An overview of pediatric obesity

As health care workers in the fields of pediatric diabetes and endocrinological problems, pediatric obesity strongly merits our concern. We need to understand the nature, extent, and implications of the problem not only in our role as advocates and carers for the child but also for the adult and parent that the child will become. Evidence is now increasing that the predilection for pediatric obesity may start in the womb (1, 2) and is affected by birth weight and maternal nutrition (3, 4), maternal glucose metabolism (5), parental smoking (6, 7), infant feeding practices (8), child rearing practices, psychological health and recreational options, and the home and school environment. Pediatric obesity can lead to immediate health problems such as early-onset type 2 diabetes mellitus and glucose intolerance (9) but is also linked to hitherto ‘adult’ problems, such as the metabolic syndrome, cardiovascular disease, and orthopedic problems (10, 11). It behooves us then to understand and address the problem of pediatric diabetes from the point of view of primary prevention (understanding the epidemiology and devising preventive strategies for the population and for the individual), secondary prevention [e.g. screening for the metabolic syndrome and obstructive sleep apnea (OSA)], and tertiary prevention (e.g. treating pediatric type 2 diabetes and hypertension).

Measuring the problem

Pediatric obesity research and meaningful comparison of data across countries are hampered by the lack of an internationally accepted measure of obesity. Some degree of fatness is physiological, and experts differ on what is the best measure of fatness, the effect of application of a variety of existing standards on the prevalence of obesity in the same population, and the role of factors such as visceral adiposity and natural history in the definition of obesity (12, 13). Most countries either express data with respect to national data sets and standards or with respect to another country’s published national standards, such as the US CDC’s growth charts. The International Obesity Task Force (IOTF) convened a workshop in 1994, and the workshop concluded that the body mass index (BMI; in kg/m²) offered a reasonable measure with which to assess fatness in children and adolescents and that the standards used to identify overweight and obesity in children and adolescents should agree with the standards used to identify grade 1 and grade 2 overweight (BMI of 25 and 30, respectively) in adults (12). Cole et al. published charts based on data from Brazil, Great Britain, Hong Kong, the Netherlands, Singapore, and the USA to provide age- and sex-specific centile curves that at age 18 yr passed through the widely used cutoff points of 25 and 30 kg/m² for adult weight and obesity (14). Concerns include the overestimation of overweight and underestimation of obesity, and various charts have been published to address body fat as opposed to BMI as measures of obesity (15, 16).

The use of a BMI of 25 and 30 in adults is related to outcome measures of morbidity and mortality such as the risk of cardiovascular disease and diabetes, but corresponding data are less available for children, especially when it comes to defining cutoff points. The relationship of BMI to body composition is complex and is affected by age, sex, ethnicity, and sexual maturation. The Western Pacific Regional Office of the World Health Organization criteria of 2000 recommended adult BMI cutoffs for Asian adults be modified to 23 kg/m² for overweight and 27 kg/m² for obesity, but problems remain because Asian body compositions are not homogenous for either fat distribution or predictiveness of cardiometabolic risk factors (17–20). Citations of the IOTF definition have increased since publication, but Chinn in 2006 reported that less than half of the papers on childhood obesity published since have used the definition. Most used the 95th centile of a national distribution to define obesity (21). In the USA, the 2000 Centers for Disease Control and Prevention Growth Charts for the United States are used to define overweight (≥95th percentile of the sex-specific BMI for age) and at risk for overweight (BMI ≥85th percentile but ≤95th percentile) and are based on national data from 1963 to 1994.

The category at risk for overweight is intended to identify children who should be referred for a second level of screening to determine if there are any additional health risks that would warrant intervention. The probability of adult obesity (BMI of >30 kg/m²) is ≥50% among children >13 yr of age, whose BMI percentiles meet or exceed the 95th percentile for age and gender. BMI-based overweight categorization for individuals, particularly for racial/ethnic minorities with differences in body composition, may have limited validity because BMI measures cannot differentiate between increased weight for height attributable to relatively greater fat-free mass (muscle, bone, and fluids) and that attributable to greater fat. Screening to categorize overweight among children under age 12 or 13 who are not clearly overweight may not provide reliable risk categorization for adult obesity (22). The IOTF-recommended BMI cutoff values may have varying sensitivity in different populations and may underestimate the local prevalence of childhood obesity (23). Sometimes, weight figures alone may be misleading as BMI comparisons may reveal that two groups with the same mean weight may have a lower mean height and thus may have a lower lean mass to fat mass ratio (24). If the objective is to screen for individuals at risk, population-specific measures rather than international cutoff values should be used.

By any measure, obesity and overweight are increasing worldwide

What is clear, however, is that childhood obesity is increasing in prevalence and severity but at different rates among different groups of individuals. Even within countries and communities, the prevalence of obesity may differ greatly between ethnic groups, age-groups, males and females, and groups of differing socioeconomic status. In most cases, the reasons are apparent to observers on the ground as they would understand the differences between food purchase and consumption patterns, exercise and activity patterns, local beliefs and taboos, and suchlike.

Data from around 1992 to 1994 showed the prevalence of obesity (BMI >95th centile by the CDC standard) among US children aged 6–19 yr was 11.1%, while the prevalence of overweight was 14.3% (25). By 1999–2002, 31.0% of US children were at risk for overweight or overweight and 16.0% were overweight, with disproportionately more African Americans and Hispanic Americans origin being affected (26). Rates of overweight increased through adolescence from 7 to 10% in the United States in girls of Caucasian origin and from 17 to 24% in their African-American counterparts (27). On the same continent, the prevalence of overweight among indigenous Canadian children was as high as 60% (28).

During the same period, 1992–1994, the prevalence of obesity was 6.0% in Russia and 3.6% in China, while the prevalence of overweight was 10.0% in Russia and 3.4% in China (25). Data from the China national surveys in 2002 showed that the prevalence of overweight and obesity in children aged 7–18 yr was 15%, prevalence of obesity was 6.6%, an increase of 28 and 4 times, respectively, between 1985 and 2000, a trend that was particularly marked in boys, and in the urban compared with the rural areas (29, 30). Increasing obesity has also been reported in Korea (31), Taiwan (32), Australia (33, 34), and Thailand (35).

The influence of socioeconomic status, sex, and age varies between regions and nations and ethnic groups. Higher socioeconomic status subjects were more likely to be obese in China and Russia, but in the USA, low socioeconomic status groups were at a higher risk (36, 37). Obesity was more prevalent in urban areas in China but in rural areas in Russia (25). Children of migrant parents were more likely to be obese than the children of German parents at school entry (38). In the United States, although parental SES is associated inversely with childhood obesity among Whites, higher SES does not seem to protect African-American, Hispanic (39), and Kuwaiti (40) children against obesity.

Some of this might be explained by different societal values and consumption patterns (30). In developing countries and in migrant communities, parents may use their newfound affluence to give their children the abundance of food they never had. Plump children are seen as desirable in many cultures (29, 41), and obesity is viewed in some cultures as a status symbol. Children from deprived inner-city backgrounds or indigenous and minority communities in rich countries may have restricted food choices and lack of exercise facilities. Studies have shown that residents in low-income areas in the United States are less likely to have stores stocking healthy foods like fruits and vegetables close at hand, thus affecting consumer choice and consumption patterns. City dwellers may on the other hand have better access to health information and have found other ways to display love and wealth other than by overfeeding their children, who may themselves be exposed to ‘thin is better’ advertising and peer pressure. In a nationally representative survey of childhood eating habits from the USA, increased fast-food consumption was independently associated with male gender, older age, higher household incomes, non-Hispanic black race/ethnicity, and residing in the South (42). School environments also play an important role – if the school canteen or lunchroom does not stock healthy foods, children would tend to choose calorie dense and less healthful foods (43, 44). Vending machines and easily available snack foods may also compete with healthy foods, even in schools with nutritional guidelines (44, 45).

The relationship between gender and childhood obesity is also often population specific. Factors influencing gender differences include societal norms for female beauty, expressing favoritism by offering larger portions of food, and society acceptance of rough and tumble play or access to sporting activities. The prevalence of overweight was 15% among girls and 14% among boys in Dresden, part of the former East Germany, but 24% in girls and 22% in boys in Munich, part of former West Germany (46). On the other hand, there are more overweight 13–18-yr-old boys compared with girls (17.8 vs. 15.8%) in Chennai, South India (47) and Spain (15.6% in boys vs. 12% in girls) (48). A Japanese study showed less sex differences over time, but a greater increase in obesity was noted in rural vs. urban settings (49). Among a sample of US college students, males were found to be less conscious of healthy food choices than females (50).

Childhood obesity and adult obesity

Childhood obesity has been shown to track through adulthood. In the Muscatine Coronary Risk Factor project, it was found that overweight adolescents have a 50–70% chance of becoming overweight or obese adults, but 30% of fat children will become leaner in adulthood, while a similar proportion of thin children become obese adults (51). Kotani et al. noted in 1976 that 36% of infants exceeding the 90th centile in weight at least once in infancy were overweight as adults, as compared with 14% of the average age and lightweight infants (52). In Japan, approximately 32% of obese boys and 41% of obese girls grow into obese adults, and the degree of obesity is a predictive factor for adult obesity (53) Girls who were overweight during childhood were 11–30 times more likely to be obese in young adulthood (27). The notion that children will grow out of childhood obesity should be put to rest once and for all.

The origins of pediatric obesity

The intra-uterine environment influences obesity in later life (1) and also the development of complications of obesity. Prevalence of childhood obesity also varied with parental smoking (6, 7), infant feeding practices (8), child rearing practices, psychological health and recreational options, the home and school environment, and parental obesity. The 2002 China National Nutrition and Health Survey found that the prevalence ratio of obesity went up up to 12.2 if both parents were obese (54). A history of previous treatment for childhood malignancy is also a risk factor for subsequent obesity and the metabolic syndrome (55).

Children who are LGA at birth and exposed to an intra-uterine environment of either diabetes or maternal obesity are at increased risk of developing the metabolic syndrome. However, Kral et al. reported that when obese mothers underwent obesity surgery and lost weight, they were able to prevent transmission of obesity to their children (2, 3). Yajnik found that Indian babies who were born small but had grown heavy (or tall) were the most insulin resistant and had the highest levels of cardiovascular risk factors. Accelerated growth in relation to mid-parental height was similarly predictive (2) Mothers with gestational diabetes are also at risk of having fetal macrosomia with ensuing fetal hyperinsulinism (56) and offspring at greater risk obesity (19%) (57) and impaired glucose tolerance (up to 36%) (58, 59). Obese women who have larger babies and not fulfilling criteria for gestational diabetes have also been reported to have children with a higher risk of the metabolic syndrome in later life (5).

Early adiposity rebound

Research on early adaptive mechanisms begun in utero for adverse nutritional environments in postnatal life has given us insights on the role of epigenetic programming and genetic differences in obesity through effects on intake, energy balance, and energy expenditure (60, 61). The age of adiposity rebound refers to the age after infancy at which the individual’s BMI is lowest and after which starts to rise to adult levels. The cumulative incidence of type 2 diabetes was 4.7 times higher in a group whose adiposity rebound occurred before 5 yr compared with those in whom it occurred after 7 yr (p < 0.001) (62). In another prospective study of 1492 men and women between 26 and 32 yr of age, those who had IGT and diabetes had a low BMI up to the age of 2 yr followed by an early adiposity rebound and an accelerated increase in BMI until adulthood, even though none of these subjects was obese at the age of 12 yr (63).

It has been suggested that low body fatness before the adiposity rebound suggests that an energy deficit occurred at an early stage of growth. It might be attributed to a low-fat, high-protein diet fed to infants at a time of high energy needs, with the high protein triggering height velocity and the low fat decreasing the energy density of the diet and hence energy intake (64). Compared with cows’ milk, human milk has higher fat and lower protein levels, and this may contribute to its beneficial effects on growth and nutrition. It is postulated that some of the insoluble cases of obesity are the result of an inborn condition of a very low muscle mass and more fat tissue in the early years of childhood, such as in SGA infants and that this will lead to an increase risk of diabetes in later life (63).

Genetic syndromes as a cause of obesity

There are reports of single gene mutations with haploinsufficiency in human subjects and single gene disorders resulting in obesity, but most cases of obesity are likely the result of subtle interactions of several related genes with environmental factors, which favor the net deposition of calories as fat, culminating in the obese phenotype. Obesity is unlikely to be caused by a single gene defect unless it is extreme (BMI > 60) or present in an isolated population group (65). There are, however, several recognizable conditions the clinician should consider when seeing a child with severe obesity. They include, Prader–Willi syndrome (66), Cohen syndrome (67), congenital leptin deficiency and leptin receptor deficiency, pro-opiomelanocortin defects, prohormone convertase 1 deficiency, human melanocortin receptor (MC4R) deficiency, cocaine- and amphetamine-related transcript defects, and so on (68).

The consequences of obesity and associated comorbidities

Obesity is associated with more chronic (continuing) medical conditions than smoking or excessive drinking (69). Obese children are at increased risk of having diabetes or impaired glucose tolerance (9, 28). Obese youth develop peripheral insulin resistance early. They exhibit increased deposition of lipid in the visceral and intramyocellular compartments, a relatively reduced adiponectin level and elevation of inflammatory cytokines, impaired glucose tolerance, and overt type 2 diabetes as a result of reduced first- and second-phase response of the beta cell to glucose loads. Certain populations, in particular ethnic minorities and native or aboriginal peoples who have undergone recent significant lifestyle changes (e.g. active to sedentary and rural to urban transition), seem to be at greater risk (5,70–75). Obese children may present with mixed features of type 1 and type 2 diabetes mellitus (76). Sometimes, it may be difficult to distinguish between the two (77), and common sense suggests that the physician treat the patient’s phenotype and not the classification (78). Magnesium deficiency has been reported to be a possible risk factor in the development of type 2 diabetes in obese children (79).

Many obese children have one or more features of the metabolic syndrome (10, 11, 31,80–83) with hypertension and hyperinsulinism, dependent on the severity of obesity (84) Obese children and adolescents have approximately fourfold higher risks for hyperlipidemia (31) and a twofold to fourfold higher risk for hypertension, although those who are overweight and fit have a slightly lower risk (32) than if they were unfit and overweight. The risk of hypertension in children increases across the entire range of BMI values and is not defined by a simple threshold effect (31, 80, 81). The prevalence of insulin resistance in obese adolescents was 52.1% (85). Clinically, healthy obese children have a higher degree of albuminuria and beta-2-microglobulinuria than normal weight children, indicating early renal glomerular and tubular dysfunction as a consequence of childhood obesity (86).

Childhood LDL-cholesterol, systolic blood pressure (BP), BMI, and smoking are all significantly associated with adult intimal–medial thickness (87). Childhood adiposity is a predictor of heart mass in the adult (88). A clustering of CVD risk factors may begin during early adolescence among the obese (89, 90). Treatment for hyperlipidemia may need to be started early, especially in families with adverse risk factors (91).

Obesity has also been linked many other health issues (92), such as to psychological problems, functional impairment in sports activities (93), and orthopedic problems such as tibia vara (94). Recent data suggest that unlike obese adults, obese children and adolescents may have decreased bone strength (95, 96). Syndromic obesity, such as that seen in Prader–Willi syndrome, may also be associated with behavior issues, low bone density, risk of fractures, and scoliosis (97).

Obesity is also a risk factor for polycystic ovarian syndrome (98–100) and OSA (101, 102). Children who are morbidly obese may have a prevalence of OSA of 13% (103) for which tonsillectomy (104), bariatric surgery (105), and continuous positive airway pressure ventilation may be helpful. Children with OSA may have daytime somnolence, and this would further reduce their ability and proclivity to exercise.

There have also been many reports in the literature linking obesity and asthma, although the precise mechanisms are not clear. It may be because of a clustering of risk factors such as urbanization, pollution, and respiratory allergies or perhaps through the action of inflammatory mediators (106, 107). Asthma is an inflammatory process, and obesity leads to an inflammatory response in the tissue (108–110). In many societies, childhood asthma is seen as a condition in which sports and physical activity should be limited. Adequate, well-informed management of childhood asthma, allowing the child to partake in normal levels of exercise and activity, would help reduce childhood obesity.

Overweight and obese children often suffer from low self-esteem, bullying, and teasing. In obese patients seeking treatment, there is an increased prevalence (40–60%) of psychiatric morbidity, most commonly depression (111). Severely obese children and adolescents have lower health-related quality of life (QOL) measures than children and adolescents who are healthy and who have similar QOL as those diagnosed as having cancer (112, 113). Put simply, some children eat because they feel depressed and unhappy, and some are unhappy because they are fat. Eating, especially eating snack foods, affords a sense of temporary relief and gratification. As eating is a form of gratification behavior, efforts to reduce childhood obesity should look at alternative means to promote feelings of gratification and self-esteem. Those who would curb smoking in obese children should do likewise.

Financial consequences

The financial costs of childhood obesity are difficult to quantify because they go beyond the cost of increased health care needs during the childhood years. For US youths (6–17 yr of age), obesity-associated annual hospital costs (based on 2001 constant US dollar value) increased more than threefold from 1979–1981 to 1997–1999 (114). In Germany alone, the cost of illness of an obese child and an obese child with type 2 diabetes were three and eight times the mean costs of treating a non-obese child of the same age (5–20-yr rehabilitation) (115). If tracking of obesity and cardiovascular risk factors, subsequent increase in adult morbidity, and loss of productivity are taken into account, the worldwide costs are astronomical.

What do we need to know about an obese child?

Beyond the need to determine the obese child’s current state of health and disease, there is the need to find out the antecedent causes and risk factors for the child’s present obesity as well as prognostic factors for future morbidity and the probability of success of any interventions. An understanding of these factors in a community would also help to guide public health messages and health care priorities.

For the individual, we would need the following:

Physical examination should include a lookout for:

The child would then need a risk assessment based on the above history and physical examination and screening tests, an assessment of food intake and feeding practice, an assessment of physical activity patterns, and a talk with the adolescent or child and the parent of a younger child to assess the psychological status, looking after depressive symptoms, anxiety, and loss of self-esteem (116).

Biochemical investigations should include a formal oral glucose tolerance test with insulin levels or at least preferably a blood glucose and insulin levels 2 h after a 75 g (or 1.75 g/kg, max 75 g) oral glucose challenge, the corresponding serum insulin level or a paired fasting insulin and glucose level and hemoglobin A1c, and a liver function test. Where indicated, a sleep study for obstructive sleep apnoea hypoventilation syndrome (OSAHS) and an echocardiogram evaluation of left ventricular mass, structure and function, and right-sided abnormalities related to increased pulmonary artery pressure may be indicated. Radiological studies may be necessary to diagnose slipped capital femoral epiphyses and Blount disease.

Strategies for pediatric obesity

We need to understand and address the problem of pediatric diabetes from the point of view of primary prevention (understanding the epidemiology and devising preventive strategies for the population and for the individual), secondary prevention (e.g., screening for the metabolic syndrome and OSA), and tertiary prevention (e.g., treating pediatric type 2 diabetes and hypertension). It appears that despite the publicity about increasing childhood obesity, health care providers may be underdiagnosing, under investigating, and undertreating childhood obesity (117).

Screening of children for overweight should begin in the first year of life. Advice should be offered to parents regarding the prevention of overweight as soon as a child begins to cross BMI percentiles and should not be postponed until the child or adolescent is at or above the 95th percentile of BMI for age and gender (118).

Treatment approaches to obesity

What measures are effective? Obesity is the end result of interactive influences of environment, biology, and behavior, and for some, it is an individual medical illness. There is relatively little evidence for effective treatment strategies for childhood obesity (64), and interventions shown to be effective in research protocols may not be translatable or scalable to clinical practice (119). The degree of parental involvement and cooperation as well as the individual family’s resources may limit what we can do. Psychological factors may be very important. There are many similarities between overeating and anorexia nervosa (120).

At the risk of oversimplification, I would like to propose the following outline for further reading and development, as it is clear from the literature and anecdotal evidence that what works in one context may have little relevance and applicability in another, although we can obviously learn from each other’s best practices.

Conclusion

The successful management of pediatric and adolescent obesity remains a challenge to caregiver and patient alike. There are no quick fixes and no easy answers. No two obese patients will be truly alike, and while whole populations may see improvements in the prevalence of obesity as a result of societal changes, we will need to carefully assess the individual etiology, complications, and possible treatment options for each obese individual who presents to us. Obesity, like diabetes mellitus, is a serious and multifaceted condition and needs our urgent and professional attention.

Conflict of interest statement

I confirm that I am not aware of any conflict of interest involved in writing of this article in my present capacity as a clinician in the KK Hospital, Singapore.

I am currently the Chairman, Diabetic Society of Singapore and Chairman, Ministry of Health Advisory Committee on Adolescent Health, Singapore.

I was previously the Member of Parliament, Sembawang Group Representation Constituency, 10th Parliament of Singapore, 2001–2006 and in that capacity was also concurrently Vice Chairman of the Government Parliamentary Committee on Community Development, Youth and Sports.

I have been on the speakers panels and have received travel support from the following companies: Novo Nordisk, Eli Lilly, Serono, MiniMed Medtronics, Lifescan J & J, Sanofi Aventis, GSK, Biorad, MSD and Bayer and I was and Bayer a member of the Executive Board of the International Diabetes Federation from 1997–2006 as well as the Advisory Council of ISPAD from 2002–2003.