Prevalence of Metabolic Syndrome among Children and Adolescents in High-Income Countries: A Systematic Review and Meta-Analysis of Observational Studies

St. Paul’s Hospital Millennium Medical College, Addis Ababa, Ethiopia Ethiopian Public Health Institute, Addis Ababa, Ethiopia College of Medicine and Health Sciences, Hawassa University, Hawassa, Ethiopia College of Health Science, Debre Markos University, Debre Markos, Ethiopia Australian Centre for Public and Population Health Research, School of Public Health, Faculty of Health, University of Technology Sydney, Ultimo, NSW, Australia College of Health and Medical Sciences, School of Nursing and Midwifery, Haramaya University, Harar, Ethiopia

The definition of MetS in children and adolescents remains unclear due to the absence of gold standard diagnostic criteria of MetS for the pediatric population [9]. Some of the diagnostic criteria used by studies include the International Diabetes Federation (IDF) criteria [10], the World Health Organization (WHO) criteria [11], the National Cholesterol Education Program Adult Treatment Panel III (NCEP-ATP III) criteria modified for age [12], the de Ferranti et al. criteria [13], the Weiss et al. criteria [14], and the Cruz and Goran criteria [15].
Globally, an estimated 3.3% with a range of 0.2% to 38.9% of children and adolescents were expected to live with MetS. The prevalence was considerably higher in the overweight (11.9%) and obese (29.2%) population [9,16,17]. Likewise, the prevalence of MetS is remarkably higher in high-income countries (HICs) due to increasing trends of childhood obesity rates [18]. The rise in obesity in the past four decades could be primarily associated with related lifestyle factors such as routine consumption of fructose in the form of soft drinks, juice, and baked goods [19][20][21][22]. Thus, obesity increased MetS in children and adolescents from 6% to 39% [23].
Metabolic syndrome has been a global pandemic affecting children and adults [24]. The burden is significantly higher in the developed world posing a high economic burden on the health care system [25]. Cardiovascular and other metabolic complications are also common consequences of MetS in children [26]. In addition, MetS also negatively impacts the mental status and over all cognitive performance of children and adolescents [27]. In spite of the fact that multiple treatment strategies were designed and implemented, the prevalence of MetS remained high in most high-income countries with a remarkable variation among countries [28,29]. Primary studies substantiated this by showing that the prevalence of MetS in the general population ranged from 0.4% [30] to 24% [31]. Similarly, the prevalence in the obese population ranged from 6% [32] to as high as 55.8% [33]. There is also considerable variation among the diagnostic methods of MetS in the pediatric population [34].
Though comprehensive systematic reviews and metaanalyses are vital for evidence-based decision making, they are scant in HICs where the burden of MetS is undoubtedly higher. Hence, this systematic review and meta-analysis is aimed at determining the pooled prevalence of MetS among children and adolescents in HICs and at giving conclusive evidence about its burden in these countries. The findings will be vital for policy makers and program planners in crafting preventive and treatment measures. The current findings will be supplementary for assessing the progress of sustainable development goals, specifically, ending all forms of malnutrition by 2030 [35]. In addition, the findings of this study will have a pivotal implication to conduct original studies on a multitude of factors related to high-burden MetS among the pediatric population.

2.
1. Data Sources and Eligibility Criteria. Studies performed in HICs with the aim of identifying MetS among children and adolescents were included in this systematic review and meta-analysis. The eligibility of the studies was verified prior to inclusion to this study using study area, study setups, title, abstract, and full texts. The Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guideline [36] was followed in the write-up process of the whole document. We explored national surveys and published and unpublished studies conducted in English. The reference lists of selected articles were also cross-checked for additional articles that were not found using search strings. Studies conducted until September 2020 were searched. Finally, observational studies reporting the prevalence of MetS among children and adolescents conducted both in clinical and community-based setups were included.
Conversely, studies with incomplete or unclear diagnostic methods and studies without full texts were excluded. We communicated with the corresponding authors using email before making the decision to exclude studies without full texts. Letters to editors, conference proceedings, and qualitative studies were also excluded. The EndNote X8 reference manager was used to manage the retrieved articles.

Search Strategies and Study Selection Process.
A comprehensive search was performed by three investigators (ZWB, ZT, and TW), independently. Literature searches were conducted for studies published up to September 2020 using databases such as PubMed, Scopus, Web of Science, CINAHL (EBSCOhost), EMBASE (Elsevier), and Medline (EBSCOhost) as well as other sources (Google Scholar and Google). The following key terms were used for searching: (a) population (children, child, school age, and adolescent), (b) exposure (associated factors and risk factors), (c) outcome (metabolic syndrome, MetS, components of metabolic syndrome, and cardiovascular risk factors), (d) study design (cohort studies, cross-sectional studies, epidemiology, observational studies, and national health surveys), (e) study setting (school, community-based surveys, and health institutions), and (f) location (high-income countries, HICs, developed countries, and names of high-income countries). The Boolean search operators "OR" and "AND" were used during the searching process, and the appropriateness of the key terms were checked before conducting the search in each of the explored databases. An example of a search string in PubMed is shown in Table 1 [37]. The tool has four options (yes, no, unknown, and not applicable). One is given for yes, and zero was given for the other options. The scores were added up and changed to percentages. The minimum score was zero, and the maximum score was eight. Studies with >50% quality scores were included in this meta-analysis (Additional file 1). The interrater agreement was computed by an author (ZT) after the critical appraisals and prior to the final decision of inclusion. The interrater agreement was computed using Cohen's kappa coefficient (κ). The findings revealed that there was substantial agreement [38] between the two raters (κ = 0:784; p ≤ 0:001).

Summary Measures.
Metabolic syndrome among children and adolescents in HICs with different diagnostic methods is the primary outcome of this study. The pooled prevalence of MetS was calculated in the general population and overweight and obese adolescents, separately. The general population included underweight, normal weight, overweight, and obese children and adolescents. The other outcomes were components of MetS and the pooled prevalence of MetS among the male and female population. The recent list of HICs was obtained from the World Bank database [39]. The prevalence was calculated by dividing the total number of events (MetS) to the total sample size and multiplying it by 100. The binomial distribution formula was used to compute the standard error for each original study. The pooled estimates were computed from prevalence and standard error of prevalence using the "metan" commands in the STATA (version 15) software. The pooled estimates were presented with their 95% CIs. The effect sizes were prevalence of MetS in HICs and the respective components.
Regarding the diagnostic criteria, the pooled estimates of MetS in HICs were computed using six diagnostic methods. The respective definitions are presented as follows. In the IDF diagnostic criteria, MetS is diagnosed if children aged between 10 and 16 years have central adiposity (≥90th centile) and two of the following: triglycerides ðTGÞ ≥ 150 mg/ dl, HDL − C < 40 mg/dl, systolic blood pressure ðBPÞ ≥ 130 mmHg or diastolic BP ≥ 85 mmHg, and fasting plasma glucose ðFGÞ ≥ 100 mg/dl or previously diagnosed type 2 diabetes [10]. Based on the WHO criteria, MetS is diagnosed when three or more of the following criteria are met: body mass index ðBMIÞ > 95th percentile, hyperinsulinemia or impaired fasting glucose or impaired glucose tolerance, BP > 95th percentile, TG > 105/136 mg/dl (1.2/1.5 mmol/l) for children aged <10 and >10 years, respectively, and HDL − C < 35 mg/dl (0.9 mmol/l) [11]. Based on the NCEP-ATP  Filters: free full text, observational study, in the last 10 years, humans, English, child: 6-12 years, adolescent: 13-18 years 3 BioMed Research International III criteria modified for age, MetS is diagnosed when three of the following criteria are met:TG ≥ 110 mg/dl, HDL − C ≤ 40 mg/dl, systolic BP or diastolic BP ≥ 90th percentile, waist circumference ≥ 90th percentile for age and gender, and FG ≥ 110 mg/dl [12]. According to de Ferranti et al., MetS is a clustering of at least three of the following criteria: FG ≥ 110 mg/dl; HDL − C ≤ 50 mg/dl (except in boys aged 15 to 19 years in whom the cut-off point is 45 mg/dl); TG ≥ 100 mg/dl; systolic BP > 90th percentile for gender, age, and height; and WC > 75th percentile for age and gender [13]. According to Cruz and Goran, MetS is defined as the presence of at least three of the following abnormalities: abdominal obesity (WC > 90th percentile for age and gender), hypertriglyceridemia (TG > 90th percentile for age and gender), low HDL-C (HDL − C > 10th percentile for age and gender), hypertension (systolic or diastolic blood pressure > 90th percentile adjusted for height, age, and gender), and impaired glucose tolerance [15]. Furthermore, Weiss et al. diagnosed MetS when three or more of the following are obtained: obesity (BMI Z score ≥ 2:0), fasting glycemia (glycemia at oral glucose tolerance test of 140-200 mg/dl), elevated BP (BP > 95th centile), low HDL-C (HDL − C < 5th centile), and high TG (TG > 95th centile) [14].
2.6. Statistical Methods and Analysis. In this meta-analysis, STATA version 15 (STATA Corporation, College Station Texas) software was used to calculate the pooled estimates. The pooled estimates were computed using both random and fixed effect models. In the presence of high heterogeneity among studies, the pooled estimates were computed using random effect models and were weighted using the inverse variance method. Subgroup analyses were performed using different parameters. The pooled estimates in the general and overweight and obese population were presented separately. For the subgroup analysis, data were extracted based on study continent, study country, and gender of study subjects. The appropriateness of each datum was verified before the analyses. Forest plots, summery tables, and texts were used to present the findings of this study.

Publication
Bias and Heterogeneity. Publication bias was assessed using the funnel plot and Egger's regression test at a 5% significant level [40]. Heterogeneity among included studies was explored using the forest plot, the I 2 test, and the Cochrane Q statistics [41]. The I 2 values of 25%, 50%, and 75% were interpreted as low, medium, and high heterogeneity, respectively [42]. In the present meta-analysis, significant heterogeneity was considered when the I 2 value was ≥50%, with a p value < 0.05. The possible sources of significant heterogeneity were addressed through subgroup and sensitivity analyses.

Characteristics of the Included Studies.
All studies included in this study were cross-sectional studies. Out of the total 77 studies, 49 studies were conducted among the general population of children and adolescents [13,30,31,. The remaining 28 studies were performed in the overweight and obese population [14,15,32,33,. In this review, 125,445 study participants were included, of which 113,742 were from the general population and 11,703 were from the overweight and obese population. In the overweight and obese population, the sample size ranged from 97 [136] to 1241 [119] children. Likewise, the sample size in the general population ranged from 234 [111] to 12,147 [30]. The age range of study subjects in both groups was between 2 and 19 years. Regarding geographic distribution of studies, 34 studies were conducted in Europe, while 23, 16, 2, and 2 studies were conducted in Asia, USA, Canada, and Latin America, respectively. The quality of articles was also assessed using the JBI checklists. Thus, 48 studies were classified under medium quality, and 29 studies had high quality (Tables 2 and 3).

Metabolic Syndrome among Overweight and Obese
Children and Adolescents. The pooled prevalence of MetS was estimated using five diagnostic methods (IDF, ATP III, de Ferranti, Weiss, and WHO). In the IDF diagnostic method, thirteen studies [32,33,119,126,[128][129][130][131][132][134][135][136]  The pooled prevalence of MetS was also estimated among males and females. The prevalence of MetS was relatively higher in males (26.62%) than in females (20.18%) in the IDF method. However, the pooled prevalence was nearly similar among males (24.75%) and females (24.97%) in accordance with the ATP III diagnostic method ( Figure 2 and Table 4).

Metabolic Syndrome among the General Population of
Children and Adolescents. In the general population of children and adolescents, the pooled prevalence of MetS was computed using the IDF, ATP III, de Ferranti, Cruz [13,31,78,106]. In accordance with the Cruz and Goran diagnostic criteria, the pooled prevalence of MetS was computed from two studies [75,78], and it was found to be 4.66% (95% CI: 3.29, 6.03; I 2 = 76:6%; p ≤ 0:01  (17) Published not in English (2) Incomplete records (5) Studies obtained from other sources = 105                The gender-based distribution of MetS in the general population was also estimated in all diagnostic methods. The pooled prevalence of MetS among males was higher than females in the IDF (3.80%, 2.37%), ATP III (6.61%, 4.65%), and Cruz and Goran (5.53%, 4.22%) diagnostic methods. On the contrary, the pooled prevalence of MetS was lower among males than among females in the de Ferrranti (16.49%, 16.76%) and WHO (2.66%, 3.03%) diagnostic criteria ( Figure 3 and Table 5).

Subgroup Analysis of the Pooled Prevalence of MetS in the
General Population. The subgroup analyses were performed for the two diagnostic methods (IDF and ATP III) using continents where the original studies were performed. In the IDF diagnostic method, the pooled prevalence of MetS was estimated in three continents (North America, Asia, and Europe). Accordingly, the highest prevalence was recorded  (Figure 4).
The heterogeneity among the included studies remained significant after subgroup analysis. Hence, the possible sources of heterogeneity were further explored for the two diagnostic methods (IDF and ATP III). Thus, the funnel plots

13
BioMed Research International for both diagnostic criteria were presented ( Figure 5). The asymmetry of the plots was objectively verified by Egger's regression test, and there was publication bias among the articles included in computing the pooled prevalence of MetS in the IDF (p ≤ 0:001) and ATP III (p ≤ 0:001) diagnostic methods. Moreover, sensitivity analyses were performed for both diagnostic methods. This was done to evaluate if the pooled estimates were altered by the exclusion of any single study. However, none of the studies had significant effects in the pooled estimates ( Figure 6).
Finally, the Duval and Tweedie trim and fill analysis, a nonparametric method of accounting for publication bias in meta-analysis, was employed to estimate the pooled prevalence. This was done to estimate the number and outcomes    of missing studies, and adjust the meta-analysis to incorporate the theoretical missing studies. Nevertheless, the pooled prevalence of MetS remained the same (3.70%) using the IDF criteria. However, the pooled prevalence of MetS among the general population of children and adolescents was reduced to 5.40% (95% CI: 4.47, 6.32) in the ATP III diagnostic criteria.
Eventually, the trend of MetS in the general population of children and adolescents in HICs was plotted in a scatter plot based on the prevalence of cases with publication year (2003 to 2020). The trend line implied that there is an increasing trend of cases in three diagnostic methods (IDF, ATP III, and de Ferranti) (Figure 7).

Discussion
This is a comprehensive systematic review and meta-analysis, determining the prevalence of metabolic syndrome among children and adolescents in high-income countries. The pooled prevalence of MetS was computed using six diagnostic methods: IDF, ATP III, de Ferranti et al., WHO, Weiss et al., and Cruz and Goran. In the current meta-analysis, 77 studies with a total of 125,445 study participants were included. Of the total studies, 49 were conducted among the general population of study subjects, and 28 were conducted among overweight and obese population.
This study revealed that the prevalence of MetS among overweight and obese study participants is considerably higher than its prevalence in the general population. The pooled prevalence of MetS in the overweight and obese children and adolescents is as follows: IDF = 25:25%; ATP III = 24:47%; de Ferranti et al: = 39:41%; WHO = 29:52%; and Weiss et al: = 33:36%. Likewise, the pooled prevalence in the general population was 3.70%, 6.08%, 14.78%, 3.90%, 4.66% with the IDF, ATP III, de Ferranti, WHO, and Cruz and Goran diagnostic criteria, respectively. The prevalence in the general population is comparable with findings of a systematic review from Iran, where the prevalence of MetS was 0-8%, 3-16%, and 0-22% in the IDF, ATP III, and de Ferranti criteria, respectively [137]. But, Iranian findings are remarkably lower than the current pooled prevalence of MetS among the overweight and obese population. A possible reason for this disparity may be explained by the fact that overweight and obese children are at greater risk of developing metabolic syndrome as compared to children with normal weight [19]. Furthermore, the higher prevalence of obesity  criteria (0.3-9.5%). But, the prevalence was relatively higher in the de Ferranti et al. criteria (4-26.4%) [140]. Thus, the current findings are in line with the findings in this study. However, the meta-analyses results of the current study are higher than most of the previous findings, which depict that MetS is having an upsurge primarily in the developed world, and it is supported by the findings of the previous reviews [9,141,142]. In general, the pooled prevalence of MetS among the obese population is higher in HICs as compared to the low-and middle-income countries, but comparable with the general population [143]. In this study, the pooled prevalence of the components of MetS was also computed using different diagnostic criteria. Abdominal obesity was the most prevalent component of MetS in the overweight and obese population ranging from 65.62% in the IDF criteria to 79.81% in the ATP III criteria. On the other hand, a high level of FG level was the most infrequent component of MetS in the overweight and obese population. The pooled prevalence ranges from 1.61% (de Ferranti et al.) to 15.53% (Weiss et al.). Similarly, the frequent and infrequent components of MetS were computed in the general population. Thus, the most prevalent components include elevated BP (27.50%), low HDL-C (23.41%), high TG level (19.05%), and abdominal obesity (14.42) with the Cruz and Goran, IDF, ATP III, and WHO diagnostic methods, respectively. However, the high FG level is the least frequent component in the ATP III (7.16%) and WHO (1.63%) criteria. Likewise, abdominal obesity and high TG level were the least prevalent components in the Cruz and Goran (10.06%) and IDF (7.10%) criteria. In general, the prevalence of MetS amongst the general population is similar between high-income and low-income countries, whereas the prevalence is not the same amongst obese children in HICs and low-income countries. The pooled prevalence of MetS in the overweight and obese population was considerably higher among children in HICs. The possible elucidation could be due to a multitude of factors like consumption of unhealthy diets such as diets low in fruit, vegetables, and grains [144,145]. Moreover, sedentary behavior and lack of physical exercise may also contribute to the rise of MetS in these countries [146].
In most of the diagnostic methods, the prevalence of MetS in males is relatively higher than that in females. The pooled prevalence of MetS in the overweight and obese males is 26.62% (IDF) and 24.75% (ATP III). Likewise, it is 20.18% (IDF) and 24.97% (ATP III) among females. The pooled prevalence of MetS in the general population was computed in both genders using five diagnostic methods. Thus, the pooled prevalence of MetS among males was higher than that among females in the IDF (3.80%, 2.37%), ATP III (6.61%, 4.65%), and Cruz and Goran (5.53%, 4.22%) diagnostic criteria. In contrast, the pooled prevalence of MetS among males was lower compared to that of females in the de Ferrranti et al. (16.49%, 16.76%) and WHO (2.66%, 3.03%) diagnostic criteria. In general, males are more highly at risk to have MetS than females both in the original studies and pooled estimates of most diagnostic methods. The current findings are in line with the findings of a meta-analysis in China which showed that males are more highly liable to have MetS than females [139]. The possible justification for gender disparities may be associated with a higher prevalence of obesity in males than females. A higher prevalence of obesity among male children and adolescents may be related to excessive energy intake due to self-and family-imposed perception of being underweight and underestimation of their weight. On the other hand, females control their weight through diet and physical activity due to a self-perception of being overweight [147].
Moreover, the pooled prevalence of MetS in HICs was computed in three continents (Asia, North America, and Europe). Thus, 4.72%, 3.95%, and 2.54% of the study subjects in Asia, North America, and Europe, respectively, are found to have MetS in the IDF criteria. Similarly, the pooled prevalence of MetS in the ATP III criteria is 6.79% (North America), 6.32% (Asia), and 3.84% (Europe). These findings pinpointed that MetS is considerably higher in HICs. This could be associated with a high burden of childhood obesity and consumption of unhealthy diets in these countries [148,149]. Childhood obesity is not only associated with childhood MetS, but with MetS in adults [150].
Eventually, the number of cases was plotted against the publication year (2003 to 2020), using five diagnostic methods. The trend line revealed that the prevalence of MetS has increased from 2003 to 2020 in all diagnostic criteria. This implies that the prevalence of MetS is increasing in a sustainable manner in the developed world.
The findings of this study may be used by program planners and policy makers to design preventive and treatment strategies against morbidities and mortalities related to MetS. These findings will also help researchers who intend to conduct original researches on multiple factors contributing to a higher burden of MetS in those high-income countries. Nonetheless, there is no specific diagnostic method for MetS, The other limitation of this study was the exclusion of the following: studies written in non-English language, studies with no full texts, and studies conducted in different study designs and with a different study population. This could cause either under-or overestimation of the pooled prevalence of MetS.

Conclusion
In conclusion, the current study revealed that the prevalence of MetS among children and adolescents is high in highincome countries with higher proportions among the overweight and obese population. The prevalence is considerably higher in overweight and obese children of Asian countries. Similarly, MetS in the general population of children and adolescents is high in North America. Male children and adolescents are also at greater risk of MetS than females. Metabolic syndrome was diagnosed in underweight, normal weight, overweight, and obese children and adolescents. This implies that MetS is a nonselective problem of children and adolescents in high-income countries. Community-based social and behavioral change communications need to be designed to promote healthy eating behaviors and physical activities. Prospective cohort studies could also help to explore all possible risk factors of MetS and to design specific interventions accordingly. Abbreviations

Data Availability
The data that support the review findings of this study are included in the manuscript and supporting files.

Conflicts of Interest
There are no competing interests.

Authors' Contributions
ZWB and AA were responsible for analysis, visualization, and writing of the manuscript; ZWB, ZT, AA, and TW made substantial contributions to data acquisition; ZWB, AA, and EGA participated in the data interpretation and made substantial revisions in the first draft; ZWB and TW contributed to the reception and the design of the work. All authors read and approved the final manuscript.