The cardiometabolic syndrome, or the metabolic syndrome, is a cluster of multiple cardiometabolic disorders characterized by central obesity and insulin resistance. The cardiometabolic syndrome has become a major public health problem worldwide, contributing to the ever-increasing mortality rate and health care costs. It has been commonly accepted that central obesity may serve as a central determining factor for the cardiometabolic syndrome, with waist circumference being essential for clinical diagnosis of the metabolic syndrome.1 Ample epidemiologic data have displayed dramatic variations in the cutoffs for central obesity among different ethnic groups.2 According to the definition from the International Diabetes Federation (IDF),1 ethnicity-specific cutoffs should be considered for reliable assessment of obesity. Table I summarizes certain ethnicity-associated cutoffs for central obesity defined by the IDF. A number of epidemiologic studies have indicated a high prevalence of the metabolic syndrome (35.2%),3 overweight (65.5%), and obesity (18.3%)4 in the Uygur ethnic population residing in Urumqi, China. Nonetheless, these diagnostic criteria were essentially made based on the global World Health Organization concept without consideration of any ethnicity specificity for the Uygur ethnic population in Urumqi, China. Given the apparent ethnicity disparity in clinical evaluation and diagnosis of the metabolic syndrome, this study was designed to determine the appropriate cutoff specific to the Uygur ethnic population in order to provide some guidance for the better prevention and management against the cardiometabolic syndrome and obesity in the Uygur population.
Table I. Ethnicity-Specific Values for Waist Circumference
Waist Circumference, cm (as Measure of Central Obesity)
aIn the United States, the National Cholesterol Education Program Adult Treatment Panel III [NCEP ATP III] values (102 cm in men; 88 cm in women) are likely to continue to be used for clinical purposes. bBased on a Chinese, Malay, and Asian Indian population. cSubsequent data analyses suggest that Asian cutoffs (90 cm in men; 80 cm in women) should be used for Japanese populations until more data are available.
Ethnic South and Central Americans
Use South Asian recommendations until more specific data are available
Use European data until more specific data are Available
Eastern Mediterranean and Middle Eastern (Arab) populations
Use European data until more specific data are available
Research Design and Methods
Participants. A total of 2526 Uygur individuals aged between 25 and 90 years (1170 men and 1356 women) were recruited. Among them, 870 persons (metropolitan residents) were sampled from 10 workplaces in Urumqi, China, using the random cluster sampling technique; 1656 persons were from the Kashgar region, using stratified cluster sampling. Stratified cluster sampling, which combines stratified and cluster sampling techniques, was used to reduce sampling errors when individuals were sampled from both urban and rural areas in the Kashgar region. Written consent was obtained from all study participants.
Methods. Questionnaires were provided by trained and certified Uygur interviewers. Information was collected on age, sex, ethnicity, marital status, occupation, nature and intensity of work, per capita income, educational background, smoking status, alcohol consumption, family history of hypertension, obesity, and diabetes.
Anthropometric measurements were obtained, including blood pressure, height, weight, and waist and hip circumference. Blood pressure assessment in the right arm was performed in a seated position in the morning using a mercury sphygmomanometer. The mean of the 2 readings was used. The waist circumference was measured at the part of the trunk located midway between the lower costal margin (bottom of lower rib) and the iliac crest (top of pelvic bone) while the individual was standing with feet 25 to 30 cm apart. The tape was fit snugly without compressing against underlying soft tissues. The circumference was measured at the end of a normal expiration. The mean of the 2 readings was used in the current study.
Clinical laboratory testing included glucose tolerance testing. Following an overnight fast, a 75-g oral glucose challenge was administered to asymptomatic participants. Fasting blood glucose, blood lipid, and postload glucose levels within 2 hours of the glucose challenge were measured in diabetic participants.
The metabolic syndrome diagnostic criteria were based on the IDF definition,1 including central obesity plus any 2 of the following 4 factors:
• Elevated triglyceride level (>150 mg/dL [1.7 mmol/L]) or specific treatment for this lipid abnormality.
• Reduced high-density lipoprotein cholesterol level (<40 mg/dL [1.03 mmol/L] in men and <50 mg/dL [1.29 mmol/L] in women) or specific treatment for this lipid abnormality.
• Elevated blood pressure (systolic blood pressure >130 mm Hg or diastolic blood pressure >85 mm Hg) or treatment of previously diagnosed hypertension.
• Elevated fasting plasma glucose (>100 mg/dL [5.6 mmol/L]) or previous diagnosis of type 2 diabetes. If fasting plasma glucose was >5.6 mmol/L (100 mg/dL), an oral gluclose tolerance test was highly recommended, although it was considered unnecessary to confirm the presence of the cardiometabolic syndrome.
Statistical Analysis. The cutoffs for waist circumference in men and women were categorized into different strata according to the waist circumference cutoffs recommended by the China Obesity Task Force5 and the distribution of the cutoffs in the Uygur population. The following brackets were used for waist circumference: <80, 80–84, 85–89, 90–94, 95–99, 100–104, and ≥105 cm for men and <75, 75–79, 80–84, 85–89, 90–94, 95–99, 100–104, and ≥105 cm for women. In contrast with the control individuals with waist circumference <80 cm (men) and <75 cm (women), the waist circumference of each stratum was analyzed to identify the odds ratio (OR) of clustering of any 1, 2, 3 or more metabolic syndrome components as well as 95% confidence interval (age-adjusted).
Accurate waist circumference values were identified by more specific strata of the cutoffs (80, 85, 90, 91, 92, 93, 94, 95, 100, and 105 cm for men and 75, 80, 85, 86, 87, 88, 89, 90, 91, 95, 100, and 105 cm for women). The sensitivity and specificity of ≥2 metabolic syndrome components were identified, as well as the positive predicative value, negative predicative value, and accuracy index (sensibility + specificity – 100). The distance in receiver operating characteristic (ROC) curve was estimated.6
The distance in ROC curve was calculated as
The cutoff corresponding to the shortest distance in ROC curve was used to estimate the sex- and age-specific prevalence of the metabolic syndrome and the ratio of metabolic syndrome components combination.
Distribution of Waist Circumference and Relevance to the Cluster of Metabolic Syndrome Components. In the study population of 2526 persons aged 25 to 90 years, 50.6% of men had a waist circumference ≥93 cm, while 44.6% of women had a waist circumference ≥89 cm (Figure 1). Regardless of sex, the ORs of clustering of metabolic syndrome components increased with the waist circumference, which showed that the clustering of metabolic syndrome components was strongly correlated with the existence of abdominal adiposity (Table II).
Table II. ORs of ≥1, ≥2, ≥3 Metabolic Syndrome Components by Different Strata of Waist Circumference (WC)
OR (95% CI)
OR (95% CI)
OR (95% CI)
Abbreviations: CI, confidence interval; OR, odds ratio. Values are age-adjusted.
Analysis of Appropriate Waist Circumference Cutoffs. Based on the IDF definition, enlarged waist circumference was deemed essential for the diagnosis of the metabolic syndrome; that is, the metabolic syndrome can be reliably identified using a waist circumference above a certain cutoff plus any 2 of the other 4 risk components. Table III displays the estimated sensitivity, specificity, positive and negative predictive values, the distance in the ROC curve, and the accuracy index of >2 metabolic syndrome components observed at different cutoffs of waist circumference. It was noted that the specificity and the positive predictive value rise while the sensitivity and the negative predicative value decline with an enlarged waist circumference. The cutoffs corresponding to the shortest distance in the ROC curve were 93 and 89 cm for men and women, respectively, which suggests that the false-positive and -negative incidence of metabolic syndrome diagnosis at this cutoff was relatively low, while the accuracy index was relatively high.
Table III. Sensitivity, Specificity, Positive and Negative Predictive Values, the Distance in ROC Curve and Accuracy Index of >2 Metabolic Syndrome Components at Different Cutoffs of Waist Circumference (WC)
Figure 2A depicts the detection rate of the metabolic syndrome estimated with the waist circumference cutoff of ≥93 cm for men and ≥89 cm for women. By these cutoffs, the detection rate of the metabolic syndrome in the study population was 26.0% in men and 26.9% in women. The detection rate rose steadily with age, with the middle-aged group at the peak. In addition, the detection rate was higher in women compared with their age-matched male counterparts. However, the detection rate decreased in both sexes in those older than 65 years. The relationship between age and waist circumference displays a somewhat similar pattern. The average waist circumference increases with age and remains stable after the age of 45 to 54 years (Figure 2B). The waist circumference averaged 92.54 and 87.29 cm, respectively, for male and female participants. As for the clustering of 2 risk components, high triglycerides plus raised blood pressure was identified most frequently in men, accounting for 53.0% of metabolic syndrome detection in men. On the other side of the coin, high triglycerides in conjunction with low high-density lipoprotein cholesterol were observed most frequently in women, accounting for 52.3% of the metabolic syndrome identified in females. In our study, the prevalence of diabetes was 13.62%, and the prevalence of impaired fasting glucose was 14.98% (breakdown data not shown).
A plethora of studies have demonstrated that visceral adiposity is not only a feature of the metabolic syndrome but also a major trigger for insulin resistance.7,8 Despres and colleagues showed that visceral adiposity carries more weight in metabolic syndrome diagnosis than does insulin resistance. Despite that insulin resistance may mediate carbohydrate intolerance, contributing to metabolic abnormalities, visceral adiposity is believed to correlate with a number of pathologic processes independent of insulin resistance.8 A study in Japanese population suggested that visceral adiposity prompted low circulating adiponectin-triggered insulin resistance, resultant dyslipidemia, endothelial dysfunction, and eventually hypertension.9 These data have provided a theoretical base for the IDF definition and are somewhat consistent with our current experimental findings. The IDF definition for the metabolic syndrome was used in our study to search for the cutoff of waist circumference most appropriate for visceral adiposity. Our goal was to better identify the high-risk group of patients with the metabolic syndrome in the Uygur ethnic population. Of interest, our study did observe the positive correlation between enlarged waist circumference and incidence of the clustering of metabolic syndrome components (in addition to waist circumference). The distance in ROC curve was found to be the shortest at 93 and 89 cm of waist circumference for men and women, respectively. The detection rate of the metabolic syndrome estimated at these points was 26.9%, which was significantly higher than the prevalence of the metabolic syndrome in Shanghai adults from a 2001 survey (17.14%)10 and in American adults based on a 2002 survey (23.7%).11
Central obesity is most easily measured by waist circumference with sex and ethnicity specificity. Pragmatic cut points taken from various sources are essential in the clinical management of the cardiometabolic syndrome in different ethnic groups. Using the cutoffs for waist circumference proposed by the IDF for the Chinese population (waist circumference ≥90 cm for men and ≥80 cm for women), higher sensitivity may be achieved in metabolic syndrome diagnosis, although the incidence of false-positives was also elevated in the Uygur ethnic population. The detection rate of the metabolic syndrome in our current study population appeared to be higher. Nonetheless, the proposed criteria by the IDF did not include consideration of the ethnicity specificity for metabolic syndrome diagnoses in the Uygur population. According to the data obtained in our study, the average waist circumference in the study population (92.5 cm for men and 87.3 cm for women) is much higher than the mean waist circumference for the Han ethnicity in China,12 but close to the European standard (waist circumference for European men ≥94 cm and the National Cholesterol Education Program Adult Treatment Panel III [NCEP ATP III] criteria for women ≥88 cm). Of interest, our estimated metabolic syndrome detection rate was accordingly higher than that in America. Other than the genotypic milieu between these 2 ethnic populations, the “thrifty genotype” hypothesis13,14 may play a major role in the high detection rate of central obesity and the metabolic syndrome in the Chinese Uygur population. As Neel13 proposed, the “thrifty” genes are capable of enhancing the capacity to store fat and thus place an individual at risk for development of insulin resistance and type 2 diabetes. The socioeconomic living standard was relatively low for the Chinese, especially the Uygurs, in Xinjiang during the 1950s and 1960s. As a consequence, babies born during those years of poverty who are in middle age now have presented with a high prevalence of obesity, type 2 diabetes, and the metabolic syndrome3; this is true even if the living standard of Uygurs is no longer an issue. Therefore, the waist circumference cutoffs may also vary with time. As the socioeconomic situation alters, the impact of nutrition on fetal and infantile development might have a disparate contribution to the ultimate determination of appropriate waist circumference cutoffs for obesity. In comparison with the currently available criteria for waist circumference, the gap in waist circumference between sex was deemed minimal in our current study, a difference of a mere 4 cm. Our result is somewhat discrepant from the Japanese population in whom waist circumference was noticeably greater in women (the waist circumference defined for the Japanese population was ≥85 cm for men and ≥90 cm for women).
In the cluster of risk components of the metabolic syndrome diagnosed at this cutoff, a high triglyceride level plus raised blood pressure was considered the most important factor in men, while a high triglyceride level plus a low high-density lipoprotein cholesterol value was deemed most pronounced in women, accounting for 53.0% and 52.3%, respectively, in both sexes. These data indicated that the Uygur population possesses overt dyslipidemia, mainly manifested as a high triglyceride level. According to the American National Health and Nutrition Examination Survey (NHANES) III study, the population with an enlarged waist circumference and raised triglycerides displayed higher circulating fasting insulin levels, homeostasis model assessment of insulin resistance index, and fasting blood glucose than those populations with normal values in these 2 aspects. This study seems to favor that enlarged waist circumference and raised triglycerides better reflect adiposity and consequently metabolic abnormalities.15 Visceral adiposity is known to lead to peripheral insulin resistance and weakened antilipolytic action of insulin.16 Individuals with visceral adiposity may often display stearolysis of lipocytes, releasing free fatty acids in large quantities to increase the overall output of glycogen and liver triglyceride secretion.17
In conclusion, data from our current study implicated that the IDF-defined cutoffs of waist circumference should be adjusted among different ethnic groups for a better and more effective screening of the risk of the metabolic syndrome in these ethnic populations. Our study has shed some light toward understanding the ethnic variation in metabolic syndrome diagnosis, although further confirmative study based on a larger population size is warranted. Hopefully, our result will provide some practical guidance in light of the diagnosis of the metabolic syndrome among Uygur adults in China. With identification of the cutoffs for waist circumference in the Uygur ethnic group, our study contributes to the comparative study on metabolic syndrome data of different ethnic groups worldwide. Last but not least, caution needs to be taken with interpretation of these data, as sampling errors are sometimes inevitable in this type of epidemiologic study.