Which vertebrae should be assessed using lateral dual-energy X-ray absorptiometry of the lumbar spine
- Cite this article as:
- Jergas, M., Breitenseher, M., Glüer, C.C. et al. Osteoporosis Int (1995) 5: 196. doi:10.1007/BF02106100
The purpose of this study was to determine precision and diagnostic capability of bone mineral density measurements using lateral dual-energy X-ray absorptiometry (DXA) of the lumbar spine in supine position. Duplicate postero-anterior (PA) and lateral DXA measurements were performed in 60 women. Precision errors of the single vertebral levels using lateral DXA ranged from 3.3% to 4.9%. The combination of all levels improved the precision errors to 2.0%. Paired PA and lateral DXA measurements (Hologic QDR 2000) including the vertebral levels L2 to L4 were performed in 331 postmenopausal women. In 42 women an overlap of L4 by the pelvis was suspected on the lateral DXA images. Vertebral fractures were assessed as a fracture/non-fracture dichotomy. L4 and combinations of vertebrae including L4 showed the best discriminatory capabilities with respect to vertebral fractures in receiver operating characteristic (ROC) analyses,t-tests andZ-scores, with smaller variability of the results when multiple vertebral levels were used. The areas under the ROC curves were 0.662 and 0.639 for lateral and PA measurements of L2 to L4, respectively when all women were included. Excluding the women with pelvic overlap on lateral DXA scans improved the ROC area for lateral scans to 0.686 while that for PA scans remained almost constant (0.641). The differences between PA and lateral measurements were not statistically significant. In 162 women of our study cohort an additional quantitative computed tomography (QCT) measurement of the vertebral levels L2 to L4 was performed and overlapping bony structures at the three levels were studied. Overlapping bony structures were found on QCT slices in 96.9% at the L2 level and in 31.5% at the L3 level. At the L4 level an overlap was found in 5.6% of the women in addition to 31 women in whom L4 overlap had been suspected on DXA images. In total, the level L4 was overlapped in 24.7% of the women. Lateral DXA measurements of the lumbar spine with the patient in supine position are meaningful for diagnosis and follow-up of osteopenia. The inclusion of a maximum number of vertebrae, i.e. L2 to L4 (if L4 is not overlapped by pelvic bone), improves precision and diagnostic capability of the method.