Preconditioning by Transient Myocardial Ischemia Confers Protection Against Ischemia-Induced Ventricular Arrhythmias in Variant Angina
Background In experimental models, ischemic preconditioning of the heart protects against ischemic damage and ventricular arrhythmias during subsequent coronary occlusion. In this study, we investigated whether protection against ischemic suffering and ischemia-induced arrhythmias may occur after spontaneous transmural ischemia in humans.
Methods and Results We performed 24-hour Holter monitoring in 10 patients with variant angina who developed complex ventricular arrhythmias (CVAs, more than five premature ventricular beats per minute or repetitive ventricular arrhythmias) during episodes of ST-segment elevation. A total of 150 episodes of ST-segment elevation were detected on Holter monitoring, 21 (14%) of which showed CVAs. Episodes separated from the previous one by a time interval of ≤30 minutes or by a time interval of >30 minutes did not differ in either magnitude or duration of ST-segment elevation, but CVAs occurred more frequently in the second group (3% versus 29%, P<.0001). The time interval from the preceding ischemic episode was longer for the episodes with compared with those without CVAs (197±275 versus 57±87 minutes, P<.001), but these two groups of episodes also had similar severities and durations of ST-segment elevation. Finally, when we analyzed 13 clusters of two to six ischemic episodes, CVAs were found much more frequently in the first (92%) than in the last (23%, P=.009) episode of the clusters, while ST-segment elevations were similar (2.1±1.6 versus 2.2±1.1 mm) and ischemia durations shorter in the first than in the last episode (3.9±3.6 versus 6.1±1.7 minutes, P=.03).
Conclusions Our data indicate that preconditioning by transient ischemia induces a significant protection against ischemia-induced CVAs in patients with variant angina. This beneficial effect was not related to a reduction in either severity or duration of ischemia, suggesting that arrhythmic protection was a direct consequence of preconditioning rather than an epiphenomenon of ischemic protection.
Preconditioning of myocardium by brief exposure to ischemia was first found to reduce myocardial damage after a subsequent prolonged coronary artery occlusion in dogs.1 Ischemic preconditioning has been consistently confirmed in a number of different animal species2 3 4 and by different experimental protocols,5 and recent studies on myocardial specimens6 and clinical investigations during coronary angioplasty7 8 suggest that this phenomenon may also occur in humans.9 The clinical relevance of ischemic preconditioning is still debated, however, although it has been suggested that preconditioning may be responsible for the reported beneficial effect of preinfarction angina on infarct size10 and that it might be used to improve myocardial protection during cardiac surgery.11 12
Several experimental studies have also found that preconditioning may reduce the incidence and severity of ventricular arrhythmias occurring during or in the recovery phase of ischemia.13 14 15 16 However, it is not clear whether this arrhythmic protection is secondary to ischemic protection or is an independent phenomenon with peculiar pathophysiological and cellular mechanisms.17 Furthermore, it is not known whether preconditioning can reduce ischemia-induced arrhythmias in humans.
Frequent, and sometimes life-threatening, ventricular arrhythmias are often detected during episodes of transmural ischemia in patients with variant angina,18 which is caused by occlusive coronary artery spasm19 and is characterized by ST-segment elevation on the ECG. In the “hot” phase of the disease, these patients frequently have recurrent episodes of transmural myocardial ischemia, often grouped in clusters and associated with ventricular arrhythmias, thus offering an ideal model to study the effect of preconditioning in humans. In this study, we investigated whether spontaneous ischemic episodes may induce ischemic and arrhythmic protection in patients with variant angina and the possible relationship between these two phenomena.
Between February 1992 and November 1995, 27 patients (18 men; mean age, 59±11 years) with a final diagnosis of variant angina were admitted to our coronary care unit. All patients underwent 24-hour ECG Holter monitoring off therapy, except for sublingual nitrates, which were given promptly to relieve chest pain.
To assess the effects of preconditioning on ventricular arrhythmias, we selected for this study only patients who had both (1) a significant number (three or more) of ischemic episodes during 24-hour Holter monitoring and (2) complex ventricular arrhythmias (CVAs) during at least one of these ischemic episodes. According to accepted clinical classifications of arrhythmias related to acute ischemic syndromes,20 CVAs were defined as more than five premature ventricular beats (PVBs) per minute of ischemia and/or repetitive forms of ventricular arrhythmias (couplets and/or runs of ventricular tachycardia, defined as three or more PVBs with a rate >100 bpm).
There were 10 patients (37%) who fulfilled these inclusion criteria (8 men; age, 59±9 years). The occurrence of CVAs was clearly related to ischemia in these patients, because they had no evidence of arrhythmias outside of the ischemic episodes. Conversely, none of the remaining 17 patients had ventricular arrhythmias apart from rare, isolated PVBs (<20 in the 24 hours).
All 10 patients included in the present study reported a typical history of variant angina, with anginal chest pain occurring at rest, lasting only a few minutes, and promptly responding to sublingual nitrates. Anginal episodes were of recent onset (days to weeks) in all these patients, none of whom were taking any anti-ischemic drugs at admission to the coronary care unit. No patient had diabetes, arterial hypertension, or basal alterations on the ECG that could interfere with ST-segment analysis, and all showed a normal global and regional left ventricular function on echocardiographic examination.
Chest pain and ST-segment elevation were also reproducible by hyperventilation and/or ergonovine infusion, either intravenous or intracoronary, in all patients with evidence of occlusive coronary spasm of a major epicardial artery during coronary angiography, which was performed within 3 days of admission. Coronary angiograms were independently evaluated visually by two expert cardiologists, who classified coronary vessels either as normal or with noncritical or critical stenoses. There was full agreement in the classification of coronary angiograms between these two investigators. Critical coronary stenoses (defined as a reduction of internal luminal diameter >70%) were present in at least one epicardial coronary vessel in 4 patients, whereas 5 of the remaining 6 patients had normal coronary arteries and 1 had noncritical stenoses. No patient had evidence of multiple epicardial coronary spasm at angiography.
All patients gave written informed consent to participate in the study, which was in agreement with the general guidelines of the Ethical Committee of our institution.
Holter monitoring was performed with two-channel tape recorders (Oxford Medilog 4500). In each patient, a CM5 lead was always recorded on the first channel, while the bipolar chest lead most similar to the standard ECG lead showing the maximal ST-segment changes during angina was selected for monitoring on the second channel (CM3 for anterior and a modified aVF lead for inferior ST-segment elevation). Patients filled in a detailed structured diary describing their activities and symptoms. The recordings were analyzed by two expert cardiologists using the Oxford Medilog Excel 2.0 device. ST-segment elevation ≥1 mm at 0.08 second after the J-point and lasting ≥1 minute was considered significant. The following parameters were obtained for each episode of ST-segment elevation: duration and magnitude of the ST shift, time interval (in minutes) from the end of the previous episode, and occurrence of CVAs. We also obtained heart rate at 5 minutes before ischemia, at the onset of ST-segment elevation, and at peak ST to evaluate its changes before and during ischemia. Instantaneous heart rate was calculated on 8-second ECG strips during sinus rhythm.
Evaluation of Preconditioning
To assess whether ischemic and/or arrhythmic protection by preconditioning occurred in these patients, the following analyses were performed.
1. We divided all the ischemic episodes that occurred during Holter monitoring into two groups: those separated from a previous episode by a time interval of ≤30 minutes and those separated from a previous episode by a time interval of >30 minutes, and then we compared the severity of ischemia and the incidence of CVAs observed in these two groups of episodes. The 30-minute cutoff to evaluate preconditioning was arbitrary, since the duration of the protective action of preconditioning in humans is not known. On the basis of experimental and clinical findings, however, we assumed that, if present, preconditioning should be operating for at least this time,21 whereas its full effects are less likely to extend beyond this time.22 23
2. We compared the time interval that separated each ischemic episode associated with CVA from the previous episode with the corresponding time interval of the episodes not associated with CVAs.
3. We carried out a further specific analysis for the episodes of ST-segment elevation grouped in clusters, which were defined as two or more episodes occurring within a time period of 30 minutes with no detectable ST changes in the 40 minutes preceding the first episode of the cluster (Fig 1⇓). These time intervals were chosen according to experimental models, showing that preconditioning by brief coronary occlusions (2 to 5 minutes, very similar to that underlying transmural ischemia in our patients with variant angina) does reduce ventricular arrhythmias during a subsequent coronary occlusion when this occurs after a time interval of ≤30 minutes16 but not when it occurs after >40 minutes.13 The first and the last episodes of each cluster were compared as to severity and duration of ST-segment elevation, prevalence of CVAs, total number of PVBs, and frequency of PVBs per minute of ischemia. All clusters were included in the analysis to evaluate the presence of ischemic protection, whereas only clusters showing CVAs in the first and/or the last ischemic episode were considered to investigate the antiarrhythmic effect of preconditioning.
In each patient, the first ischemic episode of the recording was excluded when it occurred within 30 minutes from the beginning of the recording (40 minutes in the case of cluster analysis), whereas it was included when this time was >30 minutes. In this case, the time interval from the previous episode was taken as that from the beginning of Holter monitoring.
Continuous variables were compared by Mann-Whitney U test or Wilcoxon signed-rank test, as indicated. Proportions were compared by two-sided Fisher's exact test. The changes of heart rate associated with ischemic episodes were evaluated by ANOVA for repeated measures, and multiple comparisons were done by paired t test with Bonferroni correction in case of statistical significance. Correlations were tested by Spearman rank correlation test. All values are expressed as mean±SD, when not indicated differently. A value of P<.05 was considered significant.
A total of 150 episodes of ST-segment elevation were detected during Holter monitoring (mean, 15±9 episodes per patient; range, 3 to 31 episodes), 44 (29%) of which were associated with angina. No patient had episodes of ST-segment depression. The mean duration of episodes was 3.6±2.6 minutes (range, 1 to 20 minutes), and the magnitude of the ST-segment shift was 2.3±1.3 mm (range, 1 to 6). The mean time interval between adjacent ischemic episodes was 77±139 minutes (median, 19 minutes; range, 1 to 1080 minutes). There were no changes in heart rate at the onset of ST-segment elevation compared with 5 minutes before (65±13 versus 64±10 bpm), whereas there was a mild increase at peak ST-segment elevation (69±15 bpm, P=.01 versus 5 minutes before ischemia and the onset of ST-segment elevation).
Ischemic Episodes With a Time Interval of ≤30 Minutes or >30 Minutes From the Previous One
Ischemic episodes separated by a time interval of ≤30 minutes from the preceding one (n=87) did not differ significantly in either severity (2.3±1.4 versus 2.2±1.2 mm) or duration (3.3±2.2 versus 3.9±3.2 minutes) of ST-segment elevation, compared with those separated by a time interval of >30 minutes (n=63). Yet, CVAs were detected more frequently in this latter group (3% versus 29%, P<.0001).
Ischemic Episodes With or Without CVAs
CVAs were detected in 21 ischemic episodes (14%; range, 1 to 4 episodes per patient). The main features of ischemic episodes with or without CVAs are summarized in Table 1⇓. There were no differences in severity or duration of ST-segment elevation in the two groups. The time interval between each episode with CVAs and the previous one was significantly longer than the corresponding interval of episodes without CVAs (197±275 versus 57±87 minutes, P<.001). Sinus heart rate at peak ST was higher in episodes associated with CVAs (77±20 versus 66±13 bpm, P<.002).
There was a tendency toward an inverse correlation between the prevalence of ischemic episodes with CVAs and both the total number of ischemic episodes (r=−.46, P=.18) and total ischemic time in the 24 hours (r=−.60, P=.07).
Repetitive Ventricular Arrhythmias
Similar results were observed when we performed a separate analysis considering only repetitive forms of ventricular arrhythmias (Table 2⇓). Overall, repetitive ventricular arrhythmias were detected in 9 of 47 episodes of ST-segment elevation occurring in 5 patients. Runs of ventricular tachycardia were found in 2 episodes, couplets in 2, and both in 5 episodes. In these patients, neither severity nor duration of ST-segment elevation was significantly different for ischemic episodes associated or not associated with repetitive ventricular arrhythmias, whereas the time interval from the preceding episode was longer for the episodes associated with repetitive ventricular arrhythmias (310±394 versus 79±108 minutes, P=.008). Furthermore, repetitive ventricular arrhythmias were detected more frequently in ischemic episodes separated by a time interval of >30 minutes than in ischemic episodes separated by a time interval of ≤30 minutes from the previous one (15% versus 1%, P=.0095).
A total of 75 ischemic episodes (50% of the total) occurred in 30 clusters, 22 of which were constituted of 2 episodes, 3 of 3 episodes, 4 of 4 episodes, and 1 of 6 episodes. The main characteristics of the first and last episodes of the clusters are summarized in Table 3⇓. There were no significant differences between these 2 episodes in either severity (1.8±0.8 versus 1.9±1.3 mm, P=.76) or duration (3.6±2.9 versus 4.1±2.8 minutes, P=.38) of ST-segment elevation or in heart rate before and during ischemia. On the whole, CVAs occurred more frequently during the first than during the last episodes of clusters (40% versus 10%, P<.01).
There were 13 clusters (46%) occurring in 6 patients showing CVAs in the first and/or last ischemic episode. Nine of these clusters were composed of 2 ischemic episodes, 1 of 3, and the remaining 3 of 4 episodes. The first episode of each cluster occurred after a time interval of 155±103 minutes (range, 52 to 375 minutes) from the previous ischemic episode, whereas the last episode occurred within only 11±6 minutes (range, 1 to 24 minutes) from the previous one. There was no statistically significant difference between the first and the last episodes of these clusters in severity of ST-segment elevation (2.2±1.1 versus 2.1±1.6 mm, P=.57, Fig 2A⇓), but the duration of ischemia was longer in the last episode (3.9±3.6 versus 6.1±1.7 minutes, P=.03, Fig 2B⇓). CVAs were found in the first episode of 12 of the 13 clusters (92%) but in only 3 (23%) of the last episodes (P=.005). CVAs were observed only in the first episode in 10 clusters, only in the last episode in 1, and in both the first and last ischemic episodes in 2. Furthermore, both the total number of PVBs (15±20 versus 101±100, P=.006, Fig 3A⇓) and the number of PVBs per minute of ischemia (2.5±3.5 versus 25±10, P=.002, Fig 3B⇓) were strikingly lower in the last than in the first episode of the clusters.
There was no difference in sinus heart rate before and during ischemia between the first and last episodes of the clusters (66±14 versus 63±7 bpm 5 minutes before ischemia; 69±15 versus 64±7 bpm at onset of ischemia; 75±18 versus 71±9 bpm at peak ST). The trend of heart rate was also similar for the first and last ischemic episodes even when only the 10 clusters showing CVAs in the first but not in the last episode were analyzed (69±15 versus 64±7 bpm 5 minutes before ischemia; 70±16 versus 67±8 bpm at onset of ischemia; and 77±16 versus 73±10 bpm at peak ST). Fig 4⇓ shows two examples of our observations.
Analysis of Preconditioning in Patients With or Without Coronary Stenoses
Similar results were obtained when we analyzed separately the ischemic episodes occurring in the 4 patients with and in the 6 patients without significant coronary artery stenoses at angiography.
In this latter group, the time interval between each episode with CVAs (n=14 of 91, or 15%) and the previous one was longer than the corresponding interval of episodes without CVAs (240±336 versus 48±83 minutes, P<.001), and there were 79±60 PVBs in the first compared with 17±21 PVBs in the last episode of 7 clusters (P=.02).
Similarly, in the 4 patients with significant coronary stenoses, the time interval from the previous ischemic episode was 144±114 minutes for the episodes with CVAs (n=7 of 59, or 12%) and 71±102 minutes for those without CVAs (P=.02), and there were 126±136 PVBs in the first and 13±21 PVBs in the last episode of 6 clusters (P=.04). Again, no significant differences were detected in severity or duration of ST-segment elevation between episodes with and without CVAs and between the first and the last episodes of clusters detected in both these subgroups of patients.
Our study demonstrates that, in patients with variant angina, transient myocardial ischemia confers protection against arrhythmias during subsequent episodes of spontaneous myocardial ischemia in the absence of any evidence of ischemic protection. Thus, our study shows that, at least in this human model, arrhythmic protection is not a mere consequence of ischemic protection but rather is caused directly by the ischemic preconditioning.
The protective effect of preconditioning by brief episodes of ischemia on myocardial damage produced by a subsequent prolonged coronary occlusion was first demonstrated by Murry et al1 in a canine model and was subsequently confirmed in a number of different animal species.3 4 24 Recently, ischemic preconditioning has been demonstrated in human myocardial specimens,11 and it has been suggested that it may mediate the protective effect of preinfarction angina on infarct size,10 the “warm-up” phenomenon, and “walk-through angina.”25 Convincing evidence that preconditioning can occur in humans, however, has been obtained primarily in studies performed during coronary angioplasty.7 8 In this setting, the degree of ischemia, as indicated by the severity of ST-segment elevation on surface and/or intracoronary ECG leads and by cardiac lactate production, has been found to be greater during the first than during the subsequent coronary occlusions induced by balloon inflation.
Conversely, in our study in patients with variant angina, ischemic preconditioning was not associated with a significant reduction in the degree of ischemia. Indeed, there were no differences in either severity or duration of ST-segment elevation between episodes separated by a time interval of ≤30 minutes and those separated by a time interval of >30 minutes; moreover, the severity of ST-segment elevation was not reduced at the very end of a cluster of ischemic episodes, whereas duration of ischemia was even longer.
The reason why we have found different results from those reported in previous studies during coronary angioplasty is not clear. However, differences in duration and/or severity of the preconditioning ischemic episode and/or in intervals between the potential preconditioning stimulus and the subsequent ischemic episodes may, at least in part, account for the discrepancy.
Another possible speculative mechanism to explain the absence of ischemic protection by preconditioning in our study is suggested by the findings of Cohen et al,26 who reported that repeated, intermittent ischemia may induce tolerance to ischemic preconditioning, attenuating its myocardial protection, in rabbits. This observation raises the possibility that tolerance to preconditioning might also be occurring in our patients with variant angina, who actually had frequent episodes of ischemia per day (mean, 15±9 episodes). If this were the case, however, the development of tolerance to preconditioning in our patients should have concerned only ischemic, but not arrhythmic, protection.
Finally, it should be observed that we cannot completely exclude the possibility that the lack of evidence of ischemic protection may depend on a systematic increase in the intensity of constrictor stimuli operating during episodes separated by a shorter time interval, which may offset the beneficial effects of ischemic preconditioning, but this possibility seems to be unlikely.
Several experimental studies have demonstrated that preconditioning not only reduces myocardial damage during subsequent ischemic episodes but also significantly reduces the incidence and severity of both ischemia- and reperfusion-induced ventricular arrhythmias.13 14 15 16 The protection achieved is transient and is lost quite rapidly, with a most pronounced effect in the first 10 minutes.16
Our data demonstrate that ischemic preconditioning may induce arrhythmic protection in humans as well. Indeed, in our patients with variant angina, transient ischemia was associated with a reduction in the prevalence and severity of ischemia-induced ventricular arrhythmias in subsequent ischemic episodes occurring within 30 minutes. Furthermore, the time interval from the previous ischemic episode was shorter for episodes without CVAs than for episodes with CVAs. Finally, although the duration of ischemia was even longer, the occurrence of CVAs and the number of PVBs were strikingly reduced in the last compared with the first of ischemic episodes occurring in clusters, a clinical model similar to the experimental protocols applied in animals.13 14 15 16
The mechanism(s) involved in the antiarrhythmic effect of preconditioning is not known at present17 27 and cannot be deduced from our data. The higher sinus heart rate at peak ST associated with ischemic episodes showing CVAs would have suggested a role for an antiadrenergic and/or an increased vagal effect of preconditioning.28 However, this finding was not confirmed by the analysis of the episodes of clusters, which showed no difference in the trend of heart rate between the first and last episodes even when an antiarrhythmic effect by preconditioning was evident in the last episode.
Ischemic Versus Arrhythmic Protection
Although ischemic preconditioning has been found to induce both ischemic and arrhythmic protection, it is not clear whether the protection against arrhythmias is merely secondary to the anti-ischemic protection.
Previous experimental studies have suggested that the mechanism(s) of the anti-ischemic and antiarrhythmic effects of ischemic preconditioning may actually be different.15 Indeed, although preconditioning both reduces ischemia-related myocardial damage and alters the temporal profile of the adverse metabolic effects of ischemia, thus delaying myocardial necrosis, it reduces ischemia-induced ventricular arrhythmias, but it does not change their temporal profile.29 Furthermore, the protective action of ischemic preconditioning against ischemia-induced arrhythmias can be prevented by inhibition of either cyclooxygenase30 or nitric oxide synthesis.31 This suggests a protective antiarrhythmic role for prostanoid substances and nitric oxide, which do not seem to be involved in the anti-ischemic effect of preconditioning. In our patients, ischemic preconditioning induced a significant arrhythmic protection in the absence of ischemic protection. Hence, our results support the hypothesis that arrhythmic protection is a direct consequence of ischemic preconditioning, not mediated by the ischemic protection.
A limitation of our study is that the degree of myocardial ischemia was assessed by ST-segment changes observed in only two Holter ECG leads. However, previous studies showed ischemic preconditioning during angioplasty by use of similar methods.8 Furthermore, variations in the severity and duration of ST-segment changes on Holter monitoring are well-accepted and currently used parameters to assess the ischemic burden of patients and the effects of anti-ischemic drugs in clinical trials.
To the best of our knowledge, this is the first study to demonstrate a beneficial effect of brief episodes of myocardial ischemia on the occurrence of ischemia-induced ventricular arrhythmias in humans. This arrhythmic protection in patients with variant angina was not associated with ischemic protection, thus suggesting that it is not an epiphenomenon of ischemic protection but rather is directly and independently induced by ischemic preconditioning.
- Received January 24, 1996.
- Revision received April 29, 1996.
- Accepted May 1, 1996.
- Copyright © 1996 by American Heart Association
Murry CE, Jennings RB, Reimer KA. Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium. Circulation. 1986;74:1124-1136.
Li GC, Vasquez BS, Gallagher KP, Lucchesi BR. Myocardial protection with preconditioning. Circulation. 1990;82:609-619.
Schott RJ, Rohmann S, Braun ER, Schaper W. Ischemic preconditioning reduces infarct size in swine myocardium. Circ Res. 1990;66:1133-1142.
Cohen MV, Liu GS, Downey JM. Preconditioning causes improved wall motion as well as smaller infarcts after transient coronary occlusion in rabbits. Circulation. 1991;84:341-349.
Lawson CS, Downey JM. Preconditioning: state of the art myocardial protection. Cardiovasc Res. 1993;27:542-550.
Ikonomidis JS, Tumiati LC, Weisel RD, Mickle DAG, Li RK. Preconditioning human ventricular cardiomyocytes with brief periods of simulated ischemia. Cardiovasc Res. 1994;28:1285-1291.
Deutsh E, Berger M, Kussmaul WG, Hirshfeld JW, Hermann HC, Laskey WL. Adaptation to ischemia during percutaneous transluminal coronary angioplasty: clinical, hemodynamic, and metabolic features. Circulation. 1990;82:2044-2051.
Tomai F, Crea F, Gaspardone A, Versaci F, De Paulis R, Penta de Peppo A, Chiariello L, Gioffre´ PA. Ischemic preconditioning during coronary angioplasty is prevented by glibenclamide, a selective ATP-sensitive K+ channel blocker. Circulation. 1994;90:700-705.
Kloner RA, Shook T, Przyklenk K, Davis VG, Junio L, Matthews RV, Burnstein S, Gibson CM, Poole WK, Cannon CP, McCabe CH, Braunwald E, for the TIMI 4 Investigators. Previous angina alters in-hospital outcome in TIMI 4. Circulation. 1995;91:37-47.
Galin˜anes M, Argano V, Hearse DJ. Can ischemic preconditioning ensure optimal myocardial protection when delivery of cardioplegia is impaired? Circulation. 1995;92(suppl II):II-389-II-394.
Shiki K, Hearse DJ. Preconditioning of ischemic myocardium: reperfusion-induced arrhythmias. Am J Physiol. 1987;253:H1470-H1476.
Hagar JM, Hale SL, Kloner RA. Effects of preconditioning ischemia on reperfusion arrhythmias after coronary artery occlusion and reperfusion in the rat. Circ Res. 1991;68:61-68.
Kerin NZ, Rubenfire M, Naimi M, Wajszczuk WJ, Pamatmat A, Cascade PN. Arrhythmias in variant angina pectoris: relationship of arrhythmias to ST-segment elevation and R-wave changes. Circulation. 1979;60:1343-1350.
Van Winkle DM, Thornton JD, Downey DM, Downey JM. The natural history of preconditioning: cardioprotection depends on duration of transient ischemia and time to subsequent ischemia. Coron Artery Dis. 1991;2:613-619.
Murry CE, Richard VJ, Jennings RB, Reimer KA. Myocardial protection is lost before contractile function recovers from ischemic preconditioning. Am J Physiol. 1991;260:H796-H804.
Marber MS, Joy MD, Yellon DM. Is warm-up in angina ischaemic preconditioning? Br Heart J. 1994;72:213-215.
Cohen MV, Yang XM, Downey JM. Conscious rabbits become tolerant to multiple episodes of ischemic preconditioning. Circ Res. 1994;74:998-1004.
Miyazaki T, Zipes DP. Protection against autonomic denervation following acute myocardial infarction by preconditioning ischemia. Circ Res. 1989;64:437-448.
Lawson CS, Avkiran M, Shattock MJ, Coltard DJ, Hearse DJ. Preconditioning and reperfusion arrhythmias in the isolated rat heart: true protection or temporal shift in vulnerability? Cardiovasc Res. 1993;27:2274-2281.