OBJECTIVE. The goal was to assess the association of metabolic syndrome in childhood with adult cardiovascular disease 25 years later.
METHODS. Data from the National Heart, Lung, and Blood Institute Lipid Research Clinics Princeton Prevalence Study (1973–1976) and the Princeton Follow-up Study (2000–2004) were used. BMI was used as the obesity measure in childhood, because waist circumference was not measured in the Lipid Research Clinics study. The adult cardiovascular disease status of participants and their parents was obtained through participant report. A logistic analysis was used to predict adult cardiovascular disease; pediatric metabolic syndrome, age at the Princeton Follow-up Study, gender, race, and parental history of cardiovascular disease were potential explanatory variables.
RESULTS. Ages ranged from 6 to 19 years in the Lipid Research Clinics study and from 30 to 48 years in the Princeton Follow-up Study. There were 17 cases of cardiovascular disease in the analysis cohort in the Princeton Follow-up Study. Pediatric metabolic syndrome and age at follow-up assessment were significant predictors of cardiovascular disease. Pediatric metabolic syndrome and changes in age-specific BMI percentile from childhood to adulthood were significant predictors of adult metabolic syndrome.
CONCLUSIONS. Evaluating children for metabolic syndrome could identify patients at increased risk of adult cardiovascular disease, making targeted interventions possible.
In adults, the clustered presence of ≥3 cardiovascular disease (CVD) risk factors from among elevated waist circumference, blood pressure, triglyceride levels, and glucose levels and low high-density lipoprotein cholesterol (HDL-C) levels has been designated the “metabolic syndrome.”1 Although labeling this collection of risk factors as a syndrome is controversial,2,3 their combined presence is associated with increased risk of future CVD in adults.4–7 The cutoff values used in defining abnormal levels1 are not extreme, generally falling near the 80th to 85th percentiles for adult reference populations.8 Whether the clustered presence of these factors in childhood predicts adult CVD is not known, although a number of studies have reported that individual factors track from childhood into young adulthood.9,10 Moreover, the Bogalusa Heart Study, which was initiated before Kissebah et al11 identified central adiposity as a predictor of diabetes mellitus and other metabolic disorders, reported that persistent high insulin levels12 and overweight, identified on the basis of BMI instead of waist circumference,13 were associated with higher levels of the factors in the metabolic syndrome later in life. Those reports12,13 did not assess CVD. The longitudinal nature of the data from the Lipid Research Clinics (LRC) Princeton Prevalence Study and the Princeton Follow-up Study (PFS) provides an opportunity to assess the association between the metabolic syndrome in childhood and adult CVD ∼25 years later. We hypothesized that the presence of the metabolic syndrome in childhood would predict CVD in adulthood.
The PFS participants were drawn from the Cincinnati clinic of the National Heart, Lung, and Blood Institute LRC Prevalence Study, a multistage survey of lipid levels in 10 communities in the United States and Canada that was conducted in 1973 to 1978 to describe lipid level distributions in free-living populations and their associations with other CVD risk factors. The Cincinnati clinic conducted its prevalence study in the public and parochial schools in the Princeton School District of Greater Cincinnati, targeting students in grades 1 to 12.14,15 Briefly, the original student population was 73% white and 27% black, 52% male and 48% female.14 Race was self-declared as white or black. In addition, the parents in a 50% sample of Princeton Prevalence Study households were recruited to the study. At an initial visit, total cholesterol and triglyceride levels were measured, basic demographic information was collected, and the relationships between participants and other participating members of their households were established.14 Approximately 6 weeks later, randomly selected and hyperlipidemic visit 1 subjects were recalled, independent of other family members, for a second screening, at which complete lipid profiles, blood chemistry values, anthropometric data, blood pressure, and family history of CVD were recorded.15 In 1976, the LRC Family Study (visit 3) screened all consenting first-degree relatives of selected visit 2 subjects with respect to lipid levels, anthropometric data, and blood chemistry values, to assess familial correlations of lipid levels.16 The PFS, which was conducted to evaluate long-term changes in familial lipid correlations, recruited all visit 3 participants and visit 2 participants with ≥1 sibling or parent also at visit 2.17 The time between LRC study and PFS visits ranged between 22 and 31 years, depending on when the subjects attended their LRC study and PFS visits. In the LRC study, 84% of eligible students participated at the initial LRC study visit and 91% of eligible students participated at subsequent visits; participation rates did not differ significantly between races.14–16
In both the LRC study and the PFS, data were collected by using standard protocols.15–17 Height and weight were measured with the subjects in light indoor clothing, with shoes removed. In the LRC study, single measurements of height and weight were made. In the PFS, 2 measurements of height, weight, and waist circumference were made, with a third measurement if the first 2 measurements differed by more than a set amount. The means of the 2 closest PFS measurements were used in analyses. The BMI was used to characterize body habitus. In the PFS, waist circumference was measured at the level of the umbilicus, at the end of a normal expiration. In the LRC study and PFS, blood pressure was measured on the right arm with a standard sphygmomanometer, after participants had been sitting for 5 minutes. Systolic blood pressure and diastolic blood pressure were defined with the appearance of sound (first Karotkoff phase) and the disappearance of sound (fifth Karotkoff phase), respectively. The mean of 2 readings was used for the LRC study, and the mean of the second and third readings was used for the PFS. In each study, fasting blood samples were drawn into Vacutainer tubes (Becton Dickinson and Co, Franklin Lakes, NJ) containing EDTA, kept on wet ice (LRC study) or cold packs (PFS), and delivered to the laboratory within 3 hours for processing; lipid profiles were measured in LRC/Centers for Disease Control and Prevention (CDC)-standardized laboratories. In the LRC study, glucose levels were measured with an ABA-100 system by using the hexokinase method.18 In the PFS, glucose levels were measured with a Dade Dimension Xpand system (Dade Behring Inc, Deerfield, IL) by using the hexokinase/glucose-6-phosphate dehydrogenase method.19
Definitions of Metabolic Syndrome and CVD
Adult metabolic syndrome was defined by using Adult Treatment Panel III criteria, as follows: waist circumference of ≥88 cm (female subjects) or ≥102 cm (male subjects), HDL-C levels of ≤50 mg/dL (female subjects) or ≤40 mg/dL (male subjects), triglyceride levels of ≥150 mg/dL, blood pressure of ≥130 mm Hg (systolic) or ≥85 mm Hg (diastolic), and glucose levels of ≥110 mg/dL.1 CVD (myocardial infarction, coronary artery bypass graft, angioplasty, or stroke) in participants and their parents was determined through report of the participants at the time of the PFS. BMI was used to define obesity in childhood, because waist circumference was not measured in the LRC study. Pediatric standards were used to define abnormal triglyceride levels, BMI, and blood pressure, because pediatric distributions of these factors differ markedly from adult distributions.20 Triglyceride levels of ≥110 mg/dL were defined as elevated, and BMI values at or above the age-specific 90th percentile, based on the CDC 2000 growth charts,21 were defined as high. Systolic or diastolic blood pressure at or above the age- and height-specific 90th percentile21 was defined as elevated. HDL-C levels of ≤50 mg/dL (female subjects) or ≤40 mg/dL (male subjects) were defined as low.20 Glucose levels of ≥110 mg/dL were defined as elevated.20
SAS 9.1.322 was used for all analyses. A logistic analysis for clustered sample design was used to identify factors related to CVD status (yes/no) at follow-up assessment, taking into account sibling correlations. Age at PFS, gender, race, parental CVD, and pediatric metabolic syndrome status were potential explanatory variables. Predictors with P values of <.15 were retained in the model. Because adult metabolic syndrome is associated with significantly greater risk for CVD, we also determined the risk of metabolic syndrome in adulthood associated with its presence in childhood. To evaluate intervening change in obesity as a predictor of CVD and adult metabolic syndrome, the change in the age-specific BMI percentiles between the 2 studies was tested as a covariate. The pediatric BMI percentiles in the LRC study were taken from the CDC 2000 growth charts,21 and the adult percentiles were taken from the compilation of National Health and Nutrition Examination Survey I and II data by Frisancho.23 Finally, to evaluate elevated glucose levels in childhood as predictors of metabolic syndrome and CVD (separately), we tested the associations by using LRC study glucose levels of ≥100 mg/dL and ≥110 mg/dL as cutoff values, with Fisher's exact test.
The ethnic composition of the 771 members in the analysis cohort (28% black and 72% white) was similar to that of the total school population in the LRC study (27% black and 73% white); the proportion of female subjects (55%) was greater. In the LRC study, the cohort ranged in age from 5 years to 19 years, with a mean of 12.9 years (Table 1). The mean BMI for the students was 19.8 kg/m2; 18.2% of the students were at or above the CDC 2000 age-specific 85th percentile (at risk of overweight), and 7.0% were at or above the 95th percentile (overweight). The prevalence rates of abnormal BMI (13.7%), blood pressure (11.9%), triglyceride levels (12.3%), and HDL-C levels (13.0%) in the LRC study were slightly higher than the 10% expected from the use of age-adjusted 90th or 10th percentile cutoff values (Table 2). The prevalence of abnormal glucose levels (0.7%) was low, because the cutoff value was the same as for adults (110 mg/dL). Thirty-one LRC study participants (4.0%) had ≥3 abnormal factors and were classified as having metabolic syndrome. Of the 31 students with the syndrome, 24 (77%) had BMI values above the 90th percentile and an additional 4 (13%) had BMI values above the 85th percentile but below the 90th percentile.
At follow-up assessment, the mean age of the cohort was 38.4 years, the mean BMI was 28.6 kg/m2 (approximately the 75th percentile of the National Health and Nutrition Examination Survey I and II distribution), and 25.4% of the cohort had BMI values at or above the age- and gender-specific 90th percentile (Table 1). Commensurate with the increase in BMI, the prevalence of low HDL-C levels, elevated triglyceride levels, elevated blood pressure, and elevated fasting glucose levels increased to 52.9%, 28.3%, 33.8%, and 6%, respectively. The prevalence of large waist circumferences was 49.1%, and 27.2% of the cohort had metabolic syndrome (Table 2). The syndrome was associated strongly with increased BMI; of the 210 cohort members with adult metabolic syndrome, 200 (95%) were overweight or obese (BMI of ≥25 kg/m2) and 151 (72%) were obese (BMI of ≥30 kg/m2). The overall tracking coefficient from childhood to adulthood was 0.59, and 63% of participants at risk of overweight in the LRC study were obese in the PFS. Of 31 LRC study subjects with pediatric metabolic syndrome, 21 (68%) had metabolic syndrome at follow-up assessments (χ2 = 26.7; P < .0001). In a multivariate analysis with age, gender, race, and pediatric metabolic syndrome as potential explanatory variables, pediatric syndrome status (odds ratio [OR]: 6.1; P < .0001) alone was significant; female gender (OR: 0.8; P = .10) was not significant at the P = .05 level. When the change in BMI percentile between the studies was added to the model, the factor was significant (OR: 1.024; P < .0001) but female subjects had a P value of 0.22, which suggests that part of the gender difference in the risk of adult metabolic syndrome was attributable to less weight gain among female subjects. For each increase in the age-specific BMI percentile of 10 points, the risk of adult metabolic syndrome increased 24%. Of the 10 subjects with metabolic syndrome in the LRC study but not the PFS, 2 had increases in BMI percentile (changes of +4.9 and +12.8), 7 had decreases in BMI percentile (range: −31.5 to −6.8), and 1 had no change.
There were 17 cases of CVD in the PFS analysis sample. The incidence of CVD during the intervening years for the 31 patients with pediatric metabolic syndrome was 19.4% (n = 6), compared with 1.5% for subjects without metabolic syndrome as children. In multivariate logistic analyses, pediatric metabolic syndrome (OR: 14.7; P < .0001) and age (OR: 1.2; P = .03) were significant predictors of CVD (Table 3), and gender, race, and parental history of CVD were not (P > .2). When change in BMI percentile was added to the final model, the factor was not significant (P > .5). Finally, CVD at follow-up assessment was not associated with glucose levels of ≥100 or ≥110 mg/dL in childhood. Of 5 LRC study subjects with glucose levels of ≥110 mg/dL, none had CVD; of 29 with glucose levels of ≥100 mg/dL, 2 had CVD (P = .13).
The key finding in this follow-up study of former students from the Princeton LRC School Study is that children with the cluster of factors defined as pediatric metabolic syndrome were significantly more likely to have CVD 25 years later as adults, compared with their peers. The prevalence of the syndrome in the students as children was 4%, similar to that reported by Cook et al.20 The finding that the syndrome in childhood predicts later CVD agrees with the recent report of a significant association between pediatric metabolic syndrome and subclinical atherosclerosis (determined as carotid intimal-medial thickness) in young adults with a mean age of 32 years in the Bogalusa Heart Study.24 Similar findings link combinations of these risk factors in childhood to subsequent carotid intimal-medial thickness in young adults.25,26 In the current study, the mean age (38.4 years) was 6 years older than that in the Bogalusa report, and clinical disease was becoming more prevalent.
At follow-up assessments, the prevalence of the syndrome was 27.2%, reflecting the marked increase in BMI and associated changes in obesity-related risk factors, a prevalence somewhat higher than that published by Ford et al27 (based on National Health and Nutrition Examination Survey III data from 10 years earlier). Given the well-established relationship between the constellation of adult risk factors identified as metabolic syndrome and the risk of future CVD,1,4–7 it is important to note that metabolic syndrome in adulthood was associated strongly with the presence of the syndrome in childhood (Table 3). Within this context, the change in BMI percentile between studies was also highly associated with the development of metabolic syndrome; for every increase (or decrease) in age-specific BMI percentile of 10 points, the risk of metabolic syndrome in adulthood increased (or decreased) 24%, which emphasizes the cost of greatly increasing weight in adult years and the benefit of losing relative weight in adulthood if one is overweight as a child or youth.
It is somewhat surprising that parental history of CVD was not associated with CVD, because family history is well established as a risk factor for CVD.1 The lack of an association here could be attributable to the small number of CVD cases (n = 17) in the cohort at this point or to the obesity explosion that has occurred in the past quarter century, changing the prevalence of obesity-related risk factors and potentially altering the factors explaining early CVD.
As noted above, there is controversy regarding labeling this set of factors as a syndrome. Reaven,2 who was largely responsible for focusing attention on insulin resistance as a risk factor for CVD, has argued against the label for 3 main reasons, that is, (1) it occurs only in insulin-resistant persons, (2) it has no clinical or pedagogical utility, and (3) the emphasis must be on treating the individual factors, because there is no treatment for the syndrome itself. In his review of the issue, Goetz28 suggested that using the syndrome label might be tantamount to having Big Pharma turn obesity into a disease and then seeking to develop a cure for it, but he went on to write, “a diagnosis of metabolic syndrome can get them off the couch in a way that a doctor tut-tutting about their extra pounds cannot.”28 Moreover, treating the individual factors one at a time (with prescribed drugs) may constitute a less-efficient, less-effective approach to the problem than addressing the underlying problem, which on a population basis is obesity-related insulin resistance. The strong association of the syndrome in childhood and adulthood with high BMI highlights the role of obesity in the constellation of factors, whatever it is labeled.
Limitations of the study include the use of participant reports of CVD in the participants and their parents, instead of hospital records. However, Murabito et al29 compared offspring reports of parental CVD history in the Framingham Offspring Study with confirmed medical findings and reported that positive and negative history reports were reliably accurate for heart attack and stroke. Furthermore, the authors reported no difference in accuracy between male and female offspring. A second limitation is the lack of insulin data. Insulin data were collected for a subgroup as part of an ancillary study, but only 3 of the 17 CVD cases had insulin values, which precluded meaningful statistical analyses. A third limitation is the use of casual fasting glucose levels in childhood, instead of oral glucose tolerance data, as potential predictors of adult CVD.
The findings from this 25-year follow-up evaluation of former students in the LRC Princeton Prevalence Study, with and without pediatric metabolic syndrome in childhood, stress the potential importance of preventive actions for patients during the pediatric years and into adulthood. The increased risk of adult disease associated with the pediatric syndrome clearly identifies patients at risk, as well as the need for intervention. In addition, the significant changes in the risk of the syndrome in middle age that are associated with changes in BMI percentile rankings between the 2 studies underscore the importance of weight management in early and middle adult years.
- Accepted March 16, 2007.
- Address correspondence to John A. Morrison, PhD, OSB 4, Division of Cardiology, Children's Hospital Medical Center, 3333 Burnet Ave, Cincinnati, OH 45229. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Expert Panel on Detection, Evaluation, and Treatment of High Cholesterol in Adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Cholesterol in Adults (Adult Treatment Panel III). JAMA.2001;285 :2486– 2497
- ↵Reaven GM. The metabolic syndrome: is this diagnosis necessary? Am J Clin Nutr.2006;83 :1237– 1247
- ↵Kahn R. The metabolic syndrome (emperor) wears no clothes. Diabetes Care.2006;29 :1693– 1696
- ↵Dekker JM, Girman C, Rhodes T, et al. Metabolic syndrome and 10-year cardiovascular risk in the Hoorne Study. Circulation.2005;112 :666– 673
- Rutter MK, Meigs JB, Sullivan LM. Insulin resistance, the metabolic syndrome, and incident cardiovascular events in the Framingham Offspring Study. Diabetes.2005;54 :3252– 3257
- Wilson PW, D'Agostino RB, Parise H, et al. Metabolic syndrome as a precursor of cardiovascular disease and type 2 diabetes mellitus. Circulation.2005;112 :3066– 3072
- ↵Eberly LE, Prineas R, Cohen JD, et al. Metabolic syndrome: risk factor distributions and 18-year mortality in the Multiple Risk Factor Intervention Trial. Diabetes Care.2006;29 :123– 130
- ↵Lipid Research Clinics. Lipid Research Clinics Population Studies Data Book, Vol I, The Prevalence Study. Washington, DC: US Department of Health and Human Services; 1980. NIH Publication 80-1527
- ↵Webber LS, Srinivasan SR, Wattigney WA, Berenson GB. Tracking of serum lipids and lipoproteins from childhood to adulthood: The Bogalusa Heart Study. Am J Epidemiol.1991;133 :84– 99
- ↵Freedman DS, Dietz WH, Srinivasan SR, Berenson GS. The relation of overweight to cardiovascular risk factors among children and adolescents: the Bogalusa Heart Study. Pediatrics.1999;103 :1175– 1182
- ↵Bao W, Srinivasan SR, Berenson GS. Persistent elevation of plasma insulin levels is associated with increased cardiovascular risk in children and young adults. Circulation.1996;93 :54– 59
- ↵Morrison JA, deGroot I, Edwards BK, et al. Lipids and lipoproteins in 927 schoolchildren, ages 6–17 years. Pediatrics.1978;62 :990– 995
- ↵Aronson-Friedman L, Morrison J, Daniels SR, McCarthy WF, Sprecher DL. Sensitivity and specificity of pediatric lipid determinations for adult lipid status: a longitudinal assessment. Pediatrics.2006;118 :165– 172
- ↵Patten RL, Hewitt D, Waldman GT, Jones G, Little JA. Associations of plasma high-density lipoprotein cholesterol with clinical chemistry data. Circulation.1980;62(4 pt 2) :IV31– IV41
- ↵Sacks DB. Carbohydrates. In: Burtis CA, Ashwood ER, eds. Tietz Textbook of Clinical Chemistry. 3rd ed. Philadelphia, PA: Saunders; 1999:777– 778
- ↵Kuczmarski RJ, Ogden CL, Guo SS, et al. 2000 CDC growth charts for the United States: methods and development. Vital Health Stat 11.2002;(246) :1– 190
- ↵SAS Institute. SAS/STAT User's Guide, Version 9.1.3. Cary, NC: SAS Institute; 2005
- ↵Frisancho AR. Anthropometric Standards for the Assessment of Growth and Nutritional Status. Ann Arbor, MI: University of Michigan Press; 1993
- ↵Davis PH, Dawson JD, Riley WA, Lauer RM. Carotid intimal-medial thickness is related to cardiovascular risk factors measured from childhood through middle age: the Muscatine Study. Circulation.2001;104 :2815– 2819
- ↵Goetz T. 75 million Americans may have something called metabolic syndrome: how Big Pharma turned obesity into a disease and then invented the drugs to cure it. Wired.2006;14(10) :152– 157
- Copyright © 2007 by the American Academy of Pediatrics