[Surrogate end point fallacies -- the urge for randomized trials with clinical endpoints].

Evidence increases that reliance on surrogate endpoints may be detrimental. This is particularly relevant for preventive medicine. Most impressive examples are hormone replacement therapy, vitamin therapy, and drug treatment of cardiovascular risk factors. Only a few of the oral antidiabetic agents have been studied in randomised trials using clinically relevant outcome measures. Most of these drugs or drug combinations turned out to be more dangerous than beneficial. They increased morbidity and mortality despite their blood glucose lowering effects (surrogate endpoints). New compounds are used without evidence of clinical benefit in millions of people worldwide. There is also an increasing number of examples on surrogate end point fallacies related to non-drug interventions (psychological, nursing, educational, social policy).