On the effect of body size on DEXA: (caveat, if DEXA shows severe osteoporosis and you are small, you still have concerning osteoporosis and fracture risk)

Source: https://www.inspire.com/windblown/journal/forteo-results/reply/1478797/......... from a poster called "endodoc" who is a physician, presumably an endocrinologist:

"Most patients and most otherwise good physicians who never specifically
studied bone metabolism are unaware of the fact that DXA machines do not actually measure BMD. I realize that sounds bizarre, but it is true. If you look at the units for your BMD values on DXA, they are grams per square centimeter (g/cm2). There is no universe where mass/area equals density. D=M/V, in other words, density equals mass over volume. The DXA measures your BMC (bone mineral content), based on the level of x-rays that reach the detector. More mineral in the path of the beam, less x-rays make it to the detector = higher BMC. Less mineral in the path of the beam, more x-rays make it to the detector = lower BMC.

If DXA machines divided the BMC they actually measure by the volume of the bone scanned, the units would be g/cm3, it would actually be measuring density, and we wouldn't be having this discussion. However, they divide the BMC by the area of the ROI (region of Interest) that the DXA and/or DXA tech puts over the image of your bone; hence, the units are g/cm2. If you've ever seen your hip scan, you will see a rectangle that represents the ROI for the femoral neck sub-region. You will see a smaller square box, either close to the rectangle or touching/partially within the rectangle - that's the ROI for the Ward's triangle sub-region.

The problem with DXA being an areal measurement as opposed to a true volumetric measurement is that, as a result, it systematically over-estimates BMD on individuals whose bones are larger than average (think: those 5'10" gals with the the big wrists who played rugby in college), and systematically under-estimates BMD in individuals whose bones are more petite than average (think: those 5'2" gals with the the small wrists and ankles who did ballet in college).

Because of the confound of bone size on DXA, many smallish women have t-scores of < -2.0 when they are at their peak BMD at age 25. So if their T-score is -2.5 at the time they enter the menopause, it's likely (absent other risk factors, like bone toxic drugs or diseases) to represent an approximation of their peak bone mass. Most non-expert physicians, however, make the idiotic assumption that "everyone starts at average" (trust me, half are above and half are below, and with DXA it's only the big-boned folks who are above), so they interpret your baseline T-score of -2.5 at age 50 as if you've already lost 2.5 SD [standard deviations] of your bone mass. This is absurd."

ps this also explains why the heavier strontium makes DEXA look better than it is. Also please don't use this info as a reason to avoid meds if meds are needed!