Application of survival analysis to the inception of dementia.

The age at onset of dementia of the Alzheimer, Pick or senile type in relatives of probands with early onset dementia was examined using survival analytical techniques applied to data collected by Sjögren et al. (1952). Female relatives were found to have a higher risk of dementia than males, and there was a deficit of affected brothers compared with fathers of probands. In these comparisons due allowance was made for age at the last observation of each relative. Relatives of probable and definite Pick probands had a higher risk than relatives of probable and definite Alzheimer probands, but the difference was not significant and dementia did not occur at an earlier age to the former group. For the relatives as a whole, and for subgroups of relatives, the risk of dementia increased with age, at least up to age 80. It is hypothesized that the pattern of the age-related hazard of dementia is due to the nature of the dementing process; that this slow degenerative process is widespread; and that individual differences in the rate of the process are under the influence of genes. The statistical methods are explained in detail as they have rarely been applied to dementia before, though Chase et al. (1983) have used life tables and survival analysis in testing genetic hypotheses, with an application to Alzheimer's disease.