There is a high burden of premature cardiovascular disease (CVD) among African Americans. Measures of central aortic blood pressure (CASP) and wave reflection are predictive of CVD risk in adults, but there is a paucity of data regarding the relation of these measures to target organ damage among adolescents. The objective of this study was to examine the relationship between CASP, central pulse pressure (CPP), and augmentation index (AI) with left ventricular mass index (LVMI).
A cohort of 120 African-American adolescents was examined. Study participants underwent measurement of peripheral blood pressure (BP) using auscultation, pulse wave analysis (PWA) for determination of CASP, CPP, and AI, and echocardiography for determination of LVMI.
The cohort was 55% male, with mean BP 114/62 mm Hg, mean LVMI 36 g/m2.7, mean CASP 94 mm Hg, mean CPP 31 mmHg, and mean AI was 0.5%. After adjustment for potential confounders, peripheral systolic BP (SBP) was significantly associated with LVMI (P = 0.008), but diastolic pressure was not (P = 0.887). The CASP and CPP were significantly associated with LVMI (P = 0.020 and 0.005, respectively). Peripheral SBP, CASP, and CPP had similar associations with respect to LVMI (r2 = 0.26, 0. 26, and 0.27, respectively).
Central BP is associated with LVMI among African-American adolescents, and these associations are similar to those seen with peripheral BP measurements.
Hypertension is a major risk factor for future cardiovascular disease (CVD) and is associated with changes in vascular function and structure. Although peripheral blood pressure (BP) is linked to CVD, central BP may be a more sensitive indicator of CV risk and cannot be inferred from brachial BP measures.1 Properties of wave reflection and arterial stiffness influence central BP and drive much of the disparity between central and peripheral BP measures. Currently, devices are available for noninvasive measurement of central aortic blood pressure (CASP) and parameters of wave reflection that can be used to quantify alterations in vascular structure and function. Recent evidence suggests that CASP and central pulse pressure (CPP) are more sensitive markers of CVD risk compared to peripheral (brachial) BP in some groups.2,3 CASP and CPP are more reflective of the BP experienced by major organs, such as the brain, heart, and kidneys, and therefore have a stronger association with CV risk. Central BP is associated with traditional risk factors for CVD, such as age and diabetes, independent of peripheral BP1. Augmentation index (AI) is the difference between the peak aortic pressure and the aortic pressure at the start of the reflected wave divided by pulse pressure. It is a measure of arterial stiffness, as a stiff aorta results in an early return of the reflected wave to central circulation as a result of increased pulse wave velocity.4
Recent reports indicate that these alterations in the arterial tree are evident at an early phase in life among children and adolescents with known risk factors for CVD, such as diabetes.5,6 African Americans are known to have a high burden of hypertension and associated CVD that occurs at an earlier phase in life compared to whites.7,8 Data from the Coronary Artery Risk Development in Young Adults (CARDIA) study demonstrates that among African American young adults, BP is a predictor of early onset heart failure.9 The ability to detect alterations in vascular function and structure among young African Americans may allow for early identification of those at risk for premature CVD. CPP most likely reflects the pressure against which the heart contracts during systole and may play an important role in the development of increased left ventricular mass (LVM). Therefore, we conducted a study to determine the association between central BP and left ventricular mass index (LVMI), among African American adolescents.
Healthy African American adolescents (ages 13–18 years) were recruited in Philadelphia through local advertisements and from faculty practices in the Departments of Family Medicine and Pediatrics at Thomas Jefferson University between 2009 and 2011. Adolescents with and without obesity (defined as body mass index (BMI) above the 95th percentile according to age and gender) were included, as were those with and without elevated BP (defined as BP ≥120/80 mm Hg). Consecutive subjects were enrolled. Exclusions included known secondary hypertension, diabetes, renal disease, CVD, autoimmune disease, thyroid disease, sickle cell disease, eating disorders, and use of steroids. Secondary hypertension was excluded by self-report and according to medical history provided by primary physician if available. Additionally, for each subject, the following tests were performed: fasting plasma glucose, fasting plasma insulin, fasting plasma lipids, plasma renin activity, serum aldosterone, serum creatinine, urine albumin excretion, and transthoracic echocardiogram. Finally, no subjects with stage II hypertension were enrolled. The study protocol was approved by the Institutional Review Board of Thomas Jefferson University. Written informed consent was obtained from 18-year-old participants, while for children age <18 years, consent was obtained from the parent or guardian at enrollment and assent was obtained from the child.
Data on health status, medication use, and health related behaviors were obtained by self report of each participant or guardian (for younger children). Clinical assessment consisted of BP and anthropometric measurements (height, weight, and waist circumference). BMI was calculated as weight (kg) divided by height squared (m2), and obesity was defined according to the CDC criteria for children (http://www.cdc.gov/obesity/childhood/defining.html), which are based on the child's age, sex, and BMI. BP measurements were obtained by auscultation on each subject following a 10-min rest period in a seated position using an aneroid device. Measurements were performed on the right arm using a cuff large enough to encircle 80% of the subject's upper arm. The average of three successive measurements of systolic BP (SBP) and diastolic BP (DBP) on two separate visits was used as the BP value for each participant. Peripheral pulse pressure (PPP) was computed as the difference between SBP and DBP.
Pulse wave analysis (PWA), using a SphygmoCor device (AtCor Medical, West Ryde, Australia), was performed to determine CASP, CPP, and AI. PWA was performed upon each participant in the seated position after 5 min of rest. The SphygmoCor device uses applanation tonometry of a peripheral artery (in this case the radial artery) to capture the arterial waveform. We used peripheral SBP and DBP for calibration to allow the device to use its internal transfer algorithm to calculate CASP. The device has built-in quality controls to ensure that an adequate waveform reading is captured.
Echocardiography was performed for determination of LVM. LVM was measured by 2-dimensional guided M-mode echocardiography. A trained technician made measurements of the left ventricular internal dimension, interventricular septal thickness, and posterior wall thickness during diastole according to established methods by the American Society of Echocardiography.10 LVM is calculated from measurement of the left ventricle using the equation LVM (g) = 0.81 (1.04(interventricular septal thickness + posterior wall thickness + LV end diastolic internal dimension)3 ‐ (LV end diastolic internal dimension)3 + 0.06).11 LVMI is calculated as LVM/height in meters2.7.12 We chose to adjust LVM by meters2.7 as this is the method recommended by the National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents.13 The pediatric criteria for increased LVM, in children and adolescents, are based on LVMI percentile. The 95th percentile for LVMI in pediatric patients (based on normative pediatric LVMI data) is 36.88 g/m2.7 for females and 39.36 g/m2.7 for males.14 A single observer performed and interpreted the data related to assessment of LVM by echocardiogram. Relative wall thickness (RWT) was defined as two times the posterior wall thickness, divided by the left ventricular internal dimension.
Data are presented as mean and standard deviation (s.d.) or frequencies. Preliminary analyses were based on correlations between peripheral and central BP with LVMI and RWT. The main analyses were based on linear regression. Separate regression models were fit to assess the association between each peripheral and central BP measure with LVMI and with RWT. The final models controlled for age, sex, and obesity status. Statistical analyses were carried out in Stata 11 (Stata, College Station, TX).
A total of 197 African American adolescents underwent PWA. Of these, 63 were excluded because a satisfactory waveform reading could not be captured, and 14 were excluded because they did not undergo echocardiography. Comparison of the analysis sample to those excluded because of unrecordable waveform, did not show significant differences on age (P = 0.771), obesity (P = 0.169), heart rate (P = 0.102), or peripheral BP (P = 0.485 for systolic and 0.160 for diastolic), but the analysis sample had significantly more males (55% vs. 33%, P = 0.005). Therefore, the following analyses are based on 120 subjects, whose clinical and demographic characteristics are summarized in Table 1. LVMI ranged from 14 to 68 g/m2.7. We found a number of subjects with increased LVMI. Specifically, four subjects (all males) met criteria for left ventricular hypertrophy (LVMI >51 g/m2.7). Additionally, a number of subjects had LVMI above the 95th percentile. A total of 26 (39%) males had LVMI >39.36 g/m2.7 and 16 (30%) females had LVMI >36.88 g/m2.7.
Age was not significantly associated with LVMI (P = 0.97), but sex was (P < 0.01), with boys having on average a 4.4 g/m2.7 higher LVMI than girls. Obesity was also significantly associated with LVMI (P < 0.01), with overweight and obese adolescents having on average higher LVMI than normal-weight children, by 1.4 and 4.9 g/m2.7, respectively.
Table 2 summarizes the univariable correlations of the CASP, CPP, and AI with peripheral BP measures, LVMI, and RWT. Tables 3 and 4 summarize the final multivariable regression results (adjusted for age, sex, and obesity) for associations of both central and peripheral BP with LVMI (Table 3) and RWT (Table 4).
Separate regression models were performed to determine the association between each central and peripheral BP measure with LVMI. In univariable analyses, higher peripheral SBP and pulse pressure were associated with higher LVMI. The correlations for SBP with LVMI and pulse pressure with LVMI were, 0.39 (P < 0.01) and 0.36 (P < 0.01), respectively. Specifically, a 10 mm Hg increase in SBP was associated with a 1.7 g/m2.7 higher LVMI. DBP was not significantly associated with LVMI (correlation = 0.06, P = 0.53). After adjusting for age, sex, and obesity, peripheral SBP and PP remained significant and DBP remained nonsignificant (P = 0.01, 0.01, and 0.89, respectively, Table 3). In contrast, the peripheral BP measures were not associated with RWT (P = 0.16 for SBP, 0.11 for DBP, and 0.85 for PPP). After adjustment for age, sex, and obesity, SBP became a marginally significant predictor of RWT, although the strength of the association was weak; DBP and PPP remained nonsignificant (Table 4).
In univariable analyses, CASP and CPP were significantly associated with all peripheral BP measures and LVMI (Table 2). After adjustment for age, sex, and obesity, a 10 mm Hg increase in CASP was associated with an estimated 1.9 g/m2.7 higher average in LVMI (Table 3). Similarly, a 10 mm Hg increase in CPP was associated with an estimated 2.7 g/m2.7 higher average LVMI (Table 3). CASP and CPP were not significantly associated with RWT (P = 0.08 and 0.70, respectively). After adjustment for age, sex, and obesity, CASP had a modest, but significant association with RWT (Table 4). Age, sex, and obesity were not significant predictors of RWT (P > 0.5 for all three).
Central and peripheral BP measures were both associated with LVMI, and to a similar degree. A model that included only age, sex, and obesity (but no BP measures) had an r2 = 0.21, i.e., explained approximately one fifth of the LVMI variability. Addition of peripheral SBP or PP increased the r2 to 0.26 and 0.25, respectively, while addition of CASP or CPP increased the r2 to 0.26 and 0.27, respectively (Table 3).
The mean AI for the cohort is shown in Table 1. The correlations between AI and CASP was 0.15 (P = 0.10) and for AI and CPP was 0.20 (P = 0.03). AI was not significantly associated with peripheral PP, SBP, DBP, or LVMI (Table 2). The association between AI and LVMI remained nonsignificant even after adjustment for age, sex, and obesity. However, AI was significantly associated with RWT, both in unadjusted and adjusted analyses (Tables 3 and 4). A 10% higher AI was associated with about 0.01 higher average in RWT (Table 4).
In this study, we found that both central and peripheral BP are associated with LVMI among African American adolescents. After adjustment for potential confounders, SBP, peripheral PP, CASP, and CPP remained significantly associated with LVMI. The associations of LVMI with conventional measures found in this study are similar in magnitude to prior reports.15 Therefore, elevations in central and peripheral BP are associated with LVMI among young individuals at risk for future CVD.
A key finding of this study is the association between central BP and LVMI in this cohort of adolescents. To our knowledge, we are the first to demonstrate the association between central BP (both CASP and CPP) and LVMI among African-American adolescents. Pulse pressure is traditionally thought of as a marker of arterial stiffness, a process linked to aging. Therefore, pulse pressure has been demonstrated as a marker of increased CV risk primarily among older individuals.16,17 Among adults, elevated CPP is associated with CV risk factors such as diabetes, hyperlipidemia, and hypertension.1
It is increasingly recognized that evidence of target organ damage is present in adolescents with elevated peripheral BP and other risk factors for CV such as diabetes and obesity. In a study by Lurbe et al. adolescents, grouped according to birth weight and the absence or presence of current obesity, underwent peripheral BP measurement and PWA18. PPP was highest among those with current obesity and low birth weight. However, CPP was similar among the groups.
Despite limited data on PWA in the young, data that are available suggest that PWA indices, particularly AI, are associated with CVD risk in children. In a study of over 60 children with and without hypercholesterolemia, AI was higher among those with familial hypercholesterolemia.19 Children with type 1 diabetes and end stage renal disease have been reported to have greater AI compared to controls.20,21 We did not find an association between AI and LVMI. This may have been because the majority of our cohort was healthy, without diabetes, hyperlipidemia, hypertension, or renal disease. Additionally, AI and measures of wave reflection, may be influenced more by metabolic status and inflammation,19,22 whereas LVM is linked more strongly to pressure load, particularly among children and young adults.23,24
Prior studies have reported male sex and BMI as predictors of LVMI, similar to our findings.24 The association between LVMI and BMI is of particular concern due to the childhood obesity epidemic in the United States.
This study is limited by the modest sample size. Additionally, only African-American adolescents were enrolled in our study, therefore limiting generalizability. Moreover, central BP employs a transfer algorithm to translate the peripheral arterial waveform into a central aortic waveform. This transfer function is based upon invasive measurements performed in adults, and the technique has not been validated in children. However, available data indicate that these measurements, particularly AI, in children appear to be associated with CVD similar to adults.6
In summary, we have demonstrated that central and peripheral BP are associated with evidence of target organ damage, particularly LVMI, among African American adolescents. Our results indicate that, cross-sectionally, central and peripheral BP are similarly associated with CV risk. However, it is possible that central BP indices could be more strongly predictive of future CVD than measures of peripheral BP. Prospective studies on adolescent cohorts are needed to answer this question and to examine other aspects of the natural history of central BP.
This study was supported by National Institutes of Health grant 1RO1 HL90230.
The authors declared no conflict of interest.