Serial BMD measurements are useful to monitor the efficacy of osteoporosis treatment and residual protection during a bisphosphonate holiday.1 However, changes in BMD are only interpretable if they exceed the least significant change (LSC). International Society of Clinical Densitometry (ISCD) guidelines recommend that the LSC should be based on an in-house precision study.2 Since reporting fewer than 4 vertebrae reduces spine BMD precision, LSC will be greater when based on fewer than 4 vertebrae. Recent ISCD positions suggest that once a precision study has been performed for the total spine and LSC for L1-4 has been calculated (denoted LSCL1-4), the same data should be reanalyzed to calculate the LSC for each combination of fewer than 4 vertebrae.3

These new positions raise concern that determining and remembering the LSC for all combinations of 4 vertebrae when reporting BMD is onerous. The ISCD task force proposed a simplified approximation, based on available literature, to estimate LSC for spine BMD from 3 vertebrae (LSC3VB), 2 vertebrae (LSC2VB) or a single vertebral body (LSC1VB) as 10, 35% and 85% higher than LSCL1-4, respectively.3 The ISCD expert panel declined in the end to recommend such an approach or reporting spine BMD based on a single vertebral body because of the limited supporting data, but might consider revisiting the question if there were more robust data 3.

Toward this end, the current analysis was performed to examine the effect of using combinations of fewer than 4 lumbar vertebrae on 95% LSC and to propose a pragmatic way to estimate this from LSCL1-4. The current analysis was conducted using the population-based DXA registry from Manitoba, Canada, and paired spine scans from rolling precision assessments across more than 25 years representing a variety of scanner configurations.