Starting Prednisone Long-Term: When Glucocorticoid-Induced Osteoporosis Prophylaxis Actually Begins
A single patient about to start a medication everyone agrees she needs, with a second decision riding alongside it about whether to protect her bones starting now or wait and see. The disagreement is about the actual threshold, not the general principle.
Corinne B., a 61-year-old woman, spent the last three weeks unable to lift her arms overhead to hang her holiday decorations — the kind of ordinary task that first told her something was genuinely wrong rather than just age catching up with her. Her rheumatologist diagnosed polymyalgia rheumatica after her shoulder and hip girdle stiffness, elevated inflammatory markers, and dramatic response to a diagnostic prednisone trial all lined up. The plan going forward is a taper expected to hold at moderate dose for at least six months, likely longer given how PMR courses tend to run.
Her most recent DXA, done two years ago for an unrelated reason, showed a lumbar spine T-score of −0.8, and the FRAX score calculated from it puts her 10-year major fracture risk at 8 percent — comfortably low-risk, numbers that on their own would not usually start a bisphosphonate conversation. The problem is what those numbers are a measurement of. That scan predates any steroid exposure at all, which means it describes a skeleton that no longer exists in the form it measured. Glucocorticoid-induced bone loss is fastest in the first several months of exposure, and FRAX does not take glucocorticoid dose as an independent input; the ACR's steroid-specific calculator does. Run her prednisone dose and expected duration through it and she returns a medium-risk tier on the same skeleton her FRAX score calls low-risk. Nobody in the room disputes that she needs the prednisone, and nobody disputes that six months at moderate dose will cost her bone. What is actually being decided is which of two instruments gets to describe her, and it is the dose input, not the density measurement, that carries the argument here.
Which risk number actually applies to her
I'd start bisphosphonate prophylaxis alongside the prednisone, not after her next DXA. ACR guidance recommends starting prophylaxis at initiation of glucocorticoid therapy expected to run three months or longer at moderate-to-high dose — and glucocorticoid-induced bone loss is fastest in the first several months, which is exactly the window a wait-and-see plan would miss.
Her own numbers give me pause about starting reflexively. Her DXA is normal and her FRAX is low-risk — ACR's guidance itself stratifies by fracture risk category, not glucocorticoid dose alone, and low-risk patients under that framework don't automatically get pharmacologic prophylaxis.
I take the "fastest bone loss happens early" point seriously — I'm not arguing to wait indefinitely, I'm arguing her FRAX number specifically might not be the right tool to be applying the low-risk label from.
That's the actual gap worth naming directly. FRAX doesn't factor in glucocorticoid dose as its own independent input the way the ACR's steroid-specific risk calculator does — and when I run her prednisone dose and expected duration through that calculator, she lands in the medium-risk tier despite a currently normal DXA.
That's not the calculator overriding her real risk, it's the calculator catching exactly the scenario it was built for — a normal DXA taken before steroid exposure doesn't say much about what six-plus months at this dose will do, and medium-risk by ACR's own framework is a real prophylaxis threshold.
Agreed: start alendronate now, alongside the prednisone taper, based on the ACR glucocorticoid-specific risk calculator's medium-risk output rather than waiting on FRAX or a repeat DXA to justify it retrospectively.
Not fully settled: how long to continue the bisphosphonate if her prednisone course ends up shorter than the six months currently projected — the rheumatologist and endocrinologist left that reassessment for the taper's midpoint rather than committing to a fixed stop date today.