Common Carotid Artery Stiffness and Patterns of Left Ventricular Hypertrophy in Hypertensive Patients
Abstract The aim of this study was to determine the relationship between the lumen diameter and function of the common carotid artery, a vessel representative of the capacitance portion of the circulation, and the different patterns of left ventricular hypertrophy in uncomplicated essential hypertensive patients. Carotid luminal diastolic cross-sectional area, distensibility, and compliance were derived from measurements by a high-definition echotracking system. Left ventricular dimensions were from echocardiography. The 86 hypertensive patients included 31 who had never been treated (group 1), 31 in whom treatment had been stopped for at least 2 weeks (group 2), and 24 treated patients (group 3). In multivariate analysis of the population as a whole, the following relations were statistically independent of age, blood pressure, gender, and group: Left ventricular end-diastolic volume index was positively correlated to carotid luminal cross-sectional area (r=.46, P<.0001) and compliance (r=.47, P<.0001); left ventricular mean wall thickness and mass-volume ratio were negatively correlated to distensibility (r=−.68, P<.0001; r=−.46, P<.0001, respectively) and compliance (r=−.40, P<.0001; r=−.37; P<.001, respectively); and left ventricular mass index was positively correlated to luminal cross-sectional area (r=.23, P<.02) and negatively to distensibility (r=−.26, P<.01). These results indicate that geometric and functional changes in the common carotid artery accompany geometric changes in the left ventricle. More specifically, they suggest that a reduction in distensibility paralleled cardiac concentric hypertrophy and remodeling, whereas an increase in arterial volume paralleled increased left ventricular cavity size.
Left ventricular hypertrophy (LVH) is an independent marker of increased cardiovascular risk in both the general population1 and hypertensive patients,2 although its origin in these patients is not fully understood. Genetic and neurohumoral factors may contribute to LVH, and mechanical factors are also known to be important determinants. Resistance to ejection may be divided into static and dynamic components.3 Mean blood pressure (MBP), which is partially a function of static resistance, is only weakly correlated to left ventricular mass index (LVMi), indicating that other mechanical factors could be associated with LVH. Vascular impedance, which is the correlate of static resistance in a pulsatile flow regimen, is mainly a function of the mechanical properties of the arterial system and may better represent left ventricular (LV) afterload.4 In this context, the potential associations between the mechanical properties of the arterial system and LVH have been investigated by several authors5 6 7 8 in small samples of normotensive subjects and hypertensive patients. LVH was shown to be associated with an elevation of the first harmonic modulus of impedance, measured invasively,5 and with the carotid-femoral pulse wave velocity,6 the aortic arch Peterson modulus,7 and the degree of increase in late systolic central aortic pressure,8 three noninvasive estimates of aortic stiffness.
The respective roles of high blood pressure (BP), aging, and arterial stiffness as determinants of LVH may be further analyzed in the light of the recently described patterns of cardiac adaptation in hypertension: remodeling and concentric and eccentric hypertrophy.9 This is of particular importance because these different patterns have differential consequences on morbidity, with relative wall thickness appearing as a stronger predictor of cardiovascular morbidity than eccentric hypertrophy.10 It is noteworthy that the ratio of LV mass to LV end-diastolic volume (M/V ratio), an index of concentric hypertrophy similar to relative wall thickness,11 was positively correlated with characteristic aortic impedance5 and carotid-femoral pulse wave velocity,6 whereas LVMi was not. In addition, Roman et al12 reported a positive correlation between carotid internal dimension and LV internal diameter independent of systolic BP. In the present study, we used a recently developed, highly accurate echotracking system13 to prospectively measure common carotid artery (CCA) internal diameter and its pulsatile change, allowing the derivation of cross-sectional distensibility, thus assessing the two components of cross-sectional compliance: arterial volume and volumic distensibility. Therefore, we tested the hypothesis of whether the LV dilation may be predominantly associated with CCA enlargement, whereas an increase in LV relative wall thickness may be predominantly associated with CCA stiffening, independent of age and BP.
All hypertensive patients referred to our Hypertension Unit between June and December 1992 were considered eligible. Patients were included if they presented mild to moderate uncomplicated essential hypertension. Hypertension was defined as mild to moderate by a supine auscultatory diastolic BP of 95 to 114 mm Hg on three consecutive visits.14 Essential uncomplicated hypertension was confirmed by standardized explorations performed during a day of hospitalization. Patients were checked for standard biology, plasma catecholamine levels, plasma renin activity, 24-hour urinary sodium, and catecholamine metabolite excretion. Renal angiography was performed when there was a strong clinical and paraclinical suspicion of renovascular hypertension. Atherosclerosis was assessed clinically by examination (prominent vascular bruits) and paraclinically by Doppler or echo Doppler examination of lower limb and cervical arteries. Coronary arteries were considered to be involved when the patient suffered from angina pectoris with a positive systematized myocardial scintigraphy and/or a positive coronography. The number of arterial territories affected by atherosclerosis for each patient was estimated. BP was considered well controlled if casual BP was consistently below 140/90 mm Hg. In case of discrepancies, we took into account the opinion of the referring clinician usually in charge of the patient. Patients were excluded if cardiac and/or carotid echography were not realizable or if they had significant valvular disease, transmural heart infarction, extensive bilateral carotid atherosclerosis, or secondary hypertension. The 86 hypertensive patients finally included were all ambulatory and in relative good health. All these patients underwent CCA echotracking measurements and echocardiography during a day of hospitalization. Three patient groups were distinguished for further analysis according to previous treatment history and without any overlap between the three groups. Group 1 was composed of 31 hypertensive patients who had never received any antihypertensive medication. Group 2 consisted of 31 treated patients in whom antihypertensive treatments were withdrawn (30±30 days previously; range, 8 to 90 days) to allow etiologic evaluation. Group 3 was composed of 24 patients still taking antihypertensive medication who had been referred for evaluation of target-organ damage independent of the objective degree of BP control. The protocol was approved by our institutional Ethics Committee; all patients gave informed consent before participating.
BP was measured by an auscultatory method with a mercury sphygmomanometer and adapted cuff at the level of the brachial artery after patients had rested 15 minutes in the sitting position. Korotkoff phases I and V were taken as systolic and diastolic BP values, respectively. Three consecutive measurements were performed at 5-minute intervals, and the average was recorded. These BP values were used in the correlation analysis.
BP was measured every 2 minutes at the forearm with an automatic oscillometric device (Dinamap model 845, Critikon) during cardiac and arterial measurements. Pulse pressure was calculated as the difference between systolic and diastolic BP values obtained simultaneously with arterial and cardiac measurements and was used for the calculation of arterial distensibility and compliance.
All included patients had undergone conventional cardiac echography with a standard commercially available device (Hewlett-Packard Sonos 100) equipped with a 2.25-MHz probe. All measurements were performed in the Echocardiography Department of the hospital by three experienced investigators whose measurement reproducibility has been published elsewhere.15 LV dimensions were measured from M-mode tracings obtained under two-dimensional guidance. Parasternal scans were used. If unavailable, transventricular subcostal scans were used. Measurements were performed post hoc according to American Society of Echocardiography recommendations by a single investigator blinded to the patient’s group. Measurements were made on three cycles on each of two different acquisitions, and the average was retained. The following parameters were measured: anterior wall thickness (AWT), posterior wall thickness (PWT), LV end-diastolic diameter (LVEDD), and LV end-systolic diameter (LVESD). Body surface area (BSA) was used for indexation of cardiac parameters. The following parameters were calculated according to Gaasch,11 Popp and Harrison,16 and Devereux et al17 :
Systolic function was assessed through the shortening fraction and the relation between end-systolic circumferential stress18 (method of Reichek) and shortening fraction. The following formulas were used:
where SBP is systolic BP.
We chose to analyze cardiac geometry and hypertrophy parameters as continuous variables rather than dichotomous ones because classification of LVH patterns must be based on large healthy subject samples matched to hypertensive patients. Our unit lacks such a population, and the use of published data (most of them originate from North America) is of limited interest because of obvious morphological differences between North American and European people. In the following, according to the predefined patterns of LVH,9 10 we will consider that LVMi evaluates LVH as a whole, M/V ratio evaluates LV remodeling and concentric hypertrophy, and LVEDVi is an estimate of LV dilation.
Vessel wall motion was measured by an original pulsed ultrasound echotracking system (Wall-Track system) developed to measure the wall motion of superficial large arteries after echographic localization. A complete detailed description of this system has been published previously.13 19 20 Because of the accurate determination of the Doppler frequency (phase-locked echotracking), a stereotactic device is not necessary for obtaining reliable measurements. Briefly, this system allows the transcutaneous assessment of arterial wall displacement during successive cardiac cycles and hence of the time-dependent changes in arterial wall diameter relative to its initial diameter at the start of the cardiac cycle. Based on the two-dimensional B-mode image, an M line perpendicular to the longitudinal and transverse axes of the artery is selected 2 cm below the CCA bifurcation. The radiofrequency signal over four to eight cardiac cycles is digitized and temporarily stored in a large memory. Two sample volumes, selected under cursor control, are positioned on the anterior and posterior walls, respectively. The vessel walls are continuously tracked by sample volume according to phase. Then, the displacement of the arterial walls is obtained by autocorrelation processing of the Doppler signal originating from the two sample volumes. The accuracy of the system is 30 μm for diastolic diameter (Dd) and less than 1 μm for stroke change in diameter (Ds−Dd).
Repeatability of this method was investigated in 10 subjects through calculation of the repeatability coefficient (RC) as defined by the British Standard Institution,21 according to the formula
where N is the sample size and Di the relative difference between each pair of measurements. The 95% confidence interval of the expected difference was calculated as ±1.96 RC. RC values for the long-term intraobserver repeatability (comparison of two determinations obtained at 1-month intervals by the same observer) concerning CCA diameter and pulsatile changes were 0.312 and 0.025 mm, respectively. These values were small compared with the mean values of CCA diameter and its pulsatile changes in this sample (6.55 and 0.447, respectively). The same procedure was applied for CCA distensibility and compliance. RC values for distensibility and compliance were 0.01 kPa−1 · 10−1 and 0.65 m2 · kPa−1 · 10−7, respectively. These values were small compared with the mean values of CCA distensibility and compliance in this sample (0.18 kPa−1 · 10−1 and 6.26 m2 · kPa−1 · 10−7, respectively).
Measurement were performed by the same observer after patients had rested 15 minutes in the recumbent position. The measurement site was the right CCA, 2 cm beneath the carotid bifurcation. In the presence of small plaques, the contralateral carotid was used.
The following parameters were selected to characterize CCA geometry and function. CCA dimension was assessed through the end-diastolic luminal cross-sectional area (LCSA) as (Π×Dd2)/4. Arterial stiffness parameters were derived from the analysis of the pressure-volume relationship. Volumic distensibility and compliance of a hollow structure were defined as Compliance=ΔV/ΔP and Distensibility=ΔV/(V×ΔP), where V is arterial volume, ΔV is change in volume, and ΔP is change in pressure. It is assumed that the increase in volume of an artery segment is almost exclusively caused by an increase in radius because elongation is negligible in vivo.22 Thus, arterial compliance and distensibility can be estimated through the variation in arterial cross-sectional area. The above relations could be rewritten as follows13 19 :
where PP is pulse pressure.
Since cardiac parameters used in this study are expressed in units of volume (LVMi, LVEDVi, M/V ratio), we chose to express arterial parameters in cross-sectional units (assimilable to volume units if one assumes that the length of arteries changes little during cardiac cycle in vivo22 ): luminal cross-sectional area and distensibility and compliance. Distensibility is the compliance normalized for the luminal cross-sectional area. Distensibility is the inverse of the Peterson modulus and therefore approximates the elastic modulus of the arterial wall when the assumption of a thin-walled tube is made.23 CCA compliance represents the buffering function of the artery when the change in volume in response to the change in pressure is expressed as an absolute value.
Raw data were stored in a microcomputer database. Calculations of arterial- and cardiac-derived parameters were postponed until the final analysis. Final data were transferred to ncss software24 for statistical evaluation. Data are expressed as mean±SD. Qualitative data were compared with the χ2 test. Group analysis was performed with ANOVA. Multiple robust weight regression analysis, concluded by a stepwise regression analysis, was used to assess the determinants of cardiac and arterial parameters. Discrete variables (gender and group) were used as dummy variables. All parameters correlated with a value of P<.1 were included in the stepwise regression procedure. The same procedure was used for study of the relationship between cardiac and arterial parameters after adjustment for age, MBP, BSA, gender, and group. Cardiac parameters were considered as the independent variable in the model. Univariate correlations are presented for information. The analysis was concentrated on the multivariate correlations.
Patient characteristics are presented in Table 1⇓. Groups 1 and 2 were very similar according to age, sex ratio, body size, and cardiovascular risk factors. MBP and pulse pressure were not significantly different in the two groups, indicating that the 30 days of withdrawal of antihypertensive treatment allowed BP to return to untreated patient levels. Group 3 patients were 10 years older than group 2 and group 1 patients. Their BP levels were lower, reaching nearly normal values (systolic BP, 154±26 mm Hg; diastolic BP, 89±16 mm Hg), but the large standard deviations indicate significant variability in these objective measures of hypertension control. Pulse pressure did not differ between the three groups. Hypertension was considered well controlled in a higher percentage of patients in group 3 (55%) than in group 2 patients (33%) (before treatment withdrawal) (P<.05), probably because of a more aggressive treatment regimen (1.84 versus 1.39 antihypertensive classes per patient, P<.05). Smoking habit, body surface area, total and LDL cholesterol levels, and 24-hour sodium excretion did not differ in the three groups. Fasting blood glucose was higher in group 3 patients. Atherosclerosis (as previously defined) was three times more frequent in group 3 patients than in groups 1 and 2 patients (0.85 versus 0.29 and 0.24 arterial territories presenting at least one atherosclerotic lesion, respectively, P<.05). Serum creatinine was higher in group 3 than groups 1 and 2 (106±39 versus 88±13 and 85±20 μmol · L−1, respectively, P<.05), a difference that disappeared after adjustment for age (ANCOVA).
CCA parameter values are presented in Table 2⇓. Carotid LCSA was similar in groups 1 and 2. Despite lower BP values, group 3 had higher carotid end-diastolic diameter and LCSA than groups 1 and 2, a difference that remained significant after age adjustment, but to a lesser extent. CCA distensibility and compliance did not differ between the three groups.
Table 3⇓ presents values of LV structural and functional parameters by patient groups. Among the several published LVH definitions, we chose cut points at 110 g · m−2 in women and 125 g · m−2 in men.25 On these values, LVH was relatively frequent in the entire population (23/86, 27%): 8/23 women (35%) versus 15/63 men (23%) (P=NS). LVEDVi did not differ between the three groups. Group 3 had higher LV wall thickness than groups 2 and 1, involving preferentially the anterior wall. Group 3 exhibited higher LVMi than groups 1 and 2 (118 versus 107 and 100 g · m−2, respectively, P<.05), whereas no difference in M/V ratio was noted between the three groups.
LV shortening fraction and end-systolic stress did not differ significantly between the three groups. As expected, we observed a negative correlation between shortening fraction and end-systolic stress (r=−.78, P<.0001) for the entire population. Group, gender, and age did not influence this correlation, confirming the lack of difference in systolic function between the three patient groups.
CCA Geometry and Function
Univariate determinants of CCA parameters are presented in Table 4⇓, and multivariate determinants of CCA parameters are presented in Table 5⇓. Carotid LCSA was positively correlated to age (r=.50, P<.0001) and BP (r=.43, P<.0001). Gender also influenced CCA dimensions; independently of body size, women had smaller carotids than men (P<.01). The influence of patient group was confirmed by this analysis (P<.01). This model explained 59% of the total variance.
CCA distensibility was strongly correlated to age. A negative correlation was observed (r=−.61, P<.0001), and this parameter alone explained 34% of the total variance. MBP was negatively correlated to CCA distensibility (r=−.26, P<.001). A positive correlation was observed between CCA distensibility and BSA (r=.29, P<.001). Group and gender did not influence CCA distensibility. The model explained 50% of the total variance.
CCA compliance was negatively correlated to age (r=−.22, P<.02) and MBP (r=−.18, P=.04). The major correlate of CCA compliance was BSA (r=.50, P<.0001). Group and gender did not influence CCA compliance. The predictive value of the model was poor compared with that of CCA distensibility because the parameters selected (age, MBP, gender, and group) explained only 20% of the total variance instead of 59%.
Univariate determinants of LV parameters are presented in Table 6⇓, and multivariate determinants of cardiac parameters are presented in Table 7⇓. LVEDVi was negatively correlated to age (r=−.32, P<.01) and marginally influenced by group. Group 3 had larger LVEDVi (P=.04) than groups 1 and 2. MBP and gender were not correlated with LVEDVi. The model explained only 6% of the total variance.
Mean wall thickness was mainly correlated to age (r=.47, P<.0001), which explained 13% of the total variance. Mean wall thickness was significantly correlated with BSA and MBP (r=.30, P<.02; r=.22, P=.03, respectively). Mean wall thickness was significantly but marginally associated with gender and treatment, with women having thinner mean wall thickness values than men and group 3 patients having thicker mean wall thickness values than groups 1 and 2 patients (P=.04, P=.05, respectively). The model explained 22% of the total variance.
LVMi was positively correlated to MBP (r=.48, P<.0001). Women had smaller LVMi values than men (P<.01). This analysis confirmed higher LVMi levels for group 3 patients than for groups 1 and 2 patients. LVMi was not correlated to age. The model explained 25% of the total variance.
M/V ratio was positively correlated to age (r=.54, P<.0001). Women had lower M/V ratios than men. M/V ratio was not correlated with MBP, BSA, or group. The model explained up to 22% of the total variance.
Cardiac and Vascular Interaction
All correlations are presented in Table 8⇓ after adjustment for age, MBP, gender, BSA (when nonindexed cardiac parameters were used), and group.
LVEDVi was positively correlated to carotid LCSA (r=.46, P<.0001) and CCA compliance (r=.47, P<.0001). The univariate correlation between LVEDVi and carotid LCSA was r=.27, P<.02. When LCSA was normalized to age, MBP, gender, and group according to the parameters provided by the multivariate model, a highly significant correlation existed (r=.38, P<.001) and is represented in Fig 1⇓. No correlation was observed between LVEDVi and CCA distensibility.
A highly significant correlation was observed between LV mean wall thickness and CCA distensibility and compliance (r=−.68, P<.0001; r=−.40, P<.0001, respectively). The univariate correlation between mean wall thickness (indexed on BSA) and CCA distensibility is represented in Fig 2⇓. CCA distensibility represents 70% of the total variance explained by this model (40%). No correlation was observed between mean wall thickness and common carotid LCSA.
LVMi was positively correlated to carotid LCSA (r=.23, P=.02), negatively correlated to CCA distensibility (r=−.26, P=.01), and not correlated to CCA compliance.
M/V ratio was negatively correlated to CCA distensibility (r=−.46, P<.0001) and CCA compliance (r=−.37, P<.001). No correlation was observed between M/V ratio and carotid LCSA.
Table 9⇓ presents the evolution of the cardiac multivariate models after the introduction of arterial parameters. In the case of LVEDVi, adjunction of carotid LCSA and compliance resulted in a threefold increase of total variance explained (6% up to 24%). CCA distensibility and compliance improved the power of mean wall thickness and M/V ratio models by 50% (22% to 39.7% and 22% to 33.3% of total variance, respectively). The predictivity of the LVMi model was not improved by the introduction of carotid LCSA and distensibility, but these two parameters decreased the weight of gender and group.
The main result of the present study is that cardiac geometry was strongly correlated to geometric and functional parameters of the CCA in ambulatory hypertensive patients. The main determinant of LVEDVi was carotid LCSA (positive correlation), whereas that of mean wall thickness was CCA distensibility (negative correlation). In addition, CCA distensibility was a major determinant of M/V ratio, which reflects LV remodeling and concentric hypertrophy. Thus, each component of arterial compliance (volume and volumic distensibility) was preferentially associated with a specific pattern of cardiac hypertrophy.
Consideration of Methods
The echotracking method used for the measurement of vessel wall displacement has been validated13 19 20 26 27 and shown to be repeatable enough to be used in cross-sectional studies. In addition to vessel wall displacement, pulse pressure is required for the calculation of distensibility and compliance coefficients (see “Methods”). In the present work, pulse pressure was obtained at the site of the brachial artery and therefore may not accurately reflect the pulse pressure amplitude at the site of the CCA. Amplification of pulse pressure from central arteries to a more distal one is a well-known phenomenon attributable to pulse wave reflection.3 28 Indeed, CCA pulse pressure can be measured by applanation tonometry, but this methodology was not available for all patients at the time of this study. We compared (data not shown) CCA distensibility calculated with brachial pulse pressure (DISTba) with CCA distensibility calculated with CCA pulse pressure (DISTCCA) in a sample of 58 hypertensive patients and healthy subjects (n=40 and 18, respectively) age-matched to the patients of the present study (49±16 years). A very good agreement existed between DISTCCA and DISTba (r=.86, P<.0001), but the slope differed significantly from 1 (1.22, P<.05), showing a 22% underestimation of CCA distensibility when brachial pulse pressure is used. This value is in agreement with amplification ratio values published by O’Rourke et al.28 We considered that because DISTba and DISTCCA were significantly correlated, DISTba could be used in this particular case of a cross-sectional study.
Another important issue is whether the behavior of the CCA, determined locally by our methodology, may be representative of the behavior of proximal compliant arteries during aging and hypertension. The classic work of Learoyd and Taylor,29 an autopsy study, supports the idea that age-related changes were similar in the CCA and thoracic aorta. In a recent study26 that used the same methodology used herein, we observed a parallel increase in the stiffness of the CCA and abdominal aorta with aging. In addition, we reached the same conclusion that the age-induced increase in arterial stiffness was exaggerated by hypertension, both by studying the CCA distensibility locally and by measuring the pulse wave velocity along the carotid-femoral arterial segment.30 More generally, the investigation of various superficial arteries using echotracking systems indicates that the behavior of the CCA with aging and hypertension resembles that of the thoracic and abdominal aortas, whereas there are strong discrepancies between the CCA and more distal arteries, such as the common femoral, brachial, and radial arteries.26 27 31
In the present study, we included both never-treated patients and treated patients to provide representative samples of the various categories of hypertensive patients consulting at our outpatient clinic. Treated patients were further subdivided according to previous treatment history: patients in whom antihypertensive treatment was withdrawn to allow etiologic evaluation (group 2) and patients continuing their antihypertensive medication while target-organ damage was being evaluated (group 3). These different patient groups were taken into account during the entire statistical analysis, using either ANOVA or multivariate correlations.
Interpretation of Findings
Our results outline the role of age, gender, and BP on large-artery geometry and function and are in agreement with numerous previous studies.3 19 20 26 27 29 Age was the major determinant of arterial parameters. With aging, arteries progressively enlarge and stiffen. The age-induced dilation of large arteries has been attributed to various factors, including the fatiguing effects of cyclic stress causing fracture of the load-bearing elastin fibers.3 The influence of age on compliance was less important than on distensibility, mainly because the decrease in CCA distensibility was compensated for by an increase in carotid LCSA. The effect of BP paralleled that of aging, with high BP being associated with CCA enlargement and stiffening independent of age.
Consistent with numerous previous studies,32 33 34 an increase in mean wall thickness and M/V ratio was observed with aging in our patients. Interestingly, these changes occurred independently of BP, gender, and group, suggesting that they were true effects of aging.
In the present study, independent of age and group, MBP was a main determinant of LVMi, whereas no correlation was found with M/V ratio. Indeed, significant correlations between MBP and LVMi are more often found than with concentric hypertrophy or remodeling indexes (see Reference 3535 for review), although an increase in characteristic aortic impedance, which may better represent afterload,4 has been reported to be significantly correlated to M/V ratio but not LVMi.3 Furthermore, characteristic aortic impedance largely depends on the mechanical behavior of proximal compliant arteries, and there is evidence that the M/V ratio is correlated with the carotid-femoral pulse wave velocity.6
The main result of this study is that, independently of age, arterial pressure, gender, and treatment group, LVEDVi was positively correlated to carotid LCSA, whereas LV wall thickness was negatively correlated to CCA distensibility. Consequently, the M/V ratio (expressing LV remodeling and concentric hypertrophy) was strongly correlated to CCA distensibility, whereas LVMi (expressing LVH as a whole) was only weakly correlated to the carotid LCSA and distensibility. There thus appears to be a “volume effect” and a “wall effect” involving both the left ventricle and the CCA. These correlations may be explained on the basis of cardiovascular coupling mechanics but could also result from parallel changes in heart and carotid geometry. Before discussing these two hypotheses, we must stress that we did not intend to determine CCA wall thickness in the present study. Rather, we focused on CCA function, through the determination of cross-sectional distensibility and compliance. Data from previous work27 had strongly suggested that the decreases in CCA distensibility and compliance in hypertensive patients were mainly due to the increased distending BP and that hypertension-induced vascular hypertrophy tends to increase distensibility for a given BP value, therefore acting as a compensatory factor. For these reasons, we hypothesized that the different patterns of cardiac hypertrophy may be more likely correlated with CCA function and internal diameter than with CCA wall thickness.
In general, in humans and animals, the cardiac response to systemic hypertension associates dilation and concentric hypertrophy, whereas pure pressure overload leads to pure concentric hypertrophy and volume overload to LV dilation.36 This suggests that pressure and volume overload may be associated in promoting LVH in hypertensive patients. Ganau et al37 considered that LV dilation could be representative of subclinical LV dysfunction and, because a higher LV mass is needed to bear the same load, could lead to LVH by itself. From another point of view, increased sodium intake has been hypothesized as a potential factor leading to volume overload in hypertension. Indeed, strong arguments exist for the influence of sodium intake on the development of LVH in both animal models38 and humans.39 40 In rat renal hypertension, De Simone et al38 found that elevated sodium intake was a stronger stimulus for LVH than BP level, especially through its effect on LV chamber volume. In hypertensive patients and normotensive subjects, Du Caihar et al39 reported a positive correlation between sodium intake and both LVMi and LV internal dimension. However, this latter correlation was only observed in normotensive subjects. Investigators of the Treatment of Mild Hypertension Study (TOMHS) found a significant correlation between sodium intake and LVMi but none with LV internal dimension. In the present work, we did not observe any correlation between 24-hour sodium excretion and LVMi or LVEDVi.
On the other hand, the positive correlation between LVEDVi and carotid LCSA suggests the existence of a mechanical link between these parameters. Indeed, arterial dilation leads to an increase in the volume of blood contained in arteries, a phenomenon probably exacerbated by arterial lengthening with aging. The pressure-flow relationship in large arteries (hence impedance) is modified by changes in cross-sectional area of arteries.41 42 At any given frequency, the inertial effects are greater if the caliber of the artery is increased. For low frequencies (including the fundamental frequency), inertial effect of blood predominates in large arteries, which carry most of the energy.41 Thus, larger blood-filled arteries require the heart to accelerate blood against larger inertial forces when systole begins. From the standpoint of chronic cardiac adaptation, this excess of work may act as volume overload and may represent one of the mechanisms by which hypertension leads to eccentric hypertrophy.
The M/V ratio, which expresses the concentric hypertrophy and remodeling typical of pressure overload, was correlated to CCA distensibility but not to MBP. Thus, CCA distensibility may represent the true pressure load “seen” by the ventricle better than MBP. Large-artery distensibility is a major determinant of the input impedance of the arterial system. Arterial distensibility determines not only the first minimum of impedance modulus (Windkessel model) but also wave reflections. Decreased distensibility, in response to aging and/or hypertension, causes early wave reflections and a subsequent increase in pulse pressure at the aortic root. This issue has been investigated in healthy individuals by Saba et al8 and in patients with end-stage renal disease by Marchais et al,43 who found a highly significant correlation between LVMi and the augmentation index, an estimate of wave reflections.
In conclusion, our results indicate that changes in lumen diameter and distensibility of the CCA accompany geometric changes in the left ventricle. More specifically, they indicate that a reduction in distensibility paralleled concentric hypertrophy and remodeling (reflected by an increase in the M/V ratio), suggesting that CCA distensibility is a valuable index of the pressure load “seen” by the ventricle. On the other hand, we showed that an increase in arterial volume paralleled increased LV cavity size, suggesting that larger blood-filled arteries may act as volume overload.
We particularly thank Prof René Gourgon for the pertinent advice he gave for this article and Martin Day for his excellent review of the manuscript.
Reprint requests to Stéphane Laurent, MD, PhD, Service de Pharmacologie, Hôpital Broussais, 96, Rue Didot, 75674 Paris Cedex 14, France.
- Received July 18, 1994.
- Revision received October 3, 1994.
- Accepted November 29, 1994.
Casale PN, Devereux RB, Milner M, Zullo G, Harshfield G, Pickering T, Laragh J. Value of echocardiographic measurement of left ventricular mass in predicting cardiovascular morbid events in hypertensive men. Ann Intern Med. 1986;105:173-178.
Nichols WW, O’Rourke M. MacDonald’s Blood Flow in Arteries. London, England: Edward Arnold; 1990:283-329.
Milnor W. Arterial impedance as ventricular load. Circ Res. 1975;36:565-569.
Merrillon JP, Motte G, Fruchaud J, Masquet C, Azancot I, Guiomart A, Gourgon R. Relation between physical properties of the arterial system and left ventricular performance in the course of ageing and arterial hypertension. Eur Heart J. 1982;3:95-102.
Koren M, Devereux R, Casale P, Savage D, Laragh J. Relation of left ventricular mass and geometry to morbidity and mortality in uncomplicated essential hypertension. Ann Intern Med. 1991; 114:345-352.
Roman MJ, Saba PS, Pini R, Spitzer M, Pickering TG, Rosen S, Alderman MH, Devereux RB. Parallel cardiac and vascular adaptation in hypertension. Circulation. 1992;86:1909-1918.
Hoeks APG, Brands PJ, Smeets FA, Reneman RS. Assessment of distensibility of superficial arteries. Ultrasound Med Biol. 1990; 16:121-128.
Popp RL, Harrison DC. Ultrasonic cardiac echography for determining stroke volume and valvular regurgitation. Circulation. 1970;41:493-502.
Reichek N, Wilson J, St John Sutton M, Plappert TA, Goldberg S, Hirshfeld JW. Noninvasive determination of end systolic stress: validation of the method and initial application. Circulation. 1982;65:99-108.
Laurent S, Arcaro G, Benetos A, Laflèche A, Hoeks A, Safar M. Mechanism of nitrate-induced improvement on arterial compliance depends on vascular territory. J Cardiovasc Pharmacol. 1992; 19:641-649.
Benetos A, Laurent S, Hoeks AP, Boutouyrie P, Safar ME. Arterial alterations with aging and high blood pressure: a noninvasive study of carotid and femoral artery. Arterioscler Thromb. 1993;13:90-97.
British Standard Institution. Precision of Test Methods. 1: Guide for the Determination of Reproducibility for a Standard Test Method. London, England: British Standard Institution; 1979. BS 5497, part 1.
Dobrin P. Vascular mechanics. In: Shepherd JT, Abboud FM, eds. Handbook of Physiology, Section 2: The Cardiovascular System, Volume III, Peripheral Circulation and Organ Blood Flow. Bethesda, Md: American Physiological Society; 1983:80-83.
Peterson LH, Jensen RE, Parnel J. Mechanical properties of arteries in vivo. Circ Res. 1960;8:622-639.
Hintze JL. NCSS, 05-02-1987, User Manual. Kaysville, Utah; 1987.
Devereux R, Casale P, Klingfield P, Eisenberg R, Miller D, Campo E, Alonso D. Performance of primary and derived M-mode echocardiography measurements for detection of left ventricular hypertrophy in necropsied subjects and in patients with systemic hypertension, mitral regurgitation and dilated cardiomyopathy. Am J Cardiol. 1986;57:1388-1393.
Boutouyrie P, Laurent S, Benetos A, Girerd X, Hoeks A, Safar ME. Opposite effect of ageing on distal and proximal large arteries in hypertensives. J Hypertens. 1992;10(suppl 6):S87-S91.
Laurent S, Caviezel B, Beck L, Girerd X, Billaud E, Boutouyrie P, Hoeks A, Safar M. Carotid artery distensibility and distending pressure in humans. Hypertension. 1994;23(part 2):878-883.
O’Rourke M, Blazek JV, Morreels CL, Krovetz LJ. Pressure wave transmission along the human aorta. Circ Res. 1968;23:567-579.
Learoyd BM, Taylor MG. Alteration with age in the viscoelastic properties of human arterial walls. Circ Res. 1966;18:278-292.
Asmar R, Benetos A, London G, Hugues C, Weiss Y, Laloux B, Safar M. Aortic distensibility in untreated and treated hypertensive patients. J Hypertens. 1994;12(suppl 3):S70. Abstract.
Laurent S, Girerd X, Mourad JJ, Lacolley P, Beck L, Boutouyrie P, Mignot JP, Safar M. Elastic modulus of the radial artery wall is not increased in patients with essential hypertension. Arterioscler Thromb. 1994;14:1223-1231.
Devereux R, Pickering T. Relationship between the level, pattern and variability of ambulatory blood pressure and target organ damage in hypertension. J Hypertens. 1991;9(suppl 8):534-538.
Swynghedauw B. Hypertrophies et Insuffisances Cardiaques. INSERM, ed. London, England: John Libbey Eurotext London; 1990.
Ganau A, Devereux R, Pickering T, Roman M, Schnall P, Santucci S, Spitzer M, Laragh J. Relation of left ventricular hemodynamic load and contractile performance to left ventricular mass in hypertension. Circulation. 1990;81:25-36.
De Simone G, Devereux R, Camargo M, Wallerson D, Laragh J. Influence of sodium intake on in vivo left ventricular anatomy in experimental renovascular hypertension. Am J Physiol. 1993; 264:H2103-H2110.
Du Caihar G, Ribstein J, Daures JP, Mimran A. Sodium and left ventricular mass in untreated hypertensive and normotensive subjects. Am J Physiol. 1992;263:H177-H181.
Liebson PR, Grandits G, Prineas R, Dianzumba S, Flack J, Cutler JA, Grimm R, Stamler J. Echocardiographic correlates of left ventricular structure among 844 mildly hypertensive men and women in the Treatment of Mild Hypertension Study (TOMHS). Circulation. 1993;87:476-486.
Gow G. Circulatory correlate: vascular impedance, resistance, and capacity. In: Shepherd JT, Abboud FM, eds. Handbook of Physiology, Section 2: The Cardiovascular System, Volume III, Peripheral Circulation and Organ Blood Flow. Bethesda, Md: American Physiological Society; 1983:356-358.
Marchais S, Guerin A, Pannier B, Levy B, Safar M, London G. Wave reflection and cardiac hypertrophy in chronic uremia. Hypertension. 1993;22:876-883.