To the Editor:
The chronology presented by Dr. Parfitt is correct, and we regret that we had not been more explicit about it in our Editorial.(1) At the 4th International Symposium on Osteoporosis in Hong Kong in 1993, both he(2) and our group(3) advanced the concept that high bone turnover increased the risk of vertebral fractures, not just by decreasing bone density but also by disrupting bone microarchitecture. In our Editorial, we did not cite either of these publications because they were general discussions limited to two pages. Instead, we cited the original publications on which our mutual concept was based—the study of Parfitt et al.(4) which provided histomorphometric evidence that trabecular microarchitecture is disrupted in osteoporosis, and the data from our group,(5) documenting that estrogen reduces vertebral fracture risk mainly by reducing bone turnover (based on a computer-generated graphic reanalysis of a previous treatment trial with transdermal estrogen).(6) As noted by Dr. Parfitt, it is impressive that two independent types of analysis lead to the same conclusions.
Our recent Editorial was focused on resolving the paradox that raloxifene therapy can decrease vertebral fracture rates substantially, despite only minor increases in bone density. Although we suggested that normalization of bone turnover would reduce fracture risk by eliminating perforative resorption, loss of trabecular connectivity, and other adverse effects on trabecular microarchitecture, we did not elaborate on the biomechanical consequences of antiresorptive therapy as Dr. Parfitt has done here. However, his suggestion that increased numbers of resorption cavities on remaining vertical trabeculae would destabilize them by increasing the risk of buckling and that this would result in disproportionate weakening of bone, is an important and elegant extension of the concept. It also fits well with the hypothesis advanced in our Editorial that only a minimal therapeutic threshold need be reached to reduce this adverse effect of high turnover on trabecular microarchitecture. If so, this could explain why less potent and more potent antiresorptive agents might result in similar reductions in vertebral fracture risk.
We are pleased that Dr. Parfitt has reinforced the points made in our Editorial. The concept that much, perhaps most, of the reduction in vertebral fractures induced by treatment with antiresorptive drugs is due to a reduction in bone turnover is of great theoretical and clinical importance, yet still is not widely appreciated. Although data supporting this, such as the prediction of fractures from baseline bone turnover data(7,8) continue to accrue, more studies are clearly needed. Methods for assessing trabecular structure noninvasively should soon become available,(9) and such measurements may further improve fracture prediction.