Associations between Meal Patterns and Risk of Overweight/Obesity and Metabolically Unhealthy Obesity in Children and Adolescents: A Systematic Review of Longitudinal Studies and Randomised Controlled Trials

1. Introduction

Childhood obesity remains one of the most significant global public health challenges of the 21st century. According to the Non-Communicable Diseases Risk Factor Collaboration, from 1990 to 2018 the age-standardised prevalence of childhood obesity increased in the vast majority of countries worldwide (93% of countries in girls and 98% in boys) [1]. The prevalence of overweight (including obesity) and obesity in Europe remains at alarmingly high rates (29% and 13% in boys aged 6-9 years; 27% and 9% in girls aged 6-9 years, respectively), with significant variations among countries [2] while in the U.S.A., a 20-year analysis highlighted a significant increase of about 30% in the prevalence of obesity and almost a 2-fold increase in severe obesity (defined as Body Mass Index (BMI) ≥120th percentile of U.S. Centers for Disease Control and Prevention growth charts in the U.S.A.) in children aged 2-19 years [3].

Research so far has focused on the causes and risk factors of obesity, such as early life exposures / perinatal factors, growth trajectories, socioeconomic factors, physical environment and lifestyle habits (diet, physical activity, sedentary activities, sleep) [4,5]. The role of diet in childhood overweight/obesity (OV/OB) has been variously studied in terms of nutrients and food groups intake or adherence to dietary patterns [6,7,8]. There is, however, an interest towards other approaches to examine relationships between diet and health or disease that also capture other dimensions of eating habits, such as consumption of meals and snacks. For example, the family meal has been identified as an entry point for intervention, as it constitutes a social setting with the potential to shape children’s eating behaviours from a young age [9]. A dietary approach in the context of meals, e.g. meal preparation, could complement dietary consultations and may be a more practical and easy way to convey clear messages to children and families.

A “meal pattern” has been used in the literature as a term to describe an individual’s dietary habits either at the level of meals or eating occasions [10,11]. The term “eating occasion” or “eating event” or “eating episode” describes any consumption of food based on specific characteristics, such as timing of the day, energy content, combination of foods, usually discerning eating occasions as main meals (i.e. breakfast, lunch, dinner) and snacks [10]. Meals have also been described in relation to their format (i.e. combination of foods or content of nutrients) and the context/environment the meals are consumed (i.e. with the presence of other people or while doing an activity) [11]. There is, however, heterogeneity in the assessment of meal patterns in the literature, as different definitions of a meal have been applied across studies. A meal can be participant-defined through self-reporting [12] or it is defined according to the time it is consumed within the day [13], according to the energy content [14] or a combination of both [10], which may affect the interpretation of the findings [10].

In relation to childhood OV/OB, published systematic reviews have highlighted the role of different aspects of meal patterns on the risk of OV/OB, but evidence derives primarily from cross-sectional studies. For example, a meta-analysis exploring associations between frequency of family meals and children’s health showed that having frequent family meals was associated with a lower BMI and better overall diet quality; nevertheless 67 out of the 75 included studies had a cross-sectional design [9]. Similarly, an earlier meta-analysis of 10 cross-sectional and one case-control studies showed an inverse association of eating frequency (defined as the total number of eating episodes per day) and childhood OV/OB status in boys, but not girls [15]. The systematic review conducted by Monzani et al. [16] included 37 articles, 32 of which were cross-sectional, and it reported that skipping breakfast was associated with an increased risk or prevalence of OV/OB [16]. A meta-analysis that examined the association of meal timing and adiposity showed weak associations between higher energy intake close to bedtime and evening meal skipping with adiposity, but 17 out of 20 included studies had a cross-sectional design [17]. To date, there is no systematic review to capture different dimensions of meal patterns and their associations with the risk of developing childhood OV/OB based on findings from longitudinal or randomized intervention studies that allow exploration of causality.

An emerging issue in relation to childhood OV/OB is the development of associated comorbidities, such as diabetes, hypertension, lipid abnormalities and liver dysfunction, which are often used to define metabolically unhealthy obesity (MUO) [5,18]. Children and adolescents with obesity have an increased prevalence ratio of 1.4 to develop prediabetes, 21.2 for cardiovascular disease and 26.1 for metabolic-associated steatotic liver disease, compared to children with normal weight [18]. Children with obesity are also more likely to live with obesity in adult life which is associated with comorbidities across the life course [5]. The role of diet, let alone of meal patterns, in relation to MUO in children and adolescents has been scarcely explored and very few interventions have evaluated markers of MUO [19,20,21].

Meal pattern consumption remains an interesting dimension of eating habits that could contribute in the engagement of optimal behaviours and lifestyle modification within the context of overweight and associated comorbidities prevention and management. The aim, therefore, of this review was to systematically gather all available evidence from longitudinal cohorts or randomised interventions exploring effects of meal patterns on the risk of developing OV/OB and MUO in children and adolescents.

2. Materials and Methods

The systematic review was conducted according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines [22,23] (Supplementary Materials S1: PRISMA 2020 Checklist). The study was registered in the PROSPERO international prospective register of systematic reviews, of the National Institute for Health and Care Research (Title: The association of meal patterns and risk of obesity and metabolically unhealthy obesity in children and adolescents; registration number CRD42023477708).

2.1. Search Strategy

A systematic search was conducted in May 2024 in two databases (MEDLINE/Pubmed, Scopus) to identify studies which evaluated the role of meal patterns in childhood and adolescent OV/OB risk and MUO risk, published within the last decade (in specific January 2013 – April 2024). Research questions and search keywords were guided from the “Population, Exposure, Comparator, Outcome” (PECO) model for epidemiological studies and the “Population, Intervention, Comparator, Outcome” (PICO) model for interventional studies (Table 1) [24]. Selection of exposures parameters was guided by Leech et al. [11] and included frequency of / omitting meals, consumption of meal within different contexts (e.g. while watching TV) and environments (e.g. at home) and meal quality. Outcome measures related to OV/OB included anthropometric indices, BMI/BMI z-score and body composition parameters. Indicators for MUO included blood pressure, blood lipids, glucose metabolism and metabolic comorbidities. Keywords were formulated according to the PECO/PICO model are available in Supplementary Materials S2.

2.2. Eligibility Criteria

Inclusion criteria reflected the research questions, i.e. the population, the exposures/interventions and the outcomes of interest, as well as the study design. Original, peer-reviewed articles, conducted in children and adolescents (age range 2-19 years old as defined by the World Health Organisation [25] for baseline and follow-up), published from 2013 to 2024, which evaluated outcomes related to OV/OB or MUO were included. Only longitudinal studies and randomised controlled clinical trials (RCTs) were included to ensure a better quality of methodological design that could also allow aetiological assumptions. A minimum follow-up of 12 months was also applied to both study designs, which was deemed adequate to observe meaningful effects of meal patterns on OV/OB and MUO risk. Search was also refined to include only studies conducted in countries with a westernised way of living and in populations of a similar genetic background (i.e. Caucasian origin; Europe, U.S.A., Canada, Oceania).

Studies conducted in animals and people <2 or >19 years old; studies conducted in Asia, Africa and South America; reviews, letters, editorials, review protocols and pre-prints; cross-sectional studies, in-vitro, in-vivo animals or in silico studies and non-randomised, uncontrolled clinical trials were excluded. Also, studies with multidisciplinary lifestyle observations or interventions with no clear analysis of the association between meal patterns and risk of OV/OB or MUO were also excluded. Due to the nature of the research question, studies evaluating meal patterns and MUO indicators in children without OV/OB at baseline and/or follow-up were also excluded.

2.3. Selection of Studies and Data Extraction

All studies identified from databases were imported in the Zotero software (https://www.zotero.org). Following the removal of duplicates, studies were screened for eligibility in two stages. Initially, a title and abstract screening was performed independently by two researchers (GS and AK). After exclusion based on titles and abstracts, all remaining articles were considered for full-text review by both researchers, who applied the eligibility criteria for the final selection. Disagreements at any stage were resolved by a third researcher (EB).

Data extraction was performed by one researcher (GS), with a second researcher (AK) randomly checking a sample of the eligible reports. Any disagreement was resolved by a third researcher (EB). Extracted information included:

-

Study information: first author’s name, year of publication, acronym, country, setting, duration)

-

Population: sample size, baseline age and sex distribution

-

Exposure(s) where relevant: type, definition, assessment method

-

Intervention (where relevant): groups, randomization, components, mode of delivery, duration

-

Outcome(s): type, definition, assessment method

-

Statistical analysis: analyzed sample, statistical model, covariates

-

Study results: main findings

Throughout the process, in case of missing information or uncertainties, relevant information was sought in Supplementary Materials or directly from study investigators. Extracted information is presented according to study design (prospective epidemiological studies or RCTs).

2.4. Risk of Bias

One researcher (GS) assessed the quality of all included studies and a second researcher (AK) assessed a random 20% of the sample. The quality assessment of prospective epidemiological studies was conducted with the Risk Of Bias In Non-randomized Studies - of Exposures (ROBINS-E) tool (https://www.riskofbias.info/welcome/robins-e-tool) [26]. The ROBINS-E tool comprises of seven domains: bias due to confounding; bias arising from measurement of the exposure; bias in selection of participants into the study (or into the analysis); bias due to post-exposure interventions; bias due to missing data; bias arising from measurement of the outcome; bias in selection of the reported result. Each domain and the overall study are assessed as “low risk”, “some concerns”, “high risk” or “very high risk”.

The quality assessment of randomised controlled trials was conducted with the revised Cochrane risk-of-bias tool for randomized trials (RoB 2) tool (https://www.riskofbias.info/welcome/rob-2-0-tool/current-version-of-rob-2) [27]. The RoB2 tool comprises of five domains: bias arising from the randomization process; bias due to deviations from the intended interventions (effect of assignment to intervention or adhering to intervention); bias due to missing outcome data; bias in measurement of the outcome; bias in selection of the reported result. Each domain and the overall study are assessed as “low risk”, “some concerns” or “high risk”.

3. Results

The initial search yielded 3304 results (2493 from Scopus and 811 from MEDLINE/Pubmed). After removal of duplicates (722), 2505 articles were excluded following review of title and abstract and further 49 were excluded after full-text examination. Twenty-eight reports from 25 studies were included in this review (Figure 1). All reports examined associations between meal patterns and childhood OV/OB risk. No studies were found in relation to MUO risk.

3.1. Study Characteristics

Of the 28 included reports (25 studies), 13 studies were conducted in the U.S.A. [20,28,29,30,31,32,33,34,35,36,37,38,39], 12 in Europe (U.K. [40,41,42], Netherlands [43,44], Germany [45,46], Republic of Ireland [47], Spain [48], Norway [49], multicentre across different European countries [50,51]), one in Australia [52] and one in New Zealand [53]). Finally, one study presented data from independent studies in different countries (Germany, the Netherlands, U.K., U.S.A.) [54]. All but one studies had a prospective observational design, with a follow-up range between one and 10 years and one study employed a RCT design [39], with a follow-up of 2.5 years. Analytic sample sizes ranged from 116 to 23,307 participants. The majority of studies recruited children and adolescents from school settings [29,30,31,32,33,34,35,37,38,39,44,45,46,47,48,49,51,52,54], and fewer from clinics [20,28,32,43,53] or the general population [36,40,41,42,54], while one study did not provide relevant information [50]. Most studies focused on school-aged children (n = 9) [29,30,31,37,38,45,46,47,51], while others included both pre-schoolers and school-aged children (n = 7) [20,28,35,40,41,50,54], both school-aged children and adolescents (n = 5) [36,39,44,48,49], only pre-schoolers (n = 4) [32,42,43,53] and only adolescents (n = 3) [33,34,52].

Regarding bias, most prospective observational reports had an overall high risk of bias, mainly due to risk of bias arising from measurement of the exposure (Table 2). For example, in several studies, data on exposures derived from self-reported questionnaires, often completed by parents/guardians. It was also unclear how meals were defined and whether clear instructions were provided to the participants on what constituted a meal. Only four reports raised “some concerns” in relation to bias [20,29,32,52] and no report had a low risk of bias. Most reports had a low risk of bias in relation to selection of participants (28/28), post-exposure interventions (24/28) and selection of reported results (23/28). The intervention included in this review [39] had a low risk of bias in four domains (randomisation process, missing data, measurement of the outcome and selection of reported result) and concerns raised in relation to one domain (deviation from intended intervention) (Table 3).

3.2. Main Exposures

A number of different exposures in relation to meal patterns were identified. The most commonly studied exposure was frequency of consumption of/ skipping specific meals, such as breakfast [20,29,31,33,34,36,37,41,43,44,45,46,47,51,54], dinner/evening family meals [20,29,34,43,51], and lunch at home or at school [29,30,36,43,51]. In most studies, meals were self-reported without evidence of a clear, objective definition on the timing or the content of a meal consumed. One study provided a definition of breakfast [31], that is “the first meal in the morning consisting of any solid food, beverages, or both and named by the respondent as ‘breakfast’”. In one study [34], participants were asked how often they had breakfast “which was more than a glass of milk or fruit juice”, whereas in another study [33] they were asked what they usually have for breakfast “on a weekday morning”. Regarding lunch, two studies [29,30] defined school-provided lunch as “a full meal including salad, soup, a sandwich”. Relevant information about how meals were defined and assessed are included in Table 4 and Table 5.

Included studies also examined the association of the consumption of other meal patterns with childhood OV/OB, such as snacking [36,47] (one study [47] provided examples of snacks), eating fast foods [20,34,36,47,52] (three studies provided examples [34,47,52] of fast foods) and having sugary drinks in between meals [40]. Other exposures related to meal patterns included frequency of family meals [28,36], family meal interpersonal quality [28], regular timing of meals [40], eating meals while watching TV [20,34,36], eating while doing homework [36], eating alone [36], eating with friends [36], first and last eating events [32], mealtime setting [36,42] and patterns of breakfast location (combination of the variables “frequency” and “setting” in relation to breakfast) [38]. Some studies examined the association of meal frequency/ eating occasion in childhood obesity, however the definition of meal/eating occasion varied significantly among studies. In Jaeger et al. [50], an eating occasion was defined as any occasion where food or beverages are consumed. In Taylor et al. [53] a separate eating occasion was defined as “the start of the next meal or snack that had to be more than 15 min after the end of the previous meal or snack (i.e. separated by at least four five-minute blocks)”. Stea et al. [49] defined regular breakfast/lunch/dinner/evening meal consumers if eating all meals every day. Two studies assessed meal frequency as the combined frequency of breakfast and evening meals [30,35]. In one study the frequency of eating meals with family was evaluated on a continuous scale from 0 (never) to 8 (>7 meals per week); however, it was unclear how family meals were defined [28].

Data on exposure(s) derived primarily from self-reported questionnaires completed by either parents/caregivers [20,29,30,41,42,43,45,46,47,49,51,54] or children/adolescents [20,33,34,38,44,48,52]. Two studies used dietary recalls [31,32], three studies used dietary records/diaries [36,50,53] and three studies collected data via interviews with parents/caregivers [35,37,40]. All but one studies analysed the exposure of interest as a categorical variable; Jaeger et al. [50] explored meal patterns as the amount of energy intake (kcal) in predefined time slots.

The only cluster-RCT included in this review assessed the effect of eating breakfast in the classroom and of providing breakfast-specific nutrition education in comparison to having breakfast in the school cafeteria, on overweight and obesity among urban children in low-income communities. Data on exposure were collected by teachers (intervention arm) and cafeteria staff (control arm) [39].

3.3. Main findings

Findings are presented in three axes, according to the definition of meal constructs as described by Leech et al. [11]:

a)

meal patterning, including frequency of eating occasions, regularity of meals, meal skipping, meal timing, and spacing of eating occasions

b)

meal format, referring to food type, food combinations, or food sequencing;

c)

meal context, related to the presence of others at a meal, eating while doing activities, meal location.

3.3.1. Meal Patterning

Meal patterning essentially refers to the frequency or timing of eating occasions, either examining the number or the distribution of meals/snacks within the day or focusing on the regularity/skipping of a specific eating occasion, with special interest in breakfast. In relation to meal frequency, with different definitions been used to describe the term, most studies showed no association with adiposity parameters. One study found that higher meal frequency (5 meals per day compared to 3 or fewer meals per day) at baseline was associated with a smaller increase of BMI-z score, a smaller increase in waist-to-height ratio and lower odds of developing obesity at follow up [48]. However, no association of meal frequency with obesity risk [29] or change in BMI-z score at follow up [35,49,53] was found in other studies. A dose-response association was observed between the regularity of mealtimes at age 3 and the risk of developing obesity at age 11 [40]. Specifically, compared to children who always had regular mealtimes, those who usually had regular mealtimes experienced a 23% reduction in the odds of obesity, while those who rarely or never had regular mealtimes had a 38% reduction [40]. Moreover, timing of the first eating episode at baseline was not associated with fat mass, fat-free mass or body fat% in children 3-5 years old at follow-up one year later [32]; however, the same study showed later timing of the last meal of the day at baseline to be associated with increased fat mass and body fat% at follow-up one year later [32].

Breakfast constitutes the most studied meal within meal patterning. Daily consumption of breakfast at baseline compared to less frequent consumption (<7 times/week) had a favourable association with BMI [20] and body fat mass % [20,43] at follow-up, whereas no association with odds of OV/OB [43] was observed. In the de la Rie et al. study [54], such benefit (association with lower BMI) was evident in two of the four cohorts examined. Other studies showed no association of breakfast frequency with OV/OB incidence or/and prevalence [44,47] or change in BMI-z score [34,47]. Moreover, breakfast eating habits have been also assessed as skipping breakfast, with mixed results in relation to obesity risk, and a varying definition of the term across studies. One study used a yes/no variable [41], two studies assessed meal skipping as consuming it “never/rarely” compared to “often/always” [45,46], one study included the category “frequent skippers” as having a meal 0-3 times/week [38], another named a stable breakfast skipping pattern as “eating breakfast <7 times/week” consistently over time [43], and one study assessed meal skipping if participants reported they “did not eat” that meal (option 0) on a 0-7 times/week scale [33]. In most studies, skipping breakfast was associated with weight gain [41,45], increased waist-to-height ratio [45], increased BMI percentile [45], abdominal obesity [45] , increased % of body fat mass [43], and increased overweight risk [45] and OV/OB risk [38]. However, skipping breakfast (yes vs. no) was also associated with decreasing weight trajectory, which was defined as a change in BMI category over time i.e. from overweight at baseline to normal weight at follow-up) [41]. In three studies no association of skipping breakfast with overweight risk (in males) [33], obesity risk [33,45] or abdominal obesity risk [46] was observed.

Child’s sex may also play a role, with some findings indicating a beneficial effect of breakfast consumption especially in girls. In a study with both sexes, breakfast consumption was associated with waist circumference, trunk fat mass and trunk to peripheral fat mass ratio only in girls [20]. Another study conducted only in girls, following them from childhood through adolescence, identified eating breakfast at age 9 (without specifying frequency) as a significant protective predictor against adiposity at age 19 [36]. Skipping breakfast (no breakfast consumption in any day of the week) was also associated with increased OV/OB risk only in girls in one study, but not in boys [33].

Consumption of lunch (as a yes/no variable) and frequency of its consumption were not associated with any parameter of adiposity [30,43,51], but skipping lunch at 9 years of age was identified as a significant risk predictor of adiposity in females at age 19 in one study [36]. Regarding dinner, eating dinner <6 times/week was not associated with changes in fat mass or BMI over time compared to eating dinner every day [43].

3.3.2. Meal Format

The energy and macronutrient content and consumption of specific foods have been examined within the meal format, with a special focus on snacks and fast-food consumption. One study investigated the time-of-day energy intake in relation to obesity risk and found that energy and macronutrient intake distributed in different eating occasions throughout the day, was not significantly associated with BMI-z score [50]. Regarding the content of breakfast, higher frequency consumption of Ready-to-Eat-Cereals (RTEC, 3 vs. 1 times per week of breakfast RTEC) in fourth grade was associated with a decrease of about 2 percentiles of children’s BMI in sixth grade [31].

Children who abstained from sweet and savoury snacks at baseline had decreased OV/OB prevalence, decreased OV/OB risk and decreased change in BMI-z score at follow-up, compared to regular eaters on a daily basis [47]. On the contrary, consumption of sugary drinks in between meals was not associated with obesity risk in one study [41]. Similarly, consumption of fast foods/ takeaway foods was not associated with changes in BMI [34,47,52], body fat % [52], waist circumference [52], odds of developing obesity [52] or change in OV/OB prevalence [47]. However, one study showed that consumption of fast foods less than once a week at baseline was associated with lower BMI-z score, waist circumference, body fat %, trunk fat mass and trunk to peripheral fat mass ratio at follow-up, but only in girls [20].

3.3.3. Meal Context

The aspects of meal context studied were ‘eating with whom’, ‘what doing in parallel’, and ‘place of eating’ and in particular, eating with family, eating while doing another activity, esp. screen use, and eating at school. Studies have mainly focused on breakfast and dinner, with scarce reports on lunch.

Family meals were associated in one study with reduced obesity prevalence at follow-up, with each additional family meal significantly reducing obesity prevalence by 4% [28]. Two studies assessed eating breakfast with family, and both showed beneficial effect [29,51]. Children who consumed family breakfast three to seven times a week at baseline had decreased odds of OV/OB two years later, compared to those who never had breakfast with family [51], while children with overweight at baseline and healthy weight at follow-up were more likely to eat breakfast together with at least one member of the family than children with overweight, both at baseline and follow-up [29]. Four studies examined the role of family dinner in obesity risk [20,29,34,51], two of which found a protective role, but the other two no association. In one study, children who had family dinners three to seven times per week at baseline were more likely to have lower BMI and reduced odds of overweight/obesity two years later [51]. Similarly, daily family dinners, in comparison to non-daily dinners (≤6 times/week), were associated with decreased BMI-z score, waist circumference, trunk fat mass and trunk to peripheral fat mass ratio, in girls but not boys [20]. On the other hand, two studies found no association of frequency of family dinners and BMI over time [29,34].

Four studies examined the effect of watching TV while eating meals [20,34,36,42]. In specific, one study showed lower BMI-z score, waist circumference, body fat%, trunk fat mass and trunk to peripheral fat mass ratio in boys and lower body fat% in girls that ate meals while watching TV less than once a week, compared to those who ate meals in front of TV more frequently [20]. Greater mealtime screen use and less use of a table while eating a meal (described as intermediate and informal meal settings) was associated with increasing overweight weight trajectories (i.e. being at the “normal weight” category at baseline and at the “overweight” category at follow up), compared to lower mealtime screen use and eating a meal on the table (formal meal setting) [42]. One study, however, found that eating while watching TV at age 9 was a significant protective predictor of adiposity in females at age 19 [36], which was highlighted as “surprising” by authors, while no associations were observed in another study [34]. Also, a more positive interpersonal quality of family meals at baseline, characterised by the involvement of conversations without media distractions, was not associated with obesity prevalence at follow up, compared to a less positive interpersonal quality of family meals that involved more media distractions and less conversations [28]. Finally, in one study, eating with friends, eating while doing homework and eating in the bedroom were also identified as protective factors against adiposity in females, findings which have also been highlighted as “surprising” by the authors, who nevertheless commented that the findings may imply a mitigation effect of such eating behaviours when healthy snacks are offered, assumption which might be supported by their finding of “parents fixing a snack” being a protective adiposity predictor [36].

Meals served at school have been also studied in regards to obesity risk [29,37,39]. Consumption of breakfast at school (yes/no) in fifth grade was not associated with obesity risk in eighth grade [37], neither an intervention to increase breakfast consumption at school had an effect in the combined OV/OB incidence or prevalence [39]. However, when incidence and prevalence of obesity were examined separately, the incidence of obesity alone was higher in intervention schools than in control schools (11.6% vs 4.4%; OR 2.43; 95% CI, 1.47-4.00), as well as prevalence of obesity (28.0% vs 21.2%; OR 1.46; 95% CI, 1.11-1.92) after 2.5 years of intervention [39]. Only one study examined consumption of school-provided lunch (yes/no), and found that it was associated with decreased odds of weight gain at follow up 3 years later [29]. In all studies, no information on whether a school-provided meal reduced meal skipping, replaced a meal eaten at home or was consumed in addition to a home-provided meal, was mentioned.

4. Discussion

This systematic review collected evidence on the associations of meal patterns with childhood OV/OB and MUO risk. In the absence of published studies on MUO risk, only evidence on OV/OB was presented. To our knowledge, this is the first systematic review that captures different dimensions of meal patterns (patterning, format, context) and focuses on studies with a longitudinal design (prospective studies and randomised controlled trials), aiming to explore aetiologic associations between meal patterns and OV/OB and MUO risk. Regular consumption of breakfast and family meals, as well as avoiding watching TV while eating, may be protective factors against childhood OV/OB, whereas meal skipping (primarily breakfast) may be a detrimental factor. Mixed effects of meal frequency on OV/OB risk were observed. No effects were observed regarding frequency of lunch consumption or of fast-food consumption and of meals served at school, while there was insufficient evidence to support the role of other meal patterns such as meal timing and eating in social contexts other than family in OV/OB risk.

Despite methodological considerations, frequent/daily consumption of breakfast (or avoiding breakfast skipping) has been, according to current findings, highlighted as a protective factor against childhood OV/OB longitudinally, indicating a potential aetiologic association between breakfast consumption and OV/OB risk. Recently published systematic reviews of both cross-sectional and longitudinal studies are in agreement with these findings, particularly regarding breakfast skipping [16,55]. This constitutes an important public health message for preventive policies and practices, especially taking into consideration that literature shows a discernible decline in breakfast consumption in children while entering adolescence [56]. Promotion of frequent/daily consumption of breakfast throughout childhood and adolescence would be an efficient strategy towards reducing OV/OB risk.

The self-reported data collection methods employed in most included studies might have variously impacted current findings. For example, most of the studies did not provide guidance as to what constitutes a meal and a small number of studies used varied definitions, leading to ambiguities regarding the studies’ comparison and results interpretation. Participant-identified eating occasions are prone to subjectivity, as the interpretation and allocation of each eating event may vary significantly [11]. Future studies should ensure the use of robust definitions for meals, such as the proposed definition for breakfast by O’Neil et al. [57] as “the first meal of the day that breaks the fast after the longest period of sleep and is consumed within 2 to 3 hours of waking; it is comprised of food or beverage from at least one food group, and may be consumed at any location”. Lack of a universally-accepted definition of breakfast may prove challenging in the identification of pathways through which breakfast consumption may play a significant role in the development of childhood OV/OB and MUO. Also, the use of methodologically more reliable assessment tools for the evaluation of meals in future research, such as dietary records and recalls, will not only provide more reliable results on the role of breakfast/meal frequency, but also on the role of meal format/composition in the development of OV/OB.

Results from this review also highlighted a potential protective effect of family-shared meals against OV/OB risk, which is supported by older systematic reviews/meta-analyses of both cross-sectional and longitudinal studies [9,58]. The parental influence in shaping children’s eating behaviours has been well documented, from either the perspective of parenting feeding practices, dietary habits or weight status [59,60,61]. Family shared meals could be considered a factor under the construct of home organisation within the family environment, and associations have been found between family meal routines and childhood and adolescence obesity markers, after correcting for the moderating role of socioeconomic factors [62]. It would be interesting to explore which parameters of the family meal routines and family environment mediate this protective effect towards OV/OB risk. For example, an earlier, cross-sectional study by Skafida et al. [63] showed that parameters such as children eating the same food as their parents, having conversations with parents during meals and having “enjoyable” mealtimes were positively associated with children’s diet quality, while eating with parents at the same time was not a significant predictor [63]. Only one study in this review [28] assessed interpersonal quality of family meals and specifically conversations with family during meals, compared to media distraction, with no significant results in relation to OV/OB risk. Future research should focus on the longitudinal effect of family meal routines and related parameters, which could also include food availability, food quality and consumption of homemade foods on the OV/OB risk. The family environment should be also considered as a potential target for future interventions that will focus on the improvement of the frequency and quality of family meals and explore their effect on weight status.

Included studies that investigated the effect of eating meals while being distracted by media and particularly by watching TV on OV/OB risk indicated a protective effect when abstaining from TV while eating. Findings from a systematic review showed a positive association between TV viewing and consumption of energy-dense foods, such as pizza, fried food, sweets and sugar-sweetened beverages, and a negative association with consumption of fruits and vegetables [64]. TV viewing has been hypothesised to increase OV/OB risk through increased sedentary time [65,66], influence of TV advertisements of energy-dense foods [67], promotion of mindless eating during viewing [68], and increased snacking [66].

Although this review identified several longitudinal studies involving the family environment, considerably fewer assessed meal patterns in the school environment (in the form of school-provided meals); thus, there is insufficient evidence to support a positive or detrimental effect on children’s weight status. Evidence from two large free school meal programs in the U.S.A., the Community Eligibility Provision and the Healthy, Hunger-Free Kids Act have shown mixed results on childhood obesity trends [69,70]. Provision of free breakfast and lunch to students had led to a modest decrease in obesity prevalence in the Community Eligibility Provision program within 5 years of implementation [69], whereas no effect was recorded in the Healthy, Hunger-Free Kids Act study on obesity risk [70]. These results may be due to the fact that it is unknown how school meals affect children’s daily meal and dietary intake, that is whether they replace existing meals or are added to a given meal pattern leading probably to increased dietary intake. Even though it may be unclear whether provision of meals at schools are associated OV/OB risk, such school programs have documented improvements in diet quality [71], food security [72] and academic performance [72]. It would be essential for future research studying the effect of school provided meals on childhood OV/OB risk to assess the quality and quantity of school-provided meals, including snacks, and explore whether they replace or are consumed additionally to the rest of the meals by correcting also for total energy intake as a confounding factor.

Some evidence suggests that sex might have a mediating role in meal patterns and OV/OB risk, and this was mainly observed in the literature exploring the effect of breakfast consumption. According to findings from two studies [20,33], regular breakfast consumption was associated with decreased adiposity markers, while skipping breakfast with increased OV/OB risk in girls, but not boys. One of these studies also found a protective effect of daily family dinners on obesity markers in girls but not in boys. Mahmood et al. [51] suggested that a potential explanation for sex differences might be due to differences in dietary and social behaviours. Girls tend to eat more frequently with family and friends than boys [73], whereas boys eat more takeaway meals at home, compared to girls [74]. Also, girls may be more prone to societal influences of dieting, leading to increased prevalence of meal (particularly breakfast) skipping [73]. It is unknown whether dieting and social behaviours would explain sex difference in the included studies. One of the studies in this review explored the impact of dieting on OV/OB risk and found no associations regardless of sex, but the variable “dieting” was not included as a confounder for the association of breakfast skipping and OV/OB risk. The other study did not explore other dietary or social behaviours that could potentially explain the different results according to sex. Biological dimension of sex mediating the role between frequency of consumption of meals and OV/OB risk is also largely unknown and constitutes a potential area for future research.

It should also be noted that the current review identified only one RCT regarding meal patterns (provision of breakfast) and OV/OB markers and no studies for MUO risk. Given the importance of preventative measures against obesity-related comorbidities, future research should focus on the role of meal patterns and the family environment in the development of comorbidities in children with OV/OB. Also, future interventions could target aspects of meal patterning, format and/or context in preventive interventions.

This review has strengths and limitations. In order to promote reproducibility and transparency, the PRISMA guidelines and a detailed search strategy were implemented. The inclusion of most recent studies with a longitudinal design is considered a strength of the review, so evidence on potential aetiologic associations of the role of meal patterns and OV/OB risk and MUO risk is synthesised. Findings are, however, affected by the scientific quality of the included studies, presenting in most of the cases with high or very high risk of bias due to the large heterogeneity on the assessment of exposures [particularly deriving from self (parent)-reported questionnaires]. Also, search was limited in two, widely available databases (Pubmed/MEDLINE and Scopus) and reference lists of included studies; it was not possible to conduct the search in databases where subscription was required. Finally, included studies were conducted in Westernised countries, limiting the generalisabilty of the findings in non-westernised cultural settings.

5. Conclusions

Some evidence supports a protective role of regular/daily breakfast consumption, regular shared family meals and avoiding watching TV while dining against OV/OB risk, deriving mainly from longitudinal studies, while no relevant published study reporting on childhood MUO risk was identified. There was insufficient evidence of the role of other meals such as lunch, other meal patterns such as meal timing, and other social environments such as consuming meals at school, in OV/OB risk. Overall, the quality of the findings is poor due to the high bias of the included studies. The use of harmonised definitions for the assessment of different meals, as well as better methodological approaches, are warranted to provide more robust results in future studies. Future interventions should also target the family environment with a view of determining the protective parameters of shared family meals in OV/OB risk and MUO risk.