Cardiovascular Risk Factors in 7-13 Years Old Children from Vojvodina (Serbia).

BACKGROUND: Atherosclerosis is a chronic inflammatory disease which starts early in life and depends on many factors, an important one being dyslipoproteinemia. According to several studies, atherosclerotic plaques or their precursors could be seen in children younger than 10 years. During later life, interaction with a sedentary way of life, as well as unhealthy nutrition, smoking, alcohol consumption, obesity and family history of cardiovascular disease cause the burden of atherosclerotic disease.
METHODS: Study included 624 children (316 boys, 308 girls), aged from 7-13 years. We analysed socio-demographic data (BMI, blood pressure, cardiovascular family history, smoking status), as well as lipid status with lipoprotein little a-Lp(a), and apolipoproteins: Apo AI, Apo B-100 for all children. This enabled us to calculate new atherogenic indices Tg/HDL-c, lipid tetrad index (LTI) and lipid pentad index (LPI). Cardiovascular risk for later life was estimated by using modified Risk Score for Young Individuals (RS), which divided the subjects according to the score level: low, medium and higher risk.
RESULTS: The older children (13 y) had better lipid status than the younger children, i.e. significantly lower total cholesterol, LDL-C, triglycerides and non-HDL-C concentration and significantly higher HDL-C concentration than the younger children and this was in accordance with the RS level. Children with a positive family history of CV disease had significantly higher Lp(a) concentration and blood pressure. LPI was significantly higher in children with a higher RS.
CONCLUSIONS: The results of our work could be used for cardiovascular risk assessment in apparently healthy children to provide preventive measures which could control the changeable risk factors.

Introduction

Atherosclerosis is a chronic inflammatory disease which starts early in life (1) and depends on many factors; the most important one is dyslipoproteinemia i.e. lipid metabolism disturbance. According to several studies, atherosclerotic plaques or their precursors could be seen in children younger than 10 years (2–5). During later life, interaction with sedentary lifestyle, coupled with unhealthy nutrition, smoking, alcohol consumption, obesity and family history of cardiovascular disease, causes the burden of atherosclerotic disease (2). Atherosclerosis is a prerequisite for different diseases development: acute myocardial infarction, stroke, peripheral vascular disease and many others. All of the abovementioned diseases are characterized with high mortality rate, and increased rate of patients′ disability, promoting atherosclerosis as an important economic and social problem of the modern societies, complex but preventable to a certain extent (3–5).

Increased values of total cholesterol (T-C), triglycerides (TG), LDL-cholesterol (LDL-C), non-HDL-cholesterol (non-HDL-C), and decreased HDL-C values are all together important risk factors for atherosclerotic changes in arteries in children (2). Moreover, increased Apo B and Lp(a) and decreased Apo AI also contribute to atherosclerosis risk (2). Lipid status parameters, as well as Apo AI and Apo B are dietary intervention modifiable parameters, and Lp(a) is an independent risk factor which is genetically determined and nonmodifiable during life (6, 7).

Obesity is also an independent risk factor for the development of dyslipidemia, hypertension, and thus cardiovascular disease in later life (8–10). According to the data of the Institute of Public Health of Serbia (11), people from Vojvodina (northern part of Serbia) have the highest acute coronary syndrome mortality rates per 100.000 subjects.

The aim of the present study was to explore cardiovascular risk in apparently healthy children regarding their general lipid status, Lp(a), apolipoproteins AI and B-100 concentration, obesity and cardiovascular related family history.

Materials and Methods

This study included 624 elementary school children (316 boys, 308 girls), divided in three groups according to their age: 7 years (190 subjects), 10 years (212 subjects) and 13 years (222 subjects). The study group was selected in South Banat (Vojvodina, Republic of Serbia) during regular medical check-up. All the children and parents read and signed their informed consent and all procedures were planned according to the ethical guidelines of Helsinki Declaration. The institutional ethical and review committee of the University of Belgrade Faculty of Pharmacy approved our study protocol. Data about family history regarding acute myocardial infarction and/or stroke were collected from parents by the interviewer during the same medical appointment.

Venous blood was drawn from the antecubital vein into evacuated, serum separator tubes, after 12 hours of overnight fasting. T-C and TG were assayed immediately by routine enzymatic methods using an Artax analyzer (Menarini, Florence, Italy) and Menarini reagents (Florence, Italy). HDL-C was isolated by using the precipitating method with phosphotungstic acid in the presence of Mg2+ ions. LDL-C concentration was calculated using the Friedewald formula (12).

Serum aliquots were preserved at –80 °C until analysis of apolipoproteins concentration (Apo AI and Apo B-100), and lipoprotein little a – Lp(a). Apo AI, Apo B and Lp (a) were assayed by routine immunoturbidimetric methods using the nephelometric analyser BN proSpec and spectrophotometric analyser Rxl Max. The instruments were purchased from the same manufacturer, DADA Behring (Margburg, Germany) with DADA Behring reagents (Margburg, Germany).

Non-HDL concentration is calculated as the difference between T-C and HDL-C, and atherogenic index as the ratio between TG and HDL-C concentrations. New atherogenic indices which involved several lipid parameters were calculated as lipid tetrad index (LTI) and lipid pentad index (LPI). LTI was calculated using the formula [Tc×Tg×Lp(a)] /HDL-c, and LPI was calculated using the formula: [Tc×Tg×Lp(a) × apoB]/apoA-I (13).

Body mass was measured using a portable electronic scale (Tanita, Amsterdam, Netherlands) to the nearest 0.1 kg. Height was measured to the nearest 0.1 cm using a portable wall-mounted stadiometer. Waist (just above the iliac crest) and hip (the widest part of the hips at the middle of the pelvis) circumferences were measured to the nearest 0.1 cm. Body mass index was calculated as body weight (kg) divided by height squared (m). Overweight was defined as a BMI ≥ 85th – < 95th percentile and obesity was defined as a BMI ≥ 95th percentile for children of the same age and sex. Blood pressure was measured by a mercury sphygmomanometer (Quicksilver-sphygmomanometer Riester, Jungingen, Germany).

Limits for prehypertension and hypertension status are defined according to the Expert Panel on Integrated Guidelines for Cardiovascular Health and Risk Reduction in Children and Adolescents (13).

The risk for CVD was estimated using LTI and LPI (14) and modified Risk Score for identifying young individuals with a high probability of having advanced atherosclerotic lesions, reported by McMahan et al. (15). According to risk score values, we have additionally divided the children into Low risk group (score lower or equal to 4 points), Medium risk group (5–8 points) and Higher risk group (above 8 points).

Statistical Methods

Parameters with a normal, Gaussian type of distribution are presented as means and standard deviations, but triglycerides, Lp(a), LTI, LPI deviated from normal distribution and are presented as geometrical mean and 95% confidence interval. For inter-group comparison, we have used ANOVA or Kruskal-Wallis test, with the distinct post hoc test where appropriate and chi-square test for frequency data comparison. ROC analysis was used to test the discriminatory capability of the measured and/or calculated parameters regarding low and higher risk status towards cardiovascular disease.

Results

We have compared basic lipid status parameters and parameters connected with lipid metabolism in children divided in groups according to their age, by using ANOVA and post hoc Tukey test for data with Gaussian distribution. Data which deviated from Gaussian distribution were compared with nonparametric methods (Kruskal-Wallis ANOVA followed by Mann-Whitney U test). The results are presented in .

Table I

Basic anthropometric, lipid status, and inflammation parameters in healthy children according to the age subgroups.

Parameter	Age subgroups
	7 y. (n=190)	10 y. (n=212)	13 y. (N=222)	P
Sex, boy/girl, n (%)	99/91 (52/48)	101/111 (48/52)	116/106 (52/48)	χ2=1.15, P=0.560
BMI, kg/m2	19.97±3.10	21.77±3.89aaa	19.84±3.51bbb	<0.001
Systolic BP, mm Hg	110.53±10.38	114.03±13.77a	111.02±13.46	<0.01
Diastolic BP mm Hg	72.24±10.50	75.78±12.00a	72.64±11.79b	<0.003
Positive family history for CVD, n (%)	10 (5.3%)	17 (8%)	25 (11%)	χ2=4.86, P=0.088
Glucose, mmol/L	4.62±0.53	4.68±0.47	4.70±0.53	0,158
TC, mmol/L	4.56±0.79	4.58±0.71	4.12±0.66aaa,bbb	<0.001
TG, mmol/L#	1.19 (1.12–1.27)	0.94 (0.89–0.99)aaa	0.87 (0.82–0.91)aaa	<0.001
LDL-C, mmol/L	2.44±0.93	2.69±0.70aa	2.14±0.82aa,bbb	<0.001
HDL-C, mmol/L	1.51±0.43	1.43±0.34	1.58±0.61 bb	<0.003
Non-HDL-C, mmol/L	3.04±0.88	3.16±0.74	2.53±0.75aaa,bbb	<0.001
Apo AI, g/L	1.34±0.38	1.37±0.29	1.45±0.40aa׳ b	0.003
Apo B-100, g/L	0.80±0.32	0.91±0.28aaa	0.80±0.31bbb	<0.001
Lp(a), mg/L#	222 (204–242)	171 (157–187)aaa	168 (143–197)aaa	<0.001
TG/HDL-C#	0.838 (0.555–1.203)	0.639 (0.480–0.876)aaa	0.538 (0.409–0.868)aaa,bb	<0.001
LTI#	722 (450–1695)	474 (271–757)aaa	280 (199–716)aaa	<0.001
LPI#	695 (373–1308)	417 (244–790)aaa	235 (126–600)aaa,bbb	<0.001
Fibrinogen, g/L	3.1 (2.6–3.6)	3.0 (2.6–3.5)	3.0 (2.7–3.3)	0.220

P<0.05, 0.001, compared to 7 y. group;

P<0.05, 0.001 compared to 10 y. group

– parameters with non-Gaussian distribution presented as 50th percentile (25th – 75th percentile); compared by nonparametric Kruskal-Wallis ANOVA and subsequent Mann-Whitney U test

The older children (13 y.) had better lipid status than younger children, i.e. significantly lower total cholesterol, LDL-C, triglycerides and non-HDL-C concentrations and significantly higher HDL-C concentrations than the younger children. Moreover, 13 y. old children had lower Lp(a), and higher Apo AI than the two younger groups. The oldest children also had the lowest LTI and LPI risk scores values. The number of children with positive family history for coronary heart disease increased with age, but the difference was not statistically significant. Afterward, we divided the children according to their family history for CV disease status and compared the other measured parameters. Our results showed that children with a positive family history of CV disease had significantly higher Lp(a) concentration, and blood pressure ().

Figure 1

Differences in Lp(a) concentration and blood pressure values in two categories related with presence of family history of cardiovascular disease in apparently healthy children.

*P<0.05 vs. Negative family history children (Lp(a) compared by Mann-Whitney U test; blood pressure compared with Student′s t test)

In order to estimate possible risk for cardiovascular disease development in apparently healthy children, we used the Risk Score Calculator proposed by McMahan et al. (15). This calculator uses non-HDL-C, HDL-C, smoking status, hypertension, hyperglycemia, and sex parameters for the risk score calculation.

shows risk score values in the age subgroups. Younger age groups of children have significantly higher risk score values compared to 13 y. old children (data are presented in as box plots of median values and interquartile range).

Figure 2

Risk score value in different age groups. aaa, bbbP< 0.001 vs. 7 y. and 13 y. respectively according to Kruskal-Wallis and then Mann-Whitney U test

Then, we divided the group of healthy children into three subgroups stratified by the calculated risk score level as Low, Medium and Higher risk groups. The results of this part of analysis are shown in . Unexpectedly, the youngest children had the highest level of risk for CVD development in later life when compared to the older children subgroups. These children had significantly higher BMI, and worse lipid status than the children with lower risk and also higher blood pressure. The number of positive family history for CVD children did not differ among the risk level subgroups. Regarding LTI and LPI values, children with higher risk score for CVD value had the latter index significantly higher compared to the same parameter in both lower risk groups ().

Table II

Biochemical parameters in healthy children (7–13 y.) according to risk score level subgroups.

	Risk score value			P
Parameter Points (number of children)	Low ≤ 4 (506)	Medium 5–7 (92)	Higher ≥ 8 (26)
Age	10.4±2.46	9.4±2.06aa	8.5±1.94aaa	<0.001
BMI, kg/m2	20.3±3.39	21.1±4.43	22.6±4.19aa	<0.01
TC, mmol/L	4.25±0.661	4.90±0.683aaa	5.78±0.350aaa, bbb	<0.001
TG, mmol/L	1.06±0,503	1.08±0.461	1.52±0.881aaa, bbb	<0.001
LDL-C, mmol/L	2.21±0.739	3.15±0.624aaa	3.91±0.497aaa, bbb	<0.001
TG/HDL#	0.615 (0.445–0.909)	0.811(0.580–1.153)	1.194 (0.714–1.536)a	0.011
Lp(a), mg/L#	179 (100-280)	170 (100-270)	240 (146-382)	ns
Apo AI, g/L	1.40±0.356	1.34±0.362	1.44±0.363	ns
Apo B-100, g/L	0.829±0.289	0.860±0.365	0.889±0.400	ns
Fibrinogen, g/L	3.04±0.551	3.00±0.589	3.02±0.550	ns
Systolic BP mm Hg	110±12	117±12aaa	123±12aaa	<0.001
Diastolic BP mm Hg	72±11	78±12aaa	82±16aaa	<0.001
Risk score value (points)#	2.0 (–1.0–3.0)	6.0 (5.0-6.0)aaa	8.0 (8.0-9.0)aaa, bbb	<0.001
Positive family history for CVD n (%)	41 (8%)	8 (9%)	3 (12%)	χ2=0.4, P=0.818

p<0.05, 0.01, 0.001, compared to Low Risk group;

P< 0.001 compared to Medium Risk group with ANOVA and Tukey test ns-non-significant;

– parameters with non-Gaussian distribution presented as 50th (25th – 75th percentile); compared by nonparametric Kruskal-Wallis ANOVA and subsequent Mann-Whitney U test

Figure 3

Lipid tetrad (LTI) and lipid pentad index (LPI) according to risk score values subgroups. (aaP<0.01, aaaP<0.001 vs. Low risk group, bP<0.05, bbP<0.01 vs. Medium risk group)

We also wanted to test the lipid status, inflammation and BMI parameters′ capability to discriminate between Low Risk and Higher Risk status, so we performed the receiver operating characteristics curve (ROC) analysis. The parameters analysed were those which were not used for Risk Score calculation. The results of this analysis are shown in and .

Table III

The results of ROC analysis for discriminating two risk levels i.e. risk score below 4 and risk score ≥ 8.

Variable	AreaC	95% Confidence Interval	Std. Error
TC	0.969***	0.949–0.989	0.010
LDL-C	0.968***	0.941–0.995	0.014
Tg	0.663**	0.551–0.776	0.057
Lp(a)	0.601	0.495–0.707	0.054
Apo AI	0.566	0.452–0.679	0.058
Apo B	0.469	0.352–0.587	0.060
LTI	0.562	0.447–0.677	0.059
LPI	0.680**	0.584–0.776	0.049
BMI	0.662**	0.544–0.780	0.060
Fibrinogen	0.499	0.378–0.620	0.062

P<0.01 and 0.001 respectively for AUC values compared to 0.5

Figure 4

ROC curves for lipids, lipoproteins, apolipoproteins, BMI and fibrinogen and their ratios as markers of risk of cardiovascular disease development in later life. Panel A) HDL-c, TC, LDL-C, TG, Lp(a); panel B) BMI, fibrinogen; panel C) Apo AI, Apo B-100, LTI and LPI.

Total cholesterol and LDL-C had the best discriminatory potential, which was not unexpected because non-HDL cholesterol was used for the Risk Score calculation. LPI and BMI had significantly lower AUCs and thus lower discriminating power, but still the AUCs were significantly higher than 0.5.

Discussion

Many studies, among them YUSAD study (16) and Bogalusa study (17, 18) indicate a close relationship between lipid and nonlipid atherogenic risk factors in early childhood and atherosclerosis development in later life. Identification of the risk factors and their limitation lead to significant decrease of the possibility of cardiovascular disease development (19).

BMI values among the age groups increased with age when we compared the 7 y. and 10 y. subgroups (19.97±3.10 kg/m2 vs. 21.77±3.89 kg/m2, P<0.001) and decreases in 13 y subgroup (19.84± 3.51, P<0.001 compared to 10 y. subgroup). This BMI changing with the children′s development is the same as in Ninic et al. (20), who assessed 7–15 y. old children from Vojvodina. Systolic and diastolic pressure values are comparable with Lurbe et al. data (21) for healthy children and adolescents aged 1–17 y. There was no difference in blood pressure values among different age groups ().

Triglycerides, non-HDL-C and Tg/HDL-C ratio () are higher than Ninic et al. (20) reported. Older children in our current study had better lipid status than the younger ones; significantly lower total cholesterol values, LDL-C, triglycerides, non-HDL-C and Lp(a), and also significantly higher HDL-C and Apo-AI. Lipid parameter values in children from this study are within the recommended boundaries given by Serbian National Guide for Good Clinical Practice (22–24). According to this recommendation for lipid disorder diagnosis and treatment, for children, borderline LDL-C concentrations are considered those higher than 2.85 mmol/L and high are above 3.36 mmol/L; borderline high total cholesterol value is above 4.40 mmol/L and T-C values higher than 5.17 mmol/L are considered high. Schulpis et al. (25, 26) have investigated lipid parameters′ distribution in Greek children and their results showed slightly higher levels of serum HDL-C and lower levels of TG, LDL-C and Lp(a) than in our children of similar age. Better lipid status of the Greek schoolchildren could be explained with the genetic factors and by the nutritional habits of the Greek population, regarding consumption of olive oil and other components of the Mediterranean diet. This finding compared to our results could imply that healthier nutritional habits could be protective concerning the risk for CVD development in later life.

LTI and LPI indices are calculated in order to assess more precisely the cardiovascular risk coming from a detailed lipid status. It is evident that older children had lower LTI and LPI compared to younger children (). The LTI and LPI indices have recently been described as a new form of assessment of lipid profiles, which have been analysed in some populations (27). A characteristic of these new indices is the broad approach to atherogenic and nonatherogenic lipid particles, resulting in a single value. Based on the conventional lipid profile and the emerging risk factors such as Lp(a), apolipoproteins AI and B-100, the LTI and LPI appear as models in global risk assessment, reflecting the multifactorial nature of CVD. Bogalusa (17) and YUSAD (16) studies analyzed the relation between lipid status and age of children. The YUSAD study (16) demonstrated that over a 5 and 10-year period (10 y. children compared to 15 y. and 20 y. children), BMI and waist circumference values showed significant increase. Such an observation is in correlation with the physiological changes (growth) during this period. However, the proportion of the waist circumference increase was higher in girls between 10 and 15 years of age than between 15 and 19–20 years of age, while there was less discrepancy in the difference between the proportions of the two 5-year age groups in boys′ population. Moreover, the authors of the Bogalusa study (17, 18) showed a connection between obesity and lipid status, and suggested a relation between high insulin concentration and cardiovascular risk. Children from our study were all normoglycemic, so we supposed that their insulin concentrations were among reference values (28), and did not contribute to increased CV risk.

BMI and lipid status values are comparable with values reported by Costa et al. (29), who found a negative correlation between BMI and HDL-C values, and a positive correlation between BMI and triglyceride concentration in a group of children younger than 12 y. Dai et al. (30) and Jolliffe and Janssen (31) have similar results and have commented about lipid status dependence on age. The age and gender specific threshold values reflect the natural fluctuation in lipoprotein concentration that occurs with growth and maturation. Difference in lipid status between 13 y old children and the younger ones could be explained by their puberty status and the intensive growth period of older children. Namely, the metabolism of puberty period children is devoted to rapid growth and protein synthesis (16, 21, 32). The decrease in serum lipids levels during puberty is also related to an increase in the plasma testosterone concentration in boys or estradiol concentration in girls (32).

We included 624 children in total, among them 52 (8.3%) with a positive family history regarding cardiovascular disease (presence of hypertension, acute myocardial infarction or stroke in first or second line relatives). As expected, the number of positive family history cases increased with increasing age (5.3%, vs. 8.0% vs. 11%, with a marginally statistically significant difference P=0.088), because in younger children relatives are also younger and with smaller chances of having some unwanted cardiovascular related condition. Our results showed significantly higher Lp(a) and blood pressure values in children with positive family history compared to children without family history of CVD (). Still, Lp(a) values in our group of children with CVD family history were lower than in Italian children of similar age who also had CVD family history (19), but higher compared with healthy 14 y. old Swedish children from Bergstrom′s study (22). It is well-known that Lp(a) is an independent CV risk factor (6, 7) and increased values of this parameter could predict early cardiovascular disturbance development, even if other riskfactors are within a tolerable range (18). On the contrary, Guardamagna et al. (19) considered that the mutual involvement of increased Lp(a) values with positive family history was directly connected with increased risk for cardiovascular development. Similar findings have been presented by others (29, 31), regarding positive correlation between increased Lp(a) values and LDL-c concentrations, as well as Lp(a) and total cholesterol values and Apo B-100 (19), especially in children with positive family history. According to our results, significantly higher Lp(a) concentration and blood pressure values reflected additional risk factor existence such as positive family history of CVD.

In order to assess the real cardiovascular risk level, we implemented the scale for Cardiovascular Risk Score (CRS) calculation from McMahan′s study (15). That study correlated the cumulative effects of modifiable risk factors (non-HDL-cholesterol, HDL-cholesterol, smoking, hypertension, obesity and hyperglycemia) and the level of atherosclerotic deposits on coronary arteries. Scoring system given by McMahan et al. (15) gave the ratio between RS level and probabilities of having significant stenosis on coronary arteries, where score 10 gives the probability of 10%, score 20 – 25% and score 30 – almost 60% for having significant atherosclerotic lesions. In our current study, apparently healthy children aged 7–13 y. had RS levels from –1 to 11, which means that at the moment, children had about 10% probability of having significant atherosclerosis on coronary arteries. However, in order to grade the risk level, we divided the children according to tertile values for this group; i.e. Low CRS ≤ 4 (N=506), Medium CRS 5–7 (N=92) and Higher CRS ≥ 8 (N=26) (). Along with better lipid status, older children had significantly lower CRS values than younger children (). Nonetheless, children with the highest CRS values had also significantly higher values of total cholesterol, triglycerides, LDL-c, Tg/HDL-c ratio, Lp(a) concentration and systolic and diastolic blood pressure (). These findings are in accord with other studies (32–35) which assessed risk factors in adolescents or young adults and then measured carotid artery intima-media thickness (IMT) in the same individuals, 12 to 22 years later. In all of these studies, one or more of the risk factors were associated with the carotid IMT of individuals. The study of Bild et al. (36) found a similar association between coronary arteries calcification and risk factors measured 10 years earlier. Oren et al. (37) found strong associations between current risk factors and carotid IMT in young adults. LTI and LPI values are also higher in higher CRS values groups, and this finding is in agreement with Morais, McMahan and Das′ work (14, 15, 27). These findings reinforce the idea that the conventional lipid profile in many cases may not offer an actual estimation of coronary risk. When this profile is under reference limits, it does not necessarily rule out the risk due to specific forms of dyslipidemia. Consequently, the introduction of laboratory tests related to an unconventional lipid profile would be extremely desirable in search of the real coronary risk estimate, as an addition to all the other traditionally investigated risk factors.

We wanted to estimate which of the measured parameters could predict CRS value, so we implemented a ROC analysis. The result of ROC analysis showed that, besides total cholesterol and LDL-C which are directly connected with non-HDL-C, as one of the points for the CRS calculation, LPI and BMI had good predictive potential (, ). This suggested that the values of the comprehensive parameter LPI, which included several lipid parameters, and which is the most important Lp(a), could be a reliable sign of growing cardiovascular risk. Our results are in agreement with Bogavac-Stanojevic et al. study (38). We have considered only lipid, anthropometric and family history related risk, as some kind of traditional cardiovascular risk factors. Karikas et al. (39) proposed assessment of several other new, emerging risk factors in early age. Those parameters were homocysteine and paraoxonase activity, as already established cardiovascular risk factors in the adult population. According to this, the additional work in atherosclerosis related risk factors recognition in children is certainly warranted in the future.

We believe that the results of our current work could be used by clinicians in order to arrange preventive measures to control changeable risk factors in children.