Clinical evidence to support active regulation of left ventricular compliance in ventricular hypertrophy may be found in the beneficial responses to calcium channel blockers [5, 8, 9]. Several studies have found that diastolic pressure can be reduced at similar or increased chamber volumes, suggesting improved distensibility. However, other more direct assessments of the change in diastolic dimension for a given pressure change (compliance) failed to find such drug effects [10, 11]. Part of this controversy stems from the methods traditionally used to evaluate diastolic passive properties. This analysis is usually based on pressure-dimension (or volume) data measured throughout the filling period from one beat. During initial filling, however, data may be influenced by active relaxation, whereas later filling pressures can be offset by factors extrinsic to the ventricle such as the pericardium and right heart [12-14].

To improve this assessment, we developed an approach using invasive catheter measurements of continuous and simultaneous ventricular pressure-volume data. Cardiac function curves are obtained by combining pressure-volume data from several beats during mechanical obstruction of inferior vena caval inflow [15-17] (Figure 1). Late-to-end diastolic points from each of the beats (lower boundary, Figure 1 are combined to measure passive compliance, and end-systolic pressure-volume points [upper left boundary] provide a measure of contractile function (Figure 1). We used this approach to test whether ventricular chamber compliance can be acutely changed by pharmacologic agents in patients with symptomatic ventricular hypertrophy. Both short-acting ß-receptor blockade (esmolol) and calcium channel blockade (verapamil) were used to vary two common pathways that can influence active calcium handling. Evidence supporting prominent drug influences on active processes by both agents was obtained by examining simultaneous changes in systolic contractility.

View larger version (24K): [in this window] [in a new window]   Figure 1. Measurement of systolic and diastolic chamber properties from pressure-volume relations loops and relations in humans. Left ventricular pressure-volume data are recorded continuously by conductance catheter during gradual mechanical reduction of inferior vena cava inflow. Several cycles are combined to generate systolic (upper left-hand corners) and diastolic (lower right-hand corners) pressure-volume boundaries. The systolic boundary is fit by linear regression and its slope (end-systolic elastance) provides a relatively load-insensitive measure of chamber contractility. This slope has units of millimeters of mercury per milliliter and was 3.02 mm Hg/mL in this patient. The lower boundary, defined by points taken from late diastole, is also fit by linear regression. The slope of the diastolic boundary is also elastance; however, more commonly the inverse slope (compliance) is reported (Delta V/Delta P). This slope was 2.29 mL/mm Hg for this patient.

Methods

Top Methods Results Discussion Author & Article Info References

Fourteen patients with ventricular hypertrophy were studied. Eight of the patients were considered to have primary hypertrophic cardiomyopathy because they had no history of clinically significant hypertension. At the time of the study, arterial systolic and diastolic pressures averaged 129.5 ± 30 and 74.1 ± 14 mm Hg (mean ±SD), respectively, for the group overall. Eleven of 14 patients had dyspnea on exertion (New York Heart Association Class II-IV), 4 of whom had recurrent severe pulmonary edema requiring intubation. The remaining three patients had syncope or presyncope during isometric exercise, or chest pain, or both. Five patients had predominately septal hypertrophy and the remaining nine had concentric hypertrophy. Intracavitary pressure gradients were present in six patients, with a peak value of 87.6 ± 33.1 mm Hg. The patients' ages averaged 50.2 ± 15 years (range, 21 to 73 years). Hypertrophy was documented by echocardiography (wall thickness > 14 mm in all patients), with maximal diastolic wall thickness averaging 18.8 ± 4.5 mm. All but one of the patients also met electrocardiographic criteria for ventricular hypertrophy. Patients were studied in the Cardiovascular Diagnostic Laboratories of The Johns Hopkins Medical Institutions (n = 10), and the Cardiology Division, Veterans General Hospital, Taipei, Taiwan (n = 4).

Six patients either were not receiving long-term medications or had their medications withdrawn for more than 1 week before the study. Of the remaining patients, three were treated long-term with calcium channel blockers (nifedipine or verapamil), four with ß-blockers (metoprolol, atenolol, or propranolol), six with diuretics (furosemide or hydrochlorothiazide), four with vasodilators (enalapril, lisinopril, or prazosin), and one with disopyramide. All these medications, with the exception of diuretics, were held for at least 48 hours before the study.

Contrast ventriculography showed that all patients had normal coronary arteriography and trivial to no mitral regurgitation. Informed consent was obtained from all patients, and the study protocol was approved by both the Joint Committee on Clinical Investigation of The Johns Hopkins Medical Institutions and by the Human Investigation Committee of the Veterans General Hospital, Taipei, Taiwan.

Procedure

The method for determining human pressure-volume relations has been previously reported in detail [16, 17]. All patients first had routine right and left heart catheterization. A multielectrode conductance (volume) catheter was advanced to the left ventricular apex [18]. A low-amplitude, high-frequency current was applied to electrodes located at the left ventricular base and apex, and resistances measured at multiple intervening electrodes along the catheter provided a time-varying signal proportional to intracavitary chamber volume. A micromanometer (PC-330A, Millar Inc.; Houston, Texas), placed within the lumen of the volume catheter, provided simultaneous high-fidelity ventricular pressure. In the six patients with intracavitary pressure gradients, a separate micromanometer was positioned at the cardiac base to measure chamber pressure. A custom-designed, large balloon occlusion catheter (SP-9159, Cordis; Miami, Florida) was placed in the right atrium. Rapid and reversible lowering of cardiac filling was produced by balloon inflation with 10 to 20 mL of carbon dioxide and simultaneous gentle withdrawal into the proximal inferior vena cava.

Pressure-volume data were determined under four conditions: 1) baseline; 2) intravenous esmolol [Brevibloc, DuPont Pharmaceuticals; Manati, Puerto Rico], 3 mg/kg body weight loading over 15 minutes, followed by 10 minutes at 300 µg/kg per minute; 3) second baseline [at least 30 minutes after stopping esmolol]; and 4) intravenous verapamil (Isoptin Knoll Pharmaceuticals; Whippany, New York), 10 mg. Esmolol data were obtained in 11 patients and subsequent verapamil data in 9 patients. An additional three patients received verapamil only. In those patients receiving both drugs, the order was not randomized because of the relatively long half-life of verapamil.

Analysis

Catheter volumes were calibrated to a contrast left ventriculogram (single plane, right anterior oblique projection), using a two-point calibration based on matching end-diastolic and end-systolic volumes. Ventriculogram volumes were estimated from the frames of maximal and minimal area, respectively. For the catheter signal, these volumes were averages of several volume points during isovolumic contraction and relaxation periods (sides of the pressure-volume loops). The calibrated catheter signal was then used for all subsequent data analyses.

Steady-state hemodynamic variables included end-diastolic and end-systolic pressures and volumes, ejection fraction, maximal rate of pressure increase (dP/dt_max), peak filling rate and time to peak filling rate, and relaxation time constant. These were calculated from signal averaged data, using 5 to 10 sequential cardiac cycles. Details of most of the variable estimation methods have been previously reported [17, 19]. The relaxation time constant was derived from linear regression of the natural logarithm of pressure versus time [20]. To determine peak filling rate, the volume waveform was curve fit (4-term Fourier series) and differentiated to yield peak flow. The result was divided by end-diastolic volume [21]. Time to peak filling was measured from end-systole (point of maximal chamber elastance) to the time of initial maximal filling.

Chamber systolic and diastolic function was indexed by end-systolic elastance (E_es) and end-diastolic compliance (C_ed) (see Figure 1). The end-systolic elastance was determined from the set of points [one per cardiac cycle] for which the instantaneous ratio of pressure/(volume-V_o) was maximal [16, 22]. These data were fit by linear regression, with Pearson correlation coefficients averaging 0.97. The end-diastolic compliance was determined from the set of pressure-volume points at end-diastole, obtained from the multiple cardiac cycles measured at varying filling volume. These diastolic relations are often treated as nonlinear; however, we have found that linear fits generally describe the relations well [19].

We first divided each relation into upper and lower pressure/volume ranges, applying linear regression to each half. Multiple regression analysis revealed no significant difference between these two sets of relations, and therefore a single linear fit (mean correlation coefficient, 0.89) was applied to the full measured data range and is reported. The inverse slope of this relation was C_ed. In the rare instance in which atrial systole influenced the late diastolic pressure-volume curve, data measured just before atrial contraction were used, ensuring that the points defined the lowest boundary of the pressure-volume loops and thus reflected passive chamber properties.

Statistical Methods

Data are presented as mean ± standard deviation. Steady-state parameters are compared by the Student paired t-test, with respective drug data contrasted to their immediately previous baseline values. The two baseline periods were also directly compared in nine patients. The E_es and C_ed were compared by t-tests. In addition, a multiple regression model was used in which all the raw pressure-volume data for each relation (before and after drug) were used. Drug effects were coded by a dummy variable (0 = baseline, 1 = drug), as was interpatient variability [23]. This approach is similar to covariance analysis, with all the patient data combined into a single regression model. Analysis was performed on a 386-processor personal computer, using both custom-developed software and commercial statistical software (SYSTAT; Evanston, Illinois).

Results

Top Methods Results Discussion Author & Article Info References

Mean hemodynamic responses before and after esmolol and after verapamil are provided in Tables 1 and 2, respectively. Esmolol slowed heart rate by 10% and significantly lowered cardiac output, ejection fraction, and the maximal rate of pressure development (dP/dt_max) (all P < 0.01). End-systolic volume increased by 18%, whereas end-diastolic volume remained unchanged. These responses are consistent with depressed contractility. Diastolic effects were manifest by a slowing of early peak filling rate (PFR/EDV) from 4.4 to 3.8 sec^–1 (P < 0.05). Relaxation time constant was also prolonged by 19%, consistent with known effects of ß-blockade on calcium re-uptake. The pressure at the onset of filling increased slightly from 12.4 to 15.6 mm Hg (P < 0.05), reflecting both the delay in relaxation and the increase in end-systolic volume. Verapamil did not alter heart rate, cardiac output, or ejection fraction. Contractility was lowered, however, as indicated by a reduced dP/dt_max despite similar end-diastolic chamber volume. The lack of change in cardiac output, or ejection fraction, was due to simultaneous –12.3%± 4.4% decrease in peripheral resistance (P < 0.05). Diastolic pressures at the onset of filling and at end-diastole were unchanged.

View larger version (34K): [in this window] [in a new window]   Figure 2. Pressure-volume systolic and diastolic relations before and after administration of esmolol and of verapamil. Data are from one patient and were generated in a similar fashion to those shown in Figure 1. Baseline systolic and diastolic pressure-volume relations are reproduced in each corresponding lower panel (solid lines). The left-hand panels show baseline and esmolol data, and the right-hand panels show second baseline and verapamil data. Both drugs lowered contractility, as shown by the reduced slope of the end-systolic pressure-volume relation (dashed line). In contrast, they had minimal effect on diastolic compliance, indicated by the similar slope of the diastolic pressure-volume relation.

The mean results for systolic pressure-volume relation analysis are shown in Table 3. Both esmolol and verapamil reduced E_es similarly ( –27.7%± 25.4% and –31.8%±-11.8%, respectively; P < 0.01). This is consistent with a decrease in chamber contractility, as well as systolic chamber stiffness. In contrast, neither drug significantly altered diastolic chamber compliance. Figure 3 displays data from three "example" patients before and after esmolol (left panels) and verapamil (corresponding right panels). Individual points show raw data taken from multiple pressure-volume loops (at late-to-end diastole) measured during preload reduction. These data were virtually superimposable over the measured data range.

View larger version (20K): [in this window] [in a new window]   Figure 3. Minimal influence of esmolol or verapamil on end-diastolic pressure-volume relation. End-diastolic pressure-volume data before and after esmolol (upper panels, left, middle, and right) and verapamil (lower panels, left, middle, and right) are shown for three typical patients. The individual points were derived from the latter 30% of chamber filling, using the multiple pressure-volume loops measured during transient preload reduction (as shown in Figures 1 and 2). These diastolic data reflect passive chamber properties and were not altered by either drug. Group compliance data are provided in Table 3. = baseline 1; = esmolol; = baseline 2; = verapamil.

The group compliance analysis is also provided in Table 3, showing the linear analysis applied over the full volume range. The mean compliance was 3.7 ± 1.9 mL/mm Hg at baseline, lower than values reported in normal persons [19]. Direct, paired comparisons of compliance before and after esmolol or verapamil revealed no significant change. These values were measured at virtually identical diastolic pressures (mean change in pressure from either drug was approximately 1 mm Hg, P > 0.2). Multivariate regression (covariance analysis) was used to test for differences in the pressure-volume data before and after drug was given and confirmed the lack of change from verapamil. With esmolol, this analysis suggested a small ( –8.1%)reduction in compliance.

Because the study group included six patients with a history of hypertension, five patients with predominant septal hypertrophy (versus concentric hypertrophy), and seven with intracavitary pressure gradients, we examined these subgroups separately to determine if they responded differently. The results showed no significant difference in contractile, compliance, relaxation, or early peak filling response when comparing those patients with each of these respective characteristics to the corresponding group without them. In patients with intracavitary pressure gradients, both drugs reduced gradients by an average of 33% ± 40% (P < 0.05). However, the extent of gradient reduction did not correlate with quantitative changes in early relaxation, peak filling rate, or compliance.

Discussion

Top Methods Results Discussion Author & Article Info References

Our primary goal was to test the extent to which chamber compliance can be altered by acute pharmacologic interventions in patients with marked ventricular hypertrophy. We hypothesized that if active processes responsible for relaxation and resting diastolic tone played a prominent role in determining chamber compliance, then compliance should change in the short term from ß-blockers or calcium channel blockers, which are known to alter these processes. The results showed almost 30% reductions in systolic stiffness and contractility from both agents, consistent with substantial active effects. However, neither drug altered diastolic chamber compliance or the diastolic pressure-volume relation. Both drugs had some substantial effects on early diastolic properties, such as relaxation or filling rates, or both; however, these changes did not translate into altered compliance.

Beta-blockers and calcium channel blockers are commonly used to treat hypertrophic cardiomyopathy and severe ventricular hypertrophy with hypertension. Studies have documented favorable diastolic effects of each drug on isovolumic relaxation and early chamber filling rates [5, 8, 9, 21, 24-27]. Beta-blockers also slow the heart rate, prolonging the time available for chamber filling [28, 29]. Calcium channel blockers are thought to favorably influence intracellular Ca²⁺ handling thereby speeding relaxation, lowering chamber resting tension, and improving chamber distensibility [5, 8, 9, 30, 31]. The clinical evidence for this last effect, however, is inconsistent [10, 11] and varies among patients within a given study [5, 8]. When present, the typical response is a nearly parallel downward displacement of the diastolic pressure-volume relation [5, 8, 9]. To some extent, this effect can be seen in the verapamil data for the patient in Figure 2. However, parallel upward or downward shifts in this relation do not necessarily indicate changes in the left heart; they can also stem from alterations in right heart and pericardial pressures [12, 13, 17]. For example, similar downward shifts are reported during acute nitrate administration [13, 14]. Direct evidence supporting an acute increase in compliance (that is a greater change in volume for a given change in diastolic pressure) is scant. Indeed, our data indicate there is no such change.

The inferior vena caval inflow obstruction maneuver for determining the diastolic pressure-volume curve (EDPVR), used in this and several recent studies from our laboratory [17, 19], represents a novel approach to diastolic compliance analysis in humans. One major advantage over traditional single-beat measures is that it takes advantage of the series nature of right-left heart (and pericardial) interaction by unloading the right heart before the left. By the time the left ventricular volume starts to decrease, right atrial pressure is often near zero, and peak right ventricle pressures have decreased by 40% to 50% [17, 19]. This unloading can substantially remove positive pressure offsets caused by the right heart and pericardium. In addition, unlike traditional single-beat analysis, EDPVR analysis is confined only to late diastole, minimizing confounding effects of early relaxation and rapid filling and thus more likely representing passive chamber properties.

One possible concern is that venous inflow obstruction, used to measure the EDPVRs, might alter ventricular diastolic properties. However, relaxation time constants (_RELAX) were not significantly altered even at maximal preload reduction (change of _RELAX = +2.3 ± 8.0 ms; P > 0.2). Further, the volume range of the EDPVRs nearly always fell well within volumes for the resting cardiac contraction (see Figure 1) so these data were obtained at physiologic volumes. Another concern is that the doses of each drug used in our study were simply insufficient to induce an alteration in chamber compliance. However, both drugs had substantial effects on contractility and, in particular, the end-systolic elastance [E_es], which indexes systolic chamber stiffness. Further, both drugs had some (albeit variable) effects on early diastolic relaxation and filling. Baseline relaxation was slightly prolonged and was further delayed by esmolol. Verapamil shortened the time to peak filling and had variable effects on relaxation time (some shortening was observed in 8 of the 12 patients). Despite these changes, passive chamber properties were not altered.

The notion that the diastolic pressure-volume relation is predominately determined by passive structural elements under minimal active and thus pharmacologic control is supported by several earlier studies. For example, Hess and colleagues [10] examined compliance responses using pressure-dimension data (the latter determined by echocardiography) in hypertrophic patients before and after the administration of verapamil. They found no significant compliance change. However, they also found no evidence for significant systolic effects from the drug (for example, dP/dt_max did not change), and thus it was unclear if this finding was caused by a drug-dosing problem. This was less true of an earlier study [32] in which dP/dt_max did decrease by 24% after verapamil, yet there was also no change in compliance. These and all other previous studies used single beats at a given chamber volume for analysis. Our study more accurately assessed these chamber properties by excluding early filling phases, relying on late diastolic pressure-volume data from multiple cardiac cycles, and minimizing extrinsic loading factors.

Although we examined acute pharmacologic effects, it remains possible that oral administration of similar agents over the longer term could influence chamber compliance. Long-term calcium channel blockade is known to lower ventricular mass [33]; however, whether this is associated with a change in chamber compliance remains unknown.

Slightly more than one half the patients in the study group had primary hypertrophic cardiomyopathy, whereas the remaining had at least a history of hypertension. Most of the patients in the latter group had continued to develop marked hypertrophy despite medical therapy for hypertension and had visited a physician because of symptoms of exertional dyspnea and acute pulmonary congestion despite normal systolic function. Subgroup analysis performed to test whether the drugs had influenced these two groups differently did not show significant differences. The sample size is admittedly small for such analysis, however, and the possibility remains that some underlying differences may exist.

The conductance catheter technique for measurement of pressure-volume relations in humans has been evaluated by several groups including our own [16, 18, 34, 35]. It does not provide an absolute signal and therefore must be calibrated to some other "gold standard" volume. Its advantage is that, once calibrated, it provides a continuous analog voltage that is proportional to chamber volume with reasonable (although not perfect) accuracy. Our signals were calibrated to ventriculography at steady state. Studies have found that conductance catheter volumes are almost linearly proportional to chamber volume over the ejecting range. This includes the volumes of the EDPVRs, and thus the steady-state calibration should be quite accurate for these relationships.

We conclude that in humans with symptomatic left ventricular hypertrophy, chamber compliance and the end-diastolic pressure-volume relation is not acutely altered by pharmacologic agents that simultaneously reduce systolic chamber stiffness and lower contractility. This supports the theory that the relation between cardiac diastolic pressures and volumes late in diastole is primarily determined by passive or structural elements, rather than active calcium control. The results do not indicate that current treatment of hypertrophic cardiomyopathy or hypertensive hypertrophy with calcium and ß-receptor blockers is without benefit. Rather, they suggest that active relaxation and contractile depression mechanisms play a more prominent role in their efficacy. The extent to which chamber compliance can be altered in humans in the long term still remains largely unknown, but the data we have presented suggest that structural adaptations, reflected by changes in chamber size, collagen content, and mass, are most important. Therapies that specifically target components of the extracellular space in hypertrophied hearts seem more likely to directly affect passive chamber compliance.