ENDORSED 18 SEPTEMBER 2003Clinical Practice Guidelines for the Managementof Overweight and Obesity in Children and Adolescents Commonwealth of Australia 2003Paper-based publicationsThis work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be reproduced by any process without prior written permission from the Commonwealth available from the Department of Communications, Information Technology and the Arts. Requests and inquiries concerning reproduction and rights should be addressed to the Commonwealth Copyright Administration, Intellectual Property Branch, Department of Communications, Information Technology and the Arts, GPO Box 2154, Canberra ACT 2601 or posted at http://www.dcita.gov.au/cca. Commonwealth of Australia 2003Electronic documentsThis work is copyright. You may download, display, print and reproduce this material in unaltered form only (retaining this notice) for your personal, non-commercial use or use within your organisation. Apart from any use as permitted under the Copyright Act 1968, all other rights are reserved. Requests for further authorisation should be directed to the Commonwealth Copyright Administration, Intellectual Property Branch, Department of Communications, Information Technology and the Arts, GPO Box 2154, CanberraACT 2601, or posted at http://www.dcita.gov.au/cca.ISBN Print: 1 864961 90 2 ISBN Online: 1 864961 96 1DisclaimerThis document is a general guide to appropriate practice, to be followed only subject to the clinicians judgement in each individual case.The guidelines are designed to provide information to assist decision-making and are based on the best information available at the date of compilation.It is planned to review this Guideline in 2006.For further information regarding the status of this document, please refer to the NHMRC web address: http://www.nhmrc.gov.auFor copies of this document contactPhone: 1800 020 103 extension 8654 (toll free number)Email: phd.publications@health.gov.auWebsite: www.obesityguidelines.gov.auCONTENTSPreface vSummary viiEvidence-based statements and recommendations xi1 The prevalence and tracking of overweight and obesity in children and adolescents 11.1 Prevalence 11.2Tracking 32 Defnition and measurement of overweight and obesity in children and adolescents 72.1Introduction72.2Measures of adiposity72.3Measurements as outcome indicators in weight-management183 Risk factors 193.1Genes 193.2Television viewing213.3Energy expenditure 233.4Artifcial infant feeding303.5Dietary intake, including fat, dietary carbohydrate and eating patterns323.6Single-gene defects and obesity syndromes 363.7Ethnicity 393.8Early adiposity rebound 403.9Single-child, single-parent, urban versus rural, and socio-economic status443.10Endocrine disease 453.11Central nervous system pathology 463.12Acute lymphatic leukaemia therapy 46 3.13Medications 47CONTENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents iiiCONTENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents ivPREFACEClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents v4 Morbidity and childhood and adolescent obesity 494.1Psychosocial morbidity494.2Cardiovascular morbidity 514.3Insulin resistance and diabetes534.4Other associated medical morbidities in childhood and adolescent obesity554.5Disordered eating 614.6The effect of weight loss on obesity-related morbidity 625 Conventional weight-management strategies 655.1General strategies 655.2Dietary change 705.3Increasing physical activity 735.4Reducing sedentary behaviours 755.5Behaviour change 765.6Parental involvement 785.7Frequency of review 805.8Settings for weight management816 Non-conventional weight-management strategies 856.1Very low energy diets856.2Pharmacotherapy for obesity in childhood and adolescence 886.3Other pharmacotherapeutic agents 916.4Bariatric surgery 93Appendix A The guideline-development methodology95Appendix B References103CONTENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents ivPREFACEClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents vPREFACEIn recent decades the number of Australiansboth adults and childrenwho are overweight or obese has continued to increase. The prevalence of obesity and overweight in children and adolescents has doubled in the last 15 years, and it is estimated that between 20 and 25 per cent of children and adolescents are now overweight or obese. Many of these people will go on to become overweight or obese adults, with all the health costs that this implies. In 1997 the National Health and Medical Research Councils (NHMRC) Expert Panel on Prevention of Obesity and Overweight prepared Acting on Australias Weight: a strategic plan for the prevention of overweight and obesity. The primary goals expressed in that plan were to prevent further weight gain in adults and eventually reduce the proportion of the adult population that is overweight or obese; and to ensure healthy growth of children. During the development of Acting on Australias Weight and the subsequent strategy, the need for clinical practice guidelines for the management of overweight and obesity in Australian adults and children became apparent. In 2000, in collaboration with the Population Health Division of the Commonwealth Department of Health and Ageing, the NHMRC initiated the development of the guidelines.In working on the project, and having determined that separate guidelines were required for adults and for children and adolescents, the NHMRC researched practices for managing overweight and obesity and ensured that the practices identifed were multi-facetedfor example, strategies that encompass physical activity, diet and self-esteem.These guidelines for children and adolescents are the result of a comprehensive assessment of the current scientifc evidence. They provide detailed evidence-based guidance for assessing and managing overweight and obesity in Australia. They have been published separately from the guidelines for adults because the health and psychosocial factors pertaining to the management of overweight and obesity in children and adolescents are very different. Management of these conditions in this age group is expected to become increasingly common as greater numbers of children and adolescents fall into the categories of overweight and obese. The publication of guidelines specifcally for children and adolescents emphasises the importance of appropriate clinical practice in this area. The guidelines themselves highlight important health concerns associated with overweight and obesity in childhood and adolescence, as well as the associated future health risks. The guidelines focus primarily on the majority population in Australia. It must be recognised that the problem of overweight and obesity among specifc groups, and Aboriginal and Torres Strait Islander children and adolescents in particular, has distinct characteristics that are currently less well understood and need urgent, detailed examination. PREFACEClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents viSUMMARYClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents viiThe guidelines are designed for use by general practitioners and allied health professionals when providing advice to patients in the clinical setting. Information for consumers is also being developed. It is stressed that the guidelines are for clinical practice. They do not represent a comprehensive population-based approach to overweight and obesity in children and adolescents: that was the task of Acting on Australias Weight, and it will be addressed again in future NHMRC publications.It is recommended that these guidelines be updated and revised by 2006.PREFACEClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents viSUMMARYClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents viiSUMMARYTHEPREVALENCE AND TRACKI NGOFOVERWEI GHT ANDOBES I TYI NCHI LDREN AND ADOLES CENTSThere is evidence that the prevalence of overweight and obesity in children and adolescents in Australia has increased in the last 15 years: an estimated 20 to 25 per cent of children and adolescents are now overweight or obese. The trend is similar in other developed and developing countries.Epidemiological studies provide evidence that relative body weight tracks from childhood to adulthood and that the predictive power of this association increases with age. Not all overweight and obese adults were overweight or obese as children. However, once a child or adolescent is on an overweight or obese percentile, it is unlikely that they will revert spontaneously to a lower weight percentile.The prevalence of overweight and obesity in children and adolescents is high enough to warrant both intervention and preventive action.DEFI NI TI ON ANDMEAS UREMENTOFOVERWEI GHT ANDOBES I TYI NCHI LDREN AND ADOLES CENTSMany defnitions of overweight and obesity in children and adolescents are used in the literature and in clinical practice. Unlike the situation with adults, though, these defnitions are not based on morbidity and are thus more arbitrary.There is general consensus that an age-related body mass index (BMI) should be used, since BMI is signifcantly associated with body fatness in children and adolescents. BMI has also been validated against more direct measures of adiposity.An Australian reference standard based on adult BMI cut-off points has been developed for children and adolescents, but it is not considered suitable for clinical use. An individuals BMI can, however, be compared with a BMI-for-age centile chart. When this method is used, a child can be followed over time with serial measurement. For children and adolescents in the clinical setting, the Centers for Disease Control and Prevention BMI percentiles are recommended, with a BMI above the 85th percentile suggesting overweight and a BMI above the 95th percentile suggesting obesity.Australia has no growth reference charts and at present would have to use an alternative reference on which to base aged-related BMI defnitions of overweight and obesity in children and adolescents.viiiThere is a growing body of evidence that waist circumference can be used to assess cardiovascular risk in children and adolescents and to assess response to weight management.RI S KFACTORSA number of risk factorssome of them relatively easy to modify and some nothave been identifed as being associated with child and adolescent overweight and obesity: There is a signifcant genetic predisposition to obesity. Parental obesity is a strong risk factor for future, if not present, obesity. American studies show a positive correlation between television viewing and overweight. There is no evidence available for other forms of small-screen entertainment. Television viewing is a highly modifable risk factor. Reduced physical activity energy expenditure may play a role in weight gain over time in children and adolescents and is another modifable risk factor. The role of diet composition in the development of overweight and obesity in children and adolescents is unclear. There is, however, evidence for breastfeedings protective effect against obesity and evidence that disordered eating in a parent may be associated with excess body weight in the child. A number of single-gene abnormalities have been described in which obesity is the predominant fnding, but these contribute little to the total number of obese children and adolescents. Ethnicity, birthweight and early adiposity rebound are all risk factors for obesity in children and adolescents, but they are not highly modifable. Certain well-characterised endocrine disorders, hypothalamic damage, treatment for acute lymphatic leukaemia, and the use of certain pharmacotherapeutic agents are non-genetic obesity risks, but they contribute little to the total number of obese children and adolescents.Socio-economic status and urban living are not identifed as special risk factors for obesity in children and adolescents in Australia.MORBI DI TY ANDCHI LDHOOD AND ADOLES CENT OBES I TYCardiovascular risk factors and obesity cluster in childhood and track through to adolescence. Cardiovascular risk factors in parents increase the risk of such factors being present in their offspring.Childhood and adolescent obesity that persists into adulthood confers increased adult morbidity and mortality risks. Childhood and adolescent obesity is, however, also associated with morbidity in childhood.SUMMARYClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsThere is evidence from descriptive studies that obese children and adolescents suffer a higher level of physical discomfort than their non-obese peers.The prevalence of type 2 diabetes is increasing in children and adolescents from certain ethnic groups (including those of Aboriginal and Torres Strait Islander and Middle Eastern backgrounds); the increase appears related to the high prevalence of obesity in these populations.Obese children and adolescents suffer from an increase in other medical morbidities that impair their current health. Among these morbidities are obstructive sleep apnoea, hepatic steatosis, slipped capital femoral epiphyses, and the polycystic ovarian syndrome.Obese children and adolescents exhibit impaired psychosocial function; the impairment is greater in females and with increasing age.A signifcant proportion of children and adolescents use unhealthy dietary practices for weight control; these practices are more common in those who are overweight and female.CONVENTI ONAL WEI GHT- MANAGEMENTS TRATEGI ESConventional weight-management strategies have not been studied as extensively in children and adolescents as they have in adults. The conventional strategies are a reduction in energy intake, by dietary means and using conventional food items; an increase in energy expenditure, by increasing physical activity and decreasing sedentary behaviours; behaviour modifcation; and family involvement in the process of change.Published results of weight-management programs using conventional therapies show modest success in children and adolescents in the medium to long term. There is some evidence that children and adolescents maintain weight loss better than their parents. Most of these studies are limited to a few investigating groups, however, and there is a need for the studies to be reproduced in other settings. There is some limited evidence that similar weight-loss outcomes can be achieved in a number of different settings and using different types of programs.There is no direct evidence on either optimal dietary prescription or behaviour-modifcation strategies in the management of obesity in children and adolescents. There is some limited evidence that increasing physical activity or reducing sedentary behaviours improves weight-loss outcomes in children and adolescents.For obese children and adolescents, weight-management programs that involve parents have better outcomes than programs that do not. There is also evidence, in children of primary school age, that a program that involves parents alone does better than one that requires regular attendance by their children as well.SUMMARYixClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsSUMMARYClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xiNON- CONVENTI ONAL WEI GHT- MANAGEMENT S TRATEGI ESFor extreme degrees of obesity and associated co-morbidity, particularly in adolescents, it may be necessary to consider weight-management strategies additional to those therapies that are deemed conventional. Most of these non-conventional therapies should, however, be attempted only in tertiary institutions and within a specialist, multi-disciplinary team. Very low energy diets have been shown to produce rapid weight loss in adolescents in a short period, but it is not so clear that they provide any long-term weight-loss beneft.There is no evidence for the use of sibutramine and very limited evidence for the use of orlistat in adolescents. Nevertheless, there are enough adult trial data to warrant consideration of using these medications in obese adolescents with obesity-related co-morbidity.There is limited evidence that gastric bypass or gastric restrictive surgery in obese adolescents induces a weight loss comparable to that shown in adult studies. There are, however, no established criteria for determining which subjects would beneft from such a procedure. Targeted therapy has potential for single gene-defect obesity, as exemplifed by the effcacy of leptin-replacement therapy in leptin defciency. SUMMARYClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xiEVI DENCE- BASEDSTATEMENTS AND RECOMMENDATI ONSThe information in this publicatin is summarised in two formats, as shown in Appendix A.The evidence-based statements are founded on the levels of evidence for clinical interventions set by the NHMRC. The grades of redommendation are less formally determined, being based on previous guidelines.THEPREVALENCE AND TRACKI NGOFOVERWEI GHT ANDOBES I TYI NCHI LDREN AND ADOLES CENTSThe number of overweight and obese children and adolescents in Australia has increased in the last 15 years. An estimated 20 to 25 per cent of children and adolescents are now overweight or obese. Children and adolescents who are overweight or obese tend to stay overweight or obese. Evidence-based statementEvidence levelA very signifcant percentage of children and adolescents in III-2Australia are overweight or obese. The prevalence of the condition has increased signifcantly in the last 10 years. The trend is similar in the majority of other developed and developing countries.Recommendation: level BTheprevalenceofoverweightandobesityinchildrenandadolescents in Australiaishighenoughtowarrantbothinterventionandpreventive action.Research recommendation Heightandweightshouldbemeasuredrecurrentlyinanationally representative sample of children and adolescents. Evidence-based statementEvidence levelRelative body weight tracks from childhood to adulthood, III-2and the predictive power of this association increases with age. Once a child or adolescent is on an overweight or obese percentile, spontaneous track-down is unlikely.Recommendation: level BTracking data support intervention in childhood overweight and obesity.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xiiEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xiiiDEFI NI TI ON ANDMEAS UREMENTOFOVERWEI GHT ANDOBES I TYI NCHI LDREN AND ADOLES CENTSAny defnition of overweight and obesity in children and adolescents will not be based on the risk of illness (morbidity) and is arbitrary. There is general consensus that an age-related BMI should be used. Waist circumference can be used to assess cardiovascular risk in children and adolescents.Evidence-based statementEvidence levelBMI is signifcantly associated with body III-3fatness in childhood and adolescence.Recommendation: level CBMI is a reasonable, easily determined surrogate measure for adiposity in children and adolescents.Research recommendation Studies should be performed to correlate current medical morbidity with BMIinAustralianchildrenandadolescents.Aboriginaland TorresStrait Islander children and adolescents may need to be considered separately. StatementThere is no evidence on which to base a decision about which BMI measure to use.Recommendations: level DBMI should be used as the standard measure of overweight and obesity for 2 to 18 year olds in Australia.The Australian reference standard based on the work of Cole et al. (2000) should be used in population and clinical research.BMI-for-agepercentilechartsshouldbeusedinclinicalpracticeandin non-health care settings. A BMI above the 85th percentile is indicative of overweight and a BMI above the 95th percentile is indicative of obesity. It should be noted that this defnition of overweight and obesity is arbitrary and that a more appropriate defnition is needed but not yet available. The CentersforDiseaseControlandPreventionBMIpercentilechartsare recommended for use until local BMI growth charts are developed.There isaneedforAustraliatodevelopsuchchartsforclinicalpracticevery soon.Research recommendationThereisaneedtodeveloplocalBMIgrowthchartsforusein Australia. Separate BMI growth charts may need to be considered in Aboriginal and Torres Strait Islander people.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xiiEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xiiiEvidence-based statementEvidence levelAnthropometric markers and direct measurement of III-2central fat in children and adolescents positively correlates with cardiovascular risk.Recommendation: level BWaist circumference appears to be the best clinical determinant of truncal obesity, and hence metabolic risk, in children and adolescents and can be used for longitudinal assessment in management.Research recommendationsLocal standards for clinical use for body mass index and waist-circumference cut-off points should be developed.Research aiming to characterise the association between overweight and obesity (and particularly excess abdominal fat) in children and adolescents and adverse health outcomes should be a priority.Research into the appropriate body mass index and waist-circumference cut-offpointsinAboriginalandTorresStraitIslanderchildrenand adolescents is required. RI S KFACTORSA number of risk factors have been identifed as being associated with childhood and adolescent overweight and obesity. Not all are modifable and some affect a very small proportion of children. Parental obesity is a strong risk factor for future, if not present, obesity. It is important to identify children for whom obesity is secondary to an underlying disorder. Not much is known about how dietary fat and carbohydrate intake infuence body weight in children or adolescents. There is a positive correlation between television viewing and overweight in children and adolescents. Lack of physical activity probably plays a role in weight gain over time in children and adolescents and is another modifable risk factor.Evidence-based statementEvidence levelThere is a signifcant genetic predisposition to obesity. III-2Parental obesity is a risk factor for future, if not present, obesity.Recommendation: level BIfonechildisoverweightorobese,determinewhetherotherfamily members are overweight or obese. EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xivEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xvEvidence-based statementEvidence levelThe (mainly American) data on television viewing III-2indicate a positive correlation between hours of viewing and overweight.The correlation is stronger in older children and adolescents and clearer at low or high (less than two or greater than fve) hours of viewing per day.Studies on other forms of small-screen entertainment are awaited.Recommendation: level BTelevision viewing is a highly modifable risk factor and should be assessed.Recommendation: level DTherelationshipbetweenphysicalactivityandbodyfatnessneedstobe better understood in physiological terms, as do the metabolic effects per seofphysicalactivityinchildhood. Theseneedtobetakenintoaccount whenplanningandevaluatingphysicalactivityinterventioninobesity management. StatementsJust as measurement of obesity in children is limited by the lack of immediate morbidity and mortality data, so is the ability to develop physical activity-management guidelines that are based on evidence related to positive health outcomes.Measuring physical activity in the clinical setting is diffcult. Until cheap, small and robust motion monitors become available, it will depend largely on self-reporting.Recommendation: level DAnunderstandingofhowchildrenandadolescentsareactiveshould aidintheprescriptionofactivityforobesitymanagement.Nostudyhas considered this in detail. Evidence-based statementEvidence levelReduced physical activity energy expenditure may play a III-3role in weight gain over time.Recommendation: level CPhysical activity energy expenditure is a modifable risk factor and should be assessed.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xivEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xvEvidence-based statementEvidence levelThe majority of cohort studies support the fnding III-2that breastfeeding plays a small protective role against subsequent overweight.Recommendation: level BBreastfeeding confers a number of advantages and should be promoted as the infant-feeding method of choice.Evidence-based statementsEvidence levelThe evidence that dietary fat intake is a signifcant IVrisk for obesity in children and adolescents is minimal.There is minimal evidence that carbohydrate intake IVinfuences body weight in children and adolescents.There is no clear evidence that any particular dietary IVcomposition infuences overweight or obesity in children or adolescents.Research recommendationsGood-quality studies examining the amount and type of fat consumed in the diets of children and adolescents are needed.Good-qualitystudiesexaminingthedietsofAustralianchildrenand adolescentsinrelationtooverweightareneeded,includingtheamount and type of carbohydrate.Evidence-based statementEvidence levelParents infuence food choices and other eating behaviours III-3in their children. Disordered eating in a parent may be associated with excess body weight in the child.Recommendation: level CUnhealthyparentaleatingstylescaninfuencethechildsweight.Family eating styles should be assessed.Evidence-based statementEvidence levelThere are a number of single-gene abnormalities in III-2which obesity is the predominant feature.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xviEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xviiRecommendation: level BChildren with severe and very early onset obesity should be referred for specialist assessment. Evidence-based statementEvidence levelThere are a number of rare congenital syndromes III-3that have obesity as a component and in which intellectual impairment is a common feature.Recommendation: level CIfachildpresentswithobesityinassociationwithintellectualdisability andmultiplephysicalabnormalities,thechildshouldbeassessedbya paediatrician, an endocrinologist and/or a geneticist.Evidence-based statementEvidence levelInternational and local data suggest that certain III-3ethnic backgrounds entail a higher predisposition to obesity. For Australian children, the evidence suggests that those of Middle Eastern or Mediterranean origin have a higher prevalence of obesity.Recommendation: level CBe aware that children and adolescents from certain ethnic backgrounds may be at greater risk of obesity. Effective management may necessitate an understanding of the particular ethnic groups beliefs about, and attitudes to, weight management.Research recommendationData on the prevalence of overweight and obesity among Aboriginal and Torres Strait Islander children and adolescents are urgently required. Evidence-based statementsEvidence levelThere is evidence from population studies that early III-3adiposity rebound is associated with higher adolescent and adult BMIs.No matter how overweight is defned in the individual III-3study, there is a signifcant association between higher birthweight and higher weights in childhood.Additional risk is conferred by an average, rather than tall, birth lengthand by parental overweight. Small-for-gestational-age babies who exhibit III-3catch-up growth are at risk of obesity in childhood.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xviEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xviiRecommendation: level CBe alert to an upward change in BMI percentileparticularly in association with other obesity risks in children and adolescents who experience early adiposity rebound, who are known to have been large for their gestational age and have overweight parents, or who are known to have been small for their gestational age and exhibit catch-up growth. StatementThere is not enough evidence to determine that single-child or single-parent status or urban living in Australia are signifcant risk factors for overweight.Evidence-based statementsEvidence levelSocio-economic status is not a strong risk for obesity III-3and overweight in children and adolescents in Australia.Clinical observation confrms an association between III-2obesity and a number of endocrine disorders. Height-growth failure is the feature that should alert the clinician.Recommendation: level BAn obese child or adolescent with height-growth failure should be referred to a paediatrician or an endocrinologist, or both.Evidence-based statementEvidence levelHypothalamic damage can result in a severe form of III-2obesity in children and adolescents.Recommendation: level BConditionsthatcausehypothalamicobesityarerareandshouldbe managed in a tertiary institution.Evidence-based statementEvidence levelThere is general agreement that, at the end of therapy III-2for acute lymphatic leukaemia, there is a higher prevalence of obesity among subjects than at the commencement of therapy and that obesity persists.Recommendation: level BOncologistsandotherswhocareforchildrenandadolescentswith acutelymphaticleukaemiashouldbeawareoftheriskofobesityafter treatment.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xviiiEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xixEvidence-based statementEvidence levelThe role of pharmacological agents in causing weight gain IIin children and adolescents has not been extensively studied.Recommendation: level BIf a child or adolescent is placed on pharmacotherapeutic compounds that have been clinically identifed as causing excess weight gain, remain alert to this side effect. Research recommendationResearch into the effects of psychoactive agents on weight gain in children and adolescents is needed. MORBI DI TY ANDCHI LDHOOD AND ADOLES CENT OBES I TYChildhood and adolescent obesity that persists into adulthood confers increased adult morbidity and mortality risks. Childhood and adolescent obesity is, however, also associated with current morbidity, which requires identifcation and possible investigation and treatment.Evidence-based statementEvidence levelSelf-esteem has been measured in different ways in III-2different subject groups. Overall, the evidence suggests that increasing age and female gender are associated with lower self-esteem in obese children and adolescents.Recommendation: level BBe aware that impaired psychosocial function is often a feature in childhood and adolescent obesity and is more likely in older girls.Research recommendationThe social impact of being an obese child or adolescentboth at the family level and at the broader societal levelneeds to be studied. Evidence-based statementEvidence levelCardiovascular risk factors and obesity cluster in III-2childhood and track through to young adulthood. Cardiovascular risk factors in parents further predict the presence of such risk factors in their children.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xviiiEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xixRecommendation: level BAfastinglipidprofleshouldbeconsideredinobesechildrenand adolescents, particularly those who have a family history of cardiovascular risk factors. Evidence-based statementEvidence levelThe prevalence of type 2 diabetes is increasing in III-2children and adolescents, particularly in certain ethnic groups. This increase appears associated with high levels of obesity in these populations.Recommendation: level BFastinginsulinandglucoseshouldbeconsideredinobesechildrenor adolescents, particularly those with a family history of type 2 diabetes, those with Acanthosis nigricans, and those from certain ethnic backgrounds.Research recommendationResearchisneededtodeterminethecontributionoftheincreasing prevalenceofobesityinchildrenandadolescentstotheincreasing incidence of type 2 diabetes in these age groups. Studies on type 2 diabetes in Aboriginal and Torres Strait Islander Australian children and adolescents are specifcally needed. StatementOn the basis of mainly descriptive studies and expert opinion, there is some evidence of a higher level of physical discomfort in the obese child.Recommendation: level DBe aware that, despite their minor medical signifcance, problems of physical discomfortamongthemmusculoskeletaldiscomfort,heatintolerance, and shortness of breathappear to have the potential to greatly affect the lifestyle of obese children and adolescents. Evidence-based statementEvidence levelObesity in childhood and adolescence is IVassociated with signifcant medical morbidity, which impairs current physical health.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxiRecommendation: level CBeawarethatmedicallysignifcantdiagnosessuchasobstructivesleep apnoea,hepaticsteatosis,type2diabetes,slippedcapitalepiphysisand polycystic ovariesappear in children and adolescents who are obese.Research recommendationThe health costs of medical morbidity in obese children and adolescents requirefurtherstudy.Costswillincludethosetotheindividual(both medium and long term) and those to the community, including direct health care costs. Evidence-based statementEvidence levelOverall, the evidence suggests that a signifcant III-2proportion of children and adolescents use unhealthy dietary practices for weight control; these practices are more common in those who are overweight and in females.Recommendation: level BBeawarethatdisorderedeatingpracticescanbeafeatureofchildhood and adolescent obesity.Research recommendationAbnormal eating behaviours in childhood and adolescent obesity should be addressed both before and during weight-management programs. Evidence-based statementEvidence levelAll studies on weight management have used III-2some measure of weight as the primary outcomeindicator. Few studies have attempted to evaluate secondary outcome indicators. There is evidence that cardiovascular risk factors can be altered with successful weight management.Recommendation: level B Outcome indicators of success additional to weight change should be part of any weight-management program in children and adolescents. EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxiCONVENTI ONAL WEI GHT- MANAGEMENTS TRATEGI ESConventional weight-management strategies have not been studied as extensively in children and adolescents as they have in adults. Conventional strategies are a reduction in energy intake, by dietary means and using conventional food items; an increase in energy expenditure, by an increase in physical activity and a decrease in sedentary behaviours; behaviour modifcation; and family involvement in the process of change. Children and adolescents maintain a weight loss better than do their parents. Evidence-based statementEvidence levelIn the long term, children may have a higher degree III-3of success with weight management than adults. There are no similar data for adolescents.Recommendation: level CIfachildisobese,weightmanagementshouldstartinchildhood,rather than being deferred until adolescence or adulthood. Absolute weight loss isgenerallynotnecessaryinyoungchildren.Inyoungchildren,weight management generally requires only weight maintenance until the weight percentilemovesclosertotheheightpercentileusingnormalheight-growth potential. Evidence-based statementEvidence levelPublished results of weight-management programs III-1using dietary modifcation, increased physical activity and behavioural change show modest success in children and adolescents in the medium to long term. Most of these studies are limited to a few investigating groups, however, and were performed mostly under clinical research conditions. They provide direction for more universal weight-management strategies, but there is as yet no evidence that such programs would be successful in community and primary care settings. The relative contributions of diet, exercise and behavioural modifcation cannot be determined.Recommendation: level BMakeuseofalltheconventionalcomponentsofweightmanagement-dietary modifcation, increased physical activity, decreased sedentary activity, familyinvolvement,andbehaviourmodifcation-whentreatingobesityin children and adolescents, since the relative importance of each component is unknown.

EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxiiEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxiiiResearch recommendationSimple, well-designed intervention studies in obese children and adolescents, which can be translated into usual clinical practice, are urgently needed in Australia. StatementThere is no direct evidence for which dietary modifcation works best in weight management in children and adolescents.Recommendation: level DFor weight management, children and adolescents should be encouraged to follow the Dietary Guidelines for Children and Adolescents in Australia and the Australian Guide to Healthy Eating.Research recommendationWell-designedstudiesareneededtodeterminetheoptimaldietary prescription for weight management in children and adolescents. Evidence-based statementEvidence levelThere is some limited evidence that increasing physical III-2activity improves weight-loss outcomes in children or adolescents and may be effective by itself if vigorous.Recommendation: level BMorephysicalactivitythaniscurrentlybeingengagedinshouldbe prescribedforthemanagementofobesityinchildrenandadolescents. Becausethereisnoevidenceastotheamount,intensityortypeto prescribe, and because it is known that a child's activity and play varies with age and stage of development, the prescription should be based on age-appropriate activity.Research recommendationThere is a need for research into the optimal physical activity prescriptions for effective weight loss in obese children and adolescents. Evidence-based statementEvidence levelThere is short-term evidence that reducing III-3sedentary behaviours in obese children is as effective for weight management as increased activity.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxiiEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxiiiRecommendation: level CTo help manage obesity in children and adolescents, reduce the time they spend in sedentary behaviours each day.Research recommendationThereisaneedforresearchintotheoptimalprescriptionsforreducing sedentary behaviours in obese children and adolescents. StatementThere is no evidence about the detail of how, when and what behaviour-modifcation approaches should be usedfor childhood and adolescent obesity.Recommendation: level DSimple age-appropriate behaviour modifcation should be incorporated in any weight-management program for obese children and adolescents.Research recommendationThereisaneedforresearchtoidentifyage-appropriatebehaviour-modifcationapproachesthatpromotelong-termmaintenanceofweight control in obese children and adolescents.Evidence-based statementEvidence levelFor children and adolescents, there is evidence III-2that weight-management programs that involve parents achieve better outcomes than programs that do not. For children of primary school age, there is also evidence that a program that involves parents alone does better than one that requires regular attendance by their children as well.Recommendation: level BInvolve parents in the management of overweight and obesity in children and adolescents. Parents can alter environments substantially, especially for children of primary school age.Research recommendationThere is a need for research to identify the role of effective parenting in treatment approaches to childhood and adolescent obesity. EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxivEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxvStatementThere is no strong evidence as to the optimal frequency or duration of clinic visits in the management of childhood and adolescent obesity.Recommendation: level D Whenplanningthefrequencyofclinicvisits,beawarethatobesity managementinchildrenandadolescentsisamedium-tolong-term intervention. Evidence-based statementEvidence levelThere is some limited evidence that a weight-III-3management program for children and adolescents can be delivered in a variety of settings and achieve similar outcomes. The majority of such programs use a group format.Recommendation: level CWeight management for children and adolescents should not be limited to a hospital setting.Research recommendationResearch is needed to identify weight-management programs that can be effectively delivered in a variety of settings.NON- CONVENTI ONAL WEI GHT- MANAGEMENT S TRATEGI ESFor more extreme degrees of obesity and associated co-morbidity, particularly in adolescents, it may be necessary to consider weight-management strategies additional to conventional interventions. Conventional interventions will always accompany additional strategies, among which are very low energy diets, anti-obesity drug therapy, and surgery.Evidence-based statementEvidence levelVery low energy diets (VLED) produce a rapid weight III-2loss over a short period in adolescents.Recommendation: level BVLEDtherapyinadolescentsshouldbeundertakenonlybyspecialist obesity-management teams. VLED therapy is never indicated for children.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxivEVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxvEvidence-based statementEvidence levelThe evidence that very low energy diets IVproduce any long-term weight-loss beneftis unclear.Recommendation: level C Aftercessationofaverylowenergydiet,thereshouldbeacontinuing weight-management plan. StatementThere is no evidence that sibutramine has a role in the management of adolescent obesity.Recommendation: level DBecauseoflackofdata,includingdataonshort-andlong-termharm, sibutramine should be used in obese adolescents with co-morbidity only in a specialist centre and only when there is a reasonable expectation of beneft over risk.Research recommendationLong-term use of sibutramine in adolescents with obesity associated with signifcant co-morbidity should be trialled. Evidence-based statementEvidence levelShort-term uncontrolled trial data suggest that IVorlistat may assist with weight loss in obese adolescents.Recommendation: level CBecauseoflimiteddata,includingdataonshort-andlong-termharm, orlistat should be used in obese adolescents with co-morbidity only in a specialist centre and only when there is a reasonable expectation of beneft over risk.Research recommendationWell-designed clinical trials of the use of orlistat in obese adolescents with obesity-related co-morbidity are needed.Evidence-based statementEvidence levelMetformin appears to have a potential role in therapy III-3in obese non-diabetic hyperinsulinaemic adolescents.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxvi1THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 1Recommendation: level CConsidermetformintherapyintheobeseadolescentwithsignifcant hyperinsulinaemia and who has a family history of diabetes.Evidence-based statementEvidence levelThere is evidence that gastric restrictive or gastric IVbypass surgery induces a weight loss in adolescents, with a reduction in obesity-related co-morbidity that is comparable to that found in adult studies. The overall numbers are low, however, and long-termfollow-up data are limited. Not every subject experiences signifcant weight loss, and there are no good data to suggest who will be successful. Post-operative morbidity is common.Recommendation: level CBariatricsurgerymightbeconsideredasthelastpossibleoptionina severelyobeseadolescentwithobesity-relatedco-morbidity.Sucha procedureshouldbeundertakenonlyinanexperiencedsurgicalcentre afterextensiveconsultation,lengthyeducationofthepatientandtheir family, and full psychological assessment. Continuing post-operative care in an experienced weight-management service would be mandatory.EVIDENCE-BASED STATEMENTS AND RECOMMENDATIONSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents xxvi1THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 11 THEPREVALENCE AND TRACKI NG OFOVERWEI GHT ANDOBESI TYI N CHI LDREN AND ADOLESCENTS1. 1PREVALENCEAlthough Australia does not routinely conduct national, representative and repetitive surveys of childrens height and weight, from which a prevalence of obesity might be derived, there have been some small surveys in the last 25 years, albeit with differing methodologies (Court et al. 1976; Harvey & Althaus 1993; Sciberras & Darnton-Hill 1985).The most extensive survey of height and weight, among other measures, was the 1985 Australian Health and Fitness Survey conducted by the Australian Council for Health, Physical Education and Recreation (ACHPER) on a representative sample of 8492 Australian school children aged between 7 and 15 years (Pyke 1987). This study has thus become the baseline against which other studies can be compared. Even older Australian data, from the 1969 Australian Youth Fitness Survey, were recently reported: obesity prevalence rates in 13 to15 year old boys and girls were 0.7 per cent and 2.3 per cent respectively; for overweight, the rates were 7.1 per cent and 10.5 per cent (Booth et al. 2003). From 1969 to 1985 there was no change in overweight or obesity in girls, but some increase in boys. From 1985 to 1997 the population prevalence of overweight increased by 60 to 70 per cent and obesity increased 200 to 400 per cent.There is evidence that the prevalence of obesity in children and adolescents has increased globally. The most consistent evidence comes from the United States, notably from the American National Health and Nutrition Examination Survey (NHANES) (Freedman et al. 1997; Troiano & Flegal 1998). Recent British and Spanish cohort studies in preschool and school-age children (Rudolph et al. 2001; Moreno et al. 2000; Reilly 1999) indicate an increasing prevalence of obesity. The prevalence of overweight more than doubled between NHANES II (1976-1980) and III (1988-1994) (Flegal & Troiano 2000). There is also a positive skewness occurring in the heavier children, and the heavier children are becoming heavier. This observation has been reported for an Australian data set on 2277 primary school children aged between 7 and 12 years who took part in the Health of Young Victorians Study (Lazarus et al. 2000). When compared with the 1985 school-based ACHPER survey, an increase of slightly over one unit of BMI in the 12-year interval was recorded, with the greatest change being at the heavier end of the distribution. Two other published studies have compared the results of the 1985 ACHPER survey with more recent data (Magarey et al. 2001; Wilckens et al. 1996). The Wilckens et al. study was primarily a cardiovascular risk study in Sydney primary school children and does not show as dramatic a change as the other two studies. The Magarey et al. study is a re-analysis of data against the international standards for 1THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 21THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 3overweight and obesity (Cole et al. 2000); the original data were from the National Nutrition Survey (McLennan & Podger 1998) and the 1985 ACHPER survey (Pyke 1987). In the ACHPER survey 11.8 per cent of males and 10.7 per cent of females were overweight and of this group only 10 per cent were obese. In the National Nutrition Survey, 19.5 per cent of males and 21 per cent of females were overweight and a quarter of this group were obese. Controlling for age and gender, these rates represent a relative risk of overweight and obesity of 1.79 and 3.28 respectively in 1995, compared with 1985. When compared with recently published British trend data on children over the same period (Chinn & Rona 2001), the decade increase is similar, but more Australian children were overweight at both points.The three recent data setsfrom the Health of Young Victorians Study (Lazarus et al. 2000), the National Nutrition Survey (McLennan & Podger 1998) and the New South Wales Fitness and Physical Activity Survey (Department of Education and Training 1997)have been re-analysed against the Cole et al. (2000) cut-off points (Booth et al. 2001). These studies were not designed to be comparative and have methodological differences, but the results are similar. The combined average prevalence of obesity and overweight was a little over 20 per cent in boys and about 22 per cent in girls. Thus, one in fve children and adolescents were above a healthy weight; a quarter of this group were obese. No consistent age trends were identifed. Table 1.1 summarises the data from the three studies. Hesketh et al. (2002) have reported further 1997-2000 data from the Health of Young Victorians Study, showing that 11 per cent and 0.4 per cent of those who were in the healthy weight range in 1997 had progressed to overweight and obesity respectively, demonstrating a continuing trend to increased fatness. Table 1.1The weight status of children from three Australian studies: a summary SFPASNNSHOYVSn = 5518n = 2962n = 3104BMI categoryBoysGirlsBoysGirlsBoysGirlsAcceptable79.978.780.777.778.976.4Overweight14.916.314.416.915.817.8Obese5.24.94.95.45.35.7Overweight/obese20.121.319.322.321.123.5 (per cent)Note: SFPAS = NSW Schools Fitness and Physical Activity Survey; NNS = National Nutrition Survey;HOYVS = Health ofYoung Victorians Study.Source: Booth et al. (2001).1THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 21THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 3Evidence-based statementEvidence levelA very signifcant percentage of children and adolescents III-2in Australia are overweight or obese.The prevalence of the condition has increased signifcantly in the last 10 years. The trend is similar in the majority of other developed and developing countries.Recommendation: level BTheprevalenceofoverweightandobesityinchildrenandadolescents in Australiaishighenoughtowarrantbothinterventionandpreventive action.Research recommendationHeight and weight should be measured recurrently in a nationally repre- sentative sample of children and adolescents. 1. 2TRACKI NGTracking is a well-established concept in the paediatric literature; it implies that an abnormality in childhood will continue, at least in some degree, into adulthood. If this was not true for obesityin other words, if the persistence of obesity from childhood to adulthood was a random effectthe case for intervening to prevent childhood obesity would be weak. The case would also be weakened if there was no established morbidity associated with childhood and adolescent obesity. The studies on tracking of body weight in childhood are in general agreement. Not all the studies use the same defnition of obesity, and many started well before the worldwide increase in obesity began. Further, there have been variable predictions of adult obesity tracking through from childhood obesity. Nevertheless, no study has demonstrated the absence of a relationship between childhood and adult weight. The persistence of overweight is a universal phenomenon, but what varies from study to study is the degree of effect and at the age from which the prediction is valid.The studies that were evaluated for this reviewGuo et al. (2000); Whitaker et al. (1997); Power et al. (1997a); Kelly et al. (1992); Freedman et al: (1987); Braddon et al. (1986); Peckham et al. (1983); and Stark et al. (1981)had more than 500 subjects; longitudinally followed the subjects through to at least adolescence, and measured subjects height and weight (as opposed to using self-reported measurement).The risk for an obese child becoming an obese adult is considered high in these studies. They found, however, that less than a third of obese adults were obese in childhood. Up to 50 per cent of obese adolescents remained obese in adulthood, and the later in adolescence that overweight persisted and the greater the degree of overweight, the more likely it was that an individual would be an obese adult.1THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 41THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 5Stark et al. (1981) followed 5362 British children born in 1946 and measured at ages 6, 7, 11, 14, 20, and 26 years. The risk of being overweight in adulthood was related to the degree of overweight in childhood. For subjects who were not obese at age 7 years, the risk of being overweight later in childhood was less than one in ten. Most overweight adolescents became overweight after the age of 7 years. For a child aged 7 years or more, who had a relative weight of 130 per cent, the risk of remaining overweight was six in ten. Stark et al. did not identify an optimal age for prediction.Braddon et al.s data come from a Medical Research Council National Survey of Health and Development in Britain. Subjects were born in 1946 and were studied from birth to age 36 years. The prediction of adult obesity from weight at 11 years was approximately 40 per cent, rising to over 50 per cent at 26 years, showing that childhood obesity was not as good a predictor of adult obesity as was obesity in young adulthood.Peckham et al. (1983) considered the Medical Research Council National Survey of Health and Development (1946) and the National Child Development Study (1958)both British-born cohorts. There were 16 994 subjects for the 1958 cohort and they were followed to age 16 years. The prevalence of overweight at age 7 years in the second cohort had doubled compared with the frst cohort. The correlation between relative weights at different ages were generally higher for girls than for boys.Kelly et al. (1992) looked at 1037 New Zealand children who were measured at ages 3, 5, 7, 9, 11, and 13 years, starting in 1973. About two-thirds of 7- and 11-year-olds were still in the same BMI category at age 13 years. Correlations between present and later BMI scores increased with age and were established by 7 years of age. Tracking stability (i.e. the individual maintaining a percentile) is high for those individuals who commenced at very high or very low BMI percentiles. Children with signifcant early obesity continue to maintain their obesity. Only 1 per cent of signifcantly obese children spontaneously reverted to a more normal BMI over the course of the study.Guo et al. (2000) followed 505 children from the Fels Longitudinal Study and concluded that overweight at age 35 years can be predicted from BMI at a younger age. Increasing BMI percentile and increasing age strengthened the prediction of adult BMI. The prediction was excellent at 18 years, good at 13 years, and moderate at less than 13 years. Stronger correlations appeared earlier in females than in males.Cardiovascular risk factors track in childhood. This has been best demonstrated in the Bogalusa Heart Study (see Section 4.2). Obesity in childhood is associated with increased adult cardiovascular morbidity and mortality, regardless of the adult weight. In a 40-year personal follow-up of overweight children, Mosberg (1989) demonstrated increased mortality in adult life in children whose standard deviation weight score was >+3 in puberty and a signifcant increase in chronic disease in adulthood, as well as an earlier appearance of chronic disease. 1THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 41THE PREVALENCE AND TRACKING OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents 5In a review of the third Harvard Growth Study, Must et al. (1991) demonstrated that the relative risks for all-cause mortality and coronary heart disease mortality were 1.8 and 2.3 respectively for males who were overweight in adolescence, compared with those who were not. In a retrospective study, Nieto et al. (1992) determined that relative risk for mortality for children in the top versus bottom quintile was 1.5 pre-pubertally and 1.6 post-pubertally for both males and females. Evidence-based statementEvidence levelRelative body weight tracks from childhood to adulthood, III-2and the predictive power of this association increases with age. Once a child or adolescent is on an overweight or obese percentile, spontaneous track-down is unlikely.Recommendation: level BTracking data support intervention in childhood overweight and obesity. 72DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2 DEFI NI TI ON ANDMEASUREMENT OFOVERWEI GHT ANDOBESI TYI N CHI LDREN AND ADOLESCENTS2. 1I NTRODUCTI ONMany defnitions of overweight and obesity in children and adolescents have been used in the literature and in clinical practice, and the relationship between different measures is unclear. Although there is a dictum that if a child looks fat then they are fat, this is unsound advice. In early childhood the appearance of fatness as judged by skinfold size may be normal. It is open to viewer bias (Moynihan et al. 1986) and is a qualitative assessment. Clinicians need a defnition so they can identify children who are most at risk of adverse health outcomes and in whom intervention is required. The consensus for clinical practice is that the age-related BMI should be used. There are currently no Australia-specifc growth charts for use in clinical practice. Until such charts are available, another countrys data set must be used in clinical practice. Development of an Australian reference would take into consideration the unique ethnic mix of the Australian community. There are no defnitions of obesity and overweight for children under the age of 2 years, and this review is thus limited to children older than 2 years.2. 2MEAS URES OF ADI POS I TYPower et al. (1997b) defned the ideal measure of body fat as accurate in its estimate of body fat; precise, with small measurement error; accessible, in terms of simplicity, cost and ease-of-use; acceptable to the subject; and well-documented, with published reference values. They add, however, that no existing measure satisfes all these criteria. Indirect (anthropometric) and direct methods of measuring body fatness have been used in children. 2.2.1Anthropometric measuresAmong the anthropometric measures of relative adiposity, or fatness, are waist, hip and other girth measurements; skinfold thickness; indices derived from measured height and weight (such as BMI); the ponderal index; and Benns index. The accuracy of anthropometric measurements depends on the skill of the measurer and the precision of the measuring equipment. Accuracy must be validated against a gold standard, or reference measure of adiposity. Height is measured standing in children over the age of 2 years. There should be a fxed wall stadiometer and the measurement should be to the nearest millimetre.Weight is measured by standing on scales; measurement should be to the nearest 0.1 kilogram.8 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS92DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsWaist is measured at the minimum circumference between the iliac crest and the rib cage, using an anthropometric tape. Body mass indexBMI is a weight-to-height ratio defned as weight (kilograms)/height (metres)2. It is widely used as an index of relative adiposity in adults on the basis of evidence that there is an increased risk of morbidity and mortality associated with higher BMIs. The methodology cannot be used in children and adolescents since the correlation with morbidity has not been attempted.The accuracy of BMI as a screening test in children has been compared with that of a more direct measure of adiposity, dual-energy X-ray absorptiometry (DEXA). The correlation between BMI and more direct measures is variable, ranging from 0.5 to 0.85 (Sardinha 1999; Pietrobelli et al. 1998; Himes et al. 1994). The specifcity, or false-positive rate, ranges from 0.03 to 0.13. Thus, although some overweight children would be wrongly classifed as being of normal weight when BMI is used as a screening test, very few children would be classifed as overweight when they were not. BMI may not be as sensitive a measure of body fatness in children and adolescents who are particularly short or tall for their age or have an unusual body-fat distribution. It may also misclassify children and adolescents who have highly developed muscles. Further, there are racial differences in the relationship between the true proportion of body fat and BMI, and appropriate cut-off points may vary as a result (Must et al. 1991).Evidence-based statementEvidence levelBMI is signifcantly associated with body III-3fatness in childhood and adolescence.Recommendation: level CBMI is a reasonable, easily determined surrogate measure for adiposity in children and adolescents.Research recommendationStudies should be performed to correlate current medical morbidity with BMIinAustralianchildrenandadolescents.Aboriginaland TorresStrait Islander children and adolescents may need to be considered separately. BMI percentilesBMI changes with age and with gender. An absolute BMI calculation for a child or for an adolescent below the age of 18 years must be evaluated against age and gender reference standards. The US Centers for Disease Control and Prevention have recently developed new growth reference charts, which include gender- and age-specifc BMI ranges. The new charts are based on data from fve national health examinations (NHANES) 8 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS92DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescentsbetween 1963 and 1994 and fve supplementary data sources. During the three decades of measurement there was a signifcant increase in the prevalence of obesity and overweight: weight has increased more than height. This change would have caused an upward shift of the weight and BMI curves. To avoid this, data from the most recent survey were excluded for children over 6 years of age (CDC 2000).Figures 2.1 and 2.2 show the CDC reference charts for males and females aged 2 to 20 years.Figure 2.1BMI-for-age percentiles: boys, 2 to 20 yearsPublished May 30, 2000. SOURCE:Developed by the National Center for Health Statistics in collaboration with the National Center for Chronic Disease Prevention and Health Promotion (2000). 97th95th90th85th75th50th25th10th5th3rd234567891011121314151617181920kg/ms kg/msBMI343230282624222018161412BMI343230282624222018161412 CDC Growth Charts: United StatesAge (years)k g / m2k g / m210 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS112DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsFigure 2.2BMI-for-age percentiles: girls, 2 to 20 yearsTable 2.1 shows the Australian reference standard cut-off points for overweight and obesity in males and females aged 2 to 18 years.Published May 30, 2000. SOURCE:Developed by the National Center for Health Statistics in collaboration with the National Center for Chronic Disease Prevention and Health Promotion (2000).97th95th90th85th75th50th25th10th5th3rd234567891011121314151617181920kg/ms kg/msBMI343230282624222018161412BMI343230282624222018161412 CDC Growth Charts: United StatesAge (years)k g / m2k g / m210 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS112DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsTable 2.1Australian reference standard BMI cut-off points for overweight and obesity: males and females, 2 to 18 yearsBMI equivalent to 25 in adultBMI equivalent to 30 in adultsAge (years)MalesFemalesMales Females218.4118.0220.0919.812.518.1317.7619.8019.55317.8917.5619.5719.363.517.6917.4019.3919.23417.5517.2819.2919.154.517.4717.1919.2619.12517.4217.1519.3019.175.517.4517.2019.4719.34617.5517.3419.7819.656.517.7117.5320.2320.08717.9217.7520.6320.517.518.1618.0321.0921.01818.4418.3521.6021.578.518.7618.6922.1722.18919.1019.0722.7722.819.519.4619.4523.3923.461019.8419.8624.0024.1110.520.2020.2924.5724.771120.5520.7425.1025.4211.520.8921.2025.5826.051221.2221.6826.0226.6712.521.5622.1426.4327.241321.9122.5826.8427.7613.522.2722.9827.2528.201422.6223.3427.6328.5714.522.9623.6627.9828.871523.2923.9428.3029.1115.523.6024.1728.6029.291623.9024.3728.8829.4316.524.1924.5429.1429.561724.4624.7029.4126.6917.524.7324.8529.7029.841825.0025.0030.0030.00Source: Cole et al. (2000).As noted, Australia does not have locally derived BMI-for-age charts: most paediatric institutions currently use charts provided by a pharmaceutical company, with the data source from Hammer et al. (1991).If BMI-for-age percentiles are used, a child can be described as being above or below certain percentile linesfor example, the 85th or 90th percentiles, which correspond roughly to 110 and 120 per cent overweight. A child can be followed over time with serial BMI measures, reinforcing the concept of serial measurement. Any child who 12 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS132DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescentsis rapidly crossing up percentiles is detected, as is a child who is persistently at an elevated percentile. There are, however, several potential diffculties with the use of BMI-for-age percentiles: Data are derived from a reference population. Classifying a child as overweight or obese on the basis of BMI being above a certain percentile is an arbitrary decision and not based on known medical or health risk. Clinicians might wrongly interpret the percentiles as representing an ideal population, rather than a reference population, when they may in fact have been developed from a population with a greater prevalence of obesity. A reference population is a population standard of which inquires can be made as to the integrity, capacity and the like, of another.Diffculties may occur if the reference population is not representative of the target population.In an Australian study, Lazarus et al. (1996b) examined children and adolescents aged 4 to 20 years using the 85th percentile as the BMI cut-off point. BMI had a true-positive rate of 0.67 and a false-positive rate of 0.06 for detecting total body fat >=85th percentile as measured by dual energy XR absorptiometry (DXA).BMI-for-age Z scoresIn a tertiary setting, BMI could be compared with a reference data set and reported as a Z score. The BMI-for-age Z score needs to be calculated using a suitable software program, according to the following formula:(observed value)(median reference value of a population)standard deviation of reference populationThus, a Z score of 0 is equivalent to the median, or 50th percentile, value and a Z score of +2.00 is approximately equivalent to the 97th percentile. Use of BMI-for-age Z scores allows a more detailed statistical description of individuals, particularly individuals at extremes of BMI. The statement that an individual has a BMI greater than the 97th percentile does not describe how far above the percentile that individual is, but the BMI Z score does.There is no accepted defnition of severe or extreme obesity in children and adolescents. Arbitrary cut-off points of >+3.0 and >+4.0 BMI Z scores have been proposed. Most studies in the past have used the percentage of ideal body weight, or IBW. The IBW is calculated using the standard weight and height percentile curves and the following formula:childs weightpercentile weight for height for a child of the same age and gender A BMI Z score of +2.0 would approximate 125 per cent of IBW.x 10012 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS132DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsBMI based on adult cut-off pointsAn expert committee convened by the International Obesity Task Force in 1999 determined that, although BMI was not perfect as a measure of adiposity, it had been validated against more direct measures of body fatness. BMI may therefore be appropriate for defning overweight and obesity in children and adolescents (Bellizi & Dietz 1999). The BMI level at which adverse health risk factors increase in children is not known, but the expert committee recommended cut-offs based on the accepted adult cut-offs of 25 for overweight and 30 for obesity.Cole et al. (2000) developed a reference population by using data from national surveys in six countriesGreat Britain, Brazil, the Netherlands, Hong Kong, Singapore, and the United Stateswith widely divergent obesity prevalence rates. The data from each country were used to construct gender-specifc BMI percentile curves that at 18 years passed through the points of 18.5, 25 and 30. The curves from each country were then averaged to provide age- and gender-specifc BMI cut-off points to defne overweight and obesity. These cut-off points correspond to the adult cut-off points for overweight and obesity, points that are related to morbidity and mortality. Because the cut-off points were developed using a variety of data sets, they represent an international reference that can be used to compare populations worldwide. Cole et al. caution that the cut-off points are not for clinical use: the data are not available as BMI-for-age reference charts. The combined data set does, however, go some way in resolving the problem of the known ethnic differences in body composition (Wang et al. 1994).BMI summaryThe National Health Data Dictionary recommends that in the Australian context the BMI curves in Cole et al. (2000) be used for research purposes, but that the Centers for Disease Control and Prevention (CDC 2000) curves be used in the clinical setting.BMI percentiles in clinical practiceAustralia will now use the latest BMI percentile charts from the US Centres for Disease Control and Prevention (CDC 2000), with the 85th centile and above as overweight and the 95th centile and above as obese. Different BMI percentile cut-off points for the classifcation of overweight and obesity could be chosen, since cut-off points are arbitrary, but this approach provides consistency with the literature.Australia could have chosen to use growth reference charts from another well-characterised population, such as the United Kingdom or the Netherlands. However, the CDC charts are the most accessible and Australia more closely resembles the United States in terms of overweight and obesity prevalence.The World Health Organization (WHO) is currently considering the development of an international growth reference for clinical practice, and Australia could choose to use this when it becomes available.14 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS152DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsThe fnal recommendation for BMI percentiles was that: it is desirable that the same data set be used for height and weight assessment, as well as for measuring fatness the data set be a true reference standard (not normative). It may be that children and adolescents measured in the last decade already have too high a prevalence of excess fatness and an older data set should be considered the percentiles be freely available for use in Australia there be suffcient percentile lines the same percentiles can be used to assess under-nutrition.StatementThere is no evidence on which to base a decision about which BMI measure to use.Recommendations: level DBMI should be used as the standard measure of overweight and obesity for 2 to 18 year olds in Australia. The Australian reference standard based on the work of Cole et al. (2000) should be used in population and clinical research.BMI-for-age percentile charts should be used in clinical practice and in non-healthcaresettings,aBMIabovethe85thpercentilebeingindicativeof overweight and a BMI above the 95th percentile being indicative of obesity. It should be noted that this defnition of overweight and obesity is arbitrary and that a more appropriate defnition is needed but not yet available.The CentersforDiseaseControlandPreventionBMIpercentilechartsare recommended for use until local BMI growth charts are developed.There is a need for Australia to develop such charts for clinical practice very soon. Research recommendation ThereisaneedtodeveloplocalBMIgrowthchartsforusein Australia. Separate BMI growth charts may need to be considered for Aboriginal and Torres Strait Islander people.Waist circumferenceUsing anthropometric measures, a number of authors have demonstrated that there is a genetic infuence on the distribution of body fat (Esposito-Del Puente et al. 1994; Donahue et al. 1992; Bouchard et al. 1988). Using NHANES II cross-sectional data, Gillum (1987) demonstrated that the waist-hip ratio was lower in females aged 6 to 17 years than in males of that age and that the ratio fell with age, levelling out at mid-adolescence. This fnding was similar to that of Rolland-Cachera et al. (1990), who used skin folds to describe the evolution of body fat distribution in a longitudinal study of French subjects between the ages of 1 and 20 years. They found the sub-scapular-triceps skinfold ratio a particularly good discriminator of centralised fat distribution. A high childhood BMI and high trunk skinfold values were predictive of centralised obesity as an adult. 14 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS152DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsHigh central adiposity is strongly correlated with the risk of cardiovascular disease in adults and is particularly related to the presence of visceral (intra-abdominal) fat. Waist circumference measures in childhood track well into adulthood (Goran 1998b). The correlation between waist circumference and DEXA measure of trunk fat is >0.8 (Taylor et al. 2000).Using magnetic resonance imaging, Brambilla et al. (1994) reported on 23 obese and 21 normal children aged 10 to 15 years. Subcutaneous fat was predominant in both obese and non-obese children when compared with intra-abdominal fat. Intra-abdominal fat did not correlate with waist-hip ratio in obese or non-obese subjects, but it did correlate with trunk-arm skinfold ratio in obese children. The obese children had a signifcant correlation between total and LDL cholesterol and triglycerides and intra-abdominal fat. They also had a near-signifcant correlation with area under the curve of insulin, in response to a glucose tolerance test. These results were confrmed by Goran et al. (1995b) using computerised tomography, where the ratio of intra-abdominal fat to subcutaneous fat was low. Intra-abdominal fat correlated with trunk-extremity skinfold ratio but not with waist-hip ratio.Caprio et al. (1996, 1995) used MRI to quantify intra-abdominal fat depots in obese and non-obese girls. The obese girls had a greater waist-hip ratio, a lower ratio of abdominal to subcutaneous fat, and a threefold increase in total fat depots. Caprio et al. demonstrated that abdominal fat showed a positive correlation with basal insulin and triglycerides and a negative correlation with HDL cholesterol. The obese girls were less insulin sensitive using the glucose clamp technique and had a greater stimulated insulin release, both of which correlated with abdominal fat. Unlike the situation with adults, there are no universally accepted cut-off points for waist circumference in children and adolescents because the relationship between waist measure and metabolic complications in children and adolescents remains undefned. In a cross-sectional study of 87 children, Higgins et al. (2001) used receiver operating characteristic (ROC) analysis to determine body fat percentage and waist circumference cut-off points in pre-pubertal children and thus defne cardiovascular risk. They found that children with 33 per cent or more body fat and a waist circumference of 71 centimetres or more were likely to have an adverse cardiovascular risk profle. Children with 20 per cent or less body fat and a waist circumference of 61 centimetres or less were likely to have a healthy cardiovascular risk profle. Evidence-based statementEvidence levelAnthropometric markers and direct measurement III-2of central fat in children and adolescents positively correlate with cardiovascular risk.Recommendation: level BWaist circumference appears to be the best clinical determinant of truncal obesity, and hence metabolic risk, in children and adolescents and can be used for longitudinal assessment in management.16 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS172DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsWaist-hip ratioThe waist-hip ratio is not as easy to measure as waist circumference alone; it is not considered here as a clinical measure.Skinfold thicknessSkinfold thickness measures subcutaneous fat. It is most often measured at the triceps and in the sub-scapular area. Skinfold thickness measures taken at multiple sites (with or without circumferential measures) can be used in equations to predict total body fat (Van den Broek & Wit 1997). Taking the measurement requires specifc equipment and operator skill; further, the skinfold caliper may not be wide enough to obtain a reading in an older fatter child.Skinfold thickness varies with race and gender (Goran et al. 1998b) and does not necessarily correlate with other adipose deposits (Wells 2001). There are no data on metabolic risks and an association with skinfold thickness.Other anthropometric measuresA number of other anthropometric measures, such as Benns index, the ponderal index and the conicity index, have been proposed for use as measures of adiposity (Lazarus 1996a). (The ponderal index is sometimes used for infants; it is defned as weight (kilograms)/height (metres)3.) These measures offer no benefts over BMI, so are not considered here.Research recommendationsLocalstandardsforclinicaluseforBMIandwaist-circumferencecut-off points should be developed.Research aiming to characterise the association between overweight and obesity (and particularly excess abdominal fat) in children and adolescents and adverse health outcomes should be a priority.ResearchintotheappropriateBMIandwaist-circumferencecut-off pointsin Aboriginaland TorresStraitIslanderchildrenandadolescentsis required. 2.2.2Direct measures of adiposityThere are a number of direct measures of body composition in which fat mass and various components of fat-free mass can be estimated. These are largely research tools or, at least, are restricted to tertiary care centres. Among the direct measures are bioelectrical impedance analysis, dual energy X-ray absorptiometry and underwater weighing (hydrodensitometry). Magnetic resonance imaging and computerised axial tomography can measure specifc fat depots but not total body fat. Direct measures can be used as a gold standard to validate indirect (anthropometric) measures of body fatness (Van den Broek & Wit 1997).16 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS172DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and AdolescentsBioelectrical impedance analysis (BIA)BIA can be used to evaluate body composition, including the fat mass and the fat-free mass (Heyward 1998). The equations used to convert resistance to body fat should be specifc for the population in which they are used, but operators are usually reliant on the manufacturers equation, which might not even be available. Additionally, BIA body-composition equations incorporate weight, height, age, and sex and may add little to anthropometric measures. Other shortcomings are that measurements can vary with hydration status and can be unreliable at body-weight extremes, and the method has limitations for longitudinal measurement. BIA is now a low-cost, portable method that is used in settings where its limitations may not be understood by the operators.Dual energy X-ray absorptiometry (DEXA)DEXA is based on the observation that X-rays (in this instance transmitted at two energy levels) are differentially attenuated by bone mineral tissue and soft tissue. It allows percentage and regional fat and lean mass to be determined, as well as bone mineral density. The radiation dose is low, so the technique can be used in children and adolescents. Regional fat identifed by DEXA has a high correlation with CT scan assessment of abdominal fat (Goran et al. 1995b). The drawbacks are that the technique requires expensive equipment and trained operators. Further, the subject must lie still for about 20 minutes and this may limit the techniques use in the under-6 age group. It is used only in tertiary centres.Magnetic resonance imaging(MRI)MRI provides a visual image of adipose tissue and non-fat tissue in the scanned sections of the body (Kamel et al. 2000). A highly specialised technique not suited to routine clinical use, it is excellent for the measurement of intra-abdominal fat but does not allow measurement of total body fat. There is no radiation exposure, but the procedure itself-which requires the subject to lie still in an enclosed scanneris unsuitable for young children. This is a research tool only.Computerised axial tomography (CAT)CAT scans produce high-resolution X-ray derived images and can identify small deposits of adipose tissue (Jensen et al. 1995). They are useful for quantifying intra-abdominal fat. Radiation is involved, however, so the technique is not suitable for use in children; in addition, the procedure itself is lengthy. This, like MRI, is a research tool only.Underwater weighingHydrodensitometry is based on the premise that the body can be divided into two compartmentsthe fat mass and the fat-free mass (Withers et al. 1996). Fat has a lower density than lean tissue and, by measuring the density of the whole body, the relative proportions of each component can be determined. Hydrodensitometry has traditionally been considered the gold standard in the assessment of body composition. The drawbacks are that the technique requires a person to hold their 18 Clinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescents2DEFINITION AND MEASUREMENT OF OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS193RISK FACTORSClinical Practice Guidelines for the Management of Overweight and Obesity in Children and Adolescentsbreath underwater, so it is unsuitable for use in young children or in adolescents who lack water confdence (Hills et al. 2001), and there are limited facilities available. This is a research instrument only.2. 3MEAS UREMENTSAS OUTCOMEI NDI CATORS I N WEI GHT- MANAGEMENT 2.3.1Weight lossabsoluteWeight change is the primary indicator of change in body fatness, but there are no data to c