Objective.-To evaluate the published data on the association between apolipoprotein E genotype (APOE) and Alzheimer disease (AD) and determine whether the data support the use of genetic testing for diagnosis or prediction of disease. This statement is intended for neurologists, psychiatrists, geneticists, primary care providers, diagnostic laboratories, and the public. Participants.-The joint American College of Medical Genetics (ACMG) and American Society of Human Genetics (ASHG) Test and Technology Transfer Committee developed a 10-member ACMG/ASHG Working Group to assess available data on the association of AD with APOE alleles. To ensure inclusion of clinical specialists primarily involved with AD patients and families, the American Academy of Neurology (AAN) and the American Psychiatric Association (APA) appointed liaisons to the Working Group. Evidence.-Peer-reviewed journal publications obtained from an Index Medicus search or known to members of the Working Group were the source of data on which the statement is based. Consensus Process.-Following discussions with all members of the Working Group, a draft statement was prepared by the chair and circulated among all members until a consensus was reached. The consensus draft was sequentially reviewed and endorsed by the appropriate scientific and executive committees of the ACMG, ASHG, AAN, APA, and the National Institutes of Health-Department of Education Working Group on Ethical, Legal, and Social Implications of Human Genome Research. In some instances, suggestions from these committees were incorporated into the final statement. Conclusions.-There is general consensus that APOEε4 is strongly associated with AD and that when present may represent an important risk factor for the disease. However, at the present time it is not recommended for use in routine clinical diagnosis nor should it be used for predictive testing. Studies to date indicate that the APOE genotype alone does not provide sufficient sensitivity or specificity to allow genotyping to be used as a diagnostic test. Because AD develops in the absence of APOE ε4 and because many with APOE ε4 seem to escape disease, genotyping is also not recommended for use as a predictive genetic test. The results of a collaborative study under way will clarity some of these issues. Whether APOE genotypes have other uses in the management of AD will become apparent over the next few years.