OUP user menu

★ Editor's choice ★

Translating cardioprotection for patient benefit: position paper from the Working Group of Cellular Biology of the Heart of the European Society of Cardiology

Derek J. Hausenloy , Hans Erik Bøtker , Gianluigi Condorelli , Peter Ferdinandy , David Garcia-Dorado , Gerd Heusch , Sandrine Lecour , Linda W. van Laake , Rosalinda Madonna , Marisol Ruiz-Meana , Rainer Schulz , Joost P.G. Sluijter , Derek M. Yellon , Michel Ovize
DOI: http://dx.doi.org/10.1093/cvr/cvt004 7-27 First published online: 19 January 2013

Abstract

Coronary heart disease (CHD) is the leading cause of death and disability worldwide. Despite current therapy, the morbidity and mortality for patients with CHD remains significant. The most important manifestations of CHD arise from acute myocardial ischaemia–reperfusion injury (IRI) in terms of cardiomyocyte death and its long-term consequences. As such, new therapeutic interventions are required to protect the heart against the detrimental effects of acute IRI and improve clinical outcomes. Although a large number of cardioprotective therapies discovered in pre-clinical studies have been investigated in CHD patients, few have been translated into the clinical setting, and a significant number of these have failed to show any benefit in terms of reduced myocardial infarction and improved clinical outcomes. Because of this, there is currently no effective therapy for protecting the heart against the detrimental effects of acute IRI in patients with CHD. One major factor for this lack of success in translating cardioprotective therapies into the clinical setting can be attributed to problems with the clinical study design. Many of these clinical studies have not taken into consideration the important data provided from previously published pre-clinical and clinical studies. The overall aim of this ESC Working Group Cellular Biology of the Heart Position Paper is to provide recommendations for optimizing the design of clinical cardioprotection studies, which should hopefully result in new and effective therapeutic interventions for the future benefit of CHD patients.

  • Cardioprotection: Ischaemia
  • Reperfusion
  • Acute myocardial infarction
  • Cardiac surgery

1. Introduction

Coronary heart disease (CHD) is the leading cause of death and disability worldwide. According to the World Health Organisation (WHO), each year CHD accounts for the deaths of 3.8 million men and 3.4 million women. The global burden of CHD is projected to increase from 47 million DALYs (disability-adjusted life years or ‘healthy years of life lost’) in 1990 to ∼82 million DALYs in 2020.

Many of the major complications of CHD, such as myocardial infarction (MI) and heart failure, arise from the detrimental effects of acute ischaemia–reperfusion injury (IRI) on the myocardium. As such novel therapeutic interventions are required to protect the myocardium against acute IRI in order to preserve cardiac contractile function, reduce the onset of heart failure, and improve clinical outcomes in patients with CHD. In this article, the term ‘cardioprotection’ is used to refer specifically to the protection of the myocardium against the detrimental effects of acute IRI. Over the years, the research field of cardioprotection has consistently failed to produce any effective therapeutic strategy for protecting the myocardium against acute IRI in the clinical setting. The failure has not been due to a shortage of potential cardioprotective strategies discovered in the pre-clinical experimental setting, but has been due to the inability to successfully translate many of these promising therapies into interventions that actually improve patient outcomes, a topic of much discussion in the recent literature.14 In this regard, the overall aim of this ESC Working Group Cellular Biology of the Heart Position Paper will be to critically assess the translational process which takes place in the transition from the bench to the bedside, and to suggest recommendations for the future design of clinical cardioprotection studies, which take into consideration the important findings from both pre-clinical and clinical data in the research area of cardioprotection. Specifically, in this position paper we focus on the ways of optimizing the design of the clinical studies for testing novel cardioprotective interventions in two major clinical settings of acute myocardial IRI: patients presenting with an acute ST-segment elevation myocardial infarction (STEMI), treated by either thrombolytic therapy or primary percutaneous coronary intervention (PPCI) and patients undergoing coronary revascularization by coronary artery bypass graft (CABG) surgery. In particular, we critically analyse the contributions of patient selection, co-morbidities, concomitant medication, the timing of the therapeutic intervention, and the endpoints used for assessing cardioprotection, to the outcome of the clinical study. This should hopefully improve the chances of successfully translating future cardioprotective strategies for the benefit of CHD patients.

1.1 Major signalling pathways underlying cardioprotection

Elucidation of the major signal transduction pathways underlying endogenous cardioprotective strategies such as ischaemic preconditioning (IPC),5 ischaemic postconditioning (IPost),6,7 and remote ischaemic conditioning (RIC),8,9 in which the heart is ‘conditioned’ either directly or indirectly by brief episodes of ischaemia and reperfusion, has identified two endogenous cardioprotective pathways, the Reperfusion Injury Salvage Kinase (RISK)10,11 and the Survival Activating Factor Enhancement (SAFE) pathways.1215 These are recruited at the time of myocardial reperfusion and mediate cardioprotection. The RISK pathway includes the pro-survival kinase cascades MEK1/2-Erk1/2 and PI3K-Akt, whereas the SAFE pathway is made up by the TNF-α receptor and STAT3.1419 These two pathways relay the cardioprotective signal underlying the ‘conditioning’ strategies mentioned above, from cell membrane receptors to the mitochondria where protective mechanisms subsequently occur such as mitochondrial permeability transition pore (MPTP) inhibition,2023 mitochondrial connexin-43 channel activation, and mitochondrial ATP-dependent potassium channel opening.24 The elucidation of these cardioprotective signalling pathways in pre-clinical studies has been pivotal in identifying therapeutic targets for cardioprotection in the clinical setting.

2. Opportunities for cardioprotection

In this section, the major clinical settings in the which the CHD patient is subjected to the detrimental effects of acute myocardial IRI and so potentially benefit from novel cardioprotective strategies, are reviewed.

2.1 Acute STEMI patients undergoing myocardial reperfusion

The clinical scenario, which most typically represents a classical example of acute myocardial IRI, is the patient presenting with an acute STEMI, treated by either thrombolytic therapy or PPCI.

In-hospital mortality of unselected STEMI patients in the national registries of the ESC countries varies between 6 and 14%.25 There has been a reduction in both acute and long-term mortality following STEMI, due to greater use of reperfusion therapy, PPCI, anti-thrombotic therapy, and secondary prevention treatments, although the number of patients developing heart failure has increased.26 However, despite this, mortality post-STEMI remains substantial with ∼12% of patients being dead within 6 months,27 with an increased mortality rate in higher-risk patients.28 In developed countries, ∼1–2% of the adult population suffer from heart failure, with the prevalence increasing to ≥10% among persons 70 years of age or older.29 Therefore, these data underscore the importance of discovering novel therapeutic targets for protecting the heart against acute IRI so as to limit the MI size, prevent the onset of heart failure, and reduce cardiac mortality.

For patients presenting with an acute STEMI, early myocardial reperfusion using either thrombolytic therapy or PPCI remains the most effective treatment strategy for limiting the MI size, preserving cardiac function, and reducing the onset of heart failure. Where facilities are available, myocardial reperfusion by PPCI, as opposed to thrombolysis, is the preferred therapeutic strategy. Vast improvements have already been made in reducing the duration of acute myocardial ischaemia (the chest pain onset to PPCI time) with improved patient awareness (to reduce the time to first medical contact with the emergency medical services), minimizing the transit time to the PPCI centre, and reducing the door to PCI time at the PPCI centre.30,31 Importantly, translation of such progress into improvement in patient outcomes has been documented.32

Improvements in both anti-platelet and anti-thrombotic therapy and advances in PCI technology to maintain the patency of the infarct-related coronary artery have further optimized the process of myocardial reperfusion. Although these therapeutic approaches clearly protect the coronary vasculature and reduce the risk of coronary re-thrombosis in PPCI patients, there is preliminary experimental evidence suggesting that both anti-platelet and anti-thrombotic therapy may actually confer direct protection on cardiomyocytes against acute IRI (see later section).

Paradoxically, the process of myocardial reperfusion can itself induce myocardial injury and cardiomyocyte death, a phenomenon which has been termed ‘myocardial reperfusion injury’.7,33 The reversible forms of myocardial reperfusion injury which include reperfusion arrhythmias and myocardial stunning are usually short-lived and easily managed.7,33 However, the irreversible forms of myocardial reperfusion injury, which include microvascular obstruction (MVO) and lethal myocardial reperfusion injury (‘reperfusion-induced necrosis’’),34 contribute to the final myocardial infarct size and diminish the benefits of myocardial reperfusion in terms of myocardial salvage.7,33 MVO describes the ‘inability to reperfuse a previously ischemic region’.35 The underlying cause of MVO is unclear although it has been attributed to capillary damage with impaired vasodilatation, external capillary compression by endothelial cell and cardiomyocyte swelling, microembolization of friable material released from the atherosclerotic plaque, platelet microthrombi, and neutrophil adhesion and/or plugging.3640 Lethal myocardial reperfusion injury refers to the reperfusion-induced death of cardiomyocytes which were viable or reversibly injured at the end of ischaemia.7,33,34 The mechanisms underlying this form of cardiomyocyte death are multiple and include oxidative stress, calcium overload, MPTP opening, cardiomyocyte hypercontracture, apoptosis, necrosis, necroptosis, and inflammation (reviewed in7,33,41).

2.2 Patients undergoing cardiopulmonary bypass surgery

Patients undergoing coronary revascularization by CABG surgery are subjected to global acute myocardial IRI. When the aorta is clamped prior to going onto cardiopulmonary bypass, the heart is made acutely ischaemic and when the heart is taken off cardiopulmonary bypass and the aorta is unclamped, the heart is subjected to acute myocardial reperfusion injury. This global acute myocardial IRI contributes to the peri-operative myocardial injury and infarction that occurs during CABG surgery. The incidence and magnitude of peri-operative myocardial injury and infarction can be measured using serum cardiac enzymes such as CK-MB,42 Troponin-T,43 and Troponin-I44 and have been linked to worse clinical outcomes post-surgery. Guidelines for defining MI related to CABG have been recently published in the ‘Third universal definition of myocardial infarction’.45 Myocardial infarction related to CABG has been termed as Type 5 MI and has been defined as an elevation of cardiac biomarker values >10 × 99th percentile URL in patients with normal baseline cardiac Troponin values (<99th percentile URL), along with either (i) new pathological Q-waves or new left bundle branch block (LBBB), or (ii) angiographic documented new graft or new native coronary artery occlusion, or (iii) imaging evidence of new loss of viable myocardium or new regional wall motion abnormalities.45

Other factors that can result in peri-operative myocardial injury during CABG surgery include coronary embolization, manual handling of the heart, and inflammation.46,47 As such, the discovery of novel cardioprotective strategies for minimizing this form of myocardial injury and infarction during CABG surgery would be expected to preserve cardiac function and improve clinical outcomes in this clinical setting, particularly in those high-risk patients who are most vulnerable to this form of myocardial injury and infarction.48 Any cardioprotective intervention shown to be effective in the setting of CABG surgery may also be expected to be beneficial in other surgical settings in which the heart is subjected to acute global myocardial IRI, such as in major vascular and intra-abdominal surgery. In these latter settings, in which the pathophysiology of acute IRI is often unclear (and which include low cardiac output, coronary spasm, regional hypoperfusion, and so forth), additional studies are required to determine the relative contributions of acute ischaemia and reperfusion to the damage which occurs during surgery, in order to optimize cardiac protection.

2.3 Other opportunities for cardioprotection

2.3.1 Cardiopulmonary resuscitation

In a cardiopulmonary arrest, the whole body including the heart is subjected to acute global ischaemic injury. Successful cardiopulmonary resuscitation (CPR) results in the restoration of spontaneous circulation (ROSC) following the cardiac arrest which then subjects the whole body and the heart to acute global reperfusion injury. Following ROSC, the acute global myocardial IRI results in myocardial necrosis and post-resuscitation myocardial dysfunction, factors which, together with brain, kidney, and liver damage, are associated with worse clinical outcomes post-arrest.

There is an opportunity to administer a therapeutic intervention after the onset of cardiopulmonary arrest to minimize the acute global ischaemic injury and protect the heart and other vital organs. In this regard, a number of pre-clinical studies using animal models of cardiac arrest have investigated the role of a variety of cardioprotective interventions administered prior to cardiac arrest including mechanical interventions (therapeutic hypothermia49) and pharmacological ones [β-adrenergic blockade,50 iNOS inhibition,51 KATP channel activation,52 sodium-hydrogen ion exchanger inhibitor,53 erythropoietin,54 and cyclosporin-A (CsA)55].

Importantly, a therapeutic intervention applied to protect the heart against acute IRI could also provide systemic organ-wide protection against acute IRI, benefiting the post-cardiac arrest function of other vital organs such as the brain, kidney, and liver. Clinical studies investigating novel cardioprotective strategies in the clinical setting of CPR are yet to be undertaken.

2.3.2 Cardiac transplantation

Acute myocardial IRI sustained during cardiac transplantation is a major cause of graft failure. In the setting of cardiac transplantation, the donor heart is subjected to cold myocardial ischaemic injury at the time of graft procurement, storage, and transportation, which exacerbates the inflammatory response and the chance of rejection, contributing to graft vasculopathy and failure.56 At the time of graft implantation, injury to the graft is exacerbated by the acute global myocardial reperfusion injury which occurs on reperfusion of the graft.

There is an opportunity to administer a therapeutic intervention at the time of graft procurement, storage, and transport to minimize the cold ischaemic injury and protect the donor heart. Similarly, there is an opportunity to administer a therapeutic intervention to the recipient to protect the donor heart against acute global myocardial reperfusion injury that occurs at the time of graft implantation. In this regard, a number of pre-clinical studies have been published investigating a variety of cardioprotective interventions applied to the donor heart including pharmacological agents (adenosine analogue, sodium–hydrogen exchange inhibition, KATP channel activation, sildenafil, PKC-δ inhibition, and isoflurane) and mechanical interventions (IPC, IPost, and RIC) (reviewed in56). So far, no clinical studies have investigated cardioprotection in the setting of cardiac transplantation.

3. Optimizing the design of clinical cardioprotection studies

The failure to translate novel cardioprotective strategies discovered in pre-clinical studies into the clinical setting for patient benefit can be attributed to a number of different factors, the majority of which fall into three main categories: (i) the failure to develop a study intervention for human use against validated targets; (ii) inadequate or insufficient pre-clinical testing of the therapeutic intervention before clinical testing; and (iii) the design of the clinical cardioprotection study.1,3,4,57,58

3.1 The study intervention

The first point to consider in planning a clinical trial on cardioprotection is the selection of the therapeutic intervention to be tested. Only treatments providing consistent and robust benefit in pre-clinical studies involving different models and laboratories should be considered. Although this may seem an obvious pre-requisite, the failure to take this factor into consideration has led to a large number of negative clinical trials (see Table 1 for summary). This issue was discussed in a recent NHLBI Workshop and resulted in the formation of the CAESAR: NIH Cardioprotection Consortium, a network of research laboratories which are using a variety of clinically relevant pre-clinical animal MI models to test the efficacy of novel therapeutic agents to ensure they confer consistent and robust cardioprotection before entering the clinical arena.3,59

View this table:
Table 1

Clinical studies which have failed to demonstrate any beneficial effect in STEMI patients with a therapeutic intervention administered at myocardial reperfusion

Clinical studyTherapeutic interventionn, numberOutcomeNotes
Anti-oxidant therapy
EMIP-FR 2000106IV bolus of trimetazidine given prior to thrombolysis followed by 48 h infusion19 725No difference in mortality at 35 daysAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Magnesium
MAGIC107IV bolus of magnesium given prior to reperfusion followed by 24 h infusion6213No difference in mortality at 30 daysAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Glucose insulin potassium (GIK) therapy
Mehta et al., CREATE-ECLA108IV GIK infusion for 24 h started after reperfusion in the majority of cases20 201No difference in mortality at 30 daysAnterior STEMI only: no

Only PPCI or thrombolysis: no

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Sodium–hydrogen ion exchange inhibitors
Zeymer et al., ESCAMI109Iv eniporide as a 10 min infusion prior to PPCI or after thrombolysis2118No difference in the MI size (72 h AUC alph-ahydroxybutyrate dehydrogenase)Anterior STEMI only: no

Only PPCI or thrombolysis: no

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Nicorandil
Kitakaze et al., J-WIND-KTP67Iv nicorandil bolus then 72 h infusion started after reperfusion545No difference in the MI size (72 h AUC total CK) or 6 month LVEFAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: no
Anti-inflammatory agents
Armstrong et al., APEX-MI110Iv pexelizumab bolus given prior to PPCI followed by infusion for 24 h5745No difference in all-cause death at 30 daysAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Atar et al., FIRE111Iv FX06 bolus given prior to PPCI and then repeated 10 min later232No difference in the MI size by CMR at 5 days or 4 monthsAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
PKC-δ inhibitor
Lincoff et al. 2011 PROTECTION-AMI, UnpublishedIv delcasertib infusion for 24 h started prior to PPCI1083Anterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: yes

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Erythropoietin (EPO)Large animal studies inconclusive

Potential off-target effects
Voors et al., HEBE-III112IV EPO epoetin-alpha 60 000 IU after (within 3 h) PPCI529No difference in the LVEF at 6 weeks. No difference in the MI size (AUC CK-MB or TnT)

More major adverse cardiac events occurred with EPO
Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: no
Ott et al., REVIVAL-3113IV EPO epoetin-beta 33 000 iU immediately after PPCI repeated 24 and 48 h later138No difference in LVEF at 6 months assessed by CMR. No difference in the MI size (5 days and 6 month CMR)Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: no
Ludman et al.114IV EPO epoetin-beta 50 000 iU prior to PPCI repeated 24 h later52No difference in the MI size at 3 days using CMR and or 24 h AUC Trop T. Doubling of incidence of MVO on CMRAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Rao et al. 2011, REVEAL NCT00378352IV EPO epoetin-beta 60 000 iU immediately after PPCI repeated 24 and 48 h later138No difference in the MI size on CMR within 6 days and at 3 monthsAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: no
Atorvastatin*
Hahn et al.115Oral atorvastatin 80 mg prior to PPCI and 10 mg daily thereafter173No difference in the MI size at 5–14 days using SPECTAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Post et al., REPARATOR116Oral atorvastatin 80 mg prior to PPCI and daily thereafter42No difference in LVESI at 30 daysAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Iron chelation
Chan et al.117Iv bolus of desferoxamine given prior to PPCI followed by 12 h infusion60No difference in the MI size (48 h AUC CK-MB and Trop I and CMR). No difference in myocardial salvageAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Ischaemic postconditioning
Tarantini et al.100Four-60 s angioplasty balloon inflations/deflations78Non-significant increase in the MI size

IPost protocol was delivered within the stent, increasing the risk of coronary microembolization
Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: yes

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Freixa et al.101Four-60 s angioplasty balloon inflations/deflations79Reduced myocardial salvage. No difference in the MI size at 1 week or 6 months by CMR.

IPost protocol delivered within the stent, increasing the risk of coronary microembolization
Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: yes

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
  • Anterior STEMI only, only anterior STEMI patients included; only PPCI or thrombolysis, only either PPCI or thrombolysis patients included; AAR measured, area at risk measured; collateral flow excluded, coronary collateralization to the AAR excluded; TIMI flow grade <1, TIMI flow grade <1 in the infarct-related artery prior to PCI; treatment prior to or at reperfusion, study intervention given prior to or at reperfusion,*although oral atorvastatin was given prior to reperfusion, therapeutic levels would not have been achieved by this time.

On the other hand, the translation to patients of pharmacological treatments for which there is strong pre-clinical evidence is often limited by the non-availability of drugs which can be used safely in humans, or the lack of interest in myocardial reperfusion injury by the companies who own these drugs. For example, pharmacological approaches which have solid pre-clinical evidence, but which lack drugs for human use are contractile blockers,60 calpain inhibitors,61 or particulate guanylate cyclase stimulators.62 Even treatments which are available for human use have often been developed for other actions, many of which are undesired when applied to reduce reperfusion injury—an example of this is CsA, which was developed as an immunosuppressant agent and has been used to prevent myocardial reperfusion injury based on its effect on the MPTP (see Table 2). Overcoming these limitations will require a change in the perception of the pharmaceutical industry regarding the economic potential of developing and testing treatments against myocardial reperfusion injury.

View this table:
Table 2

Clinical studies which have demonstrated beneficial effects in STEMI patients with a therapeutic intervention administered at myocardial reperfusion

Clinical studyTherapeutic interventionn, numberOutcomeNotes
Atrial natriuretic peptide
Kitakaze et al., J-WIND-ANP67IV carperitide 72 h infusion started after reperfusion56915% reduction in 72 h AUC total CK and 2.0% absolute increase in the LVEFAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: no
Ischaemic postconditioning
Staat et al.65Four-60 s angioplasty balloon inflations/deflations3036%↓ in 72 h AUC CK

34%↓ in peak CK

MBG↑ 1.7–2.4
Anterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: yes

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Thibault et al.85Four-60 s angioplasty balloon inflations/deflations3841%↓ 72 h AUC CK-MB

39%↓ MI size at 6 months by SPECT

7%↑ EF by echo at 1year
Anterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: yes

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Lonborg et al.118Four-30 s angioplasty balloon inflations/deflations118No difference in troponin T or LVEF

19%↓ MI size at 3 months by CMR

31%↑ in the myocardial salvage index

Less heart failure
Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: no

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Sorensson et al.119Four-60 s angioplasty balloon inflations/deflations76No difference in 48 h AUC CK-MB/TnT or myocardial salvage by CMR at Day 7–9

Increase in myocardial salvage in patients with large AAR (>30% of LV).
Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: yes

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Cyclosporin A
Piot et al.68IV CsA (2.5 mg/kg) 10 min prior to PPCI5844%↓MI size (72 h AUC total CK)

20% ↓MI size (CMR in subset of 27 patients)

28% ↓MI size and smaller LVESV on CMR at 6 months120
Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: yes

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Therapeutic hypothermia
Gotberg et al.,63 RAPID-MI-ICECooling by IV infusion of 1–2 L of cold saline and central venous catheter cooling with Philips InnerCool RTx Endovascular System prior to PPCI to achieve a core body temperature of 35°C20Significant reduction in the MI size as % of AAR on CMR at 4 days 43% reduction in peak and cumulative trop T releaseAnterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: Yes
Erlinge et al., CHILL-MI, NCT01379261Cooling by IV infusion of 1–2 L of cold saline and central venous catheter cooling with Philips InnerCool RTx Endovascular System prior to PPCI to achieve a core body temperature of 35°C120MI size (as a % of AAR) by CMR at 4 daysAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: Yes
Therapeutic hyperoxaemia
O'Neill et al.,64 AMIHOT IIC hyperbaric hyperoxaemic reperfusion started after PPCI and continued for 90 min269No difference in primary endpoint (14 days MI size by SPECT)Anterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Stone et al.,121 AMIHOT IIIC hyperbaric hyperoxaemic reperfusion started after PPCI and continued for 90 min281No adverse events

No difference in the MI size by SPECT at 14 days or peak CK-MB or trop. pooled analysis of AMIHOT I and II trials suggested beneficial effects on the MI size and MACE
Anterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Remote ischaemic conditioning
Botker et al.66Four 5-min inflations/deflations of an upper arm cuff delivered in ambulance by paramedics prior to PPCI142Increase in the myocardial salvage index at 30 days. No difference in the MI size (SPECT or Peak Trop). Ant STEMI subgroup had greater myocardial salvage, smaller MI size, and better LV function at 3 days122Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Rentoukas et al.123Three-4 min inflations/deflations of an upper arm cuff delivered on arrival at the hospital prior to PPCI93Better ST resolution and lower peak Trop I. Synergistic effects with morphineAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Exenatide
Lonborg et al.69IV infusion of exenatide started 15 min prior to PPCI for 6 h107Increase in the myocardial salvage index at 90 days by CMR. Reduced MI size as % of AAR at 90 days by CMR. Patients presenting with short ischaemic times (≤132 min) had greater myocardial salvage124Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: no

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Glucose insulin potassium (GIK) therapy
Selker et al., IMMEDIATE72Iv GIK infusion for 12 h started by paramedics in ambulance—prior to reperfusion357No difference in progression to MI

Reduction in the MI size and less in-hospital mortality and cardiac arrest
Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
  • Anterior STEMI only, Only anterior STEMI patients included, Only PPCI or thrombolysis: Only either PPCI or thrombolysis patients included, AAR measured: Area at risk measured; Collateral flow excluded: Coronary collateralization to the AAR excluded; TIMI flow grade <1, TIMI flow grade less than 1 in the infarct-related artery prior to PCI; Treatment prior to or at reperfusion, Study intervention given prior to or at reperfusion.

Several of the failed study interventions listed in Table 1, including anti-oxidants, calcium-channel antagonists, adenosine, and erythropoietin had not shown conclusive cardioprotection in the pre-clinical animal studies, which may in part explain why they failed in the clinical setting. Another reason for the negative studies may be that many of them were designed to target only one proponent of myocardial reperfusion injury such as oxidative stress, calcium channel accumulation, apoptosis, and inflammation (see Table 1).

3.2 Experimental animal MI models

Many of the experimental animal MI models used to investigate study interventions in the pre-clinical setting do not adequately represent the clinical setting of a patient presenting with an acute MI undergoing myocardial reperfusion (for a summary of the major factors, see Supplementary material online, Table S1). This topic has been discussed in detail in several comprehensive reviews.1,3,4,57,58

3.3 Design of the clinical cardioprotection study

It is essential that the design of the clinical cardioprotection study takes into consideration the findings of previously published pre-clinical and clinical studies.

4. Confounding factors in STEMI cardioprotection studies

There currently exists no recognized effective therapeutic intervention for protecting the cardiomyocyte from the detrimental effects of either MVO or lethal myocardial reperfusion injury in acute MI patients. Over the last two to three decades, a large number of therapeutic interventions have been investigated as adjuncts to myocardial reperfusion. However, the results from the majority of these studies have been largely disappointing in terms of finding an effective therapy for reducing myocardial reperfusion injury and improving clinical outcomes in STEMI patients undergoing PPCI. Table 1 provides a summary of the major clinical studies which have failed to demonstrate any benefit in reperfused STEMI patients, and highlights some of the potential reasons for their failure, many of which include not taking into account confounding factors to cardioprotection.

A number of novel therapeutic interventions have been reported in small proof-of-concept clinical studies to prevent lethal myocardial reperfusion injury in STEMI patients undergoing PPCI (Table 2). These include mechanical therapeutic strategies such as therapeutic hypothermia,63 therapeutic hyperoxaemia,64 IPost,65 RIC,66 and pharmacological therapies such as atrial natriuretic peptide (ANP),67 CsA,68 and exenatide.69 Large multicentre clinical studies are now required to determine whether these promising therapeutic interventions can actually improve major clinical endpoints in STEMI patients treated by PPCI. In this regard, for CsA, RIC, and IPost these studies are currently underway (see Supplementary material online, Table S2).70,71

In addition to applying the cardioprotective strategy at the time of PPCI to prevent lethal myocardial reperfusion injury, there is also the opportunity of intervening at an earlier time-point, in the ambulance while in transit to the PPCI centre, in order to protect against acute myocardial ischaemic injury. This approach has recently been shown to be beneficial in proof-of-concept clinical studies investigating RIC and glucose–insulin–potassium therapy administered in the ambulance66,72 and is currently being investigated using metoprolol (Ibanez et al. METOCARD-CNIC NCT01311700). Table 3 provides a summary of some of the major therapeutic interventions which are currently being investigated as cardioprotective therapies for reducing lethal myocardial reperfusion injury in PPCI patients.

View this table:
Table 3

Clinical studies investigating therapeutic interventions administered at myocardial reperfusion which have potential promise in STEMI patients

Clinical studyTherapeutic interventionn, numberOutcomeNotes
Adenosine
Garcia-Dorado et al.,60 PROMISEIntracoronary adenosine 4 mg prior to PPCI201MI size on CMR at 5–10 days. Ongoing studyAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: yes

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Bendavia (MTP)
Chakrabarti et al., EMBRACE, NCT01572909Bendavia at time of PPCI.200Primary endpoint is the MI size (72 h AUC CK-MB)Anterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Impella 2.5
Moses et al., MINI-AMI, NCT01319760Impella 2.5 after PPCI for 24 h50Primary endpoint is the MI size at 3–5 days by CMRAnterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Insulin-like growth factor-1
Caplice et al., RESUS-AMI, NCT01438086Intracoronary rhIGF-1 (mecasermin)45Serum glucose and change in the LVEF on CMRAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Mangafodipir (Teslascan)
Karlsson et al., MANAMI, NCT00966563Iv infusion over 2–5 min prior to PPCI20The primary endpoint is the MI size (Trop T/CK-MB)Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Melatonin
Dominguez-Rodrigeuz et al., MARIA, NCT00640094Iv infusion at time of PPCI272The primary endpoint is the MI size (72 h AUC alpha-hydroxybutyrate dehydrogenase)Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Halladin et al., NCT01172171Intracoronary and iv infusion at time of PPCI60The primary endpoint is the MI size (CMR at 1 month)Anterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Nitric oxide (inhaled)
Janssens et al., NOMI, NCT01398384Inhaled nitric oxide prior to PPCI230The primary endpoint is the MI size as a % of LV at 3 days by CMRAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Nitrite (sodium)
Frennaux et al., NIAMI, NCT01388504Iv bolus of sodium nitrite given 5 min prior to PPCI200The primary endpoint is the MI size as a % of AAR at 10–14 days by CMRAnterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Mathur et al., NITRITE-AMI, NCT01584453Intracoronary bolus of sodium nitrite over 30–60 s at the time of PPCI80The primary endpoint is the MI size (48 h Trop T AUC)Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
RIC and local IPost
Prunier et al., RIRE-1, NCT01390142Four 5-min inflations/deflations of the upper arm cuff prior to PPCI plus four-1min inflations/deflations of angioplasty balloon after PPCI50The primary endpoint is the MI size (72 h CK-MB AUC) and MI size at 3 months (CMR)Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Sevoflurane
Lavi et al., SIAMI, NCT00971607Inhaled sevoflurane during PPCI50The primary endpoint is the MI size (serum biomarkers over 72 h).Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
Thymosin Beta 4
Strobeck et al., NCT00378352Iv injection of RGN-352 (Thymosin Beta 4)75The primary endpoint is the MI size on CMR at 28 daysAnterior STEMI only: yes

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: no

Treatment prior to or at reperfusion: yes
TRO40303
Atar et al., MitoCare, NCT01374321Peripheral IV infusion of TRO40303 started at 5–15 min prior to PPCI180MI size (72 h AUC CK and Trop I)Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
Metoprolol
Ibanez et al., METOCARD-CNIC, NCT01311700Iv metoprolol three-5 mg boluses administered in ambulance prior to PPCI220MI size (5–7 days by CMR)Anterior STEMI only: no

Only PPCI or thrombolysis: yes

AAR measured: no

Collateral flow excluded: no

TIMI flow grade <1: yes

Treatment prior to or at reperfusion: yes
  • Anterior STEMI only, only anterior STEMI patients included; only PPCI or thrombolysis, only either PPCI or thrombolysis patients included; AAR measured, area at risk measured; collateral flow excluded, coronary collateralization to the AAR excluded; TIMI flow grade <1, TIMI flow grade <1 in the infarct-related artery prior to PCI; treatment prior to or at reperfusion, study intervention given prior to or at reperfusion.

Based on extensive experimental data, and the findings from recent proof-of-concept clinical studies, particularly those which have investigated IPost in STEMI patients, our new understanding of the pathophysiology of acute IRI now allows us to propose recommendations for optimizing the design of clinical ‘cardioprotection’ trials. To increase our capacity to successfully transfer basic science knowledge into clinical practice for the patient's benefit, one may consider two distinct categories of confounding factors: (i) those factors which can be controlled for, and (ii) those that cannot be controlled for (see Figure 1). It is also important to realize that the confounding factors will vary according to the clinical situation, i.e. they are not the same for the STEMI and CABG setting.

Figure 1

Summary of confounding factors which impact on the sensitivity to ischaemia–reperfusion injury (IRI) and the response to the study intervention in STEMI patients. These can be divided into those factors which can be controlled for and those factors which cannot be controlled for when designing a clinical cardioprotection STEMI study.

4.1 Confounding factors which can be controlled for

Some factors are known as major determinants of MI size and must therefore be measured or taken into account in MI size reduction studies. Not doing so will either decrease the statistical power of the trial and/or result in a misinterpretation of the results, most often by concluding that the study is negative, thereby missing the opportunity for discovering new therapies for acute MI patients.

4.1.1 Patient selection

It must be appreciated that many of the clinical cardioprotection STEMI studies often exclude the most ill STEMI patients—these include those with critical life-threatening conditions such as cardiac arrest, cardiogenic shock, severe ventricular arrhythmias, and comorbidities. In this regard, mechanical cardioprotective strategies, such as RIC, may be particularly beneficial in this patient group, as they have the potential to mediate multiorgan protection.

4.1.2 Choice of reperfusion strategy

One can hypothesize that the choice of reperfusion strategy between PPCI and thrombolysis may impact on the severity of MVO and lethal myocardial reperfusion injury experienced by the STEMI patients, and therefore have an effect on the cardioprotective efficacy of the study intervention. Pre-clinical data suggest that gradual or low-pressure reperfusion can limit the MI size when compared with unimpeded myocardial reperfusion.7375 In fact, this phenomenon76 underlies the therapeutic basis of IPost, in which myocardial reperfusion occurs in a stuttered manner as it is interrupted by short-lived episodes of myocardial ischaemia, which has been reported to improve myocardial reperfusion, prevent endothelial dysfunction, reduce inflammation, attenuate apoptotic cell death, and limit MI size.6 Therefore, in PPCI, where myocardial reperfusion occurs both abruptly and completely, one may expect there to be a greater degree of myocardial reperfusion injury when compared with thrombolysis, in which myocardial reperfusion takes place more gradually and less completely. Furthermore, the precise time and adequacy of reperfusion are unknown in patients treated with thrombolytic agents, uncertainties which will make it difficult to have comparable control and treatment groups. Alternatively, one should, however, acknowledge that previous trials directly comparing the efficacy of thrombolysis vs. PPCI in STEMI patients have not established that any form of IRI (e.g. MI size, clinical outcome) was significantly attenuated by thrombolysis with respect to PPCI. Yet, one cannot rule out that study interventions administered at the time of myocardial reperfusion may result in different outcomes depending on whether PPCI or thrombolytic therapy is employed to restore the coronary flow in the infarct-related artery. Therefore, clinical cardioprotection studies of STEMI patients should include only one of these two modes of reperfusion therapy, either thrombolysis or PPCI, as the myocardial reperfusion strategy. Interestingly, many of the early failed attempts to reduce myocardial reperfusion injury in the clinical setting were undertaken in the pre-PPCI era with the majority of patients receiving thrombolytic therapy (Table 1). Whether a different outcome would have been observed in the setting of PPCI is not known. On the other hand, since PPCI is indeed poorly accessible in most non-Western countries, it is important that cardioprotective interventions be tested using the two different reperfusion strategies in separate studies, To provide potential benefit in the largest possible number of patients worldwide. However, it must be appreciated that myocardial reperfusion by thrombolytic therapy is not the ideal model for investigating the efficacy of novel cardioprotective strategies in STEMI patients because of the issues outlined above.

4.1.3 Timing the therapeutic intervention

Timing the administration of the therapeutic intervention in STEMI patients undergoing myocardial reperfusion using either thrombolytic therapy or PPCI is essential. The detrimental effects of myocardial reperfusion injury occur in the first few minutes of myocardial reperfusion, with pre-clinical animal MI studies demonstrating that unless the study intervention is administered prior to myocardial reperfusion, it is ineffective.33 The failure to administer the study intervention prior to myocardial reperfusion in some clinical studies may explain in part some of the negative data shown in Table 1.

The study treatment may be administered at any time between first patient contact and the time of reperfusion, provided the pharmacokinetics of the drug allow sufficient delivery to the target organ as soon as the myocardial blood flow is turned on again. This may explain the negative results observed in clinical studies investigating oral atorvastatin as a cardioprotective intervention in STEMI patients (see Table 1). The study intervention may, for example, be administered in the ambulance to the suspected STEMI patient while in transit to the hospital. This therapeutic approach has been employed with GIK therapy72 and RIC,66 and is currently being investigated for metoprolol therapy. However, one specific limitation of this treatment strategy is that a significant proportion of suspected STEMI patients (perhaps 20–30%) will end up not having a diagnosis of STEMI, and will have therefore received the therapeutic intervention un-necessarily. The same problem applies to administering the cardioprotective strategy on immediate arrival at the hospital. One potential approach for selecting STEMI patients is to only select patients for study after coronary angiography has taken place. This approach will also allow one to exclude those patients with TIMI>1 coronary flow in the infarct-related artery and significant coronary collateralization to the area at risk (AAR) (Rentrop grade>1). The obvious disadvantage of waiting until coronary angiography has taken place is the limited time remaining to then administer the therapeutic intervention before myocardial reperfusion takes place. Pharmacological study interventions should be administered either using the iv or intracoronary route to ensure that therapeutic concentrations of the study agent are achieved prior to myocardial reperfusion. The intracoronary route will achieve higher local concentrations within the myocardium, which may allow a lower dose of the drug to be used should the pharmacological agent have systemic hemodynamic effects. In our future daily practice, it is likely that the optimal timing of administration of a proven protective drug will have to comply with its modalities of administration and its pharmacokinetics.

4.1.4 Major determinants of MI size

For clinical cardioprotection trials investigating the MI-limiting effects of a study intervention, it is essential to assess for the major determinants of MI size in STEMI patients undergoing myocardial reperfusion.

4.1.4.1 Ischaemic time

The duration of acute myocardial ischaemia is a major determinant of final MI size. In pre-clinical animal MI studies investigating novel therapeutic interventions the ischaemic time can obviously be chosen to generate a relatively fixed MI size. However, in STEMI patients, the ischaemic time can vary between 0 to 12 h, depending on the chest pain onset to reperfusion time, resulting in widely variable MI sizes. Myocardial reperfusion accrues the most benefit in terms of myocardial salvage in those patients presenting within 3 h of chest pain onset. Whether MI size reduction with a study intervention is greater in patients presenting early (within 3 h) or later (3 h and beyond) is not clear. Two clinical studies have reported greater benefit with the pharmacological agents adenosine or exenatide in terms of myocardial salvage in patients presenting within 2–3 h of chest pain onset, suggesting that the former may be true (see Table 2). Whether IPost or RIC is more beneficial when administered to patients presenting with shorter or longer ischaemic times is not clear. One pre-clinical study suggests that IPost was actually harmful if applied following a short episode of index ischaemia,77 suggesting that IPost may be more beneficial in patients with longer ischaemic times. Moreover, laboratory studies suggest the relative importance of different mechanisms of reperfusion injury may depend on the duration of ischaemia, with mitochondrial permeability transition playing a more prominent role after prolonged ischaemia.78

However, it is important to consider that the benefit obtained in terms of myocardial salvage does not necessarily result in patient benefit expressed in terms of clinical outcomes. Although myocardial salvage following a protective intervention may be greater in patients reperfused within the first 3 h of onset of symptoms it is most likely related to a reduction of the ischaemic damage that has developed rapidly in the first hours of ischaemia. However, these patients usually display small infarcts with good clinical prognosis, so that the improved myocardial salvage may even not be clinically visible. On the other end of the spectrum, one cannot exclude that even mild myocardial salvage in patients with a prolonged (>6 h) ischaemic insult may translate into a significant clinical benefit, including limitation of adverse LV remodelling for example. Additional studies are required to actually understand the impact of the ischaemia time on IRI and clinical outcome.

4.1.4.2 The area at risk

The size of the AAR is a major determinant of the final MI size.79 Because of this, it is essential to take into account the size of the AAR when assessing MI size reduction with novel therapeutic interventions in clinical studies. The ability to measure the AAR is particularly important in STEMI patients, where the size of the AAR can vary greatly (from 10 to 50% of the LV) depending on which the coronary artery is involved (LAD, RCA, or Cx) and where along the vessel the occlusion has occurred (proximal, mid-vessel, or distal).80 However, in most clinical cardioprotection STEMI trials, the size of the AAR is not measured. In clinical studies, the MI size can be expressed as a percentage of the size of the AAR in a similar manner to the pre-clinical studies. However, the more conventional approach is to calculate the myocardial salvage index (MSI, which is defined as the size of the AAR subtract MI size divided by size of the AAR).

There are several different techniques available for estimating the size of the AAR and calculating the MSI in STEMI patients undergoing myocardial reperfusion (Supplementary material online, Table S3).81 The current gold standard method for measuring AAR is 99mTechnetium-Sestamibi single-photon emission tomography (SPECT), with cardiac MRI emerging as a potential alternative approach, although both these imaging techniques have their drawbacks (Supplementary material online, Table S3).58,82,83 With respect to the potential use of cardiac MRI for delineating the AAR in STEMI patients undergoing PPCI, the field is particularly controversial with several issues of concern including the technical limitations surrounding T2-weighted imaging84 and the possibility of the cardioprotective intervention reducing the size of the AAR by decreasing the extent of myocardial oedema.83

Previous clinical cardioprotection studies in STEMI patients suggest that the patients most likely to benefit from a study intervention administered as an adjunct to myocardial reperfusion are those presenting with a large AAR (>30% of the LV—usually proximal LAD and RCA STEMI patients).65,66,68,69,85 Therefore, in those clinical studies in which all STEMI patients were included, irrespective of the size of the AAR, there is a possibility that any cardioprotective effect associated with the study intervention is diluted and this may account in part for some of the negative clinical cardioprotection studies in which all-STEMI patients were included (Table 1).86 Therefore, it is essential that the AAR is measured when designing a clinical cardioprotection study comprising STEMI patients.

4.1.4.3 Coronary collateralization to the area at risk

The presence of coronary collateralization to the AAR may provide some residual blood flow to the ischaemic bed after a coronary artery occlusion and reduce MI size.87 Therefore, collateral flow needs to be measured in clinical studies when investigating MI-reduction study interventions. About 15–20% of STEMI patients will have significant coronary collateralization to the AAR.58 These patients sustain smaller myocardial infarcts and have better clinical outcomes, when compared with those patients with little collateralization. However, measuring collateral flow reliably in patients presenting with a STEMI is challenging. At the time of coronary angiography, the Rentrop grading system can be used to assess whether significant coronary collateralization to the AAR is present, and these patients should therefore be excluded from clinical cardioprotection studies of STEMI patients, as they are less likely to benefit from a study intervention. Including patients with significant coronary collateralization to the AAR may in part explain the negative findings of the clinical cardioprotection studies listed in Table 1.

4.1.4.4 Coronary artery flow prior to myocardial reperfusion

Because modern efficient anti-platelet and anti-thrombotic therapies are instituted early, >40% of STEMI patients presenting to the hospital will have spontaneously reperfused prior to PPCI and will already have a significant coronary flow (TIMI flow >1) within the culprit coronary artery.66 For these patients, in whom myocardial reperfusion has already taken place, the prognosis is improved when compared with those patients presenting with a fully occluded culprit artery. For the study intervention to be effective against myocardial reperfusion injury, it needs to be administered to the STEMI patient while the culprit artery is still occluded and prior to myocardial reperfusion. This would explain why STEMI patients presenting with an occluded culprit artery accrued the most benefit in terms of MI size reduction with RIC.66 On this basis, for clinical cardioprotection studies, it is advisable to only include those STEMI patients with an occluded culprit artery.

One may question whether these four major confounding factors ought to be measured in clinical outcome studies. Here, the endpoint is not MI size, but rather death, or hospitalization for heart failure for example. In these trials, it remains essential not to include patients who display a spontaneously re-opened coronary artery on admission coronary angiography, since they have already undergone myocardial reperfusion injury (before the protective intervention could be administered). Not considering patients with visible collaterals should also be recommended, as mentioned above. But, whether AAR is a strong predictor of such clinical events remains to be demonstrated and the recruitment of a large number of patients together with the randomization process will balance between the placebo and the active treatment group the distribution of ischaemia time and sizes of AAR. Where it is difficult to perform such a large clinical study one may consider stratifying the study with respect to the AAR. One might yet want to include patients with large AAR (e.g. anterior infarcts) who constitute the high-risk population which would benefit the most, in terms of clinical outcome, of cardioprotective interventions.

4.1.4.5 Endpoints of cardioprotection

For a clinical cardioprotection study in STEMI patients, it is essential to choose study endpoints which are most relevant to the MI size limiting effects of the study intervention being investigated. For proof-of-concept clinical studies, this will most likely include endpoints of cardioprotection such as MI size (using either 48 h AUC cardiac troponins or late gadolinium enhancement on cardiac MRI), left ventricular systolic function, and indexed left ventricular volumes. If the AAR is measured in the clinical study, then MI size should take into account the size of the AAR, which will increase the statistical power of the clinical trial for detecting a significant reduction in MI size, thereby reducing the number of patients required for the study.

In terms of designing the larger clinical outcome studies, it is crucial to choose major adverse cardiac events (MACEs) which are relevant to the MI size limiting effects of the study intervention. In this regard, the combined rates of cardiac death and hospitalization for heart failure are most relevant to MI size limitation in STEMI patients as a combined primary study endpoint, whereas rates of coronary revascularization and non-fatal MI are less relevant and unlikely to be influenced by a MI size limiting study intervention.

4.2 Confounding factors which cannot be controlled for

In CHD patients, there are a large number of confounding factors which can potentially alter the sensitivity of the heart to acute IRI and/or interfere with the efficacy of a particular cardioprotective study intervention. A major cause of these confounding factors relates to the fact that CHD is caused by or associated with known cardiovascular risk factors and co-morbidities, including ageing, hypertension, hyperlipidaemia, diabetes, left ventricular hypertrophy, heart failure, and uraemia.88 Pre-clinical animal studies suggest that these diseases and their pharmacological treatments induce fundamental molecular alterations in the heart that can potentially affect the cytoprotective signalling pathways, thereby affecting both the sensitivity to IRI and the response to a particular cardioprotection strategy (reviewed in88,89). Currently, most of the animal MI models which are used to assess the efficacy of a novel cardioprotective strategy use healthy juvenile animals which are free of any co-morbid disease.1 Furthermore, pre-clinical studies have reported that human atrial tissue harvested at time of cardiac bypass surgery from aged patients90 and diabetic patients,91,92 were resistant to IPC applied ex vivo in a model of simulated IRI. Large-scale cohorts of STEMI patients are needed to analyse how much age, gender, co-morbidities, and co-treatments may affect IRI and response to protective interventions. In addition, it has to be taken into account that most patients display several comorbidities. Specific analyses will then help adapt future therapies to specific subgroups of patients.

Another major confounding factor for cardioprotection is concomitant medication, which patients are on for their cardiovascular risk factor, co-morbid condition, or as part of the treatment of the ongoing acute MI. These pharmaceutical agents have been shown in pre-clinical and clinical studies to either block the cardioprotective effect (for example, certain oral anti-diabetic sulphonylureas, nitrates when nitrate tolerance develops, certain statins) or induce cardioprotection themselves (for example, insulin and some anti-diabetic medications, some statins, ACE-inhibitors, anti-platelet agents, volatile anaesthetic agents, opioids, and so on (reviewed in4,88,89,9395). So far, the animal MI models which are used to assess the efficacy of a novel cardioprotective strategy do not investigate the effect of concomitant medication. Treatment with oral sulphonylureas such as glibenclamide96 and the anti-anginal agent, nicorandil,97 also interferes with IPC protection in human atrial tissue. With respect to concomitant anti-platelet therapy, there is preliminary animal data suggesting that these agents may actually confer direct protection against acute IRI,98,99 an observation which is likely to have a significant impact on clinical cardioprotection studies, as most CHD patients will be on anti-platelet therapy.

Whether these confounding factors can actually interfere with the efficacy of a cardioprotective intervention or an endogenous cardioprotective phenomenon in the clinical setting has only been investigated directly in relatively few clinical studies (for a summary of the major clinical studies, see Supplementary material online, Table S4). None of these confounding factors was pre-specified but all were investigated as retrospective post hoc subgroup analyses. In a recent retrospective analysis of proof-of-concept IPost trials, Roubille et al.93 reported that clopidogrel administered before PCI may indeed be a confounder both for sensitivity to IRI and response to angioplasty postconditioning in STEMI patients. This may in part explain some of the recently published clinical studies which failed to demonstrate any cardioprotective effect with IPost.100,101

5. Confounding factors in CABG cardioprotection studies

A number of different therapeutic interventions have been investigated in the setting of CABG and major vascular surgery, with many of them failing to report any beneficial effects on peri-operative myocardial injury/infarction or clinical outcomes (Table 4). Recently, several proof-of-concept clinical trials have reported cardioprotective effects with therapeutic strategies including RIC, IPost, glucose–insulin–potassium (GIK) therapy and volatile anaesthetics (Table 5). However, even with RIC, not all clinical studies have been positive, an issue which is discussed in a later section.102 Whether RIC can improve clinical outcomes in patients undergoing CABG surgery is unknown, and is currently being investigated in two ongoing large multicentre randomized clinical trials such as the ERICCA (NCT01247545)103 and RIPHeart trials (NCT01067703).104 Large multicentre clinical trials are required to investigate whether GIK or volatile anaesthetics can improve clinical outcomes in patients undergoing CABG surgery.

View this table:
Table 4

Clinical studies which have failed to demonstrate any beneficial effect in CABG patients with a cardioprotective intervention

Clinical studyTherapeutic interventionn, numberPrimary outcomeNotes
Sodium-hydrogen ion exchange inhibitors
Boyce et al., GUARDIAN125Oral cariporide or placebo prior to surgery1477, CABGLess CK-MB release

25% ↓risk of death and non-fatal MI at 36 days

Beneficial effect maintained at 6 months post-surgery
Mentzer et al., EXPEDITION126Oral cariporide or placebo prior to surgery5761, CABGReduction in the primary endpoint of death and MI

However, increase in mortality due to cerebrovascular events
Off-target cerebral effects
Acadesine
Mangano et al.127Acadesine given as IV infusion and in the cardioplegic solution2695, CABGNo difference in the primary endpoint of cardiac death, MI, or stroke at 4 daysNo difference in peri-operative MI (PMI), but in the 100 patients who did have a PMI (3.7%) patients, acadesine reduced patient death
Newman et al., RED-CABG128Acadesine given as IV infusion and in cardioplegic solution3080, CABGNo difference in the primary endpoint of all-cause mortality, non-fatal stroke, or severe left ventricular dysfunction at Day 28Trial stopped early because of futility analysis indicating a very low likelihood of a statistically significant efficacious outcome
Pexelizumab
Verrier et al., PRIMO-CABG129IV pexelizumab bolus prior to CABG followed by 24 h infusion2476, CABGNon-significant reduction in the primary combined 30-day endpoint of death and non-fatal MIOnly targets the anti-inflammatory component of acute IRI
Smith et al., PRIMO-CABG2130IV pexelizumab bolus prior to CABG followed by 24 h infusion4254, CABGNo difference in the primary combined 30-day endpoint of death and non-fatal MIOnly targets the anti-inflammatory component of acute IRI
  • MI, myocardial infarction; TnI, Troponin I; TnT, Troponin T; h, hours.

View this table:
Table 5

Clinical studies investigating therapeutic interventions which have shown benefit in the setting of cardiac bypass or major vascular surgery

Clinical studyTherapeutic interventionn, number, surgeryPrimary outcomeNotes
Positive studiesRemote ischaemic conditioning
Cheung et al.131Four 5-min inflations/deflations of thigh cuff37 children, CHD43% reduction in 72 h AUC TnTReduced ventilation time and inotrope requirements
Hausenloy et al.132Three 5-min inflations/deflations of the upper arm cuff58 adults, CABG43% reduction in 72 h AUC TnT
Ali et al.133Clamping of the right common iliac artery for 10 min followed by clamping of the left common iliac artery for 10 min82 adults, AAA98% reduction in 7 days AUC Trop ILess acute kidney injury
Venugopal et al.134Three 5-min inflations/deflations of the upper arm cuff45 adults, CABG42% reduction in 72 h AUC TnT
Wagner et al.135Three 5-min inflations/deflations of the upper arm cuff 18 h prior to surgery67 adults, CABG+AVRReduction in peak Trop-I at 8 hFirst demonstration of delayed RIC in this clinical setting

St Thomas’ crystalloid cardioplegia
Thielmann et al.136Three 5-min inflations/deflations of the upper arm cuff53 adults, CABG45% reduction in 72 h AUC TnIBretschneider crystalloid cardioplegia used
Li et al.137Three 5-min inflations/deflations of the upper arm cuff after aortic cross-clamp82, valve surgery onlyReduction in peak TnT at 30 minFirst demonstration of delayed RIC in this clinical setting

However, when RIC protocol delivered prior to CABG surgery, no difference in peak Trop T
Choi et al.138Three 10-min inflations/deflations of thigh arm cuff76, valve surgeryReduction in peak CK-MB at 24 h
Wu et al.139Three 5-min inflations/deflations of the upper arm cuff with two-10 min inflations/deflations of thigh cuff75, MVRReduced peak TnI at 6, 12, 24, 48, 72 hRIC of arm did not reduce TnI

First demonstration that combining arm and leg RIC more effective that arm alone
Kottenberg et al.140Three 5-min inflations/deflations of the upper arm cuff72 adults, CABG50% reduction in 72 h AUC TnI with Isoflurane but not Propofol.Bretschneider crystalloid cardioplegia used
Xie et al.141Three 5-min inflations/deflations of the upper arm cuff73 adults, valve surgery43% reduction in 72 h AUC TnI
Heusch et al.18Three 5-min inflations/deflations of the upper arm cuff23 adults, CABGA significant reduction in 72 h AUC TnIBretschneider crystalloid cardioplegia used
Negative studiesRemote ischaemic conditioning
Rahman et al.142Three 5-min inflations/deflations of the upper arm cuff162 adults, CABGNo difference in 48 h AUC TnTRIC protocol administered at the time of the surgical incision

IV GTN given to all patients

Sevoflurane and propofol both used for maintenance anaesthesia
Li et al.137Three 5-min inflations/deflations of the upper arm cuff82, valve surgery onlyNo difference in peak TnT at 30 minIsoflurane and propofol both used for maintenance anaesthesia

However, when RIC protocol delivered after aortic cross-clamp there was a significant reduction in peak Trop T
Karuppasamy et al.143Three 5-min inflations/deflations of the upper arm cuff53 adults, CABGNo difference in 48 h AUC TnI or CK-MBIsoflurane and propofol both used for maintenance anaesthesia
Wu et al.139Three 5-min inflations/deflations of the upper arm cuff75, MVRNo different in peak Trop IRIC did not reduce TnI unless except for combined arm and leg RIC
Luchinetti et al.144Four 5-min inflations/deflations of thigh cuff57, CABGNo difference in hsTnTPropofol used during induction and isoflurane used for maintenance anaesthesia
Young et al.145Three 5-min inflations/deflations of the upper arm cuff96 adults, CABGHigher plasma levels of hsTnT at 6 and 12 h with RICRIC protocol administered at the time of surgical incision

Isoflurane and propofol both used for maintenance anaesthesia
Lomivorotov et al.146Three 5-min inflations/deflations of the upper arm cuff80 adults, CABGNo difference in Trop I or CK-MBIsoflurane and propofol both used for maintenance anaesthesia

Improved cardiac index
Ischaemic postconditioning
Luo et al.1472 × 30 s cycles of aortic cross-clamping24 children, CHDReduction in TnI and CK-MB at 4 h post-IPostFirst study to show efficacy with IPost in cardiac bypass surgery
Luo et al.1483 × 30 s cycles of aortic cross-clamping50 adults, AVRReduction in CK-MB but not TnIIPost also resulted in a reduction in inotrope requirement
Li et al.1492 × 30 s cycles of aortic cross-clamping99 children, CHDReduction in TnI at 4 h post-IPostIPost also resulted in a 44% reduction in ventilation time and 40% reduction in inotrope requirement
Glucose insulin potassium (GIK) therapy
Ranasinghe et al.150Iv GIK given prior to CABG surgery until 6 h after aortic clamp removal plus or minus IV T3 (given on removal of aortic clamp for 6 h)440 adults, CABGBoth GIK and T3 therapy increased cardiac index, and reduced TnI at 6 and 12 h
Inhaled anaesthetics
Symons and Myles151, Meta-analysis of 27 studiesIsoflurane, sevoflurane, desflurane and enflurane2979, CABGBetter LV function

Less inotropes

Lower Trop I levels
Shorter duration of mechanical ventilation

Shorter hospital stay
Yu and Beattie152, Meta-analysis of 32 studiesSevoflurane, desflurane2841, CABGLess TnI at 6, 12, 24, and 48 h
Landoni et al.153, Meta-analysis of 21 studiesSevoflurane, desflurane1922, CABGDecreased incidence and magnitude of PMI

Shorter ITU and hospital stay
Less inotrope and ventilation requirements

A reduction in mortality
  • AVR, aortic valve replacement; CHD, congenital heart disease surgery; PMI, peri-operative MI; MI, myocardial infarction; MVR, mitral valve replacement, TnI, Troponin I; TnT, Troponin T; h, hours.

5.1 Patient selection

The heterogeneity of patients undergoing cardiopulmonary bypass surgery and the number of confounding factors which can potentially interfere with cardioprotection make patient selection for clinical cardioprotection studies in CABG patients quite challenging. The type of surgery is clearly important, with patients undergoing CABG alone very different from patients undergoing valve surgery. For example, the myocardium of patients with severe aortic stenosis or aortic regurgitation may be significantly hypertrophied, the presence of which may impact on the cardioprotective effect of a study intervention. For patients undergoing CABG surgery, it must be acknowledged that direct injury to the myocardium during surgery can happen and may contribute to cardiac enzyme release; this has to be taken into account when interpreting the results. Therefore, it may be advisable to select patients undergoing either CABG alone or valve surgery alone in a clinical cardioprotection study.

Of course the type of surgery will impact on the aortic cross-clamp time (the duration of acute global myocardial ischaemia) with longer cross-clamp times in patients undergoing more complex surgery (for example, CABG+valve surgery). This would be another reason for separating CABG alone patients from valve surgery patients in clinical cardioprotection studies.

Whether a patient has stable CHD or unstable CHD may also affect the response of the myocardium to the study intervention. Unstable patients may be sicker and have experienced episodes of chest pain at rest, which may have inadvertently preconditioned the myocardium against acute IRI.105 Therefore, for clinical cardioprotection studies it would be advisable to recruit either stable or unstable patients. Specifically, it is important in either stable or unstable patients to exclude patients who have experienced anginal symptoms in the 24 h prior to surgery, to exclude the confounding effect of IPC.

5.2 Peri-operative factors

5.2.1 Concomitant medication

A wide variety of pharmacological agents used during cardiopulmonary bypass surgery may interfere with the cardioprotective efficacy of a study intervention. Volatile anaesthetic agents (such as isoflurane and sevoflurane) and the iv anaesthetic agent, propofol, have been reported to either confer cardioprotection themselves or interfere with RIC cardioprotection (see Table 5). Furthermore, the use of iv GTN, nitroprusside, and opioid analgesics may also interfere with the cardioprotective effects of a study intervention. However, it may be difficult to standardize the anaesthetic regimen and concomitant medication in clinical cardioprotection studies given the variations in practice. Providing the study is adequately powered and properly randomized these confounding factors should distribute themselves equally between the study intervention and control treatment groups. Where it is difficult to perform such a large clinical study one may consider stratifying the study with respect to these confounding factors.

5.2.2 Myocardial preservation strategy

During cardiopulmonary bypass surgery, it is essential to create a blood-free and motionless operative field, in order to improve visibility, facilitate the surgical procedure and prevent air-embolism. This is achieved by cross-clamping the aorta (to isolate the heart from the systemic circulation) and inducing electrochemical cardiac arrest (to stop the heart beating) using cardioplegic solution, respectively. The choice of myocardial preservation strategy (blood cardioplegia, crystalloid cardioplegia, or cross-clamp fibrillation) may impact on the cardioprotective efficacy of the study agent, but again providing the study is adequately powered and properly randomized this should not be a major issue.

5.3 The therapeutic intervention

In clinical cardioprotection studies in cardiopulmonary surgery, there is the opportunity to apply the study intervention at several different time-points: either before CABG surgery begins (prior to acute myocardial ischaemia), by adding a pharmacological intervention to the cardioplegic solution after aortic cross-clamping (after the onset of acute myocardial ischaemia), or at the time of aortic declamping (at the time of myocardial reperfusion). It is important to ensure that the time elapsed between administering the preconditioning study intervention and the time of aortic declamping (acute myocardial reperfusion injury) does not exceed the 2–3 h, as this corresponds to the cardioprotective window of protection elicited by IPC.

5.4 Endpoints to assess cardioprotection

For a clinical cardioprotection study in CABG patients, it is essential to choose study endpoints which are most relevant to the cardioprotective effects of the study intervention being investigated. For proof-of-concept clinical studies, this will most likely include surrogate endpoints of cardioprotection such as the magnitude of peri-operative myocardial injury (using 72 h AUC cardiac troponins), inotrope requirements, and left ventricular systolic function (which should be measured both acutely and also after 3–4 months to allow for any recovery of LV contractile function from the effects of CABG surgery). In the recently published ‘Third Universal Definition of Myocardial Infarction’,45 MI associated with CABG has been arbitrarily defined as an elevation of cardiac biomarker values >10 × 99th percentile URL in patients with normal baseline Troponin values (<99th percentile URL), associated with either (i) new pathological Q-waves or new LBBB or (ii) angiographic documented new graft or new native coronary artery occlusion or (iii) imaging evidence of new loss of viable myocardium or new regional wall motion abnormality.

In CABG patients, where vital organs other than the heart are also subjected to acute IRI, there is an opportunity to investigate whether the study intervention can also confer protection against the detrimental effects of IRI in the lung (measure the ventilation time), brain (assess cognitive function), and kidney (incidence of acute kidney injury).

Once the proof-of-concept clinical study has demonstrated a beneficial effect with a particular study intervention with respect to surrogate endpoints of cardioprotection, the next objective is to determine whether the study intervention can actually improve clinical outcomes in cardiopulmonary bypass surgery patients as evidenced by reducing MACEs. When designing the larger clinical outcome studies, it is preferable to choose MACEs which are most relevant to the cardioprotective effects of the study intervention. Combined rates of cardiac death and hospitalization for heart failure may be the most relevant for CABG patients as a combined primary study endpoint, whereas rates of coronary revascularization, non-fatal MI, and stroke may be less relevant and unlikely to be influenced by a cardioprotective intervention, but these data can be used to assess the safety of the study intervention.

6. Summary and conclusions

Previous attempts to protect the heart against the detrimental effects of acute IRI in patients with CHD have been largely disappointing. One major contributing factor for this failure to translate cardioprotective interventions discovered in animal studies into the clinical setting can be attributed to problems with the clinical study design. In this ESC Working Group Cellular Biology of the Heart Position Paper, we provide recommendations to help optimize the design of clinical cardioprotection studies in STEMI and CABG patients (see Tables 6 and 7), which take into account the experience from previously published pre-clinical and clinical data. The hope would be to improve the translation of cardioprotective strategies into the clinical setting for the benefit of CHD patients.

View this table:
Table 6

Recommendations for designing MI-limiting studies in STEMI patients

Patient selection
  • Select patients with a large area at risk (AAR) (>30% of the left ventricle)

  • Select patients with no significant coronary collateralization to the AAR (Rentrop grade <1)

  • Select patients with an occluded culprit artery at the time of study intervention administration (TIMI flow grade 0 or 1)

The study intervention
  • Select a study intervention which has shown conclusive cardioprotection in pre-clinical studies

  • Administer the study intervention as an iv or intracoronary bolus prior to myocardial reperfusion

Choose MI-limiting-related study endpoints
  • MI size (48 h AUC cardiac enzymes or late gadolinium enhancement cardiac MRI)

  • Myocardial salvage index (AAR-MI size/AAR)

  • Incidence and extent of microvascular obstruction (cardiac MRI)

  • Indexed left ventricular end systolic/diastolic dimensions (echocardiography or cardiac MRI)

  • Left ventricular systolic function (echocardiography or cardiac MRI)

  • Hospitalization for heart failure

  • Cardiac death

View this table:
Table 7

Recommendations for designing cardioprotection studies in CABG patients

Patient selection
  • Select patients undergoing on-pump cardiac surgery

  • Select patients undergoing either CABG or valve surgery

  • Select either stable or unstable patients

Peri-operative factors
  • Standardize the anaesthetic regimen and concomitant medication where possible

  • Select one myocardial preservation strategy only (blood cardioplegia, crystalloid cardioplegia, or cross-clamp fibrillation)

The therapeutic intervention
  • Select a study intervention which has shown conclusive cardioprotection in pre-clinical studies

  • Administer the study intervention either:

    1. prior to aortic clamping (pre-ischaemia)

    2. in the cardioplegic solution (after the onset of ischaemia)

    3. at the time of aortic declamping (at reperfusion)

Choose cardioprotection-related study endpoints
  • Incidence of type 5 CABG-related MI (see text for definition)

  • Peri-operative myocardial injury or infarction (72 h AUC cardiac enzymes)

  • Inotrope requirements

  • Left ventricular systolic function (echocardiography)

  • Hospitalization for heart failure

  • Cardiac death

Funding

The authors are grateful for the support of the British Heart Foundation (grant number FS/06/023). This study was carried out at the Heart Hospital/University College London Hospital and University College London, which received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme.

Acknowledgments

Conflict of interest: none declared.

References

View Abstract