You are What You Drink? How Associations Between Profiles of Beverage Consumption and Type 2 Diabetes Risk are Mediated by Biomarker Networks

Abstract

Background: Multiple studies have independently investigated the associations of the consumption of individual beverage types and specific plasma biomarkers with the risk of type 2 diabetes (T2D). However, as individuals do not consume single beverage types exclusively and plasma biomarkers do not act in isolation, it remains unclear how patterns of beverage consumption and plasma biomarker networks associate both with each other and T2D risk.

Objectives: We aimed to elucidate potential dietary determinants of T2D risk by defining a model that describes habitual beverage consumption profiles in relation to identified networks of circulating plasma biomarkers.

Methods: This study included 1,461 case and 1,568 control participants from case-control studies of T2D nested within the Nurses' Health Study. Participants completed validated semiquantitative food frequency questionnaires that assessed habitual beverage consumption, and they provided blood samples from which 27 plasma biomarkers of cardiometabolic risk were identified. Common exploratory factor analysis (EFA) identified factors that separately described beverage consumption profiles and biomarker networks. Multivariable-adjusted regression elucidated the relationships between beverage and biomarker factors and T2D risk.

Results: EFA revealed five factors describing unique beverage consumption profiles and seven factors describing biomarker networks. The factor describing alcoholic beverage consumption was associated with a reduced risk of T2D (odds ratio [OR]: 0.50 [0.40, 0.64], P<0.001) mediated, in part, by the factor describing increased concentrations of adiponectin biomarkers (19.9% [12.0, 31.1] P = 0.004). The factor describing low-calorie sweetened beverage (LCSBs) consumption was associated with an increased risk of T2D (OR: 1.33 [1.03, 1.72], P = 0.021), and the factor describing lower concentrations of insulin-like growth factor binding proteins 1 and 2, and soluble leptin receptor, and increased leptin concentrations (P = 0.005).

Conclusions: Moderate alcohol consumption was associated with reduced T2D risk, mediated in part by increased circulating adiponectin. LCSB consumption was associated with both increased T2D risk and perturbed insulin-like growth factor and leptin signaling.

Keywords: Nurses’ Health Study; beverage consumption; mediation; plasma biomarkers; type 2 diabetes.