Background: There is a lack of evidence of the role of cardio-metabolic risk factors in brain health, especially those which are comprised of physiological information. The aim of this study is to investigate whether urinary metabolites representing systemic metabolism and brain magnetic resonance imaging (MRI)-derived parameters are prognostic markers regarding cognition with a cardio-metabolic risk factor.
Materials and methods: The study focuses on sub-cohort data from the Coronary Artery Risk Development in Young Adults (CARDIA) Study, with targeted metabolite data by high-resolution nuclear magnetic resonance spectroscopy and liquid-chromatography mass spectrometry (N=606), and brain MRI data with 280 participants (48–60 years) at Year 30, 2015–16; brain MRI data for 185 participants were also acquired at Year 25, 2010–11. A path analysis was conducted to investigate mediators of adjusted cognition with cardio-metabolic factors measured at Year 30. Bayesian multilevel models were employed to examine the degeneration of cognition recorded at Years 25 and 30.
Results: Phenylalanine, aminoadipic acid, and tryptophan showed positive associations with fasting glucose in terms of hyperglycaemia, while indole-3-acetic acid showed a negative association. Measures of brain physiology, i.e. cerebral blood flow (CBF) in temporal lobe white matter left, and CBF in left entorhinal area showed negative associations with waist circumference, respectively, in terms of (abdominal) obesity in the cohort (N=74). For cognitively impaired individuals, significant exposures (race, education, family income, sex, age, center for epidemiologic studies-depression, smoking), metabolites (glycine, methionine, aminoadipic acid, tryptophan and kynurenine) and brain MRI-derived parameters (the number of activated voxels within temporal lobe, right anterior cingulate gyrus, and fractional anisotropy in right medial frontal cortex) were identified as physiological markers.
Conclusions: This study, employing a fusion of biostatistical analytic methods using longitudinal epidemiological data, was able to discover complex multivariate epidemiologic, pathologic and phenotypic relationships across neurodegenerative disease networks that were previously unidentified.