Body fat distribution and early arterial changes in healthy 5-year-old children

Abstract

Background. It is unknown at what age overweight starts to takes its toll on the vasculature. We studied the relation between body size measures and vascular characteristics in healthy 5-year-old children. Methods. In 306 5-year-old children from an on-going birth cohort, body size characteristics were measured, including sonographic measurement of abdominal fat. Ultrasonographic measurements of the carotid artery were performed to obtain intima-media thickness (CIMT), arterial wall distensibility, and elastic modulus (EM). Results. Increased body-weight was related to thicker CIMT (linear regression coefficient 2.25 μm/kg; P = 0.003), increased EM (2.73 kPa/kg; P = 0.01), and lower distensibility (−1.23 MPa⁻¹/kg; P = 0.03). Similar relations were found for increased BMI with CIMT and EM. Increased intra-abdominal fat was related to thicker CIMT (9.19 μm/cm; P = 0.02), and increased waist circumference with thicker CIMT (2.17 μm/cm; P = 0.02), lower distensibility (−1.70 MPa⁻¹/cm; P = 0.01), and higher EM (2.77 kPa/cm; P = 0.02), independent of BMI. Conclusion. For the first time it is demonstrated that increased general body mass and particularly waist circumference and intra-abdominal fat are related to thicker and stiffer arteries already early in life.

Key words::

• Arterial stiffness
• body fat distribution
• body mass index
• carotid intima-media thickness
• early life

Key messages

• Overweight is related to vascular changes in young children.

• Intra-abdominal fat exerts its effect on the vasculature, already in early life.

• Preventive measures against overweight should be taken from an early age.

Introduction

Overweight in childhood is a major and increasing public health problem (1). This trend is worrisome, particularly with its implication for the development of atherosclerosis and subsequent cardiovascular disease (CVD) risk. It has been shown that childhood overweight is associated with coronary heart disease and premature mortality (2–4). Post-mortem studies show that atherosclerosis starts early in life with fatty streaks and fibrous plaques in the aorta and coronaries, a process in which childhood overweight plays a role (5). The mechanism of how overweight affects the vasculature is ambiguous. It may be that overweight has its adverse effects through its relation with other cardiovascular risk factors (6). Also, adipose tissue is metabolically active and known to release adipokines and inflammatory cytokines that could directly affect vascular function or structure (7).

It is unknown how early in childhood overweight affects the vascular system. Overweight adolescents have a thicker intima-media of the carotid artery, a marker for vascular damage and CVD risk in adulthood (8). Recently, similar associations were found in healthy children and adolescents (9). These observations were further extended to obese children with a mean age of 8 years (10). Intima-media thickening of the carotid artery was previously observed in high-risk children (11,12).

There is an increasing sense of urgency for early life prevention of overweight (13). Once overweight has become manifest, it is difficult to counteract even in children (14). A substantial proportion of overweight children, particularly those with overweight parents, will still be overweight as adults (15). Obviously, overweight in childhood has more sequelae than the development of cardiovascular disease risk only. A challenge will be to convince the public that early-life action may help prevent cardiovascular diseases occurring only many decades later in life. In the present study, the main objective was to assess if general body mass, and body fat distribution in particular, affects the cardiovascular system already in early childhood, as that may underscore interventions early in life to be a window of opportunity.

Within an on-going population-based birth cohort study, a cross-sectional study was conducted. At 5 years of age, vascular characteristics were measured using non-invasive ultrasound scanning, along with extensive cardiovascular risk factor profiling. Anthropometric measurements were performed, including measurement of abdominal fat using ultrasound.

Our research question was whether general body mass and body fat distribution in particular are associated with arterial wall structure and function already in early life.

Methods

Study design and study population

The present study is part of the WHeezing Illnesses STudy LEidsche Rijn (WHISTLER) study, a large prospective population-based birth cohort study on determinants of wheezing illnesses that was initiated in December 2001 and is still on-going. Study design and rationale were described in detail elsewhere (16). Briefly, healthy infants born in Leidsche Rijn, a new residential area near the city of Utrecht in the Netherlands, were enrolled at the age of 2 or 3 weeks. Exclusion criteria were gestational age < 36 weeks, major congenital abnormalities, and neonatal respiratory disease. Currently, over 2,000 infants have been included. In November 2007 the study was extended to address cardiovascular research questions (WHISTLER-Cardio). All children who had reached the age of 5 years (until February 2010: n = 603) were invited according to the last-known telephone number and address for a second follow-up visit. Of these, 82/603 (14%) subjects were lost to follow-up (due to incorrect telephone numbers and non-response despite mailing or incorrect addresses). Of the remaining 521 subjects, 145/521 (28%) declined to take part, and 374/521 (72%) were willing to participate. Vascular measurements were performed in 306/374 (82%) subjects. In the 68 subjects without vascular measurements, the focus was solely on respiratory measurements, or they had their visit planned in the near future (n = 19). The present results pertain to these subjects. The WHISTLER-Cardio study was approved by the pediatric Medical Ethical Committee of the University Medical Center Utrecht. Written informed parental consent was obtained.

Neonatal visit

Parents of eligible children were invited to visit our out-patient clinic when their offspring were approximately 4 weeks of age. Information on the child was obtained by a questionnaire completed by the parents with regard to pre-, peri-, and postnatal factors, including general characteristics (gestational age, gender), breast or formula feeding, recorded birth-weight and height. Body-weight was measured using standard electronic scales and body length using an infant stadiometer. Children further underwent a lung function measurement that is described in detail elsewhere (17).

After the visit, parents were asked to report on the child's health status, including measures of weight and height of their babies as measured at the infant welfare center, by filling out a questionnaire monthly, for a period of 12 months.

Data on parental demographics, social background, and parental anthropometric measurements were obtained from the linked database of the Utrecht Health Project (UHP), a large health-monitoring study in Leidsche Rijn (18), which aims to generate data from all inhabitants on determinants of disease.

Follow-up visit at age 5 years

Children participating in the WHISTLER birth cohort were reinvited to visit the out-patient clinic at the age of 5 years. Lung function measurements were performed, not further described here. Prior to the visit a questionnaire was completed by the parents to collect information of both child and parental determinants concerning the previous years.

Anthropometric measurements were obtained, including body-weight and height with the participants wearing indoor clothes without shoes. Standing with the feet slightly apart, waist circumference was measured in duplicate at the level midway between the lowest rib border and the iliac crest, and hip circumference at the widest level over the major trochanter, to the nearest mm. Body mass index (BMI; kg/m²) and waist-to-hip ratio were calculated. In addition, intra-abdominal and subcutaneous fat were measured using ultrasound according to a previously described procedure (19,20), with a Picus Pro system (Esaote, Italy), using a CA 421 convex transducer. For intra-abdominal fat, the distances between the posterior edge of the abdominal muscles and the lumbar spine were measured using electronic callipers. Distances were measured from three different angles: medial and left and right lateral, with the transducer placed longitudinally on a straight line drawn between the left and right mid-point of the lower rib and iliac crest. Measurements were performed at the end of a quiet expiration. The mean distance was calculated from the three angles. Placing the probe transversely at the level of the umbilicus, sonographic subcutaneous fat was measured with electronic callipers, from the external face of the rectus abdominis muscle (linea alba) to just below the skin. The measurement was repeated three times, and the mean was used for analysis. For all measurements, minimal pressure was applied to eliminate the distortion of tissue. Coefficients of variation based on measurements by one observer in 10 and respectively 11 subjects at two different occasions for intra-abdominal and subcutaneous fat were 8.9% and 12.6%, respectively.

Intra-abdominal and subcutaneous fat measurements were successfully assessed in 271/306 (89%) and 298/306 (97%) subjects, respectively. Vascular measures did not differ between children with and without successful fat measurement.

In all subjects, vascular properties of the right common carotid artery (CCA) were studied using high-resolution echotracking technology (Art.lab (21), Esaote, Italy) including the use of a 128 radio frequency (RF) line multiarray, with a L10-5 40-mm linear array transducer. Rough RF data were analyzed online, and 6-second cineloops were stored without compression (120 Mbytes) for offline analysis. This novel technology gives access to all major mechanical parameters for 4-cm arterial segments: diastolic diameter d, the change in diameter as function of time (distension), and carotid intima-media thickness (CIMT). CIMT and diameter were measured with 2.1-μm resolution, and distension was measured with 1.7-μm resolution (22). Measurements were performed with subjects in a supine position, after at least 10 min of resting. The measurement procedure was repeated a maximum of four times during one session, for both the distension and CIMT measurement. Reproducibility was evaluated for the measured common carotid distension, diameter, and CIMT. Coefficients of variation based on measurements by one observer in 10 subjects at two different occasions for distension, CIMT, and diameter were 7.1%, 4.3%, and 2.4%, respectively.

During ultrasonography, blood pressure was recorded twice at the brachial artery using a semi-automatic oscillometric device (DINAMAP; Critikon, Tampa, FL, USA). The mean of these values was used to calculate the blood pressure in the carotid artery (see Supplementary Appendix) in order to estimate the local pulse pressure, assuming mean arterial pressure (MAP) minus DBP constant throughout the large artery tree. The means of the acquired parameters (CIMT, diameter, distension) for every session per individual were computed. From these data, the elastic properties of the artery as a hollow structure were assessed through cross-sectional distensibility (DC). The elastic properties of the arterial wall material were estimated with Young's elastic modulus (EM). Formulas and units for the separate outcomes can be found in the Supplementary Appendix.

CIMT and distension measurements were successfully assessed in 301/306 (98%) and 275/306 (90%) subjects, respectively. In 13/275 cases DC could not be calculated, as blood pressure was lacking (2 persistently refused, and in the first 11 participants blood pressure was not measured during the distension measurement).

All ultrasonic assessments were performed by the investigator (C.G.) and a trained pediatric research nurse, both blinded to data on confounders. The ultrasonographic measurements took approximately 40 minutes to complete. During the measurements children were allowed to watch their favorite motion pictures.

Confounding factors

Infant feeding was considered as possible confounder because of its putative relation with body size measures and vascular characteristics. Formula-fed children have long been known to grow according to a different pattern compared to breast-fed children (23), while breast feeding was shown beneficial to the adult vasculature (24,25) although not yet confirmative (26,27). Secondly, the associations with body size measures and vascular characteristics were adjusted for birth-weight in a separate model, in order to study if birth-weight can explain associations, whereas birth-weight has been related to overweight in childhood and to CVD at adult age (28,29). Furthermore, we accounted for age and sex in the analyses and for height and BMI in analyses with abdominal fat measures. The association of body-weight, BMI, and abdominal fat measures with CIMT was additionally adjusted for systolic blood pressure (SBP). Blood pressure could be an intermediate factor in the association, as it is well known to be related to both body size (30) and CIMT in childhood (31).

Data analysis

For descriptive purposes and for evaluation of possible confounding, means and variance measures of parent and child characteristics at several time intervals were calculated by BMI tertiles at age 5 years. In addition, base-line characteristics (of the neonatal visit and before) of non-participants (including those lost to follow-up) were compared to participants of the follow-up visit at age 5 years. Differences were tested using analysis of variance or chi-square tests whenever appropriate.

The relations between the different vascular parameters and measures of body size and body fat were assessed with univariable linear regression analysis, separately, with CIMT, carotid arterial wall distensibility, and EM as dependent variables, and the different measures of body size and body fat as independent variables. The same models were used to adjust the associations for possible confounders. Magnitudes of the relations of CIMT (μm), distensibility (MPa⁻¹), and EM (kPa) were estimated by the linear regression coefficient for every unit increase in BMI (kg/m⁻), body-weight (kg), height (cm), waist and hip circumference (cm), waist-to-hip ratio (WHR), and intra-abdominal fat (cm) and subcutaneous fat (mm), separately.

All results are expressed as regression coefficients with 95% confidence intervals and corresponding P values. Statistical significance was considered reached at P_2-sided < 0.05. All analyses were performed with SPSS version 17.0 for Windows.

Results

Overall, mean CIMT was 385.9 μm (SD 37.3), mean distensibility was 96.7 per MPa (SD 24.0), and mean EM was 155.4 kPa (SD 43.6). Table I shows characteristics by BMI tertiles at age 5 years. There were more boys in the higher tertiles. Furthermore, across the successive BMI tertiles, the different measures of body size increased, including birth-weight, weight at the neonatal visit, and all measured body size parameters at age 5 years. Children in the upper BMI tertile also had higher systolic blood pressure. On average, fathers had higher body-weight with similar height, and mothers were heavier and shorter, but this was not statistically significant. Parental BMI increased across the increasing BMI tertiles of their offspring. Across the increasing BMI tertiles, CIMT and EM increased, and distensibility decreased.

Table I. Base-line characteristics of WHISTLER offspring by BMI category.

	BMI tertiles at age 5 years (kg/ m^2)
	First tertile, 11.4 thru 14.6 (n = 102)	Second tertile,14.6 thru 15.8 (n = 102)	Third tertile,15.8 thru 23.3 (n = 102)	P value
WHISTLER offspring at birth
Gender (% males)	36.3	42.2	54.9	0.02
Birth-weight (g)	3363.2 (483.4)	3531.9 (467.3)	3574.7 (557.1)	0.01
Birth length (cm)	50.7 (2.3)	51.0 (2.1)	51.0 (3.0)	0.68
Gestational age (weeks)	39.5 (1.6)	39.4 (1.3)	39.5 (1.4)	0.96
Mother's age (years)	32.8 (3.6)	32.1 (3.6)	32.0 (3.7)	0.22
WHISTLER neonatal visit
Age first visit (weeks)	4.6 (1.3)	4.5 (1.1)	4.4 (1.4)	0.57
Body-weight (g)	4137.9 (561.0)	4322.4 (593.5)	4408.7 (657.7)	0.01
Height (cm)	54.0 (2.1)	54.2 (2.2)	54.7 (2.4)	0.09
Systolic blood pressure (mmHg)	85.7 (11.0)	83.6 (10.8)	83.1 (11.6)	0.52
Diastolic blood pressure (mmHg)	39.1 (9.2)	36.9 (9.5)	35.8 (9.4)	0.25
Infant feeding (%)
Breast feeding	55.5	55.5	56.2	0.72
Formula feeding	26.7	28.3	21.4
Breast and formula feeding	17.8	16.2	22.4
WHISTLER follow-up age 5 years
Age (years)	5.3 (0.2)	5.4 (0.4)	5.4 (0.3)	0.19
Body-weight (kg)	18.1 (1.7)	19.9 (1.9)	22.8 (2.7)	< 0.001
Body height (cm)	113.6 (4.5)	114.5 (5.4)	116.0 (4.5)	0.003
Intra-abdominal fat (mm)	35.5 (5.2)	37.4 (6.4)	38.9 (6.5)	0.001
Subcutaneous fat (mm)	5.4 (1.7)	6.3 (1.8)	9.2 (4.6)	< 0.001
Waist circumference (cm)	49.8 (2.3)	51.7 (2.5)	55.3 (4.2)	< 0.001
Hip circumference (cm)	55.8 (2.7)	59.2 (3.0)	64.1 (4.4)	< 0.001
Systolic blood pressure (mmHg)	104.7 (7.2)	105.5 (7.3)	108.6 (8.2)	0.001
Diastolic blood pressure (mmHg)	54.9 (7.5)	54.8 (6.1)	56.2 (7.3)	0.28
Carotid intima-media thickness (μm)	380.4 (37.4)	387.4 (32.8)	389.9 (41.0)	0.17
Distensibility (MPa^−1)	98.2 (25.8)	97.5 (21.1)	94.5 (25.2)	0.58
Elastic modulus (kPa)	152.0 (41.7)	150.4 (40.9)	163.8 (47.3)	0.09
Mothers
Number	95	85	91
Body-weight (kg)	69.1 (11.3)	71.6 (12.2)	72.2 (11.4)	0.16
Body height (cm)	170.2 (6.2)	170.3 (6.6)	168.9 (6.8)	0.29
BMI (kg/m^2)	23.9 (3.8)	24.7 (4.0)	25.3 (3.8)	0.04
Fathers
Number	77	66	72
Body-weight (kg)	83.0 (12.0)	85.2 (11.0)	88.4 (11.0)	0.01
Body height (cm)	183.2 (7.8)	183.0 (7.3)	183.2 (7.5)	0.97
BMI (kg/m^2)	24.7 (3.1)	25.5 (2.9)	26.3 (2.6)	0.001
Socio-economic status (%)
Low education (primary school)	6.0	7.6	4.7
Middle education (secondary school)	38.6	43.0	36.5	0.79
Higher education	55.4	49.4	58.8

Values are means (SD) unless otherwise indicated.

In Table II it is shown that gestational age is shorter in non-responders (not willing to participate or lost to follow-up), mothers were younger, and their babies were shorter at birth, compared to those who participated at age 5 years. Furthermore, non-responders were less often breast-fed, and mothers of non-responders were shorter and had lower body-weight; BMI did not differ.

Table II. Base-line characteristics of the non-responders (No) compared to the participants of the follow-up visit at age 5 year (Yes).

	Yes (n = 374)	No (n = 249)	P
WHISTLER offspring at birth
Gender (% males)	45.3	51.1	0.17
Birth-weight (g)	3507.0 (512.3)	3431.1 (621.2)	0.12
Birth length (cm)	51.0 (2.4)	50.5 (2.8)	0.05
Gestational age (weeks)	39.5 (1.4)	39.2 (1.7)	0.04
Mother's age (years)	32.6 (3.7)	31.9 (3.5)	0.03
WHISTLER neonatal visit
Age first visit (weeks)	4.5 (1.3)	4.7 (1.3)	0.03
Body-weight (g)	4306.6 (620.1)	4375.6 (722.3)	0.24
Length (cm)	54.4 (2.3)	54.3 (2.7)	0.50
Infant feeding (%)
Breast feeding	57.7	47.8
Formula feeding	24.5	27.9	0.049
Breast and formula feeding	17.9	24.3
Mothers
Age at birth child (year)	32.3 (6.0)	32.0 (3.6)	0.60
Body-weight (kg)	71.1 (12.0)	68.5 (13.2)	0.03
Body height (cm)	169.9 (6.5)	168.3 (7.3)	0.02
BMI (kg/m^2)	24.6 (4.0)	24.2 (4.3)	0.23
Smoking during pregnancy (% yes)	8.1	9.5	0.56
Fathers
Body-weight (kg)	85.3 (11.6)	83.5 (11.8)	0.12
Body height (cm)	183.4 (7.6)	183.3 (7.9)	0.88
BMI (kg/m^2)	25.4 (3.0)	24.8 (3.0)	0.08
Socio-economic status (%)
Low education (primary school)	6.3	8.8
Middle education (secondary school)	39.2	36.5	0.61
Higher education	54.5	54.7

Values are mean (SD) unless otherwise indicated.

Table III shows that increasing body-weight, waist circumference, and intra-abdominal fat were each statistically significantly related to thicker CIMT, lower distensibility, and higher EM. These associations were independent of sex, age, and infant feeding and additionally, with regard to intra-abdominal fat and waist circumference, independent of height. Furthermore, an increase of 1 kg/m² BMI was related to 2.90 μm thicker CIMT (95% CI 0.04, 5.8) and 5.34 kPa increase in EM (1.5, 9.2) independent of other factors. Greater hip circumference was related to 1.09 MPa⁻¹/cm (−2.0, −0.2) lower distensibility and 2.58 kPa/cm (0.9, 4.2) increase in EM, in both uni- and multivariable analysis. The association between hip circumference and CIMT disappeared after adjustment for sex, age, infant feeding, and height. Height was independently associated with CIMT (1.26 μm/cm; 95% CI 0.3, 2.2). Subcutaneous fat and waist-to-hip ratio were not related to any of the vascular outcome parameters. Adjustment for birth-weight in addition to sex and age, and height for abdominal measurements (Table III, model 2), could not explain the associations.

Table III. The relation between different measures of body size and vascular parameters in 5-year-old children.

	Crude			Model 1			Model 2
	Linear regression coefficient	95% CI	P	Linear regression coefficient	95% CI	P	Linear regression coefficient	95% CI	P
CIMT (μm)
Body-weight (kg)	1.97	0.5, 3.4	0.01	2.25	0.8,3.7	0.003	2.44	0.9,4.0	0.002
Body height (cm)	0.91	0.06, 1.8	0.04	1.26	0.3, 2.2	0.01	1.33	0.4, 2.3	0.006
BMI (kg/m^2)	3.00	0.2, 5.8	0.04	2.90	0.04, 5.8	0.047	3.10	0.2, 6.0	0.04
Waist circumference (cm)	1.64	0.6, 2.7	0.003	1.27^a	0.08, 2.5	0.04	1.36^a	0.16,2.6	0.03
Hip circumference (cm)	1.22	0.2, 2.3	0.02	0.91^a	−0.4, 2.2	0.16	0.94^a	−0.4, 2.2	0.15
Waist-to-hip ratio	59.9	−49.1, 168.8	0.28	39.9	−75.1, 154.9	0.50	46.7	−69.4, 162.8	0.43
Subcutaneous fat (mm)	0.35	−0.9, 1.6	0.58	−0.06^a	−1.4, 1.3	0.93	−0.04^a	−1.4, 1.3	0.96
Intra-abdominal fat (cm)	10.3	3.0, 17.6	0.01	10.1^a	2.9, 17.3	0.01	9.87^a	2.6, 17.2	0.01
Carotid end-diastolic diameter (mm)	22.6	12.2,33.0	< 0.001	23.2^b	12.1,34.3	< 0.001
Distensibility (MPa^−1)
Body-weight (kg)	−1.44	−2.5, −0.4	0.01	−1.23	−2.4, −0.1	0.03	−1.20	−2.4, −0.05	0.04
Body height (cm)	−0.58	−1.2, 0.04	0.07	−0.45	−1.1, 0.2	0.18	−0.40	−1.1, 0.3	0.24
BMI (kg/m^2)	−2.39	−4.5, −0.3	0.03	−2.10	−4.2, 0.04	0.05	−2.07	−4.2, 0.09	0.06
Waist circumference (cm)	−1.45	−2.3, −0.6	< 0.001	−1.33^a	−2.2, −0.4	0.004	−1.31^a	−2.2, −0.4	0.004
Hip circumference (cm)	−1.10	−1.9, −0.4	0.004	−1.09^a	−2.0, −0.2	0.02	−1.09^a	−2.0, −0.2	0.02
Waist-to-hip ratio	−35.3	−112.1, 41.5	0.37	−33.6	−114.2, 47.1	0.41	−32.9	−114.2, 48.3	0.43
Subcutaneous fat (mm)	−0.34	−1.2, 0.6	0.46	−0.40^a	−1.4,0.6	0.42	−0.38^a	−1.3, 0.6	0.43
Intra-abdominal fat (cm)	−6.36	−11.5, −1.3	0.02	−6.06^a	−11.2, −1.0	0.02	−6.04^a	−11.2, −0.9	0.02
Elastic modulus (kPa)
Body-weight (kg)	3.22	1.3, 5.2	0.001	2.73	0.7,4.7	0.01	2.60	0.5, 4.7	0.01
Body height (cm)	1.02	−0.1, 2.2	0.08	0.73	−0.5, 1.9	0.23	0.60	−0.6, 1.8	0.34
BMI (kg/m^2)	6.13	2.4, 9.9	0.002	5.34	1.5, 9.2	0.01	5.19	1.3, 9.1	0.01
Waist circumference (cm)	3.06	1.6, 4.5	< 0.001	2.85^a	1.3, 4.5	0.001	2.79^a	1.2, 4.4	0.001
Hip circumference (cm)	2.35	1.0, 3.7	0.001	2.58^a	0.9, 4.2	0.002	2.63^a	1.0, 4.3	0.002
Waist-to-hip ratio	68.3	−71.5, 208.1	0.34	49.9	−95.6, 195.5	0.50	44.1	−102.5, 190.8	0.55
Subcutaneous fat (mm)	0.82	−0.8, 2.5	0.33	1.15^a	−0.6, 2.9	0.19	1.15^a	−0.6, 2.9	0.19
Intra-abdominal fat (cm)	11.3	2.0, 20.6	0.02	10.6^a	1.3, 19.9	0.03	10.7^a	1.3, 20.0	0.03

Model 1: adjusted for sex, age, and breast feeding. Model 2: adjusted for sex, age, and birth-weight.

^aWaist and hip circumference and subcutaneous and intra-abdominal fat measures are adjusted for height in addition.

^bCarotid end-diastolic diameter is adjusted for sex, age, and body height.

Adjustment for SBP, in addition to age and sex, and height for intra-abdominal fat and waist circumference, did not influence the association between CIMT and body-weight (2.36 μm/kg; 0.7, 4.0; P = 0.004), waist circumference (1.57 μm/kg; 0.3, 2.9; P = 0.02), and intra-abdominal fat (10.9 μm/cm; 3.4, 18.4; P = 0.005). The association with BMI was slightly attenuated (3.00 μm/kg/m²; −0.1, 6.1; P = 0.06).

In Table IV it is shown that adjustment for BMI in addition to age, sex, and breast feeding did not change the association with waist circumference and the vascular outcome measures, nor did the association with intra-abdominal fat and CIMT. The association with intra-abdominal fat and distensibility and, separately, EM was attenuated.

Table IV. The association between vascular characteristics and abdominal fat with adjustment for BMI.

Adjusted for BMI and sex, age, infant feeding	Linear regression coefficient	95% CI	P value
CIMT (μm)
Waist circumference (cm)	2.17	0.4, 4.0	0.02
Intra-abdominal fat (cm)	9.19	1.4, 17.0	0.02
Distensibility (MPa^−1)
Waist circumference (cm)	−1.70	−3.0, −0.4	0.01
Intra-abdominal fat (cm)	−4.91	−10.3, 0.4	0.07
Elastic modulus (kPa)
Waist circumference (cm)	2.77	0.5, 5.1	0.02
Intra-abdominal fat (cm)	7.13	−2.6, 16.8	0.15

Discussion

Our study shows that increased general body mass and particularly intra-abdominal fat are related to thicker and stiffer arteries already in young childhood.

To appreciate these findings, some issues need to be discussed. Participants in our study were an unselected population of healthy young children. It might be that more health-concerned parents were more likely to participate. However, it seems unlikely that this would have biased our findings, as that would require the association in non-responders to be in the opposite direction. There is no reason to assume that the biological mechanism would be different in those not participating. We believe we have accounted for all likely confounders. Although the investigators were not masked to the general appearance of the participants, automated vascular measurements avoided observation bias in the associations. The inter-session variation in the distension, CIMT, and diameter of the common carotid artery in our study is low and comparable to values estimated in 12 healthy adult volunteers with a similar ultrasound device (6%, 5%, and 1%, respectively). Moreover, the RF-based method was shown to be reliable in measuring CIMT in clinical practice (32). Whereas CIMT and distensibility are acknowledged proxies of cardiovascular disease risk in adulthood (33), associations between these measures at age 5 years and manifest cardiovascular disease in later life have never been studied and can therefore only be assumed. We performed a cross-sectional study which includes the possibility of reverse causality. However, although we cannot exclude that possibility, we find it plausible to assume that body size influences vascular characteristics rather than the other way around.

To our knowledge, this is the first study to demonstrate an association between childhood body size measures and vascular characteristics at the age of 5 years. Previous studies with similar observations were performed with children as young as 8 years. Those studies showed strong evidence for a relation of BMI and both arterial stiffness and endothelial damage (34–37). Also, a majority of studies report an increased CIMT in obese children, compared to lean children, as well as an association between BMI and CIMT (9,10,34,36–41). The present study shows the effect of increasing BMI, but especially of increasing waist circumference and intra-abdominal fat, on both functional and structural vascular properties in unselected young healthy individuals. Our findings add that associations are already present at a very young age. Moreover, most studies compared obese or overweight children with lean controls, whereas the present study shows evidence for a graded relation between general body mass and body fat with vascular parameters.

The results of the present study have to be interpreted realizing that much variation in body size, as well as in vasculature, is compatible with varying growth rates of children. We found body height to be a determinant of CIMT, in conformity with other studies (42,43). It seems plausible that increased arterial wall thickness is an adaptation in order to maintain, or even increase, blood flow through the lengthening arterial tree. Nevertheless, body height did not fully explain the associations we found between CIMT and intra-abdominal fat, indicating that intra-abdominal fat has an independent effect. The independent association of CIMT with lumen diameter observed in our study agrees with findings in younger adults (44). It could reflect an adaptive response of the arterial wall to changing hemodynamic circumstances in order to preserve lumen diameter and restore normal blood flow, known as compensatory lumen enlargement (45). Most likely in these young children, it is an adaptive response due to changes in local blood pressure and as such in local tension in the arterial wall.

The ultrasonographic assessment of abdominal fat was validated in adults. For subcutaneous fat this was confirmed in children aged 6–7 years, whereas the association between ultrasonographic measurement of intra-abdominal fat and CT was less correlated. However, this study showed good correlations with waist circumference and intra-abdominal fat measured with CT (46). In our study too, waist was associated with the vascular measurements, suggesting that the associations found with ultrasonographically measured intra-abdominal fat are reliable. Thereby, we did find a relation between ultrasonographically measured intra-abdominal fat and vascular characteristics, whereas if the measurement was completely not measuring what it should, it seems highly unlikely that any association would be identified.

Waist-to-hip ratio was not associated with any of the vascular outcome measures. However, in contrast to waist circumference, waist-to-hip ratio correlates poorly with central adiposity in children (47). Our findings indicate that fat localization in children may be a better risk marker than general measures such as waist-to-hip ratio or body mass index.

Indeed, our study shows that increased visceral abdominal fat is associated with adverse vascular health, independent of BMI. Specifically, the association between visceral abdominal fat and CIMT was also independent of blood pressure. Adipose tissue is known to secrete vasoactive compounds such as leptin, TNF-α, and interleukin-6, known as adipokines, which possibly directly affect vascular function (7). Moreover, many of the adipokines act as inflammatory mediators (48). Accumulation of visceral fat is associated with elevated inflammatory markers, such as C-reactive protein, even in children as young as 3 years of age (49). Inflammation plays an important role in all stages of atherosclerosis, including the initial activation of endothelial cells, to progression and, ultimately, its final complication of thrombosis (50). Furthermore, visceral abdominal fat has a direct circulatory connection to the liver. Excessive release of free fatty acids from visceral adipose tissue directly in the portal circulation might lead to insulin resistance and hyperlipidemia, both established risk factors for CVD (51).

Our findings emphasize the importance of preventive measures against overweight to be taken from an early age. Beneficial effects of weight loss on the child's vasculature have been shown previously, suggesting that the changes are reversible (52). However, since weight loss is hard to achieve (14), the very occurrence of excessive weight gain should probably be prevented (13). From a cardiovascular disease point of view, however, the huge time gap between such preventive action and manifest disease is a challenge. Our finding that overweight, particularly visceral adipose tissue, can cause early-life pre-atherosclerotic changes provides support for early-life prevention.

Acknowledgements

The authors gratefully acknowledge all the parents and children who participated, Liesbeth van der Feltz-Minkema for her dedicated assistance with the field-work, Myriam Olling-de Kok for secretarial support to the WHISTLER project, and Jildou Zwerver for data management. The core laboratory ‘Vascular Imaging Center’ personnel are acknowledged for their contribution in the vascular measurements.

Declaration of interest: The WHISTLER birth cohort was supported with a grant from the Netherlands Organization for Health Research and Development (grant no. 2001-1-1322) and by an unrestricted grant from GlaxoSmithKline. WHISTLER-Cardio was supported with an unrestricted strategic grant from the University Medical Center Utrecht, The Netherlands. The authors declare no other conflicts of interest.