8.5 FRAX Limitations & Clinical Adjustments
Key Takeaways
Conventional FRAX does not encode glucocorticoid dose, individual fall frequency, or the full timing and burden of earlier fractures.
The ACR glucocorticoid guideline gives relative factors of 1.15 for major osteoporotic fracture and 1.20 for hip probability when the dose is more than 7.5 mg/day in its specified setting.
Use the named source and method for any adjustment. A published spine–neck method uses the rounded T-score difference, and a valid osteoporosis-range spine score may independently support treatment consideration.
When several adjustments are possible, FRAXplus advises considering the dominant factor because combined accuracy has not been established.
Check what conventional FRAX leaves out
Conventional FRAX represents some clinical factors as yes/no inputs. It records whether a prior fracture occurred, but does not encode the full number, site, severity and recency of all previous fractures. It has no individual falls-frequency field. In someone who has fallen repeatedly or fractured recently, a seemingly reassuring 10-year estimate can understate the relevant clinical concern. Review the history and address immediate risk separately.
| Clinical factor | How it affects interpretation |
|---|---|
| Glucocorticoid dose | Standard FRAX does not accept a continuous dose. Check whether applicable guidance calls for a dose correction. |
| Femoral neck and lumbar spine discordance | Conventional FRAX uses femoral neck BMD. A substantially lower valid spine score can suggest underestimated major fracture probability. A higher valid spine score can have the opposite implication. Check the scan for artifacts before using discordance. |
| Falls and recent or multiple fractures | Consider the details in clinical assessment even though conventional FRAX lacks corresponding quantitative fields. |
| Type 2 diabetes or eligible lumbar TBS | Consider an appropriate validated adjustment if useful to the decision. Document what was used. |
FRAXplus offers separate post hoc adjustments for some of these factors. The developers advise selecting the most clinically relevant adjustment when several are possible, because their combined accuracy is not established. Keep the conventional result and the method of adjustment clear. FRAX can support risk assessment when a therapy change is being considered. A serial probability change does not measure treatment response.
Apply a named glucocorticoid correction
Standard FRAX reflects an average glucocorticoid exposure. The 2010 ISCD/International Osteoporosis Foundation positions describe approximately 2.5–7.5 mg/day prednisone equivalent as the average range represented in the model. Higher doses can lead to underestimation. The 2022 American College of Rheumatology (ACR) guideline gives a practical adjustment for adults in its glucocorticoid-induced osteoporosis assessment who take more than 7.5 mg/day. Multiply the 10-year major osteoporotic fracture probability by 1.15 and the hip probability by 1.20. Follow the method being applied. In particular, the FRAXplus corticosteroid batch tool uses its own category boundary for the high-dose option.
A 68-year-old man has taken prednisone 15 mg daily for nine months. With glucocorticoids entered, his supplied conventional FRAX probabilities are 18.0% for major osteoporotic fracture and 3.2% for hip fracture. Under the named ACR correction:
- Major osteoporotic fracture: .
- Hip fracture: , or about 3.8% to one decimal.
A lower glucocorticoid exposure than the range represented by the model can have the opposite effect on the conventional estimate. The published dose studies use assumptions about exposure and provide estimates, not a substitute for a medication history. Record the actual dose and which correction was applied.
The factors are relative adjustments, not percentage points to add. Interpret both corrected probabilities within the applicable glucocorticoid guideline alongside fracture history, BMD and the exposure pattern. The unadjusted hip probability was already above 3%. Do not describe this case as crossing both thresholds or apply an additional glucocorticoid factor to an estimate that already incorporates the correction.
Interpret spine–hip discordance without changing the input site
ISCD/IOF recognize that a lumbar spine T-score more than one unit lower or higher than the femoral neck can affect the interpretation of FRAX. A published approach by Leslie and colleagues uses the rounded difference between the lumbar spine and femoral neck T-scores to adjust the major osteoporotic fracture estimate up or down by about 10% of its original value per unit. That published method is distinct from the current FRAXplus spine–neck implementation. Do not label the manual method as a universal ISCD requirement.
In an illustrative case with valid measurements, the femoral neck T-score is −1.2, the lumbar spine is −2.8, and the supplied conventional MOF probability is 16.0%. The spine is 1.6 units lower, rounded to two under the published rule. This gives a relative increase of 20%, yielding . That value is still below 20%. Separately, the valid spine T-score already meets the BMD criterion for treatment consideration under U.S. BHOF guidance in the relevant population. Never substitute the spine measurement for the standard FRAX neck input.
ISCD/IOF FRAX Positions · ACR Glucocorticoid-Induced Osteoporosis Guideline · Leslie and colleagues’ discordance method · FRAXplus adjustment guidance
Use the published Leslie rounded-discordance method. A postmenopausal woman has a valid femoral neck T-score of −1.2, spine T-score of −2.8 and supplied conventional MOF probability of 16.0%. What is the adjusted estimate and what separate fact matters for treatment consideration?
18.56%, and the lumbar spine value has no separate role.
20.0% after rounding, which alone determines treatment eligibility.
32.0% because osteoporosis doubles the estimated probability.
19.2%. The valid spine T-score independently meets the U.S. BHOF BMD criterion.
A 68-year-old man has taken prednisone 15 mg/day for nine months. His supplied conventional FRAX probabilities with glucocorticoids entered are 18.0% MOF and 3.2% hip. What are the results using the ACR correction for this dose?
20.7% MOF and 3.84% hip.
27.0% MOF and 4.8% hip.
16.2% MOF and 2.88% hip.
23.0% MOF and 5.2% hip.
A patient’s history suggests that recent fracture, recurrent falls and TBS could each influence the conventional FRAX estimate. What does FRAXplus advise about multiple adjustments?
Apply all three corrections and add their absolute results.
Apply corrections repeatedly until the estimate crosses a treatment threshold.
Identify the most clinically relevant factor for an adjustment because combined accuracy is not established.
Enter each missing variable as a separate femoral neck BMD result.
Sections you finish are checked off in the contents.