Relationship of C-Reactive Protein to Adiposity, Cardiovascular Risk Factors and Subclinical Atherosclerosis in Healthy Children

INTRODUCTION

The available evidence suggests that the pathogenesis of atherosclerosis begins in childhood. Anatomic pathology studies have shown the presence of early precursors of atherosclerosis, such as fatty streaks or intimal thickening, in the arteries of children.¹ Currently, the guidelines for primary prevention in children and adolescents published by the American College of Cardiology and the American Heart Society include the study of the traditional risk factors, based on the idea that these continue into adulthood and are related to atherosclerotic changes during this stage of life.²

It is now known that atherosclerosis is an inflammatory disease.³ C-reactive protein (CRP), an acute phase protein, is a highly sensitive marker of general inflammation. Experimental studies have detected the presence of CRP in arteries with atherosclerotic lesions. It has also been shown that CRP induces the production of inflammatory cells and that it reduces the expression of nitric oxide synthetase.⁴ Thus, from a biological point of view, CRP participates in the atherogenic process. In adults, CRP detected with ultrasensitive techniques (ultrasensitive CRP or CRPus) is associated with traditional risk factors and its concentration predicts cardiovascular events. Few studies have been performed, however, on the concentration of CRPus and its relationship with classical risk factors, adiposity variables, endothelial function, and carotid thickness in children. From a biological point of view, the advantage of studying children is that established coronary artery disease, a confounding factor of great importance in adults can be ruled out.

The aim of the present study was to determine the CRPus concentration of healthy children and to seek possible correlations with anthropometric variables such as adiposity, cardiovascular risk factors, and endothelial function (estimated by flow-mediated dilation of the brachial artery [FMD] and the intima-media thickness [IMT] of the carotid artery).

METHODS

The study subjects were the children or grandchildren of adults who took party in an epidemiological study performed by the Departamento de Salud Pública (Dept. of Public Health) of our university in 2004. These adult subjects belonged to nuclear families residing in different neighborhoods of Santiago; all answered a questionnaire regarding their cardiovascular risk factors, and all underwent tests to determine their lipid profile, blood sugar level, blood pressure, and CRPus concentration. The different neighborhoods were randomly selected by multistage sampling and their inhabitants stratified by age, gender, and socioeconomic level. The children and grandchildren of the healthy adults in this earlier study (ie, those with no cardiovascular risk factors nor atherosclerotic disease, and who had a normal CRPus concentration) were invited to take part in the present study, which was undertaken between October 2005 and December 2006. Thus, the same type of sampling method was used to obtain a representative sample of the children from these neighborhoods. All those enrolled were aged between 6 and 13 years. The adults legally in charge of these children all gave their signed, informed consent for their charges to take part. The study was approved by the University's ethics committee.

A nurse contacted the selected children by telephone to invite them to take part and to arm their parents or grandparents with some general recommendations. These included: a) that the children should always come to the hospital accompanied by an adult, b) that they should have fasted for at least 12 h before tests, c) that a child not attend an appointment if he/she had been ill between 1 day and 2 weeks prior to the appointment date, and d) that no medications be taken for at least 12 h before attending an appointment.

Data Collection

All the children underwent testing between 8.00 h and 10.00 h. All answered questionnaires to record their demographic data, clinical history, physical activity, and nutritional background. Information on the risk factors affecting the parents or grandparents were also recorded. Children with high blood pressure, hypercholesterolemia, or who suffered a metabolic or inflammatory disease were excluded.

All the children were examined physically by a pediatric cardiologist. Two cardiologists and a nurse trained in vascular ultrasonography recorded the subjects' FMD and IMT data. Blood samples were collected after these echographic studies were completed.

Anthropometric Measurements and Pubertal Development

A pediatric nurse recorded the weight and height of the children (barefoot, dressed in their underwear, and covered by a robe) using a lever balance and a SECA^® stadiometer respectively. Height/age (H/A) and weight/height (W/H) ratios were then calculated, along with the body mass index (BMI) (weight in kg/height in meters squared), and expressed as percentiles and z values (z=value recorded­median/1 standard deviation). The NCHS-CDC⁵ curves for the year 2000 were used as references and obesity defined as BMI ≥p95, overweight or risk of obesity as p85 to p94, eutrophy between p10 and p84, and underweight as below p10. The BMI was expressed as a z value to transform it into a continuous variable adjusted for age and sex. Since body composition changes during childhood, the same NCHS-CDC⁵ reference was used.

Central obesity was determined by measuring the waist with a non-stretch plastic tape; the mean of 2 measurements was recorded. As a reference, a percentile distribution for a multiracial child population⁶ was used, with waste circumference values expressed in terms of the percentage of the 50th percentile (PCp50), ie, (true measurement/p50)ƒ100. Percentile distributions for the waist were used given the changes that occur in body shape and composition with age and development.

Finally, the thicknesses of four skin folds were measured the biceps, triceps (TC), subscapular (SS), and suprailiac folds­using Lange^® calipers, along with the brachial circumference. All measurements were made following international norms.⁷ The SS/TC ratio was used as an indirect indicator of trunk fat.⁸ The percentage total fat mass was determined using the formulae of Slaughter⁹ (which require TC and SS values be known).

The sexual maturity of each subject was determined using the Tanner index.⁶

Blood Pressure

Blood pressure was measured with subjects in the supine position following a 10 min rest period in a temperature-controlled room. All measurements were taken on the right arm using a Dynamap Pro 100 (Criticon^®) device with an adult and pediatric cuff. Three measurements were made at intervals of 5 min and the mean calculated. High blood pressure was defined as a systolic or diastolic pressure ≥p95 for the subject's sex, age, and height, according to international standards (NIH, 1996).

Blood Tests

All blood samples were obtained by venipuncture after a 12 h fast. Ultrasensitive C-reactive protein was determined by nephelometry using a Dade Behring BN II nephelometer (detection limit 0.1 mg/L). Total cholesterol, high density lipoprotein cholesterol (HDL-C) and triglycerides were determined by standard enzymatic methods using a Yaci analyzer. Low density lipoprotein cholesterol (LDL-C) was determined using the Friedewald formula, and blood sugar by the glucose oxidase method.

Assessment of Subclinical Atherosclerosis

For the measurement of IMT and FMD a Hewlett Packard Sonos 5500 apparatus with a high frequency (5-13 MHz) linear transducer was used. The operator remained blinded to the blood test results during these determinations.

Carotid Intima-Media Thickness

This was measured according to the 2004 Mannheim consensus.¹⁰ The image was centered on the posterior wall of each common carotid artery, and a 1 cm segment selected near the bifurcation. Measurements were made using software that automatically detected the carotid borders (M'Ath^® Std). All measurements were made off-line at the end of diastole. The highest IMT value measured on either the right or left side was used in statistical analyses. The intraoperator coefficient of variation was 3.8%.

Endothelial Function

This was determined according to the 2002 recommendations of the International Brachial Artery Reactivity Task Force.¹¹ The child assumed the supine position and a blood pressure cuff was placed on the non-dominant forearm. The brachial artery was explored at the elbow fold, and its baseline diameter recorded. The cuff was then inflated to a pressure at least 50 mm Hg above that of the child's resting (5 min) systolic blood pressure (hyperemic phase); it was then deflated and the posthyperemic response measured. Both diameter measurements (baseline and posthyperemic) were made during diastole over 3 cardiac cycles, involved intima-intima measurements, and were taken using an electronic caliper. The FMD value was calculated as the percentage change in the posthyperemic diameter compared to baseline. Nitric oxide-independent dilation with nitroglycerine was not performed given the age of the children and the danger of inducing low blood pressure.

Statistical Analysis

The results are expressed as means (standard deviation). The distribution of the CRPus values was markedly symmetrical; this variable was therefore presented in terms of medians and interquartile ranges. In all analyses involving the CRPus values in continuous form, log(CRPus) values were used to correct the asymmetry and stabilize the variance. The associations between the log(CRPus) and all other variables were determined by Pearson correlation coefficient analysis, using multiple linear regression models adjusted for age and sex. Logistic regression models adjusted for age and sex were used to calculate odds ratios (OR), using a CRPus concentration of >0.43 mg/L (the sample median) for all analyses. A P value less than .05 was considered significant.

RESULTS

The study sample (n=112) was composed of 58 boys and 54 girls; 30% of the boys and 55% of the girls were prepubertal (P<.01, Fisher exact test; such a difference might be expected in a representative population of children of this age). According to their BMI, 65% of the children were eutrophic, 22% were overweight, 10% obese, and 3% underweight; this is comparable to the distribution for the Chilean school-age population.⁸ Only 1 girl had a high CRPus concentration (14.7 mg/L), a sign of an infection; she was excluded from the analysis. In 9 children the images of the brachial artery obtained in FMD analysis were unsatisfactory and excluded from analysis.

Table 1 shows the demographic, anthropometric, and biochemical results for the children studied; the mean age of the subjects was 11.3 (1.9) years. No differences were seen between the boys and girls in terms of demographic variables. The SE/TC ratio of the girls was significantly higher (P<.01). The median CRPus value was 0.43 mg/L (interquartile interval, 0.69 mg/L). No significant differences were seen between the CRPus of boys and girls. Table 2 shows the Pearson correlation coefficients between CRPus concentration and the anthropometric, biochemical, and subclinical atherosclerotic variables for the sample as a whole. A separate analysis of the boys and girls revealed a positive correlation between CRPus concentration and BMI, total fat mass, waist circumference, the SE/TC ratio, and LDL-C level in the boys, both pubertal and prepubertal. In the girls, a significant relationship was found between CRPus concentration and BMI, zBMI, total fat mass, and PCp50­always in those who were pubertal. No association was found between CRPus and age, total cholesterol, HDL-C, or triglyceride levels, blood pressure, blood sugar, IMT, or FMD, either in boys or girls.

Log(CRPus) was significantly associated with BMI, zBMI, total fat mass, waist circumference, and blood sugar, after adjusting for age and sex. No significant associations were seen with cholesterol or triglyceride levels, blood pressure, IMT or FMD. Table 3 shows the values of these variables with respect to the CRPus concentration divided into tertiles.

The risk factors associated with an elevated CRPus concentration (CRPus >0.43 mg/L, the sample median) were identified by logistic regression adjusted for age and sex. Significant associations were found with BMI (OR=1.28; 1.1-1.5), zBMI (OR=2.04; 1.3-3.2), total fat mass (OR=1.1; 1.04-1.15), and waste circumference (OR=1.09; 1.03-1.14) (Table 4).

DISCUSSION

The results indicate that the CRPus concentration is directly and significantly associated with anthropometric indicators of adiposity, but not with subclinical atherosclerosis indicators in healthy children.

Knowledge of CRPus levels in children is being seen as increasingly important, although only recently have studies on the relationship of this marker with cardiovascular disease begun in children and adolescents. The CRPus concentrations of the children in the present study were significantly lower than those seen in adults.¹² In healthy children the median CRPus concentration ranges from 0.2 to 0.5 mg/L.¹² The present results are comparable to those of Ford et al¹³ who examined 3348 children in the NHANES study in 2003; the median CRPus reported was 0.4 mg/L with no significant difference between boys and girls.

From a clinical point of view the importance of clarifying the pathogenic role of CRPus in children lies in its relationship with atherosclerotic lesions established in children before the age of 15, independent of other risk factors.¹⁴ In adults it is accepted that CRPus has a pathogenic role in the formation of atherosclerotic plaques: among other properties it seems able to activate complement and endothelial cells,^15,16 alter the production of nitric oxide,¹⁷ and mediate the production of tissue fluid.¹⁸ Currently, the American College of Cardiology and the American Cardiology Society recommend the measurement of CRPus in adults to improve the prediction of cardiovascular risk in persons with an intermediate risk (10%-20%) according to the Framingham scale.¹⁹

The most important determinants in the CRPus values of the present subjects were BMI, zBMI, and other clinical indicators of adiposity such as the percentage total fat mass, the waist circumference, and the SS/TC ratio (an indicator of trunk obesity).

Few studies have investigated the relationship between the CRPus concentration and the traditional risk factors and anthropometric variables in "healthy" children. However, the present results agree with those of these studies.^20-22 Cook et al²⁰ reported the ponderal index to be the major factor determining the CRPus concentration, both in girls and boys, while Ford²¹ found this to be the BMI. In the present study, the magnitude of the correlation coefficients between the CRPus concentration and the adiposity variables were similar to those published by these other authors. The fact that this magnitude is low is explained in that healthy children have fewer risk factors and low CRPus concentrations.

The supposed relationship between the CRPus concentration and adiposity is based on numerous studies in adults that have shown obese and overweight individuals acquire a proinflammatory status.²³ It is currently thought that their higher CRPus concentrations are due to the increased expression of interkeulin-6 or tissue necrosis factor in the intra-abdominal tissue. This would explain the close, significant relationship between BMI or zBMI and CRPus, both in adults²³ and in overweight and obese children.^24-26 It has recently been reported that CRPus is also a marker of insulin resistance and diabetes in adults.^27,28 Finally, obesity is associated with high concentrations of leptin and low concentrations of adiponectin (an anti-inflammatory and anti-atherogenic protein),²⁹ both of which are produced exclusively by adipose tissue. Thus, a state of chronic inflammation (high CRPus concentrations) along with low adiponectin concentrations and a greater prevalence of metabolic risk factors could favor the development of atherosclerosis in obese adults and children.

Unlike in adults, no relationship has been consistently shown between the CRPus concentration and dyslipidemia in children. In the present sample, the CRPus concentration was only significantly associated with the LDL-C level in boys; no such relationship was seen with other plasma lipids. This finding is interesting from a pathophysiological point of view since it has been shown that CRPus mediates the capture of LDL-C by macrophages in atherosclerotic plaques, and is thus involved in the development of atherosclerosis.¹⁷ Unlike that which occurs in adults, the lack of correlation between CRPus and other risk factors (eg, blood pressure or blood sugar) seen in the present study shows that this association may begin to appear later in life or require more inflammation for it to become manifest.

In the Chilean setting, evidence exists that metabolic complications accompany childhood obesity­associations once believed to exist only in adults.^8,30 Our concern for studying these variables in children came about because of these reports and because some 38% of the Chilean school-age population is overweight. As reported by Freedman et al¹⁴ in follow-up investigations associated with the Bogalusa Study, there is a high probability that these children will continue to be obese and develop cardiovascular risk factors as they approach adulthood.^31,32 Juonala et al³³ have recently reported that, in addition to traditional risk factors, childhood CRPus concentrations tend to be retained into adult life. In a recent study on 51 healthy Spanish children and adolescents, all the offspring of at least 1 hypertensive parent, it was reported that all their CRPus values were high,³⁴ perhaps indicating a genetic effect. As in the present study, a significant relationship was also found between obesity markers and the CRPus concentration. If it is confirmed that subclinical inflammation is a predictor of cardiac events, as it is for diabetes mellitus type II, knowledge of the behavior of this marker in childhood would gain in clinical importance.

Finally, no association was found between the CRPus concentration and the subclinical atherosclerotic variables studied. In contrast, Järvisalo et al,²² who studied 79 Finnish children, showed an elevated CRPus concentration to be associated with abnormal endothelial functioning and carotid thickening. It is hard to offer an explanation for the differences between this and the present study since the values for CRPus, FMD, and other risk factors were very similar. One explanation might be the existence of protective genetic factors in our population. Certainly, the scarcity of publications regarding the association between the CRPus concentration and vascular anatomy and function in healthy children shows there is still much work to be done. What has been shown is that, compared to eutrophic children, children with morbid obesity show a relationship between the CRPus concentration and changes in the dilation of the brachial artery and the IMT. However, the information currently available is insufficient to be able to say for certain whether these arterial abnormalities in obese children are caused by the inflammation present alone, or whether they are the consequence of a greater risk factor load with the degree of inflammation being a mere epiphenomenon.

The present results should stimulate the development of new programs focused on the prevention of and early intervention in childhood obesity. Physical activity could be the "ideal medication" - one that should be used from infancy to old age.³⁵

Study Limitations

The present study suffers the following limitations: a) small sample size; b) the fact that this was a cross sectional study makes it impossible to establish cause-effect relationships; prospective studies are therefore essential; and c) given the small sample size, the lack of correlation between age and the CRPus concentration may be subject to some error.

CONCLUSIONS

This study provides evidence of a close relationship between the CRPus concentration (a marker of subclinical inflammation) and clinical markers of obesity. Only the follow-up of these children can clarify whether the CRPus concentration is the cause or consequence of the atherogenic process, and whether damage is reversible. The relationship between this marker and adipose variables in children shows it is necessary to investigate its relationship with metabolic variables that might increase the risk of diabetes or other chronic diseases of adult life.

ABBREVIATIONS
BMI: body mass index
CRPus: ultrasensitive C-reactive protein
FMD: flow-mediated dilation (of the brachial artery)
IMT: intima-media thickness (of the carotid artery)
PCp50: percentage of the 50th percentile (for waste circumference)

---

Correspondence: Dr. M. Acevedo.
Departamento de Enfermedades Cardiovasculares.
Pontificia Universidad Católica de Chile.
Lira 85, 1.^er piso. Santiago. Chile.
E-mail: [e-mail]

Received January 2, 2007.
Accepted for publication June 22, 2007.