Consistent with previous reports, we observed a significant association of the APOE epsilon 4 allele with Alzheimer's Disease (AD) in a series of 91 autopsy-confirmed cases. The epsilon 4 allele frequency was higher in cases with a family history of AD-like dementia (0.54 +/- 0.07), although the epsilon 4 allele frequency in the AD cases with a negative family history (0.38 +/- 0.05) remained significantly greater than that for the non-AD control group (0.13 +/- 0.03). A similar increase in epsilon 4 allele frequency (0.54 +/- 0.07) was observed in the AD cases with amyloid angiopathy, compared to those who did not have amyloid angiopathy (0.36 +/- 0.04). Contrary to previous reports, no effect of the dosage of the epsilon 4 allele was found on the age of onset of dementia among the AD cases and, contrary to reports suggesting an association of epsilon 4 and atherosclerosis, the epsilon 4 allele frequency was similar in cases with or without concurrent brain infarcts. Modest but consistent correlations were observed between the dosage of epsilon 4 alleles and the cortical density of senile plaques, but not neurofibrillary tangles. The last finding suggests that the pathogenic events mediated by the epsilon 4 allele may be more directly involved in the formation of senile plaques, the identifying lesions in AD, than neurofibrillary tangles. A robust association of both the presence of an epsilon 4 allele and a family history of AD-like dementia with concurrent amyloid angiopathy occurred within our sample of AD cases. This association arose from an interaction of the epsilon 4 allele with a separate familial factor for which a family history of dementia served as a surrogate. These results suggest that amyloid angiopathy may be a common or central feature of a form of familial AD that is associated with the transmission of the APOE epsilon 4 allele.