[Assessment of cognitive function of the elderly by serum metabolites of brain-gut axis].

ABSTRACT

Objective: To investigate the feasibility of assessing cognitive function of the elderly by serum metabolites of brain-gut axis. Methods: Convenience sampling was used to select 100 and 60 participants from the healthy population cohort and microecological balance cohort of the longevity population in Guangxi, to constitute subset of healthy population and longevity population, respectively. A questionnaire was used to investigate the demographic characteristics of the subjects, 2-5 ml of fasting venous blood was collected from the subjects, and the serum untargeted metabolomics was determined by liquid chromatography tandem mass spectrometry. The biomarkers related to the brain-gut axis were collected through literature retrieval, and the results were intersected with the untargeted metabolites and annotated. Spearman correlation analysis was used to screen serum metabolites of brain-gut axis associated with aging, and multiple linear regression method was used to construct biological age model. The mini mental status examination was used to evaluate the cognitive function of longevity population subsets. The differences of biological age and chronological age of longevity population subsets with different cognitive function were compared. Results: The M (Q1, Q3) of subset of healthy population and longevity population were 64 (38, 72) and 97 (95, 99) years old, respectively, and there were 50 (50.0%) and 44 (73.3%) females, respectively. Nine serum metabolites of brain-gut axis were obtained by initial screening, which were propionic acid, glutamic acid, γ-aminobutyric acid (GABA), lactic acid, 5-hydroxytryptamine (5-HT), tryptophan, trimethylamine oxide, dopamine and canine urea. Spearman correlation analysis showed that glutamic acid and dopamine were positively correlated with aging (r values were 0.208 and 0.524, respectively, all P values<0.05), and tryptophan, 5-HT and GABA were negatively correlated with aging (r values were -0.308, -0.533 and -0.213, respectively, all P values<0.05). The biological age model was constructed as y=49.81-1.18×10-5× GABA-1.82×10-4×5-HT+1.99×10-3×dopamine+1.65×10-6×glutamic acid -2.04×10-6×tryptophan+2.36×gender, where y was the biological age (years), the items on the right were the intercept item, the relative concentration of each metabolite, and gender (male=1, female=2). The coefficient of determination of model was 0.50 (P<0.001). The M (Q1, Q3) of the chronological age of the subset of longevity population with poor, moderate and good cognitive function were 97 (94, 100), 97 (93, 101) and 96 (94, 101) years old, respectively, and there was no statistical significance in pairwise comparison (all P values>0.05). The M (Q1, Q3) of the biological age of the subjects with better cognitive function was 51 (38, 54) years old, which was lower than that of the subjects with poor cognitive function [57 (47, 61)] (P=0.040). Conclusion: The biological age model can be constructed based on serum metabolites of brain-gut axis and used to evaluate the cognitive function of the elderly.