Vertebral morphometry by X-ray absorptiometry: Which reference data for vertebral heights?

IntroductionThe recent improvement in the resolution of dual-energy X-ray absorptiometry (DXA) images enables most vertebral levels to be seen adequately and thus DXA may be a worthwhile alternative to radiologic morphometry for the identification of vertebral fractures (VF). In this multicenter study, we have derived reference data for vertebral heights and their ratios in Italian women using morphometric X-ray absorptiometry (MXA).MethodsDXA scans were acquired in 1254 consecutive pre- and postmenopausal women, (mean age 63.7 ± 11.3, range 26-88 yrs), referred to six osteoporosis centers. MXA analysis of these images was performed by the same operator measuring vertebral heights and height ratios from L4 to T4. We calculated measures of central tendency and dispersion of vertebral heights and vertebral ratios using different approaches (mean and standard deviation as well as median and interquartile range of raw data, mean and standard deviation of trimmed data using an iterative algorithm, and mean and standard deviation of not fractured vertebrae).ResultsIndependently of the approach that we used, all the measures of central tendency were similar, while significant differences were found when compared with reference ranges in other populations. The vertebral heights of our sample at every vertebral level were significantly smaller than both Rea population and the Lunar reference values, even after normalization. Splitting data according to age groups, there was a decrease in the vertebral heights and ratios between the younger and older women.ConclusionsThis study demonstrates that reference data for MXA should be population specific and age matched.

► Morphometric X-ray absorptiometry (MXA) is worthwhile to identify vertebral fractures. ► Diagnosis of vertebral fractures depends on the availability of reference data. ► We derived reference data for MXA using several approaches. ► Vertebral heights of our sample were significantly smaller than other populations. ► This study demonstrates that reference data for MXA should be specific to population.