How should a DEXA scan be used to evaluate bisphosphonate therapy for osteoporosis?

Evidence summary

Testing the effectiveness of therapy for osteoporosis by measuring changes in bone mineral density (BMD) is difficult because changes are often small and occur slowly, and a decrease in BMD does not necessarily mean treatment failure. Testing patients after starting bisphosphonate therapy has been part of many drug trials to assess the effectiveness of therapy. Follow-up testing in clinical practice has not been the focus of a prospective trial and therefore remains controversial.¹

DEXA is considered the gold standard because it is the most extensively validated test for predicting fracture outcomes.² Understanding the rate of bone density response to therapy, and the precision error of DEXA, helps to determine monitoring intervals. The larger the responses in BMD to therapy and the more precise the DEXA scan result, the shorter the period between testing in which clinically relevant differences can be found. Precision error rates are estimated at <1% for the anterior-posterior spine and 1% to 2% for the hip.³ The BMD change after the initiation of treatment must escape the precision error of the testing device or exceed the least significant change (LSC) value.⁴ The LSC—roughly analogous to a 95% confidence interval—is 2.8 times the precision error of the test on a specific machine and site of measurement. If the precision error for DEXA of the femoral neck BMD is 2%, then the LSC is 5.6%.⁵ Changes in BMD of <2%–4% in the vertebrae and 3% to 6% at the hip could be due to inherent measurement error.⁶

A clinician must also understand the anticipated response to the prescribed therapy. It is not clinically useful to retest BMD before a therapy would have time to affect bone turnover. Alendronate and risedronate increase lumbar spine BMD by 5% to 7% and hip BMD by 3% to 6% when used for approximately 3 years.^7,8 These increases in BMD are associated with 30% to 50% reductions in vertebral and hip fractures.⁶ Alendronate continues to increase BMD: following 10 years of treatment, it increased BMD by 13.7% in the lumbar spine, 6.7% in the total hip, and 5.4 % in the femoral neck.⁹

Frequent testing, as seen in bisphosphonate clinical trials, demonstrates the phenomenon of regression to the mean. One analysis of the FIT trial, which compared alendronate with placebo in postmenopausal women with low BMD and at least 1 vertebral fracture, focused on the early evaluation of BMD. The study found a high degree of variability in BMD when tested after 1 year of treatment. This wide variety of response in the first year normalized in the second year.¹⁰ A second analysis showed that when women were divided into 8 groups, the group with the greatest increase in BMD in the first year (10.4%) also had the greatest decrease (1.0%) in year 2. In addition, the group with the greatest decrease in year 1 (6.6%) had the greatest increase in year 2 (4.8%). The variability in response among the 8 groups was approximately 17% (+10.4% and –6.6%) in year 1 and narrowed to a 6% difference in year 2 This regression to the mean leads to a normalization of bone density results.^11,12 This patient variability in BMD response to the prescribed therapy should be considered when deciding to retest.

In summary, limitations in DEXA precision mean any changes in BMD of less than 5.6% at the femoral neck may be due to measurement error, and BMD response to bisphosphonates vacillates in the first few years of use but can be expected to increase femoral neck BMD 3% to 6% over 3 years. Therefore, if serial DEXA scanning is preformed on patients prescribed bisphosphonate therapy, it should be considered no earlier than 2 to 3 years after therapy begins. When monitoring osteoporosis therapy, a BMD change within the LSC should be interpreted as “no change” and should not lead to changes in patient management. If the BMD has decreased beyond the LSC there is cause for concern and reevaluation of diagnosis and treatment are warranted.⁴