Accuracy of clinical criteria for AD in the Honolulu-Asia Aging Study, a population-based study.

OBJECTIVE: To determine diagnostic accuracy for AD in a population-based study of Japanese-American men. AD is neuropathologically confirmed for more than 80% of cases at major referral centers (primarily Caucasians); however, information on diagnostic accuracy in population-based studies and studies of different ethnic groups is limited.
METHODS: There were 3,734 men who participated in the Honolulu-Asia Aging Study 1991 through 1993 dementia examination and 2,603 in the 1994 through 1996 examination. Diagnoses were based on published criteria. Neuropathologists blinded to clinical data quantified neurofibrillary tangles (NFT) and neuritic plaques (NP).
RESULTS: Of 220 autopsied subjects, clinical evaluation revealed 68 with normal cognition, 73 intermediate, and 79 with dementia: 20 AD, 27 vascular dementia, 19 AD + other, and 13 other dementia. Among 20 cases with pure AD, the median value for maximum neocortical NFT density was 6.9/mm(2) and for neocortical NP density was 8.0/mm2. Corresponding densities for other groups were <3.0/mm2. Using established neuropathologic criteria, 25% (5/20) of clinical AD cases had enough NP to meet definite AD criteria, whereas 65% (13/20) had sufficient NP to meet neuropathologic definite or probable AD criteria. Among nine AD cases with moderately severe dementia, only two (22%) had NP densities great enough to meet definite neuropathologic criteria, whereas seven (78%) met neuropathologic criteria for probable AD.
CONCLUSIONS: Neuropathologic confirmation and NP density among decedents with clinical AD in this population-based study were lower than reported by referral centers and similar to reports from two other community studies. Ethnic differences in propensity for amyloid deposition as well as differences in clinical severity and representativeness of cases might contribute to these findings.