Outcome variables in osteoporosis trials.

The conduct of controlled clinical trials examining the anti-fracture efficacy of potential therapies for osteoporosis is a relatively new and developing science. We have reviewed several aspects of data acquisition, emphasizing that bone mass measurement and, similarly, back pain can only be regarded as ancillary outcome variables. Serial measurement of stature when performed in a precise manner may be an inexpensive, quick, and convenient surrogate outcome variable, most applicable to large multi-center studies. The only true end-point of such trials is the radiographic documentation of new vertebral fracture occurrence. Techniques for obtaining serial radiographs and assessing vertebral morphometry with good quality control have been described. Important questions still need to be resolved concerning the most appropriate criteria for defining incident fractures and for calculating fracture frequency. Implications regarding trial sample size requirements are discussed.