Randomized, Double-Blind Crossover Study to Investigate the Effects of Amlodipine and Isosorbide Mononitrate on the Time Course and Severity of Exercise-Induced Myocardial Stunning
Background—Myocardial stunning may cause prolonged left ventricular dysfunction after exercise-induced ischemia that can be attenuated by calcium antagonists in animal models. To assess their effects in humans, we performed a randomized, double-blind crossover study comparing the calcium antagonist amlodipine (10 mg once daily) versus isosorbide mononitrate (ISMN, 50 mg once daily) on postexercise stunning.
Methods and Results—Twenty-four men with chronic stable angina and normal left ventricular function underwent serial quantitative exercise stress echocardiography after 3 weeks on each treatment to assess the degree of postexercise stunning with simultaneous sestamibi single-photon emission computed tomography perfusion scans at peak stress to quantify the ischemic burden. Exercise time (P=1), maximum ST depression (P=0.48), and sestamibi single-photon emission computed tomography scores (P=0.17) were unchanged between treatments. Stunning occurred more often with ISMN than amlodipine (82% versus 48%). The global and segmental stress echocardiography parameters of stunning were attenuated in patients while taking amlodipine compared with ISMN. Shortening fractions and ejection fractions were less impaired 30 minutes after exercise in patients receiving amlodipine (3.5±1.4% versus 2.5±1.4%, P=0.014, and 59.7±5.4% versus 54.5±8%, P<0.001); similarly, the isovolumic relaxation period was less prolonged with amlodipine (93±15.5 versus 106.3±14.9 ms, P=0.018).
Conclusions—Despite comparable levels of ischemia, amlodipine attenuated stunning when compared with ISMN. This beneficial effect may relate to a prevention of the calcium overload implicated in the pathogenesis of stunning.
Myocardial stunning, or postischemic dysfunction, is the mechanical dysfunction that persists after ischemia is relieved despite the absence of irreversible damage.1 2 It is thought to be an important phenomenon in clinical settings of transient ischemia and reperfusion including myocardial infarction, unstable angina, and coronary artery bypass grafting.3 4 Stunning has been shown to occur after ischemia induced by increased myocardial oxygen demand in animals with partial coronary stenoses,5 and several echocardiographic studies in humans have suggested that stunning occurs after treadmill exercise.6 7 Recently by simultaneously measuring left ventricular (LV) function and perfusion, reversible myocardial dysfunction persisting after exercise in patients with coronary artery disease has been demonstrated to be due to myocardial stunning.8 Abnormalities of contraction may persist for up to 1 hour, whereas diastolic abnormalities may persist for several hours after a single episode of effort angina. The duration of contractile abnormalities after exercise-induced ischemia suggests that patients may suffer important LV dysfunction as the result of myocardial stunning during daily activities.
Intracellular calcium plays a key role in myocardial contraction and relaxation, and alterations in calcium homeostasis are thought to be involved in the pathogenesis of stunning.9 10 11 12 Animal studies have shown that the administration of calcium antagonists during ischemia and before reperfusion can attenuate myocardial stunning.13 14 15 16 17 18 19 20 To determine whether the calcium antagonist amlodipine could attenuate exercise-induced myocardial stunning in humans, we performed a double-blind crossover study in patients with significant coronary artery disease and chronic stable angina. Quantitative stress echocardiography was used to assess postexercise LV dysfunction, and nuclear perfusion imaging was used to quantify exercise-induced ischemia. Amlodipine was compared with an equivalent anti-ischemic dose of isosorbide mononitrate (ISMN), which has no direct effect on myocardial function, in an attempt to match the ischemic burden on both treatments. Potential beneficial effects of amlodipine on stunning would therefore not be attributable to an anti-ischemic action because both treatments should reduce ischemia equally.
We studied 24 patients with chronic stable angina, normal resting LV function as assessed by echocardiography, and significant coronary artery disease (≥70% stenosis in 2 major epicardial vessels demonstrated on coronary arteriography <1 month before randomization). Patient demographics are shown in Table 1⇓. Patients were excluded if there was a history of recent myocardial infarction, unstable angina, uncontrolled hypertension, or concomitant disease that precluded treadmill exercise. Patients were also excluded if it was not possible to obtain echocardiograms of sufficient quality to perform quantitative analysis (1 patient). The study protocol was approved by the Research Ethics Committee of the University Hospital of Wales, and all patients gave full written informed consent.
After initial evaluation, patients were randomly assigned, in a double-blind fashion, to receive either oral amlodipine or ISMN for 3 weeks. After a 2-week washout phase, the patient crossed over to the alternative treatment for a further 3 weeks. The amlodipine regime comprised 5 mg once daily for 7 days followed by 10 mg once daily for 2 weeks. The ISMN regime comprised 25 mg once daily for 7 days followed by 50 mg once daily for 2 weeks. With both regimens, adjustment to the higher dose was subject to adverse events and clinical discretion. During the study, all patients received metoprolol at a stable dosage of 50 mg twice daily (although discontinued before exercise testing) and aspirin 75 mg once daily. All other antianginal medication was discontinued during the study period.
Exercise testing was performed according to the Standard Bruce Protocol with a Marquette treadmill at the end of each treatment period 4 to 8 hours after the previous dose of study medication. β-Blockade was withdrawn before exercise by halving and discontinuing metoprolol at 48 hours and 24 hours, respectively. Patients were advised to avoid activities that may precipitate angina for 24 hours before the test and to use prophylactic glyceryl trinitrate up until 4 hours before study. Patients were excluded if they had angina or used prophylactic glyceryl trinitrate within 4 hours before the test. Three ECG leads (II, V2, and V5) were continuously monitored during exercise and recovery. Blood pressure (BP) reading and 12-lead ECG were performed every 3 minutes during exercise and every 2 minutes in recovery until resolution of ECG changes and a return of heart rate and BP to baseline values. Patients exercised until they experienced chest pain and continued until limited by their symptoms. Exercise was also discontinued if a drop in systolic BP of ≥10 mm Hg or significant arrhythmias occurred with exercise.
Two-dimensional echocardiograms were obtained with the patient in the left lateral decubitus position with commercially obtainable equipment (Hewlett-Packard Sonos 1500, model 77025A) with a 2.5-MHz phased array transducer. All studies were performed by the same investigator (C.A.R.). To ensure reproducibility, the position of the patient on the couch was documented at baseline and used for each image acquisition and the position of the transducer on the patient’s chest was marked. To minimize beat-to-beat variation, all recordings were made in gently held expiration. Studies were recorded on super VHS videotape for later off-line analysis. LV contractile function was assessed in the apical 2-chamber (AP2CH) and apical 4-chamber (AP4CH) views as described in the guidelines from the American Society of Echocardiography.21 Pulsed wave (PW) Doppler tracings of LV diastolic inflow at the tips of the mitral valve leaflets were recorded at rest with simultaneous recording of the phonocardiogram (in the second left intercostal space). Images were recorded at rest, immediately after exercise, and then at 15-minute intervals for 4 hours after exercise.
The videotaped images were analyzed by two experienced observers (C.A.R, A.Z.L.) using an off-line PC-based digitizing program (Thoraxcenter, Erasmus University, Rotterdam, the Netherlands).22 Three consecutive beats (excluding extrasystolic and postextrasystolic beats) were analyzed for each time point in the AP2CH and AP4CH views. Endocardial contours were traced (excluding papillary muscles) at end diastole timed as the closure of the mitral valve leaflets and at end systole defined as the point of maximal inward excursion of the endocardial contour. The centerline method was used to assess regional LV function23 24 in which the computer superimposes the end-diastolic and end-systolic endocardial tracings and calculates a centerline between the two. The deviation from the centerline of 100 chords around the LV circumference is calculated after correction for the end-diastolic circumference and expressed as a percentage shortening fraction (SF). Each apical view of the LV is divided into 6 segments and the SF of the chords in each segment is averaged so that a total of 12 values are obtained (6 AP4CH+6 AP2CH). We calculated SF for all 12 segments (SFmean), for segments with a wall motion abnormality (SFwma), and in the segments that developed the most severe wall motion abnormalities (WMAs) after exercise (SFworst). In preliminary studies we obtained a 10% variability of SF in normal regions; for this study a change in SF from baseline of >20% was taken as abnormal (ie, WMA). A persistent WMA abnormality was one defined as present 15 and or 30 minutes after exercise. LV volumes at end diastole and end systole were calculated by using the biplane disk method, and global ejection fraction (EF) (%) was derived by using the formula Analysis of diastolic indexes were performed by using a digitizing tablet and tracing PW Doppler LV inflow signal (5 consecutive beats for each time point). Multiple diastolic parameters were derived including the isovolumic relaxation period (IRP), measured as the interval between aortic valve closure from the phonocardiogram and the onset of the E wave on the Doppler trace. Other indexes included E wave maximal velocity, A wave maximal velocity, E/A wave ratio, time velocity integral of the E wave, deceleration rate, and deceleration time of the E wave.
Myocardial Perfusion Scans
Single-photon emission computerized tomography (SPECT) with 99-Technetium sestamibi (sestamibi SPECT) was performed on 3 separate occasions for each patient. A baseline scan was initially performed on a separate day at rest with the patients taking no active study medication to demonstrate resting myocardial perfusion. The two peak exercise scans were performed at the end of each treatment period. On each occasion a single peripheral intravenous injection of 400 MBq sestamibi was given. For the peak stress scans sestamibi was given at peak exercise and imaging was performed 70 minutes after exercise. Tomographic imaging was performed with an Elscint SP4 gamma camera and high-resolution low-energy collimator. The data were quantified with a dedicated PC-based workstation. Two experienced observers, (P.G.A., E.J.) blinded to the clinical and echocardiographic data analyzed the SPECT images. A 4-segment model dividing the ventricle into anterior, posterior/lateral, inferior, and septal regions was used. Each segment was given a score from 0 to 3, in which 3=normal perfusion, 2=mild perfusion defect, 1=moderate perfusion defect, and 0=severe perfusion defect. Peak exercise studies were compared with the rest study, and a change in a segment score of >1 was considered abnormal. Total perfusion scores were calculated by adding the sum of each segmental score.
The primary efficacy analyses were based on a comparison of the effects of amlodipine and ISMN on the echocardiographic parameters of LV dysfunction after exercise. These were SFworst, SFmean, SFwma, EF, and IRP. Secondary efficacy parameters were time from stunning to recovery, total sestamibi SPECT scores, maximum ST-segment depression, and total exercise duration. The safety of amlodipine and ISMN was assessed from adverse events and intercurrent illnesses during treatment.
A sample size of 24 patients was calculated to have sufficient power to give an 80% chance of detecting a significant difference between treatments at the 95% confidence level. The evaluation of efficacy was based on the intent-to-treat population with a subanalysis performed on the efficacy evaluable population. The echocardiographic parameters and maximum ST-segment depression were subjected to ANOVA if data met the homogeneity of variances assumptions; Wilcoxon rank sum tests were used to analyze data that did not meet those assumptions. Recovery times for patients who showed stunning were analyzed with logistic regression. Sestamibi SPECT scores were analyzed with Wilcoxon rank sum test. Significance values were 10% for period by treatment interactions and 5% otherwise. Adverse event data were analyzed with McNemar’s test.
Patient demographic details are shown in Table 1⇑. The mean age of patients was 60.25 years (range 47 to 75). All patients had a primary diagnosis of angina (mean duration 3.54 years). Seventeen (71%) patients had 3-vessel coronary disease and 7 (29%) had 2-vessel disease. All patients except 1 were of Caucasian origin (patient 24: Asian origin). The mean weight was 85.35 kg (range 56 to 124 kg). Most patients had at least 1 concomitant condition, most commonly hypertension and hypercholesterolemia.
Primary Efficacy: Echocardiography
Postexercise stunning was defined as a new wall motion abnormality (change in SF >20% from baseline) present at 15 and/or 30 minutes after exercise in myocardial segments that had matched perfusion defects at peak stress. Overall, fewer patients exhibited stunning during amlodipine treatment than while receiving ISMN (48% versus 82%, P=0.028).
Most wall motion abnormalities were present at 15 minutes after exercise, and the majority of abnormalities had recovered by 45 minutes after exercise.
Table 2⇓ and Figures 1⇓, 2⇓, and 3⇓ demonstrate the systolic and diastolic echocardiographic parameters after exercise. The difference between treatment groups in time to recovery was not statistically significant (P=0.887). Analysis of percentage change from preexercise to 15 or 30 minutes after exercise (whenever lower SFworst occurred) showed that exercise-induced changes were less with amlodipine than ISMN for all echocardiographic parameters.
Mean echocardiographic parameters at 15 and 30 minutes after exercise and the P values for analysis of preexercise to postexercise changes are summarized in Table 3⇓.
All echo cardiographic data met the assumption of homogeneity of variances required for using ANOVA.
There was a significant difference between treatments in favor of amlodipine at 15 and 30 minutes after exercise for SFworst (3.3±1.7% versus 2.3±1.3% and 3.5±1.5% versus 2.5±1.4%, P=0.014) and SFwma (2.8±0.8% versus 2.5±0.8% and 3.0±0.7% versus 2.6±0.9%, P=0.008). The differences between SFmean were also highly significant in the same direction (3.7±0.4% versus 3.1±0.5% and 3.6±0.4% versus 3.2±0.5%, P=0.001), as were differences in EF after exercise (59.5±4.9% versus 53.1±7.9% and 59.7±5.4% versus 54.5±8%, P=0.0001). Nonparametric analysis of IRP revealed the effect of treatment was also highly significant in favor of amlodipine (88.5±15.6 versus 106.6±18.9 ms and 93±15.5 versus 106.3±14.9 ms, P=0.0018).
Figure 4⇓ demonstrates an example of the beneficial effect of amlodipine on postexercise stunning in a study patient with 3-vessel disease.
Secondary Efficacy Parameters
The mean sestamibi scores for each segment and mean total scores at baseline and at peak stress on both treatments are shown in Table 4⇓. Nonparametric analysis of total MIBI SPECT scores revealed no significant difference in the amount of ischemia at peak stress between amlodipine and ISMN (8.5±2.2 versus 7.9±1.8, P=0.172, Koch’s test). (See Figures 5⇓ and 6⇓.)
Exercise Test Parameters
Exercise test results were similar on both treatments. All exercise parameters including ST-segment depression, chest pain, BP, and heart rate returned to baseline within 10 minutes of exercise cessation in all cases. There was no significant difference for maximum ST-segment depression (1.8±1.3 versus 1.6±.9 mm, P=0.482, ANOVA). Similarly, total exercise duration (397±170 versus 404±171 seconds, P=1, Wilcoxon rank sum test) was unchanged between amlodipine and ISMN. (See Table 5⇓.) There was no difference between blood pressure measurements between treatments during and after exercise.
Adverse events were more frequent with ISMN than with amlodipine (46% versus 15%, P=0.008, McNemar’s test), with the most commonly reported adverse event being headache.
The exercise-induced ischemia produced in this study caused significant regional wall motion abnormalities, reduced global LV ejection fraction, and lengthening of IRP, which was prolonged but eventually recovered and thus fulfills the criteria for stunning. These abnormalities did not occur or were significantly less marked when the patients were receiving amlodipine compared with ISMN.
Broadly, this very marked attenuation in stunning could have two explanations that are not mutually exclusive. The amlodipine may have reduced the ischemia more than the nitrates, and reduced stunning was simply due to less causative ischemia, or it may have an intrinsic antistunning effect. The results of this study and the results of previous basic and animal work strongly favor the second explanation. Most importantly, exercise time was the same in the two situations, whereas the tests of the degree of ischemia both by myocardial perfusion imaging and ST-segment analysis did not show any evidence for the reduced ischemia hypothesis, and in fact there was a slight trend in the opposite direction. Furthermore, there was no evidence of beneficial changes in hemodynamics in the amlodipine group (such as reduced afterload or negative inotropic effect) that could have reduced ischemia. In our study the favorable hemodynamic and anti-ischemic effects (defined and quantified by sestamibi SPECT) of amlodipine and ISMN were matched by virtue of the crossover design.
Previous experimental work also favors the hypothesis that calcium antagonists have a direct antistunning effect, possibly because calcium overload has a pathogenetic mechanism. Calcium antagonists confer protection from stunning in isolated hearts with improved recovery of function after global ischemia and reperfusion.13 14 Nifedipine, diltiazem, and verapamil reduce stunning in vivo, but it is unclear whether this is due to favorable hemodynamic changes that reduce cardiac work or a direct effect on the myocyte by a reduction in intracellular calcium.17 18 In conscious dogs instrumented to produce chronic coronary stenoses, recovery of exercise-induced contractile dysfunction is shortened (5 versus 30 minutes) in dogs pretreated with diltiazem,25 although this benefit may have been due to increased subendocardial flow in the ischemic zone. Intracoronary injection of calcium antagonists (thereby avoiding systemic hemodynamic effects) during coronary angioplasty has shown a protective effect on stunning, suggesting a direct effect caused by a reduction in intracellular calcium.26
In a recent study in humans, systolic and diastolic dysfunction after exercise was attenuated by sublingual nifedipine but not nitroglycerin administered at peak exercise,27 and our findings are in agreement with this. This previous study, however, had several limitations. It was unblinded and did not use patients as their own controls. Furthermore, not all patients had definite coronary disease confirmed by arteriography, and there were favorable changes in BP after exercise with nifedipine that may have contributed to its beneficial effect by a reduction in cardiac work, which might have reduced ischemia and indirectly the degree of stunning. Similarly, not all patients underwent myocardial perfusion scans and therefore a significant difference in exercise-induced ischemia between the groups may have contributed to the reduction in stunning seen with nifedipine. Our study has controlled for these factors to a much greater degree and does not have these shortcomings. Its other main difference is that it is a study of the long-term therapy of the type that would be given to such patients in real life.
This finding may have important therapeutic implications, since it is likely that stunning occurs after episodes of ischemia during everyday life, which could result in prolonged episodes of LV dysfunction. It has been postulated that repetitive stunning of the type shown by us may result in chronic LV impairment.28 Identification of clinically available therapeutic strategies that attenuate stunning and therefore lessen the impact of ischemia on LV function represent an important step in improving the medical management of patients with coronary artery disease.
There are limitations to the use of endocardial wall excursion as the sole criterion of abnormal contractile function; neither wall thickening nor transmyocardial heterogeneity is taken into account. However, it does provide a quantitative method that is repeatable and it is likely that only subtle abnormalities of regional LV function are missed. The use of a semiquantitative scoring system for myocardial perfusion scintigraphy in SPECT may be subject to error, but again it is likely that only subtle abnormalities of myocardial perfusion were missed.
In conclusion, our data demonstrate a beneficial effect of amlodipine on postexercise LV dysfunction, and is to our knowledge the first ever randomized, double-blind study to show a beneficial effect of calcium antagonists (or any other therapy) on stunning. It is very unlikely that this benefit was mediated by a mechanism directly related to its anti-ischemic or hemodynamic effects because these parameters did not change significantly, although the exact mechanism remains undetermined.
This study was supported by a grant from Pfizer Ltd.
- Received December 15, 1997.
- Revision received March 30, 1998.
- Accepted April 21, 1998.
- Copyright © 1998 by American Heart Association
Braunwald E, Kloner RA. The stunned myocardium: prolonged, postischemic ventricular dysfunction. Circulation. 1982;66:1146–1149.
Bolli R. Mechanism of myocardial stunning. Circulation. 1990;82:723–738.
Bolli R. Myocardial “stunning” in man. Circulation. 1992;86:1671–1691.
Patel B, Kloner RA, Przyklenk K, Braunwald E. Postischemic myocardial “stunning”: a clinically relevant phenomenon. Ann Intern Med. 1988;108:626–628.
Homans DC, Laxson DD, Sublett E, Pavek T, Crampton M. Effect of exercise intensity and duration on regional function during and after exercise-induced ischemia. Circulation. 1991;83:2029–2037.
Ambrosio G, Betocchi G, Pace L, Losi MA, Perrone-Filardi P, Sorricelli A, Piscione F, Taube J, Squame F, Salvatore M, Chiariello M. Prolonged impairment of regional contractile function after resolution of exercise-induced angina: evidence of myocardial stunning in patients with coronary artery disease. Circulation. 1996;94:2455–2464.
Kusuoka H, Porterfield JK, Weisman HF, Weisfeldt ML, Marban E. Pathophysiology and pathogenesis of stunned myocardium. J Clin Invest. 1987;79:950–961.
Kitakaze M, Weisman HF, Marban E. Contractile dysfunction and ATP depletion after transient calcium overload in perfused ferret hearts. Circulation. 1988;77:685–695.
Kitakaze M, Weisfeldt ML, Marban E. Acidosis during early reperfusion prevents myocardial stunning in perfusion ferret hearts. J Clin Invest. 1988;82:920–927.
Steenbergen C, Murphy E, Levy L, London RE. Elevation in cytosolic free calcium concentration early in myocardial ischemia in perfused rat heart. Circ Res. 1987;60:700–707.
du Toit EF, Opie LH. Modulation of severity of reperfusion stunning in the isolated rat heart by agents altering calcium flux at onset of reperfusion. Circ Res. 1992;70:960–967.
Przyklenk K, Kloner RA. Calcium antagonists and the stunned myocardium. J Cardiovasc Pharmacol. 1991;18(suppl 10):S93–S101.
Fujibayashi Y, Yamazaki S, Chang BL, Rajagopalan RE, Meerbaum S, Corday E. Comparative echocardiographic study of recovery of diastolic versus systolic function after brief periods of coronary occlusion: differential effects of intravenous nifedipine administered before and during occlusion. J Am Coll Cardiol. 1985;6:1289–1298.
Bush LR, Buja LM, Tilton GD, Wathen M, Apprill P, Ashton J, Willerson JT. Effect of propranolol and diltiazem alone and in combination on the recovery of left ventricular segmental function after temporary coronary occlusion and long-term reperfusion in conscious dogs. Circulation. 1985;72:413–430.
Tilton GD, Bush LR, Apprill PG, Buja LM, Willerson JT. Effect of diltiazem and propranolol on left ventricular segmental relaxation during temporary coronary occlusion and long-term reperfusion in conscious dogs. Circulation. 1985;71:165–175.
Sheehan FH, Bolson EL, Dodge HT. Advantages and applications of the centerline method for characterizing regional ventricular function. Circulation. 1986;74:293–305.
Matsuzaki M, Gallagher KP, Patritti J, Tajimi T, Kemper WS, White FC, Ross J Jr. Effects of a calcium entry blocker (diltiazem) on regional myocardial flow and function during exercise in conscious dogs. Circulation. 1984;69:801–814.
Sheiban I, Tonni S, Benussi P, Marini A, Trevi GP. Left ventricular dysfunction following transient ischemia induced by transluminal coronary angioplasty: beneficial effects of calcium antagonists against post-ischaemic myocardial stunning. Eur Heart J. 1993;14(suppl A):14–21.
Marinho NVS, Keogh BE, Costa DE, Lammerstma A, Ell PJ, Camici PG. Pathophysiology of chronic left ventricular dysfunction. Circulation. 1996;93:737–744.Myocardial stunning can be attenuated by calcium antagonists in animal models. To assess their effect in humans, we performed a randomized, double-blind crossover study of 24 patients comparing amlodipine (10 mg once daily) versus isosorbide mononitrate (50 mg once daily) on postexercise stunning. Patients underwent serial quantitative exercise stress echocardiography on each treatment to assess the degree of stunning with simultaneous sestamibi single-photon emission computed tomography to quantify ischemia. Exercise parameters were unchanged between treatments. Stunning occurred more often while patients were receiving ISMN than amlodipine (82% versus 48%). Systolic and diastolic dysfunction was attenuated in patients while taking amlodipine compared with ISMN despite comparable ischemia burdens.