Discordance in lumbar spine T-scores and nonstandardization of standard deviations

Previous studies have demonstrated differences in proximal femur bone mineral density T-scores depending on the reference range used. This subsequently was addressed by the recommended adoption of the National Health and Nutrition Examination Survey III reference range. There is, however, no accepted reference range for interpretation of lumbar spine bone mineral density (BMD), and the use of different reference populations by different manufacturers could result in inconsistencies in diagnosis of osteopenia or osteoporosis. We compared lumbar spine BMD, as well as T- and Z-scores, in 59 women measured using Lunar DPXL and Norland Excel densitometers. BMD measured by the instruments was highly correlated (r = 0.98, p < 0.0001). The instruments however assigned significantly different values when BMD was expressed as T-scores. There were also significant differences in BMD assignments between instruments, when expressed as Z-scores. The observed differences relate to the different young normal mean, and SD employed in calculating the T- and Z scores. To conclude, in the lumbar spine, two commonly used DXA instruments provide comparable absolute values but there are significant differences in derived T-scores due to differences in manufacturer- specific reference ranges. There is a need for standardization of the reference ranges used in the lumbar spine.