Cohort Profile: The International Childhood Cardiovascular Cohort (i3C) Consortium

Abstract

This is a consortium of large children's cohorts that contain measurements of major cardiovascular disease (CVD) risk factors in childhood and had the ability to follow those cohorts into adulthood. The purpose of this consortium is to enable the pooling of data to increase power, most importantly for the follow-up of CVD events in adulthood. Within the consortium, we hope to be able to obtain data on the independent effects of childhood and early adult levels of CVD risk factors on subsequent CVD occurrence.

How did the study come about?

During the 1950s and 1960s, it became evident that atherosclerosis, a slowly progressive systemic disease that causes narrowing and hardening of large- and medium-sized arteries, begins in childhood. Holman and colleagues demonstrated not only the presence of fatty streaks in the aortas of children as young as 3 years, but also that these fatty streaks progress to clinically significant lesions (fibrous plaques) by young adulthood.^1–3 At that time, it was anticipated that this process was likely to be influenced by the same risk factors that predicted adult coronary heart disease. This was later shown to be correct.^4–6

Over the subsequent two decades, investigators in a number of different countries initiated cohort studies to gain insights into the tracking of lifestyle and biological factors through childhood and the relationship of these factors with cardiovascular disease (CVD). These studies included the Muscatine Study,⁷ the Bogalusa Heart Study (BHS),^8–10 the Princeton Lipid Research Clinics (LRC) Study,^11,12 the Minneapolis Children's Blood Pressure Study,^13,14 and the National Heart, Lung, and Blood Institute Growth and Health Study (NGHS)¹⁵ from the USA and the Cardiovascular Risk in Young Finns Study from Finland.¹⁶ Each of these studies continued to follow participants into early and middle adulthood. In 1985, the Australian Schools Health and Fitness Survey collected extensive information on CVD risk factors.^17,18 Repeat measures were obtained on this cohort in young adulthood. The combined study became known as the Childhood Determinants of Adult Health (CDAH) Study.^19–21

By 2002, the possibility of linking data collected in childhood through adulthood to future CVD events, in particular myocardial infarction and stroke, was becoming a realistic focus in several of these large cohorts—the intention being to look at the separate effects of childhood and adult risk factors on the occurrence of CVD. However, because the cohorts were still relatively young, an extended follow-up period would be necessary to accumulate sufficient CVD events to address the key hypotheses in any individual cohort.¹⁶

The idea of pooling data from these large cohorts arose. Initially the Young Finns, the CDAH, the Bogalusa and the Muscatine studies were approached to participate. Through the increased power that this would provide, the follow-up period to ascertain sufficient early CVD events could be reduced. There was considerable enthusiasm across the studies and collaboration steadily developed. Detailed examination revealed that all four studies had included similar lifestyle and biological risk factors, among them the major CVD risk factors—blood pressure, serum lipids and adiposity measures, permitting pooling or meta-analysis involving these factors across studies.^22–24

Researchers associated with each cohort agreed to proceed with establishing a formal consortium ultimately named: ‘The International Childhood Cardiovascular Cohort (i3C) Consortium’. A Steering Committee was established that was composed of primary investigators from the four cohorts, and other investigators to be included as necessary, to assist with the consortium activities. This consortium brought together international researchers involved in the four long-standing cohort studies, as well as experts in epidemiology, paediatric cardiology, genetics and other disciplines.

Who is in the sample?

Initially, the consortium consisted only of the largest four cohorts that had undertaken measurement of major CVD risk factors in childhood and had followed them through to adulthood. As the capacity of the consortium and the extent of its network developed, other similar or moderately smaller studies were identified. These were the Princeton LRC school cohort^11,12 and its follow-up, the Princeton Follow-up Study (PFS),^25,26 two Minneapolis Children's Studies^13,14,27,28 and the NGHS¹⁵ (Table 1, Figure 1). The proportion of the sample on which all major CVD risk factors were measured varied between cohorts. The total number of participants with major CVD risk factors measured at least once in childhood and adulthood is approximately 12 000.

Table 1.

Summary of the i3C Consortium Studies

			Baseline in childhood and adolescence				Follow-up in adulthood
Study	Country	Sampling frame	Sample with one major CVD risk factor (N)a	Sample with three major CVD risk factors (N)b	Study years	Age (years)	Sample with three major CVD risk factors (N)b	Study years	Age (years)
Muscatine Studyc	USA	School based	11 377	11 377	1970–81	5–18	2547	1982–91	20–39
							865	1992–2008	29–55
BHSc,d	USA	School based	12 164	12 164	1973–94	4–17	1203	2001–02	23–43
							1052	2003–05	26–47
							914	2007–10	29–51
YFS	Finland	Random sample from five centres	3596	3596	1980	3–18	2283	2001	24–39
							2204	2007	30–45
							Ongoing	2010	33–48
CDAH Study	Australia	School based	8498	1714	1985	7–15	2410	2004–06	26–36
Minneapolis Children's Studiesc,d	USA	School based	1207 (10 423 screened)	0	1978–89	6–9 to 17–20	679	1993–96	21–24
							359	2007–11	35–38
		School based	357 (12 043 screened)	357	1996	11–14	230	2004	19–24
Princeton LRC/PFSc,d	USA	School based	6775 (Visit 1)	1729 (Visit 2)e	1973–76	5–19	623e	1999–2004	30–48
							600 (ongoing)	2011–12	43–57
NGHSd	USA	Cincinnati, OH: School based	871	705–871	1987–96	9–19	653	1997–2002	20–24
							∼650	2002–07	24–28
		Richmond, CA: School based	879	823–871			–	–	–
		Washington, DC: Sample from HMO	629	550–629			–	–	–

BHS: Bogalusa Heart Study; CDAH Study: Childhood Determinants of Adult Health Study; NGHS: National Heart, Lung, and Blood Institute Growth and Health Study; YFS: Young Finns Study; Princeton LRC/PFS: National Heart, Lung, and Blood Institute Princeton Lipid Research Clinics Study/Princeton Follow-up Study; HMO: health maintenance organization.

^aSample with measurement of at least one major CVD risk factor (blood pressure, lipids or adiposity measures including skinfold thickness or BMI).

^bSample with measurement of all three major CVD risk factors.

^cOther generations and family data available.

^dBiracial cohort: black and white.

^eAn additional 557 participants aged 5–19 years were measured in the LRC family study (Visit 3) along with their siblings already counted in the 1729 above. Blood pressure was not measured at Visit 3. Of these 557 siblings, 221 were recontacted in the 1999–2004 follow-up.

Figure 1.

Geographical locations of the participating longitudinal cohorts in the i3C Consortium (1–7). 1: Muscatine Study, Muscatine, IA, USA; 2: BHS, Bogalusa, LA, USA; 3: five dots in Finland represent study sites in the Cardiovascular Risk in Young Finns Study, Oulu, Kuopio, Tampere, Turku and Helsinki, Finland; 4: all dots in Australia represent study sites in the CDAH Study, Hobart, Melbourne, Sydney, Canberra, Adelaide, Brisbane, Darwin, Perth, small cities and towns, Australia; 5: Minneapolis Children's Studies, Minneapolis, MN, USA; 6: Princeton LRC Study, PFS, and NGHS, Cincinnati, OH, USA; 7: NGHS, Richmond, CA, and Washington, DC, USA

What does it cover?

The i3C Consortium offers the potential to extend our knowledge about the childhood origin of adult cardio-metabolic diseases and will focus on, but not be limited to, the following three key research areas.

Follow-up of CVD morbidity and mortality

The central question on which the consortium will focus is: is there an effect of childhood risk factors which is independent of the already known effect of adult risk factors on CVD? The possibility that there might be a residual effect of childhood exposures on CVD risk has enormous significance for public health programmes aimed at preventing this major cause of death and disability, and such evidence is currently unavailable. Although the oldest participants in the consortium samples are only in their sixth decade of life, a time when incident CVD events are just beginning to occur, by combining events data, analyses linking childhood risk factors to adult CVD can be conducted sooner than could otherwise be accomplished by any individual cohort. A key future research priority is therefore to initiate and maintain records of morbidity and mortality from each cohort.

Genetic data collection and analyses

In the past 5 years, a number of large-scale genome-wide association (GWA) studies have led to the identification of many inherited variants that predispose individuals to CVD.^29–32 The combined effect of the variants identified to date on susceptibility to CVD, however, is small, suggesting that other yet-to-be-identified variants may contribute to the heritable component of CVD.³³ Gene–environment interaction effects, where relevant, may also have obscured genetic associations in a way that can only be revealed with appropriate data and longitudinal measures.³⁴ The greater power provided by access to repeated measurements of risk factors afforded by longitudinal studies can lead not only to the identification of new genes that predispose to CVD, but also better characterization of the effects of these genes, in a causal manner, throughout life.

Non-invasive vascular measures

Another key priority is to update data collection in each cohort with harmonized non-invasive subclinical measures of vascular structure and function. The collection of data on these intermediate vascular end points will enable us to better understand the evolution of CVD across the lifespan and to more confidently identify causal pathways to CVD events, as these accumulate over the coming decades. Such measures include the ultrasonic assessment of carotid intima-media thickness (IMT), arterial endothelial function using flow-mediated dilatation, arterial elastic properties, ventricular geometry and function, coronary artery calcification measured through computed tomography and retinal microvasculature assessed using digital retinal photographs. Most studies in the consortium have measured adult carotid IMT and presence of plaques, providing data directly linking childhood risk factors to subclinical markers of atherosclerosis in adulthood.^23,35–37 Recently, the Young Finns and the Muscatine studies began collection of digital retinal photographs.

How often have they been followed up and what is the attrition like?

In Muscatine, IA, between 1970 and 1981, baseline measures were obtained among 11 377 schoolchildren (5–18 years) in biennial surveys.⁷ Approximately 70% of eligible students participated in each of the six surveys.³⁸ Beginning in 1981, previous survey participants were recruited at ages ranging from 20 to 30 years for young adult follow-up examinations; 2547 (67% of those eligible) participated in at least one examination between 1981 and 1991.³⁹ Beginning in 1992, a representative subset (865 individuals, 29–43 years) of participants in early adulthood had repeated follow-up examinations every 2–4 years. At baseline, the 865 participants were similar to the entire cohort for height, weight, blood pressure, triceps skinfold thickness, total cholesterol and triglycerides (adjusted for age, sex and calendar year).^35,40,41

In Bogalusa, LA, between 1973 and 1994, baseline measures were obtained from seven cross-sectional surveys of 12 164 children (4–17 years) in this biracial community. Young adults (18–37 years) who participated earlier as children and remained accessible were re-examined over six surveys between 1977 and 1996. Other follow-ups were conducted in 2001–02, 2003–05 and 2007–10. The panel design, based on repeated examinations conducted every 3–4 years, resulted in serial observations from childhood to adulthood.⁴² Follow-up in early adulthood (1420 individuals, 19–37 years) commenced in the 1995–96 survey and ultrasound examination of the carotid artery was introduced (635 individuals). At baseline, participants who had carotid IMT measures were similar to the rest of the study cohort for race, sex, body mass index (BMI), systolic blood pressure, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol and triglycerides.³⁶

In 1980, the Young Finns Study commenced its baseline study with 3596 participants (3–18 years) who have then been followed up approximately every 3–6 years.¹⁶ Participation has been dynamic and many participants lost to early follow-up have returned to the study later on. Blood samples and physical measurements were obtained for the entire cohort in 1983 (2991 participants, 83% of the baseline), 1986 (2799, 78%), 2001 (2283, 64%) and 2007 (2204, 61%). Participants in the 2001 follow-up study were more often girls and older at baseline than non-participants. Otherwise, baseline total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, blood pressure, BMI, parental education and physical activity were similar in both sexes between those who did and did not participate as adults (adjusted for age).¹⁶

After the baseline study of 8498 schoolchildren (7–15 years) in 1985, the CDAH Study followed up 61% (5170 individuals) of its baseline sample in 2001–04 to obtain basic demographic and lifestyle characteristics from questionnaire; in addition, physical measurements and blood samples were collected in 2004–06 (2410 participants, 28% of the baseline). Compared with non-participants, at baseline participants in the CDAH Study were slightly older, had lower BMI and were more often girls, but were similar for blood pressure, total cholesterol, HDL cholesterol and triglycerides (adjusted for age and sex).

In Cincinnati, OH, the Princeton LRC Study recruited 6775 schoolchildren (6–17 years) in 1973–76 (baseline Visit 1).¹¹ In all, 1729 of these children (91% of eligible participants) attended baseline Visit 2 examination 6 weeks later where major CVD risk factors were obtained;¹² an additional 557 children (siblings of LRC participants) attended a family-based study 18 months later (Visit 3), where lipids and anthropometry, but not blood pressure, were obtained. Between 1999 and 2004, the PFS followed up 623 Visit 2 and 221 Visit 3 participants (68% of those eligible) in adulthood.^25,26 Previous students who attended the PFS were more often girls and non-Hispanic white, and were older at baseline than non-participants. Boys in the PFS were slightly older and had higher baseline total cholesterol than non-participants who were eligible. Girls in the PFS had slightly higher baseline BMI than girls who did not participate. Other baseline anthropometric and lipid values were similar between those who did and did not participate as adults.

In Minneapolis, MN, between 1978 and 1989, baseline blood pressure and anthropometric measurements were obtained among 1207 children from public schools over 18 visits in the Minneapolis Children's Blood Pressure Study. This cohort has been followed up in 1990–93 (817, 68% of the original), 1993–96 (679, 56%) and 2007–11 (359, ongoing).^13,14 In addition to the anthropometric and blood pressure data, fasting blood samples were obtained in the 1993–96 follow-up. Another Minneapolis study recruited a sample of 357 schoolchildren in 1996.²⁷ Follow-up studies were conducted in 1998 (330, 92%), 2000 (257, 72%) and 2004 (230, 64%). Compared with non-participants, participants who attended each follow-up were similar for baseline anthropometric and blood pressure in both studies.

The multi-centre NGHS was a 10-year study beginning at ages 9 and 10 years in 2379 girls (1987–96).¹⁵ It continued clinical evaluation only in the Cincinnati cohort with annual visits during 1997–2002 and two clinic visits between 2002 and 2007 on 650 of 871 enrolees. The NGHS Cincinnati cohort has experienced 75% retention after 20 years of follow-up, and attrition rates did not differ between black and white girls. White girls had slightly lower baseline BMI, and black girls had slightly higher BMI, compared with the white and black girls, respectively, who did not participate in the study as adults.

These attrition analyses indicate that those who participated as adults within the i3C Consortium studies were very similar to non-participants for baseline major CVD risk factors.

What has been measured?

CVD-related biological and lifestyle measures available for each cohort in childhood (baseline) and adulthood (follow-up) include: anthropometry, blood pressure, blood biochemistry, lifestyle factors and subclinical vascular measures. Table 2 presents a list of available measures and the extent to which these measures have been collected across the cohorts. All consortium studies have also collected biospecimens (serum/plasma samples) for all or representative portions of the cohorts and some have DNA and genotype data (Table 3).¹⁶

Table 2.

Consistent key measures across the i3C Consortium Cohorts in childhood, adolescence and adulthood

Childhood and adulthood columns side by side for each study with those for adulthood on a light grey background. Tick mark represents variable available. Blank area represents variable unavailable or not applicable.

BHS: Bogalusa Heart Study; CDAH Study: Childhood Determinants of Adult Health Study; NGHS: National Heart, Lung, and Blood Institute Growth and Health Study; YFS: Young Finns Study; Princeton LRC/PFS: National Heart, Lung, and Blood Institute Princeton Lipid Research Clinics Study/Princeton Follow-up Study; WC: waist circumference; HC: hip circumference; SFT: skinfold thickness; Apo A-I: apolipoprotein A-I; Apo B: apolipoprotein B; CRP: C-reactive protein; SES: socio-economic status; IMT: intima-media thickness; PWV: pulse wave velocity; FMD: flow-mediated dilatation; LVG: left ventricular geometry; ECG: electrocardiogram; OC: oral contraceptive; SHBG: sex hormone-binding globulin; SGA: small for gestational age.

^aIn childhood and adolescence: the participants' own feetal growth information and pregnancy-related variables of their mothers; in adulthood: the participants’ own reproductive characteristics.

^bRetrospective collected in adulthood.

^cStarted or funded for the current follow-up.

^dApo A-I and Apo B levels were not measured directly in these cohorts but can be estimated using Friedewald inputs (total cholesterol, triglycerides and HDL cholesterol).

^eYFS: for males only; CDAH: for female non-OC users only.

Table 3.

Biospecimens collected for genetic and biomarker studies in the i3C Consortium

	Existing biospecimens
Cohort	Samples	Participants	Type	Collected in both childhood and adulthood
Muscatine Study	∼865a	∼865	Serum/DNA	Yes
BHS	∼2500b	∼2500	Blood/DNA	Yes
YFS	∼50 000c	3596–2404	Serum/DNA	Yes
CDAH Study	2538d	2485	Blood/DNA	No
Minneapolis Children's Studies	1241	1241	Serum/DNA	Yes
Princeton LRC/PFS	2850e	950	Blood/DNA	No
NGHS	∼20 000	2300	Serum/DNA	Yes

BHS: Bogalusa Heart Study; CDAH Study: Childhood Determinants of Adult Health Study; NGHS: National Heart, Lung, and Blood Institute Growth and Health Study; YFS: Young Finns Study; Princeton LRC: National Heart, Lung, and Blood Institute Princeton Lipid Research Clinics Study.

^aA total of 780 Muscatine participants have DNA and candidate genotypes data available and multiple stored serum samples for many participants.

^bA total of 1216 BHS participants have genome-wide SNPs data available (IIIumina 650 K), and additional 1300 DNA samples stored.

^cSpare serum samples collected in study years 1980, 1983 and 1986 (stored at −20°C) and study years 2001 and 2007 (stored at −80°C). A total of 2442 participants have genome-wide SNPs data available (IIIumina 660 K).

^d2538 CDAH participants have blood samples, 2485 have DNA samples collected as blood spots stored on FTA cards.

^eCollected at the 1999–2004 follow-up.

What has it found? Key findings and publications

Beyond the extensive findings from each individual cohort during the past 40 years, specific analyses of the combined consortium data to this point have focused on the following.

A resurgent interest in screening children and adolescents for dyslipidaemia to identify those at high risk of CVD led us to examine the two proposed paediatric cut points to predict those individuals who will develop abnormal levels in adulthood. The consortium data showed that the separate use of National Health and Nutrition Examination Survey (NHANES) cut points for HDL cholesterol and National Cholesterol Education Program (NCEP) cut points for total cholesterol, LDL cholesterol and triglycerides provide the most accurate classification for adolescents who developed dyslipidaemia in adulthood. Those employing either classification need to consider that a substantial number of adolescents identified as being at high risk will not have high-risk factor levels in early adulthood. These findings emphasize that the benefit of screening needs to be carefully considered with possible detriments such as falsely labelling some youth as at risk.²²

Using consortium data, we also demonstrated that NHANES and NCEP paediatric dyslipidaemia classifications perform equally in the prediction of adolescents who are at increased risk of high IMT in early adulthood; and that adolescent lipid levels are more strongly associated with subclinical atherosclerosis as measured by carotid IMT in adulthood than change in lipid levels from childhood to adulthood; and that dyslipidaemia in the presence of overweight or obesity places affected adolescents at substantially higher risk of increased subclinical atherosclerosis as adults compared with those who do not have both risk factors. These findings emphasize the importance of both population-wide and individualized prevention programmes to improve paediatric dyslipidaemia-related causes of early atherosclerosis.²³

The pooled data also showed that paediatric metabolic syndrome predicts adult high carotid IMT and type 2 diabetes, and does so at a level that is either equivalent or inferior to classification of paediatric BMI. These findings suggest that in clinical settings, efforts to identify youth with heightened future risk of cardio-metabolic outcomes might be made most simply by using BMI alone.⁴³

Using the pooled data, we established the optimal age for CVD risk factor screening when childhood risk levels begin to associate with adult subclinical atherosclerosis. We examined the influence of age on associations between childhood risk factors and adult carotid IMT. Risk factors measured before the age of 9 years appear to have only weak or non-significant associations with adult carotid IMT, whereas measurements in participants aged 9–18 years showed significant associations with increased adult carotid IMT. These findings suggest that risk factor screening allows youth who are at increased risk of subclinical atherosclerosis in adulthood to be identified.²⁴

Recent GWA studies have pinpointed many loci associated with CVD risk factors in adults. It is unclear, however, if these loci predict trait levels at all ages, if they are associated with how a trait develops over time, or if they could be used to screen individuals who are pre-symptomatic to provide the opportunity for preventive measures before disease onset. We completed a GWA study on participants in the BHS that is the first GWA study to evaluate the role of common genetic variants in the development of CVD risk factors in children as they advance through adulthood. We reported seven genome-wide significant associations involving CVD risk factors, two of which had been previously reported. Top regions were tested for replication in the Young Finns Study and two of the associations strongly replicated: rs247616 in cholesteryl ester transfer protein gene (CETP) with HDL cholesterol levels (combined P = 9.7 × 10⁻²⁴), and rs445925 at apolipoprotein E gene (APOE) with LDL cholesterol levels (combined P = 8.7 × 10⁻¹⁹). We showed that single-nucleotide polymorphisms (SNPs) previously identified in adult case–control studies tend to show age-independent effects in the BHS with effect sizes consistent with the previous reports. This study highlights the utility of using longitudinal studies to identify genetic predictors of adult traits in children.⁴²

Our analysis of consortium data showed that childhood overweight or obesity is predictive of type 2 diabetes, hypertension, dyslipidaemia and high-risk carotid IMT in adulthood. The data also showed that persons who had normal BMI in childhood but who become obese as adults have adverse risk factor profiles, whereas those who were overweight or obese as children but who become non-obese as adults have a cardiovascular risk profile that is similar to that of persons who are never obese.⁴⁴

What are the main strengths and weaknesses?

Using pooling or meta-analysis of consortium data, we are able to provide insights into the evolution of CVD risk factors and subclinical atherosclerosis in early life with greater power than has been available previously. The consortium data also allow for comparisons between study populations in the USA, northern Europe and Australia.

A weakness anticipated when the consortium was initiated is that even with the increased power afforded by pooling of the cohorts, there would be insufficient events to estimate CVD risk for another decade. However, it was apparent that the data available could contribute substantially to understanding the tracking of risk factors from childhood to adulthood. It could also provide a greater understanding of the role of lifestyle in childhood in determining the evolution of CVD risk factors from childhood to adulthood. An apparent weakness of this consortium, like all long-term cohort studies, is loss to follow-up. Nevertheless, these data represent what is available internationally to study the effect of exposures over the child to adult life course, a time dimension that is not amenable to randomized controlled trials. With its rare longitudinal data collection spanning several decades, this consortium also offers exceptional advantages to address key hypotheses in genetic analyses.

The consortium also faces a number of challenges. These include apparent heterogeneity of measures across cohorts, including variation in methodology and technology, questionnaire data, available capacity, diagnosis and ethical requirements. Nonetheless, these studies provide data on similar key exposures and outcome measures and feasibility of data pooling or meta-analysis is evidenced by the first six joint publications.^{22,23,42–44} In addition to the initial focus on CVD, the consortium holds promise for examining other disease outcomes (e.g. cognitive function, cancer).

Can I get hold of the data? Where can I find out more?

Mechanisms for data sharing to maximize the potential for pooling or meta-analysis across different cohorts are currently being developed. All enquiries should be addressed to the Steering Committee while mechanisms for data sharing and website portal are being developed (contact Chairman, Terry Dwyer by e-mail: terry.dwyer@mcri.edu.au, or Director of i3C Consortium Data Coordinating Centre, Olli Raitakari by e-mail: olli.raitakari@utu.fi). Only analyses agreed to by the Steering Committee can be undertaken using the consortium data. The process for gaining approval for data analysis to test specific hypotheses is also under development. Institutional review board clearance for accessing data will be necessary if potentially identifying information is to be provided to a researcher.

A website portal of the i3C Consortium is under construction (http://www.i3cconsortium.org). This will provide centralized individual cohort information and facilitate communication and document management.

Summary

This is a consortium of large children's cohorts that contain measurements of major cardiovascular (CVD) risk factors in childhood and have followed these cohorts into adulthood. The oldest participants in the consortium samples are now in their fifth decade of life. The purpose of this consortium is to enable the pooling of data to increase power, most importantly for the follow-up of CVD events in adulthood. The consortium includes five cohort studies from the USA, one study from Finland and one study from Australia. CVD-related biological and lifestyle measures available for each cohort in childhood (baseline) and adulthood (follow-up) include: anthropometry, blood pressure, blood biochemistry, lifestyle factors and subclinical vascular measures. The total number of participants with major CVD risk factors (adiposity, blood pressure and lipids) measured at least once in childhood and adulthood is approximately 12 000. Currently, all data enquiries should be addressed to the Steering Committee while mechanisms for data sharing and the website portal are being developed.

Funding

The Muscatine Study was financially supported by National Institutes of Health (NIH) Grants (HL-14230) Specialized Center for Research in Atherosclerosis, (HL48050, HL-54730 and HL61857) from the National Heart, Lung, and Blood Institute, and (RR-00059) from the General Clinical Research Centers Program. The BHS was financially supported by NIH Grants (AG-16592) from the National Institute of Aging, (HL-38844) from the National Heart, Lung, and Blood Institute, and from the American Heart Association. The Cardiovascular Risk in Young Finns Study was financially supported by the Academy of Finland, the Social Insurance Institution of Finland, the Turku University Foundation, Special Federal Grants for the Turku, Tampere and Kuopio University Central Hospital, the Juho Vainio Foundation, Paavo Nurmi Foundation, the Finnish Foundation of Cardiovascular Research, Orion-Farmos Research Foundation and the Finnish Cultural Foundation. The CDAH Study was supported by the Commonwealth Departments of Sport, Recreation and Tourism, and Health; The National Heart Foundation; and the Commonwealth Schools Commission at baseline and the National Health and Medical Research Council, the Heart Foundation, the Tasmanian Community Fund and Veolia Environmental Services at follow-up. The National Health, Lung, and Blood Institute LRC Princeton Study and PFS were supported by NIH/NHLBI (N01-HV-2-2914-L and HL62394), the American Heart Association (National-97 50129N) and NIH/NHLBI (R21DK085363). The Minneapolis Children's studies were supported by NIH Grants (HL 52851, M01-RR-00400 and DK72124). The National Health, Lung, and Blood Institute Growth and Health Study was funded by NIH grants (HL52911, HC55025, HL66430 and HL48941). C.S. and C.G.M. are supported by the Australian National Health and Medical Research Council Early Career Public Health Fellowship (1013538 and 1037559, respectively). T.D. and C.S. are supported by the Victorian Government's Operational Infrastructure Support Program.