Article Text


Targeting reperfusion injury in the era of primary percutaneous coronary intervention: hope or hype?
  1. Jacob Thomsen Lønborg
  1. Correspondence to Dr Jacob Thomsen Lønborg, Department of Cardiology, Rigshospitalet, University of Copenhagen, Blegdamsvej 9, Copenhagen 2100, Denmark; Jacoblonborg{at}


Introduction of reperfusion therapy by primary percutaneous coronary intervention (PCI) has resulted in improved outcomes for patients presenting with ST-segment elevation myocardial infarction. Despite the obvious advantages of primary PCI, acute restoration of blood flow paradoxically also jeopardises the myocardium in the first minutes of reperfusion—a phenomenon known as reperfusion injury. Prevention of reperfusion injury may help to improve outcome following primary PCI. This review focuses on the clinical evidence of potential therapeutic cardioprotective methods as adjuvant to primary PCI. Despite overall disappointing, there exists some promising strategies, including ischaemic postconditioning, remote ischaemic conditioning, pharmacological conditioning with focus on adenosine, cyclosporine A, glucose–insulin–potassium, exenatide, atrial natriuretic peptide and metoprolol and cooling. But hitherto no large randomised study has demonstrated any effect on outcome, and ongoing studies that address this issue are underway. Moreover, this review will discuss important clinical predictors associated with reperfusion injury during primary PCI that may interfere with a potential protective effect (pre-PCI thrombolysis in myocardial infarction flow, preinfarction angina, collateral flow, duration of ischaemia and hyperglycaemia). This paper will also provide a short overview of the technical issues related to surrogate endpoints in phase II trials. Based upon these discussions, the paper will provide factors that should be taken into account when designing future clinical studies.

Statistics from


During the last decades, dramatic changes in the management of patients with ST-segment elevation myocardial infarction (STEMI) have evolved, resulting in improved outcomes. The major advance has been the introduction of reperfusion therapy by fibrinolysis and subsequently primary percutaneous coronary intervention (PCI), which today is the recommended treatment. However, the mortality rate following a STEMI has reached a plateau with 1-year mortality of 10%,1 and a substantial amount of patients develop clinical heart failure. It is a fundamental dogma that the benefit of reperfusion therapy is exerted through increased myocardial salvage and thereby reduction in infarct size. Thus, there is a need for further improvement in the treatment of patients with STEMI to improve myocardial salvage and drive the event rates down.

Despite the obvious advantages of reperfusion by primary PCI, acute restoration of blood flow paradoxically also jeopardises the myocardium in the first minutes of reperfusion. This phenomenon is known as reperfusion injury, which encompasses several distinct pathophysiological components including reversible impaired myocardial contractility (stunning), arrhythmias, no-reflow and death of cardiomyocytes (lethal reperfusion injury).2 Experimental data implicate a number of factors contributing to lethal reperfusion injury independently of no-reflow such as opening of the mitochondrial permeability transition pore (mPTP), rapid normalisation of pH, intracellular calcium overload and generation of reactive oxygen species (ROS).2 The lethal reperfusion injury may account for 50% of the final myocardial damage following an acute myocardial infarction,2 and prevention of reperfusion injury is considered pivotal for improving outcomes in patients with STEMI.3 ,4 This review evaluates potential cardioprotective treatments and determinants of reperfusion injury during primary PCI.

Protecting the heart during reperfusion by primary PCI

Ischaemic preconditioning is the most potent protection against reperfusion injury of all investigated methods. However, the clinical application in patients with STEMI is limited since it has to be applied before to the ischaemic period. Possible potentially applicable strategies for cardioprotection are illustrated in figure 1. The exact mechanisms of action are unclear, but several intracellular signalling pathways are described that all have inhibition of the mPTP opening as effector (figure 2).

Figure 1

Concepts of cardioprotection during reperfusion by primary percutaneous coronary intervention in patients with ST-segment elevation myocardial infarction.

Figure 2

Cardioprotective signalling pathways. CAMP, cyclic adenosine monophosphate; mPTP, mitochondria permeability transition pore; RISK, reperfusion injury salvage kinase; TNF, tumour necrosis factor.

Ischaemic postconditioning

In dogs, Zhao et al5 discovered that ischaemic postconditioning resulted in 44% smaller infarct size. The cardioprotective effects of ischaemic postconditioning have been evaluated in several clinical studies (table 1 and see online supplementary table S1 for complete overview). Initial reports of ischaemic postconditioning were promising with marked reduction in infarct size (36%) and improvement in LVEF.6 The first study to use cardiovascular magnetic resonance (CMR) as endpoint found a smaller reduction in infarct size adjusted for area at risk, but no effect on peak troponin T or LVEF,7 which indicated that ischaemic postconditioning might not be as effective as shown in initial smaller studies. More recently, several studies have found that ischaemic postconditioning protects the heart,8 ,9 whereas others found no effect.10 ,11 A meta-analysis concludes that ischaemic postconditioning confers cardioprotection assessed by cardiac enzymes.12 In the largest study hitherto, Hahn et al11 found no effect of ischaemic postconditioning, but the study was not powered to draw firm conclusion in terms of hard endpoints and infarct size was not assessed. In one study, ischaemic postconditioning tended to increase the infarct size, indicating that this manoeuvre in arteriosclerotic vessels might not be without side effects.10

Table 1

Ischaemic postconditioning: overview of the most important clinical studies on ischaemic postconditioning in patients with STEMI undergoing primary PCI

The discrepancy between existing studies might be explained by important differences between inclusion and exclusion criteria and differences between postconditioning algorithms. The most frequently used algorithm is four cycles of 1 min inflation and 1 min deflation. There is evidence suggesting that the tissue pH is normalised as early as 40 s after reperfusion, resulting in loss of cardioprotection by applying ischaemic postconditioning after this time point.13 In contrast, reoxygenation of the tissue stimulates production of ROS, which is important for protection of the mitochondria.13 Thus, an algorithm consisting of 30 s would secure that the pH does not increase abruptly and would allow enough time for proper production of ROS.13 Accordingly, most of the clinical studies using algorithms based on a 30 s cycle show cardioprotection to some extent.

It seems that the effect of ischaemic postconditioning in the newer studies is diminished, which may reflect a general improvement in the treatment of patients with STEMI blunting the effect of postconditioning. Moreover, the effect of ischaemic postconditioning is hampered by thrombectomy, which has become adopted widely since the initial clinical studies on ischaemic postconditioning. Thrombectomy might result in a less abrupt reperfusion, which per se is cardioprotective,14 and the time used for thrombectomy delays ischaemic postconditioning leading to loss of cardioprotection.13 Increasing use and greater efficacy of antiplatelet drugs, ageing and comorbidities also blunt the effect of ischaemic postconditioning.2 Thus, we do not have a clear unambiguous answer as to whether or not this simple and cheap manoeuvre should be recommended as routine during primary PCI in patients with STEMI, or whether it should be considered of historical interest. We must await the results from the DANish Study of Optimal Acute Treatment of Patients With ST-elevation Myocardial Infarction (DANAMI-3) trial (NTC01435408), expected in 2016.

Remote ischaemic perconditioning

Przyklenk et al discovered that ischaemic preconditioning in a remote vascular bed protected the heart during ischaemia.15 It was later demonstrated that limb ischaemia reduced myocardial infarct size in animals and prevented endothelial dysfunction.16 These findings led to the concept of remote ischaemic conditioning. It is still unknown how the signal is transferred from the remote organ to the heart, but possible mechanisms include humoral, neuronal and cellular factors. Bøtker et al found a larger myocardial salvage following primary PCI pretreated with remote conditioning (table 2),17 which translated into improved clinical outcome.18 However, there is a high risk of a type I error due to small sample size, and the effect was mainly driven by a lower non-cardiac mortality rate. Recently, two studies confirmed the cardioprotective effect of remote ischaemic conditioning (table 2).19 ,20 There seems to be no additive effect of remote ischaemic conditioning and ischaemic postconditioning.21 Thus, results from the currently enrolling large multicentre trials Effect of RIC on Clinical Outcomes in STEMI Patients Undergoing pPCI (CONDI II) (NCT01857414) and Effect of Remote Ischaemic Conditioning on Clinical Outcomes in STEMI Patients Undergoing PPCI (ERIC-PPCI) (NCT02342522) are awaited with great expectation.

Table 2

Remote ischaemic conditioning: overview of clinical studies on remote ischaemic conditioning in patients with STEMI undergoing primary PCI

Pharmacological postconditioning

An alternative to ischaemic conditioning is pharmacological conditioning, which may be very effective if the right target is addressed.22 Also, these agents may be administered in the prehospital phase and may potentially protect the approximately 40% of patients with STEMI that achieve reperfusion before PCI.23 On the other hand, since the artery is occluded, the drugs reach the target myocardium during reperfusion, and whether this is sufficient to prevent reperfusion injury may be questioned. Several drugs with different biological effects/targets have been tested in clinical settings, most with disappointing results (see online supplementary table S1), but a few have been found protective (table 3).22 ,24–33 Most of these studies are small proof-of-concept studies and not powered to evaluate hard endpoints. Moreover, in some studies the effect was only observed on secondary endpoints or in post hoc analysis. The effect of adenosine was neutral in the overall population, but protective in the subgroup of patients with short duration of symptoms.25 ,30 Similar, the anti-inflammatory compound FX06 treatment did only reduce infarct size in patients with short duration of symptoms.26

Table 3

Pharmacological conditioning: overview of clinical studies demonstrating an protective effect of pharmacological conditioning in patients with STEMI undergoing primary PCI

Cyclosporine A is believed to mediate its effect through inhibition of the mPTP opening, a major determinant of cell death following ischaemia reperfusion. MPTP is a non-selective pore whose opening results in equilibrium between the mitochondria matrix and cytosol, leading to swelling of mitochondria, depletion of ATP and cell death due to necrosis. Treatment with cyclosporine A in addition to primary PCI resulted in a marked reduction in infarct size.22 This study was the first to target a downstream target in comparison to the majority of studies that target specific upstream targets, which also may explain the overall neutral results in previous studies. The study only included 58 patients,22 but the results from the large multicentre CIRCUS trial (NCT015202774) are expected soon.

Prehospital-initiated glucose–insulin–potassium (GIK) treatment in patients with suspected acute coronary syndrome was associated with a marked reduction in infarct size in the subgroup of patients with STEMI,32 but these findings must be confirmed in prospective settings. In contrast, insulin treatment without glucose–potassium in patients with STEMI with hyperglycaemia tended to increase the infarct size.34 The cardioprotective effect of GIK is controversial; some authors state that it involves increased glucose uptake, whereas others report a glucose-independent signal mechanism.

Glucagon-like peptide-1 (GLP-1) is an incretin hormone that regulates plasma glucose, and GLP-1 analogues (eg, exenatide) are used for treating diabetes. Moreover, GLP-1 and its analogues have cardioprotective actions mediated through receptor-dependent and receptor-independent pathways, insulinotropic, as well as cardiac and non-cardiac insulinomimetic effects including an increased glucose uptake.35 Interestingly, the cardioprotective effect of exenatide seems to be independent of the presence of hyperglycaemia.36 Moreover, the effect was restricted to patients treated within 132 min after first medical contact.37

β-Blockers have been used for decades in patients with acute myocardial infarction, but the potential cardioprotective effects in patients with STEMI remained undiscovered until recently.27 However, metoprolol cannot be administered in high-risk patients presenting with Killip class 3 or 4.

In spite of overall disappointing results with clinical pharmacological conditioning, there are some promising agents, but large, multicentre studies are warranted. Moreover, new potential treatments are scrutinised in phase II trials (danegaptide (NCT01977755); mineralocorticoid (NCT01882179) and melatonin (NCT01172171), mangafodipir (NTC00966563); nitric oxide (NTC01398384), sevofluran (NTC00971697)). Since ischaemic and pharmacological conditioning may have different effects and mechanism of action, another important unsettled issue is potential additive effect (eg, ischaemic remote conditioning combined with exenatide or cyclosporine A). Also, the optimal dosage and administration route need to be determined.


Cooling to 32–33° before reperfusion reduced infarct size in animal models.38 The initial report on rapid cooling in patients with STEMI was promising, showing an impressive cardioprotective effect,39 an effect that was dependent on achieving a target temperature of <35° prior to reperfusion. Nevertheless, despite achieving <35° in 76% of the patients rapid cooling did not show any effect in a large well-designed multicentre study.40 It may be argued that 35° is insufficient, and if this method should regain interest it is pivotal to develop a method for achieving 32–33° before reperfusion.

Determinants of reperfusion injury in patients with STEMI

Several factors influence the effect of cardioprotection: (1) in experimental studies, ischaemia is induced by ligation of a vessel and not by plaque rupture and thrombus formation; (2) comorbidities;2 (3) thrombectomy and more effective antiplatelet drugs; (4) general improvement for patients with STEMI; and (5) preinfarction angina, pre-PCI thrombolysis in myocardial infarction (TIMI) flow, collateral flow, duration of ischaemia and hyperglycaemia are suggested to be determinants of reperfusion injury. The impact of this last set of factors on reperfusion injury and cardioprotection is discussed in the following paragraphs, and recommendations for future clinical studies are shown in table 4.

Table 4

Factors to be taken into account in future studies on prevention of reperfusion injury in patients with STEMI undergoing primary PCI

Preinfarction angina

Angina preceding an acute myocardial infarction may resemble the mechanism of ischaemic preconditioning and lead to cardioprotection. Observational clinical studies support this notion since preinfarction angina improved clinical outcomes,41 probably through increasing myocardial salvage index and reducing infarct size.23 The cardioprotective mechanism of preinfarction angina may involve ischaemic preconditioning, recruitment of existing collaterals, increasing use of medication and accelerated thrombolysis. Ischaemic preconditioning includes an early protective window (1–2 h before infarction) and a late window (24–48 h before infarction). Conversely, existing data suggest that preinfarction angina at any time (<90 days) is associated with cardioprotection, maybe because these patients experience symptomatic and silent ischaemic episodes on a regular basis. Whether preinfarction angina interferes with protective treatments is uncertain.

Pre-PCI TIMI flow

Poor preprocedural flow regarded as TIMI flow 0/1 vs 2 or vs 3 in the infarct-related artery is present in approximately 60% of patients with STEMI treated with primary PCI and is independently associated with adverse outcomes, smaller myocardial salvage and larger myocardial infarct size in a stepwise manner.23 ,42 The time window for cardioprotection is restricted to the very first minutes of reperfusion, and conditioning only supposed to work in TIMI 0/1, because it will be applied at the time of reperfusion. This is confirmed in clinical studies since the effects of several cardioprotective strategies are lost in patients with pre-TIMI flow 2/3.17 ,23 ,27 ,28 ,43

Collateral flow

Experimental data show that collateral flow to the ischaemic myocardium is a major determinant of myocardial infarct size and salvage. Approximately one-fifth of patients without coronary artery disease and one-third of patients with coronary artery disease have well-developed collateral flow sufficient to prevent ischaemia.44 It is therefore believed that collateral flow to the infarct-related artery during a STEMI prevents reperfusion injury. This effect may be more pronounced in humans since the collateral flow constitutes a larger proportion of the antegrade flow in comparison to most other species.45 Collateral flow in patients with STEMI is traditionally assessed using the angiographic Rentrop score. However, the cardioprotective consequence of angiographic visible collateral in patients with STEMI seems to be smaller than expected since it is not related to either myocardial salvage or mortality.23 ,46 The discrepancy between experimental and clinical studies may be explained by the following: (1) collateral flow during coronary occlusion is recruited by angiographic invisible small vascular overlay or angiographic visible larger arteries.45 Thus, angiography does not quantitate the collateral flow or take invisible collaterals into account. (2) The importance of collaterals may be blunted by differences in baseline characteristics. (3) Patients with collateral flow sufficient to prevent ischaemia may not present with a STEMI, but with a non-STEMI or chronic total occlusion. Altogether, the presence of angiographic visible collaterals in patients with STEMI should not influence the revascularisation and cardioprotective strategy.

Duration of ischaemia

It seems that longer time of ischaemia results in loss of cardioprotection,25 ,26 ,30 ,37 and therefore, focus on the early presenters in clinical trials is important. This may be because interventions generally are most effective in the first two to three hours after onset of ischaemia, patients with long duration of ischaemia demonstrate large areas of irreversible myocardial damage and subsequently fewer cardiomyocytes are exposed to reperfusion injury with a potential of reversibility, and longer duration of ischaemia leads to changes in the mitochondria, rendering them more resistant to cardioprotection.

Duration of ischaemia in patients with STEMI is traditionally assessed as time from symptom onset until wire/balloon (treatment delay). However, observational clinical studies demonstrate a weak nearly horizontal correlation between treatment delay and mortality.1 Also, the influence of treatment delay on infarct size and myocardial salvage is controversial.47 There are several explanations for this discrepancy. First, in observational studies late presenters typically are characterised as low-risk patients that have survived the initial critical phase and benefit less from reperfusion and early presenters as high-risk patients. Second, treatment delay is hampered by patient memory bias. Third, symptom onset may represent preinfarction angina. Fourth, the proportion of the infarct size caused by reperfusion injury is relatively smaller in patients with longer time of ischaemia due to less viable myocardium. Alternatively duration of ischaemia can be assessed as time from first medical contact to balloon (system delay), which is associated with mortality, infarct size and myocardial salvage.1 ,47

Hyperglycaemia upon admission

Hyperglycaemia upon admission is observed frequently in patients with STEMI and may be unfavourable during reperfusion and linked to the subsequent injury.2 Previous studies have demonstrated larger infarct size and poorer prognosis in patients with hyperglycaemia upon admission compared with patients without hyperglyacemia.36 ,48 However, hyperglycaemia is also related to larger area at risk, and hence not to a smaller myocardial salvage index or infarct size adjusted for area at risk or other predictors.36 Similarly, hyperglycaemia upon admission fails to predict prognosis adjusted for other important predictors.48 These findings imply that the larger myocardial damage and adverse prognosis reported in patients with hyperglycaemia are the result of larger area at risk and not smaller salvage. Thus, it seems plausible that hyperglycaemia is a high-risk indicator and an innocent bystander rather than being a detrimental factor per se.

Assessment of reperfusion injury during primary PCI

Cell death owing to lethal reperfusion injury cannot be distinguished from that induced by the preceding ischaemia-induced cell death, but since it leads to death of cardiomyocytes a decrease in myocardial salvage and a larger infarct size are expected. Thus, assessment of infarct size and myocardial salvage is fundamental. Moreover, adjusting infarct size for area at risk increases the statistical power and assessment of area at risk allows for inclusion of anterior as well as non-anterior infarctions. Modalities such as CMR and single-proton emission CT (SPECT) are used to assess infarct size and area at risk, but may lead to high dropout rates and selection bias.

CMR has a higher spatial resolution than SPECT and is thus considered the method of choice to assess infarct size in clinical settings. However, the timing of acquisition is important since the infarct size changes rapidly within the first week following the infarction and slowly until 30 days, but thereafter remains almost constant. The infarct size measured by CMR in the acute and chronic phase following primary PCI is related to outcome,3 ,4 and thus serves excellently as surrogate endpoints.

In terms of area at risk, SPECT allows for assessment of area at risk before reperfusion, but this method is problematic in patients with TIMI 2 or 3 flow and difficult owing to logistical reason. CMR can also be used to assess area at risk by visualising myocardial oedema (T2-weighted imaging), increased extracellular space (T1-weighted images) or the extent of the infarct surface area. However, CMR visualises area at risk postreperfusion and treatments commenced prior to CMR may per se lead to a falsely reduced area at risk. This has been observed following ischaemic postconditioning and remote ischaemic conditioning9 ,20 but not exenatide.24 ,31 It is therefore of utmost importance that the area at risk is not affected by the treatment, and if so, another measurement should be introduced.20

Another omen of reperfusion injury during primary PCI is microvascular damage and impaired microvascular perfusion. There are several surrogate markers suggested to be related to microvascular dysfunction including angiography,49 ST-segment resolution, increased ST-segment elevation (>1 mm) during reperfusion (ST peak)50 and microvascular obstruction on CMR image.4


STEMI remains a major cause of mortality and morbidity and further infarct-sparing treatments are needed. Conditioning prevents reperfusion injury in experimental studies, but most strategies fail in clinical studies, highlighting the importance of focus on study design. Nevertheless, some treatments have demonstrated cardioprotection, but hitherto no large randomised study has demonstrated any effect on outcome, but ongoing studies that address this issue are underway.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Funding Rigshospitalets Research Foundation and Danish Heart Foundation.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.