Abstract EP05: Plasma Water T 2 Is A Global Marker Of Cardiometabolic Health: The Premier Study

Abstract

Introduction: Plasma water T 2 detects the multi-component pathophysiology of metabolic syndrome: insulin resistance, dyslipidemia and a pro-inflammatory, pro-coagulation state. In this study, we aimed to determine factors associated with plasma water T 2 to determine if it could serve as a marker of cardiometabolic health. To do so, we analyzed bio-banked samples from the PREMIER study. Hypothesis: Plasma water T 2 is associated with multiple measures of cardiometabolic health among PREMIER participants at baseline. Materials and Methods: This is an auxiliary study of PREMIER with an observational, cross-sectional design. The parent study recruited 810 subjects: 63% female, 35% black, 65% non-black. The exclusion criteria were diabetes, heart failure, prior CV event, cancer, or psychiatric hospitalization within 2 years. In the present study, we analyzed a subset of participants (n=455), based on the availability of bio-banked samples that had been frozen only once. Transverse relaxation time (T 2 ) decay curves were recorded at 37°C for unmodified human plasma using a CPMG pulse sequence and a Bruker mq20 benchtop nuclear magnetic resonance relaxometer. Water T 2 was resolved from non-water components using a discrete inverse Laplace transform. Total plasma protein and albumin were determined, as well as a panel of acute phase proteins, apolipoproteins and metabolic markers. Proteomic analysis was performed using Luminex xMAP technology with Milliplex plates. Statistical analyses were conducted using JMP Pro v16.1.0 (SAS, Inc.). The association of plasma water T 2 (outcome variable) with predictors of cardiometabolic health was analyzed in four stages: (1) screening of predictors using a bootstrap forest machine learning algorithm, (2) clustering of predictors using principal component analysis, (3) associations of predictors with T 2 using multiple linear regression, and (4) validation of regression models using bootstrap forest. Effect size was quantified as standardized beta coefficients, with standard errors and p-values. Results: The final regression model includes the following predictors for plasma water T 2 : albumin, globulins, LDL-cholesterol, C-reactive protein, hepatocyte growth factor, serum retinol, race, post-treadmill heart rate, household income and fasting C-peptide. This model had an adjusted R 2 of 0.57; all terms were significant. Thus, plasma water T 2 was independently associated with markers of dyslipidemia, inflammation, hyperinsulinemia, physical fitness, nutrition (vegetable intake), race and household income. Conclusion: The association of plasma water T 2 with multiple predictors of cardiometabolic health in PREMIER validates T 2 as a global biomarker of cardiometabolic health in individuals without diabetes. This measurement is practical for the point-of-care and shows promise for cardiometabolic health screening and monitoring.