Owing to the involvement of Italian Centres in a multicentre, German-Italian therapeutical trial with Alzheimer’s dementia patients, to be assessed with the Alzheimer’s Disease Assessment Scale (ADAS), it was decided that the Italian centres would use an Italian version of the scale, derived from that used by the German centres. However, the lists of words for exploring verbal memory are not merely translated from the German version, but are composed of selective Italian words chosen according to linguistic criteria. This Italian version was validated following the same procedure adopted for validating the German version. We submitted this Italian version to an interrater reliability, test-retest reliability, concurrent validity, internal consistency and sensitivity evaluation, using demented patients. Based on the results of these tests this Italian version of the ADAS proved valid and reliable. Moreover, the results were strikingly comparable to those from the validation of the German version. Our work supports the validity, reliability and transnational comparability of national versions of the ADAS constructed following definite linguistic criteria.