

J Bone Miner Res. 1991 Sep;6(9):997-1002. doi: 10.1002/jbmr.5650060914.


This study evaluated the performance of dual-energy x-ray absorptiometry (DEXA) with regard to (1) the correlation with dual-photon absorptiometry (DPA), (2) the ability to discriminate between normal and osteoporotic patients, and (3) long-term reproducibility. The bone mineral density (BMD) of the spine in 112 subjects, both normal and osteoporotic, was measured with DPA and DEXA (Lunar Corporation, Madison, Wisconsin) of the spine. The femur BMD of 22 cases was also measured with both machines. The results for the two techniques were highly correlated (r greater than 0.9, SEM = 0.02 to 0.04 g/cm2). BMD was measured using DEXA in 80 women (mean age = 61 years) with established spinal osteoporosis and 110 normal age-matched controls. The osteoporotic patients had significantly reduced spine and femur BMDs compared to the controls: -23% for L2-4 BMD (Z score = 2.6) and -13 to -20% for femur BMD (Z score = 1.1-1.3). L2-4 BMD had the best discriminative value, with an area under the ROC curve of 94%; the Ward's triangle BMD had an area of 84%. The precision error in vitro in a phantom over a 1-year period was 0.7%. The measured precision in vivo with young adults was approximately 1% (SD = 0.012 g/cm2) for L2-4 BMD and 1.7-2.3% (SD = 0.015-0.022 g/cm2) for femur over the 1 year period. The reproducibility was not as good for osteoporotic patients (SD = 0.017 g/cm2).

Publication types

  • Comparative Study

MeSH terms

  • Absorptiometry, Photon*
  • Adult
  • Aged
  • Bone Density*
  • Female
  • Femur
  • Humans
  • Male
  • Middle Aged
  • Osteoporosis / diagnosis*
  • Regression Analysis
  • Reproducibility of Results
  • Spine