Serum Metabolomic Markers of Artificially Sweetened Beverage Consumption

Abstract

Background: The consumption of artificially sweetened beverages is on the rise. Use of artificial sweeteners has been associated with adverse health outcomes. There is a need to identify novel objective biomarkers of artificially sweetened beverages in order to improve dietary assessment and to provide insight into their metabolic impact.

Objectives: We aimed to identify serum metabolites that are associated with artificially sweetened beverage consumption.

Methods: In the Atherosclerosis Risk in Communities (ARIC) study, consumption of artificially sweetened beverages was assessed using a food frequency questionnaire and fasting serum samples were collected during the first study visit (1987-1989). Participants were categorized as nonusers if they reported almost never consumption of artificially sweetened beverages, moderate users for 1 glass/mo to 6 glasses/wk, and heavy users for ≥1 glasses/d. Untargeted metabolomic profiling was conducted in 2 subgroups (subgroup 1: n = 1866, profiled in 2010; subgroup 2 profiled in 2014: n = 2072), and 360 metabolites were analyzed. In this secondary data analysis, multivariable linear regression models were used, adjusting for demographics, health behaviors, health status, and dietary factors. Analyses were conducted in each subgroup and results meta-analyzed.

Results: In a meta-analysis of 3938 generally healthy participants (mean age, 54 y; 60% women; 62% Black participants) from ARIC study visit 1, 11 serum metabolites were significantly associated with artificially sweetened beverage consumption. Heavier consumption of artificially sweetened beverages was associated with higher concentrations of 10 metabolites (saccharin, threonate, erythronate, glycerate, gluconate, mannitol, glucose, tryptophan betaine, trehalose, and N6-acetyllysine) and lower concentrations of glycocholenate sulfate.

Conclusions: Eleven serum metabolites are related to artificially sweetened beverage intake, which consist of known sugar substitutes, processed food additives, glucose-related compounds, and gut microbiome-related metabolites. These findings enhance our knowledge of the metabolic activity of artificial sweeteners and suggests new biomarkers for monitoring intake.

Keywords: artificially sweetened beverages; biomarkers; dietary assessment; metabolomics; nutrition.

FIGURE 1

Volcano plot of β coefficients and P values for the associations of serum metabolites with intake of artificially sweetened beverages in the Atherosclerosis Risk in Communities (ARIC) Study. Linear regression models adjusted for age, sex, race and center (in subgroup 2), total energy intake, BMI, physical activity, education level, sugar-sweetened beverage consumption, coffee consumption, and tea consumption. Associations were meta-analyzed across the 2 subgroups using fixed-effects regression models. The significance threshold for statistical analysis, after adjusting for multiple comparisons using the Bonferroni method, is represented by the red dashed horizontal line [y = −ln(0.05/360) = 1.39 × 10⁻⁴]. The null value of β = 0 is represented by the red dashed vertical line. Artificially sweetened beverage consumption was modeled as an ordinal variable representing the 3 categories (nonuser, moderate user, and heavy user). Asterisk indicates tier 2 metabolite that did not have a reference standard available but was identified based on physiochemical properties or spectral similarities to the metabolite.