Waist circumference and body mass index in Chinese children: cutoff values for predicting cardiovascular risk factors

Article metrics

  • A Corrigendum to this article was published on 26 February 2007



Body mass index (BMI) and waist circumference (WC) correlate with cardiovascular (CV) risk factors in childhood which track into adulthood. WC provides a measure of central obesity, which has been specifically associated with CV risk factors. Reference standards for WC, and for WC and BMI risk threshold values are not established in Chinese children.


To construct reference percentile charts of WC, establish relationships between WC, BMI and other risk factors, and propose WC and BMI threshold values predictive of CV risk factors in Hong Kong ethnic Chinese children.


Weight, height, waist and hip circumference were measured in 2593 (52% boys, 47% girls) randomly sampled Hong Kong school children aged 6–12 years. In 958 of these and 97 additional overweight children (n=1055), the relationships between WC, BMI, waist/hip and waist/height ratio and six age-adjusted CV risk factors (>85% percentile levels of blood pressure (BP), fasting triglycerides, low-density lipoprotein (LDL) cholesterol, glucose and insulin levels, and <15% percentile levels of high-density lipoprotein (HDL) cholesterol) were studied. Receiver-operating characteristic analysis was employed to derive optimal age-adjusted sex-specific WC and BMI thresholds for predicting these measures of risk.


WC percentiles were constructed. WC correlated slightly more than BMI with CV risk factors and most strongly with insulin and systolic BP, but poorly or not with LDL and glucose. Optimal WC and BMI risk thresholds for predicting four of these six CV risk factors were ca. the 85th percentiles (sensitivities 0.8, specificities 0.87) with age-specific cutoff values in girls/boys from 57/58 to 71/76 cm and 17/18 to 22/23 kg/m2.


These are the first set of WC reference data for Chinese children. WC risk cutoff values are proposed which, despite a smaller waist in Chinese children, are similar to those reported for American children. WC percentiles may reflect population risk.


Obesity is associated with clustering of cardiovascular (CV) risk factors – insulin resistance, dyslipidaemia and hypertension – empirically recognized as ‘The Metabolic Syndrome’ and with CV outcome.1, 2, 3, 4 Clustering of these CV risk factors in children and adolescents tends to track into adult life.5, 6, 7, 8, 9

CV disease appears to be related specifically to intra-abdominal fat (visceral adipose tissue)10, 11 of which waist circumference (WC) provides a good and clinically useful measure,12, 13, 14, 15, 16, 17 supporting the growing view that WC is a better measure of CV risk than body mass index (BMI) in both children and adults.18, 19, 20, 21

Ethnic differences of WC have been described in adults22, 23 but have not been studied in children, although reference data for WC are available from a number of different countries.24, 25, 26, 27, 28, 29, 30, 31 Data for ethnic Chinese children have not previously been reported.

The purposes of the present study were to develop WC percentiles for ethnic Chinese children; to confirm the relationships between WC, BMI and other CV risk factors in these children; and to develop optimal WC and BMI risk threshold percentiles and cutoff values for this population, to compare with reports in other ethnic groups.



The ‘Population Group’ studied comprised 2593 ethnic Chinese children aged 6–13 years inclusive. They were recruited from 90 classes randomly sampled in eight primary schools participating in University-based health promotion activities in different districts of Hong Kong between 2002 and 2004. Letters sent to all children's parents in randomly selected classes via their school teachers resulted in an 82% positive response. The main reason for non-participation was that not all families could attend the schools as requested 1 hour before normal school opening, whereas a small number refused venepuncture. Responders may be considered to be representative of the population studied.

Written consent for physical examination and blood sampling was obtained from the parents. Information about pubertal status was not sought to avoid prejudicing consent to recruitment. The body weight, height, WC and hip circumference data from these 2593 children (the population group) were used to construct reference WC and BMI percentile charts for this population. Nine hundred and fifty-eight children from this Group were randomly selected to have blood pressure (BP) measured and fasting blood samples taken for lipids, glucose and insulin measurements.

In order to increase the number of overweight children in whom WC and BMI were to be related to CV risk factors, data from 97 additional children (from the same community) who were voluntarily participating in a University-based weight control programme were included – randomly selected and with 100% agreement. These formed the ‘Clinic Group’ and their data were pooled with those of the 958 children from the population group, bringing the number of children in whom CV risk factors were measured to 1055 (the ‘CV Risk Evaluation Group’).

The study protocol was approved by the University ethics committee and informed consents were obtained from all participants and their parents.

Anthropometric and BP measurements

BMI and waist/hip and waist/height ratios were derived from measurement of weight, height and waist and hip circumference. Weight was measured to the nearest 0.1 kg and height to the nearest 0.5 cm, wearing light T-shirt and shorts without shoes. WC was measured midway between the lowest rib and the superior border of the iliac crest. Hip circumference was measured at maximal protrusion of the buttocks. The circumferences are given as the mean of two measurements to the nearest 0.1 cm.

BP was measured by sphygmomanometry to the nearest 2 mm Hg, twice in the right arm, seated, after 10 minutes rest, and a third time if the two readings were >4 mm Hg apart. Diastolic pressure was defined as the point of disappearance of Korotkoff sound (5th phase).

The intra-class (within-observer) correlation coefficients, based on pairs of replicate measurements made by the same observer on the same day, were >0.99 for BMI, 0.97 for WC, 0.94 for systolic BP and 0.95 for diastolic BP.

Blood lipids, glucose and insulin levels

Venous blood samples were taken after 12 h fasting for determination of blood lipids, glucose and insulin concentrations. Plasma lipids were assayed enzymatically using the Boehringer Mannheim Hitachi 911 analyzer. Triglyceride (TG) was measured using the lipase/glycerol kinase method. High-density lipoprotein (HDL) cholesterol was measured after phosphotungstate-magnesium precipitation and low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald equation. The intra-assay coefficients of variation (CVs) were 1.9% for total cholesterol, 1.9% for TG and 5.4% for HDL. Plasma glucose was measured by a glucose oxidase method (Diagnostic Chemicals reagents kit, Diagnostic Chemicals Ltd, Prince Edward Island, Canada). The intra-assay CV of glucose was 2% at 6.6 mmol/l. Serum insulin was determined using radioimmunoassay (Pharmacia Diagnostics Lab, Uppsala, Sweden). The lowest limit of detection was <2 μU/ml.

Age- and sex-specific WC percentiles

Reference sex- and age-specific percentile curves, smoothed by the LMS method,32 were constructed for WC and BMI from the data provided by the population group (n=2593).

CV risk factors

CV risk factors included in this study covered the three main parameters of high BP, dyslipidaemia and insulin resistance. They comprised high (i) BP (systolic and/or diastolic), fasting (ii) LDL cholesterol, (iii) TGs, (iv) glucose, (v) insulin and low (vi) fasting HDL cholesterol – see Statistical analysis for derivation of these risk thresholds. TG and insulin were logarithmically transformed to correct for skewness. Clustering of CV risk factors was empirically defined by the presence of either three-or-more or four-or-more of these six conventional risk factors.

Statistical analysis

WC and BMI of the 958 children of the population group in whom blood samples were obtained were regressed with respect to age (with up to a cubic polynomial of terms: age, age2 and age3) using stepwise regression procedures to provide gender-specific regression equations. These provided norms which were then applied also to age-adjust the data from the 97 children in the Clinic Group. The actual WC and BMI of each of the 1055 children in the CV risk evaluation group minus the predicted WC and BMI calculated from the gender-specific regression equations were regarded as the age-adjusted value of WC and BMI. Each CV risk factor value was similarly regressed with respect to age in these 958 children to provide age- and gender-specific regression equations. The standardized residuals of all 1055 children were retained to represent age-adjusted values.

The gender-specific 85th percentile (15th percentile for HDL) for each of the age-adjusted CV risk factors of the 958 blood-tested population group was chosen as the threshold level of ‘high CV risk’ for all children.

Receiver-operating characteristic (ROC) analysis was employed to investigate the diagnostic ability of age-adjusted WC and BMI to identify the presence or absence of risk factor clustering in the CV risk evaluation group (n=1055). The area-under-ROC curve (AUC) was used to give a measure of the global performance of using WC and BMI as effective diagnostic indicators. The value (age-adjusted WC and BMI) corresponding to the nearest point of the ROC curve to the top left-hand corner was chosen as the optimal threshold for predicting clustering of CV risk factors in that it maximizes both sensitivity and specificity (given equal weighting). Sex- and age-specific WC and BMI cutoffs were read from the corresponding percentiles generated using LMS method from the population group (n=2593).22

The relationships between the anthropometric and the CV risk factors in the CV risk evaluation group were examined by age-adjusted partial correlation analysis.

Statistical analyses were performed with SPSS 13.0 (SPSS Inc., Chicago, IL, USA) and LMS programme Version 1.16 (Tim Cole and Huiqi Pan, Institute of Child Health). Data were presented as mean (s.d.). All statistical tests were two-sided and P-value <0.05 was considered statistically significant.

Additional analyses of the data

In view of uncertainties about the nature and definition of the ‘Metabolic Syndrome’ and the differing groupings of risk factors used in different studies, the WC data from the present study were also analysed using (i) the four factors used in the ATP III report (BP, HDL, TG and glucose),33 and (ii) the four factors found to be most strongly associated with WC in the present study.

In addition, for comparison with the closely comparable US Bogalusa study,34 the present data were re-evaluated using the outer quintile (20%) of risk factor percentiles instead of the outer 15%.


The descriptive characteristics of the population group (n=2593, 47% girls, 53% boys) are shown in Table 1. 10.8% of girls and 13.1% of boys were obese (BMI>95th percentile of the Hong Kong Standards).35 Smoothed WC percentile curves are presented in Figure 1.

Table 1 Characteristics of the population group (n=2593, mean (s.d.))
Figure 1

Smoothed percentile curves for WC. (a) Girls (n=1227) and (b) boys (n=1366).

Pearson's correlation coefficients of the six age-adjusted CV risk factors with the four age-adjusted anthropometric measures are shown in Table 2. All the CV risk factors correlated significantly both with WC and with BMI, except LDL cholesterol in girls and blood glucose in boys. Fasting insulin level was the CV risk factor which correlated most strongly with both WC and BMI, followed by systolic pressure, HDL cholesterol (negative), TG and diastolic pressure, in both girls and boys. Correlations with waist/hip and waist/height ratios were weaker.

Table 2 Pearson's correlation coefficients among age-adjusted BMI, WC, waist/hip ratio, waist/height ratio and CV risk factors in the CV risk evaluation group – correlation coefficients (P-values)

The percentages of children classified as having three-or-more and four-or-more of the six CV risk factors were ca. 10% and ca. 3%, respectively. ROC analysis showed that the findings with WC were generally similar to those with BMI (Table 3), although gender- and age-specific ROC curves suggest a slightly greater AUC for WC than BMI (Figure 2).

Table 3 Optimal WC and BMI threshold for high CV risk (ROC analysis) – (a) Three of the six CV risk factors and (b) four of the six CV risk factors
Figure 2

ROC curves for age-adjusted WC and BMI with high CV risk. (a) Three-or-more out of the six CV risk factors (solid line: WC; dotted line: BMI) and (b) four-or-more out of the six CV risk factors (solid line: WC; dotted line: BMI).

Risk threshold percentiles for WC were 12 and 15 percentiles higher using four than using three of the six risk factors, with consequently fewer children ‘at risk’. Calculated AUCs (reflecting the optimal composite of sensitivity and specificity) were 0.16 (22%) and 0.05 (6%) greater and risk cutoff WC values averaged 5% (2–4 cm) and 7% (3–5 cm) higher with four than with three of the six risk factors (Table 4). Similar results were obtained with BMI – risk threshold percentiles were 9 percentiles and 16 percentiles higher, AUCs were 0.14 (18%) and 0.05 (6%) greater, and risk cutoff values averaged 6% (4–7 kg/m2) and 9% (7–11 kg/m2) higher with four than with three of the six risk factors.

Table 4 WC/BMI cutoff values for high CV risk – (a) three of the six CV risk factors and (b) four of the six risk factors

Analyses of the present data using (i) the four risk factors identified in the ATP III report and (ii) the four CV risk factors most strongly associated with WC – fasting insulin, systolic BP, HDL and TGs are shown in Table 5. Compared with the use of the six risk factors or the four ATP III CV risk factors, the use of the selected four risk factors showed a larger AUC relative to the number of risk factors considered and the number of children at risk. Among the children with three-or-more of the six risk factors, the use of systolic pressure was compared with the use of diastolic BP and shown to give a consistently greater sensitivity (mean of all children: 0.73 cf. 0.69), selectivity (0.80 cf. 0.75) and AUC (0.82 cf. 0.78), supporting a greater contribution of systolic than diastolic BP to the association with WC.

Table 5 Summary of ROC analyses for WC cutoffs using different groupings of CV risk factors

The results of analysing the present data using three-or-more of the six risk factors and the outer quintile (20%) of risk factor percentiles are shown for comparison with the Bogalusa data in Table 5.


This study presents smoothed reference percentile charts for WC and BMI for ethnic Chinese children aged 6–12 years, inclusive. The 11–13% proportion of obesity in the random population group of 2593 children is similar to that reported annually for primary school children in the Territory by the Hong Kong Department of Health (SHS communication, 2002–2003). BMI percentiles are higher than in the first population data for Chinese children's BMI collected in 1993.35 These are the first WC percentile charts for Chinese children to have been reported.

Overall, CV risk may be characterized in terms of the CV risk factors present, although their relative contributions and whether there is synergy between the individual factors underlying the concept of the Metabolic Syndrome have been questioned, as have the diagnostic criteria.36, 37 High WC and BMI levels were confirmed in the present study as being associated with clusters of three-or-more or four-or-more of the six CV risk factors measured. The most strongly associated individual factor of the six was fasting insulin, followed by TG and HDL, with systolic more strongly associated than diastolic BP, and LDL cholesterol and blood glucose only weakly or insignificantly associated – a hierarchy of individual associations similar to that noted previously.38

WC showed a slightly stronger association than BMI with the lipid and insulin risk factors, supporting the growing use of WC in preference to BMI39, 40 and consistent with the fact that WC is a measure of central obesity which is specifically related to the presence of CV risk factors.41, 42 Weight/hip ratio and weight/height ratios, as previously used in adults43, 44 albeit controversially in children,15, 21, 45, 46 were shown to be less strongly associated with the CV risk factors.

Age- and sex-specific optimal threshold levels of WC and BMI for predicting CV risk factors were derived by ROC. Predictably, fewer children had four-or-more than three-or-more of the six risk factors, with higher risk threshold percentiles and greater sensitivity, specificity and AUC. Results with BMI were similar.

In view of unresolved differences in the definition of the Metabolic Syndrome and its diagnostic components, its underlying cause and whether combinations of risk factors are more predictive of CV outcome than the sum of individual risk factors,36 the data were also analysed using two alternative combinations of risk factors used by others, the better to compare findings – (i) using the four risk factors identified in the ATP III report and (ii) using the four risk factors shown in this study to be those most strongly associated with WC which were also those originally considered to be the main features of the Metabolic Syndrome. Differences were small but suggest greater sensitivity and specificity relative to the number of risk factors used with the latter than were found using either the six risk factors or the four ATP III risk factors. Inclusion of LDL and of glucose contributed little to their selectivity. Both correlation coefficients and ROC analysis showed a greater contribution of systolic than diastolic BP, supporting the relative value of systolic pressure in characterizing the Metabolic Syndrome. Nevertheless, the cutoff levels were similar, suggesting that the choice of risk factors does not greatly affect the cutoff levels derived.

Comparable data for percentiles in children from different countries and ethnic groups are limited, but there appear to be differences in WC percentiles between different populations and ethnic groups. In particular the representative 50th percentile for WC (which reflects waist size in the population) was higher in the USA national data collected from 1988 to 199429 than in these ethnic Chinese children in Hong Kong (Figure 3). Data relating WC to CV risk factors in different child populations are scant15, 34, 38, 39, 47, 48 and valid comparisons are limited by differences in methodology (non-random population sampling, different sites of WC measurement, selection and number of CV risk factors) and/or failure to include more than one of the inter-related variables (WC percentiles, WC risk threshold percentiles and cutoff values) in the same study. The only methodologically comparable study to derive WC risk thresholds percentiles and gender- and age-related cutoff values is the Bogalusa study of 2597 Black and White American children,34 in which three-or-more of the same six CV risk factors as in the present study were used. Data from the present study, re-analysed using outer quintiles to define individual risk factors as in the Bogalusa study (instead of the outer 15% in the initial analysis of the present study), showed that WC risk thresholds (which were at the 69th and 66th percentile in girls and boys, respectively) were 15% higher than in either the White or the Black American children (Table 5), whereas WC risk cutoff values differed little (Figure 4), consistent with the smaller WC percentiles in ethnic Chinese children than in White and Black Americans. AUCs, sensitivities and specificities were similar in the two studies. This suggests that the relation between central abdominal fat and risk factors may not differ fundamentally between different ethnic and racial populations despite population differences in WC percentiles, supporting the possibility that an international standard for children's WC as a predictive measure of CV risk factors might be feasible. Further closely comparable studies will be needed to establish the extent to which this might be the case.

Figure 3

50th percentile for WC. (a) Girls and (b) boys. Bogalusa data (US White and US Black) 34 and HK (Chinese) data from present study – both using three of the six risk factors and outer quintiles.

Figure 4

WC thresholds for predicting high CV risk (a) Girls and (b) boys.

In summary, this study provides the first percentile charts for ethnic Chinese children's WC. For initial clinical screening, we suggest that optimal thresholds based on the presence of four-or-more of the six risk factors have the advantage over three of six factors of high sensitivity and specificity, and also of higher WC values, which will reduce the number of children who will require weight management. Further analysis of the present study suggests, however, that selective use of the four risk factors originally considered central to the Metabolic Syndrome may provide a more focused approach. Contrary to expectations, standardized risk threshold cutoff values for WC may differ little between different ethnic groups despite anthropometric differences reflected in population WC percentiles. WC percentiles may serve as epidemiological markers of population risk, although this remains to be established.


  1. 1

    Reaven GM . Banting lecture 1988. Role of insulin resistance in human disease. Diabetes 1988; 37: 1595–1607.

  2. 2

    Freeman DS, Dietz WH, Srinivasan SR, Berenson GS . The relation of cardiovascular risk factors among children and adolescents: The Bogalusa heart Study. Pediatrics 1999; 103: 1175–1182.

  3. 3

    Wilson PW, D'Agostino RB, Sullivan L, Parise H, Kannel WB . Overweight and obesity as determinants of cardiovascular risk: the Framingham experience. Arch Intern Med 2002; 162: 1867–1872.

  4. 4

    Ford ES . The metabolic syndrome and mortality from cardiovascular disease and all-causes: findings from the National Health and Nutrition Examination Survey II Mortality Study. Atherosclerosis 2004; 173: 309–314.

  5. 5

    Webber L, Srinivasan SR, Wattigney W, Berenson GS . Tracking of serum lipids and lipoproteins from childhood to adulthood: the Bogalusa Heart Study. Am J Epidemiol 1991; 133: 884–899.

  6. 6

    Raitakari OT, Porkka KV, Rasanen L, Ronnemaa T, Viikari JS . Clustering and six year cluster-tracking of serum total cholesterol, HDL-cholesterol and diastolic blood pressure in children and young adults. The Cardiovascular Risk in Young Finns Study. J Clin Epidemiol 1994; 47: 1085–1093.

  7. 7

    Porkka KV, Viikari JS, Taimela S, Dahl M, Akerblom HK . Tracking and predictiveness of serum lipid and lipoprotein measurements in childhood: a 12-year follow-up. The Cardiovascular Risk in Young Finns study. Am J Epidemiol 1994; 140: 1096–1110.

  8. 8

    Bao W, Srinivasan SR, Wattigney WA, Berenson GS . Persistence of multiple cardiovascular risk clustering related to syndrome X from childhood to young adulthood. The Bogalusa Heart Study. Arch Intern Med 1994; 154: 1842–1847.

  9. 9

    Berenson GS, Sathanur RS, Weihang B, Newman WP, Tracy RE, Wattigney WA . Association between multiple cardiovascular risk factors and atherosclerosis in children and young adults. N Eng J Med 1998; 338: 1650–1656.

  10. 10

    Gower BA, Nagy TR, Goran MI . Visceral fat, insulin sensitivity, and lipids in prepubertal children. Diabetes 1999; 48: 1515–1521.

  11. 11

    Weiss R, Dufour S, Taksali SE, Tamborlane WV, Petersen KF, Bonadonna RC et al. Prediabetes in obese youth: a syndrome of impaired glucose tolerance, severe insulin resistance, and altered myocellular and abdominal fat partitioning. Lancet 2003; 362: 951–957.

  12. 12

    Schreiner PJ, Terry JG, Evans GW, Hinson WH, Crouse III JR, Heiss G . Sex-specific associations of magnetic resonance imaging-derived intra-abdominal and subcutaneous fat areas with conventional anthropometric indices. The atherosclerosis risk in communities study. Am J Epidemiol 1996; 144: 335–345.

  13. 13

    Goran MI, Gower BA, Treuth M, Nagy TR . Prediction of intra-abdominal and subcutaneous abdominal adipose tissue in healthy pre-pubertal children. Int J Obes Relat Metab Disord 1998; 22: 549–558.

  14. 14

    Kamel EG, McNeill G, Han TS, Smith FW, Avenell A, Davidson L et al. Measurement of abdominal fat by magnetic resonance imaging, dual-energy X-ray absorptiometry and anthropometry in non-obese men and women. Int J Obes Relat Metab Disord 1999; 23: 686–692.

  15. 15

    Taylor RW, Jones IE, Williams SM, Goulding A . Evaluation of waist circumference, waist-to-hip ratio, and the conicity index as screening tools for high trunk fat mass, as measured by dual-energy X-ray absorptiometry, in children aged 3–19 years. Am J Clin Nutr 2000; 72: 490–495.

  16. 16

    Maffeis C, Corciulo N, Livieri C, Rabbone I, Trifiro G, Falorni A et al. Waist circumference as a predictor of cardiovascular and metabolic risk factors in obese girls. Eur J Clin Nutr 2003; 57: 566–572.

  17. 17

    Brambilla P, Bedogni G, Moreno LA, Goran MI, Gutin B, Fox KR et al. Crossvalidation of anthropometry against magnetic resonance imaging for the assessment of visceral and subcutaneous adipose tissue in children. Int J Obes Relat Metab Disord 2006; 30: 23–30.

  18. 18

    Lemieux I, Pascot A, Couillard C, Lamarche B, Tchernof A, Almeras N et al. Hypertriglyceridemic waist:a marker of the atherogenic metabolic triad (hyperinsulinemia; hyperapolipoprotein B; small, dense LDL) in men? Circulation 2000; 102: 179–184.

  19. 19

    Carr DB, Utzschneider KM, Hull RL, Kodama K, Retzlaff BM, Brunzell JD et al. Intra-abdominal fat is a major determinant of the National Cholesterol Education Program Adult Treatment Panel III criteria for the metabolic syndrome. Diabetes 2004; 53: 2087–2094.

  20. 20

    Savva SC, Tornaritis M, Savva ME, Kourides Y, Panagi A, Silikiotou N et al. Waist circumference and waist-to-height ratio are better predictors of cardiovascular disease risk factors in children than body mass index. Int J Obes Relat Metab Disord 2000; 24: 1453–1461.

  21. 21

    Janssen I, Katzmarzyk PT, Ross R . Waist circumference and not body mass index explains obesity-related health risk. Am J Clin Nutr 2004; 79: 379–384.

  22. 22

    Misra A, Wasir JS, Vikram NK . Waist circumference criteria for the diagnosis of abdominal obesity are not applicable uniformly to all populations and ethnic groups. Nutrition 2005; 21: 969–976.

  23. 23

    Zhu S, Heymsfield SB, Toyoshima H, Wang Z, Pietrobelli A, Heshka S . Race-ethnicity-specific waist circumference cutoffs for identifying cardiovascular disease risk factors. Am J Clin Nutr 2005; 81: 409–415.

  24. 24

    Zannolli R, Morgese G . Waist percentiles: a simple test for atherogenic disease? Acta Paediatr 1996; 85: 1368–1369.

  25. 25

    Moreno LA, Fleta J, Nur L, Rodriguez G, Sarria A, Bueno M . Waist circumference values in Spanish children – gender related differences. Eur J Clin Nutr 1999; 53: 429–433.

  26. 26

    Savva SC, Kourides Y, Tornaritis M, Epiphanious-Savva M, Tafouna P, Kafatos A . Reference growth curves for Cypriot children 6–17 years of age. Obes Res 2001; 9: 754–762.

  27. 27

    McCarthy HD, Jarrett KV, Crawley HF . The development of waist circumference percentiles in British children aged 5.0–16.9 years. Eur J Clin Nutr 2001; 55: 902–907.

  28. 28

    Katzmarzyk PT . Waist circumference percentiles for Canadian youth 11–18 years of age. Eur J Clin Nutr 2004; 58: 1011–1015.

  29. 29

    Fernandez JR, Redden DT, Pietrobelli A, Allison DB . Waist circumference percentiles in nationally representative samples of African-American, European-American, and Mexican-American children and adolescents. J Pediatr 2004; 145: 439–444.

  30. 30

    Fredriks AM, van Buuren S, Fekkes M, Verloove-Vanhorick SP, Wit JM . Are age references for waist circumference, hip circumference and waist-hip ratio in Dutch children useful in clinical practice? Eur J Pediatr 2005; 164: 216–222.

  31. 31

    Eisenmann JC . Waist circumference percentiles for 7- to 15-year-old Australian children. Acta Paediatrica 2005; 94: 1182–1185.

  32. 32

    Cole TJ, Green PJ . Smoothing reference centile curves: the LMS method and penalized likelihood. Stat Med 1992; 11: 1305–1319.

  33. 33

    Third report of the National cholesterol Education Program (NCEP). Expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). Final report. Circulation 2002; 106: 3134–3421.

  34. 34

    Katzmarzyk PT, Srinivasan SR, Chen W, Malina RM, Bouchard C, Berenson GS . Body mass index, waist circumference, and clustering of cardiovascular disease risk factors in a biracial sample of children and adolescents. Pediatrics 2004; 114: e198–e205.

  35. 35

    Leung SS, Cole TJ, Tse LY, Lau JT . Body mass index reference curves for Chinese children. Ann Hum Biol 1998; 25: 169–174.

  36. 36

    Kahn R, Buse J, Ferrannini E, Stern M . The metabolic syndrome: time for a critical appraisal. Joint statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetologia 2005; 48: 1684–1699.

  37. 37

    Golley RK, magarey AM, Steinbeck KS, Baur LA, Daniels LA . Comparison of metabolic syndrome prevalence using six different definitions in overweight pre-pubertal children enrolled in a weight management study. Int J Obes Relat Metab Disord 2006; 30: 853–860.

  38. 38

    Higgins PB, Gower BA, Hunter GR, Goran MI . Defining health-related obesity prepubertal children. Obes Res 2001; 9: 233–240.

  39. 39

    Moreno LA, Pineda I, Rodriguez G, Fleta J, Sarrfa A, Bueno M . Waist circumference for the screening of the metabolic syndrome in children. Acta Paediatr 2002; 91: 1301–1312.

  40. 40

    McCarthy HD, Jarrett KV, Emmett PM, Rogers I . Trends in waist circumferences in young British children: a comparative study. Int J Obes Relat Metab Disord 2005; 29: 157–162.

  41. 41

    Larsson B, Svardsudd K, Welin L, Wilhelmsen L, Bjorntorp P, Tibblin G . Abdominal adipose tissue distribution, obesity, and risk of cardiovascular disease and death: 13 year follow up of participants in the study of men born in 1913. BMJ (Clin Res Ed) 1984; 288: 1401–1404.

  42. 42

    Flodmark CE, Sveger T, Nilsson-Ehle P . Waist measurement correlates to a potentially atherogenic lipoprotein profile in obese 12–14 years old children. Acta Paediatr 1994; 83: 941–945.

  43. 43

    Yusuf S, Hawken S, Ounpuu S, Bautista L, Franzosi MG, Commerford P et al. INTERHEART Study Investigators. Obesity and the risk of myocardial infarction in 27 000 participants from 52 countries: a case–control study. Lancet 2005; 5: 1640–1649.

  44. 44

    Hsieh SD, Yoshinaga H . Waist/height ratio as a simple and useful predictor of coronary heart disease risk factors in women. Intern Med 1995; 34: 1147–1152.

  45. 45

    Hara M, Saitou E, Iwata F, Okada T, Harada K . Waist-to-height ratio is the best predictor of cardiovascular disease risk factors in Japanese schoolchildren. J Atheroscler Thromb 2002; 9: 127–132.

  46. 46

    Neovius M, Linne Y, Rossner S . BMI, waist-circumference and waist-hip-ratio as diagnostic tests for fatness in adolescents. Int J Obes Relat Metab Disord 2005; 29: 163–169.

  47. 47

    Janssen I, Katzmarzyk PT, Srinivasan SR, Chen W, Malina RM, Bouchard C et al. Combined influence of body mass index and waist circumference on coronary artery disease risk factors among children and adolescents. Pediatrics 2005; 115: 1623–1630.

  48. 48

    Asayama K, Hayashibe H, Endo A, Okada T, Hara M, Masuda H et al. Threshold values of visceral fat and waist girth in Japanese obese children. Pediatr Int 2005; 47: 498–504.

Download references


This study was supported by the Jump Rope for heart project, Hong Kong College of Cardiology and Polar Electro Oy. We are grateful to Professor AH Henderson, Emeritus Professor of Cardiology and Professor RR West, Department of Statistics and Epidemiology, at the University of Wales College of Medicine, for their critical review of the manuscript. Special thanks go to Professor Tony Nelson for his valuable comments.

Author information

Correspondence to R Y T Sung.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Sung, R., Yu, C., Choi, K. et al. Waist circumference and body mass index in Chinese children: cutoff values for predicting cardiovascular risk factors. Int J Obes 31, 550–558 (2007) doi:10.1038/sj.ijo.0803452

Download citation


  • waist circumference
  • BMI
  • cardiovascular risk factors
  • children
  • Chinese
  • Hong Kong

Further reading