Intended for healthcare professionals

Clinical Review

Obesity in children. Part 1: Epidemiology, measurement, risk factors, and screening

BMJ 2008; 337 doi: (Published 15 October 2008) Cite this as: BMJ 2008;337:a1824
  1. Ruth R Kipping, research fellow1,
  2. Russell Jago, senior lecturer2,
  3. Debbie A Lawlor, professor of epidemiology13
  1. 1Department of Social Medicine, University of Bristol, Bristol BS8 2PS
  2. 2Department of Exercise, Nutrition and Health Sciences, University of Bristol, Bristol BS8 1TP
  3. 3MRC Centre for Causal Analysis in Translational Epidemiology, University of Bristol, Bristol BS8 2BN
  1. Correspondence to: R Kipping ruth.kipping{at}

    Summary points

    • Population changes in physical activity and diet are probably the main drivers of the obesity epidemic

    • A complex interplay of genetics; epigenetics; and intrauterine, infancy, childhood, and family non-genetic factors may also be involved

    • Obesity in children and adolescents is associated with metabolic and cardiovascular abnormalities and other adverse health outcomes

    • Modifiable risk factors for childhood obesity are maternal gestational diabetes; high levels of television viewing; low levels of physical activity; parents’ inactivity; and high consumption of dietary fat, carbohydrate, and sweetened drinks

    • Obesity is commonly measured in children by plotting body mass index on a standard growth chart to adjust for sex and age using a defined cut-off point

    • Population screening for childhood obesity is not recommended

    Obesity was first included in the international classification of diseases in 1948. Since then, an epidemic has developed internationally, affecting all age groups. This article describes the prevalence of obesity in children, its underlying risk factors, its consequences, and how it can be measured; it also discusses whether children should be screened for obesity. In a second article to be published next week we will discuss the prevention and management of obesity in children.1 The terms used are defined in box 1 (where authors of cited papers use terms differently we use the authors’ own words).

    Sources and selection criteria

    This review draws on the Foresight report, guidance from the National Institute of Health and Clinical Excellence, and a Cochrane review of preventing obesity.

    In April 2008 we searched the Cochrane Library database of reviews and the Centre for Reviews and Dissemination databases using the search term “obesity”. We also conducted a Medline search ((Child$ or paediatric or pediatric or adolescent) and (Obes$ or overweight)) limited to 1 January 2005 to 6 May 2008 and “review articles”; this identified 1105 articles. We read the abstracts of these articles and retrieved relevant papers. We also used articles from our own bibliographies collected over the past 10 years.

    The evidence to support the measurement, prevention, and management of obesity in children is still relatively weak, with few randomised controlled trials or systematic reviews.

    Box 1 Definitions of terms used

    Air displacement plethysmography system
    • A measurement of body composition using a unit with two chambers. The subject is in one chamber during testing and the second chamber contains instruments for measuring changes in pressure between the two chambers

    • The amount of adipose tissue present, with excessive accumulation leading to obesity

    Body mass index (BMI)
    • Calculated as weight (kg)/height (m2)

    BMI rebound
    • The period when a child’s BMI is at its lowest, usually at age 5-6, before it starts to increase monotonically

    • Young people aged 2-18

    • Measurement of underwater weight

    Dual energy x ray absorptiometry (DEXA)
    • A scan that uses two x ray energies to measure the amount of x rays absorbed by the bones and tissue

    • The study of heritable changes in genome function that occur without a change in DNA sequence

    • A condition of excess body fat that may harm health

    • When members of a defined population are offered a test to identify who is more likely to be helped than harmed by further tests or treatment

    • The routine collection of data at a population level to be able to describe changes in disease or risk factors by person, place, and time

    What is the prevalence of obesity in children?

    In 2007 it was estimated that globally 22 million children under 5 years were overweight, with more than 75% of overweight and obese children living in low and middle income countries.2 It is not only the scale of childhood obesity that is challenging, but also the speed at which the prevalence has increased. The greatest annual increases in obesity since 1970 in school children have been in North America and Western Europe.3

    Figures 1 and 2 show trends in the prevalence of obesity in England and the United States. The data come from cross sectional surveys and use country specific growth charts. For both countries, data from the latest surveys suggest that the increase in prevalence of obesity is levelling off, but it is too early to know if the peak has been reached or if this levelling is variation around a still ongoing increase.


    Fig 1 Prevalence of overweight and obesity in England among 2-15 year olds. Adapted from Health Survey for England 2006. The overweight category does not include those who are obese. Overweight defined as ≥85th and <95th UK sex specific body mass index (BMI) for age centiles; obese defined as ≥95th UK sex specific BMI for age centiles (1990 reference charts)


    Fig 2 Prevalence of obesity in the US among 2-19 year olds. Adapted from the national health and nutrition examination survey.4 5 Obese defined as ≥95th centile using US 2000 sex specific body mass index (BMI) for age growth charts from the Centers for Disease Control and Prevention

    What are the risk factors for obesity in children?

    Obesity is caused by an imbalance between energy input and energy expenditure. The relative contribution of physical activity, sedentary activity, and diet to the development of obesity in children is unclear, partly because these variables are difficult to measure and the balance of energy is complex.6 7 Other factors—including genetic variation, epigenetics, intrauterine exposures, and assortative mating—can also affect people’s propensity to gain weight and may contribute in some populations to the epidemic (table 1).

    Table 1

     Risk factors associated with the development of obesity in children and type of evidence available

    View this table:

    A large number of studies have reported associations between a wide variety of risk factors and overweight or obesity in children.8 9 10 However, association does not prove causation, and confounding or reverse causality could explain these associations. Randomised controlled trials and prospective studies provide the most robust evidence for identifying causal associations,6 but for obesity relatively little evidence comes from randomised controlled trials.

    Table 1 summarises the evidence for risk factors associated with childhood obesity. Genetic variants, ethnicity, parental adiposity, birth weight, timing or rate of maturation, levels of physical activity and sedentary activity, and consumption of energy dense food are all associated with childhood obesity.8 9 10 Some risk factors are common (prevalence >10%) but not modifiable, such as the combined effect of multiple genetic variants. Other risk factors are both common and modifiable. These include high birth weight (which is modifiable in cases of maternal diabetes mellitus); high levels of television viewing; low levels of physical activity; parents’ inactivity; and high consumption of dietary fat, carbohydrate, and sweetened drinks. Clearly these need to be the focus of prevention programmes.

    How is adiposity measured in children?

    Many different methods are currently used to measure adiposity in children. Direct measures, such as densitometry and scanning using dual energy x ray absorptiometry,11 are more accurate than indirect measures but are not practical for population level surveillance or clinical management. Anthropometric methods, such as measurement of waist circumference, skinfolds, and BMI, measure adiposity indirectly. Bioelectrical impedance and air displacement plethysmography (for example, BodPod and PeaPod) provide reliable indirect estimates of adiposity and are often used in research studies.12 13

    During adolescence, skinfold thickness is a better predictor than BMI of high adiposity in adulthood.14 Skinfold thickness has also been associated with clustered risk of cardiovascular disease during adolescence,15 but to our knowledge no study has found a relation between skinfold thickness in childhood and future risk of cardiovascular disease. Although BMI is an indirect measure it is practical, easy to obtain, reliable, and—to our knowledge—the only measure of adiposity in childhood that has been shown to be associated with future risk of mortality from cardiovascular disease in adulthood.16

    Who should be classified as overweight or obese?

    Internationally agreed thresholds exist for defining underweight, normal weight, overweight, and obesity in adults, but in children the effects of age, sex, and pubertal status make simple classification difficult. The assessment of obesity in children relies on plotting BMI on a standard growth chart and then defining a cut-off point for increased BMI relative to age and sex. International comparisons are difficult because countries tend to use their own standard growth charts and different cut-off points. Commonly used cut-off points for overweight and obesity include 110% or 120% of ideal weight for height; weight for height z scores (box 2) of >1 and >2; and BMI at the 85th, 90th, 95th, and 97th centiles (based on international or country specific reference populations).3 The International Obesity Taskforce has therefore developed international BMI cut-off points for overweight and obesity according to sex and age (between 2 and 18 years). The cut-off points are defined to pass through a BMI of 25 and 30 at age 18, and are based on BMI data from six countries, which enables comparison of prevalence globally.17

    Box 2 Z scores

    The z score is the distance from the mean in units of standard deviations. A positive z score indicates that the value is above the mean and a negative z score indicates that the value is below the mean. Z scores are often standardised by age and sometimes by sex. An age and sex standardised z score of 1.5 indicates that the value is 1.5 standard deviations above the mean for all children of the same age and sex in the population used to standardise the score.

    In the UK the 1990 UK reference chart for boys and girls is used to measure BMI in children of all ages. However, the Scientific Advisory Committee on Nutrition and the Royal College of Paediatrics and Child Health recommend that the 2006 World Health Organization child growth standards for infants and children are used from 2 weeks of age to 24 months and that the 1990 UK reference charts should be used only from age 2.18 The WHO growth standards are based on babies who are exclusively or predominantly breast fed for at least four months. The WHO and UK charts begin to converge at 2 years of age, but a difference exists between the two charts at 2 years that is greater than the natural disjunction that arises because of the transition from measuring supine length to measuring height. Pilot work is currently under way to identify the most appropriate method of moving between these two charts before the new recommendations of the Scientific Advisory Committee are adopted in the UK.

    The National Institute for Health and Clinical Excellence (NICE) recommends tailored clinical intervention if a child’s BMI (adjusted for age and sex) is at the 91st centile or above and that assessment for comorbidities should be considered if their BMI is at the 98th centile or above, using 1990 UK reference charts. The 85th and 95th centiles are used to define overweight and obesity for surveillance purposes in the UK and the US, although until recently in the US children at or above the 95th centile were termed “overweight.”

    In adults, many of the health risks associated with obesity have been related to excess abdominal fat, particularly visceral fat, and in children markers of central adiposity such as waist circumference are associated with higher fasting insulin and glucose concentrations and adverse lipid profiles in cross sectional studies. Measurement of waist circumference may help identify children at risk of excess centrally located weight although, as with BMI, the correct cut-off point for defining central obesity in children is unclear. Sex specific waist centile curves for UK children aged 5-16 years have been published.19 The International Diabetes Federation has recently suggested criteria for defining the metabolic syndrome in children and recommends a cut-off point of ≥90th centile of waist circumference for age, sex, and ethnic origin in children aged 6 and above for defining central adiposity.20 However, no studies have shown an association between central adiposity or the metabolic syndrome in children and future risk of cardiovascular disease.

    What are the consequences of obesity in children?

    Obesity in children and adolescents is associated with a range of adverse metabolic and cardiovascular traits,21 exacerbation of asthma,22 poor self esteem,23 and an increased likelihood of being obese in adulthood.24 25 26 Table 2 shows the estimated prevalence of obesity associated metabolic and vascular characteristics in European children. However, the evidence for the consequences of childhood obesity comes from observational, often cross sectional, studies so these associations do not necessarily mean causation. Furthermore, associations found in clinically obese populations may not reflect associations in the general population, even at the same level of adiposity. It has been reported that obese children in community samples have better quality of life and self esteem than obese children from clinical samples.28 The prevalence of raised liver enzymes, indicative of non-alcoholic fatty liver disease, is greater in obese children identified in specialist clinics than in those with similarly defined obesity who come from general population samples.29 In a large general population sample of children, higher BMI was prospectively associated with increased risk of mortality from cardiovascular disease.16 However, the extent to which this is the result of tracking BMI from childhood to adulthood rather than permanent changes in metabolic and vascular characteristics in childhood is unclear.

    Table 2

     Minimum estimated numbers of children in European Union in 2006 with obesity related disease indicators. Age range 5-17.9, unless shown otherwise27

    View this table:

    Should children be screened for obesity?

    Certain criteria need to be met before a population screening programme is introduced. For example, the clinical course of the condition should be known, screening should be acceptable and valid, an effective intervention should be available for screen positive individuals, and evidence of the effectiveness of a screening programme to reduce mortality or morbidity should be available from randomised controlled trials. A systematic review of the evidence to screen primary school children concluded that most of these criteria are not met and that it is difficult to justify population screening of children for obesity.30 In line with this, the UK National Screening Committee’s policy is that, currently, not enough evidence is available to recommend screening children for obesity.

    Should there be population surveillance of childhood obesity?

    NICE reviewed evidence from eight longitudinal observational studies in children and concluded that opportunistic monitoring of growth charts after 2 years of age may be beneficial.31 In 2005, an annual National Child Measurement Programme was introduced in England for surveillance (not screening) of two school year groups. In 2006-7 the programme measured 80% of children in state schools and the prevalence of obesity was 9.9% for children in reception (age 4-5) and 17.5% for children in year 6 (age 10-11) using the 95th centile on the UK 1990 reference charts.32 The English government has introduced legislation to give parents the results of the weight and height measurements from this surveillance programme and to allow the measurements to be used by government organisations for performance management of obesity from 2008-9. The programme is still defined as surveillance, but it treads a fine line between surveillance, screening, and performance. A qualitative study commissioned by the Department of Health in England recently found that parents would like a medical interpretation of whether their child is a healthy weight and find BMI a difficult concept to understand.33

    In the US, Arkansas was the first state to pass legislation in 2003 for mandatory BMI assessments of children in public schools with annual reporting to parents. This approach has been followed in six other states.34 To date, the effect of surveillance or screening programmes on the childhood obesity epidemic has not been evaluated.


    Childhood obesity is associated with adverse health outcomes. There is a need to adopt consistent methods of measuring BMI internationally and a need for investment in research to further understand modifiable risk factors and consequences.

    Questions for future research

    • What are children’s, parents’, and health professionals’ attitudes to monitoring height and weight or screening for obesity?

    • What are the benefits and harms associated with population wide screening for childhood obesity?

    • What is the effect of contemporary patterns of physical activity and diet on the risk of obesity in childhood persisting into adulthood? Longitudinal studies that accurately measure adiposity, diet, and activity (using accelerometry) are needed

    • What are the truly causal modifiable risk factors for childhood obesity? More randomised controlled trials are needed to look at the effect of modifying risk factors in childhood obesity

    • What are the long term effects of childhood obesity and changes in weight in obese or overweight children? Large prospective studies are needed that repeatedly measure adiposity, metabolic outcomes, vascular outcomes, and other health outcomes into adulthood

    • At what body mass index for age does childhood obesity lead to important adverse health consequences?

    Tips for non-specialists

    • Children can be assessed using body mass index (BMI) plotted on a standard growth reference chart for the child’s age and sex to give BMI for age and sex. Cut-off points for overweight and obesity specific to the chart should be used

    • The key modifiable and common (>10% prevalence) risk factors for childhood obesity include high birth weight (which is modifiable in cases of maternal diabetes mellitus); high levels of television viewing; low levels of physical activity; parents’ inactivity; and consumption of dietary fat, carbohydrate, and sweetened drinks

    Additional educational resources

    Resources for healthcare professionals
    Resources for parents and patients


    Cite this as: BMJ 2008;337:a1824


    • References numbered w1-w34 are on, labelled as extra

    • Contributors: All authors discussed the scope and approach to the review. RRK drafted the review. All authors helped write the paper. All authors are guarantors.

    • Competing interests: None declared.

    • Provenance and peer review: Commissioned; externally peer reviewed.