Objective: The aim of this study was to retrospectively differentiate the cognitive profile of subjects with geriatric depression who will later be diagnosed with Alzheimer's disease (AD) from those who will be diagnosed with other dementias, and subjects who will remain with no dementia.

Methods: Forty-four depressed patients admitted to a day hospital program for depression who participated in a historical cohort study were assessed after 7.5 years of follow-up. Fourteen of these subjects subsequently developed dementia: seven met the criteria for probable AD and seven met the criteria for dementias other than AD (Dementia-No-AD; D-NAD, such as dementia with Lewy bodies (DLB), vascular and mixed dementia). Thirty subjects remained without dementia (No Dementia, ND) at follow-up. The three groups were thus compared on their baseline cognitive performances on the six sections of the Mini-mental State Examination (MMSE) and on the five subscales of the Dementia Rating Scale (DRS).

Results: An analysis of variance (ANOVA) and post-hoc Student–Newman–Keuls analyses with an alpha of p<0.05 revealed that the subjects who received a diagnosis of dementia at follow-up had previously had more impairment on tasks measuring attention and memory (DRS-MMSE) than those who did not develop dementia (AD=D-NAD<ND). Moreover, the future AD subjects could be differentiated on the basis of their difficulties on the MMSE-orientation subtest (AD<ND=D-NAD), whereas the future D-NAD subjects initially had more problems with executive functions (DRS) and MMSE-visuospatial abilities (D-NAD<AD=ND).

Conclusion: The identification of early neuropsychological markers in elderly depressed patients highlights the need to evaluate this population broadly as soon as possible in the depression/dementia process in order to improve the prognosis.