Neuropsychological predictors of dementia in a three-year follow-up period: data from the LADIS study.

Background: White matter changes (WMC) are related to cognitive deficits and dementia. Our aim was to determine the extent to which the performance in neuropsychological tests would be able to predict the clinical diagnosis of dementia. Methods: The LADIS (Leukoaraiosis and Disability) is a prospective study that evaluates the impact of WMC on the transition of independent elderly to disability. The subjects were evaluated at baseline and yearly during 3 years with a comprehensive clinical, functional and neuropsychological protocol. At each visit, dementia was classified according to clinical criteria. The performance in the neuropsychological batteries was compared according to the clinical diagnosis of dementia. Results: From the initially enrolled 639 subjects, 480 were evaluated at year 3. Dementia was diagnosed in 90 participants. The demented subjects had worse performance in almost all the baseline cognitive tests. Using receiver operating characteristic curves, we found that the Vascular Dementia Assessment Scale (VADAS) battery had higher sensitivity and specificity rates (area under the curve = 82%) to identify dementia compared with the Mini-Mental State Examination (MMSE) and Alzheimer's Disease Assessment Scale. Worse performance on baseline MMSE (beta = 0.33; p < 0.001) and VADAS (beta = -0.07; p = 0.02) were predictors of dementia (regression analyses). Conclusion: Performance on the MMSE and the VADAS battery were important predictors of dementia at a 3-year period.