Alzheimer's disease in the National Academy of Sciences-National Research Council Registry of Aging Twin Veterans. III. Detection of cases, longitudinal results, and observations on twin concordance

J. C. Breitner, K. A. Welsh, B. A. Gau, W. M. McDonald, D. C. Steffens, A. M. Saunders, K. M. Magruder, M. J. Helms, B. L. Plassman, M. F. Folstein and al. et
Department of Psychiatry, Duke University Medical Center, Durham, NC 27710, USA.

OBJECTIVES: To detect cases of Alzheimer's disease (AD) in a large population of twins living throughout the United States and to examine concordance for AD in twins as a function of age and genotype for apolipoprotein E (APOE). SETTING: Nationwide survey. DESIGN: Multistage screening and field evaluation beginning with two telephone interviews and culminating with laboratory tests, longitudinal neuropsychological measures, physician examination, and diagnostic consensus among experts. PARTICIPANTS: Membership in 1990-1991 of intact pairs in the National Academy of Sciences--National Research Council Registry of veteran twins, then aged 62 to 73 years. MAIN OUTCOME MEASURES: Completeness of case detection was examined in collateral studies. Zygosity and APOE genotypes were determined by restriction mapping. Concordance was calculated by the proband method. RESULTS: Ninety subjects who screened positively for AD were studied in person, and 60 whose differential diagnoses included AD were followed up, as were their co-twins. Sensitivity of screening was estimated at greater than 99%, but 24% of subjects refused participation after initial screening. Seven of 38 diagnoses of AD have been confirmed at autopsy, and 31 other subjects eventually met criteria for probable or possible AD (prevalence estimate, 0.42%, 95% confidence interval, 0.29% to 0.56%), with good interrater reliability (intraclass r = .86). Excluding one discordant pair with unknown zygosity, concordance rates were 21.1% (4/19) for monozygotic and 11.1% (2/18) for dizygotic probands. Concordance was 50% for twins sharing the epsilon 4/epsilon 4 genotype at APOE, but there were no affected co-twins of 15 probands with onset before age 70 years, no epsilon 4 allele, and no family history of AD. The mean (SD) period of discordance in the latter pairs was 11.3 (3.3) years. CONCLUSIONS: The multistage case-detection approach achieved reliable and valid diagnoses of AD with high apparent sensitivity but substantial attrition after initial screening. Genetic influences in AD at this age are limited, except among homozygotes for allele epsilon 4 at APOE. Subjects with early-onset AD who lack the epsilon 4 allele are not rare, and their condition appears to have little genetic influence. They should be ideal for studies on environmental cause of AD.