Tracking Bone Health Over Time: Why a Single DXA Isn't Enough

A Snapshot vs. a Trajectory

A single DXA scan tells you where your bone density stands at one point in time. It answers the question: "How does my bone density compare to the reference population right now?" This is valuable for diagnosis and initial risk assessment. But it cannot tell you something equally important: the direction and speed of change.

Consider two women, both age 60, both with a femoral neck T-score of -2.0 on their first DXA scan. Without additional information, their management might appear identical. But if one woman's prior scan (3 years ago) showed a T-score of -1.5, she has lost significant bone density and may be on a trajectory toward osteoporosis. If the other woman's prior scan showed -2.1, her bone density has been essentially stable or even modestly improved. These are very different clinical pictures that require different responses — and only serial measurements can reveal them.

Precision and the Least Significant Change

Before interpreting changes between DXA scans, it is essential to understand the concept of measurement precision. No medical test is perfectly reproducible. Every DXA scan includes some degree of measurement error — variability introduced by differences in patient positioning, soft tissue composition, technologist technique, and inherent machine variability. If you scanned the same person on the same machine on the same day with perfect repositioning, you would still get slightly different numbers each time.

The precision error of a DXA facility is determined by performing repeat scans on a series of volunteers and calculating the coefficient of variation (CV) or root mean square standard deviation (RMS-SD). From this, the least significant change (LSC) is calculated — the minimum change in bone density that can be considered a real biological change rather than measurement noise, with 95% confidence.

Typical precision errors and corresponding LSCs are:

• Lumbar spine: Precision error approximately 1-1.5% (CV), corresponding to an LSC of approximately 2.8-4.2%
• Total hip: Precision error approximately 1-1.5%, LSC approximately 2.8-4.2%
• Femoral neck: Precision error approximately 1.5-2.5%, LSC approximately 4.2-6.9%

What this means in practice: if your spine BMD decreased by 2% over 2 years, and your facility's LSC at the spine is 3.5%, that change is within the range of measurement variability. It could represent real bone loss, or it could be noise. You cannot distinguish the two with certainty. Only a change that exceeds the LSC can be considered statistically significant at the individual level.

Best Practices for Serial DXA Monitoring

To maximize the reliability and clinical utility of serial DXA measurements, several principles should be followed:

1. Use the same DXA machine. Different DXA manufacturers (Hologic, GE Lunar, Norland) use different calibration methods, reference databases, and mathematical algorithms. Comparing results from different machines is unreliable and can lead to erroneous conclusions about bone loss or gain. Cross-calibration equations exist but are imprecise. If you must change facilities, be aware that the first scan on the new machine becomes a new baseline.
2. Use the same technologist when possible. Consistent positioning reduces measurement variability. The ISCD recommends that DXA facilities have their technologists perform precision assessments to quantify their individual measurement error.
3. Compare absolute BMD values (g/cm²), not T-scores. T-scores can shift if the manufacturer updates its reference database, even though the patient's actual bone density has not changed. Absolute BMD values are the most reliable metric for tracking change over time.
4. Ensure consistent scan analysis. The regions of interest (ROI) on the spine and hip should be placed consistently between scans. Vertebral levels should be verified (e.g., L1-L4), and any excluded vertebrae (due to artifacts, fractures, or degenerative changes) should be excluded consistently across all scans. The same hip (left or right) should be measured each time.
5. Interpret changes in the context of the LSC. Ask your facility for their precision data. If a change does not exceed the LSC, it should not be over-interpreted.

How Often Should You Get a DXA?

There is no single evidence-based answer, and recommendations vary by clinical context:

For Untreated Individuals With Normal or Mildly Low Bone Density

The study by Gourlay and colleagues, published in The New England Journal of Medicine, provided useful data on this question. Among women aged 67 and older with normal bone density at baseline, fewer than 1% transitioned to osteoporosis within 15 years. Among those with mild osteopenia (T-score -1.01 to -1.49), fewer than 5% transitioned within 5 years. These data suggest that for women with normal bone density or mild osteopenia, repeat DXA every 5-15 years is reasonable.

For women with moderate to advanced osteopenia (T-score -1.5 to -2.49), the transition to osteoporosis occurred more rapidly, and repeat DXA every 1-3 years may be more appropriate.

For Individuals Starting Osteoporosis Treatment

When pharmacological treatment is initiated, a follow-up DXA at 1-2 years helps determine whether the treatment is effective. Most osteoporosis medications increase bone density by 3-8% at the spine over 2-3 years. If bone density is stable or increasing, treatment is likely working (though some medications prevent bone loss rather than increasing density, so stability can also represent success). If bone density continues to decline significantly despite treatment, adherence should be assessed and alternative causes of bone loss should be investigated.

For Individuals on a Bisphosphonate Drug Holiday

During a bisphosphonate drug holiday, DXA monitoring every 2-3 years (along with bone turnover markers) helps determine whether bone density is being maintained or declining, which informs the decision about when to resume treatment.

Bone Turnover Markers: A Complementary Tool

Bone turnover markers (BTMs) are blood and urine tests that measure the rate of bone remodeling. They change much more rapidly than bone density — typically within weeks to months of starting treatment — and can provide earlier feedback on treatment response than DXA, which requires 1-2 years to show meaningful change.

Commonly Used Markers

• Bone resorption markers: C-terminal telopeptide of type I collagen (CTX) is the most widely used serum resorption marker. It reflects osteoclast activity and decreases rapidly with anti-resorptive therapy. CTX should be measured in the fasting state in the morning due to significant circadian variation.
• Bone formation markers: Procollagen type I N-terminal propeptide (P1NP) is the recommended serum formation marker. It reflects osteoblast activity and increases with anabolic therapy (teriparatide, abaloparatide, romosozumab).

Clinical Uses of Bone Turnover Markers

• Assessing treatment response: A significant decrease in CTX within 3-6 months of starting an anti-resorptive (or increase in P1NP with anabolic therapy) provides early evidence that the medication is having its intended biological effect, even before BMD changes are detectable.
• Monitoring adherence: If BTMs do not change as expected after starting treatment, it may indicate poor adherence, improper administration (e.g., not taking oral bisphosphonates correctly), or inadequate absorption.
• Guiding drug holiday decisions: Rising bone turnover markers during a bisphosphonate drug holiday may signal that the residual anti-resorptive effect is wearing off and treatment should be resumed, even before DXA shows significant bone loss.
• Monitoring denosumab discontinuation: A rebound increase in CTX after stopping denosumab can alert clinicians to the need for prompt initiation of a bisphosphonate to prevent rebound bone loss and fractures.

Limitations of Bone Turnover Markers

BTMs have considerable biological variability (fluctuations due to circadian rhythm, fasting state, recent exercise, and fracture healing) and analytical variability (differences between assays and laboratories). They are most useful for tracking relative changes within an individual over time (using the same assay and laboratory), rather than for one-time diagnostic purposes. The ISCD and International Osteoporosis Foundation (IOF) recommend using CTX and P1NP as reference markers, measured under standardized conditions.

Beyond DXA: Other Monitoring Tools

While DXA remains the standard for bone density monitoring, other technologies provide complementary information:

• Trabecular bone score (TBS): Derived from the lumbar spine DXA image, TBS provides an indirect assessment of trabecular microarchitecture. It can be tracked over time and has been shown to change with osteoporosis treatment, providing information about bone quality changes that BMD alone does not capture.
• Vertebral fracture assessment (VFA): A lateral spine image obtained during the DXA scan that can identify vertebral compression fractures. Since many vertebral fractures are asymptomatic (an estimated two-thirds are never clinically detected), VFA can identify fractures that would change management — for example, an asymptomatic vertebral fracture in a patient with osteopenia would reclassify the diagnosis to severe osteoporosis and strengthen the indication for treatment.

Building a Long-Term Monitoring Strategy

Bone health is a long-term proposition, and effective monitoring requires a structured approach:

1. Establish a baseline. Your first DXA scan provides the reference point against which all future changes will be measured. Ensure it is performed at a facility where you can return for future scans.
2. Determine your monitoring interval based on your baseline bone density, risk factors, and treatment status. Discuss this with your clinician — not all patients need DXA every 2 years.
3. Track absolute BMD values across scans, not just T-scores. Keep a record (or request copies of your DXA reports) so that trends can be identified over multiple data points.
4. Consider bone turnover markers as an adjunct, particularly if you are starting treatment, on a drug holiday, or discontinuing denosumab.
5. Reassess your overall risk periodically. Your fracture risk is not static — it changes with age, new risk factors, new fractures, and treatment status. Periodic FRAX recalculation and clinical reassessment ensure that your management strategy remains aligned with your current risk.
6. Maintain communication with your healthcare team. Bone health monitoring is most effective when it is part of an ongoing clinical relationship where results are interpreted in context and management decisions are made collaboratively.

The Bottom Line

A single DXA scan is a starting point, not a destination. Bone health is dynamic — it changes over years and decades in response to aging, hormonal shifts, nutrition, activity, medications, and disease. Serial monitoring allows you and your clinician to detect meaningful changes, assess treatment effectiveness, and adjust your management strategy proactively rather than reactively. The goal is not just to know your T-score — it is to understand your bone health trajectory and to act on it before a fracture becomes the first sign that something has changed.