Metabolic Abnormalities, Insulin Resistance, and Metabolic Syndrome in Children With Primary Hypertension:

Abstract

Background:

We sought to describe the prevalence of metabolic abnormalities and of metabolic syndrome (MS) and its relationship to target-organ damage in children with primary hypertension (PH).

Methods:

Patients included 113 children with untreated PH at a mean age of 14.6 years (range, 5 to 18 years). The control group consisted of 134 healthy children at a mean age of 13.5 years (range, 5 to 20 years). We performed a cross-sectional assessment of anthropometric and biochemical cardiovascular risk factors, homeostatic metabolic assessment (HOMA-IR), the insulin sensitivity index (ISI[0,120]), and adiponectin.

Results:

Metabolic syndrome, as defined by classic criteria, was present in 4 of 134 (3%) of controls versus 23 of 113 (20.4%) patients (P = .0001), but when PH was not taken as a criterion of MS, MS was diagnosed in 6.2% of patients (no significance). Left-ventricular hypertrophy (LVH) was found in 46 of 113 patients (40.7%), and severe LVH was found in 14 of 113 patients (12.5%). Patients with LVH had a greater body mass index, greater waist-to-hip-ratio, and greater number of parameters of metabolic syndrome (overall P < .05). Carotid (cIMT) and femoral superficial artery intima-media thicknesses correlated positively with HOMA-IR and negatively with ISI[0.120] and serum adiponectin (P < .05). The main predictor for cIMT was adiponectin (R² = 0.178, β = −0.466, P = .002). Left-ventricular hypertrophy was predicted (R² = 0.332) by body mass index-standard deviation score (β = 0.551, P = .005) and HOMA-IR (β = 0.380, P = .04).

Conclusions:

Metabolic syndrome, as defined by classic criteria, was diagnosed in 20% of children with PH, but when PH was not a criterion, MS was present in 6.2% of patients. Irrespective of the definition of MS, the applied markers of MS and insulin resistance were the main predictors of target-organ damage. Am J Hypertens 2007;20: 875–882 © 2007 American Journal of Hypertension, Ltd.

Primary hypertension, metabolic syndrome, adiponectin, left-ventricular hypertrophy, intima-media thickness, children

Primary hypertension (PH) is related to several metabolic abnormalities that are connected to overweight, metabolic syndrome (MS), and type 2 diabetes.^1,2 The most typical metabolic abnormalities in PH are low HDL cholesterol and high triglyceride levels. Some authors reported that a tendency toward higher serum concentrations of uric acid (UA) is also typical of children with PH.^3,4

There are only a few reports on the prevalence of metabolic abnormalities and their relationship to target-organ damage (TOD) in children with PH. In the last 2 decades, the prevalence of childhood obesity has increased enormously, and at least 30% of obese children are hypertensive.⁵ Although the prevalence of MS and type 2 diabetes in childhood is also rising, it is nonetheless still significantly lower than in adults, and presents almost exclusively in severely obese children.^6,7 The aim of our study was to describe metabolic abnormalities and their relationship to TOD in children with previously untreated PH.

Methods

Primary hypertension was diagnosed after a thorough clinical and laboratory diagnostic workup, according to recently published recommendations.⁸ Normal blood pressure values were taken from the Updated Task Force Report.⁸ In all hypertensive subjects, the diagnosis was confirmed by 24-h ambulatory blood-pressure monitoring lasting at least 20 h with at least 80% of records. Exclusion criteria were age <5 and >20 years, the presence of any significant chronic disease (except for PH in the hypertensive group), any acute disease including infections in the preceding 6 weeks, and incomplete data.

In all control and hypertensive subjects, anthropometrical measurements were assessed, including body mass index (BMI), waist-to-hip ratio (WHR), waist-to-height ratio (WHTR), blood pressure (BP), carotid and femoral superficial artery intima-media thickness (cIMT and fIMT, respectively), glucose, insulin, blood lipids, homocysteine, UA, C-reactive protein (CRP), serum adiponectin, and leptin levels. All hypertensive subjects underwent echocardiography (ECHO), albumin excretion, and oral glucose tolerance tests (OGTTs). Echocardiography and OGTT were not performed in control subjects.

Blood pressure was measured three times at 5-min intervals on the dominant arm with an automated oscillometric device (Dinamap PRO100V2; GE Medical Systems International Technologies, Inc., Freiburg, Germany) on the day of measurement of intima-media thickness between 8:00 and 10:00 AM in standard sitting position, after a 5-min rest. The average of the two last recordings was used in analyses. Blood pressure values were presented as absolutes and in a BP index (BPi) calculated as the ratio of measured systolic BP (SBP) or diastolic BP (DBP) at the 95th percentile value for age and sex.

cIMT and fIMT Measurements

The cIMT and fIMT were evaluated ultrasonographically by the same examiner (ML), using an ATL 5000 HDI device (Advanced Technology Laboratories, Bothwell, WA) and a 7.5 to 12.5-MHz probe. The examiner was not blinded to the BP status of the children. The measurements were taken from the far wall of the artery in a two-dimensional presentation of the longitudinal view of the vessel. The cIMT was measured 1 to 2 cm below the bifurcation of the common carotid artery (cca) and along a 1-cm distance. Systolic and diastolic diameters of the carotid artery were measured in M-mode presentation. The fIMT was measured on the upper third to upper half of the thigh. For every left and right artery, at least five to six measurements were taken from each arterial scan. The results were averaged for every side and are presented as the mean of 10 to 12 values. The detailed procedure of intima-media thickness measurements was published previously.^4,9–11 The median and standard deviation (SD) of normal values for the cIMT and fIMT were obtained from a study of 250 healthy children.⁹

The following parameters were calculated according to formulas described elsewhere:^10,11

• Mean systolic diameter (sD) = (LsD + RsD)/2, where LsD = left cca systolic diameter, and RsD = right cca systolic diameter.

• Mean diastolic diameter (dD) = (RdD + LdD)/2, where RdD = right cca diastolic diameter, and LdD = left ccs diastolic diameter.

• Mean lumen cross-sectional area of the artery (LCSA) = π (dD)²/4.

• Mean wall cross-sectional area (WCSA) = π (dD/2 + IMT)² − π(dD/2)².

Echocardiography

All ECHO measurements were performed according to American Society of Echocardiography guidelines by an examiner (JD) who knew the clinical diagnosis but was not aware of the severity of PH. Left-ventricular mass (LVM) was calculated according to the formula of de Simone et al.¹² Left-ventricular hypertrophy (LVH) was defined as >38.6 g/m^2.7 as the cutoff for the 95th percentile, and a value ≥51 g/m^2.7 was the cutoff for significant LVH.^12,13

Laboratory Investigations

Blood samples were taken after 12 h of fasting and sent immediately to the laboratory. Oral glucose tolerance tests were performed in all hypertensive subjects after oral ingestion of 1 mg/kg (maximum, 75 mg) of glucose. Plasma glucose levels were measured by a Dimension analyzer (LINCO Research, Billevica, MA). Plasma insulin concentrations were measured by radioimmunoassay, and serum adiponectin and leptin levels by ELISA, using commercially available kits (Diagnostic System Laboratories, Inc., Webster, TX, and LINCO Research, St. Charles, MO, respectively).

Insulin resistance (IR) was expressed as the homeostasis model assessment for insulin resistance (HOMA-IR). In subjects with PH, IR was also calculated as the insulin sensitivity index (ISI[0,120]), according to the formula developed by Gutt et al.¹⁴ The ISI[0,120] is regarded as highly correlated with insulin sensitivity as measured by the euglycemic hyperinsulinemic clamp method. The ISI[0,120] was defined as (m/MPG/logMSI), where:

• m = [ingested glucose in mg + (fasting glucose in mg/dL − 2-h glucose in mg/dL) × 0.19 body weight (kg)]/120 min.

• MPG = mean of fasting and 2-h glucose concentrations (in mg/dL).

• MSI = mean of fasting and 2-h insulin concentrations (mU/ml).

• ISI[0,120] is expressed as mg × L²/mmol × mU × min.

• A low value of ISI[0,120] indicates greater IR.

Plasma homocysteine was measured with a fluorescence polarization immunoassay (IMX; Abbott, Weisbaden, Germany) for the quantitative measurement of total L-homocysteine. Concentrations of CRP were determined using highly sensitive immunoturbidimetry (Orion Diagnostica, Espoo, Finland).

Definition of MS

Metabolic syndrome was diagnosed when at least three criteria were present. The criteria of MS were defined according to Ford at al⁶ and Weiss et al,⁷ and include a BMI ≥95th percentile for age and sex, arterial hypertension, serum triglycerides >110 mg/dL, fasting plasma glucose >110 mg/dL or >140 mg/dL at 2 h of OGTT, and HDL cholesterol <40 mg/dL.

Statistical Analysis

Because the analyzed groups included subjects of different age and sex, BMI and IMT values were expressed both as absolute values and as SD for age and sex. Homogeneity of variance was checked with the Levene test. Variables with a normal distribution were compared by Student's t-test for independent variables. Values with non-normal distribution were compared by the Mann-Whitney U test. Pearson correlation analysis was performed for variables with a normal distribution, and Spearman correlation analysis was performed for non-normal distribution. Standardized LVM and absolute and standardized IMT values were dependent variables. Comparisons between groups were performed separately for boys and girls. Because in correlationwise and stepwise regression analyses the dependent variables were standardized values of LVM and IMT, both sexes were analyzed together. Variables that differed groups or were correlated with dependent variables were taken into stepwise regression analysis, into which sex was introduced as an independent variable. P < .05 was regarded as significant.

This study adhered to the principles of the Declaration of Helsinki and was approved by the local Ethics Committee. All control and hypertensive subjects, and their parents, provided informed consent.

Results

Out of 153 children and adolescents admitted consecutively between 2004 and 2006 because of arterial hypertension and in whom PH was ultimately diagnosed, 113 subjects (29 girls and 84 boys) with a mean age of 14.6 years (range, 5 to 18 years) who completed all investigative procedures were included in the study. There were no differences between the 40 excluded subjects and those who enrolled in the study regarding age, sex, ethnicity, severity of PH, and metabolic abnormalities.

The control group comprised 134 healthy children and adolescents (66 girls and 68 boys) recruited voluntarily from schools, with a mean age of 13.5 years (range, 5 to 20 years); all of them completed the study protocol. All control and hypertensive subjects were of white descent.

Hypertensive subjects had increased body mass and height and, consequently greater BMIs, compared with controls. Body mass indices >95th percentile were found in 15.6% (21 of 134) of normotensive controls and in 57.4% (46 of 113) of patients (P = .0001).

Hypertensive subjects had greater absolute and standardized cIMT and fIMT, lower lecithin:cholesterol transferase activity, higher serum UA, higher plasma insulin concentrations, and higher HOMA-IR compared with normotensive children. Hypertensive subjects tended to have lower high density lipoprotein-cholesterol (HDL-cholesterol) concentrations and apolipoprotein A1 to apolipoprotein B (apoA1:apoB) ratios (Table 1). Metabolic syndrome, defined as the presence of >3 criteria, was found in 4 of 134 (3%) controls versus 23 of 113 (20.4%) hypertensive subjects (P = .0001) (Table 2).

Table 1

Comparison of demographic, anthropometrical, and metabolic parameters in patients and control groups

Variable	Patients (N = 113; 29 girls, 84 boys)	Controls (N = 134; 66 girls, 68 boys)
Age (y)	14.6 (5–18)	13.3 (5–20)
SBP (mm Hg)	132 ± 11*******	112 ± 11
SBP index	1.01 ± 0.07*******	0.89 ± 0.07
DBP (mm Hg)	65 ± 8*******	58 ± 7
DBP index	0.77 ± 0.09*******	0.70 ± 0.09
Pulse pressure (mm Hg)	66 ± 12*******	54 ± 10
BMI (kg/m2)	24.8 ± 4.9*******	19.8 ± 2.0
BMI-SDS	1.8 ± 2.0*******	0.35 ± 1.4
Waist-to-hip ratio	0.83 ± 0.05*	0.86 ± 0.06
Waist-to-height ratio	0.42 ± 0.04^	0.45 ± 0.05
cIMT (mm)	0.45 ± 0.07‡	0.42 ± 0.04
cIMT-SDS	1.8 ± 1.6‡	1.1 ± 1.1
WCSA (mm2)	7.8 ± 1.6¶	7.0 ± 1.1
WCSA-SDS	1.6 ± 2.0§	0.65 ± 1.9
fIMT (mm)	0.35 ± 0.06§	0.32 ± 0.03
fIMT-SDS	0.56 ± 2.0†	−0.03 ± 1.0
LVMi (g/m of height2.7)	38.3 ± 10.2	n.a.
Cholesterol (mg/dL)	174 ± 35	176 ± 35
Triglycerides (mg/dL)	90 ± 52	86 ± 65
HDL cholesterol (mg/dL)	45.8 ± 11.0^	48.6 ± 12.0
LDL cholesterol (mg/dL)	114.4 ± 34.6	110.9 ± 31.5
ApoA1 (mg/dL)	1.27 ± 0.31	1.33 ± 0.31
ApoB (mg/dL)	0.89 ± 0.27	0.85 ± 0.26
ApoA1/ApoB	1.54 ± 0.6^^	1.69 ± 0.6
LCAT (IU/L)	133.6 ± 8.7#	143.9 ± 26.9
Uric acid (mg/dL)	5.5 ± 1.7*******	3.7 ± 1.4
hsCRP (mg/L)	1.6 ± 1.3	1.4 ± 1.3
homocysteine (μmol/L)	9.6 ± 3.4	9.2 ± 3.4
Glucose (mg/dL)	87 ± 8.2##	90 ± 7
Insulin (mU/L)	14.6 ± 11.9*******	8.6 ± 2.0
HOMA-IR	3.0 ± 2.4*******	1.8 ± 0.6
Insulin sensitivity index [0,120]	5.6 ± 6.2	n.a.

ApoA1 = apoprotein A1; ApoB = apoprotein B; BMI = body mass index; BMI-SDS = body mass index standard deviation score; cIMT = carotid artery intima-media thickness; cIMT-SDS = carotid artery intima-media thickness standard deviation score; DBP = diastolic blood pressure; fIMT = femoral artery intima-media thickness; fIMT-SDS = femoral artery intima-media thickness standard deviation score; HOMA-IR = homeostatic model assessment of insulin resistance; hsCRP = C-reactive protein; LCAT = lecithin:cholesterol acyltransferase; Lp(a) = lipoprotein a; n.a. = not available; SBP = systolic blood pressure; WCSA = wall cross-sectional area; WCSA-SDS = wall cross-sectional area standard deviation score.

*

P = .05;

†

P = .04;

‡

P = .006;

§

P = .004;

¶

P = .001;

*******

P = .0001;

#

P = .04;

##

P = .02;

^

P = .07;

^^

P = .06.

Table 2

Prevalence of metabolic syndrome defined by classic criteria (χ² = 170.4, P = 0.0001) and only by anthropometrical and biochemical criteria (χ² = 25.372, P = 0.0001)

No. of MS criteria	Controls (N = 134)	Patients (N = 113)	No. of MS criteria (only anthropometrical and biochemical)	Controls (N = 134)	Patients (N = 113)
0	97 (72.4%)	0 (0%)	0	97 (72.4%)	46 (40.7%)
1 criterion (only)	23 (17.2%)	47 (41.6%)		23 (17.2%)	44 (38.9%)
2 criteria	10 (7.5%)	43 (31.1%)		10 (7.5%)	16 (14.2%)
3 criteria	4 (3.0%)	16 (14.2%)		4 (3.0%)	7 (6.2%)
4 criteria	0 (0%)	7 (6.1%)		0	0

MS = metabolic syndrome.

After dividing the control and hypertensive group into subgroups according to BMI above or below the 95th percentile, it was found that irrespective of BMI, hypertensive subjects had higher values for UA, insulinemia, and HOMA-IR compared with controls (Table 3). Serum UA correlated with central adiposity, expressed as WHTR (r = 0.412, P = .003) and BMI-SDS (r = 0.343, P = .001).

Table 3

Comparison of intima-media thickness values and biochemical variables between controls and patients in relation to BMI below and above two standard deviations

Variable	Controls, BMI <95th percentile (1) (n = 122)	Controls, BMI >95th percentile (2) (n = 12)	Patients, BMI <95th percentile (3) (n = 67)	Patients, BMI >95th percentile (4) (n = 46)
cIMT (mm)	0.42 ± 0.04	0.42 ± 0.03	0.45† ± 0.04	0.45* ± 0.08
cIMT-SDS	0.95*† ± 1.1	1.0^ ± 1.1	1.5† ± 1.4	2.0* ± 1.7
WCSA (mm2)	6.9*† ± 1.1	7.4 ± 1.2	7.6*† ± 1.2	8.2* ± 1.4
WCSA-SDS	0.46* ± 2.0	1.0 ± 1.3	1.1 ± 1.3	2.0* ± 2.3
fIMT (mm)	0.32*† ± 0.05	0.35 ± 0.03	0.35† ± 0.05	0.35* ± 0.05
fIMT-SDS	−0.06 ± 1.4	0.39 ± 0.9	0.40 ± 1.8	0.57 ± 2.1
Cholesterol (mg/dL)	178 ± 36	164 ± 27	169 ± 30	184 ± 3
Triglycerides (mg/dL)	76† ± 38	105 ± 95	79* ± 52	103 ± 52
HDL cholesterol (mg/dL)	50‡* ± 12	42‡ ± 13	49# ± 13	43*# ± 9
LDL cholesterol (mg/dL)	113 ± 33	101 ± 21	103# ± 25	121# ± 41
ApoA1 (g/L)	1.38*‡ ± 0.26	1.17‡ ± 0.40	1.31 ± 0.34	1.23* ± 0.29
ApoB (g/L)	0.86 ± 0.28	0.81 ± 0.16	0.87 ± 0.34	0.91 ± 0.24
ApoA1/ApoB	1.74* ± 0.58	1.50 ± 0.56	1.70 ± 0.75	1.43* ± 0.43
Lp(a) (mg/dL)	15.6 ± 17.5	20.8 ± 31.8	19.7 ± 18.8	19.5 ± 17.9
LCAT (IU/l)	147.2† ± 25.2	126.9 ± 28.5	128.2† ± 31.5	142.5 ± 31.5
Homocysteine (μmol/L)	9.0 ± 2.9	10.3 ± 4.8	9.8 ± 3.1	9.6 ± 3.3
HsCRP (mg/L)	1.3‡ ± 1.1	2.3‡¶ ± 1.7	1.3¶ ± 1.2	1.8 ± 1.3
Uric acid (mg/dL)	3.7*† ± 1.3	4.5§ ± 1.6	5.2† ± 1.0	5.7*§ ± 1.2
Glucose (mg/dL)	89 ± 6	91 ± 5	86 ± 7	88 ± 7
Insulin (mU/L)	9.0* ± 2.9	10.3§ ± 2.9	12.0# ± 4.4	18.6*# ± 16.1
HOMA-IR	1.7* ± 0.6	2.1§ ± 0.6	2.2# ± 0.9	3.6*§# ± 3.0
Adiponectin (μg/mL)	10.7 ± 3.7	7.8 ± 5.0	9.7 ± 3.7	8.9 ± 4.1
Leptin (ng/mL)	9.8* ± 9.1	13.1 ± 24.3	17.1 ± 16.9	29.4* ± 27.7

ApoA1 = apoprotein A1; ApoB = apoprotein B; cIMT = carotid intima-media thickness; fIMT = femoral artery intima-media thickness; HOMA-IR = homeostatic model assessment of insulin resistance; hsCRP = C-reactive protein; LCAT = lecithin:cholesterol acyltransferase; Lp(a) = lipoprotein a; WCSA = wall cross-sectional area.

*

1 v 4: P < .05;

†

1 v 3: P < .05;

‡

1 v 2: P < .05;

§

2 v 4: P < 0 .05;

¶

2 v 3: P < .05;

#

3 v 4: P < .05.

Left-ventricular hypertrophy was found in 46 of 113 patients (40.7%), and severe LVH (LVM >51 g/m^2.7) was found in 14 of 113 hypertensive subjects (12.5%). Hypertensive subjects with LVH had a greater BMI-SDS (2.6 ± 2.3 v 1.2 ± 1.5, P = .0001), birth weight (3494 ± 564 g v 3200 ± 627 g, P = .01), and WHR (0.88 ± 0.06 v 0.84 ± 0.06, P = .03) and were exposed to a greater number of MS parameters (2.1 ± 1.0 v 1.6 ± 0.7, P = .004) (Table 4). Compared with other subjects, those with severe LVH had a greater BMI-SDS (3.5 ± 3.1 v 1.5 ± 0.8, P = .0001), WHR (0.91 ± 0.05 v 0.85 ± 0.06, P = .02), WHTR (0.56 ± 0.08 v 0.48 ± 0.12, P = .05), serum cholesterol level (201 ± 35 v 174 ± 30 mg/dL, P = .005), and serum LDL-cholesterol level (137 ± 36 v 110 ± 25 mg/dL, P = .001). The LVM correlated with WHR and WHTR and with LDL-cholesterol levels (Table 5).

Table 5

Univariate correlations between makers of hypertensive target-organ injury and metabolic abnormalities

LVMi versus	cIMT-SDS versus	fIMT-SDS versus	WCSA-SDS versus
BMI-SDS:	Adiponectin:	ApoA1/ApoB:	WHR:
r = 0.402, P = .0001	r = −0.428, P = .001	r = −0.249, P = .03	r = 0.441, P = .001
Number of MS criteria:	Fasting insulin:	Adiponectin:	Number of MS criteria:
r = 0.239, P = .01	r = 0.248, P = .02	r = −0.388, P = .008	r = 0.200, P = .04
WHR:	HOMA-IR:	hsCRP:	BMI-SDS:
r = 0.393, P = .003	r = 0.288, P = .01	r = 0.298, P = .009	r = 0.224, P = .02
Waist-to-height ratio:	Birth length:	Birth length:	Adiponectin:
r = 0.377, P = .005	r = −0.234, P = .001	r = −0.300, P = .009	r = −0.332, P = .02
LDL cholesterol:	Pulse pressure:	Pulse pressure:	Fasting insulin:
r = 0.222, P = .02	r = 0.314, P = .002	r = 0.255, P = .02	r = −0.355, P = .006
			HOMA-IR:
			r = 0.341, P = .005
			ISI[0,120]:
			r = −0.280, P = .02

ApoA1 = apoprotein A1; ApoB = apoprotein B; BMI-SDS = body mass index standard deviation score; HOMA-IR = homeostatic model assessment of insulin resistance; hsCRP = C-reactive protein; ISI[0,120] = insulin sensitivity index; LVMi = left-ventricular mass index; MS = metabolic syndrome; WHR = waist-to-hip ratio.

Table 4

Relationship between number of metabolic syndrome criteria and prevalence of left-ventricular hypertrophy for classic criteria of metabolic syndrome (χ² = 9.823; P = .02) or without PH as criterion (χ² = 10.429, P = .01)

No. of MS criteria	LVMi <95th percentile (N = 67)	LVMi >95th percentile (N = 46)	No of MS criteria including only anthropometrical and biochemical variables	LVMi <95th percentile (N = 67)	LVMi >95th percentile (N = 46)
0	0	0	0	33 (49.2%)	13 (28.3%)
1 criterion(only arterial hypertension)	33 (70.2%)	14 (29.8%)	1	26 (38.8%)	18 (39.1%)
2 criteria (in addition to arterial hypertension)	26 (60.5%)	17 (39.5%)	2	7 (10.4%)	9 (13.0%)
3 criteria	7 (43.8%)	9 (56.3%)	3	1 (1.5%)	6 (13%)
4 criteria	1 (14.3%)	6 (85.7%)	4	0	0

LVMi = left-ventricular mass index; MS = metabolic syndrome; PH = primary hypertension.

Analysis of the relationship between the number of MS criteria and LVH showed that as the number of MS parameters increased, the prevalence of LVM >95th percentile increased. Among the subgroup of 23 hypertensive subjects with MS, LVH was found in 15 (65.2%), compared with 31 of 90 (34.4%) hypertensive subjects without MS (Table 4). Severe LVH was found in 6 of 23 (26.1%) subjects with MS and in 8 of 81 (9%) subjects without MS (χ² = 4.912, P = .02).

Both absolute and index values of cIMT and fIMT correlated with IR, serum adiponectin concentrations, and pulse pressure (Table 5 and Fig. 1). Leptin levels did not correlate with any markers of TOD.

Polynomial regression function between carotid artery intima-media thickness standard deviation source (cIMT-SDS) and serum adiponectin concentrations (P = .001).

Figure 1.

Stepwise regression analysis revealed that adiponectin was an independent predictor for cIMT-SDS (R² = 0.178, β = −0.466, P = .02). For fIMT-SDS, the main predictors were HOMA-IR (β = 0.669, P = .0001) and serum triglycerides (β = 0.388, P = .01, R² = 0.444). The LVM was predicted (R² = 0.332) by BMI-SD (β = 0.591, P = .005) and HOMA-IR (β = 0.380, P = .04).

Discussion

Children and adolescents with PH represent a unique group for investigation of the pathogenesis of hypertension. In contrast to adults, they are not exposed to such cardiovascular risk factors as diabetes and cigarette smoking, and clinically overt consequences of hypertension are not yet present. We found that MS was present in 20% of a nonpreselected cohort of untreated children referred for investigation because of arterial hypertension and in whom PH was ultimately diagnosed. Exposure to MS and IR were strong risk factors of LVH and intima-media thickening. A tendency toward hyperuricemia was found to be characteristic of patients with PH, irrespective of obesity.

Overweight and obesity are regarded as the most typical intermediate phenotype of children with PH.^1,2 Although there is still a difference in the prevalence of obesity between adolescent populations in the US and Europe, we found a BMI ≥95th percentile in 57% of patients, compared with 15% in the control group. However, absolute values of BMI in our hypertensive subjects and controls were still lower than in reports from the US.^1,2 We did not find significant differences in classical markers of dyslipidemia between groups. However, hypertensive subjects tended to have lower HDL-cholesterol levels and apoA1:apoB ratios. On the other hand, lecithin:cholesterol transferase activity was lower in the hypertensive group. The activity of this particular enzyme is rarely analyzed in routine practice and population studies. Nevertheless, it regulates the rate of HDL-cholesterol particle generation, and so its lower activity correlates with a higher risk of premature atherosclerosis.¹⁵

The relationship of MS and IR with LVM and the development of LVH was described in both normotensive and hypertensive adults.^16–20 Cuspidi et al found that among untreated hypertensive adults, LVH was present in 30% of those presenting with MS, in contrast to 20% among hypertensive subjects without MS.²¹ Similarly, Leoncini et al found that MS was associated with the presence of early signs of TOD in nondiabetic adults with PH.²² Other studies of hypertensive adults found that MS correlated not only with LVH but also with left-atrial enlargement.²³

In contrast to adult studies, there is a paucity of data on MS and its relationship with TOD in children with PH. However, in normotensive but overweight children, hyperinsulinemia is a risk factor for LVH, independently of any relation between insulin, obesity, and blood pressure.²⁴ Srinivasan et al showed that MS parameters were present very early in children who consequently developed PH as young adults.²⁵

Hypertensive children with LVH had a higher birth weight and BMI and demonstrated central obesity. However, they did not differ from those who had normal LVM in terms of IR indices. This is in contrast to studies in adults, in which IR was found to be the main risk factor for LVH.^16,18,21 It must be stressed that none of our hypertensive subjects had either diabetes or glucose intolerance. One may speculate that the absence of direct, strong relationships between IR indices and LVH is time related, as this relationship requires a sufficiently long exposure to insulin levels, as is the case in adults. On the other hand, we found that hypertensive subjects had higher fasting insulinemia values compared with their normotensive peers, even after adjusting for BMI. This may indicate that subjects with PH are prone to IR and that this phenomenon may already be observed in childhood and adolescence.

Increased LVM in children with higher birth weights was also observed by others and in our previous study.^4,26,27 This is in contrast with Barker's hypothesis and theory of the relationship between low birth weight and the future development of PH.²⁸ It seems that the predisposition to hypertension related to low birth weight is different and independent from the predisposition to a hypertrophic reaction of the left ventricle to hemodynamic and metabolic factors.

Although two markers of arterial injury were significantly increased in hypertensive children, the cluster of “classic” MS parameters had a relatively weak impact on both cIMT and fIMT. Hypertensive subjects with greater cIMT did, however, have higher fasting insulin concentrations and HOMA-IR and lower ISI[0,120] and adiponectin concentrations. This indicates a direct relationship between wall injury, hyperinsulinemia, and IR, which may be aggravated or caused by low adiponectin levels. The absence of a strong and direct relationship between classic MS criteria and IMT may be related to the relatively short duration of metabolic abnormalities and hypertension. One may assume that the longer the exposure to hypertension and metabolic abnormalities, the greater the probability of development of close relationships between metabolic risk factors and arterial wall injury. Retnakaran et al found that MS in childhood presents as a cluster of “nontraditional” risk factors and that a low adiponectin level is one of them, possibly suggesting that abnormalities of adipocytokine metabolism may be an early sign of MS.²⁹ Our findings are similar to those of Pilz et al, who found that low adiponectin serum levels were associated with increased cIMT.³⁰ However, in other analyses it was found that in children and adolescents with PH, both cIMT and fIMT were related to classic biochemical cardiovascular risk factors.⁴

The finding that hyperinsulinemia and IR were constant predictors for both IMT and LVH once more underscores the significant role of IR in TOD in hypertensive children. One must take into account, however, that the results of a regression analysis depend on the number of subjects analyzed and can differ between studies.

A tendency toward hyperuricemia is regarded as typical for PH, and some authors hypothesized that it may play a role in the pathogenesis of PH.³ Higher UA was found to discriminate between PH and secondary hypertension. Moreover, treatment with allopurinol alone normalized blood pressure in a small study of children with PH.³ In accordance with previous reports,⁴ we found that in hypertensive subjects, UA concentrations were increased irrespective of BMI. Lean patients with PH had higher UA concentrations than obese normotensive peers. However, UA levels correlated with central obesity. This indicates that metabolic abnormalities in PH children are interrelated and that their common pathogenetic link is obesity.

The main limitation of our study is a lack of ECHO data from control subjects. Because both control and hypertensive subjects were of white origin, our results may not be generalized to other populations, especially outside Europe. Another weak point involves the cross-sectional character of the study and the nonblinded ECHO and IMT measurements.

In conclusion, we found that in a nonselected group of children with early-stage PH, significant metabolic abnormalities related to overweight and the central distribution of fat tissue are present. Metabolic syndrome was found in 20% of children with PH, and its presence was strongly related with TOD. Markers of IR were significant and independent predictors for TOD.