Abstract
At present, the neuropathologic examination is universally regarded as essential for the final diagnosis of Alzheimer disease (AD). However, there is absence of agreement on what areas to examine, what stains to use, which lesions to report, and how to interpret other neuropathologic findings. In this study, we describe the results of reviewing the neuropathologic findings in probands from a family study of AD who had received a clinical diagnosis of βprobable ADβ (N = 33). A neuropathologic diagnosis of AD was received for 88% of the probands, a proportion entirely in agreement with that reported in the literature. An independent review of the autopsy reports found variability in whether the neuropathologic diagnosis was based on the presence of plaques or tangles or both. Further, when a subset of histologic slides was reviewed by two researchers working in the field of AD, there were substantial diagnostic discrepancies. These findings indicate the need for standardized protocols, but suggest that even they have limitations. These limitations are of particular concern when estimating vulnerability to AD of firstβdegree relatives. Thus, while neuropathologic examination continues to provide invaluable information about the causes of dementia, it cannot be regarded as a gold standard for diagnosis of individual cases.