The Relationship between Child Eating Behaviour with Body Mass Index among Toddlers Aged Two to Four Years Old

Children eating behaviour was hypothesised as one of the factors associated with growth development as well as malnutrition among toddlers. This study aimed to explore the relationship between child eating behaviour among toddlers aged two to four years old and their respective Body Mass Index (BMI) or BMI for Age Z (BAZ). A cross-sectional studywas conducted among parents and caregiverswith toddlers between two to four years old, in theKlangValley. The validated Children’s Eating Behaviour Questionnaire (CEBQ) was used to measure their child eating habit and behaviours. A total of 96 parents or caregivers of toddlers participated. Girls had lower BAZ than boys (-1.20 vs -0.15, p=0.030), and signi icantly higher satiety responsiveness and food fussiness scores (3.14 vs 2.80, p=0.005 and 2.83 vs 2.44, p=0.005 respectively). Emotional overeating had opposite effects on BMI and BAZ between genders. In this study, girls aged two to four year old had lower BMI and BAZ than boys; and demonstrated higher food avoidance behaviour. Emotional overeating is associated with opposite effects on BMI and BAZ between boys and girls. Among all toddlers, satiety responsiveness and food fussinesswere associated with lower BAZ.


INTRODUCTION
Malnutrition among toddlers includes both undernutrition and a growing disorder with overweight and obesity. The incidence of low body mass index has declined globally in the past two decades, but it remains to be prevalent in Asian and African countries (Black et al., 2013). The prevalence of stunting in children under the aged of 5 years is still an unsolvable problem in many Asian countries (Moench-Pfanner and Bloem, 2013). Malaysia, on the other hand, faces the threat of a double burden of stunting and obesity. As stated by the National Health and Morbidity Survey (National Health and Morbidity Survey, 2015) the prevalence of stunting in children under ive years of age in Malaysia has raised from 17. 2% (2006) to 20.7% (2016). Similar indings were also highlighted on childhood obesity, with 10% of children were overweight, and 12% were categorised as obese (Reeves et al., 2018). It is imperative to elucidate the association between the children eating behaviour and their body weight to design a suitable intervention for healthy eating among children in this age range, thus resolving the issue of double burden.
The Children's Eating Behaviour Questionnaire (CEBQ) is a multi-dimensional, parent or caregiverreported questionnaire, measuring children's  (Svensson et al., 2011;Tay et al., 2016). Eating behaviour, in general, is related and in luenced by psycho-physiological and environmental determinants. However, children below the age of pre-school depend on parental or caregiver feeding practices. Previous studies pointed out that the pressure either to eat or to restrict intake to the children are signi icantly associated with factors of the children weight status or the children behaviour. Children with low body mass index (BMI) are consistently associated with higher pressure to eat from parents (Gregory et al., 2010). We have, therefore, aimed to explore the relationship between child eating behaviour among toddlers aged two to four years old and their respective Body Mass Index (BMI) or BMI for age Z (BAZ).

Study Design and Participants
This cross-sectional study was conducted among parents and caregivers with children aged two to four years old, in the Klang Valley. A total of 96 healthy children were recruited. Syndromic child or children with chronic diseases, bowel or digestive system anomalies, restricted diets or on dietary supplements were excluded from this study. All procedures were conducted following Universiti Sains Islam Malaysia (USIM) Ethics Committee approval

Questionnaires
After written informed consent was obtained, the parent or caregiver completed a series of questionnaires via face-to-face interview, while their children's weight and height were measured for BMI. The demographic questions regarding the children and their parents included gender, parental educational background, monthly total household income and child's medical history, followed by the validated Children's Eating Behaviour Questionnaire (CEBQ).
The CEBQ is contained of 35 items, each evaluated on a ive-point Likert scale. The scale ranges from never to always, based on the eight subscales. The sum and mean of each subscale were calculated as per the questionnaire protocol.

Child Body Mass Index
The anthropometric measurements were performed by a trained research assistant using a standardised digital calibrated loor scale (Seca 217). The scale was placed on irm looring (tile). Bodyweight was measured without clothes, diapers or shoes. While the child stood with both feet on the scale, the recorded weight was taken.
The height was measured using a portable stadiometer (Seca 217). The measurement was taken without shoes with the child's back to the wall while ensuring the child was standing with head, shoulders, buttocks and feet lat. With feet together, heels were ensured to touch the wall with straightening legs and arms at sides. The height was measured by ensuring the corner of the eyes, and the top of the ears are level. Then, the lat headpiece was lowered until it touches the head while the investigator's eyes were at the same level as the headpiece. The height was recorded to the nearest 0.1 cm.
BMI was calculated in kg/m 2 using online BMI calculator, and was converted to standardised z-scores, according to age and gender-based on the Centre for Diseases Control (CDC) growth charts using online Paediatric Z-score calculator from Children's Hospital of Philadelphia Research Institute.

Statistical Analysis
Statistical analyses were performed using the IBM SPSS version 24. Independent t-test was used to determine any difference between boys and girls for their BMI, BAZ and CEBQ sub-scales. Pearson correlation coef icient, r, was used to determine any relationship between CEBQ sub-scales and BMI or BAZ. The signi icant value was taken if p <0.05.

RESULTS AND DISCUSSION
Ninety-six parents of toddlers aged 2 to 4 years old participated in this study. The mean (s.d.) age of the toddlers was 3.24 (0.74) (Table 1). There were slightly more boys (53.1%) than girls (46.9%). Boys had signi icantly higher BMI in comparison to girls, p=0.023 and 0.030, respectively (Table 2). Girls scored higher than boys for satiety responsiveness and food fussiness. There was no signi icant variance between boys and girls with regards to other CEBQ subscales.
For emotional overeating subscale, boys demonstrated a signi icant negative, fair correlation with BMI (r -0.259, p=0.033) and BAZ (r -0.251, p=0.038). In contrast, girls had a meaningful positive, reasonable relationship for this subscale when corresponded to their BMI (r 0.322, p=0.015) and BAZ (r 0.297, p=0.023) ( Table 3). In terms of satiety responsiveness, girls had signi icant negative, fair correlation with BMI (r -0.273, p=0.034). There was also a signi icant negative but weak correlation between food fussiness for all toddlers and their BAZ (r -0.172, p=0.047).
The CEBQ has been developed to describes early precursors of eating disorders among children. Studies have shown that overweight or obese children demonstrate healthier food approach behaviour and weaker food avoidant behaviour (Viana et al., 2008;Webber et al., 2009). Another study conducted among Malaysian schoolchildren has also shown that their body adiposity was positively correlated with food approach behaviour, and negatively correlated with food avoidant behaviour (Tay et al., 2016). These indings strongly imply the contribution of an individual's eating behaviours toward developing weight problems, including overweight and obesity. Hence, intervention strategies focusing on changes in the eating behaviours are strongly recommend.
To the best of our knowledge, this study is the irst to describe eating behaviours associated with BMI-for-age Z-scores (BAZ) among Malaysian toddlers aged two to four years old. BAZ for both genders were within the normal range as according to the Child Growth Standards classi ication by the World Health Organization (WHO) (de Onis and Lobstein, 2010). Hence, the toddlers in this study have a good nutritional status. The present study found that BAZ for girls was signi icantly lower compared to boys. This inding is in agreement with results at the national levels (Khor et al., 2006). However, the difference was not signi icant. A recent inding from the Malaysian National Health Morbidity Surveys in 2015 revealed that more than 80% of Malaysian children age 0-4 years old have normal nutritional status (National Health and Morbidity Survey, 2015). With regards to the eight subscales of CEBQ which were divided into two major dimensions of eating behaviours, that are food approach behaviours and food avoidant behaviours; the present study found that food approach behaviours were not statistically different between genders. Research conducted on similar aged children also showed the same results (Svensson et al., 2011). However, another study among older children has been demonstrated that emotional overeating subscale was signi icantly different between genders (Tay et al., 2016). This might be due to different eating styles between boys and girls.
Additionally, we have found that satiety responsiveness and food fussiness were statistically different between genders. The girls reported higher satiety responsiveness, which could be explained by lower BAZ among girls in this study. In addition to that, higher food fussiness among girls also leads to lower BAZ as compared to boys.
In general, we have found that all four subscales for food approach behaviours had a negligible relationship with BAZ among toddlers age 2-4 years old. However, emotional overeating subscale had a signi icantly negative low relationship with BAZ among boys; and had a signi icantly positive low relationship with BAZ among girls. This showed that girls were likely to eat more while experiencing negative emotions and engagement in this condition positively affects their BAZ.
Slowness in eating and emotional undereating subscales had a negligible relationship with BAZ for both genders. Meanwhile, satiety responsiveness had a signi icantly negative low relationship with BMI for girls. This could account for girls having lower BMI and BAZ in comparison to boys. Collectively, satiety responsiveness for all toddlers had a signi icantly negative low relationship with BAZ. Relationship between food fussiness and BAZ was found to be signi icant. However, the strength of the relationship is negligible.

CONCLUSIONS
In this study, girls aged two to four-year-old had lower BMI and BAZ than boys; and demonstrated higher food avoidance behaviour. Emotional overeating is associated with opposite effects on BMI and BAZ between boys and girls. Among all toddlers, satiety responsiveness and food fussiness were associated with lower BAZ.