Body Mass Index (BMI) and Childhood Obesity: A Review of Evidence and Guidelines, Study notes of Public Health

Download Body Mass Index (BMI) and Childhood Obesity: A Review of Evidence and Guidelines and more Study notes Public Health in PDF only on Docsity! FINAL VERSION Section 2: Identification and classification Obesity: full guidance FINAL VERSION (December 2006) Page 196 FINAL VERSION Identification and classification 5.1 A: Clinical 5.1.1 Evidence statements 5.1.1.1 Children (Table 5.1) Table 5.1 Evidence statements and grading No. Evidence statement Grade Body mass index (BMI) 1 BMI is a widely accepted and practical estimate of general adiposity in children 2++ 2 Different classifications using BMI centile cut-offs have been proposed for children, but there is no evidence on which are the most appropriate in practice 2++ 3 There is limited evidence on which BMI measure (BMI, percentage change BMI, BMI z-score or BMI centile) is best at measuring adiposity change 3 4 Some evidence suggests that the IOTF/Cole and the WHO BMI-based systems have high specificity which can lead to fewer non-overweight adolescents being classified as overweight 3 5 There is no evidence on ethnicity differences in the association of proxy measures of obesity with morbidity in children in UK populations N/A Obesity: full guidance FINAL VERSION (December 2006) Page 197 FINAL VERSION No. Evidence statement Grade 12 In South Asians (of Pakistani, Bangladeshi and Indian origin) living in England, a given waist circumferences tends to be associated with more features of metabolic syndrome than in Europeans (for example, higher triglycerides and lower HDLs in females and higher serum glucose in males) 2 13 In South Asians living in South Asia, a given BMI tends to be associated with higher percentage body fat than in European populations 3 14 In black populations, for a given BMI, percentage body fat tends to be higher in those living in the USA than in Jamaica. It also tends to be higher in Jamaicans compared with rural Nigerians 2 HDL, high-density lipoprotein; NA, not applicable. 5.1.2 5.1.3 Methodology We searched for high-quality systematic reviews of the evidence, and these are summarised below. We also searched for evidence published since the cut-off dates of the included reviews and evidence to answer key clinical questions not addressed in the reviews. Where appropriate, expert opinion is cited. Details can be found in the evidence review for each section. We did not retrieve any study from the update searches that modified any of the recommendations. Evidence review on different anthropometric measures for the identification of individuals who are overweight or obese There is growing evidence that links body composition, specifically the levels of fat tissue in the human body, with increased health risks and the development of certain diseases (see also section 5.1.5). The amount of body fat in the human body is called adiposity. Adiposity is defined as the amount of body fat expressed as either the absolute fat mass (in kilograms) or as the percentage of total body Obesity: full guidance FINAL VERSION (December 2006) Page 200 FINAL VERSION mass. Absolute adiposity is highly correlated with body mass, but percentage adiposity is relatively uncorrelated with body mass.1 There are many methods of directly measuring the amount of fat in the human body. These usually involve complicated procedures that can only be carried out in specialist laboratories. Indirect methods, based on the relation between height and weight, can be used in everyday clinical practice to estimate adiposity. The most common and accepted, at least in adults, measures are those of body mass index (BMI) and waist circumference. BMI is calculated as the weight (in kilograms) divided by the height (in metres) squared. For example, an individual who weighs 95 kg and is 180 cm tall has a BMI = 95/(1.80 × 1.80) = 95/3.24 = 29.32 kg/m2. So the person’s BMI is approximately 29 kg/m2. A simple measure of fat distribution is waist circumference. This can be related to the overall body shape of the individual by calculating the ratio of the waist to the hip (waist-to-hip ratio). Different methods may be appropriate in different circumstances. For example, waist-to-hip ratio may be the most accurate predictor of risk of myocardial infarction,2 and waist circumference may be the most accurate predictor of risk of type 2 diabetes.3 5.1.3.1 Identification and measurement of children who are overweight or obese We were not able to find any other systematic reviews that addressed the accuracy of anthropometric measures or bioimpedance to diagnose obesity compared with the use of BMI in children. The National Health and Medical Research Council (Australia) (NHMRC)4 stated that although there was no evidence to recommend specific cut-offs, it recommended that BMI should be the standard measure for children. BMI is a measure of weight adjusted for height and is highly correlated with adiposity. Obesity: full guidance FINAL VERSION (December 2006) Page 201 FINAL VERSION Limitations of the BMI, include: not being able to distinguish between fat or lean mass, not necessarily reflecting fat distribution (which may or may not be associated with age), and not necessarily describing the same levels of body fat in different populations because of different body proportions. Both the United States Preventive Services Task Force (USPSTF) 20055 and Freedman and coworkers6 reiterated these limitations. Freedman and coworkers6 pointed out pitfalls in their assessment of the relation of BMI to levels of fat mass and fat-free mass among healthy 5–18-year-olds. By measuring fat and fat-free mass by dual- energy X-ray absorptiometry they found that the correlation of BMI to fat mass was clearly non-linear, and that substantial differences in fat mass were only observed at BMI levels equal to or more than the 85th percentile. Thus, the authors contended that despite BMI-for-age being a good estimate of excess fat mass, BMI differences among thinner children can be partly associated with fat- free mass.6 For measurement of central adiposity, waist circumference was recommended but, as for BMI, no cut-offs were specified. The role of bioimpedance was reviewed and several limitations were highlighted: equations used to convert resistance to body fat should be population specific but these may not always be available; it may add little to anthropometric measures; hydration status can affect results; results can be unreliable at extremes of body weight. Concern was also raised that bioimpedance may be used by operators who are not aware of these limitations.4 The USPSTF also stated that ‘indirect measures of body fat, such as skinfold thickness, bioelectrical impedance analysis, and waist-hip circumference, have potential for clinical practice, treatment, research, and longitudinal tracking, although there are limitations in measurement validity, reliability, and comparability between measures’.5 The Scottish Intercollegiate Guidelines Network (SIGN) guidelines7 only considered the use of BMI as ‘there is no clear threshold for waist circumference associated with morbidity outcome in children’. However, the strict use of BMI in children can underestimate the prevalence of obesity in young people. McCarthy and coworkers8;9 compared changes in waist circumference and BMI in British youth through cross-sectional surveys in 1977, 1987 and 1997. They found that Obesity: full guidance FINAL VERSION (December 2006) Page 202 FINAL VERSION We were not able to find any other systematic reviews that addressed the accuracy of anthropometric measures or bioimpedance to diagnose obesity compared with the gold standard of BMI in adults. We found several primary studies that assessed the utility of waist circumference and/or waist-to-hip ratio to classify people as obese or overweight compared with classification by BMI.16-26 None of the included studies scored highly when quality assessed (using diagnostic study criteria), as blinding was not done, which was assumed to be a practical problem with this type of measurement. This may have affected the accuracy of the measurements, particularly with waist circumference. However, most studies did report that the assessors were trained, and in some cases, the results were validated. Overall, the utility of other measures compared with BMI varied, particularly with sex and age. In general, the use of measures such as waist circumference or waist-to-hip ratio only would not classify someone as overweight or obese who was not. However, the use of these measures would miss a proportion of people who were at increased risk if assessed using BMI alone. The use of waist-to-hip ratio appeared to be less useful than waist circumference. Since we initially reviewed the evidence, the National Guideline Clearing House has produced a synthesis of guidelines relating to obesity in adults,27 and a comparison of the different recommendations relating to measurement can be seen in Table 5.3. Table 5.3 Comparison of recommendations in the key measures (weight, body mass index (BMI), waist circumference)a ACP (2005) No recommendations offered. ACP refers to the USPSTF guidelines for screening for obesity in adults ACPM (2001) Periodic measurement of BMI (weight in kilograms/height in metres2) is recommended for all adults Obesity: full guidance FINAL VERSION (December 2006) Page 205 FINAL VERSION AGA (2002) A medical evaluation is needed to identify patients who either have, or are at risk for, obesity-related medical complications. This assessment should include a careful history, physical examination (including determination of BMI), and laboratory tests to identify eating and activity behaviours, weight history and previous weight loss attempts, obesity-related health risks, and current obesity- related medical illnesses BWH (2003) BMI. The BMI is the recommended approach for assessing body size in the clinical setting, providing a more accurate measure of body size than weight alone. However, it can overestimate body fat in people who are very muscular, very short, or who have oedema, and it underestimates it in people who have lost muscle mass, such as the elderly. Waist circumference. Excess abdominal fat carries particularly elevated health risks. Waist circumference is the most practical marker of abdominal fat. (Many patients understand this concept as ‘apple’ versus ‘pear’ shaped.) A waist circumference greater than 88 cm (> 35 inches) raises cardiovascular disease risk in women Ethnic and age-related variations in distribution of body fat affect the predictive value of waist circumference. Waist circumference may be a better indicator of risk than BMI for estimating obesity- related disease risk among certain populations, such as Asian– Americans and older people. Waist cut-offs designed for the general population may not apply to very short women (under 1.5 m [5 feet]) Obesity: full guidance FINAL VERSION (December 2006) Page 206 FINAL VERSION Singapore MOH (2004) BMI is the recommended index to define overweight and obesity. It is minimally correlated with height and highly correlated with body fat percentage and levels of disease risk of comorbidities. Body weight alone can be used to follow weight loss and to determine efficacy of therapy (grade B, level III) Waist circumference is the most practical anthropometric measurement for assessing a patient's abdominal fat content before and during weight loss treatment. Gender-specific waist circumference cut-offs should be used in conjunction with BMI to identify increased disease risk (grade B, level III) USPSTF (2003) The USPSTF found good evidence that BMI, calculated as weight in kilograms divided by height in metres squared, is reliable and valid for identifying adults at increased risk for mortality and morbidity due to overweight and obesity Central adiposity increases the risk for cardiovascular and other diseases independent of obesity. Clinicians may use the waist circumference as a measure of central adiposity. Men with waist circumferences greater than 102 cm (> 40 inches) and women with waist circumferences greater than 88 cm (> 35 inches) are at increased risk for cardiovascular disease. The waist circumference thresholds are not reliable for patients with a BMI greater than 35 kg/m2 a Adapted from the National Guideline Clearinghouse guideline synthesis on the assessment and treatment of obesity and overweight in adults.27 ACP, American College of Physicians; ACPM, American College of Preventive Medicine; AGA, American Gastroenterological Association; BWH, Brigham and Women’s Hospital; MOH, Ministry of Health; USPSTF, United States Preventive Services Task Force. 5.1.3.3 Effectiveness of opportunistic screening on health outcomes We did not find any guidelines that issued recommendations on the effectiveness of opportunistic screening in the identification of people who are overweight or obese. Obesity: full guidance FINAL VERSION (December 2006) Page 207 FINAL VERSION over a wide range of adiposity in children (Cole and Rolland-Cachera cited by Neovius et al., p10732). The evidence reviews below report how BMI and waist measurements can be used to classify the weights and body shapes of individuals into groups at increased risk of health problems (Table 5.4). Table 5.4 Classification of overweight and obese (BMI) from key references Source Classification Definition and notes Overweight British childhood BMI charts show 91st, 98th and 99.6th centile lines RCPCH/NOF 200233 Obese The 2002 charts show IOTF cut-offs corresponding to adult definitions of overweight and obesity ‘Body mass index (BMI) is the most practical measure of obesity/overweight, provided values are related to reference standards for age’ Overweight > 85th centile (CDC 2002) NHMRC 20034 Obese > 95th centile (CDC 2002) CDC BMI percentile charts recommended for use (in the clinical setting) until local BMI charts are developed (Australia) Overweight ≥ 91st centile (UK 1990) SIGN 20037 Obese ≥ 98th centile (UK 1990) ‘Despite … limitations, there is widespread international support for the use of BMI to define obesity in children, expressed in non- systematic reviews and consensus statements’ Obesity: full guidance FINAL VERSION (December 2006) Page 210 FINAL VERSION Source Classification Definition and notes At risk of overweight BMI between 85th and 95th percentile for age and sex AAP 2003* Overweight or obese BMI ≥ 95th percentile Singapore MOH 2004 ‘BMIs-for-age and gender equivalent to adult WHO BMI cut-offs for obese and overweight (at ≥ 30.0 or ≥ 25.0 kg/m2) respectively can be used as thresholds, although BMI cut-offs for action among Asians of 27.5 kg/m2 and 23.0 kg/m2 respectively may be eventually used’ AHA 2005 Overweight ≥ 95th percentile (CDC age- and sex-specific nomograms for BMI) ‘By late adolescence, these percentiles approach those used for adult definitions; the 95th percentile is approximately 30 kg/m2’ RNAO 2005 Overweight BMI > 85th percentile and < 95th percentile ‘Research studies often use recommended international cut-offs corresponding to a BMI of 25-29.9 used in adults’ * The AAP 2003, Singapore MOH 2004, AHA 2005, RNAO 2005 and USPSTF 2005 are all adapted from the National Guideline Clearinghouse guideline synthesis on the assessment and treatment of obesity and overweight in children and adolescents. Obesity: full guidance FINAL VERSION (December 2006) Page 211 FINAL VERSION Source Classification Definition and notes Obese BMI for age and sex above 95th percentile using CDC growth curves ‘There is no direct measure of body fat in childhood that is readily applicable in the clinical setting …. A new international cut-off for BMI which corresponds to the adult levels of 25 and 30 for overweight and obesity respectively are recommended for population studies’ At risk of overweight BMI between the 85th and 94th percentile for age and sex USPSTF 2005 Overweight Overweight as a BMI at or above the 95th percentile for age and sex ‘BMI percentile for age and sex is the preferred measure for detecting overweight in children and adolescents because of its feasibility, reliability, and tracking with adult obesity measures’ AAP, American Academy of Pediatrics; AHA, American Heart Association; CDC, Centers for Disease Control and Prevention; MOH, Ministry of Health; NOF, National Obesity Forum; NHMRC, National Health and Medical Research Council (Australia); RCPCH, Royal College of Paediatrics and Child Health; RNAO, Registered Nurses Association of Ontario; SIGN; Scottish Intercollegiate Guidelines Network; USPSTF, United States Preventive Services Task Force. In children, weight must be adjusted for height. These adjustments are made by comparing the child’s measurements with reference standards. BMI varies with body proportions, age and puberty status. To assess individual children, measurements need to be adjusted to compare them with those of other children of the same age. Obesity: full guidance FINAL VERSION (December 2006) Page 212 FINAL VERSION used, compared to use of the international reference data. Sensitivity of the definition of obesity using the international reference data differed significantly between the sexes, with low sensitivity in girls and extremely low sensitivity in boys. International BMI cut offs for BMI in children have not been related to obesity related morbidity in childhood.’37 ‘They require further testing, with evidence of external validity, before they are adopted.’36 The Growth Reference Review Group, a working group convened by the Royal College of Paediatrics and Child Health, published a review of growth reference charts for use in the UK.35 The Group considered the data on which the references were based and their current validity, and made recommendations about which reference was to be used in defined settings. Viner and Nicholls38 made clear their use of the IOTF cut-offs to identify obesity. As there is no accepted definition of obesity they considered those with a BMI of greater or equal to 3 standard deviations (SD) above the mean (≥ 99.86th centile) as extremely obese and at potential high risk. Moreover, they acknowledged the use of waist circumference as an additional indicator of potentially high risk of abdominal obesity.38 Cole and coworkers39 aimed to identify the best possible BMI measure for change (BMI, BMI%, BMI z-score or BMI centile) for children across a range of adiposity. To do so, they measured BMI three times over a period of 9 months in 135 Italian preschool children aged 29–68 months. The authors concluded that BMI centile is (i) useful for classifying children’s adiposity, although poor at quantifying change in adiposity and (ii) sensitive to changes in the middle of the adiposity range but insensitive to changes at the extremes. BMI z-score is also useful for assessing adiposity cross-sectionally, and, unlike BMI centile it can be summarised across populations for statistical purposes. Despite these, disadvantages appear as its variability gets progressively smaller the more obese the child.39 Obesity: full guidance FINAL VERSION (December 2006) Page 215 FINAL VERSION Cole and co-workers also analysed percentage change in BMI, stating that it performs better than BMI centile or z-score. They stated that, in practice, adiposity change over time is virtually equivalent when measured either with percentage change BMI or BMI. Thus, both can be used interchangeably. To conclude, Cole and co-workers contended that adiposity change should be measured in BMI (kg/m2) or BMI (%). Nevertheless they acknowledged that this should be qualified, as the adiposity measures for change over time are all highly associated and the advantage of BMI or BMI% over BMI z-score is tenuous.39 In 2002, the ‘Health survey for England’40 focused on the health of children and young people, and on the health of infants (aged under 1 year) and their mothers. One of the ‘core topics’, which is included in all health surveys, was anthropometry. Emmanuel Stamatakis produced a chapter for ‘Health survey for England’ on the anthropometric measurement of overweight and obesity in children.41 He discussed the establishment of a standard definition for child overweight and obesity using BMI reference data from six different countries around the world.42 This linked childhood and adult obesity/overweight standards using evidence of clear associations between the adult BMI cut-off values of 25 kg/m2 and 30 kg/m2 and health risk. However, Stamatakis reported that a re-analysis of children’s BMI data using similar methods to the international classification but UK-only reference data showed that the international BMI cut-offs exaggerated the differences in overweight and obesity prevalence between boys and girls by underestimating prevalence in boys. Other possible limitations of the international classification included concerns about its sensitivity (ability to identify all obese children as obese), the limited sample size of the reference population and the lack of BMI cut-off points for underweight. However, in summary, the report concluded that ‘the issue of childhood obesity definition is far from resolved and there is an urgent need for further work’.41 5.1.4.2 Classification of overweight and obesity in adults This section describes how BMI and waist measurements can be used to classify the weights and body shapes of individuals into groups at increased risk of health problems (Tables 5.5 and 5.6). Obesity: full guidance FINAL VERSION (December 2006) Page 216 FINAL VERSION Table 5.5 Classification of overweight and obese (body mass index [BMI]a) from key references Source Classification (BMIa) Adult Overweight ≥ 25 NOF 200243 Obese ≥ 30 Overweight ≥ 25 ≥ 30 NHMRC 200313 Obese (≥ 40 severely obese) Overweight ≥ 25 NIH 199814 Obese ≥ 30 a BMI unit of measurement: kg/m2. NOF, National Obesity Forum; NHMRC, National Health and Medical Research Council (Australia); NIH, National Institutes of Health. Obesity: full guidance FINAL VERSION (December 2006) Page 217 FINAL VERSION Obesity: full guidance FINAL VERSION (December 2006) Page 220 Singapore MOH (2004) Current World Health Organization and international guidelines recommend BMI cut-offs of 25 kg/m2 and 30 kg/m2 to define overweight and obesity, respectively. Based on body fat equivalence and comorbid disease risk, BMIs of 23 kg/m2 and 27.5 kg/m2, respectively have been recommended as cut-off points for public health action in Asians (grade C, level IV). Note: BMI cut-off points are currently being reviewed in the light of new data Current international guidelines recommend waist circumference cut-offs of 102 cm and 88 cm to define excess risk in males and females, respectively. Based on an Asian-Pacific consensus and our national health survey and comorbid disease risk, cut-offs of 90 cm and 80 cm, respectively, are probably more appropriate for Asians (grade C, level IV) USPSTF (2003) The USPSTF found good evidence that BMI, calculated as weight in kilograms divided by height in metres squared, is reliable and valid for identifying adults at increased risk for mortality and morbidity due to overweight and obesity Persons with a BMI between 25 kg/m2 and 29.9 kg/m2 are overweight, and those with a BMI of > 30 kg/m2 are obese. There are three classes of obesity: class I (BMI 30–34.9), class II (BMI 35–39.9) and class III (BMI 40 and above). Men with waist circumferences > 102 cm (> 40 inches) and women with waist circumferences > 88 cm (> 35 inches) are at increased risk for cardiovascular disease. The waist circumference thresholds are not reliable for patients with a BMI > 35 kg/m2 ACPM, American College of Preventive Medicine; AGA, American Gastroenterological Association; BMI, body mass index; BWH, Brigham and Women’s Hospital; MOH, Ministry of Health; USPSTF, United States Preventive Services Task Force. FINAL VERSION BMI There is little disagreement about the classification of overweight and obese using BMI in adults; a BMI between 18.5 kg/m2 and under 25 kg/m2 is accepted to be within normal ranges, whereas a BMI of between 25 kg/m2 and under 30 kg/m2 is classified as overweight and a BMI of 30 kg/m2 and over as obesity. Further classifications, linked with morbidity, can be seen in Table 5.8. Table 5.8 Classifications of obesity44 Classification BMI (kg/m2) Risk of co-morbidities Underweight < 18.5 Lowa Healthy weight 18.5– 24.9 Average Overweight (or pre-obese) 25–29.9 Increased Obesity, class I 30–34.9 Moderate Obesity, class II 35–39.9 Severe Obesity, class III ≥ 40- Very severe a Other health risks may be associated with low body mass index (BMI)., These cut-offs are based on epidemiological evidence of the link between mortality and BMI in adults. Waist circumference and waist-to-hip ratio This agreement on classification is also reflected in the cut-offs used for waist circumference: a waist circumference of 102 cm or over in men and 88 cm or over in women is associated with substantially increased health risks (Table 5.9). Table 5.9 Classification using waist-to-hip ratio and waist circumference44;45 At increased risk Men Women Waist-to-hip ratio > 1.0 > 0.85 Waist circumference (increased risk) ≥ 94 cm ≥ 80 cm Waist circumference (greatly increased risk) ≥ 102 cm ≥ 88 cm Obesity: full guidance FINAL VERSION (December 2006) Page 221 FINAL VERSION BMI and waist circumference The WHO recommended that an individual’s relative risk could be more accurately classified using both BMI and waist circumference. These can be seen in Table 5.10. Table 5.10 Combining body mass index (BMI) and waist measurement to classify the risks of type 2 diabetes and cardiovascular disease13;44 Waist circumference (cm) Classification BMI (kg/m2) Men 94–102 > 102 Women 80–88 > 88 Underweight < 18.5 – – Healthy weight 18.5–24.9 – Increased Overweight 25–29.9 Increased High Obesity > 30 High Very high The Agency for HealthCare Research and Quality (AHRQ) undertook a systematic review of the diagnosis and treatment of obesity in older people.15 The review addressed the following questions: Are there limitations in diagnosing obesity in the elderly† with BMI? Should another measurement be used with BMI or in place of BMI for diagnosing obesity in the elderly? The review concluded that: ‘Overall, among office-based diagnostic tests for obesity, BMI and WC showed very similar correlation with body fat percentage in men and women.… While WC correlates closely with body fat percentage and aims to measure central adiposity, it showed low sensitivity when used as a single tool to identify older patients with either generalized (by BMI) or central (by WHR) obesity. Gender did not appear to strongly affect these analyses’ diagnostic accuracy, but the utility of diagnostic measures may differ across ethnic/racial groups.’ † Defined as people aged 60 years or older. Obesity: full guidance FINAL VERSION (December 2006) Page 222 FINAL VERSION UK evidence No studies investigating ethnicity differences in the association of proxy measures of obesity with morbidity in children in UK populations were found. However, there is evidence that young adult South Asians tend to have greater truncal adiposity than their European counterparts. One study developed body mass reference curves based on a representative sample of the UK population from birth to 23 years.59 However, it was not stated if ethnicity was considered in ensuring the sample was representative. International evidence The WHO review of obesity in the Asia-Pacific region published in 2002 stated that the international standard for BMI-for-age chart42 was unlikely to be appropriate for Asian and Pacific children.60 Summary Some evidence appears to suggest that Afro-Caribbean and black African girls might be at greater risk of overweight and obesity. This is also observed in some Indian boys. Evidence also suggests that Indian children have higher body mass adjusted pressure levels than white children, and are predisposed to IRS, which is associated with excess body fat, abdominal adiposity and excess truncal subcutaneous fat. 5.1.5.2 Classification of obesity in adults from different ethnic groups Background It is now generally accepted that the different ethnic groups have higher cardiovascular and metabolic risks at lower BMIs, and this may be because of differences in body shape and fat distribution. In 2001, an international meeting of researchers discussed the simplified use of anthropometry to assess the risk of chronic disease associated with overweight and body fat distribution in adults.61 The researchers concluded that: ‘for its potentially important role in health promotion and primary health care activities, WC [waist circumference] should be adopted as Obesity: full guidance FINAL VERSION (December 2006) Page 225 FINAL VERSION a valuable tool for assessing the health risks of overweight, provided that appropriate cut-off points are established’.61 (Our emphasis) Although ethnicity was discussed, the main groups were those not directly applicable to the UK. Although the UK data included in the pooled evidence presented at the meeting did include people of South Asian ancestry, no detailed discussion of this group was reported. The concept of different cut-offs for different ethnic groups has also been proposed by the WHO, but there is ongoing debate62-66 and at present, there are no commonly accepted cut-offs or indeed, methods to determine specific cut- offs.66;67 However, research is currently being undertaken,68 and any update of this guidance will consider this new evidence as appropriate. For this guidance, we have therefore looked for evidence on how different cut-offs are associated with mortality and morbidity in ethnic populations (appropriate to the UK) both in the UK and in the countries of origin. The Newcastle Heart project69 compared coronary heart disease (CHD) risk factors in Indians, Pakistanis and Bangladeshis, and also compared South Asians (as a group) with people of European origins. The participants were aged between 20 and 74 years, and lived in Newcastle, UK. Measurements included biochemical markers (including fasting insulin, lipids, blood glucose) and anthropometry, and other clinical factors (including blood pressure and electrocardiograms). Another aim of the project was to determine the association between ethnic and socioeconomic inequalities, physical activity, social networks and cardiovascular risk factors.70-74 The authors reported (in several papers) that: The risk of CHD was not uniform among South Asians but that, overall, South Asians had a higher level of CHD than Europeans.69 South Asians did not appear to have higher levels of lipoprotein (a) levels (which, in combination with high insulin resistance, was hypothesised to explain the increased level of heart disease).74 Obesity: full guidance FINAL VERSION (December 2006) Page 226 FINAL VERSION South Asians had lower levels of habitual physical activity than Europeans, and this was likely to contribute to the higher levels of diabetes and cardiovascular risk.72 The authors suggested that for South Asians living in Newcastle, the European pattern of inequalities (where social class, education and deprivation were associated with disease and risk factors) were becoming established, with different rates of establishment occurring in different ethnic groups.70 When different models of predicting cardiovascular disease were applied to the different ethnic groups, a variety of results were seen. However, overall, the authors concluded that ‘the potential gains from controlling major established risk factors could be substantial in South Asians and greater than in Europeans’.75 There remains uncertainty about how ethnic, migrant populations may or may not adapt over time to the patterns of risk of the indigenous population. Lean and co- workers compared anthropometric measures and behavioural associations in migrant and British-born South Asians and Italians and the general population of British women living in the west of Scotland. No differences were found in anthropometry between the British-born South Asian women and the general population women. The authors concluded that these results offered ‘hope that some of the high cardiovascular risks in South Asians in Britain may be overcome by lifestyle modification, and that the risks may reduce over generations through acculturation’.76 The influence of maternal nutrition, birth weight and initial weight gain on future health and risk of obesity in adulthood is also unclear (although evidence is emerging).77-79 Methods The evidence review was based on relevant, identified systematic reviews and primary studies assessing whether the association between BMI and waist circumference and morbidity is different between different ethnic groups in UK populations. This review considered only Asian and black populations. Due to the lack of evidence and ongoing international debate on this topic, we asked Obesity: full guidance FINAL VERSION (December 2006) Page 227 FINAL VERSION On the basis of available data in Asia, Asians generally have a higher percentage of body fat than white people of the same age, sex and BMI. The proportion of Asian people with risk factors for type 2 diabetes and cardiovascular disease was substantial even below the existing cut-off point of 25 kg/m2. Current (WHO) cut-off points do not therefore provide an adequate basis for taking action on risks related to overweight and obesity in many populations in Asia. However, the available data do not necessarily indicate one clear BMI cut-off point for all Asian population groups for overweight or obesity. Cut-offs for observed risk varied from 22 kg/m2 to 25 kg/m2 and for high risk from 26 kg/m2 to 31 kg/m2. Two key recommendations were as follows: Trigger points for public health action should be 23 kg/m2 (increased risk) and 27.5 kg/m2 (high risk). Where possible, in populations with a predisposition to central obesity and related increased risk of developing the metabolic syndrome, waist circumference should also be used to refine action levels on the basis of BMI. There is some limited evidence that for a given BMI or waist circumference, morbidity risk in South Asian populations (of Pakistani, Bangladeshi and Indian origin) resident in the UK may be higher. Black population UK evidence Only one UK-based study was found that investigated the measurement of obesity in the male black population.80 However, the focus of this study was on differences in the relation of central obesity with cardiovascular risk, insulin resistance and diabetes prevalence between European and South Asian populations. The sample size of the Afro-Caribbean group was considerably Obesity: full guidance FINAL VERSION (December 2006) Page 230 FINAL VERSION smaller (European = 1515, South Asian = 1421, Afro-Caribbean = 209). Unlike the European and South Asian groups, data on risk factors for the Afro-Caribbean group were not controlled for waist-to-hip ratio so it is difficult to say whether the findings have a bearing on appropriate cut-offs for Afro-Caribbeans. However, a general comparison of the Afro-Caribbean population sample with the European population found that: waist-to-hip ratio was not significantly different but BMI was significantly higher diabetes prevalence was significantly higher but serum insulin levels were not significantly different median systolic and diastolic blood pressures were significantly higher plasma triglyceride was significantly lower and HDL cholesterol significantly higher. International evidence Elsewhere, a large-scale study95 of the relation between BMI and body fat in black populations in Nigeria, Jamaica and the USA concluded that within populations bioelectrical impedance analysis as a measure of percentage body fat was not a better predictor of blood pressure, or waist or hip circumference. However, for similar levels of BMI, body fat varied substantially. Nigerians had a greater fat-free mass than Jamaicans and Jamaicans had a greater proportion than African Americans. The study did not make comparisons with white populations. A smaller-scale study96 set in the USA compared the association between upper body obesity and cardiovascular and diabetic risk in white and black pre- menopausal women. This found that upper body obesity (as assessed by waist- to-hip ratio) is not as potent a risk factor for diabetes and coronary heart disease in black women as it is in white women. Also, whereas in white women upper body obesity was associated with significantly greater glucose intolerance, hyperinsulinaemia and insulin resistance, this was not significant in black women Obesity: full guidance FINAL VERSION (December 2006) Page 231 FINAL VERSION (that is, upper body fat distribution has less impact on carbohydrate metabolism). The sample size for this study was small (black women = 22, white women = 20). Summary In summary, the evidence base for differences in the association of BMI, waist circumference and bioimpedance with morbidity is limited, particular in black ethnic groups. Available evidence for South Asian groups was consistent with findings from studies in populations living in South Asia. This may not be surprising as UK studies have focused particularly on first generation migrants. The findings on South Asian populations in the UK were consistent with those from the WHO expert consultation which assessed populations living in South Asia (although these also included a wider range of populations). Therefore, there is probably insufficient evidence to make any clear recommendations about separate cut-offs for ethnic groups in the UK, as distinct to the cut-offs recommended for Asian populations by the WHO. Obesity: full guidance FINAL VERSION (December 2006) Page 232 FINAL VERSION evidence tables. An additional 561 guidelines were identified, of which 44 were assessed in detail and of which 14 met the criteria for inclusion (5 clinical practice guidelines, 2 recommendation statements, 4 policy statements, 2 reports and 1 briefing paper). The inclusion and exclusion criteria for the review adhered to the standard public health review parameters for interventions. In addition, it was agreed with the Guidance Development Group (GDG) that studies should only be included in this review if: the paper reports an intervention to identify adults and/or children who are potentially at risk for developing obesity and who would benefit from participation in a prevention/public health intervention to manage weight the paper is a recommendation or guideline for identifying adults and/or children who are potentially at risk for developing obesity and who would benefit from participation in a prevention/public health intervention to manage weight the paper concludes that the tools evaluated have the potential for use in identification interventions. Tooth et al 2005107 was used to appraise observational longitudinal studies and the AGREE instrument (www.agreecollaboration.org/instrument) was used to appraise guidelines and recommendation documents. For the purposes of this review a clinical or practice guideline was defined as a document that aimed to identify, summarise and evaluate the best evidence and was based on a systematic review of the current research evidence. Public/policy statements and recommendations were defined as documents that aimed to provide advice or recommendations and were likely to have been developed based on consensus agreement by an expert panel. Please note that the Department of Health (DH) has recently issued ‘Measuring childhood obesity: guidance to primary care trusts’. However as this was published in January 2006 (that is, after the agreed search dates) it Obesity: full guidance FINAL VERSION (December 2006) Page 235 FINAL VERSION has not been appraised for this version of the review. The DH guidance will be appraised before final publication of this NICE guidance. 5.2.3 Identification of individuals who may benefit from participation in public health interventions to manage weight There is limited evidence on the effectiveness of interventions to identify children and adults who are likely to become overweight or obese and would benefit from interventions. This is particularly the case in adults. All studies had some confounders. Only one study101 was carried out in the UK. No UK- based corroborative data were identified other than one accelerometer study, but it is likely that the findings are applicable to the UK. No cost-effectiveness data were identified. 5.2.3.1 Children Eight moderate quality observational longitudinal studies in children97–104 suggest that those at risk of becoming overweight or obese may be identified from opportunistic monitoring of growth charts after 2 years of age (including larger than expected weight gain and early ‘rebound’),98,99 potentially using anthropometric measures in addition to BMI (height and waist circumference) and from assessing measures of habitual activity levels (for example, through an accelerometer) and diet.101 Attempting to identify children at risk before 2 years of age had poor predictability.97 Of the four studies measuring anthropometric measures, one102 concluded that measurement of waist circumference at age 8 may be a promising index to predict overweight at puberty, and two linked studies103,104 concluded that a measurement of height at age 7–8 could be used to identify more accurately children who are likely to become overweight adults, although this may only be true for those children already overweight. A further study,100 which measured anthropometric measures and examined lifestyle factors such as diet and physical activity, concluded that large-scale involvement of primary schools in screening programmes could identify those children at risk of being overweight and obese in adulthood and for whom strategies to prevent overweight and obesity would be most effective. No studies were found which considered identifying children by their parent’s weight/obesity. Obesity: full guidance FINAL VERSION (December 2006) Page 236 FINAL VERSION 5.2.3.2 Adults Two studies with some confounders,105,106 one large retrospective cohort study and one relatively small ongoing prospective study, examined interventions to identify adults at risk of overweight and obesity. The results suggest that considering an individual’s weight history (for example, previous weight gain or loss, previous attempts at dieting) and monitoring more recent weight gain (for example over 2.3 kg) may help identify adults at risk of becoming overweight or obese in future. 5.2.4 Existing guidelines and recommendations There is currently no available formal guidance in the UK and there is a lack of consensus in the existing ‘recommendation’ papers on whether to regularly monitor or screen BMI, particularly in children. No corroborative data were identified, but it is likely that the findings are applicable to the UK. No cost data were identified. 5.2.4.1 UK-based guidance and recommendations No usable UK guidelines were identified. Following the advice of SIGN, the existing SIGN guidelines for adults108 were not considered due to methodological problems. These guidelines will be updated in the future. SIGN guidance for children7 was considered, but excluded as it discussed identification of overweight and obesity only. The conclusions of three UK recommendation papers suggest that there is currently no consensus available for the screening of children for unhealthy weight gain. A policy statement form the UK’s National Screening Committee in 2005,109 based on expert consensus opinion, recommended that screening should not be offered whereas the evidence from a briefing paper prepared by the Child Growth Foundation110 firmly recommended universal serial BMI assessments for children at least until the end of primary school. One further report from the House of Commons Select Committee on Health in 2004,111 supported the guidance suggested by the Child Growth Foundation and suggested that BMI measures should be recorded annually for school-aged children. Obesity: full guidance FINAL VERSION (December 2006) Page 237 FINAL VERSION 5.3 Review limitations No review level or controlled trial evidence was found for this review question, resulting in an evidence base of observational longitudinal studies. Obesity: full guidance FINAL VERSION (December 2006) Page 240 FINAL VERSION Reference List: (1) Cole TJ, Rolland-Cachera MF. Measurement and definition. In: Burniat W, Cole TJ, Lissau I, Poskitt EME, editors. Child and Adolescent Obesity. Causes and consequences, prevention and management. Cambridge: Cambridge University Press; 2002. 3-27. (2) Yusuf S, Hawken S, Ounpuu S, Bautista L, Franzosi MG, Commerford P et al. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case-control study. Lancet 2005; 366(9497):1640-1649. (3) Mamtani MR, Kulkarni HR. Predictive performance of anthropometric indexes of central obesity for the risk of type 2 diabetes. Arch Med Res 2005; 36(5):581-589. (4) Sorensen G, Stoddard A, Peterson K, Cohen N, Hunt MK, Stein E et al. Increasing fruit and vegetable consumption through worksites and families in the treatwell 5-a- day study. AM J PUBLIC HEALTH 1999; 89(1):54-60. (5) Agency for Healthcare Research and Quality, National Guideline Clearinghouse (NGC) Guideline Synthesis. OVERWEIGHT AND OBESITY IN CHILDREN AND ADOLESCENTS:ASSESSMENT, PREVENTION, AND MANAGEMENT. 2005. (6) Freedman DS, Wang J, Maynard LM, Thornton JC, Mei Z, Pierson J et al. Relation of BMI to fat and fat-free mass among children and adolescents. International Journal of Obesity 2005; 29(1):1-8. (7) Scottish Intercollegiate Guidelines Network. Management of obesity in children and young people: a national clinical guideline. 69. 2003. (8) McCarthy HD, Ellis SM, Cole TJ. Central overweight and obesity in British youth aged 11-16 years: cross-sectional surveys of waist circumference. British Medical Journal 2003; 326. (9) McCarthy HD, Jarrett KV, Emmett PM, Rogers I. Trends in waist circumferences in young British children: A comparative study. International Journal of Obesity 2005; 29(2):157-162. (10) Rudolf MCJ, Greenwood DC, Cole TJ, Levine R, Sahota P, Walker J et al. Rising obesity and expanding wasitlines in schoolchildren: a cohort study. Arch Dis Child 2004; 89:235. (11) Singapore Ministry of health. Obesity. 2004. t (12) Neovius M, Linne Y, Rossner S. BMI, waist-circumference and waist-hip-ratio as diagnostic tests for fatness in adolescents. International Journal of Obesity 2005; 29(2):163-169. (13) National Health & Medical Research Council. Clinical practice guidelines for the management of overweight and obesity in adults. 2003. t (14) Glasgow RE, Terborg JR, Hollis JF, Severson HH, Boles SM. Take heart: results from the initial phase of a work-site wellness program. AM J PUBLIC HEALTH 1995; 85(2):209-216. (15) Cockcroft A, Gooch C, Ellinghouse C, Johnston M, Michie S. Evaluation of a programme of health measurements and advice among hospital staff. Occup Med (Oxf) 1994; 44(2):70-76. Obesity: full guidance FINAL VERSION (December 2006) Page 241 FINAL VERSION (16) Dasgupta S, Hazra SC. The utility of waist circumference in assessment of obesity. Indian Journal of Public Health 1999; 43(4):132-135. (17) Gill T, Chittleborough C, Taylor A, Ruffin R, Wilson D, Phillips P. Body mass index, waist hip ratio, and waist circumference: Which measure to classify obesity? Sozial- und Praventivmedizin 2003; 48(3):191-198. (18) Goodman-Gruen D, Barrett-Connor E. Sex differences in measures of body fat and body distribution in the elderly. Am J Epidemiol 1996; 143(9):898-906. (19) Han TS, Seidell JC, Currall JE, Morrison CE, Deurenberg P, Lean ME. The influences of height and age on waist circumference as an index of adiposity in adults. Int J Obes Relat Metab Disord 1997; 21(1):83-89. (20) Lean ME, Han TS, Morrison CE. Waist circumference as a measure for indicating need for weight management. British Medical Journal 1995; 311(6998):158-161. (21) Molarius A, Seidell JC, Sans S, Tuomilehto J, Kuulasmaa K. Varying sensitivity of waist action levels to identify subjects with overweight or obesity in 19 populations of The WHO MONICA Project. Journal of Clinical Epidemiology 1999; 52(12):1213- 1224. (22) Molarius A, Seidell JC, Sans S, Tuomilehto J, Kuulasmaa K. Waist and hip circumferences, and waist-hip ratio in 19 populations of the WHO MONICA Project. Int J Obes Relat Metab Disord 1999; 23(2):116-125. (23) Molarius A, Seidell JC, Visscher TL, Hofman A. Misclassification of high-risk older subjects using waist action levels established for young and middle-aged adults-- results from the Rotterdam Study. Journal of the American Geriatrics Society 2000; 48(12):1638-1645. (24) Sonmez K, Akcakoyun M, Akcay A, Demir D, Duran NE, Gencbay M et al. Which method should be used to determine the obesity, in patients with coronary artery disease? (body mass index, waist circumference or waist-hip ratio). Int J Obes Relat Metab Disord 2003; 27(3):341-346. (25) Taylor RW, Keil D, Gold EJ, Williams SM, Goulding A. Body mass index, waist girth, and waist-to-hip ratio as indexes of total and regional adiposity in women: Evaluation using receiver operating characteristic curves. AM J CLIN NUTR 1998; 67(1):44-49. (26) Logue E, Smucker WD, Bourguet CC. Identification of obesity: waistlines or weight? Nutrition, Exercise, and Obesity Research Group. Journal of Family Practice 1995; 41(4):357-363. (27) Guideline synthesis: assessment and treatment of obesity and overweight in adults. National Guideline Clearinghouse (NGC) . 2005. t (28) Harvey EL, Glenny A, Kirk SF, Summerbell CD. Improving health professionals' management and the organisation of care for overweight and obese people. Cochrane Database of Systematic Reviews 2001;(2):CD000984. (29) Simkin-Silverman LR, Wing RR. Management of obesity in primary care. Obesity Research 1997; 5:603-612. (30) Gambera PJ, Schneeman BO, Davis PA. Use of the food guide pyramid and US dietary guidelines to improve dietary intake and reduce cardiovascular risk in active- duty Air Force members. J AM DIET ASSOC 1995; 95(11):1268-1273. (31) Quality and outcomes framework guidance. 2006. London, BMA. t Obesity: full guidance FINAL VERSION (December 2006) Page 242 FINAL VERSION (64) Stevens J. Ethnic-specific revisions of body mass index cutoffs to define overweight and obesity in Asians are not warranted. Int J Obes Relat Metab Disord 2003; 27(11):1297-1299. (65) Stevens J. Ethnic-specific cutpoints for obesity vs country-specific guidelines for action. Int J Obes Relat Metab Disord 2003; 27(3):287-288. (66) James WP. Assessing obesity: are ethnic differences in body mass index and waist classification criteria justified? Obes Rev 2005; 6(3):179-181. (67) Deurenberg P, Deurenberg-Yap M. Validity of body composition methods across ethnic population groups. Forum Nutr 2003; 56:299-301. (68) Huxley R, Barzi F, Stolk R, Caterson I, Gill T, Lam TH et al. Ethnic comparisons of obesity in the Asia-Pacific region: protocol for a collaborative overview of cross- sectional studies. Obes Rev 2005; 6(3):193-198. (69) Bhopal R, Unwin N, White M, Yallop J, Walker L, Alberti KG et al. Heterogeneity of coronary heart disease risk factors in Indian, Pakistani, Bangladeshi, and European origin populations: cross sectional study. BMJ 1999; 319(7204):215-220. (70) Bhopal R, Hayes L, White M, Unwin N, Harland J, Ayis S et al. Ethnic and socio- economic inequalities in coronary heart disease, diabetes and risk factors in Europeans and South Asians. J Public Health Med 2002; 24(2):95-105. (71) Fischbacher CM, Bhopal R, Blackwell CC, Ingram R, Unwin NC, White M et al. IgG is higher in South Asians than Europeans: does infection contribute to ethnic variation in cardiovascular disease? Arterioscler Thromb Vasc Biol 2003; 23(4):703-704. (72) Hayes L, White M, Unwin N, Bhopal R, Fischbacher C, Harland J et al. Patterns of physical activity and relationship with risk markers for cardiovascular disease and diabetes in Indian, Pakistani, Bangladeshi and European adults in a UK population. J Public Health Med 2002; 24(3):170-178. (73) Pollard TM, Carlin LE, Bhopal R, Unwin N, White M, Fischbacher C. Social networks and coronary heart disease risk factors in South Asians and Europeans in the UK. Ethn Health 2003; 8(3):263-275. (74) Tavridou A, Unwin N, Bhopal R, Laker MF. Predictors of lipoprotein(a) levels in a European and South Asian population in the Newcastle Heart Project. Eur J Clin Invest 2003; 33(8):686-692. (75) Bhopal R, Fischbacher C, Vartiainen E, Unwin N, White M, Alberti G. Predicted and observed cardiovascular disease in South Asians: application of FINRISK, Framingham and SCORE models to Newcastle Heart Project data. J Public Health (Oxf) 2005; 27(1):93-100. (76) Lean ME, Han TS, Bush H, Anderson AS, Bradby H, Williams R. Ethnic differences in anthropometric and lifestyle measures related to coronary heart disease risk between South Asian, Italian and general-population British women living in the west of Scotland. Int J Obes Relat Metab Disord 2001; 25(12):1800-1805. (77) Fall CH, Yajnik CS, Rao S, Davies AA, Brown N, Farrant HJ. Micronutrients and fetal growth. J Nutr 2003; 133(5 Suppl 2):1747S-1756S. (78) Yajnik CS, Fall CH, Coyaji KJ, Hirve SS, Rao S, Barker DJ et al. Neonatal anthropometry: the thin-fat Indian baby. The Pune Maternal Nutrition Study. Int J Obes Relat Metab Disord 2003; 27(2):173-180. Obesity: full guidance FINAL VERSION (December 2006) Page 245 FINAL VERSION (79) Fall CH, Pandit AN, Law CM, Yajnik CS, Clark PM, Breier B et al. Size at birth and plasma insulin-like growth factor-1 concentrations. Arch Dis Child 1995; 73(4):287- 293. (80) McKeigue PM, Shah B, Marmot MG. Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancet 1991; 337:382-386. (81) Patel S, Unwin N, Bhopal R, White M, Harland J, Ayis SAM et al. A comparison of proxy measures of abdominal obesity in Chinese, European and South Asian adults. Diabetic Medicine 1999; 16(10):853-860. (82) Lovegrove JA, Brady LM, Lesauvage SV, Lovegrove SS, Minihane AM, Williams CM. Lack of association between central adiposity and lipaemia in UK Sikh men. Int J Obes Relat Metab Disord 2003; 27(11):1373-1382. (83) Bose K. A comparative study of generalised obesity and anatomical distribution of subcutaneous fat in adult White and Pakistani migrant males in Peterborough. Journal of the Royal Society of Health 1995; . 115(2). (84) Bose K. Generalised obesity and regional adiposity in adult white and migrant Muslim males from Pakistan in Peterborough. Journal of the Royal Society of Health 1996; . 116(3). (85) Williams R, Bhopal R, Hunt K. Health of a Punjabi ethnic minority in Glasgow: a comparison with the general population. J Epidemiol Community Health 1993; 47(2):96-102. (86) Williams R, Bhopal R, Hunt K. Coronary risk in a British Punjabi population: comparative profile of non-biochemical factors. Int J Epidemiol 1994; 23(1):28-37. (87) Dudeja V, Misra A, Pandey RM, Devina G, Kumar G, Vikram NK. BMI does not accurately predict overweight in Asian Indians in northern India. Br J Nutr 2001; 86(1):105-112. (88) Misra A, Athiko D, Sharma R, Pandey RM, Khanna N. Non-obese hyperlipidemic Asian northern Indian males have adverse anthropometric profile. Nutr Metab Cardiovasc Dis 2002; 12(4):178-183. (89) Misra A, Pandey RM, Sinha S, Guleria R, Sridhar V, Dudeja V. Receiver operating characteristics curve analysis of body fat & body mass index in dyslipidaemic Asian Indians. Indian J Med Res 2003; 117:170-179. (90) Misra A, Vikram NK. Clinical and pathophysiological consequences of abdominal adiposity and abdominal adipose tissue depots. Nutrition 2003; 19(5):457-466. (91) Misra A, Vikram NK. Insulin resistance syndrome (metabolic syndrome) and obesity in Asian Indians: evidence and implications. Nutrition 2004; 20(5):482-491. (92) Vikram NK, Misra A, Pandey RM, Dudeja V, Sinha S, Ramadevi J et al. Anthropometry and body composition in northern Asian Indian patients with type 2 diabetes: receiver operating characteristics (ROC) curve analysis of body mass index with percentage body fat as standard. Diabetes Nutr Metab 2003; 16(1):32-40. (93) Vikram NK, Pandey RM, Misra A, Sharma R, Devi JR, Khanna N. Non-obese (body mass index < 25 kg/m2) Asian Indians with normal waist circumference have high cardiovascular risk. Nutrition 2003; 19(6):503-509. (94) WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004; 363(9403):157-163. Obesity: full guidance FINAL VERSION (December 2006) Page 246 FINAL VERSION (95) Luke A, Durazo-Arvizu R, Rotimi C, Prewitt TE, Forrester T, Wilks R et al. Relation between body mass index and body fat in black population samples from Nigeria, Jamaica, and the United States. Am J Epidemiol 1997; . 145(7). (96) Dowling HJ, Pi-Sunyer FX. Race-dependent health risks of upper body obesity. Diabetes 1993; 42(4):537-543. (97) Toschke AM GVKBvKR. Identifying children at high risk for overweight at school entry by weight gain during the first 2 years. Archives of Pediatrics & Adolescent Medicine 2004; 258(5):449-452. (98) Guo SS, Wu W, Chumlea WC, Roche AF. Predicting overweight and obesity in adulthood from body mass index values in childhood and adolescence. AM J CLIN NUTR 2002; 76(3):653-658. (99) He Q, Karlberg J. Probability of adult overweight and risk change during the BMI rebound period. Obesity Research 2002; 10(3):135-140. (100) Barba G GRCGVARPGCeal. The BRAVO project: screening for childhood obesity in a primary school setting. Nutrition Metabolism & Cardiovascular Diseases 2001; 11(4 suppl):103-108. (101) Metcalf BS, Voss LD, Wilkin TJ. Accelerometers identify inactive and potentially obese children (EarlyBird 3). Archives of Disease in Childhood 2002; 87(2):166-167. (102) Maffeis C GAPAPSTL. Does waist circumference predict fat gain in children? International Journal of Obesity & Related Metabolic Disorders 2001; 25(7):978-983. (103) Freedman DS KKLSSBGS. BMI rebound, childhood height and obesity among adults: the Bogalusa Heart Study. International Journal of Obesity and Related Metabolic Disorders 2001; 25(4):543-549. (104) Freedman DS, Khan LK, Mei Z, Dietz WH, Srinavasan SR, Berenson GS. Relation of childhood height to obesity among adults: the Bogalusa Heart Study. Pediatrics 2002; 109(2):E23-E29. (105) Kroke A, Liese AD, Bergmann MM, Klipstein-Grobusch K, Hoffman K, Boeing H. Recent weight changes and weight cycling as predictors of subsequent two year weight change in a middle-aged cohort. International Journal of Obesity and Related Metabolic Disorders 2002; 26(3):403-409. (106) St Jeor ST BRHMeal. A classification system to evaluate weight maintainers, gainers, and losers. J AM DIET ASSOC 1997; 97:481-488. (107) Tooth L, Ware R, Bain C, Purdie DM, Dobson A. Quality of reporting of observational longitudinal research. Am J Epidemiol 2005; 161(3):280-288. (108) Scottish Intercollegiate Guidelines Network. Obesity in Scotland, integrating prevention with weight management. 1996. Edinburgh, Scotland, Scottish Intercollegiate Guidelines Network. t (109) National Screening Committee. National Screening Committee - obesity screening. 2005. UK, National Screening Committee. (110) Child Growth Foundation, Fry T. Identification and assessment & audit of obesity in children: briefing prepared for the NICE Guidance Development Group. 2004. London, Child Growth Foundation. (111) House of Commons Health Committee. Obesity: third report of session 2003-04. 2004. London, The Stationery Office Limited. Obesity: full guidance FINAL VERSION (December 2006) Page 247