Dietary Macronutrient Composition and Protein Concentration for Weight Loss Maintenance

Abstract

Objective: To examine the association between dietary macronutrient composition and 12-month weight loss maintenance (WLM) in adults who achieved initial weight loss (≥ 5%).

Methods: This prospective cohort analysis used 12-month follow-up data from the Navigating to a Healthy Weight trial. Macronutrient composition (%) was assessed using a 4-day, 24-h dietary recall. Food sources were categorized as discretionary foods, lean meat, vegetables, fruit, grains, and dairy. Primary outcomes included 12-month changes in body weight, fat mass index (FMI), waist-to-height ratio (WHtR), and hip-to-height ratio (HHtR). A nutritional geometry approach was used to examine individual and interactive associations of macronutrient intake, visualizing as response surfaces.

Results: Among 1518 participants (69.8% women; mean age 45 ± 12 years), mean macronutrient composition was 20.6% protein, 33.8% fat, and 43.1% carbohydrate. Protein energy percentage was inversely associated with energy intake (β: -0.33; 95% CI: -0.39, -0.27). Response surfaces revealed that lower proportional energy from protein, diluted by high fat and/or carbohydrate, was associated with higher total energy intake and greater 12-month increases in body weight, WHtR, and HHtR, but not FMI. Consumption of discretionary food, not other food sources, increased energy intake by reducing proportional energy from protein.

Conclusions: Maintaining dietary proportional energy from protein, particularly by limiting discretionary food consumption, was associated with reduced energy intake and improved WLM.

Keywords: cohort study; dietary protein; obesity prevention; weight loss maintenance.

FIGURE 1

Dietary macronutrient composition in relation to total food and energy intake and 12‐month weight loss maintenance. Right‐angled mixture triangle (RMT) plots show the association between macronutrient composition (% of energy) and predicted (A) total food intake (dry weight, g), (B) total energy intake (MJ), (C) fat and carbohydrate intake (MJ), and 12‐month changes in (D) body weight (kg) and (E) waist‐ and (F) hip‐height ratio. In the RMT, proportional energy from protein (%P), fat (%F), and carbohydrate (%C) sum to 100%. While %P and %F increase along their respective axes, %C increases across the diagonal lines with decreasing distance from the origin. On the plots, red represents the highest, while blue represents the lowest value of the outcomes. Mixture models estimated colored surfaces. Models were adjusted for age, sex, initial weight loss, fiber intake, energy expenditure, BMI at baseline, fat‐free mass at baseline, education, frequency of alcohol consumption, trial arm, and country. Statistics are given in Tables S2a and S2b. For reference, the red solid polygon represents the Acceptable Macronutrient Distribution Range (AMDR) for Australia and New Zealand (%p = 15–25, %F = 20–35, %C = 45–65), and the red dotted polygon in the background shows the AMDR for the United States (%p = 10–35, %F = 20–35, %C = 45–65). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 2

Macronutrient composition of discretionary foods and the five food groups. (A–F) Right‐angled mixture triangle (RMT) plots show the macronutrient composition of each food group. In the RMT, proportional energy from protein (%P), fat (%F), and carbohydrate (%C) sum to 100%. While %P and %F increase along their respective axes, %C increases across the diagonal lines with decreasing distance from the origin. The data points in RMT indicate the proportion of energy from protein, fat, and carbohydrate of foods in each food group. Description values are given in Table S4. (G–H) RMT plots show the association of macronutrient composition (% of energy) with the predicted energy proportion from discretionary foods and the five food groups. On the plots, red represents the highest, while blue represents the lowest value of the outcomes. Mixture models estimated colored surfaces. Models were adjusted for age, sex, initial weight loss, fiber intake, energy expenditure, BMI at baseline, fat‐free mass at baseline, education, frequency of alcohol consumption, trial arm, and country. Statistics are given in Tables S5a and S5b. For reference, the red solid polygon represents the Acceptable Macronutrient Distribution Range (AMDR) for Australia and New Zealand (%p = 15–25, %F = 20–35, %C = 45–65), and the red dotted polygon in the background shows the AMDR for the United States (%p = 10–35, %F = 20–35, %C = 45–65). [Color figure can be viewed at wileyonlinelibrary.com]