Changes in Body Composition in Relation to Metabolic Syndrome: A Compositional Analysis in Adults with Overweight and Obesity.

Abstract

Current knowledge of the association between body composition and health outcomes is based on traditional regression techniques, where the components of body composition are treated as noncompositional independent variables. Mounting evidence suggests that body tissues are biologically co-dependent and therefore, require a statistical technique that considers this. This study used a compositional data analysis framework to explore the longitudinal association between body composition and a continuous metabolic syndrome score. Participants included 288 physically inactive adults (age: 56 ± 12 yr [mean ± SD]; 56% female) with overweight or obesity (body mass index: 31.3 ± 3.5 kg·m 2 ) who participated in randomized controlled trials that determined the effects of exercise on adipose tissue (visceral, abdominal subcutaneous, peripheral subcutaneous, other adipose tissues) and lean tissues (skeletal muscle, other lean tissues) assessed by whole-body magnetic resonance imaging. Visceral adipose tissue, relative to the mass of the remaining tissues, was significantly associated with the metabolic syndrome score preintervention and postintervention ( P < 0.05). The slopes and intercepts of the preintervention and postintervention regression lines between relative visceral adipose tissue mass and metabolic syndrome did not differ ( P > 0.2). For a given weight loss, the greater the relative reduction in visceral adipose tissue, the larger the decrease in the predicted metabolic syndrome score. This novel compositional data analysis reinforces that visceral adipose tissue is an important marker of cardiometabolic risk and should be a primary target for therapeutic strategies in individuals with overweight or obesity.