Role of Atrial Contraction in Diastolic Pressure Elevation Induced by Rapid Pacing of Hypertrophied Canine Ventricle
Abstract The mechanism of diastolic pressure elevation induced by acute rapid pacing in pressure-load hypertrophied left ventricles (LVs) remains incompletely understood. It has been ascribed to abnormalities of coronary flow, metabolism, and calcium cycling. However, rapid pacing also alters the timing of atrial and ventricular stimulation relative to the diastolic filling period, and this could also influence diastolic pressures. To test the role of such mechanical factors, LV pressure-volume hemodynamics were measured in closed-chested anesthetized dogs during and after abrupt cessation of rapid atrial pacing. Twenty-one dogs were studied: 6 dogs with LV hypertrophy (LVH) induced by perinephritic hypertension, 5 sham-operated normotensive dogs, and 10 acute normotensive control dogs. In LVH dogs, but not in sham-operated or control dogs, end-diastolic pressure rose progressively with increasing heart rate from 5.6±3.1 mm Hg at baseline to 22.6±8.1 mm Hg at 220 beats per minute. In all hearts, rapid pacing shifted the timing of left atrial contraction so that it occurred near the onset of LV filling rather than at end diastole. However, in LVH hearts, early LV diastolic pressure and peak atrial pressure were also markedly elevated. Most striking, immediately after terminating the pacing, diastolic pressure declined to near baseline. This rapid pressure decline occurred just when atrial systole would have ensued and before ventricular activation would have followed had pacing continued. Thus, diastolic pressure elevation resolved before a change in ventricular pacing rate. The role of atrial contraction was further explored by simultaneous atrioventricular pacing. This shifted the time of atrial systole so that it occurred during LV isovolumic contraction, while maintaining the identical LV pacing rate. This change eliminated the diastolic pressure elevation found previously. Further analysis revealed that the pressure increase during rapid pacing was not due simply to partial LV filling imposed on a relaxing ventricle or to hypertension or an intact pericardium. These data indicate that mechanical effects of atrioventricular interaction play an important role in tachycardia-induced diastolic dysfunction in this model of LVH and can be more causative than ischemia or metabolic factors in this setting.
Pressure-load left ventricular (LV) hypertrophy (LVH) is associated with abnormalities of diastolic function, including prolonged isovolumic relaxation, depressed peak LV filling rate, and reduced chamber compliance.1 2 3 4 5 With increased heart rate, LVH hearts often display marked elevation of diastolic pressures and stiffness.6 7 8 9 Studies performed in animal models of hypertrophy have also reported pacing-induced abnormalities of high-energy phosphate metabolism7 and relative subendocardial hypoperfusion6 9 10 consistent with reduced coronary flow reserve.11 These changes have been proposed to explain tachycardia-induced elevation of LV diastolic pressure and stiffness. Support for this hypothesis derives from data demonstrating persistence of end-diastolic pressure elevation for seconds after pacing is discontinued.8 10 In addition, studies have reported general correlations between the extent of pacing-induced diastolic dysfunction and increased myocardial lactate extraction9 or a reduced endocardial phosphocreatine to ATP ratio.7
However, there are several potentially important mechanical alterations induced by rapid pacing that have received far less attention. For example, as heart rate increases, atrial contraction occurs earlier in diastole and can become synchronous with the onset of LV filling. This could augment initial diastolic filling pressures and thereby influence pressures throughout a rate-limited diastolic period. Furthermore, metabolic effects may have secondary hemodynamic consequences. For example, prolonged isovolumic relaxation can interact with atrial filling, further elevating pressures.
The present study was undertaken to determine the relative contribution of these mechanical factors during rapid pacing in LVH. Moderate LVH was induced by perinephritic hypertension, and rapid pacing mechanics were studied by using pressure-volume analysis in acutely instrumented intact closed-chested anesthetized animals. To help differentiate metabolic from mechanical mechanisms, we principally focused on the characteristics and rate of diastolic functional recovery after abrupt cessation of pacing. Mechanical alterations are expected to resolve immediately, whereas metabolic factors are expected to resolve more slowly after the termination of chronotropic challenge.
Materials and Methods
Perinephritic Hypertension-Induced LVH
Induction of perinephritis was attempted in 11 mongrel dogs of either sex by unilateral kidney cellophane wrapping, as described by Ferrario et al.12 The procedure was performed through a midepigastric incision by use of sterile technique and sodium pentobarbital anesthesia (25 mg/kg). Two weeks later, the animal was again anesthetized, and the contralateral renal artery was cannulated via femoral arterial access under fluoroscopic guidance. The artery was then embolized by use of a radiopaque slurry of contrast and gel-foam powder. This procedure was followed by placement of two or three small (diameter, ≈1 mm; length, 3 cm) stainless steel coils. Upon release in the artery, these take on a more or less spherical form (diameter, 5 mm). Animals were allowed to fully recover and were maintained in a chronic care facility for 9 to 12 months for development of perinephritic hypertension and ventricular hypertrophy. Arterial pressure was measured in the control condition (initial baseline), 4 to 6 months after renal embolization, and at the time of study via femoral cannula with the animals under pentobarbital anesthesia.
Of the 11 animals prepared in this fashion, 5 failed to develop hypertension and LVH. When autopsies were performed in these animals, the wrapped kidney appeared grossly normal, and the cellophane was separated from the kidney. This group served as a set of “sham” preparations (group II), and these were separately compared with the 6 animals with LVH (group III). In addition to these animals, a third group of 10 normotensive animals of similar age but with no prior surgery were studied as controls (group I). All animals were maintained in accordance with the guidelines for the “Care and Use of Laboratory Animals” of the Institute of Laboratory Animal Resources, National Council (Department of Health and Human Services publication No. [NIH] 85-23, revised 1985). The protocol was approved by the Animal Care and Use Committee of The Johns Hopkins University.
Studies were performed by cardiac catheterization in intact closed-chested dogs subjected to sodium pentobarbital anesthesia.13 The dogs were intubated and maintained on oxygen-enriched positive-pressure ventilation, with respiration held at end expiration for the few seconds required to measure data. Oxygen status was monitored by a pulse oximeter, and adjustments to ventilation were made as required. A midline neck incision was made to expose left and right internal jugular veins and the carotid arteries. Both left and right brachial arteries and femoral vessels were also exposed.
Under fluoroscopic guidance, bipolar pacing leads with passive fixation tips were advanced from the right internal jugular vein to the right atrial appendage and right ventricular apex. A flow-directed thermodilution catheter (No. 7350, Arrow) was advanced to the main pulmonary artery from the left internal jugular vein to enable measurement of cardiac output. A 5F micromanometer-tipped catheter (SPC-350, Millar Instruments) was advanced from the right brachial artery to the ascending aorta, and a similar catheter was placed in the mid-LV cavity via a femoral artery. Micromanometer catheters were presoaked in saline and calibrated against a mercury manometer before placement. A guiding catheter was advanced over a preformed guide wire through the opposite femoral artery, advanced retrograde across both aortic and mitral valves, and positioned in the left atrium. This catheter enabled measurement (by fluid-filled transducer) of left atrial pressure (LAP) and served as a route for injection of radiolabeled microspheres. Comparison with micromanometer pressure confirmed only a trivial phase delay of this fluid catheter, on the order of 5 ms. A second catheter was placed in the left brachial artery to enable arterial blood sampling during microsphere injections. A conductance catheter (Webster Laboratories) with a pigtail tip (electrodes spaced at 1-cm intervals) was advanced via the left carotid artery to the LV apex and connected to a stimulator/processor (Sigma-5, CardioDynamics) to provide LV volume measurement.14 15 The parallel conductance was determined by the hypertonic saline method14 and subtracted from the catheter signal. The gain of the signal was calibrated such that cardiac output at baseline sinus rhythm matched that determined by thermodilution.
All hemodynamic signals, including aortic blood pressure, LV pressure (LVP), LV volume (LVV), LAP, and surface ECG were amplified, digitized at 200 Hz with 12-bit precision, and stored on removable magnetic media by using customized software on an 80386 processor-based microcomputer. This system provided real-time display of signals versus time and pressure versus volume.
Data were recorded during sinus rhythm and atrial pacing. The cycle lengths of atrial pacing were 350, 300, and 275 ms, corresponding to heart rates of ≈170, 200, and 220 beats per minute. Not all dogs could be paced at the fastest rate because of mechanical alternans; and such data were excluded from analysis.
In 6 of the 10 animals in group I and in all of the group II and III animals, ≈2×106 15-μm microspheres labeled with 141Ce, 113Sn, 103Ru, or 95Nb were injected through the left atrial catheter during stable hemodynamics at each pacing rate to allow for the subsequent determination of regional myocardial blood flow. The atria were paced for at least 5 minutes before microsphere injection. Spheres were suspended in 10% dextran mixed with Tween detergent and agitated in a vortex mixer for 20 minutes before injection. Arterial reference blood was simultaneously withdrawn, and myocardial flows were calculated by standard technique.
Once steady state hemodynamic data and microsphere flows were obtained at each heart rate, pacing was abruptly terminated. Data collection included a period of stable hemodynamics before pacing cessation and continued until pressure-volume loops returned to their baseline appearance at sinus rhythm. Two to five pacing runs were performed at each rate, and results were averaged. In 6 dogs of group I, 3 dogs of group II, and 4 dogs of group III, hemodynamics were also studied during atrioventricular (AV) sequential pacing. These pacing runs were performed with cycle lengths of 300 and/or 275 ms and with AV intervals of 120 or 0 ms.
In 4 of the dogs in the control group, atrial pacing was performed during continuous infusion of angiotensin II (10 to 40 ng/kg per minute), titrated to raise systolic arterial pressure to 190 mm Hg during sinus rhythm. This provided an analysis of the contribution of afterload elevation itself on AV pacing mechanics.
Last, in 9 dogs (4 in group I, 3 in group II, and 2 in group III), pacing data were repeated after opening the chest and widely incising the pericardium. These data were obtained to examine the importance of an intact pericardium to the observed hemodynamics during pacing.
Upon completion of the experimental protocol, dogs were euthanatized by lethal injection of sodium pentobarbital. The heart was excised, the atria were removed, and the right ventricular free wall and LV were weighed. The ratio of LV to body weight was determined on the basis of body weight measured at the time of the final acute study. In hearts injected with radiolabeled microspheres, the ventricle was subsequently subdivided to provide multiple endocardial, midwall, and epicardial samples from all regions of the ventricle. The myocardium was weighed, and radioactivity was counted in a scintillation gamma counter. Raw counts were corrected for background and crossover and compared with a reference sample to provide regional flow in units of milliliters per minute per gram. The ratio of endocardial to epicardial flow was calculated by averaging results from all samples within each respective layer and determining the ratio. In animals in which cellophane wrapping/renal artery embolization was performed, both kidneys were excised and examined grossly.
Steady State Pacing
Digitized signals were analyzed off-line on a workstation (SPARCstation, Sun Microsystems) by using custom-developed software. Points of end systole, onset of filling (end-isovolumic relaxation), and end diastole were identified as the upper left, lower left, and lower right corners of the pressure-volume loop, respectively. These points were determined by placing appropriate tangent lines to the loop intersecting at these corners, as previously described.13 16
Average LV end-diastolic pressure and volume (EDP and EDV, respectively), end-systolic volume (ESV), cardiac output, and peak LAP (pLAP) were determined from analysis of 5 to 10 consecutive pressure-volume loops. The time constant of pressure relaxation was determined from the inverse negative slope of the relation between pressure and the first derivative of pressure.17 Pressure data from the point at −dP/dtmax to the onset of filling (lower left corner of the pressure-volume loop) were used for this calculation. Mean ejection midwall fiber stress (αf) was calculated during sinus rhythm by the method of Arts et al,18 given by: αf=3 · LVP/ln (1+[VW/LVV]), where VW is LV wall volume estimated from LV mass.
Upon cessation of rapid pacing, the diastolic pressure-volume relation could either continue along a trajectory similar to that observed during pacing or, alternatively, shift to a different relation (Fig 1⇓, dashed lines). To quantify such a shift, termed ΔPraw, LV pressures were compared at the onset of filling for the first postpaced beat [LVP(ta), where ta is time point a; Fig 1⇓] and at an equivalent time point measured from the preceding QRS complex in the last paced beat [LVP(tb), where tb is time point b]: Although the pressures used to determine ΔPraw were measured at the same relative time in the cardiac cycle, they reflected different LV volumes. Specifically, LVP(tb) occurred after some filling, whereas by definition, LVP(ta) occurred at the end of isovolumic relaxation. Since filling superimposed on a relaxing ventricle can elevate diastolic pressure,19 some shift would be expected from this mechanism alone. To subtract the component of ΔPraw due to this difference in filling, we used the following strategy: The ratio of LVP(t) to LVV(t) was calculated during the early-diastolic period and used as a measure of chamber elastance [E(t)]. Using E(t), one can estimate what the LVP would have been for the paced beats had partial filling not occurred. This is done by multiplying E(tb) by the volume at the end of isovolumic relaxation for the postpaced beat [LVV(ta)]. The difference between this pressure and LVP(ta) provides a measure of pressure decline that cannot be directly explained by differences in chamber filling. This corrected pressure shift (ΔPcor) was thus given by the following: If ΔPraw were entirely due to differences in LV filling superimposed on ongoing pressure relaxation, then ΔPcor would be near zero. A positive ΔPcor, on the other hand, indicates that a portion of ΔPraw results from a change in chamber diastolic stiffness (or elastance). Such altered stiffness could result from viscous behavior or AV mechanical coupling.
Data are given as mean±SD. Baseline parameters for the three animal groups were compared by the Kruskal-Wallis test with the Bonferroni correction for multiple comparisons. Variables studied as a function of pacing rate were first tested by two-way ANOVA, with heart rate and experimental group serving as categorical variables. If this analysis revealed a significant influence of the group, then individual pairwise tests between groups were performed. A Bonferroni correction was applied for multiple comparisons. The effect of heart rate on each variable within a given animal group was tested by two-way ANOVA, treating both animal and heart rate as categorical variables. Individual differences between group means at each heart rate were evaluated by paired t tests using Bonferroni correction. Statistical significance was accepted at the P<.05 level.
Development of Hypertension and LVH
Table 1⇓ provides baseline and follow-up arterial pressures for the chronically prepared animals and pressures, heart rate, chamber and body weights, and heart weight–to–body weight ratio at the time of final acute study. Arterial pressures in groups II and III were similar at baseline, and both groups demonstrated an increase in pressure after 4 to 6 months, with minimal further change measured at the time of final acute study (≈12 months). However, the change in pressure was much higher in group III, and only this group had significantly elevated pressures at the time of acute study when compared with group I.
LV mass was somewhat higher in group III compared with the other two groups, although these differences fell just short of statistical significance (P=.08). However, the LV weight–to–body weight ratio for group III dogs was significantly higher than for the other two groups. There was no significant difference in either body weight itself or right ventricular weight among the three groups. The extent of LVH observed in the group III animals is similar to that reported by other investigators using a similar model2 20 but less than that achieved with chronic aortic banding.6 7 9 Sinus rates at the time of acute study were not different among the three groups.
Atrial Pacing Hemodynamics
Table 2⇓ provides steady state ventricular pressure-volume data at baseline and during rapid atrial pacing for the three animal groups. There was a striking disproportionate rise in EDP and pLAP during rapid pacing in group III (LVH) animals (P<.001 for each). This was associated with a modest decline in EDV (P<.05) without a significant change in cardiac output. The rise in pLAP was interesting, since (as demonstrated below) pLAP did not occur at end diastole and thus could not be directly related to an elevated EDP. In fact, at rapid pacing rates, pLAP occurred near the beginning of the diastolic filling period.
ESV declined in groups I and II but was unchanged in group III by tachycardia pacing. Isovolumic relaxation time did not significantly change in any group. However, relaxation was consistently prolonged at baseline and at faster heart rates in LVH animals by ≈30% compared with the other groups (P<.01). ESV and EDV were lower in sham-operated than in control animals, but there were no differences in EDP, pLAP, cardiac output, or relaxation time between these two groups.
Fig 2A⇓ displays an example of the pressure-volume response during and immediately after termination of rapid pacing (200 beats per minute) in a control (group I) dog. Time plots are shown on the left, and the corresponding pressure-volume loops are shown on the right. As noted above, rapid pacing led to a marked reduction in EDV, with minimal change in early- and late-diastolic pressures. Upon cessation of pacing, the initial pressure-volume trajectory followed a curve virtually identical to that observed during pacing. A similar result was obtained in sham-operated hearts (group II, Fig 2B⇓).
The response in group III ventricles was strikingly different (Fig 2C⇑). In this instance, pacing induced marked preload reduction as well as a rise in early- and late-diastolic pressures. However, rather than remaining elevated upon termination of pacing, pressure fell immediately on the first nonpaced beat. Importantly, this pressure decline occurred before the moment that ventricular activation would have ensued had pacing continued (arrow). Thus, recovery of diastolic function with cessation of atrial pacing took place before any alteration in the frequency of ventricular excitation.
Group data for this early rapid decline in diastolic pressure (ΔPraw) are shown in Fig 3⇓. At a pacing cycle length of 275 ms, ΔPraw was 10.1±3.3 mm Hg in group III (LVH) compared with only 2.6±1.4 and 2.5±0.7 mm Hg in groups I and II, respectively (P<.05). Significant disparities were observed at all pacing rates.
Timing of Atrial Contraction
Fig 2⇑ also identifies the time of pLAP on each time plot (vertical dashed line) and pressure-volume plot (plus sign [+]). At rapid pacing rates, as mentioned earlier and demonstrated in this figure, pLAP occurred near the onset of early-diastolic filling rather than near end diastole. At a pacing cycle length of 300 ms (200 beats per minute), pLAP followed the onset of ventricular filling by only 8.2±8.7 ms in group I and 17.7±8.7 ms in group II and preceded the onset of filling by 5.8±7.9 ms in group III (P<.05 group III vs group II). This means that atrial systolic pressure generation occurred primarily during ventricular isovolumic relaxation, with minimal antegrade emptying of the atrium. This timing of atrial systole suggested that its sudden absence upon cessation of atrial pacing might have led to the abrupt change in the diastolic pressure-volume excursion and to ΔPraw.
The importance of this timing was further explored by means of AV sequential pacing. Data during simultaneous pacing of the atrium and ventricle were compared with AV pacing with a normal PR interval (120 ms), both at an identical rapid ventricular rate. Fig 4⇓ shows an example of the results in an LVH dog. As with Fig 2⇑, the time of maximal atrial pressure is indicated by a plus sign (+). When the AV interval was 120 ms (Fig 4A⇓), atrial contraction occurred near the onset of filling for the paced beat, and cessation of pacing led to an immediate decline in LVP. This was identical to the result previously noted with normal His-Purkinje AV conduction (Fig 2⇑). However, when the AV interval was shortened to 0 ms (Fig 4B⇓), atrial contraction now occurred during ventricular isovolumic contraction. Interestingly, the early-diastolic pressure was no longer elevated during rapid atrial pacing, and upon termination of pacing, the pressures at the onset of filling for both the last paced and first nonpaced beat were identical. It is important to note that only the pressure at the onset of filling for the paced beat was altered by varying the PR interval, demonstrating the importance of atrial timing in the development of diastolic LVP elevation. A similar result was found in all LVH dogs tested.
AV sequential pacing runs were also performed in dogs in groups I and II. ΔPraw was generally <1 mm Hg with cessation of AV pacing with an AV delay of 120 ms and <0.5 mm Hg when the AV delay was set to zero in these dogs (data not shown).
Contribution of Filling to ΔPraw
To examine the contribution of early ventricular filling on the early-diastolic pressure rise between paced and immediately postpaced contractions, ΔPraw was corrected by analysis of E(t) (see “Materials and Methods”) to yield ΔPcor. If elevated pressures during rapid pacing were solely due to chamber filling superimposed on a continuing decline of E(t), then ΔPcor would be near zero. Results are displayed in Fig 5⇓. ΔPcor was significantly greater than zero in all cases, and more than three times greater in LVH hearts than in the other two groups. Furthermore, ΔPcor was ≥80% of ΔPraw in all three groups (compare with Fig 3⇑). Thus, the majority of the diastolic pressure decline after pacing was terminated was not attributable to differences in filling alone but reflected a change in E(t) during pacing.
Elastance-versus-time plots for the last two paced beats and the first nonpaced beat for the three examples of Fig 2⇑ are shown in Fig 6⇓. In this figure, the vertical axis is amplified to highlight the diastolic behavior of elastance with cessation of atrial pacing. In groups I and II, diastolic elastance fell to the same level during atrial pacing as when pacing was terminated. In contrast, elastance was elevated during pacing in group III animals and fell to normal levels immediately upon termination of pacing.
Influence of Pericardium
To explore the possibility that the rise in diastolic pressures in LVH dogs may have been caused by an exaggerated pericardial constraint, data were also obtained in a subset of animals in which the chest was opened and the pericardium was widely incised. Examples of the pressure-volume trajectories after cessation of atrial pacing (at cycle lengths of 350 ms) are shown in Fig 7⇓. The upper two panels (Fig 7A⇓ and 7B⇓) show data from two control animals, and the lower two panels (Fig 7C⇓ and 7D⇓) show data from two group III LVH animals, all measured after pericardiotomy. There is a strong qualitative similarity between these plots and those shown for groups I and III in Fig 2⇑. Thus, the diastolic pressure elevation during rapid pacing did not require an intact pericardium.
The rapid decline in diastolic pressure after suspension of atrial activation and loss of pressure elevation by varying the AV interval seemed inconsistent with a significant role of endocardial hypoperfusion. Table 3⇓ provides the endocardial-to-epicardial flow ratios for the present experiments. Overall, the ratio was slightly but significantly lower in LVH hearts. However, despite the marked rise in EDP and pLAP in LVH animals, the endocardial-to-epicardial flow ratio did not decline below 1.0, nor was there a significant influence of heart rate on the ratio. This was similarly true for the other two animal groups.
Influence of Hypertension (Angiotensin II)
To test the influence of acute hypertension itself on the pacing-induced diastolic pressure changes, particularly ΔPraw and ΔPcor, angiotensin II was administered to four group I animals at a dose that matched the systolic pressure of group III animals (190 mm Hg). Corresponding diastolic and mean pressures (144±14 and 165±6 mm Hg, respectively) were also similar to those for group III dogs (see Table 1⇑). Table 4⇓ summarizes baseline and rapid pacing data before and after angiotensin II administration. EDP and pLAP both rose significantly with angiotensin II levels at baseline heart rate, with only minimal further change during rapid pacing. Importantly, these changes were much less than those observed in group III hearts. Furthermore, there was no significant change in ΔPraw or ΔPcor when pacing was superimposed with angiotensin II infusion. Thus, hypertension alone did not reproduce the magnitude of early- and late-diastolic pressure rises observed during tachycardia in animals with chronic LVH.
We further tested the hypothesis that the acute rise in hypertension induced by angiotensin II was associated with an even greater rise in mean systolic wall stress than that developed after chronic hypertension (ie, group III). Wall stress in control hearts (group I) was 330.8±58.7 g/cm2, and this was not significantly different from the stress measured in the group III animals (381.9±111.6 g/cm2, P=.22). In contrast, in the animals exposed to angiotensin II, wall stress rose significantly from 347.4±32.7 to 486.4±108 g/cm2 (P<.05).
The principal striking finding in the present study was the rapidity with which marked diastolic pressure elevation, observed during acute atrial tachycardia of hypertensive LVH hearts, resolved after cessation of pacing. The fact that this occurred before a change in ventricular stimulation frequency and that it was eliminated by altering the relative timing of AV stimulation suggests a prominent role for mechanical effects of AV interaction. To our knowledge, this is the first study to specifically examine this very acute time course of pressure recovery and to demonstrate the importance of the timing of atrial contraction relative to ventricular diastole in this behavior.
Comparison With Prior Studies
Diastolic pressure elevation induced by pacing stress in hypertrophied myocardium is well established in both human and animal studies.6 8 9 21 Two major mechanisms have drawn the most attention for explaining this behavior: (1) impairment of coronary flow and high-energy phosphate metabolism7 9 10 and (2) dysfunction of myocyte calcium homeostasis.22 However, neither of these mechanisms could explain the marked pacing-induced increase of diastolic pressure in the present study. Rate-dependent pressure elevation reversed immediately after the termination of pacing. This finding and data showing minimal change in myocardial endocardial-epicardial flow distribution at these same heart rates indicate that ischemia was not likely to be a dominant mechanism. Also, the rapid recovery of LV diastolic pressure after pacing was terminated occurred before the ventricle was due to be restimulated had pacing continued, ie, before slowing of the ventricular activation rate. This is not consistent with a frequency-dependent abnormality of myocardial calcium handling. However, we cannot rule out the possibility that these metabolic abnormalities may have been present to some extent in the LVH hearts. In particular, the fact that ESV did not fall with pacing in this group suggests a modest decline in systolic function. Similar changes have been reported in human ventricular hypertrophy and appear to be related to altered calcium handling.23
In experimental and clinical coronary insufficiency, pacing-induced diastolic abnormalities may persist many seconds after return to baseline heart rate.21 24 25 26 Similar physiology has been suggested for hypertrophy of the heart due to inadequacies of coronary flow and energy metabolism, as previously indicated. Sustained diastolic dysfunction has indeed been reported in patients with aortic stenosis21 and idiopathic hypertrophic cardiomyopathy,8 in whom angina was induced by rapid pacing in the absence of significant coronary artery disease. However, these may both represent examples where a mismatch between myocardial supply and demand is exacerbated by the presence of pressure gradients between the LV and aortic root. Such gradients are not present in hypertensive LVH or in hearts with supra-aortic banding. The majority of experimental LVH studies reporting diastolic pressure elevation due to atrial tachycardia have only reported data measured during pacing, so there is very little direct comparative data in the literature. One recent study did report sustained regional wall motion abnormalities and reduced endocardial flows after pacing was terminated,9 although interestingly, LAP was similar to control values at this time.
One potential source for the apparent discrepancy between present and prior data, particularly with respect to the role of endocardial hypoperfusion, may lie in differences in the severity of LVH. The majority of studies demonstrating subendocardial hypoperfusion and ischemia have used aortic-banding models of LVH,6 7 9 10 in which the degree of hypertrophy is substantially greater than that induced by perinephritic hypertension. Thus, the present data do not necessarily conflict with these prior results. Rather they indicate that similar macroscopic diastolic abnormalities during rapid pacing may be observed even in ventricles with more modest LVH; but in this setting, the behavior derives from different mechanisms.
Role of Atrial Contraction
Inspection of the relative timing of atrial and ventricular activation during rapid pacing suggested an important role for AV interaction in explaining the diastolic pressure elevation in LVH hearts. At heart rates >200 beats per minute, pLAP was nearly synchronous with the onset of diastolic filling (see Fig 2⇑). This meant that LAP rose abruptly during late relaxation, prematurely opening the mitral valve and initiating filling at higher pressures. Interestingly, once pressure was increased at the onset of filling, it remained elevated, with only minimal further change throughout the remainder of diastole. For example, in LVH dogs paced at 200 beats per minute, the pressure at the onset of diastolic filling was increased by 12.1±2.4 mm Hg over baseline, and EDP rose by 12.6±7.0 mm Hg, a difference of only 0.5 mm Hg. This suggests that at rapid heart rates with a short diastolic period, mechanisms responsible for an increase in early-diastolic pressure substantially contribute to EDP elevation. The importance of the precise timing of atrial contraction relative to LV diastole was more directly tested by synchronous AV pacing. With this intervention, the ventricular stimulation frequency was unchanged, so that mechanisms for LV diastolic dysfunction related to the LV itself should have persisted. Yet both early- and late-diastolic pressure elevation, observed in LVH hearts when paced using a normal AV interval, were eliminated by this maneuver.
The sudden loss of atrial systole with cessation of pacing delayed the onset of LV filling. This by itself could produce a sudden decline in LV pressure by allowing the ventricle more time to complete relaxation before filling began.27 If, in fact, the diastolic LV pressure rise during rapid atrial pacing were solely due to greater blood volume within the heart, then chamber stiffness (or elastance) would minimally differ between equivalent time points of paced and postpaced beats. This was tested by calculation of ΔPcor, which revealed that filling per se contributed only ≈20% to the total pressure increase. The remaining 80% was ascribed to an increase in LV E(t) (or stiffness) itself (see Fig 6⇑). This percentage depends to some extent on our computation of elastance as the instantaneous ratio of LVP to LVV, assuming a volume-axis intercept (Vo) of zero. We found little difference in the calculation of ΔPcor when Vo ranged from −10 to 10 mL. When Vo was estimated from isochronal points of multiple pressure-volume loops measured at varying preloads, it was generally negative. If anything, using a Vo of zero resulted in an underestimation of ΔPcor.
The results from the angiotensin II infusion studies support the notion that the rise in diastolic pressure (and ΔPcor) during rapid pacing in LVH dogs reflected a pacing-induced change in LV elastance rather than being merely the consequence of increased preload. In control dogs, angiotensin II acutely increased EDP both in sinus rhythm and during rapid pacing. However, both ΔPraw and ΔPcor remained small and were not significantly different from values measured in the absence of angiotensin II, despite a nearly 160% increase in EDP (Table 4⇑). Thus, simply raising diastolic pressures (or stresses) in control animals was insufficient to reproduce the phenomenon.
Several mechanisms could explain an elastance change. Viscous properties28 29 of the still-relaxing ventricle could yield an increase in pressure with the sudden onset of filling. In a study of normal canine ventricles, Nikolic et al30 clamped LV volume at end systole, allowed LV pressure to fully decay, and then refilled the heart at varying inflow rates. Under these conditions, there were negligible viscous effects. However, similar data do not exist when filling is initiated in an incompletely relaxed or hypertrophied heart. Both situations might exacerbate viscous behavior, resulting in an increase in chamber stiffness during filling. Diastolic suction31 may have contributed to the early-diastolic pressure decline because of the acute deferral of ventricular filling. However, it is unlikely that the increase in ΔPcor with heart rate in group III dogs was due to a suction effect, since the magnitude of this phenomenon should be constant at a fixed ESV, and as shown in Table 2⇑, ESV did not significantly change with the pacing rate.
Another mechanism involved in elastance change is that the atria itself could directly influence LV chamber stiffness. Atrial systole can be characterized by a time-varying elastance,32 33 and mechanical coupling of the atrium to the ventricle via the fibrous exoskeleton34 could result in an LV stiffness change during atrial contraction. This AV coupling is apparently not attributable to pericardial effects, because the phenomenon persisted after pericardiotomy (Fig 7⇑). It remains unclear why only LVH hearts demonstrated the diastolic abnormality. One possibility is that viscous effects or atrial stiffness were themselves augmented by the hypertrophic process. Alternatively, the different magnitude of this effect could relate to a greater force of atrial contraction with LVH, evidenced by the much higher pLAP. LVH also led to a slightly but significantly earlier occurrence of pLAP relative to the onset of ventricular filling than in the other two groups, exposing the ventricle to this tethering force at a point when early diastole is more readily influenced.
Role of Arterial Hypertension
LV systolic pressure elevation in group III animals could itself have contributed to rate-dependent diastolic dysfunction. Acutely increased afterload delays relaxation in both isolated muscle and intact ventricles.35 36 Furthermore, as shown by Gelpi et al,2 diastolic dysfunction that characterizes early evolving perinephritic hypertension can be mimicked by acute administration of a vasoconstrictor. Hypertension alone, however, could not explain the difference in pacing response between LVH and non-LVH hearts in the present study. After matching elevations in arterial pressure in the normal control animals, we found that relaxation was prolonged and EDP rose. However, diastolic pressures were minimally further altered by rapid pacing, and ΔPcor was not increased, in striking contrast to the LVH response. Thus, a hypertrophic response to hypertension is required to observe the pathological behavior.
The angiotensin II data also demonstrated that elevation of mean ejection stress with attendant prolongation of relaxation time was insufficient to generate the diastolic pressure rise during rapid pacing. Stress increased by 40% with angiotensin II infusion compared with a much smaller (15%) and statistically insignificant rise after chronic LVH. This result suggests that systolic stress itself was not the cause of the prolonged relaxation time in group III hearts and is not the mechanism that underlies the rise in diastolic pressure with rapid pacing.
The effect of the anesthesia on the observed hemodynamics should be considered. Sodium pentobarbital is a mild cardiodepressant, and it raises resting sinus rate through its vagolytic activity. This prevented us from obtaining data at cycle lengths >350 ms in the majority of animals, which may have masked the rate-dependent shortening in the relaxation time constant that is reported at longer cycle lengths. There did not appear to be a direct influence of the anesthesia on relaxation times themselves, as our group I data were very similar to values reported in normal intact conscious animals at similar heart rates.37 Changes in relaxation times were likely minimized by our use of an intact (nonsurgically invaded) chest preparation. Anesthesia also may have acted to reduce baseline LV EDP to the normal levels observed in group III dogs, although similar LV EDP values have been reported in mild to moderate LVH in conscious dogs.2 6 38 Thus, the anesthesia itself likely had little influence on the observed diastolic behavior during rapid pacing.
The extent of ventricular hypertrophy achieved in the group III animals was modest, so our findings cannot necessarily be extrapolated to other models of more severe LVH. However, this form of hypertrophy is very clinically relevant, since patients with LVH often develop a similar modest degree of LVH, and it is usually the consequence of chronic hypertension as opposed to aortic outflow obstruction.
Finally, the computation of ΔPcor is dependent on our definition of E(t), which may in fact be a function of loading conditions. Specifically, instantaneous elastance during relaxation has been found to vary as a function of afterload and preload volumes.39 40 However, the two cardiac cycles compared for this analysis started at identical EDV and developed identical mean ejection wall stress. Thus, there was no difference in prior loading for these two cycles.
In summary, we found that elevation in diastolic LVP induced by rapid atrial pacing in dogs with mild-to-moderate LVH due to chronic hypertension resolves immediately upon cessation of pacing and thus cannot be attributed to ischemic or metabolic mechanisms. Rather, this pressure rise depends on the occurrence of atrial contraction just before end-isovolumic relaxation during rapid atrial pacing. Atrial systole increased early-diastolic pressures partly by injecting blood into the ventricle before the completion of ventricular relaxation. However, a much larger component of the increase was due to stiffening of the LV chamber, consistent with enhanced viscous behavior, or an internal tethering linking atrial contraction force with ventricular diastole, both exacerbated by LVH.
These studies were supported by National Institutes of Health (NIH) grants HL-34519 (Dr Anderson) and HL-47511 (Dr Kass) and an NIH fellowship training award (Dr Berger). Dr Kass is an Established Investigator of the American Heart Association. We thank Drs Lewis C. Becker and W. Lowell Maughan for their helpful review of the manuscript and gratefully acknowledge the excellent technical assistance of Richard Tunin in performing these studies.
- Received January 30, 1995.
- Accepted March 9, 1995.
- © 1995 American Heart Association, Inc.
Gelpi RJ, Pasipoularides A, Lader AS, Patrick TA, Chase N, Hittinger L, Shannon RP, Bishop SP, Vatner SF. Changes in diastolic cardiac function in developing and stable perinephritic hypertension in conscious dogs. Circ Res. 1991;68:555-567.
Cuocolo A, Sax FL, Brush JE, Maron BJ, Bacharach SL, Bonow RO. Left ventricular hypertrophy and impaired diastolic filling in essential hypertension. Circulation. 1990;81:978-986.
Blaustein AS, Gaasch WH. Myocardial relaxation. Circ Res. 1981;49:633-639.
Gaasch WH, Bing OHL, Mirsky I. Chamber compliance and myocardial stiffness in left ventricular hypertrophy. Eur Heart J. 1982;3:139-145.
Fujii AM, Gelpi RJ, Mirsky I, Vatner SF. Systolic and diastolic dysfunction during atrial pacing in conscious dogs with left ventricular hypertrophy. Circ Res. 1988;62:462-470.
Zhang J, Merkle H, Hendrich K, Garwood M, From AHL, Ugurbil K, Bache RJ. Bioenergetic abnormalities associated with severe left ventricular hypertrophy. J Clin Invest. 1993;92:993-1003.
Cannon RO III, Rosing DR, Maron BJ, Leon MB, Bonow RO, Watson RM, Epstein SE. Myocardial ischemia in patients with hypertrophic cardiomyopathy: contribution of inadequate vasodilator reserve and elevated left ventricular filling pressures. Circulation. 1985;71:234-243.
Bache RJ, Arentzen CE, Simon AB, Vrobel TR. Abnormalities in myocardial perfusion during tachycardia in dogs with left ventricular hypertrophy: metabolic evidence for myocardial ischemia. Circulation. 1984;69:409-417.
Nakano K, Corin WJ, Spann JF Jr, Biederman RWW, Denslow S, Carabello BA. Abnormal subendocardial blood flow in pressure overload hypertrophy is associated with pacing-induced subendocardial dysfunction. Circ Res. 1989;65:1555-1564.
Jeremy RW, Fletcher PJ, Thompson J. Coronary pressure-flow relations in hypertensive left ventricular hypertrophy. Circ Res. 1989;65:224-236.
Ferrario CM, Page IH, McCubbin JW. Increased cardiac output as a contributory factor in experimental renal hypertension in dogs. Circ Res. 1970;27:799-810.
Baan J, Van der Velde ET, de Brun HG, Smeenk GJ, Koops J, van Dijk AD, Temmerman D, Senden J, Buis B. Continuous measurement of left ventricular volume in animals and humans by conductance catheter. Circulation. 1984;70:812-823.
Kass DA, Yamazaki T, Burkhoff D, Maughan WL, Sagawa K. Determination of left ventricular end-systolic pressure-volume relationships by the conductance (volume) catheter technique. Circulation. 1986;73:586-595.
Liu CP, Ting CT, Lawrence W, Maughan WL, Chang MS, Kass DA. Diminished contractile response to increased heart rate in intact human left ventricular hypertrophy: systolic versus diastolic determinants. Circulation. 1993;88(pt 1):1893-1906.
Raff GL, Glantz SA. Volume loading slows left ventricular isovolumic relaxation rate. Circ Res. 1981;48:813-824.
Fifer MA, Bourdillon PD, Lorell BH. Altered left ventricular diastolic properties during pacing-induced angina in patients with aortic stenosis. Circulation. 1986;74:675-683.
Gwathmey JK, Warren SE, Briggs GM, Copelas L, Feldman MD, Phillips PJ, Callahan M Jr, Schoen FJ, Grossman W, Morgan JP. Diastolic dysfunction in hypertrophic cardiomyopathy. J Clin Invest. 1991;87:1023-1031.
Liu CP, Lawrence W, Ting CT, Chang MS, Maughan WL, Kass DA. Diminished contractile response to increased heart rate in intact human left ventricular hypertrophy: systolic versus diastolic determinants. Circulation. 1993;88:1893-1906.
Serizawa T, Carabello BA, Grossman W. Effect of pacing-induced ischemia on left ventricular diastolic pressure-volume relations in dogs with coronary stenoses. Circ Res. 1980;46:430-439.
Bourdillon PD, Lorell BH, Mirsky I, Paulus WJ, Wynne J, Grossman W. Increased regional myocardial stiffness of the left ventricle during pacing-induced angina in man. Circulation. 1983;67:316-323.
Aroesty JM, McKay RG, Heller GV, Royal HD, Als AV, Grossman W. Simultaneous assessment of left ventricular systolic and diastolic dysfunction during pacing-induced ischemia. Circulation. 1985;71:889-900.
Weisfeldt ML, Frederiksen JW, Yin FCP, Weiss JL. Evidence of incomplete left ventricular relaxation in the dog: prediction from the time constant for isovolumic pressure fall. J Clin Invest. 1978;62:1296.
Rankin JS, Arentzen CE, McHale PA, Ling D, Anderson RW. Viscoelastic properties of the diastolic left ventricle in the conscious dog. Circ Res. 1977;41:37-45.
Pouleur H, Karliner JS, LeWinter MM, Covell JW. Diastolic viscous properties of the intact canine left ventricle. Circ Res. 1979;45:410-419.
Nikolic SD, Tamura K, Tamura T, Dahm M, Frater RWM, Yellin EL. Diastolic viscous properties of the intact canine left ventricle. Circ Res. 1990;67:352-359.
Nikolic S, Yellin EL, Tamura K, Vettner H, Tamura T, Meisner JS, Frater RWM. Passive properties of canine left ventricle: diastolic stiffness and restoring forces. Circ Res. 1988;62:1210-1222.
Alexander J Jr, Sunagawa K, Chang N, Sagawa K. Instantaneous pressure-volume relation of the ejecting canine left atrium. Circ Res. 1987;61:209-219.
Hoit BD, Shao Y, Gabel M, Walsh RA. In vivo assessment of left atrial contractile performance in normal and pathological conditions using a time-varying elastance model. Circulation. 1994;89:1829-1838.
Liu CP, Ting CT, Yang TM, Chen JW, Chang MS, Maughan WL, Lawrence W, Kass DA. Reduced left ventricular compliance in human mitral stenosis: role of reversible internal constraint. Circulation. 1992;85:1447-1456.
Su JB, Hittinger L, Laplace M, Crozatier B. Loading determinants of isovolumic pressure fall in closed chest dogs. Am J Physiol. 1991;260:H690-H697.
Cheng C-P, Freeman GL, Santamore WP, Constantinescu MS, Little WC. Effect of loading conditions, contractile state, and heart rate on early diastolic left ventricular filling in conscious dogs. Circ Res. 1990;66:814-823.
Bache RJ, Dai XZ, Alyono D, Vrobel TR, Homans DC. Myocardial blood flow during exercise in dogs with left ventricular hypertrophy produced by aortic banding and perinephritic hypertension. Circulation. 1987;76:835-842.
Sagawa K, Maughan WL, Suga H, Sunagawa K. Cardiac Contraction and the Pressure-Volume Relationship. New York, NY: Oxford University Press; 1988.
Burkhoff D, De Tombe PP, Hunter WC. Impact of ejection on magnitude and time course of ventricular pressure-generating capacity. Am J Physiol. 1993;265:H899-H909.