Double trouble: Portion size and energy density combine to increase preschool children's lunch intake

Abstract

Background: Both portion size and energy density (ED) have substantial effects on intake; however, their combined effects on preschool children's intake have not been examined when multiple foods are varied at a meal.

Objective: We tested the effects on intake of varying the portion size and ED of lunches served to children in their usual eating environment.

Design: In a crossover design, lunch was served in 3 childcare centers once a week for 6weeks to 120 children aged 3-5y. Across the 6 meals, all items were served at 3 levels of portion size (100%, 150%, or 200%) and 2 levels of ED (100% or 142%). The lunch menu had either lower-ED or higher-ED versions of chicken, macaroni and cheese, vegetables, applesauce, ketchup, and milk. Children's ratings of the foods indicated that the lower-ED and higher-ED meals were similarly well liked.

Results: The total weight of food and milk consumed at meals was increased by serving larger portions (P<0.0001) but was unaffected by varying the ED (P=0.22). Meal energy intake, however, was independently affected by portion size and ED (both P<0.0001). Doubling the portions increased energy intake by 24% and increasing meal ED by 42% increased energy intake by 40%. These effects combined to increase intake by 175±12kcal or 79% at the higher-ED meal with the largest portions compared to the lower-ED meal with the smallest portions. The foods contributing the most to this increase were chicken, macaroni and cheese, and applesauce. The effects of meal portion size and ED on intake were not influenced by child age or body size, but were significantly affected by parental ratings of child eating behavior.

Conclusion: Strategically moderating the portion size and ED of foods typically consumed by children could substantially reduce their energy intake without affecting acceptability.

Trial registration: ClinicalTrials.gov NCT02299531.

Keywords: Eating behavior; Energy density; Energy intake; Obesity; Portion size; Preschool children.

Figure 1

All foods at the experimental meals were served at 3 levels of portion size (100%, 150%, or 200%) and 2 levels of energy density (100% or 142%). Children were served one experimental meal per week for 6 weeks; thus, they never saw the different meals together.

Figure 2

Mean (±SEM) food and milk intakes (g and kcal) at lunch by meal portion size and energy density (ED) in 120 preschool children. Figure 2A: There was a significant effect of portion size (P<0.0001) but not energy density (P=0.22) on total meal intake by weight. Across the lower- and higher-ED meals, intake was significantly greater when children were served the 150% and 200% portion size conditions compared to the 100% condition. Figure 2B: There were significant and independent effects of portion size and energy density (both P<0.0001) on total meal energy intake. Across the lower- and higher-ED meals, energy intake was significantly greater when children were served the 150% and 200% portion size conditions compared to the 100% condition (both P<0.0001). Across portion size conditions, increasing meal ED by 42% led to an increase in energy intake at the meal (P<0.0001). The effects of portion size and energy density combined to increase meal energy intake by 79% or 175 ± 12 kcal when the children were served the higher-ED meal with the largest portions compared to the lower-ED meal with the smallest portions.

Figure 3

Frequency distribution of liking ratings (faces from left to right represent “super yucky,” “yucky,” “just okay,” “yummy,” and “super yummy”) for the lower- and higher-energy dense versions of a meal consisting of chicken, macaroni and cheese, green vegetable, applesauce, and milk. There was no significant difference in distribution of the ratings between the two versions of the meal, according to ordinal repeated measures logistic regression (p=0.15). ED, energy density.

Figure 4

Effect of parental ratings of child food responsiveness on the relationship between meal portion size and weight of the meal consumed. Scores on the food responsiveness scale indicate the mean frequency of 5 responsive behaviors and range from 1 (never) to 5 (always). Analysis of covariance showed that the slopes of the regression lines for the 150% and 200% portion conditions were significantly greater than the slope for the 100% condition (both P<0.01). Thus, children who were rated as being more responsive to food had larger increases in intake when served the 150% and 200% portions of the meal (compared to the 100% condition) than children who were less responsive.