Differential diagnosis in dementia. Principal components analysis of clinical data from a population survey

To reduce all the clinical data, collected from an unselected group of subjects, to a small set of factors and to see how these factors correspond to standard clinical diagnosis of dementing disorders.

Population survey.

General community: elderly older than 74 years, from an area in Stockholm, Sweden.

Population-based sample including (1) all the screened positive subjects using the Mini-Mental State examination; and (2) a random sample of the screened negative subjects, matched by age and sex. A clinical examination and an informant interview were carried out. Cases were identified using Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition diagnostic criteria for Alzheimer's disease (AD) and other dementias.

Independently from the clinical diagnosis, a principal components factor analysis was carried out to investigate groupings among the clinical data (factors). Factor scores, calculated as a weighted sum of the symptom variables and converted to a standard score form.

Four major factors were found: cognitive impairment, cerebrovascular disease, disturbed behavior, and depressive symptoms. The comparison of these factors with the clinical diagnoses showed that (1) the cognitive impairment factor discriminated demented cases from nondemented; (2) the cerebrovascular disease factor discriminated vascular dementia from AD cases and nondemented; (3) the disturbed behavior factor discriminated AD cases from vascular dementia cases and nondemented, indicating behavioral changes characteristic of AD.

This finding, if replicated, would have implications for the construction of diagnostic criteria for AD.