|Home | About | Journals | Submit | Contact Us | Français|
Hypertension (HT) is associated with increased LVM and carotid intima-media thickness (cIMT) which predict cardiovascular (CV) events in adults. Whether target organ damage (TOD) is found in prehypertensive youth (PreHT) is not known. We measured BMI, BP, fasting glucose, insulin, lipids and CRP, LVM/ht2.7 (LVMI), diastolic function, cIMT, carotid stiffness, augmentation index (AIx), brachial artery distensibility (BrachD), pulse wave velocity (PWV), in 723 subjects 10–23 yrs (29% type 2 diabetes mellitus, T2DM). Subjects were stratified by BP level (normotensive: NT=531, PreHT=65, HT=127). Adiposity and CV risk factors worsened across BP group. There was a graded increase in cIMT, arterial stiffness, and LVMI and decrease in diastolic function from NT to PreHT to HT. In multivariable models adjusted for CV risk factors, status as PreHT or HT remained an independent determinant of TOD for LVM, diastolic function, internal cIMT, carotid and arterial stiffness. PreHT is associated with CV TOD in adolescents and young adults.
Hypertension (HT) is an established risk factor for target organ damage (TOD) in adults. High blood pressure (BP) is associated with increased left ventricular mass (LVM),1 carotid intima-media thickness (cIMT)2 and arterial stiffness.3 Since TOD is known to predict hard cardiovascular (CV) events,4–6 screening for TOD has become an established practice in preventive care for adults at risk for complications related to elevated blood pressure.7, 8 Recent data suggest that no safe cut-point for blood pressure exists as prep-hypertension (between the 90th and 95th percentile for <18 years or between 120/80 and 140/90 mmHg for adults) may progress to HT9 and TOD may begin at pre-hypertensive (PreHT) levels.10 Studies in youth have documented TOD with sustained hypertension,11 but few data exist showing an increase in left ventricular hypertrophy in PreHT youth12 and data on arterial abnormalities are lacking. Determining the prevalence of TOD in youth at borderline levels of BP is important as current pediatric guidelines13 determine treatment levels based on arbitrary cut-points without reference to hard CV events or intermediate non-invasive outcomes. Therefore, we performed non-invasive imaging in adolescents and young adults to determine if TOD could be documented in PreHT subjects before the onset of clinical HT.
These analyses were performed on data collected for a study examining effects of obesity and type 2 diabetes mellitus (T2DM) on CV structure and function. By design, 1/3 were T2DM (N=258), 1/3 obese (≥95th percentile for BMI) but non-diabetic (O, N=234), and 1/3 lean (L, N=231 <85th percentile for BMI).14 First, subjects with T2DM (provider diagnosed) were recruited from the Cincinnati Children’s Hospital diabetes clinic (average duration of diabetes 3.6 ± 2.6 years). Each diabetic subject was then matched by age, race and gender to two controls (one L and one O). All obese subjects underwent a 2-hour oral glucose tolerance test to rule out sub-clinical T2DM according to ADA guidelines.15 Pregnant females were excluded from the study. The final study population consisted of 723 subjects with mean age 18 years, 60% non-Caucasian, 34% male, 29% T2DM.
Prior to enrollment in the study, written informed consent was obtained from subjects ≥18 years old or the parent or guardian for subjects < 18 years old. Written assent was also obtained for subjects < 18 years old according to the guidelines established by the Institutional Review Board at Cincinnati Children’s Hospital.
After a minimum 10 hour overnight fast, participants had questionnaire, anthropometric, blood pressure (BP), laboratory and arterial stiffness data collected. Trained personnel obtained two measures of height using a calibrated stadiometer (Veeder-Rood, Elizabethtown, NC). Each subject’s height was measured with the subject in the standing position wearing socks, heels together, toes apart at a 45 degree angle and the head in the Frankfort horizontal plane. Two height measurements were obtained with a third measurement taken if the first two were more than 0.5 cm apart. Weight was measured using a Health-O-Meter electronic scale. The scale was calibrated once per month and was used exclusively for this investigation. Two weight measurements were obtained. A third measurement was taken if the first two differed by more than 0.3 kg. Body mass index was calculated as kilograms per meter squared.
Blood pressures were measured using a standardized protocol according to the standards of the Fourth Report on BP in Children with mercury sphygmomanometery.13 Blood pressure examiners were certified after receiving 16 hours of instruction and evaluation. Examiners were recertified annually. Participants were seated with feet resting flat on a surface and right arm resting at heart level. The appropriate cuff was selected based on arm circumference and placed around the upper arm. Using a standard mercury sphygmomanometer (Baum Desktop Model with V-Lok cuffs), three blood pressures were measured by rapidly inflating to the maximum inflation level and deflating at a rate of 2 mmHg per second, with 60 seconds rest between each determination. The first appearance of two consecutive beats determined the first Korotkoff phase (k1); the point at which a sound became muffled determined K4; and the sound disappearing determined K5. The pulse rate was measured for 30 to 60 seconds between blood pressure determinations. Three blood pressure measurements were obtained. The three blood pressure determinations were averaged to calculate the mean systolic and diastolic blood pressure (K5). If any 2 of the 3 readings varied by more than 10mmHg, a fourth reading was performed and included in the average. The mean of 3 resting measures was used. Subjects were stratified as normotensive (NT = 531), pre-hypertensive (PreHT = 65) or hypertensive (HT = 127) by BP level according to the 4th Report on BP in children (based on gender, age and height)13 or JNC7 cut-points7 if they were 18 years or older.
Physical activity was assessed using an Actical accelerometer (Phillips Respironics) worn on the waist during waking hours over a 7 day period. This device is an omni-directional detector that provides counts of movement in all directions.16 Counts of activity per minute worn were calculated and averaged over the 7 days.
Fasting plasma glucose was measured using a Hitachi model 704 glucose analyzer (Roche Hitachi, Indianapolis, IN) with intra-assay and inter-assay coefficients of variation of 1.2% and 1.6% respectively.17 Plasma insulin was measured by radio-immunoassay using an anti-insulin serum raised in guinea pegs, 125I labeled insulin (Linco, St. Louis, MO) and a double antibody method to separate bound from free tracer. This assay has a sensitivity of 2 pmol and has intra- and interassay coefficients of variation of 5% and 8%.18 Assays of fasting plasma lipid profiles were carried out in a laboratory which is NHLBI-CDC standardized with the LDL cholesterol concentration calculated using the Friedewald equation. High sensitivity c-reactive protein (CRP) was measured using a high sensitivity enzyme-linked immunoabsorbent assay. HbA1c was measured in red blood cells using HPLC methods.
Echocardiography was performed with a GE or Philips Sonos 5500 (Andover, Massachusetts) system with the patient in the left decubitus position. Para-sternal long, short axis and apical 4 chamber views were recorded with 3 cardiac cycles averaged for each variable. Left ventricular end-diastolic dimension, end-systolic dimension, end-diastolic septal thickness and end-diastolic and end-systolic posterior wall thicknesses were measured off-line by either of two sonographers using a Cardiology Analysis System (Digisonics, Houston, Texas). LVM was calculated with the formula of Devereaux et al19 and LVM index (LVMI = LVM/ht2.7) by De Simone’s method.20 Relative wall thickness (RWT) at end-diastole was also calculated. The cutpoints of 51 g/m2.7 and a RWT of 0.41 were used to define geometry as previously described.11
For diastolic function, mitral inflow velocities were obtained with pulsed wave Doppler in the apical 4-chamber view. The Doppler cursor was placed parallel to mitral inflow and maximal velocity was measured with the sample volume at the mitral valve leaflet tips. The mitral peak E (early filling) and A (inflow with atrial contraction) waves were measured off-line and an E/A ratio was calculated. Tissue Doppler Imaging myocardial flow velocities were acquired in the apical 4-chamber view. The peak (Ea) and late velocities (Aa) of mitral annular flow were recorded at both the septal and lateral annulus and both lateral and septal Ea/Aa ratios and their average were calculated. Other diastolic variables calculated include: E/Ea lateral and septal average and E over average of Ea/Aa ratio from the lateral and septal aspects of the valve.
Carotid ultrasound studies were performed by a single registered vascular technologist with high-resolution B-mode ultrasonography (GE Vivid7, Milwaukee, Wisconsin) with a high resolution linear array vascular transducer (7.5 MHz). A 2-D image of the carotid artery was obtained from the far wall for measurement of intima-medial thickness in the common, bifurcation and internal carotid segments. Then images of the common carotid with both the near and far wall visualized were obtained for M-mode evaluation of peak and minimal diameters for calculation of arterial stiffness.21 All digital images were read off-line using the Camtronic Medical System software. Calculations included Peterson’s elastic modulus (PEM),22 Young’s elastic modulus (YEM).23 Due to pulse wave amplification along the arterial tree resulting in overestimation of brachial SBP,24 central BPs obtained with the SphygmoCor device (see arterial stiffness below) were used in the calculations of carotid stiffness. The central BPs were obtained on average, no more than 30 minutes prior to the carotid ultrasound.
The average of three measures of all vascular function measures were used in analyses. Each measure was conducted after 5 minutes of rest in the supine position. A DynaPulse Pathway instrument (Pulse Metric, Inc., San Diego, CA) collected brachial artery distensibility (BrachD) as previously described.25 This device derives brachial artery pressure curves from arterial pressure signals obtained from a standard cuff sphygmomanometer assuming a straight tube brachial artery and T-tube aortic system.25 Repeat measures in our laboratory show excellent reproducibility with coefficients of variability less than 9%.26
A SphygmoCor SCOR-PVx System (Atcor Medical, Sydney, Australia) was used for measurement of carotid-femoral pulse wave velocity (PWV) and augmentation index (AIx), an arterial stiffness measure incorporating features related to arterial stiffness and provides additional information concerning wave reflections.27 This device employs a tonometer applied on the artery of interest to obtain ECG-gated pressure data. PWV is calculate as the difference in the carotid-to-femoral path length (measured directly and entered into the device) divided by the difference in the R-wave from the ECG to the foot of the pressure wave taken from the superimposed ECG and pressure tracings. For AIx, the pressure waves are calibrated using MAP and DBP obtained in the same arm. A validated generalized transfer function is then applied for estimation of the central aortic pressure tracing and calculation of AIx.28 Since AIx is affected by HR, values are adjusted to a HR of 75 beats per minute. Repeat measures in our laboratory show excellent reproducibility with coefficients of variability less than 7% for PWV and intraclass correlation coefficients between 0.7 and 0.9 for AIx variables.26
All analyses were performed with Statistical Analyses Software (SAS®, version 9.2)29 Average values for demographic, anthropometric, and laboratory data were obtained by BP group. Analysis of variance was performed (or chi square analyses for categorical variables) to look for differences by BP group. Variance stabilizing measures to transform non-normally distributed variables were performed as needed. Bivariate correlations were calculated between TOD measures and all covariates overall and by BP group. General linear models were constructed using important covariates from correlation analyses to determine if BP group was an independent determinate of TOD even after inclusion of CV risk factors in the models.
The full model contained age, demographics (race, sex), anthropometric (waist/height ratio (WHR), BMI z-score), hemodynamic (MAP to adjust for baseline distending pressure, HR except for the model for AIx, which is already adjusted for HR), laboratory (C-reactive protein and fasting LDL-C, HDL-C, triglycerides, glucose, insulin,), presence of T2DM and physical activity (average activity counts per non-zero minute) measures. Total cholesterol was highly collinear with LDL-C, and HbA1c with fasting glucose so these covariates were omitted to ensure stability of the models. Height was added to the model for AIx since height directly influences distance of wave reflection sites from the heart. Height is used in the calculations for BrachD and PWV so it was omitted from models for those outcomes. Significance of each covariate in the initial model was assessed and non-significant terms were removed by backward elimination until all remaining covariates or their interaction (effect modifier) terms were significant (p<.05).
The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
Subject characteristics are displayed in table 1. NT participants were slightly younger than PreHT and HT subjects. There were no race differences but there were fewer males in the HT group. The prevalence of T2DM and measures of adiposity and BP worsened from NT to PreHT to HT. NT subjects tended to have a better lipid profile, metabolic control and level of inflammation than the other groups. HT had the lowest activity levels (all p≤0.05 for comparisons listed above).
LVMI increased across the BP groups (table 2 and figure 1). HT subjects demonstrated a higher prevalence of abnormal geometry, 23.6% (for all abnormal patterns combined) compared to 7.7% for NT and PreHT (p≤0.05). There were no group differences in systolic function (shortening fraction, velocity of circumferential fiber shortening or wall stress, data not shown). NT subjects had better diastolic function than HT for all measures. NT were also better than PreHT for mitral E/A ratio, TDI Ea/Aa septal ratio, average septal/lateral Ea/Aa ratios and E/average Ea/Aa TDI lateral and septal ratios (all p≤0.05).
NT had significantly lower IMT than the other BP groups for the bulb and internal carotid artery segments and they had more flexible common carotid arteries as measured by PEM and YEM (table 3 and figures 2 and and3,3, all p≤0.05). There was a graded increased in AIx and PWVf with a similar decrease in BrachD among the BP groups (table 3) indicating progressively stiffer vessels across the BP strata.
Multivariable models demonstrated that BP group remained a significant predictor of LVMI, E/Ea lateral ratio and average septal/lateral Ea/As and E/average Ea/Aa TDI lateral and septal ratios (table 4). Bp group was also an independent predictor of (table 5) vascular damage for the internal carotid IMT, common carotid stiffness (PEM, YEM), AIx, BrachD and PWV even after adjusting for CV risk factors and distending pressure (MAP). Plotting the age by BP group interaction for BrachD revealed a steeper decline in BrachD across BP groups for adolescents than young adults (data not shown).
Our data demonstrate that significant abnormalities in cardiac and vascular measures can be identified in youth with pre-hypertension (increased LVM, carotid thickness, arterial stiffness and decreased diastolic function). Although a deteriorating risk factor profile was seen across the BP distribution, the adverse cardiac and vascular changes are largely independent of other traditional CV risk factors. This is evident from the observation that classification as PreHT was an independent predictor of many measures of TOD (LVMI, E/Ea, Avg lat-sept Ea/Aa, internal cIMT, PEM, YEM, AIx, BrachD, PWV) even after adjusting for CV risk factors including BMI and presence of T2DM. This suggests that even mild elevation in BP is an important etiology for TOD.
In hypertensive adults, elevated LVM is a well described independent risk factor for adverse CV events30 and is associated with development of depressed LV systolic function, a precursor of heart failure.31 Concentric hypertrophy, the geometric pattern most frequently seen in sustained hypertension, is also associated with a poor prognosis.32 However, cardiac abnormalities can be found in pre-hypertensive adults. Recent studies found depressed diastolic function in pre-hypertensives33, 34 and two large studies found higher LVM in these patients even after adjustment for other CV risk factors.35, 36 Prehypertension may also lead to more ageing-related increase of LVM10 Furthermore, progression from pre- to sustained hypertension in the Strong Heart was predicted by both baseline systolic BP and also by baseline LVM37 with the probability of developing incident hypertension increasing 36% for each standard deviation of LVMI.38 The finding that development of mild LV thickening may accelerate progression to higher BP levels suggests that pre-hypertension is not a benign condition.
LVH can also be identified in youth with hypertension.39, 40 Using the adult cut-point of 51 g/m2.7, Daniels, et al, found the prevalence for hypertension-related LVH to be 8% in a clinic population11 while a multi-center study found the prevalence to be as high as 15.5%.41 If the pediatric definition of ≥95th percentile of LVM is used, the prevalence may be as high as 30 to 40%.41–43 Important epidemiologic studies of CV risk factors in youth also confirm a strong association between BP levels and LV thickness in non-hypertensive subjects. The Muscatine Heart Study demonstrated that resting SBP exerted an independent influence on LVM in children44 while the Bogalusa Heart Study found the cumulative burden of SBP from childhood to adulthood was a significant predictor of LVMI in young adults.45 Other cross-sectional studies of normal children confirm the independent relationship between BP and LVM.46, 47 Therefore, it is not surprising that youth diagnosed with pre-hypertension may also exhibit LVH12, 48 with odds for having elevated LVMI increasing by 54% for each incremental increase in the standard deviation score for 24-hour ambulatory systolic BP.49 Higher ambulatory BP is also significantly associated with a higher prevalence of abnormal LV geometry in children and adolescents50 and BP also relates to LA diameter51 and decreased diastolic function in youth.52, 53 Our data confirm the adverse effect of pre-hypertensive BP levels on LV structure and function in a larger cohort of adolescents and young adults.
As with LVM, carotid structure is also adversely affected by hypertension. Among all the metabolic syndrome components, hypertension carried the greatest odds ratio (1.43, CI 1.27–1.60) for presence of carotid plaque, a risk factor for stroke, in a large study of Japanese subjects age 19 to 88 years.54 However, hypertension is also linked to early carotid changes. Carotid IMT increased across BP categories in all race and gender groups in the Atherosclerosis Risk in Communities (ARIC) Study,55 a finding replicated in other large population-based studies.56, 57 Presence of hypertension also predicts progression of cIMT,58, 59 so it is not surprising that greater carotid thickness can be found in adults with pre-hypertension35, 60 and normal adults with multiple CV risk factors,61 with a 0.058 mm increase in cIMT seen per 1 SD (21 mmHg) increase in BP in a multi-ethnic study by Psaty et al.62
The adverse changes in carotid structure seen in hypertensive adults are accompanied by parallel deterioration in carotid function. The ARIC study found increased carotid stiffness predicted development of hypertension3 and hypertension was associated with increased carotid stiffness.63 However, as in earlier studies,64 the increase in stiffness was dependent upon baseline distending pressure. In contrast, other investigators have found the hypertensive-related increase in carotid stiffness to be independent of baseline pressure but only in younger hypertensives.65 Hypertension may have a stronger effect on arterial stiffness in younger individuals while age and other CV risk factors may be more important at older ages. Data demonstrating that pre-hypertensive men, if young, have lower carotid distensibility than controls66 support this hypothesis. It is possible that other age-related risk factors have a more powerful effect on carotid stiffness than BP, at older ages.
Recent studies have demonstrated a relationship between BP and carotid structure and function in youth. Children referred to a hypertension clinic were found to have thicker common carotid artery cIMT compared to controls. In two studies, the relationship was not independent of BMI.67, 68 However, other investigators have found the relationship between BP and cIMT to be significant even when adolescents are matched by BMI,69 or statistical adjustment for adiposity is performed.70 In a large recent study from our group, we found the obesity-independent relationship between BP and cIMT also existed for the carotid bulb and internal carotid artery segments.71 A few studies have related increased carotid stiffness to hypertension in youth.72 However, one investigator found the relationship only when lean controls, rather than obese controls, were compared to the hypertensive youth.68 Two other studies, found the relationship to be obesity-independent.71, 73 Our current paper extends these observations by providing data on all 3 carotid artery segments and examining the effect of both hypertension and pre-hypertension on carotid artery thickness and stiffness.
The majority of studies relating BP to arterial stiffness measure PWV. PWV is a robust measure which not only predicts CV events,74, 75 but also CV mortality.76 Indeed, increased arterial stiffness, including faster PWV,77 has been found with greater CV risk such as in hypertension. There are also some limited data relating hypertension in adults to higher AIx78, 79 and lower BrachD.80 Unfortunately, treatment of hypertension in adulthood may not normalize PWV81 and annual rates of progression of PWV are higher in hypertensives compared to controls even if BP is well controlled.82 Underlying abnormalities in arterial stiffness may be contributing to development of hypertension83 which then causes further deterioration in arterial elasticity. Higher PWV84 and AIx85 have also been documented in pre-hypertensive adults. PWV gradually increased as a function of BP classification from normal to pre- to stage II hypertension in one study.86 Furthermore, studies of normotensive young adults with a positive family history of hypertension have demonstrated lower BrachD,80 higher PWV87, 88 and Aix,89 suggesting an underlying genetic tendency for vascular dysfunction that may impact risk for developing hypertension. Therefore, to prevent development of sustained hypertension, it may be useful to assess arterial stiffness in high risk individuals.
Arterial stiffness assessment is being performed in increasing numbers of pediatric studies. As in adults, most pediatric studies focus on PWV although normative data remain spars. A recent study by Reusz, et al, provided PWV results on 1008 healthy subjects (6 to 20 years) obtained with a similar method as employed in our study.90 They found a strong correlation between BP and PWV although no multivariable analyses correcting for other CV risk factors were performed.90 Our previous data on 670 adolescents and young adults demonstrated that mean arterial pressure remained a predictor of PWV (and AIx and BrachD) even after correcting for adiposity, metabolic abnormalities (glucose, insulin, type 191 or type 226 diabetes) and inflammation. A few studies have specifically evaluated the relationship between BP classification and PWV including one that found higher PWV in pre-hypertensive adolescents compared to controls, but only in Caucasians.92 Our data found higher PWV in pre-hypertensive non-Caucasians, however, we measured the standard carotid-femoral PWV and in the Zhu, et al, paper, carotid to dorsalis pedis was measured,92 and it is known that PWV is higher in smaller leg vessels compared to the central aorta.26 A study examining younger children, mean age 11.4 years, demonstrated higher PWV in subjects with SBP ≥ 90th percentile, the cut-point for pre-hypertension, compared to normotensives.93 However, the investigators did not determine if differences existed between pre- and true-hypertensives. Our findings confirm the graded increase in PWV from normo- to pre- to hypertensive youth, We also provide BP level stratified data for AIx and BrachD, techniques previously employed to investigate other CV risk factors in youth such as diabetes94 and metabolic syndrome,95 but not to date used for pediatric hypertension research.
Our finding of a graded increase in the prevalence of TOD across the BP strata, although cross-sectional, suggests that progression to higher levels of BP increases CV risk at a young age. However, our cross-sectional findings need to be confirmed in longitudinal studies. Furthermore, due to our study design, our population had a high prevalence of obesity and T2DM. However, BP classification remained an independent predictor of all the TOD measures even in multivariable models where BMI and presence of diabetes were entered as covariates. Furthermore, the prevalence of both obesity and T2DM are increasing around the globe. Therefore, our data point to the importance of modifying CV risk factors in high risk youth even if only at borderline levels.
Some studies have suggested that the relationship between BP and arterial stiffness merely reflects the effect of increased distending pressure on the vessel.63 Investigations of brachial arterial compliance under isobaric conditions demonstrating impaired vascular function in hypertensives refute this assertion.96, 97 Furthermore, our model controlled for MAP and still found an effect of BP group on BrachD, suggesting the effect was independent of baseline pressure.
There is much controversy on the appropriate method to index LVM to correct for differences in body size. Some studies have shown that fat-free body mass is more closely related to LVM than other anthropometric measures.46, 98, 99 We chose to index LVM to ht2.7 because measurement of fat-free mass requires specialized equipment not readily available to many physicians and because the de Simone method of indexing LVM 20 has produced a gender-independent partition value of 51 g/m2.7 that has proven better at predicting incident CV events4, 99 compared to other allometric adjustments including indexing to ht1.7 suggested by 4, 100Chirinos, et al,101 which was only superior at predicting all-cause mortality.
Our data provide additional support for the argument that BP is having an important effect on the CV system in adolescents and young adults even with only modest elevation in BP (above the 90th%). This supports the concept that pediatricians should be prospectively identifying children and adolescents with BP above the 90th percentile and should begin lifestyle intervention earlier to prevent cardiac and vascular consequences. This also suggests that it may be necessary to consider implementation of pharmacologic intervention earlier and at a lower BP to prevent progression to sustained hypertension as documented in the adult TROPHY study.102 This is especially important as these target organ changes may well be part of a vicious cycle that leads to further increases in BP and greater target organ disease. Longitudinal trials addressing the issue of earlier treatment based on intermediate non-invasive CV endpoints rather than using arbitrary cut-points are needed.
We would like to acknowledge the work of the entire Cardiovascular Disease in Type 2 Diabetes Study team. We would also like to thank the participants of the Cardiovascular Disease in Type 2 Diabetes Study and their families, without whose support this study would not be possible.
Funding sources: This study was supported by NIH (NHLBI) R01 HL076269 (CV Disease in Adolescents with Type 2 Diabetes) and in part by USPHS Grant #UL1 RR026314 (National Center for Research Resources, NIH).
Disclosures: The authors have no other conflicts of interest or relationships with industry to disclose.