The role of serial bone mineral density testing for osteoporosis

As a whole, groups of women who gain more bone mineral density (BMD) on antiresorptive medications experience greater fracture protection, although the relationship is not clear on the individual level. A literature search (Medline 1966 to present) for randomized, controlled trials was performed with keywords serial bone density, osteoporosis, dual-energy x-ray absorptiometry, fracture, alendronate, risedronate, calcitonin, estrogen replacement therapy, and raloxifene. Also, reference lists and tables of contents from journals were searched manually for additional relevant randomized controlled trials. Trials were 2-3 years in duration, and the number of subjects ranged from 670 to 3954. Medications analyzed include alendronate, either 5 mg/day or 5 mg/day, followed by 10 mg/day; raloxifene 60 or 120 mg/day; and combination hormone replacement therapy (HRT) of four different regimen types. There have been no controlled studies showing that change in treatment based on serial bone density measurement results in improved patient outcomes. Whereas studies have shown changes in BMD during antiresorptive therapy to be predictive of fracture reduction in groups of patients, their utility in individual patients remains inconclusive. Osteoporotic women who lose BMD in the first year of alendronate or raloxifene use will likely gain BMD in the second year of treatment, illustrating regression to the mean. Effective medication for osteoporosis should not be changed solely because of BMD loss occurring after the first year of treatment. Young, healthy, postmenopausal women taking commonly prescribed doses of estrogen or estrogen/progestin (HRT) rarely lose BMD. Bone loss over the first 12 months of HRT is independent of bone loss in the next 24 months. If bone is not lost in the first 12 months of HRT, there is a significant chance that bone density will be lost later in treatment. Half of placebo-treated women do not lose BMD over 3 years. Treatment should be continued in patients who initially lose bone density on therapy because most will gain density with continued treatment and end in gaining bone overall. Also, patients who gain large amounts of bone in the first year and lose in the second year are not necessarily failing therapy but rather may be showing that a random error in the earlier bone density change corrects itself later. Loss of BMD with alendronate, raloxifene, or combination conjugated equine estrogen/medroxyprogesterone acetate is likely to convert to gain in BMD. More research is needed to confirm that this regression to the mean may apply to all densitometry techniques, antiresorptives, age groups, and genders.