Translating novel strategies for cardioprotection: the Hatter Workshop Recommendations.

Ischemic heart disease (IHD) is the leading cause of death worldwide. Novel cardioprotective strategies are therefore required to improve clinical outcomes in patients with IHD. Although a large number of novel cardioprotective strategies have been discovered in the research laboratory, their translation to the clinical setting has been largely disappointing. The reason for this failure can be attributed to a number of factors including the inadequacy of the animal ischemia-reperfusion injury models used in the preclinical cardioprotection studies and the inappropriate design and execution of the clinical cardioprotection studies. This important issue was the main topic of discussion of the UCL-Hatter Cardiovascular Institute 6th International Cardioprotection Workshop, the outcome of which has been published in this article as the "Hatter Workshop Recommendations". These have been proposed to provide guidance on the design and execution of both preclinical and clinical cardioprotection studies in order to facilitate the translation of future novel cardioprotective strategies for patient benefit.

Introduction

Ischemic heart disease (IHD) is the leading cause of death worldwide. As such, novel therapeutic strategies for protecting the heart against ischemia–reperfusion injury (IRI) are urgently needed to: reduce myocardial injury, preserve cardiac function, prevent the development of heart failure, and improve clinical outcomes in patients with IHD [42, 66]. However, a major obstacle to this process has been the inability to successfully translate novel cardioprotective strategies discovered in the research laboratory setting directly into the clinical arena [38].

This important issue was the main topic of discussion of the 6th Hatter Institute International Workshop on Cardioprotection, which was held this year in Mauritius, an Island in which diabetes and associated IHD are major contributors to overall morbidity and mortality. It was organized together with the Working Group of Cellular Biology of the Heart of the European Society of Cardiology. The main agenda of this International Workshop was to discuss and formulate a set of recommendations for the design and execution of future studies on cardioprotection in both the research laboratory and the clinical setting, in order to facilitate the translation of future novel cardioprotective strategies for patient benefit. One crucial aspect of this endeavour was to recognise the limitations in the design and execution of current experimental laboratory and clinical cardioprotection studies, a feature which was also highlighted by the NHLI Working Group in 2004 [7].

Inadequacy of the animal models of IRI

It is well accepted that the majority of animal models of IRI currently used to investigate novel cardioprotective strategies are inadequate representations of the clinical setting [7, 33, 36, 56], given the size and age of the animals used as well as their lack of co-morbidities and co-treatments. Ideally, one would have to prove efficacy of a certain cardioprotective intervention in animal experiments by the reduction in myocardial infarct (MI) size and/or improvement of prognosis under all mimicked clinical circumstances; however, this is unrealistic and it was agreed that the following recommendations should be proposed.

Animal selection

Species The response of the heart to IRI and cardioprotective strategies will vary depending on the species used. Often small animal MI models have been used to investigate cardioprotection, since knockout and/or silencing of target proteins is possible in these models. However, more expensive, large animal (canine, porcine, primate) MI models are needed to confirm results of small animal experiments before clinical testing, since the temporal and spatial development of MI as well as certain signalling pathways in small animals differ from that in larger mammals and humans [21, 53, 55]. Therefore, small animal MI models may be used for preliminary ‘screening’ of a novel cardioprotective strategy, as long as the latter is also demonstrated to be efficacious in at least one large animal MI model.

Age Patients with IHD usually present between the ages of 55 and 65 years, whereas many experimental studies use young adult rats and mice (aged 3–4 months) which are equivalent to the human age of 7–10 years [6]. Several studies have reported that with age the myocardium can become resistant to various cardioprotective strategies including ischemic preconditioning [48] and postconditioning [5, 44]. Therefore, it is essential to demonstrate that any novel cardioprotective strategy is effective in suitably aged animal hearts. The easiest and most convenient species for such experiments are mice and rats in which the human age of 55–65 years corresponds to about 21–24 months of age [6].

Sex Both male and female patients suffer from IHD, yet most preclinical cardioprotective studies are restricted to using male animals only. Several studies suggest that gender can impact on the myocardial sensitivity to different cardioprotective strategies [15, 56]. Therefore, it is necessary to establish whether or not a novel cardioprotective strategy is effective in both male and female animals.

Co-morbidities Patients with IHD are likely to have a number of co-morbid conditions at the time of presentation, many of which can influence the sensitivity of the myocardium to certain cardioprotective strategies [15]. Preclinical studies suggest that in the presence of diabetes [45, 50, 62] or the metabolic syndrome [63], the threshold for cardioprotection may be elevated or cardioprotection may be even absent (reviewed in [15]). Therefore, it is essential to establish whether or not a novel cardioprotective strategy is effective in the presence in one or more of these co-morbidities.

Medical therapy Many patients with pre-existing IHD take several types of medical therapy which may influence the effects of any novel cardioprotective strategy. Certain sulphonylureas such as glibenclamide may block the cardioprotection elicited by ischemic conditioning [39]. On the other hand, nicorandil, ACE-inhibitors, nitrates, statin therapy may inadvertently precondition the myocardium. Furthermore, on admission with an MI, additional therapy such as oxygen, morphine and nitrates may all precondition the myocardium or at least lower the threshold for protection. Similarly, patients undergoing elective coronary artery bypass graft (CABG) surgery are likely to receive inhalational anaesthetics such as isoflurane which has been reported to precondition the heart [58]. It is difficult to reproduce all concomitant therapy in preclinical animal models of cardioprotection but it is crucial to be aware of the limitations of such models without medication especially when co-morbidities are present. In designing the preclinical experiment to take all these factors into consideration, it may not be feasible or necessary to allow for every different combination and permutation. It may well be that prioritizing the confounders and medical therapy is more useful. In this respect, we would recommend that it is more important to address age before gender and the other co-morbidities. In terms of concomitant medical therapy, we would recommend the investigation of at least one relevant medical therapy.

The animal model of IRI

The type of model Acute coronary occlusion in animal models of MI is most often achieved by external occlusion of a healthy coronary artery. In contrast, in most patients presenting with an acute MI, the acute coronary occlusion is due to the formation of a thrombus at the site of a ruptured unstable atherosclerotic plaque. Furthermore, the process of reperfusion by percutaneous coronary intervention (PCI) in the clinical setting is more likely to be associated with a residual stenosis and with thrombotic embolisation than in the animal model [24]. Thus, the nature of the coronary artery occlusion and the methods used to re-establish perfusion in addition to the pro-inflammatory state of an acute MI may influence the effect of the cardioprotective strategy. Finding an animal model to reproduce these effects is a daunting task.

The clinical setting which most closely resembles the most frequently used animal model of MI is the IHD patient presenting acutely with a complete thrombotic coronary artery occlusion (a ST-elevation myocardial infarction, STEMI) undergoing myocardial reperfusion using PCI. Whether this animal MI model should be used to investigate novel cardioprotective strategies destined for use in different clinical settings such as cardiac arrest, cardiac transplantation, PCI or CABG surgery may be questioned. However, other more appropriate animal models are available to simulate the myocardial injury sustained in these clinical settings (see Table 1).

Table 1

Matching the animal IRI model for investigating the novel cardioprotective strategy to the clinical setting in which it is intended for use

Clinical setting	Animal IRI model	Example animal studies
STEMI patient undergoing PCI (regional acute myocardial ischemia and reperfusion)	Animal model of MI (regional acute myocardial ischemia and reperfusion)	Mice, rats, rabbits, dogs and pigs [56]
Stable IHD patient undergoing CABG surgery (global myocardial ischemia and reperfusion with cardioplegia, coronary microembolization and manual handling)	Use of cardio-pulmonary bypass for circulatory support (global myocardial ischemia and reperfusion ± cardioplegia)	Dogs, pigs and sheep [30, 31]
Cardiac arrest patient in which circulation has been restored (global myocardial ischemia and reperfusion)	Animal model of cardiac arrest and reperfusion (global myocardial ischemia and reperfusion)	Rats [59] and pigs [40]
Cardiac transplant patient (global myocardial ischemia and reperfusion with cardioplegia)	Heterotopic heart transplantation with global myocardial ischemia and reperfusion ± cardioplegia	Pigs [34]
Patient undergoing PCI (limited myocardial ischemic injury and coronary microembolization)	Coronary microembolization	Pigs [52]

Duration of ischemia The index ischemic time used in animal MI models is usually fixed having been selected to create a moderate to large infarct when expressed as the area at risk (AAR), but still allowing for protection to occur [42]. Furthermore, due to the fixed ischemic time the variation of the extent of MI is relatively small making any demonstration of cardioprotection easier. In contrast, MI patients present with widely varying ischemic times at the hospital (in the range of 1–12 h) and display huge variability in MI size even allowing for the same ischemic time because of the large number of confounding factors which can influence infarct size. Therefore, the demonstration of any kind of cardioprotection is complicated, requiring quite often subgroup analyses of patient data, thereby increasing the sample size.

Duration of reperfusion The reperfusion time used in animal MI models is usually fixed and ranges from 0.5 to 3 h depending on the particular model used. In rat and rabbit in vivo models the ultimate MI size by tetrazolium staining has been reported to be established after 3 h of reperfusion [67]. However, studies suggest that MI size in the canine heart continues to increase even after 6 h of reperfusion [68], although this issue is controversial [16]. For MI patients, the duration of reperfusion obviously depends on the time when one measures MI size. Therefore, the acute infarct-limiting effects of a novel cardioprotective strategy should be assessed when infarct size is fully developed but before remodelling occurs.

Endpoint of cardioprotection In most animal experiments, the reduction of MI size is taken as the endpoint of cardioprotection. MI size depends on the area at risk and on residual blood flow, and therefore MI size as a fraction of area at risk is the more robust endpoint of protection. The relation of MI size as a fraction of AAR to residual blood flow is probably the best endpoint of cardioprotection in animal experiments [53]. Recovery of contractile function is not only a function of MI size, but also of stunning and therefore less suited as an endpoint of protection. Of note, in clinical studies MI size is estimated from biomarkers and imaging and considered only a surrogate for clinical outcome.

It is clear from the previously published preclinical literature that an abundance of novel cardioprotective strategies have been discovered. However, the investigation of a particular novel cardioprotective strategy using a systematic, step-wise, collaborative approach has been lacking. Therefore, the formation of a collaborative network of research laboratories to test a single novel cardioprotective strategy in a range of animal MI models may be required to determine whether consistent cardioprotection is observed across species and laboratories. This would have the potential advantage of increasing the probability of a successful translation into the clinical setting, or alternatively, discouraging the initiation of a clinical trial that is destined to be unsuccessful. In 2002, this approach was used to investigate the cardioprotective effect of an adenosine A1 agonist. In this study, a multicentre randomised controlled double-blind experimental animal study was performed in three different laboratories [4].

Clinical settings of IRI

The pathophysiology of IRI obviously varies with the clinical setting. The purest example of classical IRI is the IHD patient presenting acutely with a complete thrombotic coronary artery occlusion (a STEMI) who undergoes PCI reperfusion therapy. Although IRI may contribute to the myocardial injury sustained in a number of other clinical settings such as CABG surgery, cardiac arrest and transplantation, other factors such as manual handling of the heart and coronary embolization [23] may come into play. They need to be taken into consideration when analysing the results. It should be noted, however, that the problem of distal embolization may also occur with PCI and the contribution of IRI per se to this form of injury is probably minor. Clearly, the appropriate animal IRI model should be selected to test the novel cardioprotective strategy to match the intended clinical setting in which it is to be applied to (see Table 1).

Designing clinical cardioprotective studies of STEMI patients undergoing PCI

The patient

Site of the infarct A recent analysis has suggested that only 25% of all STEMI patients will have infarcts large enough to realize benefit from any adjunctive therapy applied at the time of PCI [38]. Thus, the PCI patients who are most likely to benefit immediately from a cardioprotective strategy in the short term are those patients in whom the therapy is applied as an adjunct to reperfusion therapy and in those which have the larger areas at risk. In general these are patients who present with proximal LAD coronary artery territory infarcts. In contrast, patients presenting with the smaller areas at risk, such as those in the right coronary artery (RCA) and circumflex (Cx) coronary artery territories do not appear to accrue as much benefit from cardioprotective reperfusion therapy [8, 43]. Selecting patients with the smaller infarcts may therefore actually dilute any positive effect elicited by the novel cardioprotective strategy. As such, clinical studies should focus on selecting patients presenting with a proximal LAD infarct only [42]. The disadvantage of this approach is that it may impact negatively on recruitment rates as this population only makes up 30% of all presenting STEMI patients. It is therefore clear that the AAR should be measured when assessing any novel cardioprotective strategy in the clinical setting.

TIMI flow prior to PCI For cardioprotective strategies applied at the time of PCI, it is essential that the intervention is administered prior to myocardial reperfusion. Therefore, patients in which the infarct-related coronary artery has spontaneously recanalized and TIMI coronary blood flow is TIMI > 1, have already undergone myocardial reperfusion and are unlikely to benefit from the novel cardioprotective strategy and should be excluded. However, this may not be so straightforward, given that some patients can have intermittent reperfusion prior to the catheter laboratory, but have TIMI 0/1 flow prior to PCI. Recent studies suggest that the patients most likely to benefit from a novel cardioprotective strategy applied at the time of PCI are those presenting with TIMI 0/1 flow in the culprit coronary artery [8].

Coronary collaterals The presence of coronary collateralisation to the area at risk will reduce MI size in STEMI patients undergoing PCI, thereby confounding any potential beneficial effect with the novel cardioprotective strategy. Patients with visible coronary collaterals (Rentrop grade ≥ 1) [46] on coronary angiography should be excluded from the study. Therefore, it is essential to ensure that the image acquisitions taken during coronary angiography are long enough to visualise coronary collaterals.

Duration of chest pain It has been established that STEMI patients accrue mortality benefit from myocardial reperfusion strategies such as thrombolysis and PCI providing they present within 12 h of the onset of chest pain [17]. Therefore, cardioprotection studies have tended to focus on those patients presenting within 12 h of chest pain onset. The relationship between the extent of lethal myocardial reperfusion injury and the duration of index ischemia has not been fully elucidated. Whether patients with short ischemic times (<3 h chest pain to PCI time) or patients with prolonged ischemic times (6–12 h chest pain to PCI time) accrue more benefit from an adjunctive reperfusion treatment strategy is unclear. Indeed whether patients who present more than 12 h following the onset of chest pain, benefit from adjunctive therapy to PCI is unknown.

Measurement of the area at risk The assessment of the efficacy of a novel cardioprotective strategy in PCI patients requires the measurement of both the MI size and the area at risk (AAR). This allows a comparison between STEMI patients with different AAR. Historically, the AAR has been measured in clinical studies using coronary angiography jeopardy scores but the traditional gold-standard technique has been to use nuclear myocardial scanning [8]. The AAR is delineated as the areas of absent radioisotope tracer on a nuclear myocardial scan performed within 6 h of the injection. Quantification of the AAR by nuclear myocardial scanning can take into account coronary collateralisation because the radioisotope is able to distribute through the collaterals and reduce the AAR, and hence myocardial salvage can be determined correctly even in the presence of coronary collaterals. However, this technique is time-consuming, impractical when offered on a 24-h daily basis, lacks resolution and involves significant radiation to the patient.

An alternative approach to measure the AAR is to assess the hypokinetic segments of the left ventricle using venticulography at the time of PCI [57, 60], although this technique may overestimate the AAR by including areas of myocardial stunning. Cardiac MRI (CMR) may be a promising imaging technique for measuring the AAR. Animal studies have reported that enhanced signal intensity on retrospective T2-weighted CMR from increased myocardial oedema [10] correlates with the AAR in reperfused myocardial infarcts [1, 27]. Preliminary clinical studies suggest that the enhanced T2 signal intensity on CMR scans performed in the first week following PCI correlates with the AAR as measured by the BARI coronary angiography jeopardy score [65] and nuclear myocardial scans [11]. In addition, myocardial salvage as measured by CMR has been linked to clinical outcomes in PCI-treated STEMI patients [14]. The concern with CMR is whether the novel cardioprotective strategy itself may influence the extent of myocardial oedema by reducing it, thereby resulting in an underestimate of the AAR. In this case, CMR may be ineffective as a technique for measuring AAR post PCI. Whichever technique is used to measure the AAR it is essential to use the AAR as a co-variate for analysing MI size reduction, given that patients presenting with the larger areas at risk are those most likely to benefit from the novel cardioprotective strategy.

The intervention

Timing of the cardioprotective strategy The timing of the cardioprotective strategy is critical in the design of clinical studies. The refusal to take heed of preclinical data has been the root cause for some major failures in the clinical field of cardioprotection. Some strategies like sodium–hydrogen exchange inhibitors or therapeutic cooling target myocardial ischemia, while other strategies such as ischemic and pharmacologic postconditioning and remote ischemic conditioning target myocardial reperfusion injury. A cardioprotective strategy which targets myocardial ischemia may still be effective if administered soon after ischemia has started. Such an intervention could be applied by the paramedics in the ambulance, particularly where long transit times to the PCI centre are anticipated and should be encouraged [8]. The ambulance is also the ideal setting for remote ischemic conditioning but often the transit time is too short to complete the treatment protocol [8]. Whatever the case, it is essential that any cardioprotective strategy be applied prior to the opening of the infarct-related coronary artery, given the crucial events which occur in the first few minutes of myocardial reperfusion [production of oxidative stress, calcium overload and mitochondrial permeability transition pore (mPTP) opening] [42, 66]. Clearly, it is imperative that the application of the novel cardioprotective strategy does not in any way delay the onset of myocardial reperfusion.

Delivery of cardioprotective strategy A number of pharmacologic postconditioning agents have been identified (e.g. cyclosporine-A). Any such drug should be capable of being administered as a single intravenous or intracoronary bolus so that therapeutic levels will be achieved in a matter of minutes.

Clinical cardioprotective studies of patients undergoing CABG surgery

Despite excellent cardioprotection using cold blood cardioplegia, significant peri-operative myocardial injury still occurs in patients undergoing CABG surgery ± valve surgery. The cause of this myocardial injury is multi-factorial being attributed to global ischemia–reperfusion injury, coronary embolization and prolonged aortic cross-clamp time. The extent of peri-operative myocardial injury can be assessed by measuring serum cardiac enzymes such as CK-MB, troponin-T and troponin-I, the elevation of which has been associated with worse clinical outcomes post-surgery [9, 13, 29]. Surgeons are therefore continually seeking ways to minimise IRI, particularly as more high-risk patients are being operated upon and it is becoming increasingly clear that even mild to moderate elevations in CK-MB and troponin are associated with increased intermediate and long-term mortality.

Similar to the setting of a STEMI in which patients with a larger MI are most likely to benefit from a cardioprotective intervention, the same may apply to patients undergoing CABG surgery. Therefore, the patients most likely to benefit from a cardioprotective strategy during CABG surgery are those who are most at risk of sustaining significant peri-operative myocardial injury. This group includes patients undergoing 3-vessel CABG surgery with or without valve surgery, redo CABG surgery patients, patients with significant LVH or LV systolic dysfunction, patients with an additive Euroscore of ≥6 and diabetic patients. We believe that it is this group of higher-risk patients who should be selected for studies of novel cardioprotective strategies as they are more likely to experience a greater degree of myocardial IRI from the prolonged cross-clamp and cardio-pulmonary bypass times.

A variety of cardioprotective strategies have been tried in the CABG surgery setting in the past. Although the majority of these were unsuccessful, one of the most potentially promising treatment strategies was cariporide, but unfortunately it had off-target cerebral side effects [37]. In the setting of CABG surgery, the novel cardioprotective strategy can be applied either prior to myocardial ischemia (cross-clamping of the aorta), during myocardial ischemia in the cardioplegia solution or at the time of myocardial reperfusion (unclamping the aorta). As noted previously, the preclinical testing of the novel cardioprotective strategy in an animal IRI model which closely resembles the CABG setting should have been previously utilised to verify efficacy (see Table 1).

Which novel cardioprotective strategy should be pursued in the clinical setting?

There are a number of novel cardioprotective strategies which have shown promise in initial proof-of-concept clinical studies. The question is which of these should be taken forward into phase 2/3 clinical studies. Following discussion in the Workshop, it was agreed that the two most promising novel cardioprotective strategies were remote ischemic conditioning (RIC) and cyclosporine-A (CsA).

For RIC, in which cycles of brief ischemia and reperfusion applied to the upper or lower limb protect the myocardium from lethal IRI, there exist extensive preclinical data in a range of animal models including in vivo murine, rat, rabbit and porcine models of MI (reviewed in [20, 56]). RIC is a non-invasive virtually cost-free cardioprotective strategy which has been shown to be effective when applied both prior to or during the index myocardial ischemia [47] as well as at the onset of myocardial reperfusion [2], lending itself to the clinical settings of CABG surgery [18], planned PCI [25], and STEMI patients receiving PCI [8], settings in which initial proof-of-concept studies have already been successfully performed.

CsA has the advantage of targeting an end-effector of IRI, as opposed to the ever-expanding list of cardioprotective agents which tend to target G-protein coupled receptors and intracellular kinases and other mediators, which may be down-regulated or ineffective in the presence of co-morbidities. Because the main site of action of CsA is probably the mPTP, a purported mitochondrial channel which mediates cardiomyocyte death at the onset of myocardial reperfusion, most of the preclinical data in support of its role as a cardioprotective agent have been as an adjunct to myocardial reperfusion [19, 22]. Similarly, there exist extensive preclinical data in a range of animal models including in vitro and in vivo models of MI, as well as in a rabbit model of post-cardiac arrest [12]. However, the in vivo porcine MI model has produced mixed results with CsA for an unclear reason [28, 35, 54]. CsA has also been demonstrated to be effective in human atrial tissue models of simulated IRI [49]. Again, a preliminary proof-of-concept clinical study has demonstrated that a single intravenous bolus of CsA given prior to PCI can limit MI size in STEMI patients [43].

Other potentially novel cardioprotective strategies for which there exist both preclinical data and initial proof-of-concept clinical studies are glucagon-like peptide 1 [41], PKC-δ inhibition [3], atrial natriuretic peptide [32], and ischemic postconditioning [60].

Clearly, for all these novel cardioprotective strategies preclinical studies are required to determine whether cardioprotection is maintained in the presence of certain confounding factors such as age, sex, diabetes, the metabolic syndrome, hyperlipidemia, hypertension and so forth. In this regard, a recent preclinical study suggests that CsA-mediated cardioprotection at the time of myocardial reperfusion was ineffective in Zucker obese rats [26]. This intriguing finding requires confirmation and the mechanism underlying this observation needs further investigation.

In summary, for both CsA and RIC large multi-centred randomised controlled clinical trials are required to confirm their cardioprotective benefit in the clinical setting and investigate whether these interventions impact on clinical outcomes for patient benefit.

Summary

In order to overcome the obstacles to translation of novel cardioprotective strategies discovered in laboratories into the clinical setting for patient benefit, a set of recommendations may facilitate this process. These are outlined below and include preclinical cardioprotection studies and clinical cardioprotection studies in PCI patients. Clearly, similar recommendations may be introduced for other clinical settings of IRI such as CABG surgery, sudden cardiac arrest, cardiac transplantation, as more evidence becomes available from ongoing studies.

Recommendations for investigating novel cardioprotective strategies in preclinical studies

In vitro models A novel cardioprotective strategy should be first investigated using established small animal (murine, rat, rabbit) in vitro models of IRI, including human myocardial tissue models of simulated IRI [51, 64]. The actual animal IRI model used should closely resemble the clinical setting in which it intended to test the novel cardioprotective strategy (see Table 1).

In vivo models The novel cardioprotective strategy should then be examined first in small (murine, rat, rabbit) and then in large animal (pig, dog, primate) in vivo studies of IRI. Again, the actual animal IRI model used should closely resemble the clinical setting in which it is intended to test the novel cardioprotective strategy (see Table 1). For animal MI models, reperfusion must be of sufficient duration such that MI size has fully developed; on the other hand, infarct size must be determined before remodelling occurs.

Confounders Ideally, the cardioprotective strategy should be demonstrated to still be effective in the presence of one or more confounders and/or medical therapies.

Recommendations for investigating novel cardioprotective strategies in PCI studies

Proximal LAD infarcts STEMI patients presenting with a proximal LAD coronary artery infarct are preferred as these patients are those most likely to benefit from the novel cardioprotective strategy.

TIMI 0 or 1 prior to PCI Only STEMI patients presenting with TIMI 0/1 coronary flow on coronary angiography prior to PCI should be included in studies investigating novel cardioprotective strategies.

No collaterals on coronary angiography Only STEMI patients with no visible coronary collaterals on coronary angiography should be included in studies investigating novel cardioprotective strategies. The exception to this is if the AAR is being measured using nuclear myocardial scanning which takes into account coronary collateralisation.

Chest pain <12 h Only STEMI patients presenting within 12 h of the onset of chest pain should be included in studies investigating novel cardioprotective strategies.

Cardioprotective strategy applied prior to reperfusion The intervention must be applied prior to the opening of the culprit coronary artery.

Measurement of the area at risk When assessing the efficacy of a novel cardioprotective strategy it is imperative to measure the area at risk.

Assessment of the cardioprotective strategy In proof-of-concept clinical studies it may be sufficient to measure serum cardiac enzymes (CK-MB and troponin-T or I). However, more robust surrogate clinical endpoints include LV ejection fraction (measured by echocardiography, myocardial nuclear scanning or cardiac MRI) and MI size (measured by myocardial nuclear scanning or cardiac MRI). Ultimately, clinical outcome measures such as cardiovascular mortality and hospitalization for heart failure must be determined.

Recommendations for investigating novel cardioprotective strategies in CABG studies

Patient selection The investigation of a novel cardioprotective strategy should focus on those patients more likely to sustain significant peri-operative myocardial injury during surgery. These include patients with left ventricular hypertrophy (LVH) or impaired left ventricular (LV) systolic function and patients undergoing 3-vessel CABG ± valve surgery or redo CABG surgery and diabetic patients.

Anaesthetic choice The investigation of the novel cardioprotective strategy should not influence the choice of anaesthetic regimen, even if it comprises inhaled anaesthetic agents such as isoflurane, which have been reported in both preclinical and clinical studies to confer cardioprotection during CABG surgery [58]. It is important that any novel cardioprotective strategy is shown to be effective in the presence of routine medical therapy.

Assessment of the cardioprotective strategy In proof-of-concept clinical studies, peri-operative release of cardiac enzymes such as CK-MB, troponin-T or I can be used to assess the efficacy of the novel cardioprotective strategy [18, 61]. Other clinical endpoints include: inotrope score, length of ITU and hospital stay, LV ejection fraction, acute kidney injury, cognitive function, cardiovascular mortality, and hospitalization for heart failure.