INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DURING EXERCISE, the marked increase in left ventricular (LV) filling rate in early diastole mainly depends on the ability of the LV to relax rapidly and completely (4). The physiological mechanisms allowing this adaptation involve an increase in both heart rate and contractility through -adrenergic stimulation (3, 7). As a consequence, -blockers strongly alter the relaxation process as a result of the combination of their negative chronotropic and inotropic properties, both at rest and during exercise (4). In this respect, the development of a selective bradycardic agent such as ivabradine, a novel inhibitor of the sinoatrial inward hyperpolarization-activated current (I_f), represents an original alternative approach to -blockers because this drug is devoid of intrinsic negative inotropic effect and does not alter either global LV systolic function or coronary vasomotion in conscious dogs at rest or during exercise (20, 22, 23). Although nothing is presently known about the effects of ivabradine on isovolumic relaxation, it was reported previously that zatebradine, another inhibitor of I_f current, did not alter the time constant of relaxation in anesthetized pigs (2). In the study of Miura et al. (18) the relaxation process was preserved during exercise in dogs under zatebradine treatment. However, the results of the latter experiments were obscured by the fact that zatebradine also reduced arterial pressure and induced a significant negative inotropic effect, two properties that may have interfered per se with the analysis of the LV isovolumic relaxation period. Furthermore, dobutamine-induced decreases in the time constant of isovolumic relaxation were not altered by zatebradine in conscious dogs (10). Finally, direct comparison of the effects of a -blocker and a selective bradycardic agent on LV isovolumic relaxation at rest and during exercise has not been performed yet.

Accordingly, the aim of the present study was to investigate the effects of increasing doses of ivabradine and the -blocker atenolol on LV isovolumic relaxation at rest and during treadmill exercise in chronically instrumented dogs. These experiments were performed both at spontaneous and at controlled heart rate by atrial pacing, a maneuver that is known to preserve the intrinsic LV properties in the normal heart (11).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The animal instrumentation and the experiments were performed in accordance with official regulations of the French Ministry of Agriculture.

Instrumentation. Studies were performed in eight mongrel dogs (20-30 kg). The animals were anesthetized with pentobarbital sodium (30 mg/kg iv), intubated, and ventilated with room air enriched with oxygen. A left thoracotomy was performed through the fifth left intercostal space under sterile surgical conditions. Fluid-filled Tygon catheters were implanted in the descending thoracic aorta and the left atrium, and a Silastic catheter was implanted in the pulmonary artery. A solid-state pressure transducer (P7A, Konigsberg Instruments, Pasadena, CA) was introduced into the apex of the LV. Piezoelectric ultrasonic dimension crystals (3 MHz) were implanted 1) on opposed anterior and posterior endocardial surfaces of the LV to measure LV internal diameter halfway between the apex and the aortic valve and 2) on opposed endocardial and epicardial surfaces of the posterior LV wall to measure wall thickness. Finally, stainless steel wires were sewn to the left atrial appendage for subsequent electrical pacing. All catheters and wires were exteriorized between the scapulae, and the pneumothorax was evacuated. Cefazolin (1 g iv) and gentamicin (40 mg iv) were administered during the first week after surgery. Catheters were flushed daily with saline containing 2,000 IU/ml heparin. The position of all catheters and crystals was confirmed at autopsy.

Hemodynamic measurements. All hemodynamic data were recorded on a multichannel recorder (ES 1000, Gould Instruments, Cleveland, OH) and analyzed with HEM version 3.2 data acquisition software (Notocord Systems, Croissy sur Seine, France). Aortic and left atrial pressures were measured with Statham P23ID strain gauge transducers (Statham Instruments, Oxnard, CA) connected to their respective catheters. LV pressure, LV internal diameter, and LV wall thickness were digitized at 500 Hz. LV pressure was calibrated in vitro with a mercury manometer and cross-calibrated in vivo with the left atrial and aortic pressures.

Calculation of isovolumic relaxation time constant. The isovolumic relaxation period was defined as the period elapsed from the time of the peak of negative LV dP/dt to the time when LVP fell to a value of 5 mmHg above LV end-diastolic pressure of the following beat. The LV relaxation time constant  was computed by three distinct methods: 1) a monoexponential pressure decay model with zero asymptote (₀; Ref. 25), 2) a monoexponential decay model with nonzero asymptote (_D), and 3) a best fit monoexponential decay model with nonzero asymptote (_BF; Refs. 13, 17). In the first model, ₀ was evaluated using a zero-asymptote model
where  was computed by a linear regression of the logarithm of LVP during the isovolumetric phase of relaxation; t is the elapsed time from the timing of peak negative LV dP/dt where t = 0; P₀ is the value of LVP at the time of peak negative LV dP/dt; and ₀ is the relaxation time constant. In the second model, LVP was formulated as
where _D is the relaxation time constant and P_A is the pressure asymptote. In this model, _D was calculated by a linear regression of LV dP/dt against LVP (24). In the third model, LVP was modeled with the same formula as in the second model, but the relaxation time constant _BF and the pressure asymptote (P_BF) were calculated by the Levenberg-Marquart nonlinear regression algorithm to obtain the least-squares best fit curve to the measured LVP. Parameters obtained from the second method were used as initial values for the iterative procedure of the nonlinear regression to stabilize the solution. Correlation coefficients of regression were calculated for the three methods. Because the second method is the basis for improved curve fitting and correlation, only the results obtained by the Weiss method (₀) and the Levenberg-Marquart algorithm (_BF, P_BF) are presented in the present study. Finally, LVP_min, the lowest LV diastolic pressure, was also measured.

Protocol. The experiments were conducted 3-4 wk after surgery when the dogs were healthy and apyretic. While the dogs were standing quietly on the treadmill, all investigated parameters were first measured ("baseline"). Thereafter, all dogs received in random order saline, atenolol (0.25, 0.5, or 1 mg/kg) or ivabradine [0.25, 0.5, or 1 mg/kg; (3-(3-{[((7S)-3,4-dimethoxybicyclo[4,2,0]octa-1,3,5-trien-7-yl) methyl]methylamino}propyl)-1,3,4,5-tetrahydro-7,8-dimethoxy-2H-3-benzazepin-2-one hydrochloride (S-16257-2); Servier, Neuilly-sur-Seine, France]. All drugs and saline were administered as an intravenous infusion through the pulmonary artery catheter over a 5-min period. Fifteen minutes later, all investigated parameters were measured again ("rest") and a sequence of 5 min of atrial pacing at a rate of 125 beats/min was performed ("rest paced"). The dogs were then subjected to 10 min of treadmill exercise (10 km/h, slope 10%), with the first 5 min performed at spontaneous heart rate ("exercise") and the last 5 min performed under atrial pacing at a fixed rate of 250 beats/min ("exercise paced"). During these two phases of exercise, all measurements were performed at the steady state of the response. Only one treatment was administered on each experimental day, and at least 2 days elapsed between two consecutive exercises in the same animal.

Statistical analysis. All results are means ± SE. Statistical analysis was performed by a two-way analysis of variance for repeated measures. When overall differences were detected, individual comparisons were performed by Student's t-test for paired observations with Bonferroni's correction. The statistical analysis was performed with StatView version 5.0 (Abacus Concept, Berkeley, CA). A value of P < 0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hemodynamics at rest. As shown in Tables 1 and 2, baseline hemodynamic values were not significantly different among the different sequences of the protocol. None of the measured hemodynamic parameters was altered at rest after saline administration. Both atenolol and ivabradine induced a similar dose-dependent decrease in heart rate at rest. Atenolol, at all investigated doses in spontaneous and paced heart rates, significantly decreased maximum LV dP/dt (LV dP/dt_max) and V_cfc compared with saline. Resting LV end-diastolic pressure was increased after 1 mg/kg ivabradine, an effect that was abolished by atrial pacing. Finally, LV systolic pressure and LV peak systolic and LV end-systolic wall stresses were not affected by ivabradine and atenolol compared with saline.

Hemodynamics during exercise. During exercise under saline, heart rate increased from 108 ± 5 to 220 ± 6 beats/min (+103 ± 19%). Both atenolol and ivabradine similarly and dose-dependently limited this increase (e.g., 149 ± 4 and 148 ± 4 beats/min, respectively, at 1 mg/kg). The exercise-induced increases in LV dP/dt_max and V_cfc were strongly blunted by atenolol regardless of the dose administered, whereas ivabradine slightly reduced LV dP/dt_max at 0.5 and 1 mg/kg and V_cfc at 1 mg/kg. Atrial pacing during exercise restored LV dP/dt_max and V_cfc measured under ivabradine administration, but the effects of atenolol on LV dP/dt_max were not corrected by this maneuver. Ivabradine (0.5 and 1 mg/kg) increased LV end-diastolic pressure, and this effect was corrected by atrial pacing. In contrast, atenolol significantly increased LV end-diastolic pressure during exercise, but this was not corrected by atrial pacing at the dose of 1 mg/kg. Finally, the exercise-induced increases in LV systolic pressure and peak systolic wall stress were not affected by ivabradine. They were, however, limited by atenolol (P < 0.05), an effect that was not corrected by atrial pacing. Concomitantly, LV end-systolic wall stresses were significantly greater at the highest dose of atenolol at spontaneous heart rate compared with saline.

Rate of isovolumic relaxation. The effects of saline, atenolol, and ivabradine on LV relaxation parameters at rest and during exercise are shown in Table 3. In Figs. 1 and 2 phase plane plots of LV pressure vs. LV dP/dt are shown, and they illustrate the validity of the calculations. Correlation coefficients of regression reached a value of ~0.999 with the Levenberg-Marquart calculation. With the Weiss calculation, this value averaged 0.991 (0.987-0.997) in spontaneous heart rate, i.e., at baseline, at rest, and during exercise. Under atrial pacing, the mean correlation coefficients averaged 0.982 (0.968-0.989) and 0.983 (0.969-0.992) at rest and during exercise, respectively.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Left ventricular (LV) change in pressure over time (dP/dt) vs. LV pressure (LVP) loops constructed during a single representative beat at spontaneous heart rate at baseline (A), at rest (B), and during exercise (C). Recordings were performed every 2 ms after administration of saline (), ivabradine (1 mg/kg; ), and atenolol (1 mg/kg; ). During exercise, the slope of the linear relation between LV dP/dt and LVP during the isovolumic relaxation period was similar between saline and ivabradine but was less steep with atenolol. At a given LVP during the relaxation period, the corresponding value of LV dP/dt was similar between saline and ivabradine but decreased by atenolol. Any deviations of isovolumic LVP decay from a monoexponential course can be appreciated on the phase-plane plots of LV isovolumic relaxation pressure as deviations from a linear relation between LVP and LV dP/dt. Note that such a deviation is not observed in this figure. The shaded zones show the course of the pressure decay during the isovolumic relaxation period. The arrows show the direction of time and starts at the end-diastolic period.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   LV dP/dt vs. LVP loops during a single representative beat measured at rest (A) and with exercise (B) under atrial pacing at 125 and 250 beats/min, respectively. Recordings were performed every 2 ms after administration of saline (), ivabradine (1 mg/kg; ), and atenolol (1 mg/kg; ). During exercise, the slope of the linear relation between LV dP/dt and LVP during isovolumic relaxation period was similar with saline and ivabradine but was less steep with atenolol. At a given LVP during the relaxation period, the corresponding value of LV dP/dt was similar with saline and ivabradine but significantly decreased by atenolol. Note that the loops constructed for saline and ivabradine are superimposed. The shaded zones show the course of the pressure decay during the isovolumic relaxation period. The arrows show the direction of time and starts at the end-diastolic period.

View larger version (10K): [in this window] [in a new window]   Fig. 3.   Typical traces of the LVP measured during the isovolumic relaxation period vs. time in a conscious dog at baseline (A), at rest (B), and during exercise (C) at spontaneous heart rate. Data were obtained after administration of saline (), ivabradine (1 mg/kg; ), and atenolol (1 mg/kg; ). Each point was obtained every 2 ms. Ivabradine did not alter the LVP decay either at rest or during exercise. In contrast, atenolol significantly prolonged the LV isovolumic relaxation period both at rest and during exercise compared with control and ivabradine.

View larger version (8K): [in this window] [in a new window]   Fig. 4.   Relationship between heart rate and LV relaxation constant determined by the best fit method (BF) after administration of saline (), ivabradine (1 mg/kg; ), and atenolol (1 mg/kg; ) at rest and during exercise. *P < 0.05, BF different from corresponding saline value.

Extent of isovolumic relaxation. As shown in Table 3, ivabradine increased both the extent of relaxation as assessed by P_BF and LVP_min during exercise. These changes were abolished by atrial pacing except for LVP_min at the highest dose (1 mg/kg). Atenolol induced more pronounced, dose-dependent increases in P_BF and LVP_min. In contrast to ivabradine, the P_BF alterations induced by atenolol were only reduced whereas those in LVP_min remained unchanged when heart rate was controlled by atrial pacing.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study demonstrates that the exercise-induced acceleration of the rate of LV isovolumic relaxation was preserved by the selective I_f channel inhibitor ivabradine, the extent of relaxation being only altered at the highest doses tested (0.5 and 1 mg/kg). In contrast, and for a similar limitation of exercise-induced tachycardia, -blockade by atenolol dose-dependently decreased both the rate and extent of LV isovolumic relaxation. These results were observed regardless of the calculation method used (Weiss method for ₀ and best fit calculation method for _BF). Our results also suggest that heart rate is not a major determinant of acceleration of the rate of LVP fall during exercise.

As previously demonstrated (20), ivabradine, an inhibitor of the pacemaker I_f current (5), induced dose-dependent reductions in heart rate at rest and during exercise. Atenolol exhibited a similar pattern of negative chronotropic effects at all investigated doses. As expected, myocardial contractility was markedly depressed by atenolol, both at rest and during exercise, whereas ivabradine did not alter global LV function, in agreement with a previous study (20). Calculations of  demonstrated significant increases in the rate of LV relaxation and more negative values of P_BF during exercise performed under saline. Atenolol dose-dependently alleviated this acceleration of LV pressure fall during exercise, and  failed to decrease despite a 40% increase in heart rate, in agreement with previous results (4). The changes of the asymptote P_BF and LVP_min demonstrated the same pattern. In contrast with atenolol, the effects of ivabradine were similar to those observed under saline, i.e., the bradycardic effect per se did not counteract the acceleration process of LV isovolumic relaxation. Further analyses revealed that, at the highest doses, ivabradine induced a significant upward shift in P_BF and LVP_min compared with saline, suggesting that the drug could, to some extent, interfere with the LV relaxation process. It is important to emphasize that this effect was always concomitant with a slight but significant negative inotropic effect, as assessed by V_cfc and LV dP/dt_max. Given that this effect was abolished by atrial pacing, this could be related to the negative staircase phenomenon. Although both ivabradine and atenolol similarly limited the exercise-induced tachycardia, they clearly exhibited different intrinsic properties on the rate of LV isovolumic relaxation. Differences in the inotropic state between atenolol and ivabradine may in part explain these results. Changes in loading sequence, i.e., the magnitude of afterload and systolic load profiles, should also be taken into account (9, 12, 21, 26). Kohno et al. (16) demonstrated that change in LVP at aortic valve closure was a reasonable marker of the alteration of the systolic loading sequence. Ivabradine did not significantly alter this sequence compared with saline. In contrast, atenolol decreased the peak systolic stress but increased the end-systolic load, which, importantly, may contribute to a decrease in LV relaxation rate (12, 14). Preload itself does not influence the LV isovolumic relaxation rate in the intact heart (8). Finally, differential intracellular effects of atenolol and ivabradine should be mentioned. -Adrenergic stimulation is known to enhance calcium reuptake by the sarcoplasmic reticulum through the phosphorylation of phospholamban, which causes disinhibition of sarcoplasmic reticulum Ca²⁺-ATPase (19). Therefore -blockade, but not ivabradine, might slow the relaxation process by interfering with this cellular pathway.

Although tachycardia is a fundamental adaptive mechanism of myocardial performance during exercise, the present results argue against a major role for heart rate in the exercise-induced decrease in the time constant of LV relaxation, as previously suggested (6, 15, 18). Despite the ivabradine (1 mg/kg)-induced reduction in heart rate by ~70 beats/min at the peak of exercise, we did not observe any significant difference in  compared with saline. Furthermore, after matching for heart rate by atrial pacing at 250 beats/min, the values of _BF remained similar between saline and ivabradine. Therefore, acceleration of LV relaxation during exercise was mediated by mechanisms other than changes in heart rate. Accordingly, these results are consistent with previous studies showing that sympathetic stimulation contributes to the downward shift of the early diastolic portion of the LV pressure loop during exercise, independently of the simultaneous increase in heart rate (4). In addition, acceleration of LV relaxation at increasing heart rate was minimal in autonomically blocked conscious dogs (1). In the present study, the alterations in  were more likely related to the changes in the inotropic state and to the sequence of loading conditions than to changes in heart rate. Although the role of heart rate appears to be minimal in our experimental conditions, we cannot rule out its role as an important determinant of .

In conclusion, ivabradine dose-dependently limited tachycardia during exercise without simultaneously exerting major intrinsic depressant effects on exercise-induced acceleration of the rate of LV relaxation. In contrast, atenolol markedly decreased the rate of LVP fall at rest and during exercise. Atenolol more severely altered the extent of the relaxation process than ivabradine.