10.7 Proximal Femur Follow-Up Scans and Rate of Change
Key Takeaways
- Total hip is the preferred hip region for monitoring because its precision is better than the femoral neck, which improves the ability to detect real change.
- Hip BMD typically changes by only 1 to 2% per year with antiresorptive therapy, so significant change takes longer to demonstrate than at the spine.
- The same femur must be scanned at every visit, because right and left BMD differ and a side change makes the comparison meaningless.
- Internal rotation must be reproduced, since a rotation difference between visits alters projected neck area and produces a purely geometric apparent change.
- Ward's area and the trochanter alone must not be used for monitoring or for diagnostic classification.
10.7 Proximal Femur Follow-Up Scans and Rate of Change
Quick Answer: The hip changes slowly — roughly 1–2% per year under antiresorptive therapy versus 2–4% at the spine — so significant change takes longer to demonstrate. Total hip is preferred over femoral neck for monitoring because its precision is better (ISCD minimum 1.8% versus 2.5%). Reproduce the same side, the same rotation, the same shaft alignment, and the same ROI placement, and compare absolute BMD against the facility LSC.
Why the Hip Is Monitored Despite Being Slower
The hip changes less than the spine, which is a disadvantage for detecting change. It compensates in three ways:
- Hip fracture is the outcome that matters most. Hip fractures carry the highest morbidity, mortality, and cost of any osteoporotic fracture, and hip BMD is the strongest single predictor of hip fracture.
- The hip is not corrupted by spinal degenerative disease. In the older patient whose spine is inflated by osteophytes, sclerosis, and aortic calcification, the hip is often the only trustworthy site.
- Total hip precision is good. Better precision partially offsets the smaller signal.
Total Hip Versus Femoral Neck for Monitoring
| Total hip | Femoral neck | |
|---|---|---|
| Region | Femoral neck + trochanteric + intertrochanteric | Neck only |
| Area analyzed | Larger | Small |
| ISCD minimum acceptable precision | 1.8% | 2.5% |
| Sensitivity to rotation error | Lower | Higher |
| Preferred for monitoring | Yes | No |
| Valid for diagnosis | Yes | Yes |
A larger ROI averages over more bone, so random placement and positioning variability contribute proportionally less. That is the entire reason total hip monitors better. The femoral neck remains a valid diagnostic site — and is the BMD input FRAX uses — but the total hip is what is tracked over time.
Ward's area and the trochanter alone are used for neither diagnosis nor monitoring. Ward's area is a small low-density region with poor precision; it will reliably produce the lowest number on the report and the least reliable one.
Reproducing the Baseline Femur Study
| Element | Requirement |
|---|---|
| Side | Same femur, every time, without exception |
| Scanner and mode | Same unit, same acquisition mode |
| Internal rotation | Same achieved rotation, using the same positioner at the same setting |
| Shaft alignment | Same parallel alignment; no new abduction or adduction |
| Scan field | Same superior and inferior limits |
| ROI placement | Same neck box position and angle; same global ROI boundaries |
| Manual edits | Same edits reapplied, using the baseline analysis as reference |
The Side Rule
Right and left femur BMD differ in most people, often by several percent, and there is no conversion between them. A baseline on the right and a follow-up on the left produces a difference that is anatomy, not biology.
Document the side prominently, and check it against the prior study before positioning — not after acquiring.
If the previously scanned hip becomes unmeasurable, because of a new fracture or an arthroplasty, the contralateral hip becomes a new baseline, not a continuation. Say so in the report.
The Rotation Rule
Rotation is the femur's largest precision lever. Under-rotation foreshortens the neck, shrinking projected area and raising BMD. A patient rotated correctly at baseline and under-rotated at follow-up will appear to have gained bone; the reverse appears as loss.
The practical defense is the same aid at the same setting plus a check of the lesser trochanter on both images. If the baseline image shows a prominent lesser trochanter and today's does not, the comparison is compromised even though today's scan is technically better. Note it.
Calculating Rate of Change
Worked example. Baseline total hip 0.745 g/cm²; follow-up at 24 months 0.769 g/cm²; facility total hip LSC 0.028 g/cm².
- $\Delta \text{BMD} = 0.769 - 0.745 = +0.024\ \text{g/cm}^2$
- $0.024 < 0.028$, so the change is not statistically significant. Report stable bone density.
- Percent change $= 0.024 \div 0.745 \times 100 = +3.2%$ over two years.
Notice that a 3.2% rise is reported as no significant change at the hip. This is the practical reason hip monitoring intervals are longer: the biological signal at the hip is small relative to the LSC, so a two-year interval may still be insufficient to demonstrate significance.
Second example. Baseline femoral neck 0.690; follow-up 0.648; femoral neck LSC 0.035.
- $\Delta = -0.042$, and $|-0.042| > 0.035$, so this is a significant decrease.
- Percent change $= -0.042 \div 0.690 \times 100 = -6.1%$.
A loss of this magnitude at the hip prompts review: is the patient adherent to therapy, has a secondary cause emerged, has denosumab lapsed, and — technically — were both scans acquired with the same side, rotation, and ROI?
Reporting the Rate
Rate of change may be expressed as absolute change per year, percent change per year, or total change over the interval. All three are derived from absolute BMD, never from T-scores, because reference database and software changes move T-scores independently of the patient.
The ARRT specifications effective January 2027 name rate of change in BMD and in T-score and Z-score under follow-up scans. The distinction to hold is that T-score and Z-score trends may be reported descriptively, but the determination of whether change is real is made on absolute BMD against the LSC.
Timing
- One to two years after initiating or changing therapy is the standard interval, and at the hip the longer end of that range is often more informative.
- Six to twelve months in rapid-loss states: high-dose glucocorticoids, transplantation, immobilization.
- Shorter than six months is rarely useful at the hip, because expected change is far below the LSC.
- Once stability is documented on treatment, intervals lengthen further.
Baseline total hip BMD is 0.745 and the 24-month follow-up is 0.769 g/cm-squared, with a facility total hip LSC of 0.028. How is this reported?
Why is the total hip preferred over the femoral neck for serial monitoring?
A patient's baseline hip study was performed on the right. She has since undergone right total hip arthroplasty. What is the correct approach for follow-up?