Abstract
Objectives
Nutritional metabolomics to objectively assess dietary intake in aging permit the opportunity to circumvent measurement errors that accompany subjective means of dietary assessment. At the same time, they may offer insights into mechanisms of action and metabolic disturbances that are actionable targets for modulation through diet in hopes of disease prevention and treatment. However, prior to more broad deployment the pre-analytical and temporal variation over time should be documented in order to design and interpret epidemiological studies properly. We quantified and examined 155 nutrient biomarkers and metabolites selected for their potential relevance to dementia.
Methods
Blood samples from three time points, spanning a 2-year period, were obtained from older adults participating in the NIA-Layton Oregon Alzheimer's Disease Center's Nutrition and Brain Aging Study (NBAS). Blood samples were batched randomly across three time points for quantification of blood amino acids, minerals, water and fat-soluble micronutrients, lipids, one carbon, and kynurenine pathway metabolites using a variety of methods including, tandem mass spectrometry. Pre-analytical coefficients of variation (CV) were calculated for all the biomarkers and intraclass correlation coefficients (ICC) were calculated to evaluate the within-person reproducibility in a subset of 137 participants.
Results
The mean baseline age of the analytic sample (n = 137) was 85.6 (± 8.3, 57 - 101 years), 70% are female, 21% carry the ApoEe4 allele and MMSE was 28.3 (± 1.78). The pre-analytical CVs ranged from 0.9% to 55.0% and the ICC ranged from 0 to 0.87 (25%-tile/median/75%-tile 0.41/0.54/0.66). Twenty four % had ICC < 0.40, 66% had ICC 0.40 −0.75 and 10% had ICC > 0.75.
Conclusions
The pre-analytical and within-person reproducibility of nutritional metabolomics in aging ranges widely. The majority can reliably estimate average concentrations over a 2 year period from a single time point and the biomarkers with ICC's above 0.40 can be used for correction of measurement error and those below 0.40 should consider multiple samples per subject and exploring the methodological and biological explanation for the variation over time.
Funding Sources
Nestle Institute of Health Sciences, Hinda and Arthur Marcus Institute for Aging Research, NIA-Layton Aging & Alzheimer's Disease Center (P30AGO8017).
Correction to: Nutritional Metabolomics in Cancer Epidemiology: Current Trends, Challenges, and Future Directions
