Comment on Reid et al.: Effect of calcium supplementation on hip fractures

Abstract

Dear Editors, As researchers whose work is referred to in support of the hypothesis that calcium causes hip fractures, we feel entitled to comment on the ‘review’ in a recent issue of your journal [1]. The article could perhaps be better labelled ‘for debate’, since it is certainly not an objective review of the subject but rather the presentation of an idiosyncratic opinion. The authors postulate that calcium therapy weakens the femoral neck by inhibiting its expansion, but their published data on hip status assessed by DXA BMD do not support their case nor is it supported by the abstract they quote. Furthermore, they do not report the effect of calcium on the femoral neck in their own study, which would be of great interest. Periosteal expansion occurs on all long bones [2] and is likely to occur at the femoral neck, but whether the process is suppressed by calcium, and if so whether that is of any significance, is another matter. They do not quote our RCT data on the forearm, which shows that calcium supplementation is associated with increased cortical thickness and increased calculated strength in bending at the proximal radius [3]. Nor do they quote Boonen’s metaanalysis [4], which showed a pooled RR of 1.10 (0.89– 1.36; NS) in patients on vitamin D alone in four RCTs on 9,083 patients. By contrast, the pooled RR was 0.82 (0.71– 0.94; P=0.0005) in six RCTs on 45,509 patients on vitamin D with calcium. How can they reconcile their hypothesis with this much larger meta-analysis, which clearly shows the benefit of calcium in hip fracture prevention? But above all, they do not make it clear that the hip fracture data, which they derive from their original paper [5] (17/732 hip fractures on calcium and 5/739 in controls), was the intention-to-treat analysis; in those who actually took the tablets, there were six hip fractures on calcium and two on placebo (NS). How can anyone build such a far-reaching claim on such small numbers? How did it get past the reviewers? We also question the use of the word ‘trend’ by Reid and co-authors in this and their previous paper [6]. They appear to use trend to mean an effect that does not even meet the conventional level of weak statistical significance (P 800 U/day) with calcium (at least 1,200 mg per day) reduces appendicular fractures of all kinds and thus should be the first intervention to reduce fracture risk in the elderly. Osteoporos Int (2009) 20:831–832 DOI 10.1007/s00198-008-0755-3