Intended for healthcare professionals

Practice ABC of obesity

Childhood obesity

BMJ 2006; 333 doi: (Published 07 December 2006) Cite this as: BMJ 2006;333:1207
  1. John J Reilly, professor of paediatric energy metabolism1,
  2. David Wilson, senior lecturer2
  1. 1University Division of Developmental Medicine, Yorkhill Hospitals, Glasgow
  2. 2Division of Child Life and Health, University of Edinburgh, Royal Hospital for Sick Children, Edinburgh

    This article summarises recent evidence based clinical guidelines and systematic reviews and proposes strategies for managing paediatric obesity

    Obesity, an excessive body fat content with increased risk of morbidity, has become increasingly common in children and adolescents. Confusion exists, however, over basic questions such as whether paediatric obesity matters, how to diagnose it, and whether it should be treated (and if so, how best to do this). Doctors in many fields need a better understanding of these issues.


    Simplified diagram of some body mass index curves seen on UK BMI charts for boys (that for girls is similar). The charts are available from Harlow Health via the Child Growth Foundation (


    This is the 11th article in the series

    Subjective assessment methods are inaccurate, so diagnosis must be objective. A substantial and consistent body of high quality evidence has shown that body mass index (BMI; weight in kilograms divided by height in metres squared) can be used to diagnose obesity effectively.

    BMI is lower in childhood and adolescence than in adulthood and differs between boys and girls. Clinically, no diagnostic alternative exists, therefore, to considering BMI in an age and sex specific manner, either by plotting it on a BMI chart or by referring to tables of BMI for age and sex.

    Evidence based definitions of paediatric obesity

    For research and epidemiological purposes
    • Overweight should be defined as BMI ≥85th centile for sex and age

    • Obesity should be defined as BMI ≥95th centile for sex and age

    For clinical purposes in UK
    • Overweight should be defined as BMI ≥91st centile if using the UK BMI charts

    • Obesity should be defined as BMI ≥98th centile if using the UK BMI charts

    In the United Kingdom, BMI for age charts represent the distribution of BMI in children in 1990. Diagnosing obesity as a high BMI centile (such as the 98th centile or above on the UK charts) identifies the fattest children in the population with high specificity (low false positive rate). High specificity provides confidence that a child diagnosed as obese in this way actually is excessively fat, and not simply muscular. In addition, strong evidence shows that children with a high BMI centile are at high risk of comorbid conditions, so the diagnosis is clinically meaningful.

    For comparisons of obesity prevalence between nations, “international” definitions of overweight and obesity based on BMI for age are available. These are designed to provide definitions that are conceptually equivalent to adult BMI values of 25 and 30 respectively.

    Despite high specificity for excess fatness, a high BMI for age has only modest sensitivity in children and adolescents (moderate false negative rate). Modest sensitivity is a problem for public health applications such as surveillance of obesity, as estimates of prevalence of obesity based on BMI tend to be conservative.


    Development of epidemic of obesity in English children during the 1990s. Adapted from Reilly et al (Lancet 1999;354:1874-5)

    Future research may provide useful diagnostic alternatives to BMI, and the use of waist circumference in the paediatric age range is promising, though not yet sufficiently evidence based.


    In the UK the epidemic of paediatric obesity began in the late 1980s. Prevalence has continued to increase rapidly, and obesity is now the most common disorder of childhood and adolescence. The health survey for England 2004 showed that 14% of 2-11 year olds and 25% of 11-15 year olds were obese (BMI ≥95th centile).

    In the United States, prevalence of childhood obesity is highest in some ethnic minority groups; preliminary evidence suggests that this might also be the case in Europe, although further research is necessary and will require oversampling of children and adolescents from ethnic minority groups.

    Obesity is more common in older children and adolescents than in younger children; in the developed world it is slightly more common in children from less affluent families, for reasons that are unclear

    Excess obesity is a particularly important clinical issue in certain ethnic groups, such as South Asians. South Asians are more sensitive to the adverse effects of excess weight gain and account for most childhood cases of type 2 diabetes in the UK.

    Consequences of childhood obesity

    Consequences of paediatric obesity

    Short term (for the child or adolescent)
    • Psychological ill health

    • Cardiovascular risk factors

    • Asthma

    • Chronic inflammation

    • Diabetes (type 1 and 2)

    • Orthopaedic abnormalities

    • Liver disease

    Long term (for the adult who was obese as a child or adolescent)
    • Persistence of obesity

    • Adverse socioeconomic outcomes, particularly in women

    • Cardiovascular risk factors, diabetes, cancers, depression, arthritis

    • Premature mortality

    A systematic review and critical appraisal recently concluded that paediatric obesity was a major health burden not only in childhood and adolescence but also later on, in adulthood.

    Short term

    Many obese patients seeking medical care experience teasing and bullying and will have low self esteem or a poor quality of life. Paediatric obesity is also associated with many other comorbid conditions. Evidence on some important comorbidities (such as liver disease) and on the adverse economic effects of paediatric obesity has only emerged recently, and the list of adverse consequences will lengthen in the near future.


    Obesity in adolescence, particularly in girls, is associated with impaired adult socioeconomic outcomes such as poorer educational results and lower income

    From a public health viewpoint the most serious consequence of obesity in childhood may be its damaging long term effects on individuals' cardiovascular health. Cardiovascular risk factors are common, particularly in obese adolescents, and tend to “cluster” (aggregate) in obese individuals. For example, in a non-clinical sample of adolescents in the United States, 29% of those who were obese had the metabolic syndrome compared with 0.1% of those who were not overweight.

    Long term

    “Planet Health” trial*

    • Was conducted over two school years; 1295 participants aged 11 years at baseline

    • Aimed to reduce television viewing, increase lifestyle physical activity, reduce fat intake, and increase fruit and vegetable intake

    • Intervention was tested rigorously in a high quality randomised trial

    • Intervention is probably generalisable

    • Assessment of economic impact was carried out

    • School based changes were sustainable and cost effective

    • Benefits to intervention were found in girls (reduced risk of becoming obese, remission of existing obesity)

    • Benefits were attributed to reductions in television viewing

    • *Trial of model programme for school based obesity prevention (further details from Gortmaker et al. Arch Pediatr Adolesc Med 1999;151:409-15)

    Obesity tends to persist, particularly from adolescence into adulthood. Children are least likely to “grow out of” obesity when it is more severe and when they have at least one obese parent, and both of these features are much more common now than in the past. These days, at least 60% of obese children and 70-80% of obese adolescents are likely to become obese adults.


    The evidence on interventions for prevention has recently been reviewed systematically and critically appraised. Few randomised controlled trials have been conducted. Most trials were found to be methodologically weak and short term, and most of the interventions tested were unlikely to be generalisable.

    Guide to prevention interventions*

    Interventions should:

    • Do no harm

    • Lead to benefitting child health or development in other ways

    • Target behaviours that are causally related to the development of obesity or to its maintenance

    • Target behaviours that are modifiable

    • Target behaviours that are measurable because (a) families would otherwise find it difficult to make changes and (b) scientific evaluation requires measurement

    • *Modified and expanded from Whitaker (Arch Pediatr Adolesc Med 2003;157:725-7)

    Few trials found benefits linked to the intervention being tested. One notable exception, however, is the “Planet Health” intervention in US schoolchildren, which was of high methodological quality, reasonably long term, generalisable, and successful (at least in girls). The success of this particular intervention was attributed largely to reductions in television viewing.

    Further research on preventive interventions is needed urgently. A helpful guide to preventive interventions was proposed by Whitaker. One barrier to such interventions is the common perception among parents, teachers, and health professionals that the interventions may cause harm (for example, by promoting eating disorders). Such adverse effects are extremely unlikely.

    Useful tips when treating paediatric obesity

    • Treat as “intensively” as possible—more frequent and longer appointments are beneficial

    • Treat the whole family, not just the child

    • Treat only motivated families

    • Aim for dietary changes, perhaps using a “traffic light” approach—that is, greatly restrict foods high in energy (“red”); restrict foods with moderate energy content (“amber”) to meal times only; eat freely foods that are low in energy (“green”) and substitute them for red foods


    Recent systematic reviews and critical appraisal exercises have consistently concluded that the evidence on interventions to treat paediatric obesity is extremely limited. Most of the randomised controlled trials have been methodologically weak and short term, and many have been unsuccessful. Most trials tested interventions that are unlikely to be generalisable. Systematic reviews have identified promising elements of treatment from the literature.

    The possibility of causing harm is also a perceived by parents, teachers, and health professionals as a barrier to treatment, but the evidence suggests that this is unlikely. In fact, with some treatment interventions the psychosocial wellbeing of those being treated actually improved

    No high quality evidence is currently available on the medium to long term effects of surgical treatment, drug treatment, or residential treatment. Longer term randomised controlled trials of these forms of treatment are needed urgently.

    Audits of typical dietetic and paediatric treatment of obesity in the US and in the UK usually report disappointing results. A recently conducted five year audit of an obesity clinic at the Royal Hospital for Sick Children in Edinburgh found that over half of the patients who had been referred to the clinic did not attend any of their appointments. Of the patients who attended at least one appointment, only 22% maintained their weight over six months.

    Reports on management from expert committees

    • Management of obesity in children and young people. Guideline No 69 from SIGN (Scottish Intercollegiate Guidelines Network), 2003. (accessed 18 Oct)

    • Summerbell CD, Waters E, Edmunds LD, Kelly S, Brown T, Campbell KJ. Interventions for preventing obesity in children. Cochrane Database Syst Rev 2005;(3):CD001871.

    • Summerbell CD, Ashton V, Campbell KJ, Edmunds L, Kelly S, Waters E. Interventions for treating obesity in children. Cochrane Database Syst Rev 2005;(3):CD001872.

    • Canadian review of reviews 2004.

    • Clinical practice guidelines for the management of overweight and obesity in children and adolescents. Australian review and evidence based guidelines.

    • P Gibson, L Edmunds, DW Haslam, E Poskitt. An approach to weight management in children and adolescents in primary care. J Fam Health Care 2002;12:108-9.

    • SE Barlow, WH Dietz. Obesity evaluation and treatment: expert committee recommendations. Pediatrics 1998;102:E29.


    To know how to manage childhood obesity we need to know who should be treated, who should be referred, and what the treatment should aim for.

    Systematic reviews and critical appraisals have concluded that the evidence base (from reports from expert committees) for providing answers to these questions is weak. Nevertheless, these reports are likely to be extremely helpful in management.

    Summary of reports from expert committees

    Who should be treated?

    As treatment of obesity requires long term adherence to lifestyle changes, success is unlikely if patients and/or their family do not perceive obesity as a problem or if they are poorly motivated to make lifestyle changes.


    Time spent playing computer games may need to be reduced if an obese child is to achieve and maintain weight loss

    The disappointing results of treatment from past audits partly reflect a lack of understanding among families that paediatric obesity matters. Treatment should be reserved for families who perceive obesity as a problem and who show motivation to make and sustain lifestyle changes.

    Who should be referred?

    Referral from primary to secondary care would be justified for two reasons: to investigate possible underlying pathological causes of the patient's obesity; and to investigate or manage a possible comorbid condition such as type 2 diabetes.

    In the vast majority of patients obesity is caused by lifestyle; pathological causes are extremely rare. An underlying pathological cause should be suspected, however, if obesity is particularly severe in young children (where it may reflect an underlying genetic cause such as a single gene defect) or if it coexists with short stature (which may indicate a syndromic cause such as the Prader-Willi syndrome or some other endocrine cause).

    Families can measure television viewing and related sedentary behaviours; if they reduce their viewing, this may result in them also reducing their food intake and increasing their physical activity

    If referral to secondary care reveals no comorbidities requiring urgent treatment and no underlying pathological causes of obesity, patients could be discharged to primary care for treatment. However, in many parts of the UK treatment of childhood obesity is either limited or not offered at all in primary or secondary care.

    Aims of treatment

    • To resolve comorbidity if it is present

    • To achieve weight maintenance, not weight loss

    • To ensure that families monitor their own diet, activity, television viewing, and computer use

    • To introduce dietary changes

    • To reduce sedentary behaviour (particularly television viewing)—to less than two hours a day

    • To increase physical activity through lifestyle changes such as walking to and from school

    What should treatment aim for?

    Weight maintenance over six to 12 months is considered to be a success, particularly as continued excessive weight gain is common in the absence of treatment. Weight maintenance is also successful because, with growth, paediatric patients can to some extent “grow into their weight.”

    Weight loss is an unrealistic aim except if patients have achieved prolonged weight maintenance or are severely obese and so have to lose weight to enable management of comorbidity such as type 2 diabetes or sleep apnoea. Progress with management should be reviewed with the family using the BMI centile chart periodically.

    Further reading

    • Neovius M, Linne Y, Barkeling B, Rossner S. Discrepancies between classification systems of childhood obesity. Obes Rev 2004;52:105-14.

    • Rudolf MCJ. The obese child. Arch Dis Child Educ Pract 2004;89:ep57-62. (

    • Viner R, Nicholls D. Managing obesity in secondary care. Arch Dis Child 2005;90:385-90.

    • Stewart L, Houghton J, Hughes AR, Pearson D, Reilly JJ. Dietetic management of pediatric overweight: development and description of an evidence-based, behavioral, approach. J Am Diet Assoc 2005;105:1810-5.

    • Kopelman PG, Caterson ID, Dietz WH, eds. Clinical obesity in adults and children. 2nd ed. Malden, MA: Blackwell, 2005.

    The ABC series—now published by Blackwell Publishing

    From January 2007, ABC articles will no longer be serialised in the BMJ, but Blackwell Publishing is delighted to continue offering the successful ABC book series. With over 40 titles, this series provides a dependable reference on a range of topics across the major specialties. For further details, please visit


    • The photographs of adolescent girls and of a child playing a computer game are published with permission from Bubbles and Gusto/SPL respectively.

    • John J Reilly is professor of paediatric energy metabolism in the University Division of Developmental Medicine, Yorkhill Hospitals, Glasgow, and David Wilson is senior lecturer in the Division of Child Life and Health, University of Edinburgh, Royal Hospital for Sick Children, Edinburgh.

      The ABC of Obesity is edited by Naveed Sattar (, professor of metabolic medicine, and Mike Lean, professor of nutrition, University of Glasgow. The series will be published as a book by Blackwell Publishing in early 2007.

      Competing interests: For series editors' competing interests, see the first article in this series.

    View Abstract