PLASMA WATER T2 IMPROVES WITH LIFESTYLE MODIFICATION: NEW PREDICTIONS FROM LINEAR MIXED EFFECTS MODELS

Abstract

Therapeutic Area

Novel Biomarkers

Background

Plasma and serum water T2 are global markers of metabolic health. In an ancillary study of the PREMIER randomized controlled trial, we measured T2 values for 4,578 biobanked plasma and serum samples. The parent study recruited and randomized 810 participants into 3 treatment arms: advice-only (TX1), comprehensive lifestyle intervention (TX2), and comprehensive plus DASH: dietary approaches to stop hypertension (TX3). Health and fitness were measured at 0 (baseline), 6- and 18-months post-intervention. Here, we developed new linear mixed effects models to assess the plasma water T2 response for biomarker validation. The models incorporate adjustments appropriate for a repeated measures study design to minimize confounding and bias. Hypothesis: Plasma water T2 increases (improves) in response to the PREMIER lifestyle interventions.

Methods

Linear mixed-effects modeling was performed using Stata 18.0. The primary outcome was change from baseline in plasma water T2 at 18 months for each treatment arm, evaluated using pairwise marginal comparisons.

Results

Model 1 incorporated baseline plasma water T2, visit (months), and treatment group as fixed effects. Subject ID nested into cohort were random effects. After 18 months, plasma water T2 increased by 25.3 msec in TX1 (p=0.0240), 12.1 msec in TX2 (p=0.2815), and 23.5 msec in TX3 (p=0.0362). Model 2 incorporated additional covariates: baseline plasma water T2, baseline LDL, plasma total protein, fasting insulin, fitness, BMI-Race interaction, visit (months), treatment group, and the interaction between visit in months & treatment as fixed effects. Subject ID nested into cohort, varying by time, served as random effects for Model 2. After 18 months, plasma water T2 increased by 35.2 msec in TX1 (p=0.0003), 27.8 msec in TX2 (p=0.0049), and 30.0 msec in TX3 (p=0.0028). Similar but smaller increases were observed after 6 months: 17.0 msec in TX1 (p=0.0142), 14.3 msec in TX2 (p=0.0459) and 15.3 msec in TX3 (p=0.0364). The time trajectories for Model 2 are plotted in Figure 1.

Conclusions

Model 2 demonstrated substantial increases (improvements) in plasma water T2 after 6 and 18 months for each treatment group. These results validate plasma water T2 as a practical biomarker for monitoring global improvement in cardiometabolic health following lifestyle modification.