Prevalence

It is difficult to estimate the frequency with which etiologies other than AD and cerebrovascular disease (vascular dementia) cause the dementia syndrome. There are problems in the uncritical acceptance of frequency estimates based on autopsy series (Byrne, Smith & Arie, 1991) as such samples are usually highly selected, and pathological diagnosis, until recently, has been less than reliable. Other factors such as the medical specialty from which the autopsy sample was derived have been shown to influence the frequency with which different etiologic causes for the dementia syndrome occur (Jellinger et al., 1990). AD is more common in autopsy samples referred by psychiatrists, whereas Parkinson's disease is most frequently found in samples referred by geriatricians and physicians. Galasko et al. (1994), found that 12% of their sample of (predominantly) community residents with dementia who came to autopsy had non-AD, nonvascular dementia.

Epidemiological studies of dementia rarely consider differential diagnosis and those that do are largely limited to AD and vascular dementia (Hofman et al., 1990). We do know the prevalence of some of the individual conditions that cause dementia and in some instances have information on how frequently this occurs. For example, the prevalence of Parkinson's disease is 2% of the population aged over 70 years (Schoenburg, 1987) and dementia arises in around 15–20% of these patients (Brown & Marsden, 1984; Mayeux et al., 1988). Lishman (1987) lists 39 non-AD, nonvascular causes of dementia and McKeith (1994) lists 42 other dementias.