BACKGROUND: Metabolic syndrome (MetSyn) in children and adolescence is increasing worldwide; however, its pattern may be different between Asians and Americans. We compare the prevalence and patterns of MetSyn between American and Korean children and adolescents between roughly 1998 and 2007.
METHODS: Data from the American and Korean versions of the NHANES (NHANES and KNHANES) were used for this study. The main outcome is prevalence and pattern of MetSyn among participants separately in each country. In each survey, stratified multistage probability sampling designs and weighting adjustments were conducted to represent the entire population. The revised National Cholesterol Education Program criteria were used to define MetSyn.
RESULTS: Totals of 934, 1781, and 1690 Americans aged 12 to 19 participated in NHANES 1988–1994, NHANES 1999–2002, and NHANES 2003–2006, respectively; and 1225, 976, 705, and 456 Koreans aged 12 to 19 have participated in KNHANES 1998, 2001, 2005, and 2007. The age-adjusted prevalence of MetSyn in American NHANES decreased from 7.3% to 6.7% and 6.5%, whereas in Korean NHANES there was an increase from 4.0% to 5.9%, 6.6%, and 7.8% in each country’s respective study. Increases in dyslipidemia and abdominal obesity contributed to the increased prevalence in Korea, whereas in the United States, decreases in low high-density lipoprotein cholesterolemia and high blood pressure contributed to a decreased prevalence.
CONCLUSIONS: Considering different phenotype changes, different approaches should be conducted at the national level to reduce the burden and consequences of MetSyn between Korea and the United States.
- BP —
- blood pressure
- CVD —
- cardiovascular disease
- DBP —
- diastolic blood pressure
- HDL —
- high-density lipoprotein
- LDL —
- low-density lipoprotein
- MetSyn —
- metabolic syndrome
- NCEP-ATP III —
- National Cholesterol Education Program–Adult Treatment Panel III
- SBP —
- systolic blood pressure
What’s Known on This Subject:
In the United States, adolescent obesity rates have tripled in the last 3 decades, with concomitant increases in other metabolic risk factors, including the metabolic syndrome (MetSyn). However, in Asian countries, these same risks have only recently begun increasing.
What This Study Adds:
Representative data for the United States and Korea reveal trends in adolescent obesity and MetSyn moving in opposite directions. This study provides a benchmark for Korea and other Asian countries toward mitigating the upward trends in obesity and MetSyn.
Metabolic syndrome (MetSyn) is a well-known risk factor of type 2 diabetes and cardiovascular disease (CVD),1,2 with an increasing trend noted among adults.3 More recently, similar metabolic abnormalities, including obesity, in both children and adolescents have been increasing rapidly across many countries.4,5 Considering that developing these risk factors at a young age is related to a significantly increased risk of CVD morbidity and mortality during adulthood,6–8 identification of MetSyn in children and adolescence is of great importance from a clinical and public health perspective.
By using the National Cholesterol Education Program–Adult Treatment Panel III (NCEP-ATP III) criteria, the NHANES III reveals 4.2% of US adolescents aged 12 to 19 years as having MetSyn.9 Other investigators have varied in their methodology for identifying MetSyn, with a prevalence range of 6.5% in Mexican adolescents,10 18% in the obese Spanish pediatric population,11 18.6% in Native Canadian adolescents,12 and ∼50% in severely obese US adolescents.13 Thus, the diagnostic criteria and prevalence of MetSyn have not been well characterized in younger age groups across a number of different countries.
Asian populations, likely due to different genetic backgrounds, dietary patterns, and physical activity levels, have shown different metabolic profiles compared with white patients.14–16 Among Asian countries, Korea is experiencing the fastest epidemiologic switch from infectious diseases to chronic degenerative conditions among its population.17 Westernization of the diet and a decrease in physical activity have evoked metabolic imbalances, increased rates of obesity, and dramatic increases in CVD and diabetes.18–21 These recent negative lifestyle and health changes are similar to what has already occurred in the United States and present a unique opportunity for cross-cultural comparison.
Therefore, the purpose of this study was to compare changes in the prevalence of MetSyn between American and Korean adolescents aged 12 to 19 years by using the NCEP-ATP III-derived definition of MetSyn. In addition, specific risk factors related to changes in MetSyn prevalence are also discussed.
US data were obtained from (1) NHANES III (1988–1994, data release series 11, #1A), (2) the combined 4-year sample weights for NHANES 1999–2002, and (3) appropriate sample weights for NHANES 2003–2006 constructed by using the National Center for Health Statistics guidelines.22 Comprehensive details of the various NHANES surveys are described elsewhere.10–12
Korean data were obtained from public-use data sets of the KNHANES, conducted by the Korean Ministry of Health and Welfare in noninstitutionalized Korean civilians during 1998, 2001, 2005, and 2007. Details of the 1998, 2001, and 2005 surveys have been previously described,23,24 and a similar design and method was used in KNHANES 2007. In brief, a stratified multistage probability sampling design was used with selection made from sampling units based on geographical area, gender, and age groups by using household registries. In addition, there was no significant difference between participants and noncompleters in all studies, indicating no bias due to dropouts.
To give an equal probability of being sampled, weights were assigned to each respondent, enabling the results to represent the entire Korean population, guaranteeing unbiased point estimates of population parameters for the entire population and subsets.25
The girl ratio and number of participants in the NHANES and the KNHANES are displayed in Table 1. All subjects in the studies participated voluntarily and provided informed consent.
Measurement of Metabolic Risk Factors
The sample includes boy and nonpregnant girl participants 12 to 19 years who fasted for at least 8 hours. For NHANES 1988–1994, only those attending a morning examination session were used because the continuous NHANES only used blood samples drawn in the morning. The procedure described by Ford et al3 was followed to maintain the consistency of blood pressure (BP) data between surveys.
In the KNHANES, anthropometric measurements were consistent across all surveys. Height was measured to the nearest 0.1 cm by using a portable stadiometer (Seriter, Bismarck, ND); weight to the nearest 0.1 kg by using a calibrated balance beam scale (Giant-150N, HANA, Seoul, Korea); waist circumferences were taken at the end of normal expiration to the nearest 0.1 cm, measuring from the narrowest point between the lower borders of the rib cage and the iliac crest; and BP was measured twice by using a mercury sphygmomanometer (Baumanometer, Baum Co, Inc, Copiague, NY) in the sitting position after at least 10 minutes of rest. If the first 2 measurements differed by more than 5 mm Hg, additional checks were obtained.
Fasting plasma concentrations of glucose and lipids from all participants were measured enzymatically with a Hitachi 747 chemistry analyzer (Hitachi, Tokyo, Japan) in 1998 and 2001, and using an ADVIA 1650/2400 (Siemens, New York, NY) in 2005 and 2007. To confirm and compare accuracy and consistency in each survey, commutable frozen serum samples (n = 38) from normal subjects and patients with dyslipidemia was sent to the Centers for Disease Control and Prevention (Atlanta, GA) and were measured by using the standard method. A Passing and Bablok regression method was then used for the conversion rate for the KNHANES 2007 (Supplemental Fig 2). The revised high-density lipoprotein (HDL) cholesterol values were statistically substantiated based on the extrapolated fitted regression line (R2 = 0.977). Conversion rates for other surveys were obtained similarly.
With the NHANES and the KNHANES, participants were categorized based on their age at the time of interview into 3 age groups: 12 to 14 years, 15 to 17 years, and 18 to 19 years. Participants with ≥90th percentile of waist circumference for each age group were identified with abdominal obesity, separately for boys and girls. The quartile groups for height were identified among the participants for each age group separately for each gender. Participants with a BP reading ≥90th percentile of either systolic BP (SBP) or diastolic BP (DBP) for each age group, gender, and height were identified as having high BP.
Definition of MetSyn
Defining MetSyn was based on extrapolation from NCEP criteria9 as the presence of 3 or more of the following: hypertriglyceridemia (serum triglycerides ≥110 mg/dL) or current lipid lowering medication; high blood glucose (fasting blood glucose ≥110 mg/dL) or current antidiabetic medication; low HDL cholesterol (≤40 mg/dL); abdominal obesity (waist circumference ≥90th percentile for age and gender); and high BP (≥90th percentile for age, gender, and height) or current high BP medication. The triglyceride and HDL cholesterol thresholds have been previously used in the definition of pediatric MetSyn.9 The hyperglycemia threshold was based on the NCEP-ATP III cutoff point and has also been used in the definition of MetSyn in this population.9
All data are presented as means ± SE or as prevalence (% and SE). Sampling weights were used to account for complex sampling. The age-adjusted and age-specific prevalence of individual metabolic abnormalities was calculated for the NHANES 1988–1994, 1999–2002, and 2003–2006 and the KNHANES 1998, 2001, 2005, and 2007. The statistical significance of the change in MetSyn prevalence between the surveys was examined by using Student’s t test, in which the square root of the sum of the squared SEs was used to calculate the pooled SE of the difference in the mean.
The reference population used for age-adjusted prevalence was US and Korean youth between 12 and 19 in the year 2000. Age-adjusted and age-specific prevalence of 1 or more metabolic abnormalities was also calculated. The subsequent absolute and relative changes in prevalence of MetSyn among each study were also assessed in both populations. The change in prevalence of the MetSyn between the 2 surveys was examined by using Student’s t test, with the square root of the sum of the squared SEs used to calculate the pooled SE of the difference in the mean. Due to a smaller sample size and more stringent inclusion criteria, we restricted our estimation by pooled sample only. Statistical analyses to calculate the prevalence were performed by using the complex sample analysis in SPSS software version 16.0 (SPSS Inc, Chicago, IL). Statistical significance was defined as P < .05.
Table 2 reveals the age specific, unadjusted, and adjusted descriptive statistics of anthropometric and metabolic variables of both the NHANES and the KNHANES.
The mean age of the NHANES sample was between 15.4 and 15.7 years, and between 14.9 and 15.4 years for the KNHANES. The NHANES population varied from 47.7% to 48.5% girl participants, whereas KNHANES was 46.6% to 49.1% girl participants. The age-adjusted mean weight increased almost 3 kg in the NHANES (63.0 ± 0.9 kg to 65.9 ± 0.9 kg), and ∼2.5 kg in the KNHANES (55.2 ± 0.4 kg and 57.6 ± 0.6 kg).
The mean standing height of the participants did not change much in either the NHANES (∼167 cm) or the KNHANES (∼163 cm). The age adjusted mean BMI increased by 1 in both the NHANES (22.5 ± 0.3 in the NHANES 1988–1994 to 23.5 ± 0.3 in the NHANES 2003–2006) and the KNHANES (20.4 ± 0.2 in the KNHANES 1998 to 21.4 ± 0.2 in the KNHANES 2007). The age-adjusted mean value of waist circumference increased more than 4 cm in the NHANES (77.5 ± 0.7 cm in the NHANES 1988–1994 to 81.7 ± 0.8 cm in the NHANES 2003–2006) and ∼2 cm in the KNHANES (70.3 ± 0.2 cm in the KNHANES 1998 to 72.2 ± 0.6 cm in the KNHANES 2007). Age-adjusted mean value of SBP increased in the NHANES (107.4 ± 0.6 mm Hg in the NHANES 1988–1994 to 109.5 ± 0.6 mm Hg in the NHANES 2003–2006) and decreased in the KNHANES (111.7 ± 0.3 mm Hg in the KNHANES 1998 to 105.6 ± 0.6 mm Hg in the KNHANES 2007). The age-adjusted mean value of DBP decreased in both the NHANES (63.0 ± 0.7 mm Hg in the NHANES 1988–1994 to 60.6 ± 0.5 mm Hg in the NHANES 2003–2006) and the KNHANES (68.2 ± 0.3 mm Hg in the KNHANES 1998 to 67.1 ± 0.6 mm Hg in the KNHANES 2007). The age-adjusted mean value of total cholesterol decreased in the NHANES (161.5 ± 1.6 mg/dL in the NHANES 1988–1994 to 160.2 ± 1.0 mg/dL in the NHANES 2003–2006) and increased in the KNHANES (157.3 ± 0.6 mg/dL in the KNHANES 1998 to 159.7 ± 1.6 mg/dL in the KNHANES 2007). No consistent trend was noticed for the age-adjusted mean values of triglycerides, HDL cholesterol, and low-density lipoprotein (LDL) cholesterol in the NHANES, or for triglycerides in the KNHANES. The KNHANES population did reveal a decrease in HDL cholesterol (49.8 ± 0.3 mg/dL in the KNHANES 1998 to 43.5 ± 0.7 mg/dL in the KNHANES 2007) and an increase in LDL cholesterol (89.3 ± 0.8 mg/dL in the KNHANES 1998 to 96.5 ± 1.2 mg/dL in the KNHANES 2007). The age-adjusted mean values of fasting blood glucose increased in the NHANES (89.8 ± 0.6 mg/dL in the NHANES 1988–1994 to 92.7 ± 0.6 mg/dL in the NHANES 2003–2006) and decreased in the KNHANES (88.7 ± 1.2 mg/dL in the KNHANES 1998 to 87.4 ± 0.6 mg/dL in the KNHANES 2007).
Table 3 reveals the age-adjusted prevalence of MetSyn in the KNHANES significantly increasing ∼0.4% annually (1998: 4.0% ± 1.4%; 2001: 5.9% ± 1.1%; 2005: 6.6% ± 1.2%; 2007: 7.8% ± 1.4%), whereas the prevalence decreased from 7.3% ± 1.4% in the NHANES 1988–1994 to 6.7% ± 0.8% in the NHANES 1999–2002 to 6.5% ± 0.8% in the NHANES 2003–2006.
Table 3 also reveals the prevalence (95% confidence interval) of individual metabolic abnormalities of MetSyn in the NHANES and the KNHANES. In the NHANES, abdominal obesity and hypertriglyceridemia slightly increased while prevalence of low HDL cholesterol, high BP, and elevated fasting glucose decreased significantly across the 3 study periods. In the KNHANES, abdominal obesity, hypertriglyceridemia, and low HDL cholesterol increased, whereas high BP and fasting glucose remained consistent across the 4 studies. Figure 1 reveals the ∼10-year secular trend of each component of MetSyn in the NHANES and the KNHANES.
The NHANES participants had a higher prevalence of high BP, whereas the KNHANES had a higher prevalence of high fasting glucose. Component clustering analysis within the KNHANES (Table 3) revealed the prevalence of participants with at least 1 metabolic abnormality increased from 47.9% in 1998 to 55.7% in 2007, whereas the number of participants with 4 or 5 components increased from 0.9% to 2.1%, respectively. In the NHANES (Table 3), the participant prevalence of at least 1 metabolic abnormality decreased from 56.9% in NHANES 1988–1994 to 45.4% in the NHANES 2003–2006, as did the number of subjects with 4 or 5 components, from 1.9% in the NHANES 1988–1994 to 1.5% in the NHANES 2003–2006.
Few published studies exist investigating the prevalence of MetSyn in youth,26–28 with little consistency in MetSyn criteria. Our study used the widely accepted NCEP-ATP III-derived definition and indicates that age-adjusted prevalence of MetSyn in American adolescents decreased from the NHANES 1988–1994 to the NHANES 2003–2006. In contrast, Korean adolescents’ MetSyn prevalence increased almost twofold from the KNHANES 1998 to the KNHANES 2007.
These results, however, are inconsistent with previously published MetSyn rates. Ferranti et al26 used lower cutoff points for hypertriglyceridemia and waist circumference to show a MetSyn prevalence of 9.5% in boys and 8.2% in girls in the NHANES III. Johnson et al28 used the NHANES data and reported a low prevalence of MetSyn among African American adolescents; however, the current study did not differentiate by ethnicity. Deboer et al27 reported MetSyn prevalence was 8.1% in the NHANES 1999–2000 and 9.6% in the NHANES 2005–2006, using different criteria than the current study.
Among the components of MetSyn, there was no consistent trend to explain the overall decreasing prevalence in American adolescents. Low HDL cholesterol and high BP decreased, and abdominal obesity and high triglycerides maintained or slightly increased. The prevalence of high BMI revealed no significant changes between 2003–2004 and 2005–2006.29
In contrast, the Korean doubling of MetSyn may be partly explained by a 10.5% absolute increase in the prevalence in low HDL cholesterol. Hypertriglyceridemia (6.2%) and abdominal obesity (2.9%) may have also contributed due to their respective increases.
The prevalence of high fasting glucose and high triglycerides were higher in the Koreans, whereas high BP was higher in Americans. In addition, we have previously reported that the prevalence of MetSyn in Korean adults was comparable with that in US adults, with some components of MetSyn even more prevalent in Korea than in the United States.30
Overall, approximately half of the adolescents in both countries had 1 or more metabolic abnormality, with prevalence decreasing in Americans and increasing in Koreans. Furthermore, the proportion of adolescents with 4 or more abnormalities more than doubled in Koreans while decreasing in Americans.
The NHANES and the KNHANES data reveal clear and opposing trends for MetSyn prevalence between American and Korean adolescents. What is striking is that in Korea, the last decade has seen a rapid progression of the economy, including a high calorie/fat diet, and low physical activity,17,31 similar risk factors as noted in American adolescents.32–34 The authors of several studies indicated that Korean children and adolescents’ dietary behaviors are only recently changing from a traditional, rice and plant-based diet, to a Western, high fat and animal food diet.35,36 In Korean adolescents, there was an increasing trend of daily fat intake (Supplemental Fig 3), whereas the proportion of adolescents reporting regular walking has been decreasing and time to complete long distance running tests has been increasing (Supplemental Fig 4).37 The prevalence of MetSyn was positively correlated with fat intake and the long-distance running time, and negatively correlated with proportions of regular walking. In addition, physical education classes in Korean middle and high schools have been reduced by 5% to 20% due to a new curriculum system.38 Indeed, blood lipids and abdominal obesity are affected by lifestyle factors such as diet and physical activity levels,39,40 and changing to a high fat/calorie diet with low levels of physical activity in adolescents is a likely contributor to a high prevalence of dyslipidemia and abdominal obesity.41,42 Thus, gradual changes toward high-fat intake and less physical activity are occurring in Korean adolescents, which can contribute to increases in obesity and blood lipids.39–42 These changes may eventually result in adverse cardiovascular outcomes in early adulthood.
In contrast, the US increase in these factors began decades ago and have since plateaued or decreased.43 This may be partly due to a decades-long decline in the average percentage of calories from total fat and from saturated fat,44 and widespread awareness campaigns addressing obesity risks and the benefits of eating well and being active.45–47 Physical activity campaigns (eg, VERB, Walking School Bus) and efforts to improve school food environments may finally be yielding positive results in US adolescents.48,49
Judging from the substantial increase in Korean MetSyn prevalence, a national strategy to counter this trend is warranted, particularly in light of data revealing that as many as 65% of overweight children and 82% of obese children become obese adults.50 A change in dietary habit back to the more traditional low-calorie vegetable-based foods,35 along with emphasizing regular exercise, should be a central component of the national strategy. This has been shown effective elsewhere to prevent increased levels of metabolic abnormalities.39,40
There are several limitations in this study. First, our study defined MetSyn based on extrapolation from the NCEP criteria9; however, the global criteria for MetSyn have not been well characterized in adolescents. Particularly, there is no clear consensus on percentile cutoff levels for defining abdominal obesity in children and adolescents. Second, this is an analysis of cross-sectional surveys; secular changes of MetSyn associated with lifestyle changes cannot be determined.
Conversely, unlike the vast majority of studies testing obese children and adolescents,51,52 our study focused on representative data from a national survey and used criteria specifically for adolescents.9 Indeed, the increasing trends of MetSyn in Korea are obvious at the national/population level, yet population level findings do not necessarily translate to the individual level.53 However, recent individual level studies have shown corroborative data. Hong et al54 recruited 246 Korean children from local elementary schools and showed that low physical activity and high fat diet were independent predictors for the clustering of metabolic risk factors. A second study applying a semiquantitative food frequency questionnaire to 1441 preschool-aged children also revealed that animal food consumption was related with overweight.55
The prevalence of MetSyn in American adolescents decreased, in large part due to decreases in low HDL cholesterol and high BP. In stark contrast, the prevalence of MetSyn in Korean adolescents increased significantly over a similar time span, largely due to an increase in dyslipidemia and abdominal obesity. The results reveal that children in Asian countries (at least in Korea) are not safe from metabolic abnormalities and are in fact more vulnerable to certain factors than are current Western adolescents.
- Accepted August 20, 2012.
- Address correspondence to Gary Liguori, PhD, FACSM, Department Head, Health and Human Performance, University of Tennessee, Chattanooga, Chattanooga, TN 37403. E-mail:
Drs Lim and Liguori had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis; Drs Lim and Liguori contributed to the study concept and design; Drs Lim, Jang, Park, Cho, and Mozumdar contributed to the analysis and interpretation of data; Drs Lim and Liguori drafted the article; Drs Jang, Park, Cho, Lee, Joung, and Mozumdar contributed to the critical revision of the article for important intellectual content; Drs Lim, Mozumdar, and Liguori contributed to the statistical analysis; and all coauthors have participated in the current study and written the article, and the authors have approved the final version.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by the National Research Foundation grant funded by the Korea government (MEST) (2006-2005410) and Seoul National University Bundang Hospital to Dr Lim. The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the sponsors.
- Wilson PW,
- D’Agostino RB,
- Parise H,
- Sullivan L,
- Meigs JB
- Chen W,
- Srinivasan SR,
- Li S,
- Xu J,
- Berenson GS
- Rodríguez-Morán M,
- Salazar-Vázquez B,
- Violante R,
- Guerrero-Romero F
- López-Capapé M,
- Alonso M,
- Colino E,
- Mustieles C,
- Corbatón J,
- Barrio R
- ↵Annual Report on the Vital Statistics in Korea. National Statistical Office (in Korean). Ministry of Strategy and Finance, Republic of Korea. 2001
- He Y,
- Ma G,
- Zhai F,
- et al
- Jeon CY,
- Lokken RP,
- Hu FB,
- van Dam RM
- Mozumdar A,
- Liguori G
- Park HS,
- Oh SW,
- Cho SI,
- Choi WH,
- Kim YS
- ↵Brogan D. Software for sample survey data, misuse of standard packages. In Encyclopedia of Biostatistics. 1998;5. P. Armitage and T. Colton, eds. New York: Wiley; 4167–4174
- de Ferranti SD,
- Gauvreau K,
- Ludwig DS,
- Neufeld EJ,
- Newburger JW,
- Rifai N
- Lim S,
- Shin H,
- Song JH,
- et al
- Report on the Social Statistics Survey in Korea
- Kim S,
- Moon S,
- Popkin BM
- ↵Korean Statistical Information Service. KOSIS. Korea National Statistical Office. 2012. Available at: http://kosis.kr/abroad/abroad_01List.jsp. Accessed October 17, 2012
- ↵Curriculum & Textbook Information Services. Korea Ministry of Education, Science and Technology. 2011. Available at: http://cutis.mest.go.kr/main.jsp?gCd=S02&siteCmsCd=CM0001. Accessed October 17, 2012
- Young-Hyman D,
- Schlundt DG,
- Herman L,
- De Luca F,
- Counts D
- ↵Freedman DA. Ecological inference and the ecological fallacy. In: International Encyclopedia of the Social and Behavioral Sciences. Smelser NJ, Baltes PB, eds. New York. Elsevier. pp. 4027–4030, 2001
- Copyright © 2013 by the American Academy of Pediatrics