Childhood obesity remains one of the most important public health issues globally. Despite decades of efforts to control its prevalence, childhood obesity continues to cause immediate and long-term adverse health consequences(Reference Lobstein, Jackson-Leach and Moodie1,Reference Ng, Fleming and Robinson2) . In the short term, children with overweight or obesity are more likely to suffer from asthma(Reference Farah and Salome3,Reference Rhee, Love and Groth4) , fatty liver disease(Reference Shaunak, Byrne and Davis5), systemic inflammation(Reference Selvaraju, Ayine and Fadamiro6,Reference Oliveira-Santos, Santos and Moreira7) , musculoskeletal disorders(Reference Faienza, D’Amato and Chiarito8), higher metabolic and cardiovascular risk(Reference Ayer, Charakida and Deanfield9,Reference Kelishadi, Mirmoghtadaee and Najafi10) , including high blood pressure(Reference Ding, Cheung and Mak11,Reference Cai, Wu and Wilson12) , and type 2 diabetes(Reference Valaiyapathi, Gower and Ashraf13,Reference Spurr, Bally and Hill14) in addition to several psychological issues such as depression(Reference Quek, Tam and Zhang15,Reference Rao, Zong and Zhang16) , anxiety(Reference Wang, Sun and Zhai17), low self-esteem(Reference Puhl and Latner18), and low academic and cognitive performance(Reference Puhl and Latner18,Reference Yang, Shields and Guo19) . Long-term complications of childhood obesity may include increased risk of developing CVD(Reference Weihrauch-Blüher, Schwarz and Klusmann20,Reference Umer, Kelley and Cottrell21) , type 2 diabetes(Reference Ohlsson, Bygdell and Nethander22–Reference Fang, Zuo and Zhou24), some cancers(Reference Zohar, Rottenberg and Twig25,Reference Hidayat, Du and Shi26) , physical disability(Reference Lee, Pantazis and Cheng27,Reference Henriksson, Henriksson and Tynelius28) and pre-mature death(Reference Twig, Tirosh and Leiba29–Reference Carslake, Jeffreys and Davey Smith31). Furthermore, children with overweight or obesity are more likely to have overweight or obesity in adulthood compared to their normal-weight peers(Reference Whitaker, Wright and Pepe32–Reference Singh, Mulder and Twisk34). Against this background, obesity should be prevented at the earliest possible stage of life to mitigate the risk of many diseases in childhood and adulthood, particularly with the low efficacy of the available treatments(Reference Di Cesare, Sorić and Bovet35,Reference Reinehr36) .
As per the comprehensive plan for Maternal, Infant and Young Child Nutrition, the WHO aims to have zero increase in childhood overweight by 2025(37), which is consistent with the other target for obesity and type 2 diabetes between 2010 and 2025 in the ‘WHO Global Action Plan for the Prevention and Control of Non-communicable Diseases 2013–2020(38,39) . Assessing the progress on these targets requires data on the trends of obesity and overweight, which is critical for policy development(40). Although the data on recent trends of childhood obesity are available from Europe, North America, Australia and China, there is a lack of data on secular trends of childhood obesity from the rest of the world, particularly in the Middle East. When data do exist from the Middle East region, it is usually based on literature reviews of individual methodologically heterogenous studies(Reference Kelishadi, Haghdoost and Sadeghirad41–46) providing contradictory findings. The only exception is a new report from Iran where the authors examined the trends of waist-to-height ratio in children between 2003 and 2015, demonstrating an increasing trend of the prevalence of abdominal obesity in Iranian children(Reference Hosseini, Yousefifard and Baikpour47).
Prevalence of obesity has increased substantially in children and adolescents around the world(Reference Ng, Fleming and Robinson2) but recently, there has been some flattening of trends or probably a reduction in the prevalence of childhood obesity, particularly among children with high socio-economic status in high-income countries(Reference Ng, Fleming and Robinson2,48,Reference Wabitsch, Moss and Kromeyer-Hauschild49) such as the USA(Reference Ogden, Carroll and Lawman50,Reference Skinner and Skelton51) , Canada(Reference Rodd and Sharma52), Spain(Reference Sánchez-Cruz, de Ruiter and Jiménez-Moleón53), Australia(Reference Olds, Tomkinson and Ferrar54), Germany(Reference Keß, Spielau and Beger55), France(Reference Lioret, Touvier and Dubuisson56,Reference Péneau, Salanave and Maillard-Teyssier57) , the Netherlands(Reference de Wilde, Verkerk and Middelkoop58), Switzerland(Reference Herter-Aeberli, Osuna and Sarnovská59) and Sweden(Reference Bygdell, Ohlsson and Célind60). However, there is paucity of data on the recent trends of childhood obesity in Middle Eastern settings, particularly in the oil-rich Arab states in the Gulf region, where the socio-economic status has improved materially during the last few decades with strong shift towards an obesogenic environment, including high energy density foods coupled with low physical activity(Reference Aboul-Enein, Bernstein and Neary61). In these settings, several individual studies have estimated the point prevalence of childhood obesity in various age groups at different points in time and there is a general consensus that the prevalence of childhood obesity is similar to or higher than that in high-income countries in North America or Europe. In Kuwait, for example, it has been reported that 21·6 % and 30·5 % of schoolchildren have overweight or obesity, respectively(Reference Elkum, Al-Arouj and Sharifi62). In a recent study, 42 % of female university students had obesity based on their BMI and 62 % had excessive fatness as measured by bioelectric impedance analysis(Reference Al Hammadi and Reilly63). However, it is not clear whether the weight status of children and adolescents is stabilising, particularly in school aged children 5–19 years, among which the obesity had increased globally more rapidly between 1975 and 2016 compared to younger pre-school children(Reference Di Cesare, Sorić and Bovet35).
For over a decade, Kuwait Nutrition Surveillance System (KNSS) objectively measured the weight and height of thousands of schoolchildren (5–19 years) annually using standardised protocol. This provides a unique opportunity to examine recent trends of obesity in schoolchildren in Kuwait over several years and to compare them to that reported from other high-income countries. This study used individual data records collected over a 13-year period (2007–2019) by KNSS to, for the first time, examine age-specific trends in the prevalence of childhood obesity in schoolchildren in Kuwait based on BMI using various definitions of childhood obesity including the WHO(Reference de Onis, Onyango and Borghi64), the Centers for Disease Control and Prevention (CDC)(Reference Ogden, Kuczmarski and Flegal65), and the International Obesity Taskforce (IOTF)(Reference Cole, Bellizzi and Flegal66,Reference Cole and Lobstein67) .
Methods
Study site and KNSS
Kuwait is a small country with a population of 4·7 million with approximately 25 % of its population under the age of 19 years. School enrolment is extremely high for both males and females(68), with approximately 25 % of students enrolled in private schools. Unlike private schools, public schools are single-sex schools at all levels with total number of students approximately 476 000. There is no clearly recognised race or ethnic classification of the population of Kuwait.
KNSS is a public health activity funded by the government to collect data on the nutritional status of the population by trained data collectors who are employed on a long-term basis for this purpose. The aim of KNSS is to provide nationwide information on the trends of nutritional status on all age groups over time. KNSS is mainly a repeated cross-sectional study that collects data using separate data collection forms on children ≤ 24 months, children> 2 years to <5 years, schoolchildren (5–19 years) and adults. In 2014, KNSS was reviewed, and all data collection forms were updated except for schoolchildren (More details on KNSS have been published before(Reference Al-Taiar, Alqaoud and Hammoud69)). In this age group, only anthropometric measures are conducted with no other data are collected on socio-economic status, dietary habits or physical activity. Such data are collected by a different department using Global School-Based Student Health Survey (GSHS) questionnaire(70). Unlike KNSS, weight and height of schoolchildren in GSHS are self-reported. KNSS includes children aged 5–19 years from public primary, middle and high schools in all governorates (provinces) of Kuwait. In each governorate, at least one school for males and one school for females is randomly selected from each primary, middle and high schools making sure the same school is not selected in two consecutive years. For the current analysis, KNSS data from 2007 to 2019 were used.
Data collection and measurement of weight and height
Body weight was measured to the nearest 0·1 kg using calibrated digital scales (SECA®) without shoes or heavy clothing. Height was measured to the nearest 0·1 cm using a stadiometer (SECA®) with full extended knees and shoes off. Sex and date of birth were extracted from the school records. Data on socio-economic factors, diet or physical activity were not collected at any point in time.
Statistical methods
BMI was calculated by dividing the weight in kilograms by squared height in metres. Age was calculated in days by subtracting the date of birth from the date of data collection and then converted into years. First, BMI-for-age Z-scores were calculated according to the WHO growth reference curves(Reference de Onis, Onyango and Borghi64), and then overweight was defined as >1 sd but ≤2 sd of the WHO growth reference median, while obesity was defined as >2 sd of the WHO growth reference median. Second, BMI-for-age Z-scores were calculated according to the CDC growth reference curves(Reference Ogden, Kuczmarski and Flegal65), which were turned into percentiles, and the overweight and obesity were defined as ≥85th percentiles and ≥ 95th percentiles, respectively. Third, the calculated BMI was categorised according to the IOTF definition(Reference Cole, Bellizzi and Flegal66,Reference Cole and Lobstein67) . Unlike WHO definition that uses arbitrarily chosen cut-off points, IOTF definition uses smooth sex-specific BMI curves, constructed to match the values of 25 kg/m2 (to define overweight) and 30 kg/m2 (to define obesity) at the age of 18 years. The three definitions of overweight and obesity were used to facilitate international comparisons and to help create a pool of studies that allows developing algorithm that converts the prevalence of overweight and obesity from one definition to another. In this article, the term ‘overweight’ does not include obesity unless stated otherwise.
The age-specific prevalence (95 % CI) of obesity and overweight (including obesity) was plotted using the three definitions of obesity in children over the study period while stratifying by sex. Moreover, the overall and the age-specific mean BMI-for-age of Z-scores calculated from the WHO reference curves and their 95 % CI were plotted over years of study while stratifying by sex. With the large number of observations, the regular scatterplot which plots every data point became too crowded to interpret visually; therefore, binned scatterplots were used controlling for age and stratifying by sex generated by STATA command ‘binscatter’ to graphically present the change in the mean BMI-for-age Z-score over the study period.
The analysis started by calculating the odds of obesity and overweight (including obesity) over the study period while stratifying by sex and adjusting for age. Three separate analyses according to the three definitions of overweight and obesity were conducted to investigate the trends in obesity and overweight (including obesity) over the study period using log-binomial regression models. In this analysis, the prevalence ratio for obesity was estimated while adjusting for age and stratifying by sex fitting the main independent variable (year of measurement 2007 to 2019) as a continuous variable. In this analysis, a quadratic time (year of study) term was fitted to assess non-linearity, which was found to be significant. STATA 14.2(71) was used to fit models with both linear and quadratic time terms with obesity (yes = 1 and no = 0) as well as overweight (including obesity) as the binary dependent variables. In these models, the standard error was calculated while taking the clustering effect into account using the variance–covariance matrix (vce) option in STATA. Guided by this analysis, aggregated data were used in NCI’s Joinpoint software(72) to identify the location of joinpoints for each age group and sex as per the guidelines(Reference Ingram, Malec and Makuc73). In this analysis, the proportions were calculated within the joinpoint software with all other settings according to the Joinpoint software default. The joinpoints were also investigated using hockey stick regression method using STATA command ‘loghockey’.
Results
Of all data records, 392 records were excluded because BMI-for-age Z-scores were biologically implausible (BMI-for-age Z-score < −5 or BMI-for-age Z-scores > +5). The total number of schoolchildren included in this analysis was 172 603 of which 85 764 (49·69 %) were females. Table 1 shows the distribution of the study participants by age and year of study. The crude prevalence of overweight and obesity over the entire 13-year study period was 19·86 % (95 % CI 19·15, 20·61) and 25·63 % (95 % CI 23·20, 28·23), respectively, according to the WHO definition, 16·06 % (95 % CI 15·35, 16·81) and 25·01 % (95 % CI 22·62, 27·56), respectively, according to the CDC definition, and 20·57 % (95 % CI 19·59, 21·59) and 19·59 % (95 % CI 17·53, 21·83), respectively, according to the IOTF definition.
Table 1 Distribution of the study participants by age and years of the study
Trends in the prevalence of overweight and obesity in schoolchildren
According to the WHO definition, the prevalence of overweight and obesity, respectively, increased from 17·73 % and 21·37 % in 2007 to 20·19 % and 28·39 % in 2019 (P for trends in odds of each overweight and obesity <0·001). This remained evident while stratifying by sex and adjusting for age. Table 2 shows the prevalence of overweight and obesity by sex and year of data collection during the study period. Supplemental Tables S1 and S2 show the prevalence of overweight and obesity by sex and year according to the CDC definition and the IOTF definition, respectively.
Table 2 Prevalence of overweight and obesity in schoolchildren (5–19 years) in Kuwait between 2007 and 2019 according to the WHO definition of obesity in children
Obesity was defined as >2 sd while overweight was defined >1 sd but ≤2 sd of the WHO growth reference median.
P-value, testing for trends during the study period.
Figure 1 shows the prevalence of obesity and overweight (including obesity) and their 95 % CI over the 13-year period according to the WHO, the CDC and the IOTF definitions among males, females and both sexes. The age-specific prevalence of obesity and overweight (including obesity) and their 95 % CI over the 13-year study period stratified by sex are shown in the supplementary material according to the WHO definition (see online supplemental Fig. S1(a) and S1(b)), the CDC definition (see online supplemental Fig. S2(a) and S2(b)) and the IOTF definition (see online supplemental Fig. S3(a) and S3(b)). All figures clearly indicate that the prevalence of obesity generally increased over the 13-year study period and the trends have slowed down in later years.
Fig. 1 Prevalence of overweight (including obesity) and obesity and their 95 % CI over 13-year period according to the WHO, the CDC and the IOTF definitions among males, females and both sexes. CDC, the Centers for Disease Control and Prevention; IOTF, the International Obesity Taskforce
The analysis of trends in the prevalence of obesity and overweight (including obesity) in schoolchildren (5–19 years) between 2007 and 2019 after adjusting for age and stratification by sex is shown in Table 3. This analysis confirmed that the prevalence of obesity and overweight (including obesity) increased over the study period according to the WHO, the CDC and the IOTF definitions. To investigate whether the prevalence of obesity and overweight (including obesity) started to level off (nonlinearity), we fitted a quadratic term of the explanatory variable (years of the study) in addition to the linear term. This was statistically significant in the model indicating that the trend is non-linear. Guided by this and the Fig. 1, we conducted a joinpoint analysis to search for one joinpoint around which the increase of obesity has abated. We found one joinpoint around 2013 for both sexes, which was confirmed using hockey stick regression analysis. To investigate this further, we omitted data from 2013 and another joinpoint emerged around 2010. Then, we examined the trends before and after 2013 while stratifying the analysis by sex and adjusting for age (data not shown). This analysis confirmed that in males, there was a significant upward trend in the prevalence of obesity and overweight (including obesity) before and after 2013. In females, significant upward trend in the prevalence of overweight (including obesity) before and after 2013 was found, but no significant trend was found after 2013 in the prevalence of obesity, which suggest that the prevalence of obesity may have stabilised in females. These findings remain the same when the analysis was repeated using the CDC and the IOTF definitions.
Table 3 Trends in the prevalence of overweight and obesity in schoolchildren (5–19 years) in Kuwait between 2007 and 2019 using log-binomial regression
CDC: Centers for Disease Control and Prevention; IOTF: International Obesity Taskforce; PR: prevalence ratio; 95 % CI.
After taking clustering into account, quadratic time term was significant at 5 % level of significance indicated by ‘**’ or at 10% level of significance indicated by ‘*’.
† In all models, we adjusted for age as a continuous variable.
Trends in the BMI-for-age Z-score in schoolchildren
The mean BMI-for-age Z-scores and their 95 % CI calculated from the WHO reference curves over the 13-year study period in males and females and both sexes are shown in Fig. 2. Supplemental Fig. S4(a) and S4(b) show the age-specific mean BMI-for-age Z-scores over the study period in males and females, respectively. Figure 3 shows the bin scatter of BMI-for-age Z-score over the study period after controlling for age while fitting regression line with quadratic time term. All figures showed an increase in the mean BMI-for-age Z-score during the 13-year study period in a pattern consistent with the increase in the prevalence of obesity and overweight demonstrated above.
Fig. 2 Mean BMI-for-age Z-scores and their 95 % CI calculated from the WHO reference curves over the 13-year study period in males and females and both sexes
Fig. 3 Bin scatter of BMI-for-age Z-score over 13-year period after controlling for age fitting regression line with quadratic time term
Discussion
For the first time, using individual data records of thousands of children, this study examined the age-specific trends in the prevalence of overweight and obesity in male and female schoolchildren aged 5–19 years in Kuwait, an oil-rich country in the Middle East that represents a group of countries in that region where an obesogenic environment has expanded for years(Reference Aboul-Enein, Bernstein and Neary61). Overall, the prevalence of overweight and obesity in schoolchildren (5–19 years) has risen over the past 13 years in Kuwait and that it has recently started to level off in females. These findings are extremely important because there is little data on trends of childhood obesity from Middle Eastern settings except for few literature reviews of methodologically heterogenous studies that show contradictory findings(Reference Kelishadi, Haghdoost and Sadeghirad41–46).
There was an increase in the prevalence of childhood obesity as well as overweight (including obesity) in Kuwait according to the three definitions of childhood obesity and overweight (WHO, CDC and IOTF). At the end of the study period in 2019, the prevalence of obesity and overweight was 20·19 % and 28·39 %, respectively, which is similar to that in Italy(Reference Lauria, Spinelli and Buoncristiano74) but higher than that reported in several high-income countries(Reference Herter-Aeberli, Osuna and Sarnovská59,75) . In fact, a recent literature review of obesity in schoolchildren at a global level has identified Kuwait as having the highest prevalence of obesity and overweight in the Middle East and North Africa (MENA) region(Reference Di Cesare, Sorić and Bovet35). Over the last few decades, high oil revenue has contributed to substantial improvement in the standard of living for citizens that includes highly subsidised foods. Exclusive breast-feeding during the first 6 months of birth, which has been linked to lower obesity rates during adolescence(Reference Horta, Loret de Mola and Victora76), is also low; and the government provides highly subsidised formula milk(Reference Al-Taiar, Alqaoud and Hammoud69). GSHS data suggest that more than 62 % of schoolchildren spend 3 or more h/d sitting and watching television, playing computer games, or talking with friends. Also, less than 17 % of schoolchildren were physically active at least 60 min/d; and more than 52 % of students consumed carbonated soft drinks 1 or more times/d during the 30 d before the survey(70).
In this study, a significant upward trend in the prevalence of obesity and overweight (including obesity) was found in schoolchildren from 2007 until 2014 in both sexes. At a global level, it has been reported that the obesity in schoolchildren has increased over the last few decades in a higher rate compared to that in pre-school children(Reference Di Cesare, Sorić and Bovet35). It has been suggested that as children get older, that is, from pre-school age to adolescence, they gain independence and thus are capable of making their own lifestyle choices such as choosing which foods they consume and the level of physical activity that they participate in, which may explain the rapid increase in the prevalence of obesity in this particular age group(Reference Di Cesare, Sorić and Bovet35,Reference Hill77,Reference Hardcastle, Thøgersen-Ntoumani and Chatzisarantis78) . Throughout the study period, girls had a lower prevalence of obesity and higher prevalence of overweight than boys, a pattern that has been reported in Iranian children(Reference Kelishadi, Haghdoost and Sadeghirad41). Several studies have reported higher prevalence of obesity in boys compared to girls in different settings(Reference Sánchez-Cruz, de Ruiter and Jiménez-Moleón53,Reference Bancej, Jayabalasingham and Wall79) including Kuwait(Reference Elkum, Al-Arouj and Sharifi62). Currently, there are calls to investigate the rapid increase of obesity in boys and whether they are more susceptible to obesogenic pressures(Reference Di Cesare, Sorić and Bovet35).
Although the prevalence of childhood obesity has increased in the last 13 years, the data suggest that the prevalence may have levelled off in females since 2013 as there was no significant trend in the prevalence of obesity in the period 2014–2019 after adjusting for age. When the analysis was stratified by sex and age, there was no increase in obesity among girls in any age group except a slight increase in those older than 17 years in the period 2014–2019. The prevalence peaked in year 2013 with no clear explanation. Therefore, the analysis was repeated while omitting the data from 2013, which produced similar results. Although several recent reports indicate that the trend in childhood obesity may have slowed down or even decreased in high-income countries(48–Reference Sánchez-Cruz, de Ruiter and Jiménez-Moleón53,Reference de Wilde, Verkerk and Middelkoop58–Reference Bygdell, Ohlsson and Célind60,75,Reference Olds, Maher and Zumin80) , this issue remains highly controversial and poorly understood. In England, based on data from primary care electronic records, it was reported that the annual increase in the prevalence of obesity stabilised in younger (2–10 years) children but not older (11–15 years) children among which obesity continues to increase with no difference between males and females(Reference van Jaarsveld and Gulliford81). In the same setting, an earlier study showed that the prevalence of obesity among girls declined between 2005 and 2007 while it remained stable or increased in boys(Reference Stamatakis, Zaninotto and Falaschetti82). Surprisingly, in the same setting, upward trends have been reported for both males and females since 2012(75). In Swedish schoolchildren, it was demonstrated that the obesity declined in females not males(Reference Sjöberg, Lissner and Albertsson-Wikland83). Also, although within a short period of time frame (2007 to 2010), data from the WHO European Childhood Obesity Surveillance Initiative showed upward trend in some countries and downward trends in other countries in Europe over the same period of time(Reference Wijnhoven, van Raaij and Spinelli84). Finally, in children aged 3 to 17 years, a relatively stable rates of obesity were reported in France up to 2012, while since 2011 the rates of obesity have increased in boys in the USA(75).
The data suggest that obesity in female schoolchildren may be slowing down in recent years according to the WHO, the CDC and the IOTF definitions, which was supported by similar trends in the mean BMI-for-age Z-score that showed significant increase over time but started to level off around 2013–2014. This is consistent with several reports from Europe(Reference Stamatakis, Zaninotto and Falaschetti82,Reference Sjöberg, Lissner and Albertsson-Wikland83) and a recent global review of trends in childhood obesity, where a higher increase in obesity was observed in boys in comparison to girls between 1975 and 2016(Reference Di Cesare, Sorić and Bovet35). Our findings highlight the importance of sex differences, which should be considered in the intervention programmes designed to combat obesity in children and adolescents(Reference Sweeting85,Reference Simen-Kapeu and Veugelers86) . Several plausible reasons for the recent plateau in the trend of childhood obesity have been proposed including the notion that the rates have reached a point of saturation(Reference Olds, Maher and Zumin80). The other possible reason is the cumulative impact of public health campaigns, which may have resulted in body weight awareness and hence stabilisation of the prevalence of obesity in female schoolchildren. However, it is not clear whether this stabilisation will be a critical juncture or just a temporary trend. Knowing this may require further monitoring of future trends while including data from different sources. It is noteworthy that upwards trends in childhood obesity have been reported in countries where childhood obesity had plateaued or even declined(75,Reference Wijnhoven, van Raaij and Spinelli84) .
In this study, three definitions of obesity and overweight among children were used to facilitate international comparisons with our data. The WHO definition provided the highest estimate of the prevalence, while the IOTF definition showed the lowest estimate of the prevalence of both overweight and obesity among schoolchildren in Kuwait. This is similar to that reported earlier from the region(Reference Musaiger, Al-Mannai and Al-Marzog87) and Italy(Reference Vidal, Carlin and Driul88). The results of this study also suggest that the WHO, the CDC and the IOTF definitions can all be used to ascertain the trends of overweight and obesity in schoolchildren over time in the region, but the interpretation of the magnitude of the prevalence should take into account the definition used. Using national recently collected BMI-for-age reference data is not recommended to define overweight and obesity in children as it may underestimate childhood obesity providing false assurance that may undermine the need for public health interventions(Reference El-Ghaziri, Boodai and Young89).
The study has several strengths including the large sample size which allowed us to investigate age-specific trends in overweight and obesity after stratification by sex. Also, the data were based on objectively measured weight and height rather than reported weight and height by children or their parents. It has been demonstrated that adolescents and adults underreport weight and overreport height, which leads to lower BMI values(Reference Akinbami and Ogden90,Reference He, Cai and Fan91) . Moreover, weight and height measurements were conducted by a trained team who are employed on a long-term basis for this purpose hence allowing for consistency of data collection. However, the study has some limitations including the lack of data on socio-economic status. Given the objectives of this study (i.e. examining the trends of obesity over time), this should cause little concern because we are not looking for specific factors associated with the trends in childhood obesity. Nevertheless, the lack of data on socio-economic status may conceal different patterns of trend in childhood obesity between those with high and low socio-economic status. Private schools are excluded from KNSS which may hinder the generalisability of the findings. Additionally, although BMI is a valid, convenient and reliable method of assessing weight status (whereby weight becomes a surrogate for fatness), the ideal definition of obesity is based on the percentage of body fat but this is impracticable for epidemiological studies as it requires body composition analysis using reliable methods such as dual-energy X-ray absorptiometry (DEXA)(Reference Wellens, Roche and Khamis92).
In conclusion, the prevalence of obesity and overweight among schoolchildren in Kuwait has risen over the last 13 years and trends are similar across various definitions. The findings of this report indicate that obesity is no longer increasing at the same pace and there is some evidence that the prevalence of obesity in female schoolchildren has plateaued. These findings should not be interpreted as grounds for optimism as they may reflect a transient trend and nevertheless, the current prevalence is extremely high in both sexes. With this current level of childhood obesity, the country needs to go beyond increasing knowledge and awareness of obesity, hence intensifies community-based interventions as well as plans, tests and implements school-based interventions to reduce the prevalence of obesity in schoolchildren.
Acknowledgements
Acknowledgements: The authors would like to thank Dr Mona Alsumaie and the data collectors at KNSS for hardwork and dedication. Financial support: KNSS is funded by the Kuwaiti government. This research received no specific grant from any funding agency, commercial or not-for-profit sectors. Conflict of interest: There are no conflicts of interest. Authorship: A.A.-T. analysed the data and drafted the manuscript. N.A.-Q. supervised data collection, contributed to data interpretation and revised the manuscript with significant intellectual input. A.H.Z. contributed to data interpretation and revised the manuscript with significant intellectual input. F.A. supervised data collection and data entry, contributed to data management and revised the manuscript with significant intellectual input. M.S. supervised data collection and data entry, contributed to data management and revised the manuscript with significant intellectual input. R.S.A. contributed to data management and data analysis in addition to drafting the article. H.A.J. contributed to data interpretation and revised the manuscript with significant intellectual input. M.A.-E. contributed to data interpretation and revised the manuscript with significant intellectual input. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki. The study was a secondary data analysis of completely anonymous data collected by Kuwait Nutritional Surveillance System (KNSS). All procedures of KNSS were approved by the Ethics Committee at the Ministry of Health in Kuwait (No. 98). Because KNSS is a public health activity, measuring weight and height of schoolchildren does not require parental informed consent according to the research regulations in Kuwait.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980021003177
