What is Wrong With Cardiac Conditioning? We May be Shooting at Moving Targets

Abstract

Early recanalization of the occluded culprit coronary artery clearly reduces infarct size in both animal models and patients and improves clinical outcomes. Unfortunately, reperfusion can seldom be accomplished before some myocardium infarcts. As a result there has been an intensive search for interventions that will make the heart resistant to infarction so that reperfusion could salvage more myocardium. A number of interventions have been identified in animal models, foremost being ischemic preconditioning. It protects by activating signaling pathways that prevent lethal permeability transition pores from forming in the heart’s mitochondria at reperfusion. Such conditioning can be accomplished in a clinically relevant manner either by staccato reperfusion (ischemic postconditioning) or by pharmacological activation of the conditioning signaling pathways prior to reperfusion. Unfortunately, clinical trials of ischemic postconditioning and pharmacologic conditioning have been largely disappointing. We suggest that this may be caused by inappropriate use as models intended to mimic the clinical scenario of young healthy animals that receive none of the many drugs currently given to our patients. Patients may be resistant to some forms of conditioning because of comorbidities, for example, diabetes, or they may already be conditioned by adjunct medications, for example, P2Y₁₂ inhibitors or opioids. Incremental technological improvements in patient care may render some approaches to cardioprotection redundant, and thus the clinical target may be continually changing, while our animal models have not kept pace. In remote conditioning, a limb is subjected to ischemia/reperfusion prior to or during coronary reperfusion. Its mechanism is not as well understood as that of ischemic preconditioning, but the results have been very encouraging. In the present article, we will review ischemic, remote, and pharmacologic conditioning and possible confounders that could interfere with their efficacy in clinical trials in 2 settings of myocardial ischemia: (1) primary angioplasty in acute myocardial infarction and (2) elective angioplasty.

Introduction

Flow restoration soon after an acute total coronary occlusion induces myocardial salvage with a subsequent decrease in morbidity and mortality of patients with an acute myocardial infarction (AMI). ¹ In the last 40 years, there has been a substantial increase in the understanding of the mechanism of acute coronary syndromes which has led to dramatic improvements in their treatment. Cardiologists were gradually transformed from silent spectators of an acute event in the early 70s to today’s active fighters against ensuing myocardial necrosis. Effective antiarrhythmic agents and rapid recanalization of the culprit artery with thrombolytic agents and mechanical interventions have changed the course and the outcome of acute coronary syndromes dramatically. Incremental improvements were added step-by-step to our armamentarium against ischemia and reperfusion injury, and, one by one, they have been incorporated into the guidelines. However, other than early reperfusion, no treatment has been adopted to specifically reduce infarct size despite hundreds of reports in the basic science literature that indicate that this is possible. Last year, Kloner reviewed the current status of efforts to translate a number of cardioprotective interventions into clinical use. ² Here we will concentrate on just one of those approaches, the conditioning mechanism. This review tries to determine why conditioning has still not found its way into current practice.

Conditioning Protects Against Infarction

The quest for an anti-infarct intervention began with the studies by Maroko et al in 1971. ³ A number of articles were published reporting infarct size reduction in animals by interventions including antioxidant and anti-inflammatory agents, but the effects were modest and their reproducibility was poor. Then, in 1986, a revolutionary endogenous mechanism of myocardial protection was described. In that seminal study four 5-minute occlusions in dogs with each followed by a 5-minute period of reperfusion were applied to the coronary artery, which was afterward occluded for 40 minutes (the index ischemia) to induce an infarct. ⁴ The addition of the short cycles of ischemia–reperfusion reduced the volume of infarcted myocardium by 75% of that expected from a 40-minute occlusion and was termed ischemic preconditioning. The protection was independent of the presence of collateral vessels and its mechanism was a total mystery. This promising intervention has since been successfully reproduced in every species studied. Early investigations described the natural history of preconditioning, which depends on the number and duration of each brief ischemic episode as well as on the time interval between the short bouts of ischemia and the prolonged ischemic episode. ^{5 –7} A single 5-minute occlusion/reperfusion cycle is sufficient to protect most species and the protected state lasts for about 1 hour before it wanes. A somewhat less potent protection reappears without any additional intervention as a delayed form of protection 24 hours later. ^8,9 This second window of protection is thought to be the result of induction of protective proteins and remains for the next 3 to 4 days. The phenomenon of ischemic preconditioning proved that the heart could be made resistant to infarction, and investigators were convinced that it would soon lead to a new protective therapy for ischemic myocardium in man. ¹⁰

This enthusiasm was immediately transferred from experimental studies in animals to clinical studies, and a great number of observational and interventional reports indicated that preconditioning was also effective in the human heart. ^{11 –15} The major drawback with preconditioning is that it has to be instituted prior to the onset of ischemia, which is impractical in the setting of AMI when patients present only after ischemia has begun.

In 2003, another revolutionary phenomenon, ischemic postconditioning, was discovered, which was almost as protective as preconditioning against infarction. ¹⁶ Postconditioning consists of very short reperfusion–ischemia cycles (each lasting no more than a minute) applied immediately after flow is restored to a totally occluded artery. Postconditioning could actually be tested in humans, since it is an intervention that follows the restoration of flow and could thus be performed in patients with AMI treated with primary coronary angioplasty. ¹⁷ Earlier studies had shown that “gentle” or gradual reperfusion was preferable to abrupt restoration of coronary flow, ^18,19 but controlling reflow and its transmural distribution in a reproducible manner with only a catheter was a major obstacle. The staccato off–on protocol of postconditioning solves that problem, as all layers of the heart see the same reduction in average flow.

Many studies have investigated the mechanism of pre- and postconditioning. Today the evidence indicates that both phenomena use the same basic mechanism: substances released during ischemia including adenosine, opioids, and bradykinin trigger cardioprotective signaling through their receptors. A complex signaling pathway ultimately prevents formation of lethal permeability transition pores in the heart’s mitochondria (mPTP) in the first minutes of reperfusion. ^{20 –22} Opening of mPTP leads to mitochondrial destruction which results in cell death. Figure 1 presents a summary of what we think the core of this pathway looks like, but additional components seem to be reported daily. We refer to this type of protection as “conditioning" and today that signaling can easily be activated pharmacologically. For example, cyclosporin can interact with cyclophilin D and directly prevent mPTP opening. ²³

Figure 1.

Proposed signaling scheme for ischemic preconditioning. Protection is triggered by occupancy of surface receptors for adenosine, bradykinin (Brady), sphingosine, and opioids, which are released during ischemia. At the end of the preconditioning ischemia reoxygenation drives redox signaling, which completes the pathway possibly by activating protein kinase C (PKC). The final pathway involves the reperfusion injury survival kinases (RISK), extracellular signal-regulated kinase (ERK), Akt, and glycogen synthase kinase-3β (GSK3β), which inhibit permeability transition pore opening in the mitochondria (mPTP) when the heart is therapeutically reperfused after a prolonged period of ischemia. The survivor activating factor enhancement (SAFE) pathway involves signal transducer and activator of transcription (STAT) 3 and acts in parallel to the RISK pathway. Other parallel routes may exist as the system is highly redundant. Reprinted with permission from. ²² MMP indicates matrix metalloproteinase; HB-EGF, heparin-binding epidermal growth factor-like growth factor; Pro, pro-HB-EGF; Src, sarcoma tyrosine kinase; Tyr, tyrosine; PI3K, phosphatidylinositol 3-kinase; JAK, janus kinase; PI_4,5P₂, phosphatidylinositol bisphosphate; PI_3,4,5P₃, phosphatidylinositol trisphosphate; PDK1/2, 3′-phosphoinositide-dependent kinase-1/-2; MEK, mitogen-activated protein kinase; NO, nitric oxide; NOS, NO synthase; eNOS, endothelial NOS; GC, guanylyl cyclase; cGMP, cyclic guanosine monophosphate; PKG, protein kinase G; Sphingo 1-P, sphingosine-1-phosphate; PLC/PLD, phospholipase C/D; K_ATP, ATP-dependent potassium channel; TNF-α, tumor necrosis factor-α; p70S6K, p70S6 kinase.

Remote Ischemic Conditioning

In 1993, it was reported that preconditioning by occlusion/reperfusion of 1 coronary arterial branch was capable of conferring myocardial protection to the vascular territory of another coronary branch. ²⁴ This phenomenon has been termed remote conditioning and has since been extended to vascular beds outside the heart. Because remote conditioning has been primarily evaluated in humans, its mechanism is less well understood. To initiate remote conditioning, a patient’s limb (leg or arm) is made ischemic for several minutes with a blood pressure cuff. This causes the heart muscle to become resistant to infarction by some, as yet, poorly understood mechanism that may or may not involve the conditioning pathway described previously. In remote conditioning, the protection reportedly involves release of adenosine ²⁵ and bradykinin ²⁶ similar to classical conditioning through activation of neural pathways. Another proposed mechanism is that chemicals like nitrite, ²⁷ chemokines, ²⁸ or other factors are brought to the heart by the systemic circulation. ^29,30 Both direct and remote conditioning mainly target a reperfusion injury, which makes them clinically relevant. ¹ Accordingly, remote ischemic interventions applied during the index ischemia (termed perconditioning) are reported to favorably affect ischemic myocardium when the latter is eventually reperfused. ^31,32 The advantage of remote conditioning is its ease of application and apparent safety.

Can We Condition Patients?

Evidence-based medicine dictates that regardless of the effect in animals these treatments must show efficacy in patients before they can be translated to general clinical practice. To that end, a great number of clinical studies have investigated the effectiveness of conditioning in human hearts. ³³ Additional hard end points such as cardiovascular or total mortality and prevention of arrhythmic events or other surrogate end points have also been used in clinical studies. ^34,35 However, almost 30 years after the first description of ischemic preconditioning, 22 years of remote preconditioning, 12 years of postconditioning, and 10 years of remote (per) conditioning, none of these interventions has found its way into the clinical guidelines. The hard end point of conditioning is the reduction in the final infarct size after coronary ischemia–reperfusion, which should leave the patient with a stronger heart that would be less likely to fail.

The effectiveness of cardiac direct and remote conditioning in animals is beyond dispute. However, their demonstration in clinical practice has been surprisingly elusive. It is known that comorbidities such as diabetes ^36,37 or hypertension, ³⁸ which are common in coronary patients, can interfere with conditioning’s ability to protect the heart. Also, today’s patients receive many drugs including opioids, platelet inhibitors, and antiarrhythmic agents, 1 or more of which may itself be a conditioning agent. If so, this could make addition of another conditioning intervention redundant. Today’s patient is probably very different from one seen a decade ago, so we may be aiming at a moving target. Finally, the population with AMI is extremely heterogeneous with respect to the location of the thrombus, duration of ischemia, and collateralization of the heart. If an intervention only benefits a subset of that population, detecting that benefit in a clinical trial would require a careful risk stratification design. True protection may be obscured by noise in the data. In the following sections, we shall review the clinical experience with 3 forms of conditioning, namely, (1) ischemic conditioning, (2) remote ischemic conditioning, and (3) pharmacological conditioning-mimetics. This conditioning will be examined in the 2 most frequent therapeutic settings, namely, primary coronary angioplasty in evolving myocardial infarction and elective angioplasty. Although many interventions have been touted to be cardioprotective, for example, cooling and sodium/hydrogen exchange inhibitors, this review will focus only on the 3 forms of conditioning described previously.

Results of the 3 Forms of Conditioning in Humans Undergoing AMI and Primary Percutaneous Coronary Intervention

Ischemic Conditioning

An initial, very promising study by Staat et al ¹⁷ showed that ischemic postconditioning following primary angioplasty in patients presenting with evolving myocardial infarction reduced the rise in total serum creatine kinase (CK) by 36% compared to that in non-postconditioned patients. The heart was postconditioned with 4 cycles of 1 minute of reperfusion followed by 1 minute of balloon inflation immediately after stenting the culprit lesion. The investigators also devised a risk stratification analysis by which they showed that patients with a large amount of ischemic muscle greatly benefited from postconditioning, whereas the treatment offered little benefit to those with small ischemic areas since they had negligible infarction with reperfusion alone. This study raised great enthusiasm. Although additional manipulations in a successfully opened coronary artery pose a dilemma for interventional cardiologists who may be rather nervous or even reluctant to perform additional maneuvers, this approach appeared to result in a clear benefit for the reperfused myocardium. However, apart from the above-mentioned skepticism regarding its application in the catheterization laboratory, subsequent studies with very similar protocols were unable to reproduce the salutary findings of Staat et al. ¹⁷

Sörensson et al ³⁹ studied 76 patients eligible for primary percutaneous coronary intervention (PCI) who underwent the same intervention as mentioned earlier with 4 cycles of 1-minute reperfusion/1-minute ischemia and showed no difference in infarct size estimated by troponin T (TnT) and CK-MB release. In another study, 79 patients with evolving myocardial infarction and primary PCI derived no benefit from local postconditioning including infarct size estimated by troponin I (TnI) and magnetic resonance imaging (MRI), left ventricular ejection fraction, and cardiac remodeling. ⁴⁰ Instead, postconditioning resulted in a rather adverse outcome in patients with greater TnI release and smaller myocardial salvage—expressed in percentage as well as an index—and, contrary to the control group, no improvement in ejection fraction.

In a more recent trial with 700 patients, the same postconditioning protocol with 4 cycles of 1-minute deflation and inflation of the angioplasty balloon after primary PCI showed no difference in ST-segment resolution, frequency of myocardial blush, and major adverse cardiac events compared to the control group. ⁴¹ Finally, Limalanathan et al ⁴² showed no difference in infarct size measured with cardiac magnetic resonance (CMR), ST-segment resolution, peak TnT values, left ventricular ejection fraction, or rehospitalization for acute coronary syndrome or heart failure in 272 randomized patients with AMI treated with either PCI only or additional ischemic postconditioning. Therefore, of the 5 described clinical studies of local ischemic postconditioning after primary PCI, only the first was positive ¹⁷ while the other 4 were negative, ^{39 –42} despite application of the same protocol consisting of a series of 1-minute cycles of local reperfusion/ischemia.

Favaretto et al ⁴³ recently performed a meta-analysis of 14 clinical trials of ischemic postconditioning in patients undergoing primary PCI in which infarct size was estimated by either enzyme release or imaging by CMR. The overall analysis revealed a significant reduction in infarct size. However, when the authors limited the analysis to the 6 recent studies using CMR to directly measure infarct size, postconditioning’s benefit was lost. Among the 14 studies reviewed, none reported salvage on the order of that originally seen by Staat et al. ¹⁷

Thoughts

The optimal ischemic postconditioning protocol is more difficult to identify than that for ischemic preconditioning. In ischemic preconditioning once the threshold for triggering a preconditioned state is reached, added cycles cause no further modulation of the protection. That is not the case with postconditioning. Postconditioning is thought to limit reflow just enough to maintain acidosis in the heart, ⁴⁴ which inhibits mPTP formation. ⁴⁵ At the same time, the flow has to be high enough and last long enough to provide sufficient oxygen for the redox signaling required to put the heart into a conditioned state. Finally, the entire process must last long enough to fully condition the heart. Transitioning to full reperfusion too early will fail to protect. Once conditioned, the pH can be normalized without fear of mPTP formation. ⁴⁴ If a reperfusion cycle is too long, the pH may normalize before the heart is conditioned. In 1 rabbit study, even a 1-minute reperfusion period was too long to protect, whereas 30-second cycles of reperfusion–ischemia were protective. ⁴⁶ By this reasoning, the overall aim would be to reduce average reflow to 50% of normal for several minutes. Thus many rapid on–off cycles would be more likely to protect than just a few longer cycles. It is possible that all of the above-mentioned clinical investigations may have been using a relatively ineffective postconditioning protocol. Unlike what can be done in investigational animal models, it is not feasible to test either different numbers of cycles or different cycle durations in patients being treated with PCI for AMI to arrive at the most protective protocol.

The study by Staat et al ¹⁷ suggested that the effect of treatment increased with the size of the ischemic zone. This implies that inclusion of patients with a small ischemic zone will dilute the data with non-responders who would have negligible infarction even with standard treatment alone. None of the subsequent studies used such a risk stratification design. However, Hahn et al ⁴¹ did analyze a subgroup of patients with left anterior descending lesions, which should have produced larger ischemic zones, but postconditioning was still not beneficial in that subgroup.

There is another possible explanation for the variable efficacy of clinical ischemic postconditioning. Staat et al ¹⁷ published their study in 2005. In these early years, the use of P2Y₁₂ blocking agents as an adjunct to primary angioplasty was just becoming popular. Approximately half of the patients in their study received an oral clopidogrel loading dose prior to reperfusion therapy and half of those received only 300 mg, which has since been shown to be an insufficient dose. ⁴⁷ Recent evidence indicates that P2Y₁₂ inhibitors like clopidogrel are powerful cardioprotectants independent of their platelet antiaggregation effects. ^{48 –50} At optimal dosing of the platelet antagonist, combination of a P2Y₁₂ inhibitor with pre- ⁴⁸ or postconditioning ⁴⁹ offers no more protection in animal studies than that seen with the P2Y₁₂ antagonist alone, possibly suggesting similar modes of action. Thus, while only about 25% of Staat patients were protected by platelet inhibitors, ¹⁷ virtually all of the patients in subsequent studies received a proper dose of platelet inhibitor and thus may have already been conditioned.

A retrospective analysis of the database used by Staat et al ¹⁷ found that clopidogrel contributed equally with postconditioning to the protection. ⁵¹ Although combining clopidogrel and postconditioning showed the greatest protection suggesting additive effects, it can be argued that neither the dose of clopidogrel nor the postconditioning protocol was optimal. In a recent publication, Mochly-Rosen and Grimes ⁵² examined placebo groups in cardioprotection trials over the past 10 years and noted that 3-month mortality from AMI has progressively fallen from near 10% in 2000 to less than 4% today. Increased use of antiplatelet drugs over that time period could have played a large role in this improvement. If that hypothesis were true, then patients pretreated with antiplatelet agents undergoing reperfusion with primary angioplasty will already be in a conditioned state and any attempt to protect them with an additional conditioning-based intervention may well be redundant and ineffective. Whether that argument applies to remote conditioning is unknown, as its mechanism remains poorly understood and remote conditioning in the presence of a P2Y₁₂ inhibitor has yet to be studied in an animal model.

Remote Ischemic Conditioning

Remote conditioning is an innocuous procedure that may protect myocardium from reperfusion injury during AMI. This very easy and promising intervention was applied to patients with evolving myocardial infarction who were treated with primary PCI. ⁵³ They underwent 4 inflations of a blood pressure cuff to 200 mm Hg. "Conditioned" patients had significantly better myocardial salvage assessed by SPECT with ⁹⁹ technetium sestamibi. Moreover, infarct size reduction was most pronounced in patients whose culprit lesion was in the left anterior descending coronary artery, presumably because they had a larger ischemic area. This observation echoed that of Staat et al. ¹⁷ Additionally, the same group of patients had a better long-term outcome signified by lower rates of major adverse cardiac and cerebral events and all-cause mortality after a median follow-up of 3.8 years. Another promising study with a slightly different protocol demonstrated that 3 lower limb blood pressure cuff inflations applied simultaneously with thrombectomy or balloon dilatation of the culprit coronary artery rather than during the transportation of the patient resulted in less CK-MB release, better ST-segment resolution, and less myocardial edema assessed by MRI. ⁵⁴ Another recent evaluation of remote conditioning induced by upper extremity ischemia/reperfusion immediately prior to revascularization decreased peak CK-MB levels. ⁵⁵ Although there was no significant difference in areas under the curve for 72-hour CK-MB release between control and intervention groups, a plot of area under the curve against size of the risk zone yielded a regression line for remotely conditioned patients which was significantly shifted downward compared to that in the control group. Hence for any given area at risk, infarction was less following remote conditioning. Interestingly in a third group, the addition of ischemic postconditioning to remote conditioning provided no additional protection. ⁵⁵

Thoughts

Remote conditioning seems to be a promising approach during primary angioplasty in evolving myocardial infarction and we look forward to the results of additional studies. It would be very important if this easily applicable intervention conferred benefit. We may conclude that different protocols with 3 or 4 cycles of ischemia–reperfusion of the upper or lower limb prior to or at the time of coronary intervention result in a reduction in reperfusion injury. However, it should be noted that only 142 (42.7%) of the 333 patients of the initial Bøtker cohort completed the first study to provide useful information regarding the primary end point of myocardial salvage and salvage index. ⁵³ Because such interventions appear to be effective and have no known deleterious effect, some interventionalists are already including remote conditioning in their protocol for patients treated with primary angioplasty with the hope that it “can’t hurt and may help." Whether this is a wise choice should become clear when the results of ongoing large-scale trials become available.

Whether the protection of remote conditioning can be added to that of P2Y₁₂ receptor inhibitors to our knowledge has yet to be tested. Xin et al ⁵⁶ reported that remote perconditioning of both hind limbs combined with ischemic postconditioning of the coronary artery in an open-chest rat model yielded additive protection, but that protection was not greater than that seen with ischemic preconditioning. Although the Kloner laboratory repeated the Xin study, they found that neither remote perconditioning nor ischemic postconditioning was protective in their in vivo rat model. ⁵⁷ Not surprisingly the combination also failed to protect.

Pharmacological Intervention

Here we will focus on drugs that invoke the protective mechanism of conditioning through activation of appropriate receptors or signaling pathways that finally prevent mPTP opening and thus cell death by mitochondrial disruption. Adenosine, one of the main ligands triggering the conditioning pathway, was used in the Acute Myocardial Infarction STudy of ADenosine (AMISTAD) I ⁵⁸ and II ⁵⁹ trials as an adjunct to thrombolytic treatment or primary angioplasty in patients with ST-segment elevation myocardial infarction (STEMI). The results from AMISTAD I were encouraging in only those patients who had anterior myocardial infarction or in those who were treated with higher doses of adenosine. The larger AMISTAD II trial failed to confirm results of the previous study. However, in a retrospective analysis of AMISTAD II data Kloner found a significant reduction of infarction in a subgroup of patients who were treated early after onset of symptoms. ⁶⁰ In another study, intracoronary adenosine given as adjunct therapy during primary coronary angioplasty ameliorated coronary flow, prevented the no-reflow phenomenon, and improved left ventricular ejection fraction. ⁶¹ Despite these promising findings, however, entirely opposite results have been reported in 2 more recent studies that showed no effect of adenosine on infarct size or myocardial perfusion. ^62,63

Cyclosporin is extensively used in organ transplantation, in immune system disorders, as well as in other clinical conditions. A useful side effect of this drug is its effective inhibition of the opening of mPTP, thus preventing mitochondrial disruption and cell death. ²³ The experience acquired with this drug in experimental models was extrapolated to patients with evolving myocardial infarction in whom intravenous injection of the agent before primary angioplasty resulted in a significant decrease in average serum CK (but not TnI) values. ⁶⁴ It is of interest that the effect of cyclosporin on CK release was virtually identical to that observed with ischemic postconditioning during primary PCI. ¹⁷ When patients were stratified by the size of their ischemic zone, both TnI and CK release showed a highly significant reduction in the cyclosporin group. This study reveals that the risk stratification design greatly increases the statistical power of the analysis. In a subgroup of cyclosporin-treated patients who underwent MRI scanning, myocardial infarct size was smaller. Furthermore, cyclosporin also provided a long-term benefit with better left ventricular remodeling at 6 months after the acute event. ⁶⁴ A single infusion of cyclosporin had none of the unwanted effects of long-term administration observed in transplant patients. In light of this very promising small trial, the CIRCUS trial, a multicenter clinical study, was designed to answer the question whether cyclosporin can benefit patients with AMI when used as adjunctive therapy in primary coronary angioplasty.

Erythropoietin ³⁵ and atrial natriuretic peptide (ANP) ⁶⁵ activate a family of kinases that inhibit mPTP opening and can be thus effectively used as pharmacological agents that indirectly prevent mitochondrial destruction and cell death. Erythropoietin was given to patients with STEMI in 2 clinical studies, the erythropoietin (EPO) Revival ⁶⁶ and the Randomized EValuation of the Effects of Anacetrapib Through Lipid-modification (REVEAL) trial. ⁶⁷ Neither showed any reduction in infarct size. Kitakaze et al ⁶⁸ administered ANP just prior to reperfusion with primary PCI to patients presenting with AMI and found that the ANP group had significantly smaller infarcts and better left ventricular ejection fractions, although the improvement in ejection fraction was quite modest. These investigators did not incorporate a risk stratification design, so patients with small ischemic zones may have diluted the data. The same trial had a nicorandil arm, but that agent was not protective.

Statins whether administered before ischemia or just before reperfusion to experimental animals appear to be conditioning mimetics. ⁶⁹ Statins increase nitric oxide synthase (NOS) activity and nitric oxide production ⁷⁰ and also activate the phosphatidylinositol 3-kinase (PI3K)-Akt-endothelial NOS signal transduction pathway. ⁷¹ These enzymes and kinases are critical parts of conditioning’s signal transduction pathway. Pretreatment with high-dose rosuvastatin before primary PCI for AMI demonstrated no improvement in myocardial perfusion or decrease in infarct volume compared to exposure to only low-dose statin. ⁷² Although high-dose atorvastatin did improve coronary flow immediately following primary PCI, ⁷³ there was no effect on clinical events in the month following PCI. It should be noted that all patients in these 2 studies were also pretreated with 600 mg of clopidogrel, and, as indicated previously, this P2Y₁₂ platelet antagonist could have obscured any statin-related cardioprotective effect.

Thoughts

Low-dose adenosine started at reperfusion was reported to cause dramatic reduction in infarct size in a canine model of AMI ⁷⁴ and was the basis for the AMISTAD trials. However, other investigators could not reproduce the results in a similar canine model. ⁷⁵ The effectiveness of this treatment remains unknown.

Although cyclosporin has been proven to be an effective conditioning agent in experimental animals ⁷⁵ and in a small proof-of-concept study in man, ⁷⁷ its effect may be obscured in current and future clinical trials in which all patients receive standard treatment with platelet P2Y₁₂ antagonists. The latter will condition the myocardium, so any additional protection by cyclosporin which we believe involves the same mechanism as the P2Y₁₂ antagonist will likely be minimal at best. Only a portion of the patients in the study of Piot et al ⁷⁷ received a loading dose of clopidogrel. A retrospective analysis of which patients received the platelet inhibitor and its independent effect on infarction would give some insight into whether these platelet inhibitors combined with cyclosporin can have additive effects. Hopefully, such a retrospective analysis will be forthcoming.

Erythropoietin upregulates kinases such as PI3K, Akt, and ERK which prevent mPTP opening. ³⁵ Despite promising results from experimental studies, the results in humans are disappointing, possibly because of the timing of administration. Kobayashi et al ³⁵ found that erythropoietin, effective as a pretreatment, could not limit infarction in rabbit hearts when treatment was begun just prior to reperfusion. In the clinical studies, erythropoietin was given in 3 daily doses with the first one starting immediately after successful PCI in the EPO Revival study ⁶⁶ and within 4 hours after successful PCI in the REVEAL trial. ⁶⁷ This timing of initial administration may be too late to prevent the damage from reperfusion injury which starts immediately upon reperfusion. It would be interesting to see whether erythropoietin started earlier, that is, before the restoration of flow, would exert beneficial effects. Also to be considered is the potential for the drug to promote thrombosis. This could explain some findings of the REVEAL trial, which showed more in-stent thromboses in treated patients and an increased infarct size in older ones. ⁶⁷

In the Japan-Working Groups of Acute Myocardial Infarction for the Reduction of Necrotic Damage (J-WIND) trial ⁶⁸ in which ANP was protective, intravenous nicorandil had no effect on infarction, possibly because of the rather small dose that was chosen. Interestingly, oral nicorandil during the chronic recovery phase was effective in preserving ventricular function.

Finally, several agents with mechanisms of action either related to or unrelated to the conditioning effect, like δ protein kinase C inhibitors, exenatide, fibrin-derived peptide FX06, iron chelators, and others, have been used in clinical studies with divergent results. There is no convincing explanation for these discrepancies, and, therefore, there can be only speculation. A more detailed description of these studies is provided in an excellent review article by Kloner. ²

Elective Angioplasty

Ischemic Conditioning

Patients with stable coronary artery disease or with recent acute coronary syndromes who undergo coronary angioplasty on a nonurgent basis represent an heterogeneous group showing a great variation in coronary anatomy from 1-vessel disease with a single smooth lesion to 3-vessel disease with multiple, complicated lesions. This variation means that the area at risk and the ischemic burden vary widely from small to extensive. Using an algorithm of 6 cycles of 10 seconds of ischemia/reperfusion immediately prior to stent deployment in patients who underwent elective angioplasty, a group that included some of the authors found a decrease in oxidative stress markers, better restoration of coronary microcirculation, and improved wall motion score index. ^78,79

Thoughts

There is a difference in the protocol that yielded these positive results compared to that in other studies applying local ischemia. Our protocol consisted of 6 cycles of 10 seconds of ischemia/reperfusion instead of the frequently applied 4 cycles of 1 min of reperfusion/ischemia. Furthermore, there was a substantially different conditioning approach, since we applied the local ischemic insults immediately after the passing of the angioplasty wire and before stent deployment thus avoiding additional manipulations inside the stent after successful implantation.

Remote Ischemic Conditioning

Patients undergoing elective angioplasty may derive additional benefits from the application of a remote ischemic intervention. The Cardiac Remote Ischemic Preconditioning in Coronary Stenting (CRISP) Stent trial showed that 3 cycles of 5 minutes of upper extremity blood pressure cuff inflation/deflation prior to elective PCI led to an improvement in electrocardiogram changes, lessening of chest discomfort, and reduction in TnI release, and twice as many conditioned patients had no detectable enzyme release following PCI. ⁸⁰ Benefits persisted for at least 6 years. ⁸¹ Of note, all of these benefits were observed even when remote ischemia was applied approximately 1 hour before elective angioplasty, at a time when the resulting subsequent protection typically fades and is eventually lost as shown in most experimental studies. This is confusing since our group previously found that this intervention, applied in a more potent way with cuff inflation on both arms and closer to the time of elective angioplasty, was not only ineffective but also possibly even detrimental, especially in a subgroup of patients not using statins. ⁸² Possibly, the brief occlusions may trigger a detrimental inflammatory response which is prevented by statins. Another study of 232 patients undergoing elective angioplasty also showed no benefit from remote conditioning evaluated by TnI and CK-MB release in the short term as well as clinical events after 1 year. ⁸³ More recently, another large trial with 360 patients subjected to remote ischemia in the upper or lower limb by 3 cycles of 5-minute ischemia–reperfusion again showed no benefit with no difference in the release of TnT, CK, or high-sensitivity C-reactive protein or acute kidney injury. ⁸⁴

Thoughts

We observed divergent results in these clinical studies with one being positive ⁸⁰ and the other 3 negative. ^{82 –84} In the first study, average TnI release was slightly but significantly lower in remotely conditioned patients 24 hours after elective angioplasty, and there were more treated patients who had a median value of TnI below the detectable limit of 0.04 ng/mL suggesting a protective effect. Although remote conditioning was applied long enough before elective angioplasty to threaten fading of the first window of protection, this intervention resulted in fewer major adverse events in the long term. In the second study, ⁸² a greater amount of limb ischemia closer to the procedure yielded negative results suggesting excessive limb ischemia may cause loss of effect. In the last 2 large studies that were also negative, remote conditioning was applied after stent deployment. ^83,84 But reperfusion injury probably would have occurred seconds after restoration of flow—long before protection would have been established. Regardless of possible explanations, the fact is that 3 of the 4 described clinical studies were negative in patients who received remote conditioning with elective angioplasty.

Pharmacological Intervention

Adenosine and nicorandil are 2 drugs that have been extensively used in experimental and clinical studies. These drugs were given separately or in combination for 213 patients who underwent elective angioplasty. ⁸⁵ The drugs were administered through the guiding catheter into the target coronary artery before advancement of the guidewire. The study was negative showing no difference in CK-MB or TnI release between controls and treated patients, and the authors concluded that adenosine, nicorandil, and the combination of these drugs do not reduce the incidence of necrosis or myocardial infarction in patients with low-risk angina pectoris. Rosuvastatin given orally before elective PCI showed significant improvement in microcirculatory perfusion as assessed by contrast echocardiography. ⁸⁶

Thoughts

Elective angioplasty is a low-risk intervention and in most patients the culprit coronary artery is at least partially open before the procedure. This population has, therefore, a low ischemic burden and the possibility to demonstrate a clear benefit as a result of an anti-infarct intervention is limited. Adenosine is a ligand that triggers conditioning, ⁸⁷ but it is difficult to use in a clinical setting. We found that adenosine had to be given at a receptor-saturating dose to isolated hearts for a full 5 minutes to elicit a preconditioned state. ⁸⁸ That was easy to achieve in an isolated rabbit heart but impossible in vivo since adenosine will arrest the heart by causing atrioventricular block and trigger severe hypotension as it recirculates. It is doubtful whether any of these patients received a conditioning dose of adenosine. Nicorandil is an agent that activates mitochondrial adenosine triphosphate-dependent potassium channels. ^89,90 There have been relatively few animal studies demonstrating a direct anti-infarct effect of this agent. Nicorandil showed no effect against infarction in the J-WIND study, while ANP did. ⁶⁸ These agents may be insufficient for conditioning the myocardium. Statins protect the heart by activating signaling kinases which then prevent mPTP opening. ⁹¹ The hydrophilic statins given acutely may provide additional benefit, ⁹² although, as explained previously, concomitant therapy with P2Y₁₂ platelet antagonists may obscure benefits.

General Discussion

Myocardial protection against postischemic reperfusion injury in animal models is effective and may contribute to the reduction in the final infarct size by approximately 40% beyond the level that is obtained by reperfusion per se. ¹ Compelling evidence of the effectiveness of conditioning is its unequivocal reduction in infarct size. Unfortunately, the consistency and reproducibility of the effectiveness of conditioning measured with hard end points in experimental models has been difficult to repeat in clinical studies. The whole story becomes even more complicated if we consider the various conditioning protocols applied to patients. Unfortunately, we do not know what the optimal postconditioning or remote conditioning protocol is for humans. Hence much of the discrepancy in the results of these studies may be attributable to the difference of the conditioning algorithms which vary from a few very short-lived bouts of ischemia/reperfusion to multiple and more prolonged ischemic interventions. ⁹³

Based on the first clinical ischemic postconditioning protocol applied in patients undergoing primary angioplasty, most of the clinical studies have used 3 or 4 cycles of 1 minute of local reperfusion–ischemia induced by balloon deflation/inflation. ¹⁷ As mentioned previously, it is possible that this protocol is far from optimal and perhaps a greater number of ischemic bouts with shorter periods of occlusion and reperfusion would be more effective. Unfortunately, it is not easy to experiment with different protocols in a clinical setting.

There are also some additional questions related to the postconditioning protocol. To prevent reperfusion injury it is critical that the first reperfusion period does not exceed 1 minute. However, wire advancement or predilation with smaller balloons may, at least partially, restore coronary flow prior to stent deployment and initiate the reperfusion injury before the formal postconditioning protocol even begins. On the other hand, repeated short-lived periods of ischemia–reperfusion with small balloons before the deployment of a stent may act as a postconditioning stimulus at the initial reperfusion period without the need for more and potentially hazardous manipulations inside the stent. ^78,86,94 This latter protocol might be an attractive alternative.

The above-mentioned difficulties and reservations are eliminated by the use of remote conditioning. There is no danger to the coronary artery or stent, and the stenting procedure is not delayed or prolonged. The small studies that have so far been completed seem to demonstrate efficacy of remote conditioning in the treatment of AMI. We await larger clinical trials to confirm the effectiveness and applicability of remote conditioning in AMI. The results of remote conditioning are more divergent and confusing in the setting of elective angioplasty or cardiac surgery.

Patients in the CRISP Stent trial who underwent elective angioplasty with remote conditioning not only had lower troponin levels shortly after angioplasty ⁸⁰ but also had fewer adverse cardiac events over the next 6 years. ⁸¹ Why remote conditioning during elective angioplasty would have such a long-term beneficial effect is a mystery. Some studies have reported that additional local manipulations after successful primary PCI result in an improvement in renal function. ⁹⁵ Although this is a very welcome finding, it is difficult to comprehend how manipulations of coronary flow might protect the kidneys that are not ischemic.

In parallel with primary or elective angioplasty, another very interesting issue for myocardial protection is the role of conditioning in cardiovascular surgery. Preconditioning the heart with intermittent episodes of ischemia–reperfusion was performed in patients who underwent cardiovascular surgery with the assumption that these interventions would protect the heart from the iatrogenic ischemia experienced during the procedure. The surgical setting has the advantage that it allows preconditioning of the heart prior to iatrogenic ischemia. However, the ischemic burden in cardiac surgery is small, as the procedures are designed to avoid infarction by keeping the periods of ischemia as short as possible and by cooling the heart. Ischemic, remote, and pharmacological conditioning have also been applied in patients undergoing coronary artery bypass grafting, valve replacement, or surgery for congenital heart disease but again with divergent results regarding various end points. ⁹⁶ Remote conditioning achieved by limb ischemia with blood pressure cuff inflation is promising, and the results of 2 ongoing multicenter clinical studies, the Effect of Remote Ischemic preConditioning on clinical outcomes in patients undergoing Coronary Artery bypass graft surgery (ERICCA) and Remote Ischaemic Preconditioning for Heart Surgery (RIPHeart) from the UK and Germany, will be revealed in 2015 and 2016, respectively. ^97,98

A thorough approach to patients with coronary artery disease should also take into account that these patients have comorbidities and simultaneously may be already treated with drugs that either “condition” the heart or alternatively prevent any benefit that would be otherwise obtained. This criticism has been extensively reviewed ²⁹ and it remains a great obstacle to the wide application of conditioning in clinical practice. This is especially true for the P2Y₁₂ platelet antagonists which are now standard of care for the treatment of AMI and coronary stenting. As described repeatedly previously, these agents have intrinsic cardioprotective properties independent of their effects on platelet aggregation. Because these drugs are now used in virtually all patients presenting to the hospital with acute coronary syndrome, any additional intervention aiming for reduction in infarction must salvage myocardium by a mechanism different from that used by these antiplatelet agents or else be almost certainly relegated to a failed trial. This observation should also be used to revise how experimental studies are done. The effectiveness of any new cardioprotective intervention should be evaluated in an animal model simultaneously treated with a P2Y₁₂ antagonist to accurately simulate the current patient treated for AMI. Today’s patients presenting with AMI are very different from those 3 decades earlier, but our animal models have not kept pace. Only those interventions that can protect in a clinically relevant model should be considered for clinical testing. We have recently found that ischemic preconditioning was unable to further reduce infarct size in a rat model treated with a P2Y₁₂ inhibitor. ⁴⁸ However, employing multiple cardioprotective interventions, each using a unique mechanism (mild hypothermia plus a sodium/hydrogen exchange blocker plus a P2Y₁₂ inhibitor), greatly increases salvage beyond that from the P2Y₁₂ inhibitor alone. ⁴⁸ While cooling and cariporide are ineffective if given only at reperfusion, we have found that a mitochondrially directed DNA repair enzyme can provide additional protection in the presence of a P2Y₁₂ inhibitor when administered at reperfusion. ⁹⁹

Finally, there are less tangible variables at play. Of note, there is a variation in the results among different patient cohorts and different groups of investigators. ^100,101 There is no answer whether genetic, environmental, geographic, or other unknown factors could have an influence on the final outcome. For example, patients from different countries appear to have different outcomes after an acute coronary syndrome, despite the use of very similar therapeutic interventions according to the National Registries. ¹⁰²

Moreover, reduced door-to-balloon times, novel antiplatelet agents, thrombus aspiration, and technical improvements such as very soft wires, flexible catheters, and balloons and stents with advanced technology have all decreased the amount of myocardium in jeopardy thus leaving less room for additional benefit from an anti-infarct intervention. Although some may argue that we have reached a “therapeutic ceiling” beyond which we cannot go, it must be pointed out that the incidence of postinfarction heart failure, while reduced, is far from eliminated. We believe that further protection can be achieved with novel interventions and that they will significantly improve outcomes.