When Cognitive Decline and Depression Coexist in the Elderly: CSF Biomarkers Analysis Can Differentiate Alzheimer's Disease from Late-Life Depression

Claudio Liguori, Mariangela Pierantozzi, Agostino Chiaravalloti, Giulia M. Sancesario, Nicola B. Mercuri, Flaminia Franchini, Orazio Schillaci, Giuseppe Sancesario

Late-life depression (LLD) and Alzheimer's Disease (AD) are the two most frequent neuropsychiatric disorders affecting elderly. LLD and AD may clinically present with depressive and cognitive symptoms. Therefore, when cognitive decline is coupled with depression in the elderly, the differential diagnosis between LLD and AD could be challenging. The aim of the present study was to evaluate in a population of elderly patients affected by depression and dementia the usefulness of CSF AD biomarkers (tau proteins and β-amyloid42-Aβ42) and 2-[18F]fluoro-2-deoxy-d-glucose positron emission tomography (18FFDG-PET) in early differentiating LLD from AD. Two hundred and fifty-six depressed and demented patients, after performing CSF AD biomarkers and 18FFDG-PET, were distributed in two groups on the basis of the current diagnostic guidelines for AD (n = 201) and LLD (n = 55). Patients were then observed for 2 years to verify the early diagnosis. After the 2 year follow-up we compared AD and LLD patients' CSF and 18FFDG-PET data obtained at baseline to a group of age- and sex-matched controls. We found CSF Aβ42 levels significantly higher in LLD compared to AD patients. Remarkably, CSF Aβ42 levels of LLD patients (range between 550 and 1204 pg/mL) did not overlap with those of AD patients (range between 82 and 528 pg/mL). Moreover, we documented no differences in CSF AD biomarkers (Aβ42 and tau proteins) when comparing LLD patients to controls. In addition, AD patients showed the significant reduction of 18FFDG-PET uptake in temporo-parietal regions compared to both controls and LLD. Conversely, LLD and control groups did not differ at 18FFDG-PET analysis, although LLD patients showed heterogeneous patterns of glucose hypometabolism involving cortical and subcortical brain areas. It is noteworthy that at the end of the clinical follow-up, patients owing to AD group showed the expected significant decline of cognitive performances, whereas patients assigned to LLD group improved cognition as depressive symptoms recovered. Hence, in case of co-existence of cognitive impairment and depression in the elderly, we propose CSF AD biomarkers analysis to early differentiate LLD from AD and properly target the patient's therapeutic strategy and clinical follow-up.