Differential Effects of Chronic Oral Antihypertensive Therapies on Systemic Arterial Circulation and Ventricular Energetics in African-American Patients
Background A comprehensive evaluation of arterial load characteristics and left ventricular energetics in systemic hypertension has been limited by the need for invasive techniques to access instantaneous aortic pressure and flow. As a consequence of this methodological limitation, no data exist on the effects of long-term antihypertensive therapy on global arterial impedance properties and indexes of myocardial oxygen consumption (MV̇o2). Using recently validated noninvasive techniques, we compared in hypertensive patients the effects of chronic oral treatment with ramipril, nifedipine, and atenolol on arterial impedance and mechanical power dissipation as well as indexes of MV̇o2.
Methods and Results Sixteen African-American subjects with systemic hypertension were studied with a randomized, double-blind, crossover protocol. Instantaneous central aortic pressure and flow, from which arterial load characteristics can be derived, were estimated from calibrated subclavian pulse tracings (SPTs) and continuous-wave aortic Doppler velocity in conjunction with two-dimensional (2D) echocardiographic measurements of the aortic annulus, respectively. To derive ventricular wall stress and indexes of MV̇o2, left ventricular short- (M-mode) and long-axis (2D echo) images were acquired simultaneously with SPTs. Data were collected at the end of a 2-week washout period (predrug control) and after 6 weeks of treatment with each agent. Although all three agents reduced diastolic blood pressure to the same extent, different effects on mean and systolic pressures and vascular impedance properties were noted. Nifedipine reduced total peripheral resistance (TPR; 1744±398 versus 1290±215 dyne-s/cm5) and increased arterial compliance (ACL; 1.234±0.253 versus 1.776±0.415 mL/mm Hg). This improvement in arterial compliance was not entirely accounted for by the reduction in distending pressure. Ramipril also decreased TPR (1740±292 versus 1437±290 dyne-s/cm5) and increased ACL (1.214±0.190 versus 1.569±0.424 mL/mm Hg), but with this agent, the change in arterial compliance was explained solely on the basis of a reduction in distending pressure. Atenolol, in contrast, did not affect either TPR or ACL. In agreement with the compliance results, nifedipine and ramipril significantly lowered the first two harmonics of the impedance spectrum, but atenolol did not. None of these agents resulted in a significant change in characteristic impedance or in the relative amplitude of the reflected pressure wave. Total vascular mechanical power and percent of oscillatory power remained unaltered with all antihypertensive treatments. Only ramipril and nifedipine reduced the integral of both meridional and circumferential systolic wall stresses, indicating that MV̇o2 per beat was reduced with these agents. Stress-time index, a measure of MV̇o2 per unit time, decreased significantly with ramipril but not with nifedipine because of an increase in heart rate noted in 10 of 16 patients (mean increase, 10 beats per minute). Thus, a reduction in MV̇o2 coupled with unchanged total vascular mechanical power suggests improved efficiency of ventriculoarterial coupling with ramipril and with nifedipine in the subset of patients in whom heart rate remained unchanged. In contrast, there was no evidence of a reduction in wall stress, stress integral, or stress-time index with atenolol.
Conclusions The noninvasive methodology used in this study constitutes a new tool for serial and simultaneous evaluation of arterial hemodynamics and left ventricular energetics in systemic hypertension. In this study, we demonstrate the differential effects of chronic antihypertensive therapies on systemic arterial circulation and indexes of MV̇o2 in African-American subjects. Consideration of drug-induced differential responses of arterial load and indexes of MV̇o2 with each drug may provide a more physiological approach to the treatment of systemic hypertension in individual patients.
Systemic hypertension is characterized by complex alterations in arterial hemodynamics, including elevated total peripheral resistance and characteristic impedance, reduced arterial compliance, and abnormal propagation/reflection of pressure and flow waves.1 2 3 These pulsatile and steady components of arterial load impede left ventricular ejection and may adversely affect left ventricular energetics and ventriculoarterial coupling.4 Despite the high prevalence of hypertension, few studies have examined the long-term effects of different antihypertensive agents on any of these parameters.1 5 The paucity of such longitudinal studies is largely attributable to the absence of noninvasive methods for acquisition of instantaneous aortic pressure and flow required for the computation of pulsatile load characteristics and indexes of left ventricular energetics.
We recently developed and validated a new noninvasive method to estimate instantaneous aortic flow and pressure, based on simultaneous recordings of continuous-wave aortic Doppler and calibrated subclavian pulse tracings, respectively.6 In addition, by combining echocardiographic measurements of left ventricular geometry with calibrated subclavian pulse tracings, it is possible to obtain indexes of left ventricular energetics noninvasively.7
This study was designed to test the hypothesis that long-term oral therapy with different antihypertensive agents elicits differential responses in arterial load and left ventricular energetics that are not predictable from analysis of traditional hemodynamic variables. Accordingly, we studied the effects of 6-week treatment with three widely used classes of antihypertensive agents in a cohort of African-American patients with moderate systemic hypertension: an angiotensin-converting enzyme (ACE) inhibitor (ramipril), a calcium channel antagonist (nifedipine), and a β-blocker (atenolol).
Seventeen African-American patients with essential moderate hypertension were enrolled. This group included 11 men and 6 women ranging in age from 33 to 60 years (mean±SD, 49±8 years). Because of the lack of increased left ventricular mass in our study group (see below), we believe that the duration of hypertension was relatively short, although precise information regarding this issue is unavailable. Patients either were previously untreated (n=3) or had antihypertensive therapy discontinued for 2 weeks before enrollment into the study (n=14). Moderate hypertension was defined as duplicate measurements of diastolic blood pressure ranging between 90 and 110 mm Hg obtained at two consecutive clinic visits at least 1 week apart. Exclusion criteria included (1) previous history of stroke, (2) clinical or ECG findings suggestive of coronary artery disease, (3) segmental wall motion abnormalities on two-dimensional (2D) echocardiography, (4) Doppler echocardiographic evidence for significant valvular heart disease, (5) significant cardiac arrhythmias, (6) asthma or any other contraindication to the administration of atenolol, nifedipine, or ramipril, and (7) renal failure. The protocol was approved by the Institutional Review Board of the University of Chicago Medical Center, and each patient provided written informed consent before entering the study.
A double-blind, randomized, three-way crossover study design (Fig 1⇓) was used to investigate the responses to ramipril, an ACE inhibitor; a long-acting preparation of nifedipine (dihydropyridine calcium channel antagonist); and atenolol (β-adrenoceptor blocker). Each agent was given as a single daily oral dose for 6 weeks. Doses were increased at 2-week intervals (5, 10, and 20 mg for ramipril; 30, 60, and 90 mg for nifedipine; and 25, 50, and 100 mg for atenolol) to achieve a diastolic blood pressure <90 mm Hg. Each 6-week treatment cycle was followed by a 2-week washout period (during which placebo was administered) before initiation of the subsequent agent as dictated by the randomization protocol. In each subject, data were acquired at six different visits: at the end of each 2-week washout period (predrug control) and after 6 weeks of therapy with each agent (Fig 1⇓).
Data Acquisition and Processing
Patients underwent two sequential phases of data acquisition. First, aortic pressure and flow data were recorded to obtain arterial characteristics, followed by echocardiographic imaging to determine left ventricular mechanics and energetics.
Aortic Pressure-Flow Data
The following data were acquired simultaneously: ECG, phonocardiogram, continuous-wave aortic Doppler, and calibrated subclavian pulse tracing. In addition, measurements of aortic annular diameter were obtained with 2D echocardiography at each visit. The ECG, phonocardiogram, and subclavian pulse tracing were connected to the echocardiographic unit (Sonos 1500, Hewlett-Packard Inc) via a custom-made multichannel physiological signal module. An example of these data is shown in Fig 2⇓.
Noninvasive instantaneous aortic pressures were determined from subclavian pulse tracings obtained with a plastic funnel positioned over the right subclavian artery at its point of maximal impulse in the supraclavicular fossa and connected by Silastic tubing to a strain-gauge transducer (model 03040170, Cambridge Instrument Co).6 The waveform was recorded by a physiological pulse wave recording system (bandwidth, 0.05 to 50 Hz) and was calibrated according to the proximal brachial artery pressure determined with an oscillometric sphygmomanometer (model 1846 SX, Dinamap Vital Signs Monitor, Critikon Inc). Systolic and diastolic pressures were assigned to the peak and nadir of the subclavian pulse tracing, respectively, and instantaneous pressures over the cardiac cycle were derived by linear interpolation.
Noninvasive estimates of instantaneous aortic flow were generated from continuous-wave Doppler velocity recordings and 2D echocardiographic measurements of aortic annular area. Ascending aortic blood velocity was recorded with a 1.9-MHz continuous-wave Pedof Doppler transducer (Hewlett Packard Inc) positioned at the cardiac apex. The aortic annular diameter (Dao, cm) was measured from parasternal long-axis 2D echocardiographic recordings obtained with a 2.5-MHz transducer by the trailing-to-leading-edge technique. Annular cross-sectional area (CSA, cm2) was calculated assuming a circular orifice:
Instantaneous aortic flow (QAo, mL/s) was obtained as the product of the instantaneous Doppler blood flow velocity and CSA.8
Three cardiac cycles were selected for analysis according to the following criteria for beat selection6 : (1) the magnitude of the baseline shift in diastolic pressure was <10% of the pulse pressure; (2) the Doppler envelope was well demarcated, and peak velocities for the three cycles were within 10% of each other; and (3) the cardiac rhythm was sinus with interbeat RR interval variations of <15%. The pressure trace and flow velocity envelope were digitized at 200 Hz on a digitizing tablet (Bit Pad Two, Summagraphics Corp) and stored on a personal computer (Gateway 2000, 486/33 C). The propagation delay between the ascending aorta and the subclavian artery was calculated as the time interval between the first high-frequency component of the second heart sound (A2) on the phonocardiogram and the incisura of the pressure waveform. The digitized pressure tracings were offset accordingly. Pressure and flow data of the three cardiac cycles were then averaged and subsequently processed with custom software for computation of hemodynamic parameters.
Left Ventricular Echocardiographic Imaging
The following data were acquired simultaneously: ECG, phonocardiogram, calibrated subclavian pulse tracing, and 2D targeted M-mode echocardiogram (Fig 3⇓, top). Care was taken to record the largest left ventricular minor-axis dimension (D) at the level of the superior aspect of the papillary muscles (equator of the ventricle). In addition, the left ventricular long-axis dimension (L) was measured from the apical four-chamber view as the distance between the apex and the mitral annulus (Fig 3⇓, bottom). The apical endocardium was defined as the point within the left ventricle at which the septum and lateral wall form the sharpest angle. To avoid tangential imaging of the left ventricle, both L and the mitral annulus dimensions were maximized. End-diastolic dimensions and wall thickness were measured at the time of the R wave on the ECG, and end-systolic measurements were performed at the time of the first high-frequency component of A2 on the phonocardiogram.
Instantaneous left ventricular dimension, wall thickness, and aortic ejection pressure were digitized by use of the off-line data analysis system described above. Left ventricular mass as well as meridional and circumferential wall stress during cardiac ejection were derived.
Parameters calculated were heart rate (HR, beats per minute); mean aortic pressure (MAP, mm Hg), calculated as the total area under the calibrated subclavian pressure waveform divided by cardiac cycle length; stroke volume (SV, mL), calculated as the product of the aortic flow velocity–time integral and the aortic cross-sectional area; cardiac output (CO, L/min), calculated as stroke volume multiplied by heart rate; and left ventricular shortening fraction (FS), calculated as (Ded−Des)/Ded, where Ded is end-diastolic diameter and Des is end-systolic diameter.
Arterial Load Characteristics
Total peripheral resistance (TPR, dyne-s/cm5) was calculated as mean aortic pressure divided by cardiac output. The aortic input impedance spectrum was obtained for each condition by use of Fourier transformation of the pressure and flow signals.9 Characteristic impedance (Zc, dyne-s/cm5) was calculated as the average of high-frequency impedance harmonics (harmonics 4 through 10).1 To minimize the effect of noise, only those harmonics with flow amplitude of >5% of the first harmonic were included in the averaging process. Arterial compliance (AC, mL/mm Hg) was calculated according to the area method described by Liu et al10 for both linear (ACL) and nonlinear (exponential, ACNL) models of the arterial pressure-volume relation. The compliance estimation was performed over the diastolic range defined by two points, P1 and P2. The maximum aortic pressure after the dicrotic notch was denoted as P1, while P2 was the end-diastolic aortic pressure.
Ad is defined as the area under the aortic pressure curve bounded by P1 and P2, and P represents the pressure at which ACNL is calculated. The coefficient b, which characterizes large arteries, is relatively independent of vessel tone3 10 and was assumed to be a constant (−0.01 mm Hg−1) for all patients under all conditions.3
Unlike TPR, which completely characterizes the arterial load to steady flow, the load for pulsatile flow cannot be quantified in terms of a single number. Several indexes were used for this purpose, each quantifying different aspects of pulsatile arterial load: characteristic impedance (Zc), the magnitudes of the first (Z1) and second (Z2) harmonics of the impedance spectrum, and arterial compliance (ACL or ACNL).
To quantify wave reflection and propagation, two indexes were used. Pressure and flow waveforms were decomposed into forward and backward components according to the linear transmission line theory.11 The first reflection index (RI1) was the ratio of backward to forward pressure wave peak-to-peak amplitudes. The second reflection index (RI2) was calculated as the amplitude of the first harmonic of the global reflection coefficient (ΓG), derived from the aortic input impedance spectrum.11 12 Relative to control, a higher value of RI1 or RI2 indicates a greater contribution of wave reflections to measured pressure and flow waveforms.
External Mechanical Power
Total power (Ẇtot, watts), steady power (Ẇstd, watts), oscillatory power (Ẇosc, watts), and percent oscillatory power (%Ẇosc) were calculated from instantaneous aortic pressure [P(t)] and flow [Q(t)] as follows13 :
where T is the cardiac cycle duration and P and Q are the mean pressure and flow, respectively. Ẇosc represents the energy expended in pulsatile phenomena, which does not contribute to net forward flow.
Left Ventricular Mass Index
Left ventricular mass index (LVMI, g/m2) was calculated by use of the formula described by Devereux et al.14
Left Ventricular Meridional and Circumferential Wall Stresses
Left ventricular meridional wall stress (ςm, g/cm2) was computed throughout ventricular ejection according to the following angiographically validated formula15 :
where D and h are left ventricular short-axis diameter and thickness, respectively.
Left ventricular circumferential stress (ςc, g/cm2) was calculated by use of the formula of Sandler and Dodge16 :
where L is left ventricular long-axis dimension. Since L was measured at three discrete points during systole, ςc was computed at the onset of ejection (time of the initial pressure upstroke), at the time of peak meridional stress, and at end systole using the corresponding values of P, D, L, and h.
Indexes of Myocardial Oxygen Consumption
Integrals of left ventricular meridional and circumferential systolic wall stress (∫ςm and ∫ςc, respectively, g-s/cm2) were computed as the areas under the respective stress curves during the ejection period. These integrals correlate with myocardial oxygen consumption (MV̇o2) per beat.17
Meridional and circumferential left ventricular stress-time indexes (STIm and STIc, respectively, g-s/cm2-min), were calculated as the product of heart rate and ∫ςm and ∫ςc, respectively. These indexes are related to MV̇o2 per unit time.18 19
For each variable, data were compared by one-way repeated-measures ANOVA. When assumptions of normality or equal variance were violated, a nonparametric test (Friedman repeated-measures ANOVA on ranks) was used. If differences among treatment groups were found to be statistically significant (P<.05), pairwise multiple comparisons were performed by the Student-Newman-Keuls method.
Of the 17 patients initially enrolled into the study, one had to be excluded because of inadequate subclavian pulse tracings. Thus, aortic pressure-flow relations were studied in 16 patients. Left ventricular echocardiographic measurements were obtained only in patients who had technically adequate recordings at all six visits (n=12). All patients required the highest dose of each antihypertensive agent (ie, ramipril 20 mg, nifedipine 90 mg, and atenolol 100 mg) to achieve diastolic blood pressures ≤90 mm Hg, with the exception of patient 12, who became normotensive with 50 mg of atenolol and 5 mg of ramipril. A representative data set of pressure and flow tracings obtained in 1 subject before and after 6 weeks of treatment with ramipril, nifedipine, and atenolol is shown in Fig 4⇓.
All hemodynamic parameters returned to baseline values after the 2-week washout periods. Hemodynamic effects of ramipril, nifedipine, and atenolol on heart rate, blood pressure, stroke volume, and cardiac output are presented in Table 1⇓. Heart rate remained unchanged with ramipril but decreased with atenolol (−19%, P<.05). With nifedipine, a variable heart rate response was observed: 10 of 16 patients increased heart rate (mean increase of 10 beats per minute). The reduction in heart rate with atenolol was associated with an increase in stroke volume that was not observed with the other agents. As a result, cardiac output was unchanged with all three drugs. Systolic blood pressure decreased most with nifedipine and least with atenolol, but all three drugs reduced diastolic blood pressure to a similar extent.
The effects of each agent on TPR, ACL, and Zc are summarized in Table 2⇓, and the individual patient data are depicted in Fig 5⇓. TPR decreased with ramipril (P<.05) and nifedipine (P<.05) but not with atenolol. Arterial compliance, calculated with a linear pressure-volume model, increased with both nifedipine (P<.05) and ramipril (P<.05) but was unchanged with atenolol. When a nonlinear model was used, only nifedipine increased ACNL for any level of distending pressure (Fig 6⇓).
The averaged aortic input impedance spectra of the 16 patients before and after each treatment are shown in Fig 7⇓. Significant reductions in impedance moduli of the first (Z1) and second (Z2) harmonics of the impedance spectrum were observed after ramipril and nifedipine but not after atenolol (Table 2⇑). The reduction in Z1 was greater with nifedipine than with ramipril. No change in characteristic impedance was observed with any antihypertensive agent (Table 2⇑).
Although they did not reach statistical significance, the changes in reflection indexes (RI1 and RI2) tended to follow the changes in TPR noted with each agent (Table 2⇑). Neither total power nor percent oscillatory power was significantly affected by any of the three antihypertensive agents.
Left ventricular dimensions and thicknesses, as well as shortening fraction data, are shown in Table 3⇓. Although end-diastolic short- and long-axis dimensions remained unchanged with all drugs, a small but statistically significant reduction in end-systolic dimensions was noted only with ramipril and nifedipine. There was a tendency for the shortening fraction to increase with ramipril and nifedipine and to decrease with atenolol. The values of left ventricular mass index were within the normal range at the onset of the study and remained essentially unchanged after 6 weeks of antihypertensive therapy with either ramipril, nifedipine, or atenolol (Table 3⇓).
Fig 8⇓ illustrates the effects of each antihypertensive agent on circumferential and meridional left ventricular systolic wall stress at three time points: the onset of ejection, time of peak meridional stress, and end systole. Overall, ramipril and nifedipine reduced wall stress throughout ventricular ejection, but atenolol did not. Consequently, the integrals of circumferential and meridional wall stress per beat were reduced only with ramipril and nifedipine (Table 3⇑). Since there was a tendency for heart rate to increase with nifedipine and not with ramipril (Table 1⇑), circumferential stress-time index was reduced only after ramipril (Table 3⇑).
Using recently validated noninvasive techniques, we were able to identify differential hemodynamic responses to three commonly used oral antihypertensive agents in a cohort of hypertensive African-American patients. Despite similar reductions in diastolic blood pressure, nifedipine and ramipril were more effective than atenolol in reducing systolic and mean blood pressures. These effects on blood pressure were associated with differential responses in heart rate, arterial load characteristics, and left ventricular energetics. Both the steady and pulsatile components of arterial load, as well as the integrals of circumferential and meridional wall stress during ventricular ejection, decreased with ramipril and nifedipine but not with atenolol. The circumferential stress-time index, reflecting left ventricular MV̇o2 per unit time, was reduced significantly with ramipril. With nifedipine, this reduction was blunted as a result of increased heart rate noted in some patients. In terms of coupling, the maintenance of constant total vascular mechanical power at a time that MV̇o2 decreased indicated an improved efficiency of ventriculoarterial coupling.
Steady Versus Pulsatile Arterial Load
TPR, a major determinant of arterial load, is the sole parameter used by most clinicians to characterize the systemic arterial circulation. However, this parameter describes only the load or opposition to steady flow. Since arterial pressure and flow are pulsatile, additional vascular parameters should be considered to describe arterial load more accurately. As can be seen from the impedance spectra in Fig 7⇑, the opposition to flow is high for steady flow (impedance at zeroth harmonic or TPR) and rapidly falls as frequency increases. This fall is governed by arterial geometric and viscoelastic properties as well as wave propagation-reflection phenomena. Such a design ensures maintenance of adequate mean blood pressure with minimal opposition to pulsatile ejection. Therefore, it is important to examine both the steady and pulsatile components of arterial load. Although the physical properties and phenomena mentioned above are distributed over the entire vasculature, they are represented in simplified models of the arterial circulation as lumped parameters. For example, arterial compliance, as calculated in this study, incorporates arterial elasticity and geometry as well as wave propagation and reflection phenomena. Characteristic impedance of the aorta, on the other hand, is a function of local aortic material and geometric properties.20 21 The importance of including parameters other than TPR to describe instantaneous aortic pressure-flow relations is illustrated in Fig 9⇓.
Methodological Considerations for the Noninvasive Technique
Clinical application of any pulsatile model of the cardiovascular system has been limited by the requirement for invasive procedures to obtain the necessary instantaneous pressure and flow data. Invasive studies, however, are not suited for the serial assessment of chronic pharmacological interventions. As a result, the potential diagnostic and therapeutic applications of these more physiological models of the circulation are not well appreciated.
We have previously demonstrated that noninvasively acquired subclavian pulse waveforms are morphologically similar to high-fidelity ascending aortic pressure recordings.6 The subclavian pulse tracing was calibrated according to brachial measurements obtained with an oscillometric sphygmomanometer. This method of calibration has been validated against high-fidelity central aortic pressure in children22 and adults.23 The validity of this calibration procedure in hypertensive individuals derives further circumstantial support from the well-described attenuation of pulse wave amplification during chronic hypertension.24
Many investigators have demonstrated that the transaortic continuous-wave Doppler-derived spectral velocity envelope in conjunction with 2D echocardiographic estimation of cross-sectional area at the level of the aortic valve annulus provides an adequate noninvasive estimation of stroke volume.25 We have recently validated the accuracy of transthoracic continuous-wave aortic Doppler against electromagnetic flowmeters for measurements of instantaneous aortic inflow characteristics in both open-chest monkeys and humans.6 8 In these studies, Doppler flows were slightly higher than their electromagnetic counterparts during early ejection and were accurate during the remainder of systole. These minor differences in morphology did not impact significantly on the derived hemodynamic parameters of interest, such as arterial compliance, characteristic impedance, and total peripheral resistance.6
Values for arterial compliance and characteristic impedance are in the range of previously published results in untreated hypertensive patients when invasive techniques were used.1 3 5 Similarly, values for external mechanical power (total, steady, and oscillatory) and reflection indexes are comparable to those reported by Ting et al in hypertensive patients.1
Effects of Antihypertensive Agents on the Arterial Circulation
We found similar reductions in diastolic blood pressure for all three classes of antihypertensive therapy. Interdrug differences were noted for systolic and mean blood pressure responses. Atenolol appeared to be less efficacious than both ramipril and nifedipine in achieving systolic blood pressure reduction, whereas nifedipine was slightly more effective than ramipril. Overall, these differential responses are similar to those reported by Materson et al26 in African-American patients of a similar age group. It has been suggested that the low renin levels reported in African-American patients may account for the reduced antihypertensive efficacy of ACE inhibitors relative to nifedipine.27 28
Both ramipril and nifedipine markedly decreased TPR (parameter of steady arterial load) and slightly increased cardiac output. In contrast, atenolol caused proportional reductions in pressure and flow with no change in TPR for the entire group. Some patients experienced slightly increased TPR with atenolol. Interestingly, these patients were the ones who had low TPR values at baseline. Atenolol, in the doses used in this study, is largely selective for cardiac β1-receptors. Therefore, it is unlikely, yet possible, that this drug would elicit unopposed α-activity resulting in an increase in TPR. Irrespective of the mechanism, the observation of increased TPR with a β-blocker in some hypertensive patients is not unique to our study.1
Likewise, nifedipine and ramipril reduced Z1 and Z2 (parameters of pulsatile load), whereas atenolol did not. Since heart rate remained unchanged, these impedance reductions reflected true vascular changes. The failure of atenolol to alter Z1 and Z2 may be a result of the offsetting effect of the reduction in heart rate. None of the drugs altered characteristic impedance (Zc) significantly. Zc is a function of aortic material and geometric properties as well as distending pressure and is not influenced by the distal vasculature.20 Discordant responses among the multiple determinants of Zc may explain the invariance of this parameter during antihypertensive therapy. The lack of response of Zc to β-blockers has been reported previously in acute studies using invasive techniques after intravenous propranolol.1
The increase in linear arterial compliance observed with nifedipine and ramipril, but not atenolol, corroborates findings obtained in the brachial and forearm arterial circulation after chronic treatment with these agents.29 30 31 32 When an exponential pressure-volume model was used to compare compliance values independent of distending pressure, only nifedipine was found to increase arterial compliance (Fig 6⇑), indicating that the pressure drop induced by this agent does not entirely account for the increase in ACNL. These findings, also reported by Simon et al30 on the brachial circulation after treatment with nicardipine and nitrendipine, have been attributed to dihydropyridine calcium channel antagonist–induced smooth muscle tone reduction. In contrast, the increase in ACL observed after ramipril treatment can be accounted for solely by the reduction in distending pressure. This finding differs from those reported by Simon et al,30 who reported a direct effect of ramipril on arterial compliance and Levy et al,33 who noted an increase in arterial compliance after 8 weeks of oral treatment with ACE inhibitors that was associated with a reduction in medial collagen content of rat carotid arteries. The absence of direct improvement in arterial compliance after ramipril in our patients may reflect the shorter duration of treatment, differences in dosing, or ethnic differences in the populations studied.
Although not statistically significant, the reflection indexes (RI1 and RI2) tended to follow the TPR responses to each antihypertensive agent. This observation is in agreement with the findings of Ting et al,1 who reported no change in wave reflection index after vasodilatation with nitroprusside in hypertensive patients but found an increase after intravenous propranolol, which also increased TPR. It should be noted that the indexes used in this study do not examine timing aspects of wave reflections. Systemic hypertension is a condition in which pulse wave velocity is usually increased, resulting in early return of wave reflections, with potentially deleterious effects on ventriculovascular coupling.2
None of the antihypertensive treatments tested in this study resulted in a significant decrease in total vascu-lar mechanical power (Ẇtot). For ramipril and nifedipine, this was the result of a slight increase in cardiac output that counterbalanced the decrease in pressure. For atenolol, pressure and flow decrements were not sufficient to significantly reduce Ẇtot. The fraction of the total energy dissipated in pulsation (%Ẇosc) was small (<15% of Ẇtot) and was not reduced by any antihypertensive treatment. Conflicting results have been published on whether blood pressure reduction in hypertensive patients decreases %Ẇosc, thereby indicating an improvement in efficiency of power dissipation in the arterial system.1 5 34 Recent theoretical findings, however, indicate that there are multiple determinants of oscillatory power (eg, TPR, pulse wave velocity, arterial compliance).35 Thus, offsetting effects of simultaneous changes in these determinants might account for the lack of change observed in %Ẇosc after ramipril and nifedipine.
Effects of Antihypertensive Agents on the Left Ventricle
At the onset of the study, left ventricular mass index values were within the normal range (Table 3⇑). The lack of noticeable change in left ventricular mass index with all treatments is possibly related to the absence of initial hypertrophy.
Left ventricular circumferential and meridional wall stresses were significantly reduced throughout ventricular ejection after ramipril and, to a lesser extent, nifedipine treatment (Fig 8⇑). This resulted from the combined effects of blood pressure reduction and decreased systolic chamber dimensions with a tendency toward increased systolic wall thickening. In contrast, atenolol tended to increase left ventricular chamber size, which blunted the beneficial effect of pressure reduction in terms of wall stress.
Reduced wall stress and unchanged left ventricular ejection times with ramipril and nifedipine resulted in a decrease in the integrals of circumferential and meridional systolic wall stresses, indexes of oxygen consumption per beat. With atenolol, these integrals increased (meridional) or tended to increase (circumferential) as a result of the prolonged ejection time and slightly elevated wall stress.
Stress-time index, the product of heart rate and integral of systolic wall stress, has been closely correlated to MV̇o2.18 19 Since the majority of myocardial fibers are oriented in near-circumferential direction and since most of the fiber shortening occurs in the same direction, the circumferential stress-time index (STIc) may better reflect changes in MV̇o2. Thus, in this study, alterations in STIc should accurately reflect changes in MV̇o2, provided that left ventricular contractility remained relatively unchanged with all agents. Ramipril significantly reduced STIc because of its beneficial effect on the integral of wall stress without a concomitant increase in heart rate. Although this index tended to decrease with nifedipine, it did not attain statistical significance. Given that the integral of systolic wall stress was reduced with nifedipine, the increased heart rate noted in 10 of 16 patients was responsible for this finding. With atenolol, the stress-time index remained unchanged because of the offsetting effects of decreased heart rate coupled with a trend toward an increase in the integral of systolic wall stress.
In terms of ventriculoarterial coupling, the reduction in myocardial oxygen consumption with unaltered total vascular mechanical power indicates an improvement in the efficiency of the cardiovascular system after treatment with ramipril in African-American hypertensive patients. Similar improvements in efficiency were observed with nifedipine in the subgroup of patients whose heart rate remained unchanged with this agent.
Relevance and Limitations of the Present Study
To the best of our knowledge, there are no human studies that have simultaneously assessed the long-term effects of different antihypertensive agents on pulsatile and nonpulsatile arterial load and left ventricular energetics. Most previous longitudinal studies with antihypertensive agents have examined the effects on the peripheral circulation (eg, brachial and forearm).29 30 31 32 Our noninvasive methodology enables serial assessments of the systemic arterial circulation, ventricular energetics, and ventriculoarterial coupling. Since this study was performed on a selected homogeneous group, ie, African-Americans with moderate hypertension without left ventricular hypertrophy, extrapolation of our results to other hypertensive ethnic populations with and without hypertrophy should be made with caution. However, given that systemic hypertension is extremely prevalent in the African-American community, our results remain relevant.
Summary and Clinical Implications
We have shown the feasibility of noninvasive, serial quantification of arterial load characteristics and indexes of left ventricular energetics. Using these techniques, we were able to identify differential responses to three commonly used antihypertensive agents with respect to pulsatile and steady arterial load, ventricular energetics, and ventriculoarterial coupling. It may now be possible to classify hypertensive patients according to specific ventricular and arterial load profiles that would help select pharmacological therapy in individual patients. Further studies are necessary to determine whether therapeutic strategies based on this physiological approach can reduce mortality in patients with hypertension.
This study was supported in part by grants from the Upjohn Co to Dr Lang, from NIH-NHLBI (HL-36185) to Dr Shroff, and from the French Ministère des Affaires Etrangères (Fondation Lavoisier) to Dr Cholley. We are indebted to Fetima Davis, RN, Linda B. Roberts, RN, and Barbara Bates, RN, for their expert nursing assistance throughout the study. We thank Errol C. Rubenstein and Donna Barrett for their help with the preparation of the manuscript.
- Received June 17, 1994.
- Revision received September 12, 1994.
- Accepted October 2, 1994.
- Copyright © 1995 by American Heart Association
Ting CT, Brin KP, Lin SJ, Wang SP, Chang MS, Chiang BN, Yin FCP. Arterial hemodynamics in human hypertension. J Clin Invest. 1986;78:1462-1471.
O’Rourke M. Arterial stiffness, systolic blood pressure, and logical treatment of arterial hypertension. Hypertension. 1990;15:339-347.
Liu Z, Ting CT, Zhu S, Yin FCP. Aortic compliance in human hypertension. Hypertension. 1989;14:129-136.
O’Rourke MF, Avolio AP, Nichols WW. Left ventricular-systemic arterial coupling in humans and strategies to improve coupling in disease states. In: Yin FCP, ed. Ventricular/Vascular Coupling: Clinical, Physiological, and Engineering Aspects. New York, NY: Springer Verlag; 1986:1-19.
Marcus RH, Korcarz C, McCray G, Neuman A, Murphy M, Borow KM, Weinert L, Bednarz J, Gretler DD, Spencer KT, Sareli P, Lang RM. Noninvasive method for determination of arterial compliance using Doppler echocardiography and subclavian pulse tracings: validation and clinical application of a physiological model of the circulation. Circulation. 1994;89:2688-2699.
Borow KM, Neuman A, Lang RM. Milrinone versus dobutamine: contribution of altered myocardial mechanics and augmented inotropic state to improved left ventricular performance. Circulation. 1986;73(suppl III):III-153-III-161.
Spencer KT, Lang RM, Neuman A, Borow KM, Shroff S. Doppler and electromagnetic comparisons of instantaneous aortic flow characteristics in primates. Circ Res. 1991;68:1369-1377.
Milnor WR. Vascular impedance. In: Hemodynamics. Baltimore, Md: Williams & Wilkins; 1989:167-203.
Liu Z, Brin KP, Yin FCP. Estimation of total arterial compliance: an improved method and evaluation of current methods. Am J Physiol. 1986;251:H588-H600.
Berger DS, Li JKJ, Laskey WK, Noordergraf A. Repeated reflection of waves in the systemic arterial system. Am J Physiol. 1993;264:H269-H281.
Milnor WR. Cardiac dynamics. In: Hemodynamics. Baltimore, Md: Williams & Wilkins; 1989:260-293.
Sandler H, Dodge HT. Left ventricular tension and stress in man. Circ Res. 1963;13:91-104.
Weber KT, Janicki JS. Myocardial oxygen consumption: the role of wall force and shortening. Am J Physiol. 1977;233:H421-H430.
Weber KT, Janicki JS. Interdependence of cardiac function, coronary flow, and oxygen extraction. Am J Physiol. 1978; 235:H784-H793.
Dujardin JP, Stone DN, Paul LT, Pieper HP. Response of systemic arterial input impedance in volume expansion and hemorrhage. Am J Physiol. 1980;238:H902-H908.
Milnor WR. Arterial impedance as ventricular afterload. Circ Res. 1975;36:565-570.
Colan SD, Fujii A, Borow KM, MacPherson D, Sanders SP. Noninvasive determination of systolic, diastolic and end systolic blood pressure in neonates, infants and young children: comparison with central aortic pressure measurements. Am J Cardiol. 1983; 52:867-870.
Borow KM, Newburger JW. Noninvasive estimation of central aortic pressure using the oscillometric method for analyzing systemic artery pulsatile blood flow: comparative study of indirect systolic, diastolic, and mean brachial artery pressure with simultaneous direct ascending aortic pressure measurements. Am Heart J. 1982;103:879-886.
O’Rourke MF. Vascular impedance in studies of arterial and cardiac function. Physiol Rev. 1982;62:570-623.
Huntsman LL, Stewart DK, Barnes SR, Franklin SB, Colocousis JS, Hessel EA. Noninvasive Doppler determination of cardiac output in man: clinical validation. Circulation. 1983;67:593-602.
Cody RJ, Laragh JH, Case DB, Atlas SA. Renin system activity as a determinant of response to treatment in hypertension and heart failure. Hypertension. 1983;5(suppl III):III-36-III-42.
Freis ED, Materson BJ, Flamenbaum W. Comparison of propranolol or hydrochlorothiazide alone for treatment of hypertension: evaluation of the reninangiotensin system. Am J Med. 1983; 74:1029-1041.
Benetos A, Vasmant D, Thiery P, Safar ME. Effects of ramipril on arterial hemodynamics. J Cardiovasc Pharmacol. 1991;18(suppl 2):S153-S156.
Simon AC, Levenson J, Chau P, Pithois-Merli I. Role of arterial compliance in the physiopharmacological approach to human hypertension. J Cardiovasc Pharmacol. 1992;19(suppl 5):S11-S20.
Levy BI, Michel JB, Salzmann JL, Azizi M, Poitevin P, Safar M, Camilleri JP. Effects of chronic inhibition of converting enzyme on mechanical and structural properties of arteries in rat renovascular hypertension. Circ Res. 1988;63:227-239.
Nichols WW, Conti R, Walker WE, Milnor WR. Input impedance of the systemic circulation in man. Circ Res. 1977;40:451-458.
Berger DS. Repeated Reflections and the Effects of Wave Reflections on Arterio-Ventricular Function. New Brunswick, NJ: Rutgers University; 1993. Dissertation.