Structural effects of raloxifene on the proximal femur: results from the multiple outcomes of raloxifene evaluation trial

Abstract

Raloxifene improves spine bone mineral density (BMD), and its ability to reduce vertebral fractures by 40-50% suggests that it increases vertebral strength. Positive effects on hip BMD suggest a similar strengthening of the hip, but dimensional ambiguities in BMD by dual energy x-ray absorptiometry (DXA) make it difficult to infer strength effects directly. Hip fractures may be too infrequent to evaluate in practical clinical trials; even the Multiple Outcomes of Raloxifene Evaluation (MORE) study with 7,705 subjects was insufficiently powered to show a comparable reduction in hip fractures. An alternative evaluation of hip DXA data in structural terms should provide more direct evidence of treatment effects on hip strength. Hip scans from a subset of the MORE study, including 4,806 postmenopausal women with osteoporosis randomized to daily oral doses of placebo, 60 mg, or 120 mg of raloxifene were reanalyzed by the hip structure analysis (HSA) method. Scans acquired at baseline, 1, 2, and 3 years were evaluated to extract BMD and cross-sectional geometry across the narrowest point on the neck (NN), the intertrochanteric region (IT), and the proximal shaft 1.5 times the minimum neck width distal to the intersection of the neck and shaft axes. While femur outer diameter expanded during follow-up at all three regions, there were no differences in expansion between groups; treatment influenced mainly the amount and distribution of bone within cross-sections. Effects were similar at the two dose levels at the NN region although the 120 mg dose produced a greater effect on section modulus (SM) at the IT region and on BMD, bone cross-sectional area (CSA), SM, average cortical thickness (CT), and buckling ratio (BR) at the shaft region. Compared with placebo after 3 years, treatment groups showed 0.4-2% higher BMD, CSA, SM, and CT and 1-2% lower BR. The smallest treatment effects were evident at the shaft at 60 mg. We conclude that raloxifene does not influence periosteal apposition in the proximal femur but it nevertheless produces small but significant improvement in resistance to axial and bending stresses (CSA and SM, respectively) at all analyzed regions. The significant reductions in buckling ratio suggest that additional strength loss due to cortical instability is also ameliorated by treatment.