1 INTRODUCTION

The rates of childhood overweight have increased dramatically over the last decades. To illustrate, worldwide the number of children with overweight or obesity below the age of 5 years increased from 32 million in 1990 to 38 million in 2019.1 In the Netherlands in 2018, 12% of all children aged from 2 to 8 years were overweight.2

The fundamental cause of overweight and obesity is a positive physiological imbalance between energy intake and energy expenditure over an extended period of time.3 An imbalance between intake and expenditure of only 2 % on a daily basis, sustained over time, can induce overweight in growing children.4 Also, children with overweight are at higher risk of remaining overweight in adulthood.3, 5 Therefore, prevention of overweight and obesity should start early in life since eating behaviours are learned and established during childhood, especially in the first 2 years of life.6

There is a need for better understanding of behavioural risk factors contributing to overweight in childhood. For instance, individual eating behaviours affect children's responses towards food, and are related to a greater risk of overweight and obesity.7-12 Larger appetite ratings, greater enjoyment of food, greater food responsiveness, faster eating rate, and lower satiety responsiveness were found to be related to higher body weights and faster weight gain in infants aged 3 to 15 months.11 In addition, lower satiety responsiveness and greater food responsiveness have been associated with higher prevalence of adiposity in older children aged 3 to 5 years.13

Children are born with the ability to self-regulate short-term energy intake but this ability seems to diminish with age, when the effects of sensory, cognitive, and social factors on energy intake become stronger.14-16 For example, external cues such as availability of palatable foods, portion size effects, presence of other people, time of day, and parental feeding practices may overrule or have a greater influence on intake than internal cues of hunger and satiety and consequently reduce the ability to self-regulate energy intake.17, 18 A poor ability to self-regulate short-term energy intake is associated with a positive energy balance in children aged 5 to 12 years,19, 20 thereby increasing the risk of developing overweight later in life.5, 21

Eating in the absence of hunger (EAH) refers to the failure to self-regulate energy intake, and the susceptibility to eat palatable, often energy-dense foods despite experiencing satiety,10 making it a behavioural risk factor for developing overweight.7, 20, 22, 23 Previous work in preschoolers and primary school age children has shown that EAH increases with age and is consistent within individuals over time.20, 22, 24-27 Also, EAH has been observed in children as young as 21 months25, 28 suggesting it already occurs at a very young age. However, the age at which self-regulation of short-term energy intake diminishes and EAH emerges may be even younger than this. Also, it remains unclear how individual eating behaviours play a role in the emergence and dynamics of EAH over time.

Extending our knowledge on the determinants of EAH is important for early recognition of risk behaviours contributing to overeating in children, and for the timing of early targeted interventions to prevent overeating. Furthermore, such insights in risk behaviour could be translated into practical guidelines for parents and caregivers. To this aim, we performed a study among 18-month-old children (n = 217) in a home setting to assess to what extent they eat in the absence of hunger. This procedure was repeated in a subset of the sample (n = 107) when the children were 24 months of age, to examine whether EAH remained stable over time. In addition, the association of satiety of the child (as perceived by the mother) and child eating behaviours with EAH was investigated. It was expected that children who ingested more energy during the evening meal, and were perceived as being more satiated, would ingest less energy from the finger foods presented during the EAH procedure. It was also hypothesized that certain child eating behaviours, such as greater enjoyment of food, greater food responsiveness, and lower satiety responsiveness would result in a relatively higher energy intake of finger foods. Finally, it was expected that EAH at 18 months would predict EAH at 24 months in the subset of children where EAH was measured twice, at 18 and 24 months of age.

2 METHODS 2.1 Study design and participants

The current study presents data collected as part of a large longitudinal study, Baby's First Bites (BFB), of which a detailed description of the study protocol has been published elsewhere.29 In short, BFB is a four-arm randomized controlled trial that studies the role of the start of complementary feeding in infancy in promoting vegetable intake. It included 243 first-time Dutch mothers and their infants. Interventions started with a standardized feeding schedule after parents indicated their child was ready to start complementary feeding (age 4-6 months), and lasted until the age of 16 months.29 Home assessments were performed at the ages of 12, 18, 24, and 36 months. As part of this RCT, EAH was measured during the home assessment at the ages of 18 and 24 months. The home assessments were performed by trained Master and PhD students and entailed anthropometric measurements and a videotaped meal and play session between mother and child.29

Written informed consent was obtained from both parents before participation in the BFB trial and the protocol was approved by the Ethical Review Board of Education and Child Studies, Leiden University (protocol number ECPW-2015/116) and the Medical Ethical Review Board of Wageningen University and Research (METC-WU protocol number NL54422.081.15). The trial was registered during inclusion of participants at the Netherlands National Trial Register (identifier NTR6572) and at ClinicalTrials.gov (NCT03348176).

For each assessment, mothers received a €25 gift voucher and the child received a small present (eg, colouring or reading book).

The EAH measurements were conducted between July 2017 and June 2019. In total, 217 mother-child pairs participated in the EAH procedure when the child was 18 months old. Due to constraints in budget and manpower, the repeated measurement at 24 months did not include the full sample, but a subsample of 107 mother-child pairs living in the surroundings of Leiden University.

General information, such as child's date of birth, sex and maternal ethnicity, and educational level, was obtained from written and online surveys completed at the start of the BFB trial.

2.2 Child eating behaviour

Prior to the home visits at 18 and 24 months, mothers filled out the Child Eating Behaviour Questionnaire for toddlers (CEBQ-T)30 online. This is a well-validated, reliable, and widely used questionnaire consisting of 26 items to assess six eating styles that have been related to overeating and overweight: food responsiveness (FR), satiety responsiveness (SR), enjoyment of food (EF), food fussiness (FF), emotional overeating (EOE), and slowness in eating (SE).10, 31 Mothers reported on a five-point Likert scale (‘1 = never’ to ‘5 = always’) how frequently they observed their child demonstrating a range of eating behaviours on a typical day. The scales have good internal consistency with Cronbach's Alpha ranging from .74 to .91 (FR: .80, SR: .74, EF: .91 FF: .91 EOE: .79 and SE: .74).10

2.3 Anthropometrics

Children's and maternal bodyweight was measured during every home visit using a calibrated digital scale (SECA robusta 813), in kilograms to the nearest 0.1 kg. Children's length was measured on an infant measuring mat to the nearest 0.5 cm. Maternal height was measured at the start of the BFB trial using a portable stadiometer to the nearest 0.5 cm. For children, body mass index (BMI) was calculated and transformed into age and sex-standardized z-scores (BMI-z) based on the WHO Child Growth Standards for boys and girls aged 0 to 60 months.32 Maternal height and body weight were used to calculate BMI in kg/m2. In case the mother was pregnant or had recently given birth, the bodyweight of the mother measured during the home visit when the child was 12 months old was used to calculate maternal BMI (n = 37).

2.4 Measurement of child's EAH

The protocol for measuring EAH was based on the free-access procedure for children aged 3 to 5 years old in a laboratory setting as described by Fisher and Birch (1999), where children's snack food intake was measured after consuming their usual lunch. In that study, children were provided with 10 snack foods and some toys for 10 minutes and only children who indicated that they were not hungry following lunch participated in the experiment.33 For the present study, the described protocol was adapted to suit 18- and 24-month-old children in a home setting. The following adjustments were made: (a) testing was done in a home-setting instead of a laboratory setting after an evening meal prepared by the parents, (b) mothers estimated the level of satiety of their child, as children this young are not yet able to report this themselves, (c) children participated in the procedure irrespective of the satiety score, (d) intake of the child's evening meal was assessed, and (e) toys were provided and mothers were asked to play with their child as usual for 8 minutes. Thereafter four finger foods (Table 1) were presented to all children for the 10-minute EAH procedure. Children had the opportunity to continue playing with the toys or eat the provided foods. Prior to the home visits, mothers were asked for permission to offer the child the selected foods. Only if the child was allergic to a food or the mother disapproved of a food an alternative was offered (Table 1). Twenty-four children were offered at least one alternative finger food.

TABLE 1. Finger foods provided during the EAH procedure Standard foods Serving Weight (g) per serving Energy per serving (kcal) Savoury Breadstick Two sticks 14 57 Nibbits sticks One handful 15 72 Sweet Gingerbread One slice 27 84 Plain biscuit Two pieces 14 62 Alternative foods Serving Weight (g) per Serving Energy per serving (kcal) Savoury Rice cracker Two pieces 14 53 Cream cracker Two pieces 16 72 Sweet Banana 0.5 Piece 65 62 Raisins One spoonful 15 52

The EAH procedure was piloted in four sessions at a day care centre in Wageningen, The Netherlands (n = 17, age 16-24 months), to test the procedure's feasibility and children's behaviour toward multiple finger foods. In total, 11 foods were tested (breadsticks, cream crackers, Nibbit sticks, salty biscuits, cake, gingerbread, banana, Miffy shaped cookies, pouch of pureed fruit, raisins, and plain sweet biscuits) of which four were selected to be offered during the EAH procedure (breadsticks, Nibbit sticks, gingerbread, and plain sweet biscuit). Selection was based on the criteria that children were able to take and eat the food without help, the food attracted the interest of the children and was considered to be age appropriate by the Netherlands Nutrition Centre. Selected foods were frequently consumed according to the Dutch National Food Consumption Survey (2010) among 1-year-old children. The procedure itself was considered feasible and remained unchanged.

Parents were asked to prepare an evening meal for the child as part of their usual daily routine. The child's intake was assessed by obtaining a detailed description of the ingredients and preparation of the meal, drink, and desert. Everything the child ate and drank was weighed on a calibrated digital kitchen scale (Soehnle, Fiesta 65106) and photographed by the researcher before and after consumption. Photographs were taken at a 45° angle and a ruler was placed under the plate as a visual reference for portion size estimation.34 This method for assessing the weight, energy, and macronutrient content of the evening meal was evaluated within the sample of the BFB trial (data not reported) using a weighed food record as the reference method. The mean energy and macronutrient content intakes assessed by detailed description, weight, and photographs were not significantly different from the reference method. Indicating it as a reliable method for estimating energy intake of the evening meal.

The nutritional data were processed by trained dietitians. Portion sizes of each food item were estimated based on the description, weight and photographs of the meal, food items were coded and converted into total amount eaten in grams, and energy and macronutrient intake were calculated using the online program Compl-eat35 using the Dutch food composition database (NEVO) edition 2016/5.0.

Directly after the evening meal mothers were asked to rate the child's satiety on a five-point Likert scale (from ‘1 = not at all satiated’ to ‘5 = very satiated’). This was followed by an 8 minute free-play session between mother and child. After this the researcher offered the child a plate with two savoury and two sweet finger foods (see Table 1) for 10 minutes. The researcher told the child that these were for him/her to eat; ‘hello these foods are for you’. Mothers remained in the room but were asked not to interfere with the child's behaviour so the child had the opportunity to continue playing with the toys or eat the provided foods without interference. Finger foods were weighed before and after the EAH procedure and the weight was multiplied by the energy content36 of each individual food to determine the weight (grams) and energy (kcal) consumed by the child respectively. An EAH score was calculated using the following formula37:

A score of 0% indicates that the child did not consume any of the finger foods. A higher score indicates greater energy intake of the finger foods, proportionately to the energy intake of the evening meal.37

2.5 Statistical analysis

Statistical analyses were carried out in SPSS (version 25; SPSS Inc., Chicago, Illinois). For the measurements at 18 months, data of 11 participants (5.1%) were excluded from analysis because parents did not allow their child to participate in the EAH procedure as planned, for instance they did not consent to the standardized amount of finger foods or stopped the procedure within 10 minutes. At 24 months, data of four participants (3.7%) were excluded from the analysis because parents did not comply with the requirements of the procedure as described above.

Data are presented as mean values with standard deviations or numbers with percentages, unless otherwise specified. Tests were performed two-sided and P-values <.05 were considered significant. Distribution of the EAH scores was not normal and, therefore, the EAH score was log transformed. In addition, a one way ANOVA was run to test if EAH scores of the children differed between the four intervention groups, no significant differences were found and data were collapsed for further analysis. Maternal BMI was not related to EAH (r[198] = −.07, P = .31)) and not taken into account for further analysis.

Few children were rated as being ‘not at all satiated’ and ‘not satiated’, the lowest satiety scores after dinner (n < 20 for both measurements). Therefore, these categories were pooled into one group, resulting in a four-point scale for satiety (ie, 1 = not satiated; 2 = neutral; 3 = satiated; 4 = very satiated) for further analyses.

To explore the association of satiety of the child (as perceived by the mother), energy intake of the evening meal, child eating behaviours measured with the CEBQ-T at age 18 and 24 months and child BMI-z score with EAH, we used a Univariate General Linear Model (GLM) with energy intake of finger foods as the dependent variable. Child sex and intervention group were added as independent variables.

Finally, a Univariate GLM was used to investigate the association between EAH score at age 18 and 24 months in the group of children who had participated in both measurements. Sex and intervention group were added in the model as independent variables.

Assumptions for linearity, normality of residuals, homoscedasticity, and multicollinearity were checked. Boxplots revealed two outliers (>3 SD) for EAH-scores. However, these outliers were not due to measurement error and re-running the analyses without these outliers did not change the results. Therefore, outliers were included in the analysis.

3 RESULTS 3.1 Sample characteristics

The final sample consisted of 206 mother-child pairs (n = 98 boys and 108 girls) and 103 mother-child pairs (n = 45 boys and n = 58 girls) at the 18 and 24 month measurement, respectively. Mothers were on average (±SD) 31.6 (±4.6) years old and 21.7% had a university degree or higher. The majority of the participants were from Dutch Caucasian origin (71%), with an additional 4% of mixed ethnicity (one parent Dutch, the other from a different ethnic group). The remaining 25% of participants came from different ethnic groups such as Surinamese, Turkish, and Antillean. Maternal BMI (kg/m2) was on average (±SD) 26.6 (±5.8) kg/m2. The majority of mothers, 53.5%, were overweight (BMI > 25 kg/m2), 45% had a normal weight (BMI between 18.5 and 25 kg/m2), and 1.5% were underweight (BMI < 18.5 kg/m2). Characteristics of the children are presented in Table 2.

TABLE 2. Characteristics of the children at 18 and 24 months 18 Months n = 206 24 Months n = 103 Mean ± SD or n (%) Mean ± SD or n (%) Age (months) 18.0 ± 0.7 24.0 ± 0.7 Sex Boys 98 (47.6) 45 (43.7) Girls 108 (52.4) 58 (56.3) BMI-za Underweight 5 (2.4) 0 (0) Normal weight 185 (89.8) 87 (84.5) Overweight 16 (7.8) 16 (15.5) Eating behaviourb n = 182 n = 98 Enjoyment of food 4.20 ± 0.57 4.06 ± 0.57 Satiety responsiveness 2.35 ± 0.53 2.58 ± 0.49 Food responsiveness 2.66 ± 0.78 2.55 ± 0.74 Emotional overeating 1.80 ± 0.61 1.78 ± 0.67 Slowness in eating 2.94 ± 0.44 3.06 ± 0.48 Food fussiness 2.94 ± 0.26 2.94 ± 0.27 a Underweight when BMI-z ≤ −2; normal weight when BMI-z between −2 and 2; overweight when BMI-z ≥ 2. b Eating behaviour assessed with the CEBQ-T. Scores ranged from 1 to 5 on a five-point Likert scale: 1 = rarely, 2 = never, 3 = sometimes, 4 = often, 5 = always. 3.2 EI, EAH score and mothers perceived satiety

The mean time (±SD) between the evening meal and the EAH task was 21 (±5) minutes. EI from the evening meal (including drink and dessert), finger foods and total energy intake (EI from evening meal + EI from finger foods), and EAH scores are shown in Table 3.

TABLE 3. Evening meal and finger foods intake and eating in the absence of hunger score in the full sample vs satiated children at 18 and 24 months 18 Months 18 Months 18 Months 24 Months 24 Months 24 Months Full sample (206) Satiated childrena (143) Non-satiated childrena (63) Full sample (103) Satiated childrena (50) Non-satiated childrena (53) Energy intake (N) Mean ± SD (min-max) Mean ± SD (min-max) Mean ± SD (min-max) Mean ± SD (min-max) Mean ± SD (min-max) Mean ± SD (min-max) Intake evening mealb kcal 240 ± 117 (17-627) 264 ± 108 (23-597) 186 ± 122 (17-627) 209 ± 106 (8-705) 241 ± 125 (21-705) 178 ± 77 (8-344) Intake finger foods kcal 40 ± 37 (0-237) 42 ± 38 (0-237) 34 ± 34 (0-150) 33 ± 29 (0-113) 35 ± 24 (0-109) 30 ± 25 (0-112) Total energy intakec kcal 280 ± 127 (23-664) 306 ± 118 (23-642)* 220 ± 128 (32-664) 242 ± 113 (36-733) 276 ± 132 (41-733)* 208 ± 83 (36-408) EAH score %d 23.1 ± 52.5 (0-704) 18.7 ± 21.2 (0-158) 33.1 ± 89.5 (0-704) 23.4 ± 38.8 (0-346) 19.9 ± 23.9 (0-128) 26.5 ± 50 (0-346) * Statistically significant (P < .05). a Score reported by the mother, satiety score ≥3 was considered satiated. b Energy intake of the evening meal including drinks and dessert. c Total energy intake = EI intake from the evening meal + EI from the finger foods. d EAH score = EI from finger foods/EI from evening meal × 100%.

Of the full sample, 185 (89.8%) children ate from the finger foods at 18 months with a mean (±SD) intake of 40 kcal (±37), of whom 143 (69.4%) were considered satiated based on the mother's estimation (score 3 satiated or score 4 very satiated). At 24 months, 96 (93.2%) children ate from the finger foods with an average (±SD) intake of 33 kcal (±29), of whom 40 (41.7%) were considered satiated. Twenty-one children (10.2%) ate nothing from the finger foods at 18 months, of whom 17 (81%) were considered satiated. Seven children (6.8%) ate nothing from the finger foods at 24 months, of whom four (58.8%) were considered satiated.

The EAH score between satiated and non-satiated children did not differ significantly (P = .31) and (P = .28) at 18 and 24 months, respectively. Total EI of satiated children was significantly higher than that of non-satiated children (P < .001) and (P = .003) at 18 and 24 months, respectively.

Furthermore, EAH at 18 months predicted EAH at 24 months F(1, 88) = 8.520, P = .004. The proportion of variance explained by EAH at 18 months after excluding variance explained by the other predictors (sex and intervention group) was 8.8% (partial eta squared, ηp2 = 0.088, a medium effect size).

3.3 Association between EI finger foods, EI evening meal, mothers perceived satiety, and eating behaviours

Mother's perceived satiety was significantly positively related to the energy intake of finger foods at child age 18 months, F(3,166) = 3.859, P = .01 (Table 4). Figure 1 shows the energy intake of finger foods grouped according to satiety score. It shows that group 1, consisting of children whose mother reported them to be ‘not at all satiated’ or ‘not satiated’ had a lower energy intake than the other groups. The other three groups (‘neutral’, ‘satiated’, and ‘very satiated’) were quite similar concerning energy intake from finger foods intake. In addition, Figure 1 shows that every satiety group included children who did not eat any finger foods. When removing the group with the lowest satiety score from the analysis, the association between satiety and the energy intake of finger foods was no longer significant F(2,152) = 1.867, P = .16.

TABLE 4. Association between finger food intake, evening meal intake, mothers perceived satiety, and subscores of the CEBQ at 18 (n = 182) and 24 months (n = 98) 18 Months 24 Months Variable Mean intake (kcal) B 95% CI P-value Mean intake (kcal) B 95% CI P-value Intake evening meal 240 0.15 −0.34 - 0.64 .544 209 0.04 −0.03 - 0.10 .264 Satietya .011* .384 1 not satiated 18 −22.91 −47.07 to 1.25 34 8.053 −16.22 to 32.33 2 neutral 42 14.98 −4.79 to 34.75 29 1.956 −20.11 to 24.02 3 satiated 40 1.99 −14.04 to 18.04 37 13.32 −8.05 to 34.69 4 very satiated 49 - - 30 - - Sexb .103 .055 Male 9.09 −1.85 to 20.04 12.48 −0.27 to 25.22 Female - - - - Eating behaviourc 17.06 5.18 to 28.95