Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Aug 9;34(10):1904–1914. doi: 10.1002/oby.70279

Eating Behavior Latent Profiles and Their Associations With Body Composition in Early Childhood: The CORALS Cohort

Andrea Jimeno‐Martínez 1, María L Miguel‐Berges 1,2,✉, Ivie Maneschy 1,2, Mercedes Gil‐Campos 2,3, Rosaura Leis 2,4,5,6, Nancy Babio 2,7,8, Pilar De Miguel‐Etayo 1,2, Santiago Navas‐Carretero 2,9,10,11, Olga Portolés 2,12, Montserrat Fitó Colomer 2,13, Cristina Castro‐Collado 3, Rosaura Picáns‐Leis 14, Natalia Farré 8,15, Miguel Seral‐Cortés 1,2, J Alfredo Martínez 2,9,16, José Jurado‐Castro 2,3, Rocío Vázquez‐Cobela 2,4,5,6, Cristina Rey‐Reñones 17,18, Belén Pastor‐Villaescusa 3,19, Francisco Llorente‐Cantarero 2,3, Karla‐Alejandra Pérez‐Vega 2,13,20, Jordi Salas‐Salvadó 2,7,8, Luis A Moreno 1,2, Guiomar Masip 1,2; CORALS group
PMCID: PMC13615416  PMID: 42571936

ABSTRACT

Objective

This study aimed to identify sex‐specific eating behavior profiles and assess the associations between eating behaviors subscales and profiles and body composition indices in Spanish preschoolers from the CORALS cohort.

Methods

We conducted a cross‐sectional analysis of baseline data from 1208 children aged 3–6 years (50.1% girls) from the CORALS cohort. Eating behaviors were assessed using the Child Eating Behavior Questionnaire (CEBQ). Sex‐specific latent profile analyses were conducted separately for girls and boys. Associations of CEBQ subscales and profiles with BMI, fat mass index, fat‐free mass index, and waist to height ratio (z scores) were examined using multiple linear regression models adjusted for sociodemographic variables, energy intake, playing outdoors, and sleep duration.

Results

Four eating behavior profiles were identified in both sexes: controlled, avid, typical, and avoidant. Food approach subscales and avid eater profile were positively associated with body composition indices, whereas food avoidance subscales and avoidant and controlled profiles showed inverse associations. Controlled eater profile showed opposite associations with fat‐free mass by sex: positive associations in boys (β = 0.42, 95% CI: 0.21, 0.62) but negative in girls (β = −0.55 95% CI: −0.77, −0.32).

Conclusions

Eating behavior profiles were associated with body composition indices similarly in boys and girls, with the exception of the controlled eater profile in relation to fat‐free mass index, for which the association differed by sex. Combining subscale‐level and profile‐level approaches may improve understanding of early eating behavior patterns and their role in body composition development.

Trial Registration: ClinicalTrials.gov NCT06317883 (observational cohort)

Keywords: body composition, childhood obesity, CORALS cohort, eating behaviors, latent profile analysis

Study Importance

  • What is already known?
    • ○
      Eating behaviors in early childhood are associated with body composition, with food approach behaviors positively associated and food avoidance behaviors inversely associated with body composition indices.
    • ○
      Most evidence has focused on individual eating behavior traits rather than on how eating behaviors cluster together in children.
  • What does this study add?
    • ○
      Four distinct eating behavior profiles—avid, avoidant, controlled, and typical—were identified in preschool children and were associated with fat mass, fat‐free mass, and waist to height ratio.
    • ○
      Controlled eaters showed different associations with fat‐free mass by sex: a positive association in boys and an inverse association with most body composition indices in girls.
  • How might these results change the direction of research or the focus of clinical practice?
    • ○
      Evaluating eating behaviors as profiles rather than single traits may better reflect how children eat in real‐world clinical settings.
    • ○
      Identifying eating behavior profiles in early childhood may support more tailored nutritional guidance and monitoring of body composition in girls and boys.

Abbreviation

CORALS

Childhood Obesity Risk Assessment Longitudinal Study

1. Introduction

Early childhood eating behaviors play a key role in shaping body composition and long‐term health outcomes [1]. In response to food‐related stimuli, children develop early eating behaviors that are shaped by both genetic and environmental factors [2]. These behaviors are typically categorized into food approach and food avoidance traits [3, 4]. Several tools exist to assess children's eating behaviors, including the Child Eating Behavior Questionnaire (CEBQ) [5] developed by Wardle et al., which is one of the most widely used questionnaires [6]. In Spain, the CEBQ has been adapted, validated, and used to explore associations with body composition [7, 8] and diet quality [9], as well as to examine factors influencing early eating behaviors [10].

Although individual appetitive traits have been associated with variations in body composition and growth patterns [11], research on how these behaviors cluster in children remains limited. Person‐centered approaches such as latent profile analysis (LPA) may offer a more comprehensive understanding by identifying distinct profiles, based on co‐occurring traits. Recent studies using LPA with CEBQ subscales have revealed distinct eating behavior profiles in preschool children, with a UK cohort reporting avid, typical, happy, and avoidant eater profiles, but associations with body composition indices were not assessed [12, 13]. On the other hand, a study of US preschool children reported high food avoidance, high food approach, and moderated eater profiles, finding that children in the food avoidance profile had lower BMI z scores compared with the other profiles [14]. However, these previous studies [12, 13, 14] have not assessed whether eating behavior profiles differ by sex. Evidence indicates that sex‐related differences in eating behaviors emerge in early childhood, with girls and boys differing in appetitive traits and how they relate to body composition [15]. Later, during adolescence, girls appear more influenced by external motivations such as body image and social norms, while boys emphasize intrinsic factors like enjoyment, autonomy, and performance [16, 17].

Hence, the aims of this study were to determine eating behavior profiles in girls and boys, respectively, and to assess the associations between both variable‐centered and person‐centered approaches and detailed body composition indices in Spanish preschoolers from the CORALS cohort.

2. Methods

2.1. Study Population

The Childhood Obesity Risk Assessment Longitudinal Study (CORALS) is an ongoing multicenter prospective cohort study (https://corals.es). At baseline, it included children aged 3–6 years attending selected schools in seven Spanish cities: Barcelona, Cordoba, Pamplona, Reus, Santiago de Compostela, Valencia, and Zaragoza. Participants are followed annually, with an expected follow‐up of approximately 10 years. The main objective of CORALS is to identify risk factors for childhood obesity [18]. The study was registered under ClinicalTrials.gov identifier NCT06317883. A detailed description of the CORALS cohort is published elsewhere [18]. The protocol was approved by the Ethics Committee of each recruitment center (references: 051/2019, 4155/2019, 2019/18, 9/19, 09/2019, 19/27, and 2019/131) and conducted in accordance with the standards of the Declaration of Helsinki. To participate in the study, parents or caregivers had to sign a consent form and fill in some questionnaires, such as sociodemographic data and lifestyle habits questionnaires. Exclusion criteria include belonging to a family with difficulty in participating, comprehension or language difficulties, and unstable residence.

For the present work, cross‐sectional data from the CORALS baseline sample (2025‐09‐02), which included 1509 participants, was considered. After excluding participants with missing data on maternal education, playing outdoors, energy intake, sleep duration, body composition measurements, and eating behaviors, 1208 children were included in the analyses (Figure 1).

FIGURE 1.

FIGURE 1

Open in a new tab

Flowchart of the sample selection process. CEBQ, Child Eating Behavior Questionnaire; FFM, fat‐free mass; FM, fat mass; WC, waist circumference. [Color figure can be viewed at wileyonlinelibrary.com]

2.2. Assessment of Eating Behaviors

Eating behaviors were assessed using the CEBQ, adapted and validated for Spanish children, as previously described [7]. The CEBQ comprised 35 items grouped into eight subscales reflecting both food approach and food avoidance traits. The food approach subscales include enjoyment of food, food responsiveness, desire to drink, and emotional overeating. The food avoidance subscales include food fussiness, satiety responsiveness, slowness in eating, and emotional undereating. The primary caregivers were asked to select the option that best described their children's eating behavior, with five possible answer categories from “never” to “always” [5].

2.3. Assessment of Body Composition Indices

Weight and height were measured by trained and qualified professionals. A portable stadiometer (SECA 213) was used for height measurement, as well as a body composition analysis scale (TANITA MC780 MA) to measure body weight (kg), body fat mass (kg), and body fat‐free mass (kg). Waist circumference (cm) was also assessed using an elastic tape (SECA 201). Body mass index (BMI) was calculated from weight and height (kg/m2) [19]. The fat mass index (FMI) was calculated as body fat mass (kg)/height (m2), and the fat‐free mass index (FFMI) was derived similarly, expressed in kg/m2 [20]. Waist to height ratio (WtHR) was estimated by dividing waist circumference by height. Age‐ and sex‐specific z scores for BMI (z‐BMI), FMI (z‐FMI), FFMI (z‐FFMI), and WtHR (z‐WtHR) were calculated within the internal study sample distribution. Although international BMI z score references exist, sample‐based standardization was applied to all body composition indices to ensure methodological consistency and comparability across outcomes, given the lack of established external references for FMI, FFMI, and WtHR in preschoolers. Z‐BMI and z‐FMI were used for the assessment of total body fat, z‐FFMI for fat‐free mass, and z‐WtHR for abdominal fat. As a sensitivity analysis, we re‐ran the models using BMI z scores derived from the IOTF reference charts from Cole and Lobstein [19].

2.4. Assessment of Covariates

Covariates, including sex, age, center, maternal education, energy intake, playing outdoors, and sleep duration, were collected by questionnaires filled out by the parents. These covariates were selected a priori based on previous evidence of associations with both eating behaviors and body composition [21, 22, 23, 24, 25], as they may confound the eating behavior‐body composition associations. Center refers to the seven Spanish cities from which participants were recruited. Maternal education, used as a proxy for socioeconomic status, was reported by the parents and categorized as low, medium, and high levels [26]. Total energy intake was estimated using the validated COME‐Kids Food and Beverage Frequency Questionnaire, a semi‐quantitative tool with 125 items assessing usual diet over the past year in children aged 3–11 years. Parents reported the frequency of consumption for each food and beverage, using the portion size specified in the questionnaire, including fast food, and also reported any usual foods or dietary supplements not included among the predefined items through three open items. Total energy and nutrient intakes were estimated according to the Spanish CESNID. To minimize misreporting, participants with energy intake below the 5th or above the 95th percentile (F&B‐FQ3 + 3d‐DR) were excluded from analyses [27]. Playing outdoors was used as a proxy for physical activity, estimated as minutes spent playing outdoors per week, calculated from time spent on school days, weekends, and extracurricular sports activities [28]. Sleep duration, in minutes per day, was calculated as a weighted average across 5 school days and 2 weekend days based on parental reports.

2.5. Statistical Analysis

All analyses were conducted separately for girls and boys. Sex‐stratified analyses were conducted given established sex differences in eating behaviors and body composition among preschoolers [15, 21, 29], to explore potential sex‐related differences in associations.

General characteristics of the participants were analyzed using means and standard deviations (SD) for continuous variables, and categorical variables were presented as numbers and percentages (%). Differences between girls and boys were determined using t‐tests and chi‐square tests.

Latent profile analysis (LPA) was conducted using CEBQ subscale scores to identify distinct eating behavior profiles. Model fit was evaluated using the Akaike information criterion (AIC), Bayesian information criterion (BIC), sample‐size adjusted BIC (aBIC), entropy, log‐likelihood (LogL), and the Lo–Mendell–Rubin adjusted likelihood ratio test (LMRT). Lower values of AIC, BIC, and aBIC indicated better model fit, while higher entropy values reflected greater classification precision. LogL values closer to zero reflected a better likelihood of the data under the model. Significant LMRT p values suggested that a model with k profiles provided a significantly better fit than a model with k–1 profiles. Participants were assigned to latent profiles based on their highest posterior probability of class membership, following established procedures [30].

Multiple linear regression models were used to assess the associations between eating behavior subscales and profiles and body composition indices. For profile analyses, eating profiles were treated as categorical predictors with “typical eater” as the reference category. Model 1 was adjusted for center and maternal education. Model 2 was additionally adjusted for energy intake, playing outdoors, and sleep duration.

SPSS Statistics version 29 (IBM Corp.) and R Studio statistical software were used to perform all statistical analysis. False discovery rate (FDR) correction was applied to account for multiple testing. Associations were considered significant if the FDR‐adjusted p values were < 0.05.

3. Results

3.1. General Characteristics of the Study Sample

Table 1 presents the general characteristics of the study sample. Half of the study participants were girls and half of the children came from families with a high maternal education level. Girls had slightly higher FMI (4.04 ± 1.3) and WtHR (0.48 ± 0.1), as well as lower z‐FFMI (−1.16 ± 1.0) values, and engaged more in slowness in eating (2.94 ± 0.8) compared to boys.

TABLE 1.

Baseline characteristics of the CORALS population.

Total Girls Boys p
Total sample, n (%) 1208 605 (50.1) 603 (49.9) 0.90
Age, years, mean (SD) 4.67 (1.1) 4.68 (1.1) 4.65 (1.1) 0.37
Maternal education, n (%) 0.36
Low 236 (19.5) 117 (19.3) 119 (19.7)
Medium 365 (30.2) 172 (28.5) 193 (32.0)
High 607 (50.3) 316 (52.2) 291 (48.3)
Energy intake, kcal/day, mean (SD) 1771.31 (484.4) 1732.04 (473.7) 1811.10 (492.2) 0.06
Playing outdoors, min/week, mean (SD) 252.70 (107.7) 248.51 (112.5) 257.13 (102.4) 0.10
Sleep duration, min/day, mean (SD) 624.40 (58.1) 625.83 (57.0) 622.92 (59.3) 0.99
BMI, kg/m2, mean (SD) 16.31 (2.1) 16.41 (2.2) 16.21 (2.0) 0.10
FMI, kg/m2, mean (SD) 3.84 (1.3) 4.04 (1.3) 3.65 (1.2) < 0.01
FFMI, kg/m2, mean (SD) 11.90 (1.1) 11.84 (1.1) 11.96 (1.2) 0.06
WtHR¸ mean (SD) 0.47 (0.1) 0.48 (0.1) 0.47 (0.1) 0.03
z‐BMI, mean (SD) −0.03 (1.0) −0.04 (1.0) −0.01 (1.0) 0.64
z‐FMI, mean (SD) −0.01 (1.0) −0.01 (1.0) −0.02 (1.0) 0.94
z‐FFMI, mean (SD) −0.92 (1.0) −1.16 (1.0) −0.66 (1.0) < 0.01
z‐WtHR, mean (SD) −0.04 (1.0) −0.04 (1.0) −0.04 (1.0) 0.99
CEBQ scales, mean (SD)
Enjoyment of food 3.41 (0.8) 3.43 (0.8) 3.38 (0.8) 0.31
Food responsiveness 2.18 (0.9) 2.21 (0.9) 2.15 (0.9) 0.25
Desire to drink 2.28 (0.9) 2.26 (0.9) 2.31 (0.9) 0.36
Emotional overeating 1.69 (0.6) 1.71 (0.6) 1.67 (0.6) 0.32
Food fussiness 2.89 (0.9) 2.85 (0.9) 2.93 (0.9) 0.14
Satiety responsiveness 2.76 (0.7) 2.78 (0.8) 2.874 (0.7) 0.34
Slowness in eating 2.89 (0.9) 2.94 (0.8) 2.84 (0.9) 0.03
Emotional undereating 2.93 (0.6) 2.93 (0.6) 2.93 (0.6) 0.98
Open in a new tab

Note: Data are presented as numbers (%) or means (SD); p values were determined by using chi‐square and t‐tests. z‐ indicating age‐ and sex‐standardized values. Bold values indicate statistical significance: *p < 0.05.

Abbreviations: FFMI, fat‐free mass index; FMI, fat mass index; WtHR, waist‐to‐height ratio.

3.2. Associations Between Eating Behavior Subscales and Body Composition Indices

Associations between CEBQ subscales and body composition indices are summarized in the color‐coded Figure 2, providing a comprehensive overview to facilitate interpretation across sexes and models. The subscale analyses with numerical data are shown in Figure S1.

FIGURE 2.

FIGURE 2

Open in a new tab

Direction and significance of associations between CEBQ subscales and body composition indices in girls and boys. EF, enjoyment of food; FR, food responsiveness; DD, desire to drink; EOE, emotional overeating; FF, food fussiness; SR, satiety responsiveness; SE, slowness in eating; EUE, emotional undereating. Green circles indicate significant positive associations (FDR‐adjusted p < 0.05); and red circles indicate significant negative associations. [Color figure can be viewed at wileyonlinelibrary.com]

3.3. Eating Behavior Profiles Stratified for Girls and Boys

Fit indices for LPA are shown in Table S1. For both girls and boys, a four‐profile solution showed the most appropriate fit. While models with five or six profiles showed slightly lower AIC and aBIC values, the BIC was lowest for the four‐profile solution in both girls and boys. The four‐profile model showed acceptable entropy values (0.33 for girls and 0.23 for boys) and allowed for larger and balanced sample sizes.

Figure 3 presents the mean CEBQ subscale scores for the selected four‐profile solution (3.A. for girls, 3.B for boys), and means and SD can be found in Tables S2 and S3. In girls and boys, the four identified profiles were controlled eater, avid eater, typical eater, and avoidant eater. Although the behavioral patterns were consistent across sexes, the proportion of children in each profile and the numerical ordering of the profiles varied between models. One example of this variation is that the controlled eater profile corresponded to Profile 4 in girls and Profile 1 in boys. In LPA, classes are not inherently ordered, and their numbering is model‐dependent; therefore, this difference does not affect interpretation, as all comparisons were conducted using the typical eater profile as the reference category.

  1. Controlled eater (n = 217, 36% for girls; n = 279, 46% for boys). Characterized by moderate scores across most subscales, low food responsiveness and emotional overeating, and high enjoyment of food, indicating that children in this profile do not eat impulsively and exhibit greater controlled appetite regulation.

  2. Avid eater (n = 62, 11% for girls; n = 47, 8% for boys). Characterized by high enjoyment of food and food responsiveness, alongside low satiety responsiveness and slowness in eating, reflecting a pronounced food approach tendency and reduced self‐regulation.

  3. Typical eater (n = 160, 26% for girls; n = 114, 19% for boys). Characterized by balanced scores across all CEBQ subscales, reflecting normative eating behaviors without pronounced tendencies toward food approach or avoidance.

  4. Avoidant eater (n = 166, 27% for girls; n = 163, 27% for boys). Characterized by high satiety responsiveness, slowness in eating, food fussiness, and emotional undereating and low food responsiveness and emotional overeating, reflecting cautious and selective eating behavior.

FIGURE 3.

FIGURE 3

Open in a new tab

Mean scores of CEBQ subscales for the selected four‐profile solution in (A) girls and (B) boys (girls =605; boys = 603). DD, desire to drink; EF, enjoyment of food; EO, emotional overeating; EU, emotional undereating; FF, food fussiness; FR, food responsiveness; SE, slowness in eating; SR, satiety responsiveness. [Color figure can be viewed at wileyonlinelibrary.com]

3.4. Associations Between Eating Behavior Profiles and Body Composition Indices in Girls

Associations between eating behavior profiles and body composition in girls are presented in Figure 4 (typical eater profile as reference). Girls in the controlled eater profile showed negative associations across all body composition indices, except for abdominal fat as z‐WtHR, which was not significantly associated in any model. Furthermore, the association with z‐FMI was significant in Model 1 (β = −0.15, p = 0.04) but did not survive FDR correction (p values = 0.06) and was not significant in Model 2 (β = −0.17, p = 0.10). The avoidant eater profile also showed negative associations with all body composition indices in both Models 1 and 2, with the strongest associations for z‐FFMI (β = 0.75, p < 0.001 for Model 1; β = 0.77, p < 0.001 for Model 2).

FIGURE 4.

FIGURE 4

Open in a new tab

Associations between eating behavior profiles and body composition in girls (n = 605). Model 1 was adjusted for center and maternal education. Model 2 was additionally adjusted for energy intake, playing outdoors, and sleep duration. [Color figure can be viewed at wileyonlinelibrary.com]

In contrast, the avid eater profile showed positive associations with most body composition indices, except for z‐FFMI, and the association for z‐WtHR (β = 0.36, p = 0.006 for Model 1 versus β = 0.22, p = 0.2 for Model 2) was attenuated after additional adjustment.

3.5. Associations Between Eating Behavior Profiles and Body Composition Indices in Boys

Results for boys are shown in Figure 5. The controlled eater profile was positively associated with z‐FFMI in boys (β = 0.42, p < 0.001 in both models), in contrast to the negative association observed in girls. Boys in the avoidant eater profile showed negative associations across all body composition indices, except for z‐FFMI, which was not significant (β = 0.07, p = 0.4 for Model 1; β = 0.01, p < 1.0 for Model 2), reflecting a pattern similar to the pattern observed in girls.

FIGURE 5.

FIGURE 5

Open in a new tab

Associations between eating behavior profiles and body composition in boys (n = 603). Model 1 was adjusted for center and maternal education. Model 2 was additionally adjusted for energy intake, playing outdoors, and sleep duration. [Color figure can be viewed at wileyonlinelibrary.com]

Conversely, the avid eater profile showed positive associations with all the body composition indices, with the strongest associations for z‐FFMI (β = 0.82, p < 0.001 for Model 1; β = 0.96, p < 0.001 for Model 2).

Analysis using BMI z scores derived from the Cole and Lobstein reference charts [19] yielded similar results in both girls and boys, with associations remaining in the same direction (Table S4).

4. Discussion

In this cross‐sectional study of Spanish children aged 3 to 6 years, we identified four distinct eating behavior profiles in both girls and boys: controlled eater, avid eater, typical eater, and avoidant eater, indicating comparable underlying behavioral structures across sexes. However, the numerical profile ordering differed between girls and boys due to the data‐driven nature of LPA. Similarly, while profile distributions and body composition associations were broadly comparable between girls and boys, a notable exception was observed for the controlled eater profile, which showed opposite associations with z‐FFMI in boys and girls. These prepubertal differences, though more typical in older children [31], might appear unexpected considering the young age of participants, probably reflecting early neurodevelopmental trajectories and sex‐specific parental feeding patterns rather than direct hormonal effects [15]. Both variable‐centered (CEBQ subscale) and person‐centered (LPA profile) approaches yielded consistent main findings in both sexes: food approach behaviors and the avid eater profile showed positive associations across all body composition indices, whereas food avoidance behaviors and the avoidant eater profile demonstrated inverse associations, in line with previous studies [7, 8].

A notable exception was the controlled eater profile, which showed sex‐specific associations with z‐FFMI, contrasting with the more uniform inverse associations of individual avoidance subscales. This discrepancy likely reflects LPA's ability to capture co‐occurring traits—moderate‐high enjoyment of food combined with low food responsiveness and emotional overeating—that coexist with individual subscale patterns, highlighting the complementary value of integrated profile analysis for understanding complex eating behavior–body composition relationships in preschoolers.

The main analyses were based on z scores standardized within the study sample to ensure methodological consistency across body composition indices. The findings were also robust in sensitivity analyses using BMI z scores derived from the Cole and Lobstein reference charts [19], with associations remaining in the same direction (Table S4).

The avid eater profile, characterized by food approach behaviors, was positively associated with most body composition indices in both sexes. This interpretation suggests that the predisposition to respond to external food cues may represent early and sex‐independent risk factors for increased adiposity, although previous evidence has proposed that girls might be more sensitive to external food cues or social reinforcement related to eating than boys [32]. Girls showed larger associations with z‐BMI while boys exhibited particularly larger associations with z‐FFMI, suggesting that pronounced food approach tendencies may contribute to the development of both adipose and lean tissue, with sex‐specific patterns. In the same CORALS cohort, food approach behaviors were associated with different dietary patterns by sex [21]. Boys tended to show higher intake of animal protein sources (e.g., meat, fish), which may be linked to greater fat‐free mass, whereas girls showed greater adherence to a veg–seafood–legumes pattern [21]. These sex‐specific dietary patterns may contribute to differences in body composition; however, the underlying mechanisms remain unclear and are likely multifactorial. Importantly, these findings should not be interpreted as reflecting differences in diet quality or energy density alone, but rather as part of a broader behavioral and metabolic context associated with eating behavior traits [33].

Similarly, the avoidant eater profile showed inverse associations with body composition indices in both sexes, consistent with previous research findings on the avoidant eater profile reported by Fisher et al. [14].

By contrast, the controlled eater profile showed sex‐specific associations with z‐FFMI, with a negative association in girls and a positive association in boys despite the presence of avoidance‐related traits. This discrepancy may arise from the way LPA combines multiple behaviors into distinct patterns [34], from sex‐specific differences in early growth trajectories and body composition development [35], which may modulate how the same controlled eater profile affects lean mass. In the same CORALS cohort, boys' controlled patterns may coincide with higher animal protein intake, supporting FFMI despite avoidance traits [21].

This person‐centered approach integrates multiple traits into meaningful patterns, offering a more comprehensive understanding than scale‐level analyses alone. The identified latent profiles broadly align with those described in previous studies of preschool children, such as the four‐profile structure (avid, happy/emotional, typical, and avoidant eater) reported by Pickard et al. in the UK [12, 13] and the more differentiated profiles observed in a study from the US (high food avoidance, high food approach, and moderate eater) [14]. However, the emergence of two avoidance‐related profiles in our Spanish cohort suggests a finer differentiation within low‐intake behaviors. While avoidant eaters showed the expected pattern of high satiety responsiveness, slowness in eating, and food fussiness, the controlled eater profile combined low food responsiveness with relatively high enjoyment of food, indicating a form of regulated eating rather than avoidant eating. This distinction may reflect cultural and contextual influences on children's feeding environments in Spain [36]. Family mealtimes tend to be more structured [37], and parents tend to play a more direct role in feeding, which may promote self‐regulation of appetite [38]. Although previous studies identifying latent profiles in similar samples did not stratify profiles by sex at early ages, our findings suggest that sex differences may emerge from early neurodevelopmental trajectories and be reinforced over time through parental practices and social norms [15, 39]. This supports prior research indicating that person‐centered profiles may capture complex behaviors but are less stable across contexts [40]. Overall, our findings are broadly consistent with the limited available previous research [12, 14], while highlighting subtle local variations that underscore the importance of examining eating behaviors as integrated profiles.

These findings have potential clinical and policy relevance. Overall, body fat composition indices align directionally across profiles. Additionally, BMI remains more feasible in routine settings [41] where fat mass (z‐FMI) and central adiposity (z‐WtHR) cannot always be calculated, though they may offer complementary prognostic information [42]. Eating behavior profiles should be assessed early, as they reflect how individuals eat and their relationship with food, preceding dietary patterns and informing dietary intake [33]. In clinical settings, identifying eating behavior profiles in early childhood may help clinicians to detect children at risk of developing unfavorable body composition patterns before they become established. The avid eater profile shows consistent positive associations across indices (particularly z‐BMI and z‐WtHR), flagging children needing behavioral monitoring [12]. Conversely, avoidant or controlled eater profiles may signal self‐regulation patterns warranting growth surveillance [14]. These profiles—neither inherently good nor bad, encompassing overeating and undereating tendencies—offer a complete behavioral lens beyond obesity risk, relevant for eating disorders prevention [43], family guidance on child feeding, clinical practice, and public health policies.

Integrating eating behavior profile‐based assessment with traditional anthropometric measures could therefore support more personalized, developmentally sensitive approaches to obesity prevention in preschool children.

There are some limitations in our study. First, this study has a cross‐sectional design, which did not allow us to establish temporal or causal relationships between eating behaviors and body composition indices and left open the possibility of reverse causation [44]. Therefore, future longitudinal studies are needed to clarify the directionality of these associations, which were bidirectional in previous studies [45]. Second, several key variables—including eating behaviors, playing outdoors, sleep duration, and energy intake—were assessed through parent‐reported questionnaires, which may introduce reporting bias [46]. Parents of children with overweight/obesity may under‐ or overestimate behaviors due to weight concerns or social desirability, potentially affecting covariate adjustment and results. Third, although anthropometric and body composition measurements were obtained by trained staff using standardized protocols and validated devices, some degree of measurement error is inevitable, particularly in young children, and may have led to nondifferential attenuation of associations. Fourth, despite adjustment for multiple sociodemographic and lifestyle covariates, residual confounding from unmeasured or imperfectly measured factors cannot be ruled out. Finally, this sample comprised Spanish preschool children, which may limit the generalizability of findings. However, considering the similarities in dietary habits and preschool environments across Westernized countries [47], the findings may still be relevant in comparison to other Westernized populations.

Despite these limitations, our study presents several strengths. This study is the first to examine the relationship between eating behaviors and a comprehensive set of objectively measured body composition indices, enabling the differentiation of fat and fat‐free mass in young children. Moreover, the use of LPA allowed us to capture the whole eating behavior picture, providing a more integrated view than considering each eating behavior subscale individually. While the same four profiles emerged in both girls and boys, stratified analyses revealed nuanced sex‐specific associations with body composition, specifically for the controlled eater profile with z‐FFMI, highlighting subtle differences in how similar behavioral patterns may influence energy balance across sexes. Collectively, these methodological features enhance the depth, rigor, and interpretive value of our findings, contributing to a more refined understanding of the relationship between eating behaviors and body composition in children.

5. Conclusion

In conclusion, our findings indicate that eating behaviors in preschoolers are associated with body composition. Food approach behaviors are linked to higher body composition indices, whereas food avoidance behaviors are linked to lower scores across all these indices. The controlled eater profile showed opposite fat‐free mass associations by sex, suggesting that similar eating tendencies may differentially influence body composition depending on sex. These findings highlight the value of person‐centered approaches for understanding early appetite regulation and informing pediatric nutrition strategies. Future interventions should prioritize avid eaters (highest adiposity risk) while monitoring avoidant and controlled eaters for balanced growth. Future longitudinal studies are needed to test the predictive value of these profiles for long‐term health outcomes.

Author Contributions

The author's responsibilities were as follows: M.G.‐C., N.B., L.A.M., J.S.‐S., O.P., M.F.C., R.L., and S.N.‐C., conceptualization and CORALS study design. L.A.M., conceptualization and this analysis. L.A.M., G.M., and A.J.‐M., methodology. A.J.‐M. and G.M., formal analysis. A.J.‐M. and M.L.M.‐B., data curation. A.J.‐M. and G.M., writing original draft. G.M., M.L.M.‐B., and L.A.M., supervision. A.J.‐M., G.M., M.L.M.‐B., M.S.‐C., and L.A.M., draft review and editing. All authors (A.J.M., M.L.M.‐B., I.M., M.G.‐C., R.L., N.B., P.M.E., S.N.‐C., O.P., M.F.C., C.C.‐C., R.P.‐L., N.F., M.S.‐C., J.A.M., J.J.‐C., R.V.‐C., C.R.‐R., B.P.‐V., F.L.‐C., K.‐A.P.‐V., J.S.‐S., L.A.M., G.M.) revised and approved the final manuscript.

Funding

Funds for the establishment of the CORALS cohort in the first year of the study (2019) were provided by an agreement between the Danone Institute from Spain and the Centro de Investigación Biomédica en Red de la Fisiopatología de la Obesidad y Nutrición (CIBEROBN). This work was partially funded by the 2024 Intramural Projects Call of the Biomedical Research Networking Center in Physiopathology of Obesity and Nutrition (CIBEROBN) and by the Spanish government's official funding agency for biomedical research, Instituto de Salud Carlos III (ISCIII), through the Fondo de Investigacion para la Salud (FIS) and co‐funded by the European Union ERDF/ESF, “A way to make Europe”/“Investing in your future” [PI24/00711]. G.M. received funding from the “Ayudas para contratos Juan de la Cierva” funded for MCIU/AEI/10.13039/50110001103 and the European Union “NextGeneration/PRTR,” grant reference number: JDC2022‐048656‐I. J.S.‐S. is partially supported by ICREA under the ICREA Academia program.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Associations between CEBQ subscales and body composition indices stratified by sex (girls = 605; boys = 603). A. z‐BMI; B. z‐FMI; C. z‐FFMI; D. z‐WtHR. Model 1 was adjusted for center and maternal education. Model 2 was additionally adjusted for energy intake, physical activity, and sleep duration.

Table S1: Fit statistics for 1–6 profiles solutions.

Table S2: Mean scores of the eight CEBQ subscales across four profiles in girls (n = 605).

Table S3: Mean scores of the eight CEBQ subscales across four profiles in boys (n = 603).

Table S4: Associations between eating behavior profiles based on Cole and Lobstein reference charts [19].

OBY-34-1904-s001.docx (866.4KB, docx)

Acknowledgments

The authors thank all the CORALS participants and their parents or caregivers as well as the health centers and primary schools for their collaboration, the CORALS personnel for outstanding support, and the staff of all associated primary care centers for exceptional work.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

References

  • 1. WHO , “Childhood Obesity in European Region Remains High: New WHO Report Presents Latest Country Data,” published November 8 (2022), https://www.who.int/europe/news/item/08‐11‐2022‐childhood‐obesity‐in‐european‐region‐remains‐high‐‐new‐who‐report‐presents‐latest‐country‐data.
  • 2. Llewellyn C. and Wardle J., “Behavioral Susceptibility to Obesity: Gene‐Environment Interplay in the Development of Weight,” Physiology & Behavior 152 (2015): 494–501. [DOI] [PubMed] [Google Scholar]
  • 3. Wardle J., “Eating Behaviour and Obesity,” Obesity Reviews 8, no. S1 (2007): 73–75. [DOI] [PubMed] [Google Scholar]
  • 4. Dantas R. and Silva G., “The Role of the Obesogenic Environment and Parental Lifestyles in Infant Feeding Behavior,” Revista Paulista de Pediatria 37, no. 3 (2019): 363–371. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Wardle J., Guthrie C. A., Sanderson S., and Rapoport L., “Development of the Children's Eating Behaviour Questionnaire,” Journal of Child Psychology and Psychiatry, and Allied Disciplines 42, no. 7 (2001): 963–970. [DOI] [PubMed] [Google Scholar]
  • 6. Sanlier N., Arslan S., Buyukgenc N., and Toka O., “Are Eating Behaviors Related With by Body Mass Index, Gender and Age?,” Ecology of Food and Nutrition 57, no. 4 (2018): 372–387. [DOI] [PubMed] [Google Scholar]
  • 7. Jimeno‐Martínez A., Maneschy I., Moreno L., et al., “Reliability and Validation of the Child Eating Behavior Questionnaire in 3‐ to 6‐Year‐Old Spanish Children,” Frontiers in Psychology 13, no. 4 (2022): 705912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Calderón García A., Alaminos‐Torres A., Pedrero Tomé R., et al., “Eating Behavior and Obesity in a Sample of Spanish Schoolchildren,” International Journal of Environmental Research and Public Health 20, no. 5 (2023): 4186, 10.3390/ijerph20054186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Calderón García A., Pedrero Tomé R., Alaminos Torres A., et al., “Adhesión a la dieta mediterránea y conducta alimentaria en escolares españoles [Adherence to the Mediterranean Diet and Eating Behaviour in Spanish Schoolchildren],” Nutricion Hospitalaria 41, no. 1 (2024): 47–57, 10.20960/nh.04696. [DOI] [PubMed] [Google Scholar]
  • 10. Ortega‐Ramírez A. D., Maneschy I. R., Miguel‐Berges M. L., et al., “Early Feeding Practices and Eating Behaviour in Preschool Children: The CORALS Cohort,” Maternal & Child Nutrition 20, no. 4 (2024): e13672, 10.1111/mcn.13672. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Kininmonth A., Smith A., Carnell S., Steinsbekk S., Fildes A., and Llewellyn C., “The Association Between Childhood Adiposity and Appetite Assessed Using the Child Eating Behavior Questionnaire and Baby Eating Behavior Questionnaire: A Systematic Review and Meta‐Analysis,” Obesity Reviews 22, no. 5 (2021): e13169. [DOI] [PubMed] [Google Scholar]
  • 12. Pickard A., Croker H., Edwards K., et al., “Identifying an Avid Eating Profile in Childhood: Associations With Temperament, Feeding Practices and Food Insecurity,” Appetite 191 (2023): 107050. [DOI] [PubMed] [Google Scholar]
  • 13. Pickard A., Farrow C., Haycraft E., et al., “Associations Between Parent and Child Latent Eating Profiles and the Role of Parental Feeding Practices,” Appetite 201 (2024): 107589. [DOI] [PubMed] [Google Scholar]
  • 14. Fisher J. O., Hughes S. O., Miller A. L., et al., “Characteristics of Eating Behavior Profiles Among Preschoolers With Low‐Income Backgrounds: A Person‐Centered Analysis,” International Journal of Behavioral Nutrition and Physical Activity 19 (2022): 91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Keller K. L., Kling S. M. R., Fuchs B., et al., “A Biopsychosocial Model of Sex Differences in Children's Eating Behaviors,” Nutrients 11, no. 3 (2019): 682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Maneschy I., Moreno L. A., Ruperez A. I., et al., “Eating Behavior Associated With Food Intake in European Adolescents Participating in the HELENA Study,” Nutrients 14, no. 15 (2022): 3033, 10.3390/nu14153033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Deslippe A. L. B. C. and Cohen T. R., “Boys and Girls Differ in Their Rationale Behind Eating: A Systematic Review of Intrinsic and Extrinsic Motivations in Dietary Habits Across Countries,” Frontiers in Nutrition 10 (2023): 1256189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Garcidueñas‐Fimbres T., Paz‐Graniel I., Gómez‐Martínez C., et al., “Associations Between Eating Speed, Diet Quality, Adiposity, and Cardiometabolic Risk Factors,” Journal of Pediatrics 252 (2023): 231–239. [DOI] [PubMed] [Google Scholar]
  • 19. Cole T. and Lobstein T., “Extended International (IOTF) Body Mass Index Cut‐Offs for Thinness, Overweight and Obesity,” Pediatric Obesity 7, no. 4 (2012): 284–294. [DOI] [PubMed] [Google Scholar]
  • 20. VanItallie T., Yang M., Heymsfield S., Funk R., and Boileau R., “Height‐Normalized Indices of the Body's Fat‐Free Mass and Fat Mass: Potentially Useful Indicators of Nutritional Status,” American Journal of Clinical Nutrition 52, no. 6 (1990): 953–959. [DOI] [PubMed] [Google Scholar]
  • 21. Maneschy I. O.‐R. A., Jimeno‐Martínez A., Gil‐Campos M., et al., “Association Between Eating Behaviours and Food and Beverage Consumption in Male and Female Children Aged 3–6 Years: The CORALS Cohort,” European Journal of Nutrition 65, no. 1 (2026): 26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. De Craemer M., Cardon G., Decraene M., et al., “Longitudinal Changes in Preschoolers' Adiposity Indicators According to Compliance With 24‐Hour Movement Behavior Guidelines: Results From the ToyBox‐Study,” BMC Public Health 24, no. 1 (2024): 3115, 10.1186/s12889-024-20626-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Bučan Nenadić D., Štrkalj L., Zloić K., et al., “Laying the Groundwork for Health: Eating Behaviour and Physical Activity in Preschoolers in Split‐Dalmatia County, Croatia,” Children 12, no. 6 (2025): 699, 10.3390/children12060699. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Delahunt A., Conway M. C., McDonnell C., et al., “Sleep Duration and Eating Behaviours Are Associated With Body Composition in 5‐Year‐Old Children: Findings From the ROLO Longitudinal Birth Cohort Study,” British Journal of Nutrition 127, no. 11 (2022): 1750–1760. [DOI] [PubMed] [Google Scholar]
  • 25. Ayine P. S. V., Venkatapoorna C. M. K., Bao Y., Gaillard P., and Geetha T., “Eating Behaviors in Relation to Child Weight Status and Maternal Education,” Children 8, no. 1 (2021): 32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. UNESCO Institute for Statistics , ISCED Fields of Education and Training 2013 (ISCED‐F 2013): Manual to Accompany the International Standard Classification of Education 2011 (UNESCO, 2014). [Google Scholar]
  • 27. Babio N., De Miguel‐Etayo P., Pastor‐Villaescusa B., et al., “Reproducibility and Relative Validity of a Semi‐Quantitative Food and Beverage Frequency Questionnaire for Spanish Children Aged 3 to 11 Years: The COME‐Kids F&B‐FQ,” European Journal of Pediatrics 182, no. 12 (2023): 5577–5589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Jurado‐Castro J. M., Castro‐Collado C., Gil‐Campos M., et al., “Evaluation of Physical Activity, Sedentary Patterns, and Lifestyle Behavior in Spanish Preschool Children From the CORALS Cohort,” Sports Medicine ‐ Open 11 (2025): 71. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Durão C. S. M., Oliveira A., and Lopes C., “Sex‐Heterogeneity on the Association Between Dietary Patterns at 4 Years of Age With Adiposity and Cardiometabolic Risk Factors at 10 Years of Age,” Nutrients 14, no. 3 (2022): 540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Nagin D. S., Group‐Based Modeling of Development (Harvard Press, 2005). [Google Scholar]
  • 31. Santiago S., Zazpe I., Martí A., Cuervo M., and Martínez J. A., “Gender Differences in Lifestyle Determinants of Overweight Prevalence in a Sample of Southern European Children,” Obesity Research & Clinical Practice 7, no. 5 (2013): e391–e400, 10.1016/j.orcp.2012.07.001. [DOI] [PubMed] [Google Scholar]
  • 32. Cooper A. J., Gupta S. R., Moustafa A. F., and Chao A. M., “Sex/Gender Differences in Obesity Prevalence, Comorbidities, and Treatment,” Current Obesity Reports 10, no. 4 (2021): 458–466. [DOI] [PubMed] [Google Scholar]
  • 33. Maneschy I., Jimeno‐Martínez A., Miguel‐Berges M. L., et al., “Eating Behaviours and Dietary Intake in Children and Adolescents: A Systematic Review,” Current Nutrition Reports 13, no. 3 (2024): 363–376. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. He J. and Fan X., “Latent Profile Analysis,” in Encyclopedia of Personality and Individual Differences, ed. Zeigler‐Hill V. and Shackelford T. (Springer, 2019). [Google Scholar]
  • 35. Peralta G. P. F. E., Granell R., Mahmoud O., et al., “Childhood Body Composition Trajectories and Adolescent Lung Function. Findings From the ALSPAC Study,” American Journal of Respiratory and Critical Care Medicine 200, no. 1 (2019): 75–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Klerks M. R. S., Bernal M. J., Haro‐Vicente J. F., and Sanchez‐Siles L. M., “Complementary Feeding Practices and Parental Pressure to Eat Among Spanish Infants and Toddlers: A Cross‐Sectional Study,” International Journal of Environmental Research and Public Health 18, no. 4 (2021): 1982. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Fischler C., “Commensality, Society and Culture,” Social Science Information 50, no. 3–4 (2011): 528–548. [Google Scholar]
  • 38. López‐Gil J. F., Smith L., Tully M. A., Álvarez‐Pitti J., Gómez S. F., and Schröder H., “Prevalence of Daily Family Meals Among Children and Adolescents From 43 Countries,” Maternal & Child Nutrition 21, no. 1 (2025): e13706, 10.1111/mcn.13706. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Breton E., Khundrakpam B., Jeon S., Evans A., and Booij L., “Cortical Thickness and Childhood Eating Behaviors: Differences According to Sex and Age, and Relevance for Eating Disorders,” Eating and Weight Disorders 29, no. 1 (2024): 47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Derks I. P. M., Sijbrands E. J. G., Wake M., et al., “Eating Behavior and Body Composition Across Childhood: A Prospective Cohort Study,” International Journal of Behavioral Nutrition and Physical Activity 15, no. 1 (2018): 96. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. WHO , Child Growth Standards: Length/Height‐for‐Age, Weight‐for‐Age, Weight‐for‐Length, Weight‐for‐Height and Body Mass Index‐for‐Age, Methods and Development (WHO, 2006). [Google Scholar]
  • 42. Clasey J. L. E. E., Murphy M. O., Kiessling S. G., et al., “Body Mass Index Percentiles Versus Body Composition Assessments: Challenges for Disease Risk Classifications in Children,” Frontiers in Pediatrics 11 (2023): 1112920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Bourne L., Bryant‐Waugh R., Cook J., and Mandy W., “Avoidant/Restrictive Food Intake Disorder: A Systematic Scoping Review of the Current Literature,” Psychiatry Research 288 (2020): 112961. [DOI] [PubMed] [Google Scholar]
  • 44. Hernán M. A. R. J., Causal Inference: What if (Hall/CRC BRC, 2020). [Google Scholar]
  • 45. Masip G., Foraita R., Silventoinen K., et al., “The Temporal Relationship Between Parental Concern of Overeating and Childhood Obesity Considering Genetic Susceptibility: Longitudinal Results From the IDEFICS/I.Family Study,” International Journal of Behavioral Nutrition and Physical Activity 18, no. 1 (2021): 139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Saglam D. A. M., Colak G. A., and Bas M., “Validation of the Dutch Eating Behaviour Questionnaire Children (DEBQ‐C) Version in Turkish Preadolescence Children,” Nutrition Research and Practice 16, no. 6 (2022): 765. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Pereira‐da‐Silva L. R. C. and Pietrobelli A., “The Diet of Preschool Children in the Mediterranean Countries of the European Union: A Systematic Review,” International Journal of Environmental Research and Public Health 13, no. 6 (2016): 572. [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

Figure S1: Associations between CEBQ subscales and body composition indices stratified by sex (girls = 605; boys = 603). A. z‐BMI; B. z‐FMI; C. z‐FFMI; D. z‐WtHR. Model 1 was adjusted for center and maternal education. Model 2 was additionally adjusted for energy intake, physical activity, and sleep duration.

Table S1: Fit statistics for 1–6 profiles solutions.

Table S2: Mean scores of the eight CEBQ subscales across four profiles in girls (n = 605).

Table S3: Mean scores of the eight CEBQ subscales across four profiles in boys (n = 603).

Table S4: Associations between eating behavior profiles based on Cole and Lobstein reference charts [19].

OBY-34-1904-s001.docx (866.4KB, docx)

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

Articles from Obesity (Silver Spring, Md.) are provided here courtesy of Wiley

RESOURCES