The consortium to establish a registry for Alzheimer's Disease (CERAD). Part XIII. Obtaining autopsy in Alzheimer's disease.

Although autopsy rates in the United States have been decreasing steadily, the necessity for brain autopsy to confirm Alzheimer's disease (AD) remains. Of 308 consecutively deceased AD patients at 24 CERAD (Consortium to Establish a Registry for Alzheimer's Disease) sites, 167 (54%) were autopsied; 141 (46%) were not. The autopsied and nonautopsied groups were comparable in gender (men, 57.5% versus 49.7%), marital status (married, 69.3% versus 67.1%), age at entry (73 versus 74 years), age at death (76 versus 77 years), and stage of disease at entry (mild, 46% versus 43%). However, the autopsied patients were significantly more likely to be white (94.5% versus 82.1%), to be better educated (13.1 versus 11.3 years), to have been in the study longer (mean, 3.3 versus 2.6 years), and to have had longer total duration of AD (8.1 versus 6.7 years). Of the 24 CERAD sites, 13 stressed the importance of autopsy by dedicating a staff member to seek autopsy and by providing free autopsy and transportation; 11 did not. Logistic regression analysis showed that white race (odds ratio [OR] = 2.74; 95% confidence interval [CI] = 1.10-6.83), increased education (OR = 1.12; 95% CI = 1.04-1.21), and emphasis on autopsy (OR = 4.69; 95% CI = 2.67-8.25) were the only significant factors. Although race and education were important, autopsy was more likely to be obtained when sites dedicated resources to this endeavor.