Biomarcadores séricos, desempenho cognitivo e funcional em indivíduos muito idosos provenientes da comunidade: análise de corte transversal e prospectiva após um ano

Abstract

Introduction: The mechanisms underlying cognitive and functional decline associated with aging are still poorly understood. Objective: To investigate the association between blood-based biomarkers and cognitive and functional performance outcomes assessed cross-sectionally and after one year, in a cohort of older seniors from the community. Methods: Participants were recruited from the Pietà Study, conducted in the city of Caeté, Minas Gerais state, Brazil. Briefly, 639 individuals aged 75 years or older underwent a medical evaluation, comprising clinical features, neurological examination, structured psychiatric interview, functional assessment by the Functional Activities Questionnaire (FAQ), and cognitive assessment, consisting of the Mini Mental State Examination, animal category semantic fluency test and the Figures Memory Test. Tests results were normalized into Z-scores according to controls’ performance from the same population, and according to four levels of schooling, which allowed the calculation of a Global Cognitive Score (GCS). A total of 356 participants provided blood samples for serum levels analysis of cortisol, BDNF, and TNF-α soluble receptors, type 1 (sTNFR1) and type 2 (sTNFR2), and Apoliprotein E genotyping. Subjects were classified into three groups: without cognitive impairment (WCI), cognitive impairment no dementia (CIND) and dementia. Those without dementia at baseline underwent the same evaluation protocol in the subsequent year; of these, 183 were reassessed, and 146 of them had provided blood samples at baseline. In the following year assessment subjects were classified as decliners or not according to their worsening nominal performance in GCS or FAQ. Results (1-cross-sectional): Cortisol levels were found to be higher in both cognitively impaired groups (p=0,0002), and sTNFR1 levels were higher only in dementia group (p=0,0007). In logistic regression analysis adjusted for multiple covariates sTNFR1 levels were associated with prevalent dementia (OR=1,0044; CI:1,002-1,007); instead, cortisol levels were associated (OR:1,084 ; CI:1,024-1,147) with prevalent cognitive impairment (CIND + dementia), but not with dementia itself, and in a clinical covariate adjusted-model, cortisol levels did not hold statistical association with CIND status. In both groups with cognitive impairment cortisol levels did not correlate with performance in cognitive measures.(2-longitudinal): No biomarker was found to be associated with GCS and FAQ decline according to the employed logistic regression models. CIND status at baseline was the main predictor of functional decline in one year follow-up (OR=3,17 ; CI:1,19-,45). Conclusion: The absence of a graded pattern for cortisol and sTNFR1 serum levels across the three groups categorized according their functional and cognitive status, a lack of correlation between serum levels and measures of cognitive performance, and the absence of associations with the investigated longitudinal outcomes suggest that these biomarkers do not play a lead role in the mechanisms underlying age-associated cognitive and functional impairment.