[Prevention of osteoporosis: estrogen substitution paradigm]

There is an increasing belief among most physicians that hormone replacement therapy (HRT) is unquestionable efficacious to prevent osteoporotic fractures in women after the menopause. However, an analysis of the studies which are regarded as efficacy proof of HRT concerning the prevention of hip fractures--as the most important manifestation of osteoporosis--shows that the data are weaker than generally assumed: there are no randomised trials with regard to hip fracture prevention by HRT and the protective effect which is seen in the observational studies is inversely related to the methodical state of the respective type of study. Consequently, a bias cannot be excluded, while the so-called 'selection-by-health'-bias could play an outstanding role. For the efficacy proof of a treatment randomised controlled trials cannot be replaced by results of observational studies, however impressive they may be.