Twin studies of Alzheimer disease: II. Some predictions under a genetic model.

The twin method for investigating genetic and environmental causes of disease has been applied mostly in early-onset illnesses. Analysis of late-onset disorders requires reexamination of common assumptions about the relation between genetic causes and the degree of concordance expected. This paper considers Alzheimer disease (AD) as an example of a late-onset disorder with putative genetic factors. For argument it employs the strong hypothesis that AD is an autosomal dominant trait with age-dependent expression, as described by a previously published parametric model. That model encompasses 2 principal variants of disease: a rare form with onset in middle life, and a more common late-onset type which is nonetheless eventually fully penetrant. The present work then specifies the probability that, when a given member of a twin pair (the proband) is affected, an identical or fraternal co-twin also shows the disease. Such probability is expressed as a function of the age at onset of the proband and the current age of the pair. Even under strong working assumptions regarding genetic influence, the expected proportion of identical co-twins actually affected with AD will not exceed 40% until the subjects are about 80 years old. Therefore, except in very old subjects, modest twin concordance is a feeble argument against genetic causes, or in favor of exclusively environmental ones. In this sense the interpretation of results of twin studies in AD and other late-onset disorders differs substantially from studies of diseases with early onset.