Bone Mineral Density (BMD)

Bone mineral density as clinically measured is an areal quantity, not a volumetric one. Dual-energy X-ray absorptiometry passes two photon energies through a region, solves for the mineral in the beam path, then divides by projected area to give grams per square centimetre. That projection is why the number is confounded by bone size: a larger bone yields a higher areal density at the same true tissue density. Results are reported as a T-score against a young adult reference mean, with the World Health Organization threshold for osteoporosis at minus 2.5 or below.

Precision sets what a change means. Repeat scanning error at the lumbar spine is on the order of one to two percent, so the least significant change is typically around three percent, and comparisons across different machines or scan sites are not interchangeable at all. The peptide and protein anabolics were assessed against fractures rather than density alone: teriparatide, approved in 2002, abaloparatide in 2017, and the sclerostin antibody romosozumab in 2019 all carry fracture endpoint data.

This matters because density is a surrogate. Analyses of antiresorptive trials consistently find only a minority of the fracture risk reduction is statistically explained by the change in BMD, so a drug can move density modestly and prevent fractures well, or move it and do nothing useful. Strength also depends on geometry, microarchitecture and cortical porosity, none of which DXA sees.

The errors worth naming are reading a difference between scans on different machines or in different positions as a treatment effect, and the marketing of growth hormone secretagogues on density grounds. Growth hormone raises bone turnover before it raises mineral, so density can fall over the first months of treatment before rising, and a short follow-up window can be chosen to capture either direction.