Effect of calcium tablets on interpretation of lumbar spine DXA scans.

Undissolved calcium (Ca) tablets in the gastrointestinal tract at the time of a dual-energy X-ray absorptiometry (DXA) scan could conceivably affect the accuracy of the lumbar spine bone mineral density (BMD) determination. We studied phantoms and volunteers to determine the effect of Ca tablets overlying bone, Ca tablets in the soft-tissue field, and Ca tablets overlapping both bone and soft tissue. For L1-4, a 500-mg Ca tablet taped to the phantom surface or the skin of volunteers resulted in <or=2.2% mean increase in measured BMD (less than the lumbar spine least significant change), changing the T-score for L1-4 very little. However, an overlying Ca tablet had a substantial effect on BMD of a single vertebral body. Greater effects were seen with higher amounts of tablet Ca and with lower BMD, resulting in as much as a 12.6% mean increase in single vertebral BMD in the in vivo series. Recent versions of DXA software successfully identified tablets in the soft-tissue field as artifact and eliminated the tablet pixels from the analysis. However, tablets overlapping bone and soft tissue necessitated operator intervention to either exclude the affected vertebral body or adjust edges by mapping the tablet as artifact and/or neutral field. A paraffin-coated Ca tablet swallowed by a volunteer was tracked serially by DXA during gastrointestinal transit. Similar effects were seen as with the phantom and volunteer studies using external tablets; there was little effect of the Ca tablet on L1-4 BMD either with the tablet overlying bone or in the soft-tissue field. We conclude that undissolved Ca tablets introduce small artifacts on L1-L4 BMD insufficient to alter diagnostic categorization of patients. However, if only two or three vertebral levels are available, misclassification of the patients may occur due to tablet artifact. The precision of monitoring BMD over time could also be adversely affected by tablet artifact if a tablet was directly overlying bone and undetected, especially with smaller regions of interest and with lower baseline BMD.