Abstract Body

There are numerous reports about successful proof of concept testing of new treatments in differnt models of Alzheimer's disease (AD), but so far the translational value of these findings was extremely low. Initially mainly transgenic mice overexpressing 1 or more of the pathogenetically important proteins with at least one, but very often multiple mutations were used. These systems are highly artifical and may not really recapitulate the human disease. But even this speculation needs to be carefully discussed, because there was a high tranlational success if proposed disease mechanisms were addressed, e.g. gamma- or beta-secretase inhibitors decreaed Abeta production in animals and in humans. The most critical issue is the overall clinical efficacy, namely the interpretation and translation of cognitive and behavioral findings in transgenic mice and rats.Over the last few years major efforts were made to better model AD, by switching to knock-in models with normal gene expression, or by creation of mice co expressing "human risk gene" of AD. Now it is important to validate their translational value, but unfortunately so far no drug tested in these models was in clinical trials. Beside the right choice of a modle, the experimantal approach itself needs dicussion and standardisation. This includes all steps like blinding, randomisation, sample size calculation, behavioral tests and suitable positive or negative control groups. This should also allow for a better comparibility of data from different systems and a realistic judgement of the potential clinical usefulness of tested new drugs.