Carotid Artery Intimal-Medial Thickness and Left Ventricular Hypertrophy in Children With Elevated Blood Pressure

METHODS

Study subjects (n = 32) were recruited from a subspecialty pediatric hypertension clinic. All subjects were newly referred to the clinic after documentation of elevated blood pressure by a primary care provider on several preceding occasions. Elevated blood pressure was confirmed in the clinic by averaging the last 3 of 4 blood pressure measurements made at the first clinic visit with a Critikon oscillometric monitor (Tampa, FL) after 5 minutes of rest. Inclusion criteria for the study were average clinic blood pressure greater than the gender, age, and height-specific 90th percentile by current Task Force criteria²⁶; age ≥6 years and ≤18 years; no past or current use of antihypertensive medication; no concurrent use of medications known to affect blood pressure; and no known secondary causes of hypertension. The protocol was approved by an institutional review committee, and subjects and parents gave informed assent and consent, respectively.

Echocardiography was performed by sonographers who were blinded to carotid ultrasound results. A complete 2-dimensional echocardiographic examination, consisting of parasternal long and short, apical 4-chamber, subcostal coronal and sagittal, and suprasternal long and short axes, was first obtained to rule out congenital heart disease, with particular attention to ensure that the aortic arch was patent and that there was no coarctation of the aorta present. LVM was calculated from 2-dimensionally guided m-mode measurements made during diastole of the left ventricular internal dimension, interventricular septal thickness, and posterior wall thickness according to methods established by the American Society of Echocardiography. LVM was calculated using the equation reported by Devereux et al.²⁷ LVMI was calculated by dividing LVM by height in meters to the power of 2.7 to minimize the effect of age, gender, ethnicity, and overweight status.^28,29 LVH was defined as LVMI >38.6 g/m^2.7, a value reported to represent the pediatric 95th percentile of LVMI in normotensive healthy children.²⁸

Carotid artery ultrasound was performed by experienced vascular sonographers who were blinded to echocardiography results according to previously developed and validated scanning protocol. Briefly, subjects were examined in a supine position. A longitudinal view of the distal common carotid artery (CCA) was obtained using a linear 8-MHz transducer. The gain and focus settings were optimized to contrast the vessel lumen and IMT appearance. The IMT measurements were made on both CCAs on the far wall of the distal CCA at least 2 cm below the flow divider. A longitudinal B-mode image of the CCA with sharp edges of the far wall IMT complex was used to place 2 measurements with digital calipers at 1 cm apart. The measurements were carried on frozen images demonstrating the thickest IMT complex with calipers placed on a zoomed CCA image. The reproducibility of measurements of maximum IMT thickness in the distal CCA was determined for repeat assessments by the same sonographer and by 2 sonographers independently in a sample of 10 healthy volunteers yielding κ values of >0.9. In our validation studies, the difference between these measurements was <10% (average 5%).

Descriptive statistics are presented as percentages, means, and standard deviations. Univariate analyses for group comparisons of continuous variables were performed using Student t test. Multivariate analyses for group comparisons were performed using analysis of covariance. The correlations among LVMI; cIMT; and continuous demographic, clinical, and hemodynamic variables were determined using the Pearson correlation coefficient. Multiple regression analysis was used to determine the strength of association among LVMI, cIMT, and multiple independent variables. Fisher exact test was used to compare the LVH percentage between groups. P < .05 indicated statistical significance.

RESULTS

Demographic and clinical data for study subjects are shown (Table 1). The majority of subjects in the study were male (78%) with Hispanic the predominant ethnicity (47%). Fifty-three percent of subjects were obese, defined as body mass index (BMI) ≥95th percentile for age and gender, and 78% had BMI ≥90th percentile. Fifty-three percent of subjects had isolated systolic blood pressure elevation, 3% had isolated diastolic blood pressure elevation, and 44% had both systolic and diastolic blood pressure elevation. The mean time between echocardiography and carotid ultrasound was 5.9 ± 10.6 days. Twenty-three of 32 subjects had echocardiography and carotid ultrasound performed on the same day, with a maximum time between studies of 34 days.

The prevalence of LVH was 41% (13 of 32). By definition, subjects with LVH had higher LVMI than those without LVH (45.9 g/m^2.7 vs 28.8 g/m^2.7; P < .00001). Subjects with LVH had a higher BMI than those without LVH (31.2 vs 26.2; P = .037). Subjects with and without LVH did not differ in age, weight, height, clinic blood pressure values, or category of blood pressure elevation (systolic blood pressure only, diastolic blood pressure only, or both). LVMI was positively correlated with weight (r = 0.42; P = .02) and BMI (r = 0.49; P = .005) but did not vary by age, gender, ethnicity, height, or blood pressure.

cIMT was positively correlated with weight (r = 0.50; P = .003) and BMI (r = 0.43; P = .014) but did not vary by gender, ethnicity, age, height, or clinic blood pressure values. The upper quartile of maximum cIMT within the study population was bounded by a threshold of 0.8 mm. The prevalence of subjects with increased cIMT (ie, upper quartile) was 28% (9 of 32). Subjects with increased cIMT were heavier (91.3 kg vs 68.9 kg; P = .002) and had higher BMI (33.7 g/m^2.7 vs 26.1 g/m^2.7; P = .003) compared with those with cIMT in the lower 3 quartiles but did not differ in age, height, or blood pressure.

The relationship between carotid ultrasound and echocardiography parameters was determined by univariate and multivariate analysis. Subjects with increased cIMT had higher LVMI (46.8 g/m^2.7 vs 31.4 g/m^2.7; P < .0001; Fig 1). Similarly, subjects with LVH had higher cIMT (0.72 mm vs 0.63 mm; P = .047) than those without LVH. By univariate analysis, cIMT was positively correlated with LVMI (r = 0.54; P = .001; Fig 2), interventricular septal thickness (r = 0.58; P < .001), and posterior wall thickness (r = 0.54; P = .001). By multivariate analysis, the relationship between cIMT and LVMI remained significant after controlling for either BMI or weight (r = 0.42; P = .018). Stepwise linear regression analysis showed that 33% of the variability of cIMT was accounted for by LVMI and weight (r = 0.62; r² = 0.39; P < .001).

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

Comparison of LVMI between subjects with normal versus thickened IMT. Nml, carotid IMT <0.8 mm; Thick, carotid IMT ≥0.8 mm.

• Download figure
• Open in new tab
• Download powerpoint

Fig 2.

Scatterplot of cIMT versus LVMI.

The prevalence of LVH was 89% (8 of 9) among subjects with increased cIMT as compared with 22% (5 of 23) in subjects with normal cIMT (P = .0009, Fisher exact test, 2-tailed). The sensitivity and specificity for a cIMT ≥0.8 mm to be associated with LVH were 62% and 95%, respectively.

DISCUSSION

The results from the current study demonstrate for the first time in children with elevated blood pressure that early arterial wall changes related to atherosclerosis and increases in LVM occur in concert. This relationship is evidenced both by the positive correlation between LVMI and cIMT and the association between LVH and increased cIMT. Specifically, a compelling finding of the study is the high percentage of LVH in children with increased cIMT. Eight of 9 children in the upper quartile of cIMT had LVH, as compared with only 5 of 23 in the lower 3 quartiles. These findings are consistent with studies of hypertensive adults showing an association between increased cIMT and cardiovascular alterations such as LVH.^30–32 Although the conclusions that can be drawn are limited by the relatively small number of subjects studied and the lack of a normotensive control group, these results raise the possibility that carotid ultrasound, by indicating the presence of early arterial wall changes, may be useful for predicting other cardiovascular sequelae in hypertensive children.

In contrast to adults, in whom increased cIMT may be confounded by factors such as advancing age and body habitus, cIMT is reportedly independent of age, gender, BMI, fat mass, or serum cholesterol in healthy children through 18 years of age.²⁵ However, in the current study, both cIMT and LVMI were positively associated with BMI. Although an independent relationship between cIMT and LVMI was found after controlling for body habitus, other atherogenic factors that were not assessed in this study, such as glucose intolerance and dyslipidemia, and that are often associated with obesity (present in more than half of study subjects) may be important modifiers of the association between cIMT and LVMI. It is interesting to note that neither cIMT nor LVMI correlated with clinic blood pressure. This lack of correlation suggests that the severity of clinic blood pressure elevation may be unreliable for quantifying the risk of target organ sequelae in children. However, the small sample size, with its inherent lack of power to detect relatively weaker associations, creates uncertainty regarding the interpretation of negative results.

The presence of preclinical atherosclerosis in children with cardiovascular risk factors has been documented in previous studies.³³ Autopsy studies have revealed the presence of fatty streaking and plaque formation in the abdominal aorta of children and young adults with cardiovascular risk factors before the onset of clinical disease.^34–37 Studies using carotid ultrasound to measure cIMT in children at high risk for atherosclerotic disease are consistent with these autopsy findings. Increased cIMT has been reported in children with diabetes ^17–19 and in children with familial hypercholesterolemia^17,20,21 as compared with normal controls. Furthermore, prospective community-based studies have shown that total serum cholesterol measured in childhood predicts cIMT measured in the same subjects in adulthood.³⁸ In aggregate, these studies provide strong evidence that the atherosclerotic process begins in children who are at risk many years before disease becomes clinically apparent.

Although the results from the current study are consistent with previous studies showing increased cIMT in children with atherogenic risk factors, the ultrasound method for IMT does not allow for differentiation of intimal thickening because of the atherosclerotic process from medial hypertrophy (smooth muscle growth) as a result of pressure effects. It has been argued that increased cIMT in younger patients does not represent local atherosclerosis but instead reflects an adaptive response to altered flow, shear stress, and pressure.³⁹ The positive relationship between IMT and LVM found in the current study is consistent with the latter interpretation by indicating a generalized augmentation of cardiovascular growth inducing both increased LVM and medial thickening. However, an autopsy study by the Pathobiological Determinants of Atherosclerosis in Youth Study reported that intimal lesions appeared in all of the aortas and more than half of the right coronary arteries in patients aged 15 to 19 years.⁴⁰ Given that these subjects who underwent autopsy were unselected for cardiovascular risk factors, it seems likely that the children in the current study with hypertension and elevated BMI would have the same or a greater degree of intimal changes. In either case, the risk of future cardiovascular and cerebrovascular disease in adults increases gradually with increasing common cIMT.⁴¹ Thus, increased cIMT may serve as a graded marker for cardiovascular risk by showing the same type of adaptive response to chronic blood pressure elevation that is presumed to cause increased LVMI.

Although both LVH and increased cIMT are independently predictive of cardiovascular morbidity and mortality in adults, the threshold values of “abnormality” for LVMI and cIMT are not as well validated in children. In the current study, LVM was calculated by the formula of Devereux et al²⁷ and indexed to height to the power of 2.7 to minimize the effect of age, gender, and overweight status.^28,29 LVH defined as LVMI greater than the pediatric 95th percentile was found in 42% of subjects, consistent with previous studies of hypertensive children using a similar LVH definition.^12,42 In hypertensive adults, LVH defined as LVMI >51 g/m^2.7 is associated with a 4-fold higher risk for the development of cardiovascular endpoints.⁴³ In contrast, the pediatric definition of LVH used in this study has not been validated by linkage to morbid outcomes.

The definition of abnormal cIMT is less clear, particularly in children. A study of hypertensive adults found that the prevalence of abnormally thickened cIMT increased from 15% to 44% as the definition threshold decreased from 1.0 mm to 0.8 mm.³² When linked to clinical events, specific studies of adults have reported that cIMT ≥1 mm is associated with a 2- to 5-fold greater risk of coronary events,^44–46 and cIMT ≥1.18 is associated with a 4-fold greater risk for combined acute myocardial infarction and stroke.⁴⁷ A meta-analysis of studies in adults relating cIMT measurements to the risk of coronary and cerebrovascular events showed that the risk of first myocardial infarction increases with an IMT of 0.822 mm or more and the risk of stroke with an IMT of 0.75 mm or more.⁴⁸

The upper quartile of cIMT in the current study population (cIMT ≥0.8 mm) is within the range of cIMT associated with the increased risk defined in the meta-analysis of adult studies and higher than the normal values described in the largest study of carotid ultrasound in children.²⁵ However, differences in measurement technique make comparisons between studies difficult. Reported values for cIMT in normal pediatric patients vary from 0.42 mm¹⁹ to 0.64 mm,²² with several studies reporting values intermediate to these extremes.^{18,20,23–25,49} In the absence of a concomitant control group, it cannot be definitively concluded that cIMT is increased in hypertensive as compared with normotensive children. Although the threshold for abnormal cIMT in children is likely lower than the value of 0.8 mm used in the current study, 0.8 mm bounded the upper quartile of values in the current study and therefore defined the high end of the range of values within this selected patient population. The method in the current study of choosing the maximum cIMT value likely overestimates IMT thickness compared with an average of multiple measurements used in previous large-scale studies in adults along similar CCA segments by 0.01 to 0.02 mm. However, the internally consistent cross-sectional observations of 1 maximum cIMT measurement in the current study suggest that early arterial wall changes are present in hypertensive children and deserve additional study.

CONCLUSION

The current study is consistent with the results from other studies of cardiovascular risk in children that arterial wall changes may be present at an early age. The study findings are strengthened by the prospective, protocolized enrollment of subjects; the independent measurements of blood pressure, LVMI, and cIMT by separate study personnel; and the persistence of the association between LVMI and cIMT after controlling for age, gender, or body habitus. In the face of an evolving epidemic in children of hypertension, obesity, type 2 diabetes, and dyslipidemia,⁵⁰ more sensitive and specific diagnostic studies to define present and future cardiovascular risk are needed. In this regard, additional studies of carotid ultrasound in hypertensive children will help determine the extent to which increased cIMT may be used as a marker for this risk.