Neuropsychological screening tests such as the Mini Mental State Examination (MMSE) are commonly used for case finding in community studies on dementia or Alzheimer's disease (AD). However, the high proportion of false-positives is an important limitation to the feasibility of such studies. The aim of this study was to evaluate whether adding apoliporotein E (apoE) genotyping to the MMSE is followed by a significant reduction of the false-positive rate. Subjects were 70 AD patients (MMSE 13-28) and 70 normal controls (MMSE 25-30). Multivariable discriminant analysis was used to classify subjects on the basis of age, gender, MMSE score and the presence of the epsilon 4 allele of apoE. When sensitivity was set at 99%, the model including age, gender and MMSE had a false-positive rate of 13.5%, while adding epsilon 4 to the previous variables decreased this figure to 6.7%. In a hypothetical community study screening for AD in a population of 1,000,000, this would turn in a decrease of false-positives from about 19,000 to about 9,500. We conclude that the use of apoE genotyping in community case-finding studies is promising and should deserve further consideration.