Abstract 016: Artificially Sweetened Beverage Consumption, Plasma Metabolomics, And Risk Of Type 2 Diabetes Among US Adults
Danielle E Haslam,Biqi Wang,Jun Li,Marta Guasch,Liming Liang,Clary B Clish,Joann E Manson,Deirdre K Tobias,Clemens Wittenbecher,Walter C Willett,Meir J Stampfer,Mark A Herman,Josee Dupuis,Nicola M McKeown,Vasanti S Malik,James B Meigs,Frank B Hu,Shilpa N Bhupathiraju
DOI: https://doi.org/10.1161/circ.145.suppl_1.016
IF: 37.8
2022-03-01
Circulation
Abstract:Introduction: The relationship between artificially sweetened beverage (ASB) intake and type 2 diabetes (T2D) risk remains inconclusive. Few studies have evaluated whether circulating metabolites that reflect ASB consumption may unveil potential mechanisms underlying the association between ASB consumption and T2D risk. Hypothesis: We hypothesized that a novel metabolomic profile reflecting habitual ASB consumption would positively associate with incident T2D among US adults. Methods: We quantified 120 plasma metabolites with liquid chromatography-mass spectroscopy in N=3,424 Nurses’ Health Study (NHS), NHSII, and Health Professionals Follow-up Study (HPFS) discovery cohort participants, and N=1,870 Framingham Heart Study (FHS) replication cohort participants. Habitual ASB consumption (servings/day; low-calorie cola or other carbonated beverages) was estimated from food frequency questionnaires at the time of blood draw. We used elastic net regression with 10-fold-cross-validation to age- and body mass index- adjusted metabolite levels to identify a metabolite profile associated with higher ASB consumption (ln-transformed). We then derived the continuous ASB metabolomic score as the weighted sum of these metabolites and replicated the analysis in an internal testing set and FHS. Finally, we evaluated the association of quartiles (Q) of the ASB-derived metabolomic score in NHS/NHSII/HPFS and FHS cohorts with incident T2D risk after blood draw (baseline for analysis) using Cox regression models to estimate hazard ratios (HR) and 95% confidence interval (CI), adjusted for demographics, lifestyle, diet, and body mass index at blood draw. Results: The ASB-derived metabolomic profile included 61 metabolites, primarily lipids and amino acids (Pearson correlation coefficients [ r ] between ASB self-reported intake and metabolomic score: r=0.08 [95% CI: 0.05, 0.11; p <0.0001] and r=0.09 [95% CI: 0.04, 0.14; p =0.001], for discovery and replication, respectively). Mean follow-up was 20 and 16 years from blood draw, including 359 and 241 confirmed incident T2D cases for NHS/NHSII/HPFS and FHS, respectively. Compared with participants in the lowest quartile of the ASB-derived metabolomic profile, higher quartiles were incrementally and significantly at higher T2D risk in NHS/NHSII/HPFS (HR [CI] for Q2, Q3, Q4 vs. Q1: 1.45 [1.04, 2.02], 1.83 [1.33, 2.52], 1.62 [1.18, 2.22], respectively; p-trend =0.003) and FHS (1.76 [1.24, 2.48], 1.82 [1.28, 2.58], 2.45 [1.77, 3.40], respectively; p-trend <0.0001). Conclusions: A robust metabolomic profile of ASB intake was not identified, yet the objective and replicable ASB-derived metabolomic profile was associated with higher incident T2D risk. The identified metabolites may reflect ASB intake, western dietary patterns, underlying metabolic health, and/or T2D-related pathophysiology.
cardiac & cardiovascular systems,peripheral vascular disease