Childhood predictors of the metabolic syndrome in adulthood. The Cardiovascular Risk in Young Finns Study

Abstract

Background. Obese youths may be susceptible to develop the metabolic syndrome (MS) later in life.

Aim. To study childhood predictors of MS in adulthood.

Method. Prospective cohort study including 2,195 subjects, aged 3–18 years at base-line in 1980, who were re-examined in 1983, 1986, and 2001.

Results. In adults (aged 24–39 years) in 2001, the prevalence of MS (using the International Diabetes Federation criteria) was 19% in men and 12% in women. Multivariable logistic regression model selected obesity, male sex, high triglycerides, high insulin, high C-reactive protein (CRP), and family history of hypertension and type 2 diabetes, as independent predictors of adult MS. Youth obesity (body mass index (BMI)>80^th age- and sex-specific percentile) was the strongest risk factor for MS. During the 21-year follow-up, there was an increasing trend in BMI, insulin, systolic blood pressure, and triglycerides, and a decreasing trend in high-density lipoprotein cholesterol in obese subjects who developed MS in adulthood compared to those obese subjects who did not develop MS.

Conclusions. Youth determinants of adult MS included obesity, high triglycerides, high insulin, high CRP, and family history of hypertension and type 2 diabetes. Identifying these risk factors at an early stage could help identifying children and adolescence at greater risk of developing MS later in life.

• Adolescence
• atherosclerosis
• childhood
• metabolic syndrome

Introduction

Cardiovascular diseases are the leading causes of mortality in the adult population worldwide. A constellation of cardiovascular risk factors including central obesity, hypertension, dyslipidemia, and impaired glucose metabolism with insulin resistance is commonly called the metabolic syndrome 1. The clustering of metabolic risk factors begins already in childhood 2–4 and these multiple risk factors tend to persist from childhood into adulthood 5. Nevertheless, overall this cluster tracking may not be particularly strong. We have previously shown in the Young Finns cohort that on average only about 25% of children and young adults initially at high risk (with all three risk factors, serum total cholesterol, high-density lipoprotein (HDL) cholesterol, and diastolic blood pressure in the extreme tertile) remained there after 6 years 6. In the 3-year follow-up of the Princeton School District Study, approximately half of adolescents with base-line metabolic syndrome lost the diagnosis at follow-up regardless of the definitions used 7. These observations indicate that there may be substantial instability in the diagnosis of the metabolic syndrome. Dichotomous approach may also exclude too much clinically important information. Thus, the clinical utility of the diagnosis may be reduced in young people, and alternative approaches are needed in primary prevention to identify young people who are at risk of developing the metabolic syndrome as adults.

Abbreviations

Table

BMI	body mass index
CHD	coronary heart disease
CRP	C-reactive protein
EGIR	European Group for the Insulin Resistance
HDL	high-density lipoprotein
IDF	International Diabetes Federation
LDL	low-density lipoprotein
MS	metabolic syndrome
NCEP	National Cholesterol Education Panel
SAS	Statistical Analysis System
SD	standard deviation
SEM	standard error of mean
TG	triglycerides
VLDL	very-low-density lipoprotein

Obesity is strongly associated with the insulin resistance and it increases the risk of cardiovascular disease morbidity in adults 8. Previously, the Harvard Alumni Health Study has shown a direct correlation between body weight and mortality 9. The prevalence of obesity is increasing worldwide in all age groups, including children and adolescents. According to the recent epidemiological studies, there has been a 3-fold increase in the prevalence of overweight in US children and adolescents during the past two decades 10. Ogden et al. 11 have shown that 17.1% of 2–19-year-old US children are overweight. Overweight children are at increased risk for adult obesity 12, 13. We have previously reported that the risk of being obese in adulthood (body mass index (BMI) > 30 kg/m²) was increased by 3-fold in overweight or obese (BMI > 80th percentile) children (ages 3–9 years) and by 4-fold in overweight or obese youths (ages 12–18 years) 12. Alarmingly, the incidence of type 2 diabetes and other morbidities reported in obese children has increased markedly in the past 20 years. Obesity is causally related to the metabolic syndrome. Weiss et al. 14 have examined the effects of varying degrees of obesity on the prevalence of the metabolic syndrome in a large cohort of children and adolescents. They observed that the prevalence of the metabolic syndrome increased and each element of the syndrome worsened directly with the degree of obesity. However, not all obese youths will go on to develop the metabolic syndrome. The pathophysiological mechanisms that predispose some obese individuals to develop the metabolic syndrome are not fully understood.

Key messages

• Childhood determinants of the adult metabolic syndrome included obesity, family history of hypertension, family history of type 2 diabetes, high triglycerides, high insulin, and high C-reactive protein.

• Identifying these risk factors in children and adolescents could be helpful in pediatric metabolic risk assessment.

In the present analysis, we used longitudinal data from the Cardiovascular Risk in Young Finns Study to examine the association of childhood and adolescence risk variables and adulthood metabolic syndrome. In addition, to gain insight of the changes in risk profile during maturation in obese individuals who develop the adult metabolic syndrome, we compared serial changes in metabolic variables and life-style from childhood to young adulthood between obese individuals who developed and those who did not develop metabolic syndrome in adulthood.

Materials and methods

Subjects

The Cardiovascular Risk in Young Finns Study is an on-going epidemiological study of cardiovascular disease risk factors and their determinants from childhood to adulthood. The original sample size was 4,320 healthy Caucasian children and adolescents aged 3, 6, 9, 12, 15, and 18 years, and a total of 3,596 (83.2%) subjects participated in the first cross-sectional study in 1980. The follow-up studies were performed for the whole study group in 1983, 1986, and 2001, when 2,991 (83.2%), 2,799 (78.3%), and 2,624 (75.9%) subjects participated. Details of the data collection and the study design have been presented previously 15. The analyses were restricted to non-pregnant subjects. In the present analysis, our study population with complete data included 2,195 subjects (1,014 men and 1,181 women) aged 3 to 18 years in 1980, and 24 to 39 years in 2001.

Clinical characteristics

Height and weight were measured, and body mass index (BMI) was calculated. In 1980, 1983, and 1986, biceps, triceps, and subscapular skinfold thicknesses were measured in triplicate to the nearest millimeter from the non-dominating arm using Harpenden skinfolds calipers (Holtain and Bull, British Indicators Ltd, Luton, Beds, UK). Their sum variable was used in statistical analysis. Blood pressure was measured from the brachial artery with a standard mercury sphygmomanometer in 1980 and 1983, and with a random-zero sphygmomanometer in 1986 and 2001. The average of three measurements was used in statistical analysis.

Data on smoking, alcohol consumption, physical activity, family history of coronary heart disease, hypertension, or type 2 diabetes, and socio-economic status were obtained from questionnaires. The data on breast-feeding and birth measurements (height and weight) were obtained from the parents by a questionnaire in 1983 when the subjects were 6 to 21 years old. The information on smoking habits was collected in subjects aged 12 years or older. Subjects smoking on a weekly basis or more often were considered as smokers, and those reporting consuming beer, wine, or liquor once a week or more often were classified as regular users of alcohol. Physical activity measurements were done in 1980 for subjects at 9 years or older. The physical activity index was calculated as previously explained 16.

A positive family history was defined as mother and/or father having coronary heart disease, hypertension, or diabetes mellitus. Metabolic syndrome in 2001 was defined using the International Diabetes Federation (IDF) 17, the National Institute of Health Adult Treatment Panel III (NCEP) 18, and the European Group for the Study of Insulin Resistance (EGIR) 19 guide-lines. The data are shown using the IDF criteria, but similar results were seen using any of the criteria. A subject's socio-economic status was determined using information on parental education, occupation, and family's income as previously described 20.

Blood samples

For the determination of serum lipid levels, venous blood samples were drawn after an overnight fast. All of the lipid determinations were done using standard methods 21, 22. In 1980, 1983, and 1986 serum insulin was measured using a modification of the immunoassay method of Herbert et al. 23. In 2001, serum insulin was measured by a microparticle enzyme immunoassay kit (Abbott Laboratories, Diagnostic Division, Dainabot), and glucose concentrations were analyzed enzymatically (Olympus Diagnostica GmbH, Hamburg, Germany). Childhood serum samples taken in 1980 were stored at −20°C. In 2005, they were used to determine serum high-sensitive C-reactive protein (CRP) analyzed by an automated analyzer (Olympus AU400) using a turbidimetric immunoassay kit (‘CRP-UL’ assay, Wako Chemicals, Neuss, Germany). During the storage, the samples were not thawed or refrozen. The long-term stability of CRP has been previously reported. We excluded subjects with CRP levels more than 10 mg/L. Details have been presented elsewhere 24.

Statistical methods

Data are presented as means±SD for continuous variables and as percentages for categorical variables, unless otherwise stated. Group comparisons were performed using t test for continuous variables and chi-square test for categorical variables. Stepwise logistic regression analysis was done to establish the independent childhood determinants of the adulthood metabolic syndrome. In multivariable model, risk factors were defined as values at or above the age- and sex-specific 80th percentile for BMI, systolic blood pressure, triglycerides, insulin, CRP, and at or below the age- and sex-specific 20th percentile for HDL cholesterol.

The primary aim was to examine study childhood determinants of the adulthood metabolic syndrome. In addition, we also present the childhood and adolescence risk factor data by the presence or absence of adulthood metabolic syndrome and by the number of adulthood metabolic syndrome components.

The longitudinal association of childhood obesity status with risk variables was assessed by repeated measures analysis of variance (adjusted for age, sex, and pubertal status) to study childhood risk factors in obese subjects with and without the metabolic syndrome in adulthood and to compare these groups at different time points. Values for triglycerides, insulin, and CRP were log-transformed before analysis because of skewed distributions.

All statistical analyses were done stratified by two groups according to the base-line age when the risk factor assessments were done—in 3–9- and 12–18-year-olds. This age stratification paralleled pubertal staging, as 85% of 12–18-year-olds at base-line were classified having puberty on-going or completed 25.

All tests were performed with Statistical Analysis System (SAS) version 8.1, and statistical significance was inferred at a two-tailed P < 0.05.

Results

In adulthood (year 2001), the prevalence of the metabolic syndrome (using the IDF criteria) was 19% in men and 12% in women. Subjects with the metabolic syndrome had higher waist circumference than those subjects without the syndrome (mean±SD; 100±11 cm versus 81±10 cm, P<0.0001). The base-line characteristics in 1980 of the study subjects, stratified by base-line age group, with and without IDF metabolic syndrome in adulthood are shown in Table I. Subjects who had the metabolic syndrome in adulthood had higher skinfold thickness, BMI, systolic blood pressure, triglycerides, insulin, and lower HDL cholesterol in childhood than subjects without the metabolic syndrome. In males, the adulthood metabolic syndrome was associated with older age and higher CRP levels in youth. The metabolic syndrome in adulthood was generally more prevalent in subjects with positive family history of hypertension or type 2 diabetes.

Table I.  Childhood and adolescent characteristics of study subjects with and without the IDF metabolic syndrome diagnosis in adulthood.

	3–9 years at base-line				12–18 years at base-line
	Men		Women		Men		Women
Characteristics in childhood and in adolescence	IDF −	IDF +	IDF −	IDF +	IDF −	IDF +	IDF −	IDF +
n	434	59	507	53	392	129	530	91
Age (years)	6.1±2.5	7.3±2.1^e	6.1±2.4	6.3±2.6	14.9±2.4	15.0±2.4	14.8±2.4	15.0±2.5
Skinfold thickness (mm) ^a	19.9±6.2	24.4±10.4^d	23.4±7.9	30.4±13.3^e	22.4±10.0	28.8±13.2^e	33.2±11.7	38.8±14.9^d
BMI (kg/m^2)	15.8±1.5	17.1±2.4^e	15.7±1.7	17.0±2.0^e	19.4±2.7	21.2±3.5^e	19.5±2.6	20.8±3.4^e
Systolic blood pressure (mmHg)	107±11	111±9^d	107±10	110±11^c	119±12	122±13^d	116±10	118±9^c
Diastolic blood pressure (mmHg)	67±9	70±9^c	67±9	69±9	70±9	71±11	69±9	71±9
Total cholesterol (mmol/L)	5.3±0.9	5.4±0.8	5.5±0.9	5.5±0.9	5.1±0.9	5.0±0.9	5.2±0.9	5.5±0.9^c
LDL cholesterol (mmol/L)	3.5±0.8	3.5±0.7	3.7±0.8	3.7±0.7	3.3±0.8	3.2±0.9	3.3±0.8	3.6±0.8^d
HDL cholesterol (mmol/L)	1.60±0.31	1.59±0.33	1.56±0.30	1.46±0.30^c	1.53±0.31	1.40±0.29^e	1.60±0.29	1.49±0.25^d
Triglycerides (mmol/L)	0.57±0.25	0.65±0.39	0.61±0.23	0.75±0.31^e	0.68±0.30	0.80±0.41^d	0.72±0.33	0.88±0.36^e
Insulin (mU/L)	6.0±4.0	8.0±4.4^e	6.5±3.7	7.7±5.3	11.0±5.2	13.3±6.7^e	13.4±5.5	15.4±6.3^d
CRP (mg/L)	0.8±2.8	1.1±2.3^c	0.8±1.4	1.0±1.8	0.6±1.4	0.8±1.1^e	0.5±1.1	0.5±0.7
Birth weight (g)	3560±570	3570±550	3450±530	3410±440	3540±530	3610±660	3440±530	3350±560
Birth height (cm)	50.6±2.3	50.7±2.3	49.9±2.2	50.1±1.7	50.6±2.0	50.9±2.4	50.0±2.4	49.7±2.2
Breast-feeding (months)	4.0±3.6	3.6±3.6	3.8±4.2	3.4±3.0	3.6±4.0	4.2±3.8	3.8±3.7	4.1±3.5
Smoking (%)					17.3	20.3	15.8	8.0
Alcohol use (%) ^b					6.9	6.5	3.7	2.3
Physical activity index					9.3±2.1	9.6±2.0	8.5±1.7	8.3±1.5
Family history of CHD (%)	14.1	15.3	14.0	18.9	30.1	34.9	33.2	35.2
Family history of hypertension	40.2	55.4^c	44.9	61.5^c	51.8	68.9^d	52.0	62.1
Family history of diabetes (%)	8.7	18.9^c	13.0	20.0	20.7	30.0	18.2	33.3^d
Socio-economic status (%):
low	3.4	8.0	5.9	13.0	9.9	13.1	9.8	4.5
medium	79.4	78.0	78.8	76.1	75.8	73.7	76.2	86.6
high	17.2	14.0	15.3	10.9	14.2	13.1	14.0	9.0

Values are mean±SD for continuous variables and as percentages for categorical variables.

^aSum of biceps, triceps, and subscapular.

^bAlcohol use: subjects reporting consuming beer, wine, or liquor once a week or more often were classified as regular users of alcohol.

^cP<0.05. ^dP<0.01. ^eP<0.001.

The levels of risk variables measured in childhood and adolescence according to the number of metabolic syndrome components measured in adulthood are shown in Table II. Risk factor levels showed significant positive trends in BMI, skinfold thickness, blood pressure, triglycerides, insulin, and CRP, and a negative trend in HDL cholesterol with increasing number of adulthood metabolic syndrome components.

Table II.  Adjusted mean levels of youth risk variables by number of IDF metabolic syndrome components present in adulthood.^a

	Number of metabolic syndrome components in adulthood ^b
Childhood risk variable	0	1	2	3	≥4	P for trend
n	724	687	408	237	139
BMI (kg/m^2)	17.4	17.7	18.3	18.6	19.2	<0.0001
Skinfold thickness (mm)	23.9	25.3	27.6	29.1	31.8	<0.0001
Systolic blood pressure (mmHg)	112	112	114	114	116	<0.0001
Diastolic blood pressure (mmHg)	68	69	69	69	71	0.0018
Total cholesterol (mmol/L)	5.3	5.2	5.3	5.3	5.4	0.23
LDL cholesterol (mmol/L)	3.4	3.4	3.5	3.5	3.6	0.095
HDL cholesterol (mmol/L)	1.65	1.54	1.51	1.49	1.44	<0.0001
Triglycerides (mmol/L)	0.62	0.65	0.69	0.73	0.84	<0.0001
Insulin (mU/L)	9.0	9.5	10.3	11.1	12.5	<0.0001
CRP (mg/L)	0.6	0.6	0.7	0.8	0.7	0.0002

^aValues adjusted for sex and age.

^bNumber of any components of metabolic syndrome by IDF criteria (waist, blood pressure, HDL cholesterol, triglycerides, glucose).

In all subjects, significant 21-year tracking correlations were observed between childhood/adolescence and adulthood metabolic risk variables. The highest Spearman's correlation coefficient was seen with HDL cholesterol (r=0.45, P<0.0001), followed by BMI (r=0.41, P<0.0001), systolic blood pressure (r=0.30, P<0.0001), triglycerides (r=0.26, P<0.0001), and glucose (r=0.17, P<0.0001, 15-year correlation).

Youth determinants of the adult metabolic syndrome

Multivariable logistic regression models were constructed to examine the independent contributions of childhood and adolescents risk variables to the development of the adult metabolic syndrome. In children (aged 3–9 years at base-line), childhood obesity, high triglycerides, family history of hypertension, and age were independently associated with the adulthood metabolic syndrome (Table III, Model I). In adolescents (aged 12–18 years at base-line), the independent determinants of the adulthood metabolic syndrome included obesity, male sex, high insulin level, family history of hypertension, high triglycerides, family history of type 2 diabetes, and high CRP level (Table III, Model II). Risk variables that were not associated with the adulthood metabolic syndrome in the multivariable models included elevated blood pressure, low HDL cholesterol, smoking, alcohol use, physical activity index, family history of coronary heart disease, birth size, breast-feeding, and parental socio-economic status. Similar results were seen when the adult metabolic syndrome was defined by using the NCEP or the EGIR criteria (data not shown).

Table III.  Childhood (Model I) and adolescent (Model II) characteristics that predict adulthood International Diabetes Federation (IDF) metabolic syndrome (n = 1955).

	OR	95% CL	P-value
Model I
3–9 years at base-line (n = 958)
Obesity (BMI >80^th∼17.0 kg/m^2)	3.0	2.0–4.7	<0.0001
High TG (>80^th∼0.76 mmol/L)	2.3	1.5–3.6	0.0002
Family history of hypertension	1.8	1.1–2.7	0.009
Age	1.1	1.0–1.2	0.028
Model II
12–18 years at base-line (n = 996)
Obesity (BMI >80^th∼21.5 kg/m^2)	2.1	1.4–3.1	0.0002
Male sex	1.7	1.2–2.4	0.002
High insulin (>80^th∼16.8 mU/L)	1.7	1.2–2.5	0.006
High TG (>80^th∼0.95 mmol/L)	1.6	1.1–2.3	0.019
CRP (>80^th∼0.7 mg/L)	1.6	1.1–2.4	0.019
Family history of type 2 diabetes	1.6	1.1–2.3	0.021
Family history of hypertension	1.5	1.1–2.2	0.015

Initial stepwise logistic regression models included also childhood high systolic blood pressure, low HDL cholesterol, smoking, alcohol use, physical activity, family history of coronary heart disease, birth size, socio-economic status.

BMI = body mass index; CL = confidence limit; CRP = C-reactive protein; HDL = high-density lipoprotein; OR = odds ratio; TG = triglycerides.

Characteristics of obese children and adolescents who developed the metabolic syndrome

Because not all overweight and obese youths appear to be at equal risk of developing metabolic syndrome, we studied the development of risk factor levels in initially obese subjects with and without the metabolic syndrome in young adulthood. The longitudinal serial changes in study variables in initially obese subjects who developed and did not develop the adult metabolic syndrome are shown in Figure 1–6, along with the data in all study subjects. Among 3–9-year-olds at base-line, there were 169 obese children, and 38 (22%) of these subjects were diagnosed with the metabolic syndrome in adulthood. Among 12–18-year-olds at base-line, there were 125 obese adolescents, and 38 (30%) of these subjects were diagnosed with the metabolic syndrome in adulthood. Obese children and adolescents who had the metabolic syndrome diagnosis as adults had significantly higher BMI already during childhood and adolescence than those who did not have the diagnosis as adults. They also increased their BMI more rapidly, especially between 1986 and 2001, i.e. during the transition into adulthood (Figure 1). Obese subjects at base-line who later developed the metabolic syndrome had higher waist circumference in adulthood (mean±SD, 106±11 cm) than those who did not develop the syndrome (mean±SD, 87±12 cm), P<0.0001.

Figure 1.  Serial changes in body mass index (mean±SEM) a) from childhood (3–9 years) and b) from adolescence (12–18 years) to young adulthood in all subjects, and initially obese subjects (obesity status was defined in 1980) with respect to adult metabolic syndrome (using the International Diabetes Federation (IDF) criteria). Statistical comparisons between obese groups. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Regarding insulin, the physiological state of insulin resistance during puberty 26 is reflected as high insulin levels during adolescence in all subjects as well as in obese subjects (Figure 2). Normally insulin levels decrease after puberty 27 and this normalization is seen in obese youths who do not develop the metabolic syndrome. In contrast, those individuals who develop the metabolic syndrome continue presenting with high insulin levels in adulthood (Figure 2).

Figure 2.  Serial changes in insulin (mean±SEM) a) from childhood (3–9 years) and b) from adolescence (12–18 years) to young adulthood in all subjects, and in initially obese subjects (obesity status was defined in 1980) with respect to adult metabolic syndrome (using the International Diabetes Federation (IDF) criteria). Statistical comparisons between obese groups. * P<0.05; ** P<0.01; *** P<0.001; **** P<0.0001.

The trends in triglycerides levels resemble those seen in BMI: a rapid increase especially during the transition from adolescence/early adulthood to adulthood (Figure 3).

Figure 3.  Serial changes in triglycerides (mean±SEM) a) from childhood (3–9 years) and b) from adolescence (12–18 years) to young adulthood in all subjects, and in initially obese subjects (obesity status was defined in 1980) with respect to adult metabolic syndrome (using the International Diabetes Federation (IDF) criteria). Statistical comparisons between obese groups. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Development of systolic blood pressure levels is shown in Figure 4. Obese children and adolescents who develop the metabolic syndrome have a steady rise in systolic blood pressure by age. In comparison, obese individuals who do not develop the metabolic syndrome show no change by age in systolic blood pressure. Their blood pressure changes over time resemble the trend seen in all subjects 22.

Figure 4.  Serial changes in systolic blood pressure (mean±SEM) a) from childhood (3–9 years) and b) from adolescence (12–18 years) to young adulthood in all subjects, and in initially obese subjects (obesity status was defined in 1980) with respect to adult metabolic syndrome (using the International Diabetes Federation (IDF) criteria). Statistical comparisons between obese groups. * P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

The trends in HDL cholesterol levels are shown in Figure 5. Obese youths who develop the metabolic syndrome have consistently lower HDL cholesterols compared to those who do not develop the syndrome.

Figure 5.  Serial changes in high-density lipoprotein (HDL) cholesterol (mean±SEM) a) from childhood (3–9 years) and b) from adolescence (12–18 years) to young adulthood in all subjects, and in initially obese subjects (obesity status was defined in 1980) with respect to adult metabolic syndrome (using the International Diabetes Federation (IDF) criteria). Statistical comparisons between obese groups. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

The trends in physical activity levels are shown in Figure 6. No significant differences were seen in the levels of physical activity indices between obese youth who developed and did not develop the metabolic syndrome.

Figure 6.  Serial changes in physical activity (mean±SEM) from adolescence (12–18 years) to young adulthood in all subjects, and in initially obese subjects (obesity status was defined in 1980) with respect to adult metabolic syndrome (using the International Diabetes Federation (IDF) criteria). Statistical comparisons between obese groups. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

The smoking rates were similar between obese subjects who developed and did not develop the metabolic syndrome, and resembled the rates in all subjects (data not shown).

Discussion

In the present study, the independent youth determinants of the adult metabolic syndrome included obesity, male sex, family history of hypertension, family history of type 2 diabetes, high triglycerides, high insulin, and high CRP. Obesity was a strong predicting risk factor for the development of the adult metabolic syndrome. Furthermore, the serial data demonstrated that those obese youths who developed the metabolic syndrome had significantly higher BMI already during adolescence than those who did not develop the syndrome. They also gained weight more rapidly during transition into adulthood. Our findings thus confirm the results of previous studies which have shown that childhood obesity predicts the development of the metabolic syndrome in adulthood 28, 29.

In addition to obesity, elevated level of triglycerides was another individual component of the metabolic syndrome that independently predicted the risk. The serial trends in triglyceride levels in obese individuals who developed the adult metabolic syndrome demonstrated a constant rise during childhood and adolescence and a rapid increase during the transition into young adulthood. Elevation in triglyceride levels may thus be a specific marker of the early metabolic events related to the pathogenesis of the metabolic syndrome. In adults, the simultaneous occurrence of hypertriglyceridemia and obesity has been suggested as a high-risk phenotype with atherogenic and diabetogenic profiles 30.

High insulin level in adolescence was also an independent risk factor for the adult metabolic syndrome. We have previously shown in the Young Finns cohort that high fasting insulin measured in children predicted the clustering of high triglycerides, low HDL cholesterol, and elevated systolic blood pressure in a 6-year follow-up 31. The current observation with considerably longer follow-up provides more evidence for the idea that hyperinsulinemia may be causally related with the deterioration of lipid and blood pressure profiles. In the serial data, the deviation in insulin metabolism was clearly seen in obese individuals who developed the adult metabolic syndrome. Normally insulin levels decrease after puberty, but this normalization was not seen in these individuals who continued to have elevated insulin levels through out the follow-up. In line with these observations, Bao et al. 32 have previously shown in the longitudinal Bogalusa Heart Study that when insulin concentrations are increased in childhood, they tend to remain elevated in adulthood, and that those adults with consistently elevated insulin levels also tend to have a combination of increased obesity, hypertension, and dyslipidemia. There are several mechanisms that could explain the cause and effect relation between insulin and clustering of metabolic risk factors. Free fatty acid concentrations are suggested to be the link between insulin resistance and dyslipidemia 33. In patients with insulin resistance, the suppression of free fatty acid release from adipose tissue is impaired, providing increased production of very-low-density lipoprotein (VLDL) and triglycerides by the liver 33. The defect in lipoprotein lipase activity in insulin resistance also contributes to the decrease in plasma HDL cholesterol, and the LDL cholesterol pool may become enriched with small, dense, highly atherogenic LDL particles 34. Hyperinsulinemia or insulin resistance may also raise blood pressure by increasing sympathetic nervous system activation 35, stimulation of vascular smooth muscle cell growth, promoting the renal reabsorption of sodium, and altering cell electrolyte composition 36.

Recently, in the Fels Longitudinal Study of 493 study participants, Sun et al. 37 showed a direct relation between childhood blood pressure and the metabolic syndrome in adulthood. The investigators concluded that the effects of childhood BMI on the adulthood metabolic syndrome were mediated by childhood blood pressure. We could not confirm these observations in the present study. Children and adolescents who eventually developed the adult metabolic syndrome had higher blood pressure levels at base-line compared to others. In addition, obese youths who developed the metabolic syndrome had a steady rise in systolic blood pressure by age. However, elevated blood pressure in youth was not an independent predictor of the adult metabolic syndrome in models adjusted for several other metabolic risk factors.

We have previously shown in this cohort that maintaining a high level of physical activity from youth to adulthood is associated with lower risk of abdominal obesity in women 38. In the present analysis, however, the level of physical activity was not an independent determinant of the adult metabolic syndrome. We were also unable to demonstrate significant differences in the level of physical activity in serial measurements between obese youths who differed regarding the metabolic syndrome status in adulthood. These findings suggest that the assessment of habitual physical activity level in children and adolescents may not provide useful information about the future risk of the metabolic syndrome.

Increased CRP associates with obesity 39 and is an independent predictor of cardiovascular events 40. In the present study, an elevated CRP level in adolescence was an independent predictor of the adult metabolic syndrome. Previous studies have suggested that children and adolescents with the metabolic syndrome show evidence of low-grade inflammation. Ford et al. 41 analyzed data from the National Health and Nutrition Examination Survey and showed that the concentrations of CRP were higher among 12–17-year-old subjects with the metabolic syndrome than in subjects without the syndrome. In another US study, De Ferranti et al. 42 reported in 12–19-year-old adolescents that CRP concentrations increased with the increasing number of metabolic abnormalities and were higher in adolescents with the metabolic syndrome than in those without the syndrome. In the present study, the relation between elevated CRP and adult metabolic syndrome remained independent in a multivariable model after adjustments for other confounding factors such as obesity. Together, these observations point to the possibility that inflammation may have role in the pathogenesis of the metabolic syndrome.

We found that family history for hypertension and type 2 diabetes were independent predictors of the adult metabolic syndrome. A recent study by Ventura et al. 43 demonstrated that girls at the age of 13 classified as having higher risk for the metabolic syndrome and its components were more likely to have a family history of obesity, type 2 diabetes and gestational diabetes. There are various potential mechanisms explaining the relations between the metabolic syndrome and positive family history. There may be some underlying genetic variants, which predispose some vulnerable individuals to the development of the metabolic syndrome 44, 45. Genes may also modify the impact of obesity and other conventional risk factors. Further genetic studies are needed to clarify the involvement of genetic variants in the metabolic syndrome. The practical implication is that the assessment of family risk of hypertension and type 2 diabetes may be helpful in the pediatric risk assessment for the propensity to develop metabolic syndrome.

A potential limitation of the study is the non-participation in the follow-up studies. The participation rate of the latest follow-up study in 2001 was approximately 65%. We have previously reported that the base-line characteristics are similar between study subjects and those lost to follow-up 22. Therefore, the present study cohort seems to be representative of the original study population. We measured only fasting insulin level that is a crude measure of insulin resistance. We did not measure waist circumference in 1980 and 1986 and blood glucose in 1980. Information on family history of coronary heart disease, type 2 diabetes, and hypertension were collected in 2001, when participants were 24–39 years of age. Obviously, the rates for positive family histories would have been lower if collected in 1980.

In adults, the metabolic syndrome predicts the development of type 2 diabetes and cardiovascular diseases 46, 47. Among young adults, individuals with the metabolic syndrome have evidence of subclinical atherosclerosis indicated by increased carotid intima-media thickness 48. We have previously shown that exposure to metabolic risk factors in adolescence may contribute to the development of structural and functional vascular markers of subclinical atherosclerosis 25, 49, 50. Data are also emerging linking childhood exposure to metabolic risk factors to cardiovascular end-points decades later 51; Morrison et al. 51 recently reported from the longitudinal Princeton LRC School Study that those children with the cluster of factors defined as pediatric metabolic syndrome were significantly more likely to have cardiovascular disease decades later as adults. These observations thus emphasize the importance of early identification of children and adolescents at increased risk of the metabolic syndrome. In the present study, youth determinants of adult metabolic syndrome included obesity, family history of hypertension, family history of type 2 diabetes, high triglycerides, high insulin, and high CRP. Identifying these risk factors in children and adolescents could be helpful in pediatric metabolic risk assessment.

Acknowledgements

This study was financially supported by the Academy of Finland (grant no. 77841 and 210283), the Juho Vainio Foundation, the Finnish Foundation of Cardiovascular Research, the Finnish Cultural Foundation, Special Federal Grants for the Turku University Hospital, the Turku University Foundation, the Research Foundation of Orion Corporation, the Research Foundation of AstraZeneca, the Yrjö Jahnsson Foundation, the Lydia Maria Julin foundation, and the Maud Kuistila Foundation.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.