The exposure to neurotoxic trace elements has been implicated in the etiology of several neurologic diseases and some recent concern has focused the metalloid selenium (Se). Se may exists in several chemical species with markedly different and even opposite biological properties and both excess or deficiency of Se has been linked to Alzheimer’s disease. In this study we sought to evaluate the different interpretations that could be carried out depending on the time of sampling of Se. 

CSF samples from 33 AD (‘stable AD’) and 50 MCI subjects were finally collected, and of these latter after a median follow-up of 42 months, 29 remained MCI (‘stable MCI’), while 21 evolved to AD (‘converted AD’). Our results showed that patients with pre-existing stable AD had lower levels of overall Se and inorganic Se in particular compared with stable MCI and converted AD, as baseline total Se CSF level were 4.09 (Interquartile range - IQR 3.72-4.56) in stable MCI, 4.40 (IQR 3.63-5.09) in converted AD and 3.68 (IQR 2.91-4.31) in stable AD and selenate (one of the main inorganic Se species) levels were 0.12 (IQR 0.09- 0.31) in stable MCI, 0.23 (IQR 0.13-0.34) in converted AD and 0.12 (IQR 0.06-0.23) in stable AD. Levels of organic species showed opposite results with 1.88 (IQR 1.28-2.27) in stable MCI, 1.60 (IQR 1.03-2.18) in converted AD and 1.84 (IQR1.19-2.25) in stable AD. 

Our results point out an opposite behavior of Se levels depending on both species considered and time of sampling. That difference could explain such inconclusive results of previous case-controls studies that generally assessed Se levels after the onset of the disease, thus they could be hampered by the neuropathological modifications during the progression of the disease. Therefore, our findings highlight the importance of considering both the type of Se species and the time of sampling in order to avoid misinterpretation of the etiologic role of this metalloid into the process leading to AD, especially when results come from non-prospective studies.