Objectives: Most clinico-neuropathological correlative studies of Alzheimer's Disease (AD) are based on research cohorts that are not necessarily generalizable to patients seen in the general medical community. In this study, we examine the accuracy of the criteria used in diagnosing AD in a community-based case series of patients with memory complaints.
Design and participants: Clinical and neuropathological diagnoses were obtained from 134 patients evaluated for dementia who subsequently underwent autopsy.
Setting: Subjects who exhibited new symptoms of dementia and were enrolled in the University of Washington/Group Health Cooperative Alzheimer's Disease Patient Registry were eligible for this study.
Measurements: Clinico-pathological correlation was performed using NINCDS-ADRDA (National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association) and CERAD (Consortium to Establish a Registry for Alzheimer's Disease) criteria.
Results: Ninety-five of the 134 cases studied met CERAD neuropathological criteria for AD. The sensitivity of NINCDS-ADRDA "probable AD" was 83% (diagnosing AD correctly) and overall clinical diagnostic accuracy was 75%. However, there was a high rate of additional neuropathological findings. Only 34 of the 94 cases had pure AD on neuropathology, whereas the remainder frequently had coexisting vascular or Parkinson's disease lesions.
Conclusions: This study of a large series of community-based incident dementia cases provides a way of judging the adequacy of currently available clinical diagnostic criteria. It also shows that co-existing neuropathological findings are common in community-based AD.