The relationship between body mass index (BMI) and quality of life in Iranian primary school students in Tehran, Iran

Minimally Invasive Surgery Research Center, Iran University of Medical Sciences, Tehran, Iran Zabol Medical Science University, Zabol, Iran Epidemiology and Biostatistics Department, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran 4 Pediatric growth and development research center ,institute of endocrinology and metabolism, Iran University of Medical Science, Tehran, Iran


Introduction
Nowadays, one of the most important world health concerns is the increased rate of obesity and overweight through childhood [1,2] Based on CDC 2000 BMI standards; In 2012 among ages 2-19, 31.8% of them were overweight and 16.9% were obese. [3] Based on WHO-HBSC examine, overweight prevalence is more than 10% among school-developed children in most countries, with an extent of 7.6% (Latvia) to 28.8% (USA) [4]. Obesity and overweight of children are serious concerns in Iran and prevalence of obesity in children are at high incidence in this country. [5,6] Childhood overweight and obesity are related to advancement of several long-term diseases such as CAD (coronary artery disease), DM (diabetes mellitus), metabolic syndrome, stroke, dyslipidemia, hypertension and asthma [7][8][9][10]. Likewise, childhood Social functioning subscale refers to peer relationship, problems with companion and capacity to make companions.
School functioning subscale refers to cognitive functioning such as trouble in focusing and association with teachers. [23] A five-point scale (from 0 = never a problem to 4 = almost always a problem) was used for both children self-report and parent proxy-report, items were scored reversely (0 = 100, 1 = 75, 2 = 50, 3 = 25 and 4 = 0); so, a higher total score shows a better HRQOL. [24]

Data collection
First, the study protocol was approved by the Medical Research Ethical Committee of Obesity Research Center of Minimally Invasive Surgery, Iran University of Medical Science.
Then, the approvals were obtained from the Ministry of Education in Tehran and selected schools. Oral assents and signed consents were obtained from both children and their parents, and questionnaires were distributed to them. All questionnaires were completed anonymously.
Totally, 920 children and their parents were participated in this study. 53 students were excluded from study because of chronic diseases (2 for anemia,1 for migraine, 1 for CKD,1 for kidney stone,5 for heart disease, 2 for G6PD, 2 for recurrent UTI, 1 for nocturia, 2 for chronic sinusitis, 2 for GER, 3 for diabetes, 2 for hypothyroidism, 2 for epilepsy, 1 for thalassemia, 11 for asthma and 15 for allergies), 13 students were excluded because of incomplete information. Therefore, a total of 829 children and their parents participated in this study. After children's selfreport version of HRQOL questionnaires were distributed to the children, they were made a request to bring back finished parentproxy sheet of HRQOL questionnaire on the following day. A Rule was set up to clarify the method for filling in parent-proxy version of HRQOL questionaires.

Data analysis
We used SPSS (Statistical Package for Social Sciences, V 11.5) program for Statistical analysis. Mean and standard deviation were calculated for quantitative variables and frequency (percent) describes qualitative variables. Pearson correlation analysis was used to examine the correlations between BMI or age and scores of PedsQL. Independent-samples t-test was used to compared total and subscale scores of PedsQL between two sexes. One-way ANOVA test was used to determine differences between groups of body weight status. Fisher's Least Significant Difference (LSD) test as post hoc test was used to determine which four groups of body weight status were significantly difference in PedsQL scores. Less than 0.05 was accepted as indicating statistical significance.

Result
Totally 829 students were enrolled in this study, with mean age of 9.76 ± 1.48 (Rang: 8-12 years). Among included children, 463 (55.9%) were girls. The mean BMI of children were 18.60±4.02 (18.12 ± 3.72 and 19.20 ± 4.32 in girls and boys respectively; P < 0.001). The body weight status of children were classified to 61 (7.4%) as underweight, 476 (57.4%) as normal weight, 171 (20.6%), as overweight and 121 (14.6%) as obese. Table 1 shows demographic data of the children. Table 2 shows mean of children self-reported and parent subscales in obese children were at the lower level in comparison with their ordinary weight matches. Some studies have shown the strong relationship between obesity and HRQOL and some others not, there is not definitive and comprehensive study in this field in Iran. By the importance of the subject, the aim of this study was to examine the correlation between body mass index (BMI) and HRQOL in primary school students in Tehran in 2016.

Study design and sampling
This cross-sectional study was conducted among elementary school students with the age of aged 8 to12 year old in Tehran in 2016. The multi-stage sampling was used to select subjects. Firstly, Tehran was divided into five areas of the South, East, North, West and Central division. Then the Department of Education of each zone was visited and a list of non-profit and for-profit schools was observed, two elementary chosen selected using simple random sampling method from each zone, thus our study was conducted in a total of ten elementary schools.
The duration of primary schools in Iran is six years, we selected six classes from each school in order to participate in this study. Of course students with any chronic disease like Malignancies, Diabetes and kidney disease were excluded from study.

Measurements 2.2.1. Assessment of weight status
Students' height and weight were measured in collaboration with the school health educators and trained nurses. Students had minimum clothes and no shoes during scaling with the normal position of the shoulder. BMI was computed by the ratio of weight (Kg) to height squared (m 2 ).

Health-related quality of life (HRQOL)
For evaluation HRQOL of children, the Persian version Ped-sQL™ 4.0 questionnaire was used. The reliability and validity of the scale have been Repeatedly confirmed (α = 0.71-0.89). [22] The PedsQL ™ 4.0 is a self-administered questionnaire consists of Generic Core Scales that are comprised of parallel child selfreport and parent proxy-report formats and includes 23 items.
It envelops four subscales physical functioning, emotional functioning, social functioning, and school functioning. Physical subscale consists of five categories: limitations with effortful movements (like sports), (b) constraint of regular exercises, (c) physical symptoms like pain, (d) feelings about physical wellbeing and physical sensations, and (e) acute and chronic disorders like Asthma, epilepsy, diabetes, etc.
Emotional subscale consists of five categories: (a) negative feelings (feeling stressed, feeling sad and having sleep issues), (b) positive emotions (Feeling happy and cheerful), (c) confidence and body image, (d) cognitive functioning (Difficulty focusing, recollect those learned at school), and (e) general behavior.
proxy-reported total and subscale scores of PedsQL. There were no significant differences between mean of children self-reported and parent proxy-reported total and subscale scores (for all of them, P > 0.05). Based on both children self-reported and parent proxy-reported, social functioning was the highest subscale score (84.67 ± 15.07 and 85.13 ± 15.29, respectively) and emotional functioning was the lowest subscale score (77.79 ± 17.26 and 76.45 ± 17.68, respectively).
According to gender classification, based on children selfreported scores of PedsQL, school functioning was significantly higher in girls compared with boys (84.52 ± 14.49 vs. 80.78 ± 17.17, P = 0.001) while other subscales and total score were the same in girls and boys children (for all of them, P > 0.05) ( Table 3).
In overall, the correlation coefficient between BMI and total score of children self-reported (r = -0.076, P = 0.029), physical (r = -0.069, P = 0.047), and social functioning score (r = -0.087, P = 0.013) were significantly negative. Although the correlation between BMI and emotional (r = -0.029, P = 0.403) and school functioning score (r = -0.056, P = 0.109) are negative, they were not statistically significant (Table 4). Table 5 shows self-reported total and subscale scores of Ped-sQL according to BMI classification, normal weight children had highest mean total score of PedsQL and all subscale scores compared with underweight, over weight and obese groups; whereas the mean score for total, physical and social subscale scores were   significantly different by body weight status (P < 0.05, for all of them) (table 5). The result of post hoc test for total score of Ped-sQL showed that the difference between normal weight and obese were statistically significant (P = 0.004). In physical subscale score, obese (P = 0.004) and overweight children (P = 0.006) had lower mean scores compared with normal weight children; furthermore, social subscale score in obese children had significantly lower mean score than normal weighted group (P = 0.003).
In girls, self-reported total and physical scores of PedsQL were significantly different by body weight status (for both of them, P < 0.05). The result of post hoc test for total score in underweight (P = 0.046) and overweight (P = 0.017) girls was significantly lower than normal weighted, the same result was found in the physical functioning subscale, so underweight (P = 0.015) and overweight (P = 0.007) girls had significantly lower mean score compared to normal weighted.
In boys, the mean score for total and all subscale scores of PedsQL were significantly different by BMI category (P < 0.05, for all of them), except for school functioning subscale (P = 0.352) (table 5). Comparing obese boys with normal weight and overweight groups, there is significantly lower total score (P = 0.002 and P = 0.007, respectively), emotional (P = 0.003 and P = 0.026, respectively) and social score (P = 0.009 and P = 0.008, respectively); in physical functioning subscale score, obese boys had significantly lower mean score compared to normal weight (P = 0.003) and underweight (P = 0.022) ones. Based on parent proxy reports in overall, as body weight status, total score of PedsQL, and all subscale scores (P < 0.05, for all of them) except physical subscale score (P = 0.242) were significantly different ( Table 6). Result of post-hoc tests, obese children compared to normal weight and overweight groups had lower PedsQL in total score (P < 0.001 and P = 0.035, respectively), emotional score (P = 0. 01 and P = 0.028, respectively), and social subscale scores (P < 0.001 and P = 0.031, respectively).  Additionally, in school functioning subscale the difference between normal weight and obese children is statistically significant (P = 0.008).
In girls, according to parent proxy-reported total score and all subscale scores of PedsQL were not significantly different by BMI category (for all of them, P > 0.05). This finding shows that parents perceived PedsQL of girls were similar regardless of their BMI.
Based on parental perception of PedsQL in boys, obese children compared to normal weight and overweight children, had the lower scores in total score (P < 0.001 and P = 0.002, respectively) emotional (P = 0.002 and P = 0.001, respectively) and social (P < 0.001 and P = 0.004, respectively).

Discussion
In this study, we evaluate the relationship between HRQOL and BMI of primary school children. Our findings showed that obese children had lower HRQOL than other children. The mean total score HRQOL self-reported among children were 82.05 ± 12.04 and parent proxy-reported were 81.66 ± 12.81 in this study. Self-reported scores are similar to parent-proxy scores and are higher than scores that observed by Amiri et al. among school children of Tehran in 2012 (self-reported score: 83.99 ± 11.90, parent proxy-reported score: 76.75 ± 13.45). [25] There was a study by Hovsepian et al. in 2011 during a nationwide study in Iran (the Weight Disorders Survey of the CASPI-AN-IV study) investigated 23,043 students (6-18 years) and mean total PedsQL ™ score in children (age 6-12 years ) was 83,79. [26] In another study conducted in Qazvin (a city near to Tehran) in 2011, among 1107 students aged between 8and 18, children self-reported was 77.55 ± 13.02 and parent proxy report was 75.77 ± 18.42. [27] It appears that total mean PedsQL ™ score is comparative in different areas of Iran using the same score scale. The finding by Hovsepian et al. (2017) [26] In this study there was a harmony between children selfreport and parent proxy-report of HRQOL which is in agreement with the results of study by Amiri et al. (2012) high parental support in Iranian families can explain the harmony between children self-report and parent proxy-report of HRQOL. [25] this finding is not consistent with results of studies in most of the other translated versions. [28][29][30] Based on our results, emotional and social functioning subscales had the lowest and the highest HRQOL subscale scores respectively. In Previous studies, Iranian children had the lowest HRQOL score in the subscale of emotional functioning while the highest HRQOL subscale score of children was in social functioning [25]. A good capacity in making companions and peer relationships of Iranian children can explain these findings. On the other hand, emotional functioning was the poorest subscale in comparison with the three other subscales of HRQOL. This finding indicates that negative emotions and rest related issues are the most well-known issues among Iranian children and these health problems should be considered by health-care professionals. [31] In our study, age was not related to HRQOL in primary school children of Tehran. Although in previous studies (in children 8-18) HRQOL declined with age [32] this finding could be due to selecting participant from a narrow range of age in this study. (children [8][9][10][11][12] Past examines have exhibited sexual orientation contrasts in the pediatric HRQOL. [22,27] Our study showed that girls had significantly higher HRQOL in school functioning score than boys, by self-report. However, boys in emotional and social functioning had higher scores, but  [33] Amiri et al. [25] showed that girls have significantly higher scores in all subscales. Pakpour et al. in another study in 2011 demonstrated that boys have higher scores but differences weren't significant [27]. In previous studies, among adolescents, girls have lower scores in all subscales. The difference between adolescents and children might be due to the diverse impression of social and ecological weight and furthermore to pubescence changes in teenagers. [25] In general, it can be concluded that by increasing age (From childhood to adolescence) the quality of life (especially in social domain) in obese children due to social problems decreases. [26] BMI for age was inversely correlated to HRQOL total score in this study. Also HRQOL scores in normal weight children were significantly higher than obese children. Our results consisted of past investigations which demonstrate opposite impact of obesity on HRQOL. [34] The decline in HRQOL score was not significant for overweight children in comparison with normal weight children, but it was significant for those who were obese, that is in agreement with the results reported by Khodaverdi et al. in 2011 [20] poorer HRQOL scores in obese children in comparison with normal weights children can be a cause of negative effects of obesity outcomes. Also, higher BMI can predict higher levels of disordered eating attitudes and behaviors, which lead to lower levels of HRQOL. [35] Since the HRQOL is multidimensional, it is sensible that some domains might be more influenced by overweight or obesity. Various results have been obtained in various studies in this field. For example, Williams et al. demonstrated that aggregated score, physical health, and social functioning score diminished significantly as weight increased [36]. In another study overweight children had significantly lower scores on psychosocial, physical functioning and total scores in comparison with normal weight children. [37] In a study by Fiveash et al., obese children had Lower scores in social functioning compared to underweight, normal weight and overweight [38] another examination demonstrated that obese children were over five times more likely to report poor quality of life scores when in comparison with normal weight children. [17] In another study by Pinhas-Hamiel et al., demonstrated that obese children significantly poorer quality of life related to physical functioning and social domains compared to normal weight children. [18] In an investigation of clinical sample of obese children, they had significantly poorer physical health compared to control group. [39] In our study, by physical dimension, obese and overweight children had significantly lower scores than normal weight children. This finding maybe be related to restrictions associated with excess weight for obese and overweight children (like difficultly running and doing exercise), however, they were not so heavy as to seek medical treatment [20]. It could be in mind that in another study physical dimension were not different between weight groups. [38] BMI was conversely associated with physical functioning in this study. This supports the idea that reduced activity can make mismatch between production and consumption calories and will lead to weight gain. [17] In social dimension, obese children had significantly lower scores. This finding elucidates that Obesity is one of the most disparaging and least socially adequate conditions in childhood [17] Obese children are exposed to labeling of obese and ridicule and bullying from other students. [40] On the other hand, obese children had significantly lower scores in the social domain, but the difference between overweight and normal weight children isn't significant. So slight decrease in weight (i.e., shifting from obese to overweight) may improve social wellbeing. This finding is consistent with khodaveri et al. [20] In Hovsepian S et al. study, in physical and social domain, there was no significant difference between weight groups. [26] In emotional and school functioning, these relations weren't significant and emotional and school functioning dimension seem unaffected.
Although in Hovsepian S et al. study, emotional and school functioning domains had a negative association with obesity. [26] The lack of differences in emotional subscale scores across weight categories is consistent with khodaverdi et al. [20] this can revealed that obese children despite physical and social difficulties, keep up their emotional health and endeavors for understanding the nature and determinants of constancy to create preventive interventions to advance HRQOL of them will be useful. [41] It appears that different sample size of studies and also cultures of different societies may cause this variety of outcomes in several studies. In addition, in this sample of study, overweight children is higher than obese ones, and given that overweight children have less difficulty in quality of life (especially in the physical domain) compare to obese children, the effect of weight gain on quality of life may not well defined. [26] Using a valid instrument for evaluation HRQOL was the main quality of this study. On the other hand, this is a generic instrument. The generic instrument may be less sensitive than specific instruments for obesity to evaluate the relationship between body weight status and HRQOL of the children. However, PedsQL is very popular for assessing HRQOL in many countries. It examines the quality of life from the perspective of children and their parents, and allows us to compare them.
There are also some limitations in this study. We can't make a causal assessment about connections amongst obesity and HRQOL due to its cross-sectional nature.
In conclusion, our results demonstrated that obesity was associated with poor HRQOL in children of primary schools of Tehran. Future investigations should be concentrated on obesityrelated practices that may add to poor HRQOL among Iranian children, in addition, obese children have lower HRQOL scores especially physical and social dimensions. Further studies should consider the importance such dimensions in future. On the other hand, the comprehension of the flexibility of obese children in emotional and school functioning dimension may help in approaches to manage the administration of childhood obesity.

Conclusion
Our results demonstrated that obesity was associated with poor HRQOL in children of primary schools of Tehran.
Obese children have lower HRQOL scores especially in physical and social domains but emotional and school functioning domains seemed unaffected.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with re-