Statistics from Altmetric.com
- percutaneous coronary intervention
- acute myocardial infarction
- heart failure with reduced ejection fraction
Revascularisation strategies in cardiogenic shock.
Mechanical circulatory support in cardiogenic shock.
Evidence base to guide current best practice.
Cardiogenic shock (CGS) complicates 5 – 10% of cases of acute myocardial infarction (AMI) and the most common cause of CGS is AMI (80% of cases).1 When CGS occurs after ST-segment elevation myocardial infarction (STEMI), the median time delay before the onset of shock is 5–6 hours.2 CGS complicates non-ST-segment elevation myocardial infarction (NSTEMI) less frequently (approximately 2.5% of NSTEMI cases) and tends to occur later (median time delay of 76 hours).3 Fewer than 50% of patients with CGS survive up to 1 year.4
The clinical syndrome of CGS is present when there is inadequate cardiac output and systemic tissue hypoperfusion despite adequate circulating volume and filling pressure. The haemodynamic criteria used to define CGS are a systolic blood pressure (SBP) <90 mmHg for at least 30 min, or a need for vasopressors to achieve SBP ≥90 mmHg, or a fall in mean arterial blood pressure >30 mmHg below baseline, combined with an unsupported cardiac index <1.8 L/min/m2 or <2.2 L/min/m2 with cardiac support, and a pulmonary capillary wedge pressure >15 mmHg indicating elevated left ventricular (LV) filling pressures.2 In patients with AMI, an acute reduction in LV contractile function is central to the process leading to CGS along with reduced coronary blood flow, increased myocardial oxygen demand, diastolic dysfunction and elevated LV end diastolic pressure. Neurohormonal compensatory mechanisms, including sympathetic stimulation promote tachycardia and vasoconstriction, both of which increase cardiac work that in turn increases LV preload. Systemic inflammatory responses also contribute to the pathophysiology,5 and metabolic derangements such as acidosis may further impair myocardial contractility.
Timing of revascularisation of the culprit artery
In patients with STEMI and CGS, early revascularisation of the culprit artery improves outcomes.6 7 Notably, in the landmark Should We Emergently Revascularise Occluded Coronaries for Cardiogenic Shock (SHOCK) trial,2 7 6- month mortality was lower with emergency compared with delayed revascularisation (median time from randomisation to revascularisation: 1.4 hours vs 102.8 hours; mortality: 50% vs 63%, p=0.027). Even so, mortality in those patients randomised to early revascularisation in the SHOCK trial was still high.2
Culprit only versus complete revascularisation
Multivessel disease is present in around 70% of patients with CGS.8 Studies comparing multivessel percutaneous coronary intervention (PCI) to treatment of the culprit artery only have reported variable results9–16 (table 1). The ‘Culprit Lesion Only PCI versus Multivessel PCI in Cardiogenic Shock’ (CULPRIT-SHOCK) trial17 is the largest multicentre randomised clinical trial (RCT) in CGS due to AMI. In 706 patients with CGS (62% STEMI, 38% NSTEMI), initial culprit vessel PCI and if appropriate staged multivessel PCI were compared with immediate multivessel PCI. The rationale for performing the CULPRIT-SHOCK trial was that significant lesions in non-culprit vessels are associated with increased mortality in AMI.8 Therefore, a logical consideration is whether treatment of these lesions in the acute phase could improve mortality by improving myocardial perfusion and ventricular function. Adverse event rates in CGS tend to occur early and then plateau suggesting that effective interventions applied early are most likely to reduce mortality (figure 1). Exclusion criteria in CULPRIT-SHOCK included ischaemic time >12 hours, cardiopulmonary resuscitation >30 min and age >90 years. In the immediate complete revascularisation group, the protocol required that revascularisation be attempted in all vessels suitable for PCI with >70% diameter stenosis including chronic total occlusions (CTOs). In the culprit-only group, staged revascularisation was based on ischaemia demonstrated by either non-invasive testing or fractional flow reserve.
Complete revascularisation was achieved in 81% of the immediate multivessel PCI group. Staged revascularisation of non-culprit lesions was performed in 17% of the culprit lesion PCI group. The use of extracorporeal membrane oxygenation (ECMO) was numerically higher in the immediate multivessel PCI group (28% vs 18%), and although this difference was not statistically significant, it may have had an influence on the trial outcome.17 The primary composite endpoint of 30-day mortality or severe renal failure requiring renal replacement therapy was lower in the culprit lesion only group versus the immediate multivessel PCI group (45.9% vs 55.4%, relative risk [RR]: 0.83, 95% CI 0.71 to 0.96, p=0.01).17 All-cause death was also lower in the culprit lesion only group (RR: 0.84, 95% CI 0.72 to 0.98, p=0.03)17 and although the rate of renal replacement therapy was lower in the culprit only group, this difference did not reach statistical significance (RR: 0.71, 95% CI 0.49 to 1.03, p=0.07).17 The mortality reduction was mainly due to fewer deaths from hypoxic brain injury in the culprit lesion only group (7.4% vs 14.2%).17 Potential mechanisms for the increased risk of brain injury in the immediate multivessel PCI group include the prolonged duration of multivessel PCI at a time when the patient is haemodynamically compromised and additional catheter manipulation that could increase the risk of embolisation of atherothrombotic debris to the cerebral circulation.
After 1 year of follow-up, when the culprit lesion only PCI group was compared with immediate multivessel PCI, there was no longer any significant difference in mortality (50.0% vs 56.9%, RR: 0.88, 95% CI 0.76 to 1.01; figure 1) and no difference in the rate of recurrent AMI (1.7% vs 2.1%, RR: 0.85, 95% CI 0.29 to 2.5).18 Repeat revascularisation was more frequent in the culprit only arm (32.3% vs 9.4%, RR: 3.44, 95% CI 2.39 to 4.95) as was rehospitalisation for heart failure (5.2% vs 1.2%, RR: 4.46, 95% CI 1.53 to 13.04).18 The higher rate of heart failure rehospitalisation with the culprit-only approach was a surprise finding. A potential explanation is that multivessel PCI could have resulted in improved ventricular function at long-term follow-up. The counterargument is that the patients undergoing culprit lesion-only PCI who survived with poor ventricular function were those who would have died in the acute phase if they had received multivessel PCI.
The strengths of the CULPRIT-SHOCK trial are that the crossover rate was relatively low (9% of patients in the immediate multivessel PCI group had culprit only PCI and 13% of the culprit only group had immediate multivessel PCI), and only one patient was lost to follow-up. However, the reasons for 43 patients in the culprit lesion-only PCI group crossing over to multivessel PCI included lack of haemodynamic improvement and identification of new lesions after the initial PCI.
The finding that immediate complete revascularisation in patients with CGS led to harm is contrary to the results of trials comparing PCI strategies in STEMI patients without shock and requires some further explanation. It has been suggested that in AMI, inflammatory processes promote vulnerability in non-culprit arteries and that treating such non-culprit lesions by PCI could reduce event rates and potentially improve collateral blood supply to peri-infarct ischaemic regions. However, in the setting of CGS, the intensified inflammatory and prothrombotic effects may increase the risk of periprocedural infarction during PCI of non-culprit lesions, contributing to further deterioration of LV function.
One limitation of the CULPRIT-SHOCK trial is that in contrast to previous trials involving patients with haemodynamically stable AMI, the presence of a CTO was not an exclusion criterion. At least one CTO was present in 22.4% of patients in the culprit lesion only group and 24% of patients in the multivessel PCI group. Of this latter group, PCI was attempted in approximately 50% of CTOs with success in only one-third of vessels. In patients with CGS, the presence of a non-culprit CTO confers a worse prognosis than multivessel disease without a CTO.19 However, CTO-PCI is technically challenging and requires an operator with sufficient skills. Failed attempts at CTO-PCI likely contributed to the higher volumes of contrast used in the immediate multivessel PCI group (250 cc vs 190 cc, p<0.001), and this, in turn, may have led to both acute LV volume overload with a negative impact on myocardial function and the observed increased need for renal replacement therapy. The counterargument is that exclusion of CTO would have led to selection of a lower risk cohort. Furthermore, in the predefined subgroup analysis, the results for the primary endpoint were consistent for CTO presence or absence (p value for interaction=0.26).20
Importantly, despite there being more than two decades between the SHOCK and CULPRIT-SHOCK trials, mortality in patients with culprit lesion only PCI was similar. This may be partly due to lower risk patients with single-vessel coronary disease being excluded from the CULPRIT-SHOCK trial. Nonetheless, it highlights the unmet need for evidence-based strategies to reduce mortality in this high-risk cohort.
Guideline recommendations on revascularisation strategies
Prior to CULPRIT-SHOCK, the European Society of Cardiology (ESC) STEMI guidelines recommended immediate PCI of both the culprit artery and non-culprit lesions in patients with CGS.21 In the USA, appropriate use criteria considered that immediate revascularisation of non-culprit lesions when CGS persists after revascularisation of the culprit lesion was an acceptable strategy.22 However, the findings from the CULPRIT-SHOCK trial have led to a change in the ESC guidelines, which now state that immediate multivessel PCI is not recommended in AMI with CGS as it may be harmful (class III, level of evidence B).23 In this setting, the default strategy should be PCI of the culprit artery only.23 A caveat is that the ESC guidelines suggest that it is acceptable for initial multivessel PCI to be considered in some patients. We agree and would propose that initial multivessel PCI or early staged PCI should still be considered in certain clinical situations including: (1) where the culprit artery is difficult to identify, (2) multiple potential culprit lesions, (3) angiographic signs of highly unstable lesions, (4) severe flow limiting stenosis in a non-culprit artery supplying a large territory, or (5) clear persistent ischaemia after PCI of the apparent culprit lesion.
Mechanical circulatory support (MCS)
In patients with STEMI, multivessel coronary disease and CGS, it is possible that MCS devices (table 2) (figure 2) may be of greater value than revascularising a non-culprit artery. Early short-term use of MCS aims to increase systemic flow and improve peripheral tissue perfusion, while limiting the detrimental effects of exogenous inotropes and vasopressors. In turn, this may beneficially impact on final infarct size and ultimate ventricular function.
Intra-aortic balloon pump counterpulsation (IABP)
IABP generates higher diastolic coronary perfusion pressures and reduces LV afterload during systole. IABP therapy has been shown to improve coronary flow in patients with persistent ischaemia, i.e. when the vasodilatory capacity of the coronary microcirculation cannot meet the demand of the myocardium. In this setting, coronary flow becomes proportional to perfusion pressure and is augmented by increasing aortic pressure in diastole.24 However, the physiological improvements with IABP therapy have not translated into any mortality reduction in RCTs25 or recent registries,26 27 even though some benefit was shown in early registries in the context of reperfusion therapy with thrombolysis.28 Notably, the IABP-SHOCK II (Intra-Aortic Balloon Counter-Pulsation in Acute Myocardial Infarction Complicated by Cardiogenic Shock) trial25 29 showed no reduction in mortality at 30 days, or 1 year, or long-term follow-up (6.2 years)30 in patients with AMI complicated by CGS (figure 1).
It has been suggested that the population included in the IABP-SHOCK II trial could have been a lower risk group within the spectrum of CGS. Shock ranges from mild hypoperfusion to near circulatory arrest, and the definition of GCS varies from study to study. This contributes to variability in populations and findings between studies of MCS in CGS. In the IABP-SHOCK II trial, CGS was defined by all of the following criteria: (1) SBP <90 mmHg for >30 min or vasopressors required to maintain SBP ≥90 mmHg; (2) clinical signs of pulmonary congestion; (3) impaired end-organ perfusion (altered mental state, cold, clammy skin, oliguria and raised serum lactate). Advanced haemodynamic monitoring to assess cardiac index and pulmonary capillary wedge pressure was not used to select patients for inclusion in the IABP-SHOCK II trial.
In addition, one-fifth of patients crossed over into the IABP arm and only 1 in 7 patients had IABP insertion prior to PCI (when the device might be expected to give most benefit). Thus, most patients did not have the IABP in place at the time of reperfusion, which is often the point of further haemodyamic deterioration due to reperfusion injury. Nevertheless, guideline recommendations for IABP therapy in AMI patients with CGS have been downgraded from class 1 in the previous American Heart Association (AHA)/American College of Cardiology31 and ESC32 guidelines. In the most recent ESC guidelines,23 the use of IABP in CGS complicating AMI is not recommended, apart from selected patients with mechanical complications (i.e. severe mitral regurgitation or ventricular septal defect).
One of the main reasons why IABP therapy has not been shown to improve survival is because it only provides relatively modest cardiac output support (maximum 0.5–1 L/min). As such, focus has shifted to other devices that can provide higher cardiac output,33 even though current evidence to support their use is very limited (table 2).34 The Impella range of devices (Abiomed Europe, Aachen, Germany) includes the Impella 2.5 (maximum cardiac output support 2.5 L/min), Impella CP (maximum cardiac output support 3.7–4.0 L/min) and Impella 5.0 (maximum cardiac output support 5 L/min). The Impella device is placed across the aortic valve, using large-bore femoral artery access, and provides transvalvular LV assistance expelling aspirated blood from the LV into the ascending aorta, unloading the LV, increasing cardiac output and reducing myocardial oxygen consumption. The recent IMPella versus IABP Reduces mortality in STEMI patients treated with primary PCI in Severe cardiogenic SHOCK (IMPRESS in Severe SHOCK) trial showed that 30-day and 6-month mortality was similar in the Impella (CP or 2.5) group versus IABP group (n=48).35 The findings are consistent with other studies36 and with a subsequent meta-analysis of four randomised trials, incorporating 148 patients with CGS, which reported no difference in 30-day mortality with either Impella or TandemHeart (CardiacAssist Inc, Pittsburgh, Pennsylvania, USA) compared with IABP therapy (RR: 1.01, 95% CI 0.70 to 1.44, p=0.98).34 However, bleeding was increased in patients with either Impella or TandemHeart compared with IABP. Possible explanations for these findings could be the small numbers of patients in the studies, the type of Impella device used and the timing of initiating Impella therapy relative to PCI. Further ‘real-world’ long-term outcome data will be provided by the prospective catheter-based ventricular assist device (cVAD) registry, a global study that includes unselected patients (from 2009 to ongoing) receiving Impella in routine clinical care.
The TandemHeart works by removing oxygenated blood from the left atrium and returning it through a femoral artery cannula, via an extracorporeal centrifugal pump, with retrograde perfusion of the upper aorta. Thus, both the LV and the pump provide flow to the aorta simultaneously. The TandemHeart decreases LV preload, unloads the LV and decreases myocardial oxygen demand. The maximum supported cardiac output with this device is 4 L/min. Although the TandemHeart has been shown to improve cardiac output in CGS due to AMI,37 improvements in 30-day mortality compared with IABP therapy have not been demonstrated.37 38
Veno-arterial ECMO (VA-ECMO)
Peripheral VA-ECMO is a cardiopulmonary bypass system that extracts blood from a femoral venous inflow cannula, then oxygenates and decarboxylates the blood before returning it to the arterial system through a femoral artery cannula (figure 2). The maximum supported cardiac output with VA-ECMO is >6 L/min. Unfortunately, VA-ECMO increases afterload and creates high LV filling pressure, which increases myocardial oxygen demand and may aggravate pulmonary oedema. Adjunctive LV unloading can be used to mitigate the adverse effects of VA-ECMO on LV afterload, and there may be incremental benefit when ECMO is used with Impella or TandemHeart to decrease LV filling pressure.39–41 Other approaches include venting the LV by percutaneous cannulation, for example, via the interatrial septum, or surgical approaches, for example, an apical vent.42
There are some advantages to VA-ECMO over other MCS devices: (1) rapidity of implantation that does not necessarily require fluoroscopic guidance, (2) rapid oxygenation, (3) respiratory in addition to cardiac support and (4) utility in cases of serious ventricular arrhythmias.43 However, complication rates are relatively high, including limb ischaemia, thromboembolic events, infection and pulmonary oedema. Strategies aimed at preventing lower limb ischaemia include providing antegrade or retrograde perfusion via additional cannulas or perhaps using the subclavian rather than the femoral artery.44 45
There is limited evidence on the efficacy of VA-ECMO to improve outcomes in CGS (table 2). In a recent meta-analysis that included four studies with CGS patients (n=235) VA-ECMO was associated with a 33% higher 30-day survival compared with IABP therapy (95% CI 14% to 52%, p<0.001), but no significant difference was observed when compared with TandemHeart or Impella.46 47
Knowledge gaps: primary LV unloading
There is recent evidence in support of upfront MCS prior to reperfusion with primary PCI (‘primary LV unloading’). This approach is not currently recommended in routine clinical practice, and further research is warranted. Studies in swine showed that primary LV unloading for 30 min with Impella CP reduced myocardial infarct size by decreasing LV workload and reperfusion injury.48 Human studies suggest that early Impella use (pre-PCI) is safe and may be associated with improved survival, although the numbers of patients included in studies is small. In patients with AMI and GCS (n=73) Impella initiation pre-PCI versus post-PCI was associated with improved survival to discharge and at 1 year (31.3% vs 17.6%, log-rank p value=0.03).49 Another registry (n=15,259) reported that Impella use pre-PCI in patients with AMI and GCS was associated with 59% survival to discharge compared with 52% when used as a salvage strategy after inotropic or IABP therapy failure (p<0.001).50 Likewise, in 36 patients from the cVAD registry, initiation of Impella 2.5 pre-PCI versus post-PCI of unprotected left main coronary artery for AMI with GCS was associated with improved survival to discharge (55.0% vs 18.8%, p=0.041); however, STEMI as the presentation was more frequent in the post-PCI group (75.0% vs 40.0%, p=0.049).51 The Door-To-Unload pilot trial reported that in anterior STEMI (n=50) delayed reperfusion after 30 min of unloading with Impella CP was safe compared with LV unloading followed by immediate reperfusion.52 Further research is warranted to compare LV unloading before reperfusion to the current standard of care.
Knowledge gaps: optimal systems of care
Another area for development recommended by the AHA is to apply a system of care whereby individual hospitals are part of a regional network that includes a tertiary CGS centre.53 The rationale is that this may lead to improvements in survival in the same way as regional systems of care incorporating heart attack centres for STEMI patients. Furthermore, CGS team initiatives have been set up with the aim of improving outcomes by: (1) facilitating early multidisciplinary team activation, (2) strict protocol adherence, (3) rapid initiation of MCS and (4) haemodynamic guided management. The recent single-centre Implementation of a Cardiogenic Shock Team and Clinical Outcomes registry (INOVA-SHOCK) found that, compared with historical controls, the introduction of a CGS team improved survival to discharge rates for all cause CGS.54 Robust evidence is lacking on whether CGS teams and specialised CGS centres result in survival benefit.
Knowledge gaps: predicting risk and assessing prognosis
Patient selection for MCS devices needs to be improved. The current consensus recommendations for VA-ECMO55 highlight that selection criteria and techniques for delivering VA-ECMO vary across hospitals, and standardised treatment algorithms are lacking. The relative predictive utility of current risk scores for selecting patients for advanced mechanical therapies needs further evaluation. Two examples are the IABP-SHOCK II risk score,56 which incorporates age >73 years, admission glucose >10.6 mmol/L, admission creatinine >133 µmol/L, thrombolysis in myocardial infarction flow grade <3 after PCI and admission arterial lactate >5 mmol/L and the CardSHOCK risk score,57 which incorporates age >75 years, confusion at presentation, previous AMI or coronary artery bypass graft surgery, acute coronary syndrome aetiology, reduced LV ejection fraction, arterial lactate and estimated glomerular filtration rate. However, evidence is lacking on whether clinical application of such risk scores in this setting improves clinical outcomes. Further research is warranted on: (1) valid predictors of patients with AMI who are at highest risk of developing CGS, (2) predictors of patients most likely to have improved outcomes with MCS devices and (3) biomarker profiling to aid diagnosis, response to management and assessing prognosis. Clarity is also needed on which haemodynamic parameters and specific treatment targets are best for monitoring such patients.
Knowledge gaps: evaluating mechanical and novel therapies
The pathophysiology of CGS needs to be defined in more detail to help develop and test targeted therapies. Biomarkers of the systemic inflammatory response syndrome, for example, interleukins and tumour necrosis factor-α, have been associated with worse outcomes in CGS.58–61 There is interest in novel approaches with potent anti-inflammatory therapy, made all the more relevant by speculation that contact with the surfaces of MCS devices might promote systemic inflammation.43 For example, tocilizumab is a humanised monoclonal antibody against the interleukin-6 receptor, currently licenced for treatment of rheumatoid arthritis. Studies have shown elevated levels of interleukin-6 early in CGS and that interleukin-6 predicts 30 day mortality in CGS patients with AMI.59 However, whether interleukin-6 inhibitors result in a survival benefit is unknown.
Trials in CGS are difficult to perform, and the largest contribution so far to randomised trials in CGS has been in Germany. Enrolment is due to start for the Testing the Value of Novel Strategy and Its Cost Efficacy in Order to Improve the Poor Outcomes in Cardiogenic Shock (EURO-SHOCK) trial, (principal investigator: Anthony Gershlick, UK). EUROSHOCK is a large-scale, international RCT to assess whether VA-ECMO initiated in the acute phase of CGS in patients with AMI undergoing primary PCI can improve survival compared with standard care. Other ongoing studies include DanGer Shock (principal investigator: Jacob Moller, Denmark, NCT01633502) in which patients with AMI complicated by CGS are randomised to conventional care or Impella 3.5 prior to PCI for all-cause mortality at 6 months. The Randomised Trial of Early LV Venting Using Impella CP for Recovery in Patients with Cardiogenic Shock Managed with VA ECMO (REVERSE) trial (principal investigators: Christian Bermudez and Maichael Ibrahim, USA, NCT03431467), is another multicentre RCT assessing whether Impella CP improves survival to hospital discharge in CGS patients managed with VA-ECMO.
The ideal trial
The ideal trial should include patients fulfilling a standard definition of CGS, using haemodynamic criteria. In addition to looking at 30-day mortality reduction (during which period survivors of CGS are more vulnerable), long-term heart failure is another important endpoint. The optimal timing of initiating MCS, i.e. early (pre-PCI) rather than in refractory shock, is important to establish through future trials.
Summary of current best practice
First, current evidence is insufficient to recommend routine use of MCS in CGS. The current American and ESC guidelines give MCS devices (including, Impella, TandemHeart and VA-ECMO) a class IIb, level of evidence C recommendation for use as a bridge to recovery, decision or transplantation in refractory CGS.23 Second, the use of IABP in CGS complicating AMI is not supported by evidence, apart from selected patients with mechanical complications. Third, in view of the results of the CULPRIT-SHOCK trial, immediate multivessel primary PCI is not recommended.23 Based on current evidence, the approach to treating a patient with CGS would involve the following stages (figure 3):
Vasopressors/inotropes as part of initial resuscitation.
Identification of mechanical complications more suitable for a surgical approach.
Routinely, PCI should be limited to the culprit lesion but consider multivessel PCI in some scenarios, for example, where there are multiple culprit lesions.
If CGS persists despite revascularisation consider initiating short-term MCS, with device selection according to the Heart Team decision or institutional protocol. The decision to initiate MCS should also depend on patient age, comorbidities, neurological function and the prospects for long-term survival and good quality of life.
Mortality in patients with CGS complicating AMI remains high, and there are many unresolved problems. Current evidence supports limiting revascularisation to the culprit artery. Future research in this setting should address uncertainty around the efficacy of MCS devices and novel therapies, patient selection for advanced therapy, strategies for risk prediction and whether CGS centres improve survival.
The CULPRIT-SHOCK (Culprit Lesion Only PCI versus Multivessel PCI in Cardiogenic Shock) trial demonstrated that routine immediate complete revascularisation during the index percutaneous coronary intervention (PCI) procedure in acute myocardial infarction (AMI) patients with cardiogenic shock (CGS) may be harmful.
In patients with multivessel disease presenting with ST-segment elevation myocardial infarction and CGS, PCI restricted to the culprit artery should be the default strategy.
The use of intra-aortic balloon pump counterpulsation in CGS complicating AMI is not supported by evidence, apart from selected patients with mechanical complications (i.e severe mitral regurgitation or ventricular septal defect).
There is no robust evidence supporting routine use of mechanical circulatory support, and randomised trials are needed to address this issue.
Multidisciplinary care in specialised centres may be of benefit.
CME credits for Education in Heart
Education in Heart articles are accredited for CME by various providers. To answer the accompanying multiple-choice questions (MCQs) and obtain your credits, click on the ‘Take the Test’ link on the online version of the article. The MCQs are hosted on BMJ Learning. All users must complete a one-time registration on BMJ Learning and subsequently log in on every visit using their username and password to access modules and their CME record. Accreditation is only valid for 2 years from the date of publication. Printable CME certificates are available to users that achieve the minimum pass mark.
We are grateful to Professor Nawwar Al-Attar for providing the images of extracorporeal membrane oxygenation (ECMO) in figure 2.
Contributors AMM wrote the review. TJF wrote the multiple choice questions. KGO revised manuscript drafts with AMM and contributed to the writing. All authors have seen the final version of the manuscript and have contributed to the final version.
Funding AMM (FS/16/74/32573) is supported by a British Heart Foundation clinical research training fellowship. TJF is supported by grants from the British Heart Foundation (PG-17- 25-32884).
Competing interests None declared.
Provenance and peer review Commissioned; externally peer reviewed.
Patient consent for publication Not required.
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.