Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Dec 1.
Published in final edited form as: Appetite. 2021 Jul 8;167:105592. doi: 10.1016/j.appet.2021.105592

Meal composition during an ad libitum buffet meal and longitudinal predictions of weight and percent body fat change: The role of hyper-palatable, energy dense, and ultra-processed foods

Tera L Fazzino a,b,*, James L Dorling c, John W Apolzan c, Corby K Martin c
PMCID: PMC8434988  NIHMSID: NIHMS1723633  PMID: 34245802

Abstract

Background:

Foods that increase obesity risk are ubiquitous in the US food environment. Such foods may be the target of hedonic eating, which may facilitate weight gain and lead to obesity. The study tested whether meal composition during an ad libitum buffet meal was associated with 1-year weight and percent body fat changes among healthy younger adults without obesity. Hyper-palatable foods (HPF) were the study focus; comparisons were conducted with high energy dense (HED) and ultra-processed foods (UPF).

Design:

Younger adults without obesity (N=82; 43% male; mean age 26.8) completed an ad libitum buffet meal and provided body composition measurements at baseline and 1-year follow up. Multiple regression models tested associations between the proportion of the target food consumed (HPF, HED, or UPF) during the ad libitum meal and 1) weight change and 2) percent body fat change. The proportion of HPF was characterized by HPF group, specifically carbohydrate and sodium (CSOD) foods or fat and sodium (FSOD) foods.

Results:

Participants who consumed a greater proportion of CSOD HPF in their ad libitum buffet meals had significantly greater weight change (b=0.354, p=.003) and percent body fat change (b=0.247, p=.036) at 1-year follow up. In contrast, no significant associations were found between the proportion of FSOD HPF, HED, or UPF consumed and anthropometric outcomes (p values =.099-.938).

Conclusions:

Eating a greater proportion of hyper-palatable CSOD foods ad libitum appears to be a pattern of hedonic eating, which may increase an individual’s risk for weight and body fat gain in early adulthood.

Keywords: eating behavior, young adults, energy balance, hedonic eating, obesity, body fat, appetite

1.0. Introduction

Foods that are rewarding to eat are ubiquitous in the US food environment. Such foods may be the target of hedonic eating, which may facilitate weight gain and lead to obesity (1). Several terms have been used to describe such foods in the literature, and various corresponding mechanisms through which they may increase obesity risk have been proposed. The most well-established concept is energy density (calculated as kcal/g) (2). It is argued that because people typically consume the same volume of food per meal, total energy intake per meal may depend on the energy density of the food consumed (2). Greater intake of high energy density (HED) foods may increase total energy intake per meal and obesity risk (2). More recently, the NOVA system defined ultra-processed foods (UPF), which primarily consist of industrially-extracted nutrients and additives (3,4). Due to their extensive processing, UPF may contain unhealthy ingredients (e.g., salt) and have been represented as hyper-palatable and HED (3,4). As a result, they have been hypothesized to lead to overeating and obesity (36).

Both HED and UPF have been described in the literature as having heightened palatability; however, neither is a direct measure of palatability and therefore they may not be appropriate to rely on as proxy measures of palatability. For example, the original hypothesized mechanism of HED foods was clearly identified as energy density (2). However, over time, researchers have posited that the tendency to overconsume HED foods may be exacerbated by their high palatability (79). This notion has been extensively discussed in prominent reviews and studies (7,1014) to the point at which ED has almost become synonymous with palatability. Regarding UPF, the mechanism through which food processing alone may increase obesity risk has been less clearly defined. Instead, UPF have been suggested to be hyper-palatable and HED and to facilitate overeating (4). However, similar to HED, the UPF definition does not directly address palatability. A consequence of using HED and UPF as a proxy definitions for palatability may be that researchers miss identifying foods that have enhanced palatability and may increase obesity risk. For example, many pasta dishes have enhanced palatability, but have low energy density (15). In addition, enhanced palatability may occur in foods that are minimally processed. For example, in a steak cooked with oil and salt, the combination of oil and salt at moderate to high levels would synergistically enhance its palatability. Notably, none of the ingredients are UPF. Thus, overall, because HED and UPF are conceptually distinct from palatability, relying on the terms as proxy measures of palatability lacks precision and may have consequences for identifying a wide variety of foods that may increase obesity risk.

Hyper-palatability in food occurs when key ingredients are combined to create an artificially enhanced palatability experience (15). Recently, we developed a definition of hyperpalatable foods (HPF) using criteria that specify combinations of ingredients that may lead to hyper-palatability: fat, salt, sugar, and carbohydrates (15). The mechanism in hyper-palatability that may increase obesity risk is the combination of key ingredients at moderate to high levels, which may create a highly rewarding eating experience and bypass satiety mechanisms, leading to overeating (15). Thus, the HPF definition directly aligns with the mechanism in enhanced palatability and may yield the most specificity in identifying foods that increase obesity risk.

The next step needed to inform obesity prevention efforts is to test core hypotheses regarding HPF, that HPF may lead to overeating and weight/body fat gain, and to identify eating patterns specific to HPF that may increase an individual’s risk for weight gain and obesity. Thus, the analyses reported herein were designed to test core hypotheses regarding HPF by examining associations between individual-level eating behavior specific to HPF and obesity-related outcomes. In addition, associations between eating behavior specific to HED and UPF and obesity-related outcomes were evaluated for comparison. The study used a sample of younger adults without obesity at baseline, who completed an ad libitum buffet meal and provided weight and body composition measurements at baseline and one year follow up. We hypothesized that: 1) During an ad libitum buffet meal, individuals would consume significantly greater total energy, relative to their energy needs, when the meal contained a greater proportion of HPF; 2) Individuals who consumed an ad libitum buffet meal containing a greater proportion of HPF would gain more weight and percent body fat over time compared to individuals whose meals contain a smaller proportion HPF; and 3) Consumption of a greater proportion of HPF, but not HED or UPF, during the ad libitum meal would predict weight and percent body fat changes at follow up (3a). Finally, the proportion of HPF consumed during the ad libitum meal would have predictive utility for weight and percent body fat change, even when considering contributions from HED or UPF (3b).

2.0. Methods

2.1. Parent Study

The present study conducted secondary analysis of data from a longitudinal study that was originally designed to test biological, environmental, and individual-level factors implicated in weight gain and the development of obesity (clinicaltrials.gov: NCT00945633) (16,17). The study was approved by the Institutional Review Board and all participants provided written informed consent. Participants were recruited from the greater Baton Rouge, Louisiana area and inclusion criteria consisted of the following: 1) aged 20–35 years, 2) BMI < 27.5 kg/m2, and 3) fasting blood glucose < 126 mg/dL. Exclusion criteria included a history of diabetes, obesity, and other medical conditions that might influence health status (detailed in Harrington et al (18)). At baseline, participants completed a buffet-style test meal to quantify intake under ad libitum conditions and completed body composition measurements via dual-energy X-ray absorptiometry (DXA) and doubly labeled water to measure total daily energy expenditure (TDEE). Participants attended a 1-year follow up visit in which all anthropometric measures were reassessed. A total of N=88 participants consented and provided baseline measurements. The vast majority of participants (93%; 82/88) provided measures at baseline and follow-up and were included in analyses.

2.2. Measures

Hyper-palatability.

The definition of HPF, as described in the source publication (15) was used in the study. The HPF definition was developed in accordance with theoretical and empirical evidence that has indicated that the combination of palatability-inducing ingredients (e.g., fat, sugar, sodium, and carbohydrates) present at moderate to high levels may synergistically and artificially enhance a food’s palatability, thereby inducing hyper-palatability (1923). The definition was developed using a data-driven approach that examined commonalities among foods identified as highly palatable and/or difficult to stop eating in experimental laboratory studies and survey studies (15). The resulting HPF definition specified the three following groups of foods as HPF: 1) fat and sodium; 2) fat and sugar; and 3) carbohydrates and sodium (criteria are detailed in the analysis section below). Evidence supports 1) the convergent validity of the definition with fast foods/fried foods and sweets/desserts and 2) the discriminant validity of the definition for raw/fresh foods and foods hypothesized not to be hyperpalatable (15). In addition, the HPF definition has been shown to be distinct from energy density; 49% of foods identified as HPF in the US food system had low ED (<2 kcal/g) (15). Analyses characterized HPF consumption by HPF group.

High energy density.

HED foods were defined as having >2 kcal/g (24). Researchers have generally found associations with HED and 1) dietary intake behavior and 2) obesity-related indices with HED ≥2 kcal/g. More specifically, researchers have reported significant associations between HED at or above 2 kcal/g with poor diet quality in adolescents (25) and adults (24), and positive associations between HED food intake and obesity-related outcomes among adolescents (26) and adults (27,28).

Ultra-processed foods.

The NOVA classification system identified four groups under which foods can be classified based on level of processing (3,4). Group one includes unprocessed or minimally processed foods, group two includes processed culinary ingredients, group three includes processed foods, and group four comprises UPF (3,4). Foods may be classified as UPF based on whether they contain industrial ingredients rarely used in home cooking (e.g., high-fructose corn syrup and hydrogenated fats) and/or additives for the purposes of increasing palatability or visual appeal. UPF may contain ingredients that are unhealthy, such as sugar and salt, and have been hypothesized to lead to unhealthy diets and overeating (3,4). A growing body of evidence indicates that consumption of UPF is associated with poorer diet quality (4), as well as obesity; however, it remains unclear whether food processing or the nutrient contents of the foods may be driving the observed associations (29).

2.3. Procedures

Ad libitum buffet meal.

In a controlled laboratory setting, participants were provided a single ad libitum buffet meal between 11am and 12pm. Participants consumed the meal following at 10 hour fast and were instructed to avoid alcohol, caffeine, and strenuous physical activity in the 24 hours preceding the meal. The meal comprised 20 well-liked food and beverage items, including snack items (e.g., chips, candy), as well as meal-based items (e.g., chicken). Supplemental Table 1 provides the nutrients of each food served in the meal. The meal was provided in isolation and participants consumed as much or as little as they desired. To limit distractions that may modulate food and drink consumption (30), the use of cell phones, books, and computers were not permitted during the meal. All foods were weighed pre and post meal, and subsequently, energy intake was determined from manufacturer data and the USDA Food and Nutrient Database (31).

Body composition and total daily energy expenditure.

DXA was performed with a Hologics Quantitative Digital Radiography (QDR) 4500A whole-body scanner. The scans were analyzed with QDR for Windows V11.1. The coefficients of variation (CV) for the body composition measurement of lean mass, fat mass and percentage body fat were 0.8%, 1.6%, and 1.7%, respectively. Fat mass, fat-free mass, and percentage of fat were calculated.

TDEE was measured over 14 days using the doubly labeled water (DLW) procedure, as detailed previously (18).

2.4. Data Processing

A total of 20 items were offered during the buffet meal, consisting of 19 food items and one calorie-containing beverage. The beverage item was removed before analysis, because 1) the HPF definition does not apply to liquids (15), and 2) liquids substantially decrease ED estimates and may weaken ED associations with obesity-related outcomes (32). Thus, to maintain consistency across constructs examined, the 19 items were analyzed across all three definitions. The items were analyzed to determine which items met criteria for hyper-palatability, high energy density, and ultra-processing, as detailed below.

Application of HPF definition.

First, percent kcal from fat, sugar, and carbohydrates (after subtracting sugar and fiber) was calculated per procedures outlined in Fazzino et al. (15). Percent sodium per food weight in grams was calculated as sodium in grams/food weight in grams per serving (15). Buffet foods meeting HPF criteria for each group were identified using the HPF criteria: 1) fat and sodium group (FSOD; >25% kcal fat and ≥0.30% sodium); 2) fat and sugar group (FS; >20% fat and >20% sugar); and 3) carbohydrates and sodium group (CSOD; >40% carbohydrate and ≥0.20% sodium). For analysis, the proportion of HPF items from each group (FSOD, FS, or CSOD) that participants consumed during the ad libitum meal was calculated as the number of HPF items consumed from each group divided by the total number of buffet foods consumed. HPF variables included in analyses were the proportion of FSOD and CSOD foods consumed during the meal. Because only one food item met criteria for the FS group, the proportion of FS foods consumed during the buffet meal was not included in analyses.

Application of HED definition.

Foods with >2 kcal/g were identified as HED (24). For analysis, the proportion of HED foods consumed during the ad libitum buffet meal was calculated as the number of HED foods consumed divided by the total number of buffet foods consumed.

Application of UPF definition.

Food items were classified as UPF based on whether they contained industrial ingredients rarely used in home cooking (e.g., hydrogenated oils and high-fructose corn syrup) and/or additives for the purposes of increasing palatability or visual appeal (3,4). IS (acknowledgements) and JWA independently categorized each lunch buffet food item into a NOVA classification. For analysis, the proportion of UPF items consumed by participants during the ad libitum meal was calculated as the number of UPF items consumed divided by the total number of buffet foods consumed.

Food item classifications.

Among the buffet foods, 53% (10/19) of items were HPF. Specifically, six foods met criteria for the FSOD group, one met criteria for the FS group, and four met criteria for the CSOD group. Sixty-three percent of food items (12/19) were HED, and 63% were UPF (12/19). Despite some overlap across food constructs, five food items were different across HPF and HED constructs, and eight items were different across HPF and UPF. Buffet foods are presented in Supplemental Table 2, along with their classifications as HPF, HED, and/or UPF.

2.5. Data Analysis

The variables used in analyses were based on the composition of participants’ ad libitum buffet meals, specifically the proportion of the target food (CSOD, FSOD, HED, or UPF) consumed during the meal. The variable was chosen to characterize individual eating behavior under ad libitum conditions. Furthermore, ad libitum buffet meal energy (kcal) intake was not predicted to characterize eating behavior, because energy intake within a meal is largely driven by fat free mass and resting metabolic rate (3336), among other factors that can affect the amount of food consumed within a meal (e.g., gut-hormone responses, mood, social norms, etc.) (37). Thus, focusing on the composition of participants’ meals facilitated the characterization of individual-level eating patterns for analysis, while being less directly related to energy expenditure and intake.

Hypothesis 1:

The first two analytic models were designed to test the hypothesis that, during an ad libitum buffet meal, individuals would consume significantly greater total energy, relative to their energy needs, when the meal contained a greater proportion of HPF (CSOD and FSOD). A path analysis model was constructed with percent CSOD or FSOD foods consumed during the ad libitum meal regressed on total energy (kcal) intake during the meal, while accounting for energy needs via TDEE, sex, and age. Path analysis was used to specify correlations among TDEE and the other predictor variables that had moderate to strong correlations. Comparison path models were constructed with variables for the proportion of HED and UPF consumed in the ad libitum meals and are presented in the supplementary information section.

Hypothesis 2:

The second analysis tested the hypothesis that individuals who consume an ad libitum meal containing a greater proportion of HPF (FSOD and CSOD) would have significantly greater weight and percent body fat changes at 1-year follow up. Two regression models were constructed to test whether the proportion of HPF consumed from FSOD and CSOD foods during the ad libitum meal predicted 1) weight change and 2) percent body fat change at 1-year follow up. The association between proportion of FSOD and CSOD foods consumed was small in magnitude (r= −0.31, p= .004), suggesting that the two variables could be included in the same models without introducing multicollinearity. Estimated marginal means were reported to specify the expected weight and percent body fat gain at the sample mean and >1 standard deviation (SD) above the mean.

Hypothesis 3:

The final set of analyses tested the hypothesis that consumption of a greater proportion of HPF, but not HED or UPF, during the ad libitum meal would be associated with weight and percent body fat changes at follow up (3a), and that the proportion of HPF consumed during the meal would have predictive utility even when considering relative contributions from HED and UPF (3b). Two individual regression models were constructed with the proportion of HED or UPF consumed during the test meal predicting weight or percent body fat change at follow up (3a). To test relative effects, additional models included the proportion of HPF consumed from FSOD and CSOD foods and the proportion of HED (comparison model 1) or UPF (comparison model 2) during the ad libitum meal as predicting 1-year weight change or body fat change. The associations between the proportion of FSOD and CSOD foods consumed with the proportion of HED foods (FSOD: r= 0.26, p= .016; CSOD: r= 0.26, p= .021) and proportion of UPF (FSOD: r= −0.21, p= .049; CSOD: r= 0.18, p= .089) were small in magnitude.

Analytic models for weight change and percent body fat change outcomes were run without and with covariates and are reported in the results. Final analytic models were the models that included covariates. Covariates included in the models consisted of baseline weight or percent body fat, sex, and total kcal intake during the buffet meal. Race and age were tested as covariates in the models, but were not included in the final models because they were not significantly associated with the outcomes and did not substantially impact the beta estimates of the main predictor variables of interest. Unstandardized beta estimates are presented in the text of the results to allow for interpretation within a variable’s original metric. Standardized beta estimates are presented in the tables to facilitate comparisons of effect sizes across food constructs. Effects in the models are reported with covariates held constant.

Additional analyses were conducted to test whether the intake of single nutrients during the ad libitum meal, specifically percent kcal from fat, carbohydrates, or protein, was associated with weight or percent body fat changes. In addition, analyses were conducted to determine whether intake of nutrients from HPF, HED, or UPF, specifically percent kcal from fat, carbohydrates, or protein, was associated with anthropometric outcomes.

2.6. Missing Data

Participants who did not provide ad libitum meal data at baseline (n=1) or follow up anthropometric data (n=5) were excluded from analyses. One participant was missing the percent body fat measure at follow up, but did provide weight at follow up and was therefore included in analysis of weight change only. All other participants provided data on all variables required for analysis.

3.0. Results

3.1. Participant Characteristics

Table 1 provides demographic characteristics of all participants (N=82). Participants consumed a mean of 832.3 total kcal (SD= 339.5) during the ad libitum meal. The proportion of HPF, HED, and UPF eaten at the ad libitum meal is presented in Table 2, and nutrient intake data for of the three food constructs is presented in Supplemental Table 3.

Table 1.

Demographic Characteristics of Participants (N=82)

Mean (SD) or % (n/N)

Age 26.8 (4.6)

Sex (% male) 43% (35/82)

Race
White 77% (63/82)
Black/African American 15% (12/82)
American Indian/Alaskan Native 1% (1/82)
Other 6% (5/82)

Ethnicity (% Hispanic/Latino) 6% (5/82)

Weight (kg) 67.8 (10.8)

Body mass index 22.9 (2.4)

Percent body fat 23.5 (7.8)

Total daily energy expenditure 2583.9 (593.1)

Table 2.

Descriptive Statistics for Proportion of Target Foods Consumed during the Ad Libitum Meal

Variable Proportion of food consumed in the ad libitum meala Mean (SD)

HPF by group:
Fat and sodium group 29.1% (8.1)

Carbohydrate and sodium group 20.7% (7.9)

High energy density foods 58.5% (9.4)

Ultra-processed foods 64.2% (9.5)

Note: HPF= hyper-palatable foods.

a

Represents the proportion of the target food consumed during the ad libitum meal calculated as n target foods consumed/N foods consumed during the meal

At 1-year follow up, mean weight change for the sample was 0.94 kg (SD= 3.52; R= −13.8 to 10.5) and mean percent body fat change was 1.1% (SD= 2.2; R= −7.9 to 5.3).

3.2. Model Results

Hypothesis 1:

Results of the path analysis model indicated that total kcal intake during the ad libitum buffet meal was positively associated with proportion of CSOD foods consumed during the meal, when accounting for TDEE, sex, and age (beta= 0.005; p = .028; Confidence Interval (CI) = 0.001–0.009). In contrast, total kcal intake during the meal was not significantly associated with proportion of FSOD foods consumed during the meal, when accounting for covariates (b= −0.001 p = .542; CI= −0.006– 0.003).

Comparison analyses specifying the proportion of HED and UPF consumed during the ad libitum meal indicated that greater energy intake was not significantly associated with the proportion of HED or UPF consumed during the meal (see Supplementary Information section).

Hypothesis 2:

Results of the regression models revealed that the proportion of CSOD foods consumed during the ad libitum meal was significantly associated with 1-year weight change (b= 0.140, p= .006, CI= 0.041–0.238) and percent body fat change (b= 0.070, p= .024, CI= 0.004–0.135). For every one percentage increase in the proportion of CSOD foods consumed, participants gained an average of 0.14 kg and 0.07% body fat at the 1-year follow up. Estimated marginal means indicated that participants who consumed the mean proportionate CSOD foods in the ad libitum meal gained 0.92 kg and 1.08% body fat at 1-year follow up, and participants who consumed >1SD above the mean gained 2.12 kg and 1.68% body fat. The significant associations between proportion of CSOD foods consumed with weight and percent body fat change remained when accounting for baseline weight or percent body fat, sex, and total kcal intake during the ad libitum meal (Table 3). The proportion of FSOD foods consumed during the meal was not significantly associated with weight change or percent body fat change in the base models (weight change: b= −006, p= .907, CI= −0.104–0.093; percent body fat change: b= 0.004, p= .938, CI= −0.099–0.108), or in the models with covariates (Table 3).

Table 3.

Associations between the Proportion of Target Foods Consumed during the Ad Libitum Meal and One-Year Anthropometric Changes

DV: Weight change
Standardized Beta Standard Error P value Confidence Interval
Carbohydrate and sodium HPFa 0.354 0.052 .003 0.251 to 0.458
Fat and sodium HPFa 0.009 0.051 .937 −0.094 to 0.112
High energy dense foods −0.068 0.042 .552 −0.153 to 0.016
Ultra-processed foods 0.090 0.043 .442 −0.004 to 0.176
DV: Percent body fat change
Carbohydrate and sodium HPFa 0.247 0.032 .036 0.182 to 0.312
Fat and sodium HPFa −0.011 0.033 .920 −0.078 to 0.054
High energy dense foods −0.034 0.027 .769 −0.088 to 0.020
Ultra-processed foods 0.176 0.026 .130 −0.122 to 0.230

Note: HPF= hyper-palatable foods.

Models were run independently for each food construct. Models controlled for covariates: sex, baseline weight or percent body fat, and total kcal consumed during buffet meal.

a

Included in the same model because both are types of HPF

Hypothesis 3:

Results of the regression model with HED indicated that the proportion of HED foods consumed during the ad libitum meal was not significantly associated with weight change (b = −0.027, p= .526, CI= −0.110–0.057) or percent body fat change (b= −0.007, p= .798, CI= −0.060–0.046). In addition, proportion of HED foods consumed was not significantly associated with weight change or body fat change when accounting for covariates (Table 3). Similarly, results indicated that the proportion of UPF items consumed during the meal was not significantly associated with weight change (b = 0.036; p= .382, CI= −0.046–0.119) or percent body fat change (b= 0.043, p= .100, CI= −0.008–0.095) in the base models, or when accounting for covariates (Table 3).

Standardized estimates from the combined models with HPF and constructs of HED or UPF are presented in Tables 45. Findings from the multiple regression model that included HPF (CSOD and FSOD) and HED indicated that the proportion of CSOD foods consumed during the ad libitum meal significantly predicted weight change and percent body fat change, whereas FSOD and HED did not (Table 4). In addition, findings from the model that included HPF (CSOD and FSOD) and UPF indicated that the proportion of CSOD foods consumed during the buffet meal significantly predicted weight and percent body fat change, whereas FSOD and UPF did not, when accounting for covariates (Table 5).

Table 4.

Combined Models with Proportion of Hyper-Palatable Foods and High Energy Dense Foods Consumed Predicting One-Year Anthropometric Changes

DV: Weight change
Standardized Beta Standard Error P value Confidence Interval
Carbohydrate and sodium HPF 0.434 0.055 <.001 0.323 to 0.545
Fat and sodium HPF 0.099 0.056 .435 −0.013 to 0.213
High energy dense foods −0.204 0.045 .099 −0.296 to 0.113
DV: Percent body fat change
Carbohydrate and sodium HPF 0.290 0.035 .023 0.219 to 0.360
Fat and sodium HPF 0.037 0.036 .772 −0.035 to 0.110
High energy dense foods −0.118 0.030 .355 −0.179 to 0.058

Note: HPF= hyper-palatable foods. Models controlled for covariates: sex, baseline weight or percent body fat, and total kcal consumed during buffet meal.

Table 5.

Combined Models with Proportion of Hyper-Palatable Foods and Ultra-Processed Foods Consumed Predicting One-Year Anthropometric Changes

DV: Weight change
Standardized Beta Standard Error P value Confidence Interval
Carbohydrate and sodium HPF 0.350 0.052 .004 0.245 to 0.456
Fat and sodium HPF 0.014 0.053 .903 −0.091 to 0.120
Ultra-processed foods 0.030 0.042 .793 −0.054 to 0.115
DV: percent body fat change
Carbohydrate and sodium HPF 0.233 0.032 .048 0.168 to 0.299
Fat and sodium HPF 0.016 0.033 .888 −0.050 to 0.084
Ultra-processed foods 0.145 0.027 .219 −0.090 to 0.199

Note: HPF= hyper-palatable foods. Models controlled for covariates: sex, baseline weight or percent body fat, and total kcal consumed during buffet meal.

HPF= hyper-palatable foods.

Additional analyses indicated that there were no significant associations between percent kcal consumed from individual nutrients (fat, carbohydrates, protein) and weight change or percent body fat change, suggesting that consumption of individual-level nutrients was not driving the observed anthropometric changes (Supplementary Information). Furthermore, there were no significant associations between percent kcal consumed from individual nutrients (fat, carbohydrates, protein) in HPF, HED, or UPF with weight change or percent body fat change, suggesting that individual nutrients within the food types were not driving the observed anthropometric changes (Supplemental Table 4).

4.0. Discussion

Recently, a quantitative definition of hyper-palatable foods was developed to address limitations in the field regarding the lack of a standardized definition and the use of proxy terms such as high energy density and ultra-processed foods when examining the effects of palatable food intake on obesity outcomes (15). The current study tested core hypotheses regarding HPF and examined the impact of eating behavior specific to HPF during a single ad libitum buffet meal on 1) overeating, and 2) 1-year naturalistic changes in weight and percent body fat among a sample of healthy adults without obesity. The study also tested the relative contribution of eating behavior specific to HPF when compared to eating behavior specific to HED and UPF. In support of Hypothesis 1, findings revealed that during an ad libitum buffet meal, individuals consumed significantly greater total energy during the meal when their meals contained a greater proportion of CSOD HPF, relative to individuals whose meals contained a smaller proportion of CSOD HPF. Effects were present when accounting for energy needs (TDEE), indicating that participants ate more in relation to their physiological energy requirements when they selected and ate more CSOD HPF. However, in contrast to our hypothesis, individuals did not consume significantly greater total energy when their meals contained a greater proportion of fat and sodium HPF. A similar pattern of results was found for weight change and percent body fat change outcomes. Individuals who consumed an ad libitum meal that contained a greater proportion of CSOD HPF had greater weight and percent body fat changes at 1-year follow up, with small to moderate effect sizes. However, consumption of meals with greater FSOD HPF was not associated with anthropometric changes, and neither was proportion of HED or UPF consumed. Finally, in support of our hypothesis, the significant associations between proportion of CSOD foods consumed and anthropometric outcomes were observed even when accounting for the proportion of HED or UPF consumed. Thus, the findings highlight CSOD foods as a potential contributor to overeating and weight and percent body fat gain at 1-year follow up. Findings indicate that CSOD foods appear to be the target of hedonic or reward-driven eating, which may lead to passive overconsumption, excess energy intake relative to energy needs, and anthropometric changes over time.

The study findings may have important implications for obesity prevention efforts. The study identified a novel predictor of naturalistic weight and percent body fat change among healthy younger adults without obesity. There are limited predictors of naturalistic weight gain in adulthood, particularly among healthy younger adults. Furthermore, most reliable predictors are sociodemographic or environment-based and are thus not easily modifiable (e.g., sex, socioeconomic status, food environment) (3840). Some modifiable predictors of weight gain have been identified (e.g., disinhibited and restricted eating, weight suppression behavior, depression, etc) however most lack consistency in their predictive utility across studies (4144); thus new modifiable predictors of weight gain are needed. Our findings identify a predictor of naturalistic weight change that may be a marker of hedonic eating, and thus may be a modifiable risk factor for weight and body fat change in early adulthood. Furthermore, the magnitude of the observed effects was significant and meaningful within the context of the sample. Individuals who consumed >1 SD above the mean in proportion of CSOD foods during the buffet meal were predicted to gain >2 kg and 1.7% body fat over the year, which could move individuals from a healthy BMI to an overweight BMI, or an overweight BMI to an obese BMI. Given that weight gain from young adulthood to middle adulthood occurs gradually and may increase the risk for many chronic health conditions (45), obesity prevention efforts may consider targeting individuals who engage in CSOD-specific eating patterns to prevent weight and body fat gain during young adulthood, and to prevent potential chronic health conditions later in life.

Regarding the HPF construct, study results partially support the predictive utility of the quantitative HPF definition food groups (FSOD, FS, and CSOD) for obesity-related outcomes. Our finding that the proportion of CSOD HPF consumed was significantly associated with overeating is broadly consistent with results from three prior experimental studies conducted in healthy adults or children that found that elevated sodium in a carbohydrate-dense pasta meal increased consumption by approximately 10%, suggesting that enhanced sodium content in carbohydrate-dense food may increase intake within a meal (23,46,47). Thus, our results bolster the existing literature in identifying the synergistic effects of carbohydrates and sodium on total energy intake within a meal. Our null findings regarding FSOD foods and overeating are in contrast to one initial experimental study that found that elevated sodium may promote passive overconsumption of fat among healthy adults (23). However, one important distinction between the current study and the aforementioned literature is that the present study evaluated the behavioral tendency to consume types of HPF (FSOD and CSOD) when freely available, whereas the prior studies assessed the effect of a specific FSOD or CSOD food’s composition (e.g., sodium level) on intake of that food when served to participants. Thus, our study found that the behavioral tendency to consume a greater proportion of CSOD foods in a buffet meal was predictive of weight and body fat change, whereas the behavioral tendency to consume a greater proportion of FSOD foods in a buffet meal was not. It may be that individuals who eat more CSOD foods in a buffet meal may take longer to feel satiated compared to individuals who consume more FSOD foods, which may facilitate excess kcal intake within a meal. In the study, FSOD foods contained substantially higher fat and protein compared to CSOD foods, which may have facilitated participants in feeling satiated sooner than those consuming more CSOD foods. While protein or fat alone did not appear to drive the observed effects (as revealed in the supplemental analyses), there may be relative differences across types of HPF in initiating satiation based on their macronutrient contents. However, it could also be that the FSOD criteria do not adequately capture foods with fat and sodium at a threshold needed to predict weight gain. Overall, future research is needed to distinguish the degree to which behavioral approach toward HPF and the composition of types of HPF may drive overconsumption and facilitate weight gain.

Findings from the study also highlight the distinction between HPF and constructs of HED and UPF, and demonstrate the importance of using a definition specific to hyper-palatability when examining associations between food-specific eating patterns and obesity-related outcomes. The results reinforce the premise that the definition of HPF is distinct from HED and UPF, as indicated by the differences in buffet foods that met criteria for each definition, as well as the small to modest correlation sizes between the constructs. In addition, using the HPF term specific to CSOD facilitated the identification of individual-level eating behavior under ad libitum conditions that may predict anthropometric changes. Specifically, eating meals containing a greater proportion of CSOD foods may be a risk factor for longitudinal changes in weight and percent body fat. Importantly, the proportion of HED and UPF consumed, both of which have been used as proxy constructs for palatability in the literature, were not predictive of weight or percent body fat changes. Overall, results indicate that using a definition specific to hyper-palatability may be necessary to identify the role of palatable food-specific eating behavior in obesity-related outcomes, and that the definition is distinct from constructs of HED and UPF.

The study had several limitations. First, the buffet meal contained a small number of foods from each food construct examined (four to twelve foods). The distinctness of the meal may limit generalizability to other settings, such as larger buffets. However, the smaller number of foods may have improved sensitivity for detecting the effects observed in the study. Relatedly, the ad libitum buffet meal did not contain enough foods that met FS HPF criteria and we therefore could not examine associations between FS HPF with study outcomes; thus future research is needed. In addition, the study conducted a single test meal and future research should expand to replicating observed effects across multiple meals. Nonetheless, the ability to predict anthropometric changes over one year with a single test meal is noteworthy. Also, the majority of the sample was non-Hispanic white and it is unclear to what degree the results would generalize to individuals who are racial or ethnic minorities. Furthermore, though we believe that research like ours is needed to identify the causes of weight gain prior to the onset of obesity, the sample was comprised of healthy younger adults and further research is needed to determine whether the findings replicate in clinical samples with obesity. Finally, the study required that participants provided data at baseline and one year follow up to be included in analyses. However, only six participants had missing follow up data; thus, the vast majority of participants (93%) were included in analyses.

Strengths of the study included the use of standardized definitions for HPF, HED, and UPF constructs, as well as comparisons of the relative predictive utility of the constructs for anthropometric outcomes. In addition, the study utilized sophisticated measuring techniques to assess body composition and TDEE, and a 1-year follow up period was used to observe weight and body composition changes. Importantly, body mass change during that year was naturalistic, as participants were not intervened upon. Finally, the sample was comprised of healthy younger adults at baseline, who had maintained their body weight below the obesity threshold upon study entry. Lastly, an eating pattern was identified from a single test meal that was sensitive enough to predict longitudinal weight and body fat changes among healthy younger adults.

Conclusions and Implications

Our findings indicate that eating more hyper-palatable CSOD foods may result in overeating (in relation to energy requirements), which may increase an individual’s risk for excess energy intake and weight/body fat gain over time. Consequently, consumption of more hyper-palatable CSOD foods may be a form of hedonic eating, where eating behavior is driven by the palatability and other sensory characteristics of the food rather than energy needs. Notably, effects were observed in a sample of younger adults without obesity at baseline. Future obesity prevention efforts may consider targeting individuals who engage in CSOD-specific eating patterns to prevent weight and body fat gain. In addition, results highlight the distinction between constructs of HPF, HED, and UPF, and emphasize the necessity in using a term specific to hyper-palatability instead of proxy constructs such as HED or UPF when testing associations between palatable food-specific eating patterns and obesity-related outcomes. Thus, when examining associations between food-specific eating behavior and obesity outcomes, researchers should consider using the food construct that most closely reflects a study’s target food type and corresponding mechanism of action (e.g., energy density, food processing, or hyper-palatability).

Supplementary Material

1

Acknowledgements:

We thank the participants for their time and effort. Isabelle Schexnayder MS, RDN, LDN assisted with NOVA classification of food items. Kaitlyn Rohde, BS, assisted with the classification of food items as hyper-palatable foods.

InSight research group

Peter T. Katzmarzyk, Ph.D.; Eric Ravussin, Ph.D.; Steven R. Smith, M.D; Sudip Bajpeyi, Ph.D.; Claude Bouchard, Ph.D.; Stephanie Broyles, Ph.D.; Catherine Champagne, Ph.D.; Conrad Earnest, Ph.D. (deceased); Alok Gupta, M.D; William Johnson, Ph.D.; Corby Martin, Ph.D.; Robert Newton, Ph.D.; Tuomo Rankinen, Ph.D.; Leanne Redman, Ph.D.; Jennifer Rood, Ph.D.; Yourka Tchoukalova, M.D, Ph.D.; Catrine Tudor-Locke, Ph.D.

NCT Registry: NCT00945633

This work was created in the performance of a Specific Co-operative Agreement with the U.S. Department of Agriculture. The Government of the United States has a royalty-free government purpose license to use, duplicate or disclose the work, in whole or in part and in any manner, and to have or permit others to do so, for government purposes. This work was partially supported by a NORC Center Grant P30 DK072476 entitled “Nutrition and Metabolic Health through the Lifespan” sponsored by NIDDK, and NIGMS grant U54 GM104940, which funds the Louisiana Clinical and Translational Science Center. JLD is supported by the American Heart Association Grant # 20POST35210907.

Abbreviations:

HPF

hyper-palatable foods

CSOD

carbohydrate and sodium hyper-palatable foods

FSOD

fat and sodium hyper-palatable foods

HED

high energy density

UPF

ultra-processed foods

TDEE

total daily energy expenditure

Footnotes

Conflict of interest statement: The authors declare no conflicts of interest.

Declarations of Interest

None.

Ethical Statement

The study consisted of secondary analysis of deidentified participant data. The study was reviewed by an independent institutional review board at the University of Kansas (reference number: STUDY00145031) on December 5, 2019 and was determined to be ‘not human subjects research.’ For the original study, an institutional review board at Pennington Biomedical Research Center, Louisiana State University System approved the study, and all participants provided informed consent prior to participating.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Swinburn BA, Sacks G, Hall KD, McPherson K, Finegood DT, Moodie ML, Gortmaker SL. The global obesity pandemic: shaped by global drivers and local environments. The Lancet 2011;378:804–14. [DOI] [PubMed] [Google Scholar]
  • 2.Poppitt SD, Prentice AM. Energy Density and its Role in the Control of Food Intake: Evidence from Metabolic and Community Studies. Appetite 1996;26:153–74. [DOI] [PubMed] [Google Scholar]
  • 3.Monteiro CA, Levy RB, Claro RM, de Castro IRR, Cannon G. A new classification of foods based on the extent and purpose of their processing. Cad Saúde Pública 2010;26:2039–49. [DOI] [PubMed] [Google Scholar]
  • 4.Monteiro CA, Cannon G, Moubarac J-C, Levy RB, Louzada MLC, Jaime PC. The UN Decade of Nutrition, the NOVA food classification and the trouble with ultra-processing. Public Health Nutr 2018;21:5–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Beslay M, Srour B, Méjean C, Allès B, Fiolet T, Debras C, Chazelas E, Deschasaux M, Wendeu-Foyet MG, Hercberg S, et al. Ultra-processed food intake in association with BMI change and risk of overweight and obesity: A prospective analysis of the French NutriNet-Santé cohort. PLoS Med 2020;17:e1003256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Hall KD, Ayuketah A, Brychta R, Cai H, Cassimatis T, Chen KY, Chung ST, Costa E, Courville A, Darcey V, et al. Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake. Cell Metab [Internet] 2019. [cited 2019 May 16];0. Available from: https://www.cell.com/cell-metabolism/abstract/S1550-4131(19)30248-7 [DOI] [PMC free article] [PubMed]
  • 7.Crino M, Sacks G, Vandevijvere S, Swinburn B, Neal B. The Influence on Population Weight Gain and Obesity of the Macronutrient Composition and Energy Density of the Food Supply. Curr Obes Rep 2015;4:1–10. [DOI] [PubMed] [Google Scholar]
  • 8.Kearney J Food consumption trends and drivers. Philos Trans R Soc B Biol Sci 2010;365:2793–807. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Drewnowski A Energy density, palatability, and satiety: implications for weight control. Nutr Rev 1998;56:347–53. [DOI] [PubMed] [Google Scholar]
  • 10.Berthoud H-R, Lenard NR, Shin AC. Food reward, hyperphagia, and obesity. Am J Physiol - Regul Integr Comp Physiol 2011;300:R1266–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Morris MJ, Beilharz JE, Maniam J, Reichelt AC, Westbrook RF. Why is obesity such a problem in the 21st century? The intersection of palatable food, cues and reward pathways, stress, and cognition. Neurosci Biobehav Rev 2015;58:36–45. [DOI] [PubMed] [Google Scholar]
  • 12.Burger KS, Stice E. Variability in Reward Responsivity and Obesity: Evidence from Brain Imaging Studies. Curr Drug Abuse Rev 2011;4:182–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Mendoza JA, Drewnowski A, Christakis DA. Dietary energy density is associated with obesity and the metabolic syndrome in U.S. adults. Diabetes Care 2007;30:974–9. [DOI] [PubMed] [Google Scholar]
  • 14.Howarth NC, Murphy SP, Wilkens LR, Hankin JH, Kolonel LN. Dietary Energy Density Is Associated with Overweight Status among 5 Ethnic Groups in the Multiethnic Cohort Study. J Nutr Oxford Academic; 2006;136:2243–8. [DOI] [PubMed] [Google Scholar]
  • 15.Fazzino T, Rohde K, Sullivan D. Hyper-palatable foods: Development of a quantitative definition and application to the United States food system database. Obesity 2019; [DOI] [PubMed] [Google Scholar]
  • 16.Staiano AE, Marker AM, Martin CK, Katzmarzyk PT. Physical activity, mental health, and weight gain in a longitudinal observational cohort of nonobese young adults. Obesity 2016;24:1969–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Staiano AE, Martin CK, Champagne CM, Rood JC, Katzmarzyk PT. Sedentary time, physical activity, and adiposity in a longitudinal cohort of nonobese young adults. Am J Clin Nutr 2018;108:946–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Harrington DM, Martin CK, Ravussin E, Katzmarzyk PT. Activity related energy expenditure, appetite and energy intake: potential implications for weight management. Appetite 2013;67:1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.de Macedo IC, de Freitas JS, da Silva Torres IL. The Influence of Palatable Diets in Reward System Activation: A Mini Review. Adv Pharmacol Sci 2016;2016:e7238679. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.DiFeliceantonio AG, Coppin G, Rigoux L, Edwin Thanarajah S, Dagher A, Tittgemeyer M, Small DM. Supra-Additive Effects of Combining Fat and Carbohydrate on Food Reward. Cell Metab 2018;28:33–44.e3. [DOI] [PubMed] [Google Scholar]
  • 21.Stinson EJ, Piaggi P, Ibrahim M, Venti C, Krakoff J, Votruba SB. High fat and sugar consumption during ad libitum intake predicts weight gain. Obes Silver Spring Md 2018;26:689–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Drewnowski A, Greenwood MR. Cream and sugar: human preferences for high-fat foods. Physiol Behav 1983;30:629–33. [DOI] [PubMed] [Google Scholar]
  • 23.Bolhuis DP, Costanzo A, Newman LP, Keast RS. Salt Promotes Passive Overconsumption of Dietary Fat in Humans. J Nutr 2016;jn226365. [DOI] [PubMed] [Google Scholar]
  • 24.Ledikwe JH, Blanck HM, Khan LK, Serdula MK, Seymour JD, Tohill BC, Rolls BJ. Low-Energy-Density Diets Are Associated with High Diet Quality in Adults in the United States. J Am Diet Assoc 2006;106:1172–80. [DOI] [PubMed] [Google Scholar]
  • 25.O’Connor L, Walton J, Flynn A. Dietary energy density and its association with the nutritional quality of the diet of children and teenagers. J Nutr Sci [Internet] 2013. [cited 2020 Apr 8];2. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4153288/ [DOI] [PMC free article] [PubMed]
  • 26.Arango-Angarita A, Shamah-Levy T, Rodríguez-Ramírez S. Dietary energy density is associated with body mass index-for-age in Mexican adolescents. Matern Child Nutr 2019;15:e12664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Vergnaud A-C, Estaquio C, Czernichow S, Péneau S, Hercberg S, Galan P, Bertrais S. Energy density and 6-year anthropometric changes in a middle-aged adult cohort. Br J Nutr Cambridge University Press; 2009;102:302–9. [DOI] [PubMed] [Google Scholar]
  • 28.Rouhani MH, Haghighatdoost F, Surkan PJ, Azadbakht L. Associations between dietary energy density and obesity: A systematic review and meta-analysis of observational studies. Nutrition 2016;32:1037–47. [DOI] [PubMed] [Google Scholar]
  • 29.Poti JM, Braga B, Qin B. Ultra-processed Food Intake and Obesity: What Really Matters for Health—Processing or Nutrient Content? Curr Obes Rep 2017;6:420–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Ogden J, Coop N, Cousins C, Crump R, Field L, Hughes S, Woodger N. Distraction, the desire to eat and food intake. Towards an expanded model of mindless eating. Appetite 2013;62:119–26. [DOI] [PubMed] [Google Scholar]
  • 31.Agricultural Research Service Food Surveys Research Group. USDA Food and Nutrient Database for Dietary Studies. Beltsville, MD: USDA; 2008. Report No.: 3.0. [Google Scholar]
  • 32.Ledikwe JH, Blanck HM, Khan LK, Serdula MK, Seymour JD, Tohill BC, Rolls BJ. Dietary Energy Density Determined by Eight Calculation Methods in a Nationally Representative United States Population. J Nutr Oxford Academic; 2005;135:273–8. [DOI] [PubMed] [Google Scholar]
  • 33.Blundell JE, Caudwell P, Gibbons C, Hopkins M, Näslund E, King NA, Finlayson G. Body composition and appetite: fat-free mass (but not fat mass or BMI) is positively associated with self-determined meal size and daily energy intake in humans. Br J Nutr Cambridge University Press; 2012;107:445–9. [DOI] [PubMed] [Google Scholar]
  • 34.Blundell JE, Caudwell P, Gibbons C, Hopkins M, Naslund E, King N, Finlayson G. Role of resting metabolic rate and energy expenditure in hunger and appetite control: a new formulation. Dis Model Mech The Company of Biologists Ltd; 2012;5:608–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Caudwell P, Finlayson G, Gibbons C, Hopkins M, King N, Näslund E, Blundell JE. Resting metabolic rate is associated with hunger, self-determined meal size, and daily energy intake and may represent a marker for appetite. Am J Clin Nutr Oxford Academic; 2013;97:7–14. [DOI] [PubMed] [Google Scholar]
  • 36.Blundell J Biology of the drive to eat. Implications for understanding human appetite and obesity. Ernahrungs Umsch 2018;65:126–33. [Google Scholar]
  • 37.Blundell J, de Graaf C, Hulshof T, Jebb S, Livingstone B, Lluch A, Mela D, Salah S, Schuring E, van der Knaap H, et al. APPETITE CONTROL: METHODOLOGICAL ASPECTS OF THE EVALUATION OF FOODS. Obes Rev Off J Int Assoc Study Obes 2010;11:251–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Hillemeier MM, Weisman CS, Chuang C, Downs DS, McCall-Hosenfeld J, Camacho F. Transition to Overweight or Obesity Among Women of Reproductive Age. J Womens Health Mary Ann Liebert, Inc., publishers; 2011;20:703–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Bowen DJ, Jabson JM, Barrington WE, Littman AJ, Patrick DL, Moudon AV, Albano D, Beresford SAA. Environmental and Individual Predictors of Healthy Dietary Behaviors in a Sample of Middle Aged Hispanic and Caucasian Women. Int J Environ Res Public Health Multidisciplinary Digital Publishing Institute; 2018;15:2277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Lakerveld J, Mackenbach J. The Upstream Determinants of Adult Obesity. Obes Facts Karger Publishers; 2017;10:216–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Lowe MR, Marti CN, Lesser EL, Stice E. Weight suppression uniquely predicts body fat gain in first-year female college students. Eat Behav 2019;32:60–4. [DOI] [PubMed] [Google Scholar]
  • 42.Manzato E, Bolognesi A, Simoni M, Cuzzolaro M. Risk factors for weight gain: a longitudinal study in non-weight loss treatment-seeking overweight adults. Eat Weight Disord - Stud Anorex Bulim Obes 2015;20:371–8. [DOI] [PubMed] [Google Scholar]
  • 43.Lowe MR, Doshi SD, Katterman SN, Feig EH. Dieting and restrained eating as prospective predictors of weight gain. Front Psychol [Internet] Frontiers; 2013. [cited 2021 May 11];4. Available from: https://www.frontiersin.org/articles/10.3389/fpsyg.2013.00577/full [DOI] [PMC free article] [PubMed]
  • 44.Haynes A, Kersbergen I, Sutin A, Daly M, Robinson E. A systematic review of the relationship between weight status perceptions and weight loss attempts, strategies, behaviours and outcomes. Obes Rev 2018;19:347–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Zheng Y, Manson JE, Yuan C, Liang MH, Grodstein F, Stampfer MJ, Willett WC, Hu FB. Associations of Weight Gain From Early to Middle Adulthood With Major Health Outcomes Later in Life. JAMA 2017;318:255–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Bouhlal S, Issanchou S, Nicklaus S. The impact of salt, fat and sugar levels on toddler food intake. Br J Nutr 2011;105:645–53. [DOI] [PubMed] [Google Scholar]
  • 47.Bouhlal S, Chabanet C, Issanchou S, Nicklaus S. Salt Content Impacts Food Preferences and Intake among Children. PLOS ONE Public Library of Science; 2013;8:e53971. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1

RESOURCES