Reply to: Change in Bone Density and Reduction in Fracture Risk: A Meta-Regression of Published Trials.

Abstract

To the Editor We appreciate Dr Bergmann's interest in our study examining the association between changes in dual-energy X-ray absorptiometry (DXA)-based spine and hip bone mineral density (BMD) and antifracture efficacy of various osteoporosis therapies. We agree with his observation that the meta-regression results appeared to underestimate the reduction in vertebral fracture risk provided by changes in BMD for some anabolic therapies (ie, parathyroid hormone [PTH] 1-84, teriparatide, and abaloparatide).1-5 On the other hand, the meta-regression accurately predicted vertebral fracture risk reduction associated with changes in BMD in the FRAME trial of romosozumab, a potent osteoanabolic therapy. Furthermore, the meta-regression results for nonvertebral fractures were similar for anabolic and antiresorptive therapies. It is possible that treatment with osteoanabolic therapies leads to improvements in bone microarchitecture or bone tissue quality above and beyond what is captured by changes in DXA-BMD6, 7 and that perhaps these effects are more prominent in the trabecular-rich spine. Alternatively, anabolic therapies generally lead to more rapid gains in spine BMD than antiresorptive therapies, perhaps thereby allowing a longer time for vertebral fracture prevention. However, at this point, these explanations remain somewhat speculative. Our meta-regression analyses were limited to data from previously published placebo-controlled trials, each of which used slightly different definitions for their fracture outcomes and had varied trial durations. Analyses using individual patient data from these trials that are in progress by our group may address this issue, as fracture definitions and follow-up durations can be standardized. Finally, we note that if changes in BMD were used as a surrogate marker for antifracture efficacy, the underestimation for vertebral fracture reductions that one observes among trials of anabolic therapies suggests that, in the worst case, changes in BMD may provide a conservative estimate of the reduction in vertebral fracture that would be observed in a trial of a new osteoanabolic therapy.