Associations of Fruit and Vegetable Intake with Metabolic Health Status in Overweight and Obese Youth

Introduction: The objective of this study was to assess the relationship between fruit and vegetable intake with metabolic health status among overweight and obese Iranian youth. Methods: In total, 203 obese and overweight youth including 101 boys and 102 girls, 12–18 years, were randomly recruited through a multistage cluster sampling approach in this study. A validated Persian format of food frequency questionnaire was used to estimate participants’ usual food intake. Adolescents were categorized in accordance with the International Diabetes Federation (IDF) criteria as well as IDF plus insulin resistance (HOMA-IR) definition into metabolically healthy obese and metabolically unhealthy obese (MUO). Results: Based on IDF criteria, youth in the third tertile of vegetable and the combination of fruit and vegetable intake had 66% (odds ratio = 0.34, 95% CI: 0.12–0.95) and 73% (odds ratio = 0.27, 95% CI: 0.09–0.81) lower odds of MUO compared to those in the first tertile, in the fully-adjusted model. According to IDF plus HOMA-IR definition, inverse associations were also found between vegetable (odds ratio = 0.36, 95% CI: 0.12–1.02) and the combination of fruit and vegetable (odds ratio = 0.25, 95% CI: 0.08–0.79) intake with MUO phenotype. Stratified analysis indicated inverse associations between intake of vegetables and the combination of fruits and vegetables with MUO phenotype (based on IDF definition) in overweight, but not obese subjects. Conclusion: The current study demonstrated that greater consumption of vegetables as well as the combination of fruits and vegetables was related to a lower risk of MUO phenotype, particularly in overweight youth.

nation of fruit and vegetable intake had 66% (odds ratio = 0.34, 95% CI: 0.12-0.95) and 73% (odds ratio = 0.27, 95% CI: 0.09-0.81)lower odds of MUO compared to those in the first tertile, in the fully-adjusted model.According to IDF plus HOMA-IR definition, inverse associations were also found between vegetable (odds ratio = 0.36, 95% CI: 0.12-1.02)and the combination of fruit and vegetable (odds ratio = 0.25, 95% CI: 0.08-0.79)intake with MUO phenotype.Stratified analysis indicated inverse associations between intake of vegetables and the combination of fruits and vegetables with MUO phenotype (based on IDF definition) in overweight, but not obese subjects.Conclusion: The current study demonstrated that greater consumption of vegetables as well as the combination of fruits and vegetables was related to a lower risk of MUO phenotype, particularly in overweight youth.

Introduction
Overweight and obesity in children and adolescents have become a major global public health crisis in the recent century [1].As indicated in the World Health Organization report in 2016, more than 340 million pediatrics aged 5-19 years were obese or overweight [2].Pediatric obesity is associated with the occurrence of chronic diseases such as cardiovascular disorders, hypertension, type 2 diabetes mellitus, and nonalcoholic fatty liver [3].However, a particular obesity subgroup with a favorable metabolic profile that is called "metabolically healthy overweight or obese" (MHO) has been identified with a lower risk of developing obesity-related chronic diseases compared to "metabolically unhealthy overweight or obese" (MUO) individuals [4].Depending on the criteria used to define MHO, different prevalence rates of this phenotype have been reported in previous studies [5].Given the transition of metabolic health status from adolescence to adulthood, it is critical to identify determinants of MHO and its maintenance during the life span [6].In addition, it has been suggested that different phenotypes of obesity should be taken into account while applying dietary and lifestyle interventions for obesity treatment [7].
It has been suggested that modifiable risk factors such as dietary intake and dietary behaviors are associated with metabolic health status.There were some reports on the effects of dietary behaviors such as cognitive dietary restraint and regular meal intake on metabolic health status in pediatrics [8,9].Additionally, the results of previous studies indicated that adherence to a healthy diet was associated with MHO phenotype.For example, a cross-sectional survey involving 137 overweight or obese youth showed that individuals with the MHO phenotype had a higher adherence to the Mediterranean diet than adolescents with MUO [10].In addition, a higher healthy eating index-2005 score was observed in MHO adolescents aged 12-18 years compared to their MUO peers [11].However, a positive significant association has been found between unhealthy food consumption such as salty snacks, and total fat intake with MUO [12,13].
The linkage between metabolic health status and dietary intake of fruits and vegetables among children and adolescents has been examined in a limited number of prior investigations.These studies have obtained inconsistent results [13,14].Some studies that have examined the relationship between fruit and vegetable consumption and metabolic health status in children and adolescents with obesity and overweight have found inverse associations, but some others did not [15][16][17].Such that, a systematic review in this field indicated that the link between fruit and vegetable consumption and cardiometabolic health was conflicting, probably due to differences in the applied definitions for cardiovascular risk [18].Therefore, the present study was performed to evaluate the association between dietary intake of fruits and vegetables with metabolic health status among Iranian overweight and obese youth.

Participants
A cross-sectional study was performed among a representative sample of Iranian youth aged 12-18 years.Participants were randomly recruited through a stratified, multistage cluster sampling approach from students in grades 7-12.In the first stage, a total of 16 public high schools were randomly selected from six various educational districts of Isfahan city.In the second stage, 2-8 classes were randomly selected from each grade, based on the school size.Students from selected classes were chosen as participants based on participation criteria.A minimum sample size of 188 was calculated by considering a 5% significance level, 95% confidence interval, 80% power, and 7% precision (d) and the reported prevalence of MUO among obese and overweight Iranian adolescents in previous studies (60%) [13,19].All adolescents from the selected schools were screened by age-and sex-specific percentile curves of body mass index (BMI) [20] and those with a BMI greater than the 85th percentile (as overweight or obese ones) were invited to the study.Individuals with endocrine and genetic disorders such as hypothyroidism, Cushing's syndrome, and type 1 diabetes mellitus were not included.Additionally, individuals who were on a weight-loss diet or taking vitamin and mineral supplements or medications potentially affecting body weight, blood pressure, lipid profile, or blood glucose were not included.The Ethics Committee of the Isfahan University of Medical Sciences approved the study protocol and written informed consent was obtained from all participants and their parents.

Assessment of Dietary Intakes
Usual dietary intakes over the previous year were evaluated using a validated 147-item food frequency questionnaire (FFQ).A previous investigation on Iranian children and adolescents has documented that this questionnaire has a desirable validity and reliability for the assessment of food intake (with a Cronbach's alpha coefficient of 0.96 and an intra-class correlation coefficient of 0.92) [21].This questionnaire contained the most commonly consumed fruits (n = 36 items) and vegetables (n = 23 items) in a raw, dried, juice, or cooked format.The questionnaire was filled by two trained dietitians through interviews with participants.The dietary intake of each food item was documented in standard portion sizes based on frequency consumption on a daily, weekly, or monthly basis.Then, portion sizes of consumed foods were converted to grams per day using household measures.Finally, total energy and nutrient intakes were examined by the Nutritionist IV software.Previous studies demonstrated favorable validity and reliability for the FFQ for dietary intake evaluation of Iranian youth [22,23].

Assessment of Anthropometric Indices and Cardiometabolic Risk Factors
Weight and height were measured using a Seca digital scale and a stadiometer while participants were minimally clothed and wearing no shoes.BMI was calculated as weight (kg) divided by squared height (m 2 ).Adolescents were classified into obese (BMI >95th percentile) or overweight (85th < BMI <95th percentile) categories based on the World Health Organization's (WHO) sex-and age-specific growth curves [20].Waist circumference (WC) was measured by an unstretchable flexible anthropometric tape at the midpoint between the iliac crest and the lower margin of the last palpable rib and recorded to the nearest 0.1 cm.Systolic blood pressure and diastolic blood pressure were measured twice after a rest period of 15 min using a standard mercury sphygmomanometer with an appropriate cuff size placed on the right arm.The mean values of the two measurements were considered for the analyses.
Blood samples were collected after 12 h of fasting to evaluate biochemical variables.An enzymatic colorimetric method (glucose oxidase) was used to determine fasting blood glucose (FBG) concentration (Pars Azmoon Company, Tehran, Iran).Additionally, triglyceride (TG) and high-density lipoprotein cholesterol (HDL-c) were measured by standard methods.Commercial kits were used to measure serum levels of insulin (Diagnostic Biochem Canada Inc., Canada).The following formula was used to calculate homeostasis model assessment insulin resistance (HOMA-IR): HOMA-IR = [(fasting insulin (mU/L) × FBG (mg/dL)]/405.

Assessment of Metabolic Health Status
Metabolic health status was defined in accordance with the International Diabetes Federation (IDF) criteria in which individuals with two or more of the following risk factors were known as MUO: decreased HDL-c (defined as HDL-c <40 mg/dL for the age of <16 years, and <50 mg/dL in girls/< 40 mg/dL in boys for the age of ≥16 years), as well as raised TG (defined as TG ≥150 mg/dL), FBG (defined as FBG ≥100 mg/dL), and BP (defined as BP ≥130/ 85 mm Hg) [24].According to the IDF plus insulin resistance (IR) definition, individuals with HOMA-IR >3.16 and two or more of the four metabolic risk factors described above were known as MUO; while subjects with HOMA-IR <3.16, disregarding the number of their cardiometabolic risk factors, were known as MHO [25].

Assessment of Other Variables
Basic data such as sex, age, medical history of diseases, and use of dietary supplements and medications were collected by a standard checklist.The socioeconomic status of the study participants was evaluated by a validated questionnaire based on variables such as parents' education level, parents' occupation, number of family members, having a personal room, number of the family car(s), computer(s), and trip(s) per year [26].
The usual physical activity levels of participants were evaluated by a 9-item physical activity questionnaire for adolescents (PAQ-A), previously validated among Iranian adolescents [27].The physical activity level of participants was classified as sedentary or not having an orderly week activity (PAQ-A score less than 2), low active (PAQ-A score equal to or greater than 2 and less than 3), moderately active (equal to or greater than 3 and less than 4), and highly active (equal to or greater than 4).

Statistical Analysis
The Kolmogorov-Smirnov test was applied to examine the normality of quantitative demographic variables.Continuous and categorical variables were respectively presented as mean ± SD/SE and frequency (percentage).To obtain energy-independent exposures of interest, energy-adjusted fruit, vegetable, and the combination of fruit and vegetable intakes were calculated by the residual method [28].Individuals were then classified based on energy-adjusted tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes.The comparison of categorical and continuous variables across tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes were respectively performed by χ 2 and one-way ANOVA tests.Dietary intakes of participants (adjusted for age, sex, and total energy intake) among tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes were compared by analysis of covariance.Odds ratios (OR) and attributed 95% confidence intervals (95% CIs) for MUO across energy-adjusted tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes were calculated by binary logistic regression.Confounding variables in adjusted models were as follows: model 1: adjusted for age, sex, and total energy intake; model 2: additionally adjusted for physical activity and socioeconomic status; model 3: additionally adjusted for dietary intake of dairy, grains, legumes, fat, red and processed meat; model 4: additionally adjusted for BMI.In all models, tertile 1 of fruit, vegetable, or the combination of fruit and vegetable intakes was considered as the reference category.Trends were specified by considering tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes as ordinal variables in logistic regression models.Stratified analysis was performed by BMI categories.All analyses were performed using SPSS software version 20 and statistical significance was accepted at p < 0.05.

Results
The current study was performed among 203 overweight/ obese adolescents (mean age of 13.98 ± 1.61 years and mean BMI of 28.54 ± 3.91 kg/m 2 ) comprising 101 boys and 102 girls.Among these participants, 104 were overweight and 99 others were obese.Based on IDF and IDF plus HOMA-IR definitions, 38.9% (37 boys, and 42 girls) and 33.0%(35 boys, and 32 girls) of adolescents have been, respectively, known as MUO.
Participants' basic characteristics across tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes are presented in Table 1.A significant difference was observed regarding physical activity levels across tertiles of fruit (p = 0.01), vegetable (p < 0.001), and the combination of fruit and vegetable (p < 0.001) intakes.The mean systolic blood pressure (p = 0.04), FBG (p = 0.01), serum insulin (p = 0.01), TG (p = 0.01), and HOMA-IR (p = 0.01) of adolescents in the highest tertile of fruit intake were significantly lower than those in the lowest tertile.The mean FBG (p < 0.001), serum insulin (p = 0.01), TG (p = 0.01), and HOMA-IR (p = 0.01) of individuals in the top tertile of Fruits and Vegetables Intake and Metabolic Health Status   vegetable intake were significantly lower than individuals in the bottom tertile.The mean FBG (p < 0.001), serum insulin (p = 0.01), TG (p < 0.001), and HOMA-IR (p < 0.001) of participants in the third tertile of the combination of fruit and vegetable intakes were significantly lower than individuals in the first tertile.However, the mean HDL-c of individuals in the highest tertile of vegetable (p = 0.01) and the combination of fruit and vegetable (p = 0.01) intakes was significantly higher than those in the reference tertile.No significant difference was observed between adolescents across tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes regarding other basic variables (p > 0.05).
Daily dietary intakes of the studied adolescents across tertiles of fruit, vegetable, and the combination of fruit and vegetable are presented in Table 2. Dietary intake of fiber, total fruit, citrus fruits, non-citrus fruits, total vegetable, green leafy vegetables, dairies, nuts, legumes, and whole grains was significantly higher in individuals in the third tertile of fruit intake than individuals in the first tertile (p < 0.05).However, a lower consumption of refined grains was observed in participants in the highest tertile of fruit intake in comparison to participants in the lowest tertile (p < 0.05).There was no significant difference between individuals across tertiles of fruit intake regarding dietary intake of energy, protein, total fat, starchy vegetables, red and processed meat, poultry, and fish (p > 0.05).An increasing trend was observed in the dietary intake of protein, dietary fiber, total fruit, citrus fruits, non-citrus fruits, total vegetable, green leafy vegetables, dairies, nuts, legumes, whole grains, fish, and poultry across tertiles of vegetable intake (p < 0.05).Though, dietary intakes of carbohydrates and refined grains was significantly lower among participants in the top tertile of vegetable intake than those in the reference tertile (p < 0.05).Dietary intake of energy, total fat, starchy vegetables, red and processed meat was not significantly different across vegetable intake tertiles (p > 0.05).Compared to youth in the first tertile, individuals in the third tertile of the combination of fruit and vegetable intakes showed a higher dietary intake of protein, dietary fiber, total fruit, citrus fruits, non-citrus fruits, total vegetable, green leafy vegetables, dairies, nuts, legumes, whole grains, and fish (p < 0.05).However, adolescents in the highest tertile of the combination of fruit and vegetable had a significantly lower intake of refined grains (p < 0.05), whereas no significant difference was found between participants among tertiles of the combination of fruit and vegetable intake regarding dietary intake of energy, carbohydrate, fat, starchy vegetables, and poultry (p > 0.05).
The frequency of adolescents with MUO phenotype based on IDF criteria across energy-adjusted tertiles of fruit, vegetable, and the combination of fruit and vegetable intake is presented in Figure 1a-c.The frequency of individuals in tertiles 1-3 of fruit intake was 55.2%, 30.9%, and 30.9%, respectively (p = 0.01).In addition, 64.2%, 32.4%, and 20.6% of individuals were classified as MUO in tertiles 1, 2, and 3 of vegetable intake (p < 0.001).A statistically significant difference was also observed in the frequency of adolescents in tertiles 1 (61.2%), 2 (36.8%), and 3 (19.1%) of the combination of fruit and vegetable intake (p < 0.001).Figure 1d-f presents the frequency of MUO adolescents based on IDF plus HOMA-IR criteria across energy-adjusted tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes.A significant difference was found in the frequency of individuals in tertiles 1-3 of fruit (p = 0.01), vegetable (p < 0.001), and the combination of fruit and vegetable intake (p < 0.001).
Crude and multivariate-adjusted ORs and 95% CIs for MUO across tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes are presented in Table 3. Regarding the IDF definition of metabolic health, participants in the highest tertile of fruit intake had 64% lower risk of MUO compared with participants in the first tertile in the crude model (OR = 0.36, 95% CI: 0.18-0.73).The association was no longer significant following adjustment for all covariates (OR = 0.76, 95% CI: 0.27-2.14).Adolescents in the top tertile of vegetable intake showed an 84% lower likelihood of having MUO phenotype compared to adolescents in the bottom category in the crude model (OR = 0.16, 95% CI: 0.07-0.31).This association remained statistically significant after adjustment for covariates; such that, individuals in the third tertile of vegetable intake were 66% less likely to be MUO compared to individuals in the first tertile (OR = 0.34, 95% CI: 0.12-0.95).In addition, a significant inverse relation was observed between the combination of fruit and vegetable intake with MUO risk in crude (OR = 0.15, 95% CI: 0.07-0.53)and final adjusted (OR = 0.27, 95% CI: 0.09-0.81)models.As shown in online supplementary Table 1 (for all online suppl.material, see https://doi.org/10.1159/000533343), all R 2 values of multiple logistic regression were higher than 0.05 in both crude and adjusted models.Adjustments for confounding variables led to an increase in the R 2 levels to 0.4 and higher values, which indicates a relatively strong relationship between the independent and dependent variables.
Regarding the IDF plus HOMA-IR definition, significant inverse associations were found between fruit (OR = 0.29, 95% CI: 0.09-0.93),vegetable (OR = 0.03, 95% CI: 0.01-0.28),as well as the combination of fruit and vegetable (OR = 0.12, 95% CI: 0.03-0.50)intakes with MUO risk among overweight subjects in the crude model.Nevertheless, these associations became insignificant after considering covariates.In obese adolescents, a significant inverse association was found between the combination of fruit and vegetable intake with MUO phenotype in the crude model (OR = 0.29, 95% CI: 0.10-0.87).But the association was no longer significant after adjustment for confounding variables (OR = 0.25,   Obtained from ANCOVA test for adjustment of energy intake. 95% CI: 0.05-1.20).No statistically significant association was detected between the higher intake of fruits or vegetables with MUO phenotype in obese subjects in the crude or adjusted models.

Discussion
The present research revealed that 38.9% and 33.0% of Iranian overweight/obese adolescents had MUO phenotype based on the IDF and IDF plus HOMA-IR definitions, respectively.A higher intake of vegetables as well as the combination of fruits and vegetables was associated with lower odds of MUO in the studied population, even after considering the effects of confounding variables.Stratified analysis showed that the relationship was significant among overweight subjects, but not in obese ones, mostly based on IDF criteria.To our knowledge, our study is among the first that investigated the association between fruit and vegetable intake with MUO among an adolescent population.
Given the failure of many weight loss strategies in pediatrics, maintaining a metabolically healthy status among overweight and obese adolescents can be considered one of the beneficial approaches to decrease the risk of overweight/obesity-related complications and financial burden imposed on societies and healthcare systems.The present study indicated that intake of more than 700 g (or approximately 8.5 servings) of fruits and vegetables per day could decrease the risk of MUO.However, many adolescents do not meet daily dietary recommendation of 400 g (or 5 servings) per day for fruit and vegetable intake [29,30].This issue shows the importance of developing nutritional approaches for overweight and obese adolescents and their parents to improve the intake of fruits and vegetables in pediatrics.
A small number of previous investigations have examined the relationship between fruit and vegetable consumption and metabolic health status in the pediatric population.A cross-sectional study on 1,700 school-aged Korean children and adolescents showed a significant higher intake of the combination of fruits and vegetables in MHO children than in MUO children [14].Nevertheless, another study on 408 obese Iranian children and adolescents indicated no significant difference in the consumption of fruits and vegetables between individuals Fruits and Vegetables Intake and Metabolic Health Status with the MHO phenotype and those with the MUO phenotype [13].A prospective study on 8-10-year-old Canadian children with the MHO phenotype showed that increasing daily consumption of fruits and vegetables by one serving could be associated with a 39% (OR 0.61; 95% CI: 0.40-0.94)reduced risk of converting from the MHO to MUO phenotype, after 2 years of follow-up [16].Furthermore, a significant positive association has been found between fruit intake and the MHO phenotype in Canadian children aged 2-17 years (OR = 1.12; 95% CI: 1.01-1.24)[17].A cross-sectional investigation on 1,047 Saudi obese adolescents also found an inverse association between the frequency of vegetable intake per day and the MHO phenotype among boys, but not in girls.According to the results of the study, a 77% increase in MHO phenotype was observed among boys with a vegetable intake frequency of 2 or more (OR = 1.77; 95% CI: 1.07-2.91)[15].These controversies in findings of previous investigations could originate in differences in the studied populations or research design.Some of these studies were not controlled for some potential confounding variables which might affect their findings.
In addition, various types of fruits and vegetables consumed in different regions could lead to different results among investigations.It should be noted that wide 95% CI intervals in the present study can be related to the small sample size.Therefore, further large-scale studies, especially with prospective designs, are needed to determine the relationship between consumption of various groups of fruits and vegetables and metabolic health status.
In the present study, statistically significant associations between consumption of vegetables and the combination of fruits and vegetables with metabolic health were observed; however, fruit intake was not related to metabolic health status.Such a nonsignificant association might be explained by the narrow range of fruit intake among tertiles (T3 vs. T1: >382.76 vs. <260.43g/d), which made it difficult to find a relation with the outcome of interest.We also find a significant relationship in overweight adolescents, but not in obese ones.The physiological responses of overweight subjects to environmental factors (including dietary intakes) that could affect metabolic health status might be different from those of obese individuals.Further studies are required to verify this hypothesis.
The inverse correlation between fruit and vegetable intake and metabolic health status could be explained by several plausible mechanisms.Oxidative stress, which is mainly caused by excessive production of free radicals, and inflammation have been known as underlying causes of  Additionally adjusted for dairy, grains, legumes, fat, red and processed meat.
Fruits and Vegetables Intake and Metabolic Health Status metabolically unhealthy status [31][32][33].Fruits and vegetables contain natural antioxidant compounds like fibers, vitamin C, vitamin A, carotenoids, and phytochemicals which have protective roles against both oxidative stress and inflammation [34].Fiber content of fruits and vegetables probably improves metabolic status through its favorable effect on the gut microbiome [35], as a prior research has indicated that there are significant differences between MUO and MHO individuals regarding their gut microbiota composition and diversity [36].In addition, the altered composition of gut microbiota was associated with various cardiometabolic risk factors such as IR and dyslipidemia [37].
The current study has some limitations that should be stated.The cross-sectional design of the study precludes establishing a causal relationship between fruit and vegetable consumption and metabolic health status in adolescents.Furthermore, although many confounding variables have been considered in our analyses, the effect of some covariates like sleep quality, puberty stage, and genetic predisposition to obesity as important determinants of metabolic health status has been neglected.Moreover, due to the rather small sample size of the survey, it was not possible to examine the relationship between different types of fruits and vegetables with metabolic health status.Finally, although a validated interviewer-administered FFQ was applied for dietary assessment, self-reported dietary intakes were subjected to recall or interviewer bias.Nevertheless, the present study was one of the limited studies on the relationship between fruit and vegetable consumption and metabolic health status in adolescence, a critical period of life in determining adulthood metabolic health status and related complications.
In summary, the current study demonstrated that higher intake of vegetables and the combination of fruits and vegetables were linked with lower risk of MUO in an Iranian adolescent population, mainly in overweight adolescents.However, more large-scale prospective population-based studies are required to approve these findings.

Statement of Ethics
The study procedure was performed according to declaration of Helsinki and STROBE checklist.Written informed consent form was obtained from all participants and their parents.The study protocol was approved by the Local Ethics Committee of Isfahan University of Medical Sciences (approval number: 299142).

Tertiles
General characteristics and cardiometabolic factors of study participants across energy-adjusted tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes 1

Tertiles
Dietary intakes (energy, macronutrients, and food groups) of study participants across energy-adjusted tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes 1

Fig. 1 .
Fig. 1.Prevalence of MUO in energy-adjusted tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes based on IDF (a, b, and c) as well as IDF plus HOMA-IR definitions (d, e, and f).

Table 2 .
Values are mean ± SE.Energy intake was adjusted for age and sex; all other values were adjusted for age, sex and energy intake. 2

Table 3 .
Multivariate adjusted ORs and 95% CIs for MUO across energy-adjusted tertiles of fruit, vegetable, and the combination of fruit and vegetable intakes 1 5Model 4: Additionally adjusted for BMI.

Table 4 .
3ultivariate adjusted ORs) and 95% CIs for MUO across energy-adjusted tertiles of fruit, vegetable, and the combination of fruit and vegetable intake, All values are ORs and 95% CIs. 2 Model 1: Adjusted for age, sex, and energy intake.3Model2: Additionally adjusted for physical activity and SES (parental education, parental job, number of family members, having car in the family, having computer/laptop, having personal room and having trips).