Cardiogenic shock (CS) is characterised by a low cardiac output state resulting in inadequate tissue oxygen delivery and systemic hypoperfusion, representing a global health priority with high rates of mortality and morbidity. Despite advances in physiological understanding, prompt coronary revascularisation and mechanical circulatory support (MCS), inhospital mortality rates have remained up to 50%, with little improvement over the past few decades.1–3 This persistent mortality reflects both the complex and heterogeneous pathophysiology of CS, as well as variability in the organisation and delivery of care, including delays in recognition, escalation and access to specialist therapies.4
Acute MI-related CS (AMI-CS) is the most common precipitant of CS, accounting for nearly half of all cases and representing the subgroup of patients in whom timely recognition, revascularisation and coordinated systems of care are most critical to improving outcomes.5,6 International guidelines have proposed multidisciplinary AMI-CS centres with the capability to rapidly recognise, classify and deliver advanced therapies, including revascularisation, MCS and specialised intensive care.7–9 Hub-and-spoke models have been proposed to facilitate regionalised systems of care to expedite access to high-volume specialist centres.10,11
In this review we examine the organisation and delivery of care for AMI-CS within specialised CS centres and regionalised networks. We consider the pathophysiology of AMI-CS and the role of centre-based capabilities in its recognition, classification and management. Evidence for dedicated shock protocols, multidisciplinary team (MDT) care models and interhospital transfer systems is evaluated and current approaches to AMI-CS care structure are outlined, highlighting key areas of consensus, variation and unmet need.
Acute MI-related Cardiogenic Shock, Pathology, Classification and Systems of Care
Pathophysiology of Acute MI-related Cardiogenic Shock
CS represents a state in which cardiac dysfunction causes inadequate tissue perfusion and a failure to meet metabolic demand. In AMI-CS, this process is initiated by acute left ventricular (often with coexisting right ventricular) impairment due to acute coronary occlusion, leading to reduced cardiac output, subsequent hypoperfusion and progressive multiorgan failure. In a minority of patients, AMI-CS may instead result from mechanical complications of infarction, including ventricular septal defect, acute mitral regurgitation due to papillary muscle rupture or free wall rupture.
These presentations underline the need for prompt echocardiographic diagnosis and urgent management in high-volume centres with surgical expertise. At a haemodynamic level, reduced cardiac output results in compensatory vasoconstriction leading to impairment of the microcirculation and initiation of a systemic inflammatory response, further contributing to organ dysfunction.12,13
Traditionally, patients with AMI-CS formed the largest cohort of patients with CS and, indeed, many studies have previously shown a trend for an increasing number of patients with AMI presenting in CS.5,14 However, recent observational data have shown that as a proportion of all-comer CS, AMI-CS has declined, making up only 27% of 5,869 patients admitted to a medical cardiac intensive care unit (CICU) with CS.15 Despite this, it should be noted that those admitted with AMI-CS had the highest inhospital mortality of any CS aetiology, at 41%, compared with 31% for de novo heart failure (HF)-related CS (HF-CS), 25% for acute on chronic HF-CS and 31% for secondary CS.15
Unlike several other CS phenotypes, the pathological cascade of AMI-CS may be partially reversed with timely coronary reperfusion and MCS application. Delays in recognition, triage or transfer may result in irreversible MI and worse outcomes.16–18 In AMI-CS, haemodynamic and metabolic derangements such as sustained hypotension, rising lactate and evidence of end-organ injury reflect different stages of CS. The recognition of these physiological markers can guide triage, transfer decisions and timing of specialist therapies.
CS centres and networks are designed to address a broad spectrum of CS phenotypes including, but not limited to, acute on chronic HF, valvular disease and myocarditis, as well as AMI-CS. These pathophysiological and time-sensitive considerations underpin the need for organisation of care in AMI-CS, which is increasingly delivered within specialised CS centres that centralise expertise, infrastructure and multidisciplinary decision-making. Timely recognition, continuous communication across levels of care and early transfer to higher-level CS centres for specialist care when required are central principles to this approach.
Systems of Care
Centralisation of AMI-CS care within regionalised networks is supported by observational data reporting improved outcomes in high-volume specialist centres.19–23
A tiered model of care has been adopted in both US and European guidance, with defined roles for Level 1, Level 2 and Level 3 centres, which are summarised in Table 1 and illustrated in Figure 1.
Level 1 CS centres are high-volume tertiary hospitals functioning as regional CS hubs, described in more detail below. Level 2 centres offer 24/7 primary percutaneous coronary intervention (PCI) with cardiology-led teams and limited MCS capability. These centres can initiate CS pathways and stabilise patients, with early coordination and transfer to Level 1 centres when escalation of care is required.
Level 3 centres are typically non-interventional hospitals with diagnostic capability only. Their role is early recognition, initial stabilisation and rapid transfer within a hub-and-spoke network.9,10
Level 1 Centres: US Guidelines
Guidance from the American Heart Association (AHA) and American College of Cardiology (ACC) defines Level 1 CS centres as hospitals capable of providing complete and comprehensive post-resuscitation and shock care, including 24/7 coronary angiography and PCI, targeted temperature management, multimodal neuroprognostication, MCS and surgical support.8,9 High procedural volumes are emphasised, although evidence remains observational.24 Select centres may also offer durable left ventricular assist device implantation and cardiac transplantation.
Level 1 Centres: European Guidelines
European guidance similarly supports centralisation of AMI-CS care, although the description is currently less prescriptive. European Society of Cardiology (ESC) position statements emphasise management within dedicated CS centres with 24/7 PCI, multiple cath labs, on-site cardiac surgery, advanced haemodynamic monitoring and access to multiple MCS modalities, delivered by an experienced MDT with appropriate staffing ratios.7,25 Patients with out-of-hospital cardiac arrest (OHCA) complicated by AMI-CS are a particularly high-risk subgroup who may benefit from immediate and direct transfer to centres with advanced MCS capabilities.
These guidelines underpin the delivery of protocolised care and form the basis for contemporary regional AMI-CS networks. Effective triage, escalation and interhospital transfer within these systems require a shared and operational definition of CS, as well as a framework for classifying severity.
Definition and Classification of Acute MI-related Cardiogenic Shock
Despite broad agreement on the underlying pathophysiology, defining the transition from haemodynamic compromise to overt CS remains challenging. Diagnostic criteria for AMI-CS vary across randomised trials, guidelines and registries, contributing to heterogeneity in study populations and reported outcomes. A key example is the threshold for hypotension: although most randomised controlled trials (RCTs) have used a systolic blood pressure (SBP) threshold of 90 mmHg, Burkhoff et al. mandated a mean arterial pressure of <70 mmHg and DanGer Shock used an SBP of 100 mmHg.26,27 Such variation contributes to heterogeneous study populations representing different stages of AMI-CS.
Table 2 summarises this variability across major trials.
To improve standardisation, the Shock Academic Research Consortium proposed a consensus definition for AMI-CS incorporating sustained hypotension (SBP <90 mmHg for ≥30 minutes, or vasopressors required to maintain an SBP >90 mmHg) alongside markers of hypoperfusion, including reduced cardiac index, raised arterial lactate, systemic vascular resistance index and end-organ injury.28 The ACC recently proposed the mnemonic SUSPECT-C to support early recognition and triage, integrating Symptoms, Urine output, Sustained hypotension, Perfusion, ECG/Echocardiogram, Congestion and Triage, with shock team activation or transfer where required.9
Although these frameworks enhance recognition, they primarily identify established CS and fail to fully capture its dynamic progression. Importantly, a subset of AMI-CS patients may exhibit significant haemodynamic compromise despite preserved blood pressure, reflecting advanced circulatory dysfunction not detected by conventional thresholds.29
To address this, classification systems have been developed to stage disease severity (Table 3). The Society for Cardiovascular Angiography and Interventions (SCAI) classification, is the most widely adopted framework, describing a continuum from A to E (At risk, Beginning, Classical, Deteriorating and Extremis), each described by a combination of clinical, haemodynamic and biochemical parameters.30 Observational data, including the National Cardiogenic Shock Initiative, demonstrate that both baseline and dynamic SCAI stage are strongly associated with survival to discharge and 30-day mortality, supporting the role of the SCAI classification in risk stratification and guiding escalation of care within regionalised systems of care.31,32 However, limitations of the SCAI classification include its partial subjectivity and that it has not yet been prospectively embedded within an RCT design.
Beyond staging, phenotyping may further refine risk assessment. Earlier approaches relied on single parameters such as lactate, but registry data and machine learning have enabled the identification of distinct CS phenotypes.33 In 1,959 patients, Zweck et al. identified three phenotypes: non-congested (I), cardiorenal shock (II) and cardiometabolic shock (III).33 Compared with phenotype I, patients with phenotype II (OR 3; 95% CI [1.8–5.1]) and phenotype III (OR 4.6; 95% CI [2.7–5.1]) had higher rates of mortality. Within the 410-patient Cardiogenic Shock Working Group AMI cohort, in-hospital mortality was 21%, 45% and 55% across phenotypes I, II and III, respectively.34 Integration of phenotyping with SCAI staging further improves risk stratification and may support more individualised treatment strategies.35
Ultimately, the clinical utility of classification and phenotyping lies in supporting timely decision-making. The SCAI classification provides a dynamic framework for reassessment, where progression or failure to improve, such as transition to stage D, may identify a window for escalation before refractory shock develops.30 In practice, rising lactate, low cardiac power output (CPO), worsening end-organ function or progression in SCAI stage should prompt reassessment of revascularisation, haemodynamic support, MCS suitability and transfer to a specialist centre. When integrated into structured systems of care, these frameworks support early triage and more consistent escalation decisions.
Management of Acute MI-related Cardiogenic Shock within Shock Centres
Treatment of AMI-CS is typically protocolised and delivered through multidisciplinary CS teams, integrating rapid diagnostic assessment, early revascularisation, invasive haemodynamic monitoring, selective MCS use and structured critical care. Trial outcomes should be interpreted within systems of care, because timing, patient selection and centre expertise may influence treatment effect. Table 4 summarises levels of evidence for CS assessment and management across guidelines, while Figure 2 highlights the timeline of published evidence.
Echocardiography
International guidelines emphasise the importance of echocardiographic assessment for CS, where this is considered mandatory for several aspects of AMI-CS care (ESC Class 1, Level C).8,9,36,37 Immediate echocardiography can identify the underlying pathophysiology of AMI-CS in >95% of cases, including left ventricle and right ventricle impairment (in 70% and 30%, respectively), valvular function (mitral regurgitation present in 5–10% of patients) and pericardial effusion or severe mechanical complications in 5%.38 Echocardiography also provides non-invasive estimates of haemodynamic parameters, including cardiac index, pulmonary artery pulsatility index (PAPi) and CPO, and plays a key role in serial assessment, as well as in guiding MCS device implantation and weaning. Despite this, real-world uptake remains variable, with incomplete usage within the first 72 hours of critical care admission and substantial heterogeneity across centres.39 This likely reflects differences in infrastructure, the availability of advanced imaging modalities and access to experienced operators, which may be more challenging in Level 2 or Level 3 centres.
These disparities have important implications for the design of regionalised systems of care, where consistent access to early echocardiographic imaging, potentially across all levels of care, is essential to standardise assessment, identify reversible pathology and guide escalation decisions.
Pulmonary Artery Catheterisation
Pulmonary artery catheterisation (PAC) is currently used in approximately 30% of AMI-CS patients, with substantial international variability. Although specific parameters, such as PAPi and CPO, may have a role in guiding aspects of care, the routine use of PAC for AMI-CS is controversial according to current evidence. Randomised data from a number of trials, including ESCAPE, have not shown benefit for PAC in either mortality or rehospitalisation in CS compared with clinical assessment.40,41 Importantly, PAC placement was associated with complications (e.g. infection, pulmonary artery rupture, arrhythmias) in approximately 3–5% of patients.42 However, these RCTs did not specifically evaluate PAC use in AMI-CS and do not reflect contemporary practice.
Nevertheless, recent observational data suggest that PAC placement within dedicated CS initiatives for select patients with diagnostic uncertainty or nonresponsiveness to initial treatment may be of benefit.43 Clinical guidelines for PAC use in AMI-CS are consequently discrepant, with the ESC proposing a Class IIb recommendation, whereas the ACC/AHA guidelines reserve the use of PAC for refractory cases only.44,45 Ongoing RCTs, such as the PACCS trial (NCT05485376), may address these evidence gaps.
Percutaneous Coronary Intervention
Coronary revascularisation remains a cornerstone of AMI-CS care and may be delivered across Level 1 and Level 2 centres. Several RCTs have assessed its role in AMI-CS care (summarised in Table 5). The SHOCK trial randomised 302 ST-elevation MI patients with shock to medical therapy (thrombolysis plus intra-aortic balloon pump [IABP]) or early coronary revascularisation (PCI/coronary artery bypass grafting and IABP) and was the first to show benefit with early revascularisation.16 However, system-level factors, including timing of presentation, interhospital transfer and centre expertise, were not standardised or explicitly described within the trial design.
Although underpowered to meet its primary outcome of 30-day all-cause mortality, early revascularisation was found to confer a 6-month survival benefit (50.3% versus 63.1%; RR 0.80; 95% CI [0.65–0.98]; p=0.027), which persisted at both 1 and 6 years.46
Following the SHOCK trial findings, the question of revascularisation extent (complete or culprit-only revascularisation) was addressed by CULPRIT-SHOCK, which randomised 706 AMI-CS patients to either culprit-only PCI (with optional staged intervention) or immediate multivessel PCI.1 Culprit-only intervention was associated with a significantly lower 30-day composite outcome of all-cause mortality or renal replacement therapy (RR 0.83; 95% CI [0.71–0.96]; p=0.01), driven primarily by lower mortality. These findings should be interpreted in the context of trial design, where MCS use was not protocolised and broader system-of-care factors were not the primary focus.
Consequently, current ESC and AHA guidelines recommend culprit-only PCI in AMI-CS (ESC: Class I, Level B; AHA: Class I, Level B–R), with both assigning a Class III (harm) recommendation against multivessel PCI during the index procedure.47,48 As practice evolves, the role and timing of non-culprit revascularisation in the context of dynamic CS deterioration warrants re-examination.
Mechanical Circulatory Support in Acute MI-related Cardiogenic Shock
MCS is increasingly used in select AMI-CS patients across CS centres, with device choice and timing guided by haemodynamic profile and response to therapy, as well as institutional protocols. MCS has historically been used on the basis of physiological efficacy and observational data suggesting benefit.49 Primarily, this includes venoarterial extracorporeal membrane oxygenation (VA-ECMO), offering full cardiopulmonary support, and unloading devices (including IABP and microaxial flow pumps, most commonly Impella [Abiomed]). There has been no clear benefit demonstrated when MCS is delivered to allcomers in CS.50 However, specific devices targeted to specific CS subpopulations, particularly AMI-CS, have suggested potential mortality benefit (Table 6).
Initial evidence for MCS use in AMI-CS came from the IABP-SHOCK II trial, which randomised 600 patients with AMI-CS to IABP or no IABP. The trial demonstrated no reduction in 30-day mortality (HR 0.96; 95% CI [0.79–1.17]).51 These findings informed Class III recommendations in international guidelines and contributed to a global decline in routine IABP use in AMI-CS.37,44
The ISAR-SHOCK study demonstrated the haemodynamic efficacy of Impella support, increasing cardiac output by around 1.0–3.5 l/min.52 This precipitated the DanGer Shock trial, a 360-patient RCT that showed that the Impella group had a reduced 180-day mortality (45.8% versus 58.5%; HR 0.74; 95% CI [0.55–0.99]) in AMI-CS, primarily with ST-elevation MI.27 Importantly, that study excluded patients with comatose OHCA (Glasgow Coma Scale <8), established multiorgan failure and with severe right ventricular failure, thereby reflecting a highly selected population of AMI-CS. Subsequent subgroup analysis showed improved lactate clearance and reduced vasoactive drug requirement in the Impella group.53 Although there was increased use of renal replacement therapy in the Impella group, none of the patients required long-term haemodialysis.53
IABP and Impella offload the left ventricle to varying degrees, whereas VA-ECMO provides high-flow circulatory support (approximately 4–6 l/min) but at the expense of increasing left ventricular afterload and mechanical work of the heart (in approximately 20%), particularly in the absence of concurrent mechanical unloading.54 Recent data from highvolume US centres alongside two VA-ECMO RCTs have not shown benefit in AMI-CS.3,55,56 High OHCA rates, cross-over and heterogeneity in the timing and use of adjunctive strategies, including left ventricular unloading, highlight important system-level factors not fully captured in the ECLS-SHOCK trial and should be considered when interpreting its findings.3
In summary, there is no evidence to suggest VA-ECMO as first-line treatment in AMI-CS. Its role in refractory cardiac arrest remains promising, but with significant logistical and implementation challenges.
Current international guidelines assign Impella/ECMO a Class IIb (Level C) recommendation (Table 4) for refractory CS based on limited evidence. However, guidelines for VA-ECMO are likely to be iterated to a more negative recommendation in light of recent trials.44,45 A consensus statement prior to DanGer Shock publication emphasised that MCS application should be based on haemodynamic parameters including PAC, biochemical markers and overall clinical scenario.57 Dedicated AMI-CS centres may possess the volume and expertise to more efficiently decide on the suitability, timing and device selection of MCS, although definitive evidence supporting this approach is lacking. More broadly, interpretation of randomised trial data in AMI-CS should be considered within the context of systems of care.
Trial populations are often highly selected and do not reflect real-world practice. The effectiveness of therapies, particularly MCS, is influenced by timing, patient selection and centre expertise. Variations in these factors may contribute to neutral or negative trial results.
Care Beyond the Cardiac Cath Lab
The management of AMI-CS patients following revascularisation is highly heterogeneous, despite being a critical determinant of outcome. Although guideline recommendations exist for components of haemodynamic monitoring and management, there is a lack of randomised evidence or protocolisation in this setting.
In broader critical care populations, closed models (led by dedicated critical care teams rather than the admitting team) have been associated with improved outcomes, and emerging data suggest similar in the CICU setting.58 Observational data also support cardiology intensivist comanagement in the coronary care unit.59 Respiratory support is often needed, ranging from non-invasive strategies, such as high-flow nasal oxygen and continuous positive airway pressure or bilevel positive airway pressure, to invasive mechanical ventilation in severe cases, and may reduce the work of breathing and improve gas exchange, although AMI-CS-specific data are limited. Renal support, antimicrobials and nutrition also require intensivist-led MDT care.25
Inotropic and Multiorgan Support
Vasopressors, most commonly norepinephrine, are used to maintain perfusion pressure, whereas inotropes are used to augment cardiac output. Inotropes are used in up to 90% of AMI-CS patients after PCI despite risks of worsening ischaemia through increased myocardial work.51 Although the SOAP II and OPTIMA-CC trials have suggested the superiority of norepinephrine over dopamine and epinephrine, generalisability to AMI-CS is limited.60,61 There is no clear evidence to support the superiority of one vasoactive strategy over another in AMI-CS. Renal dysfunction complicates approximately 30% of cases of AMI-CS, with 20% requiring renal replacement, although early initiation in the absence of classical indications confers uncertain benefit.8,27,62
Dedicated Centre Protocols for Acute MI-related Cardiogenic Shock Treatment
Contemporary AMI-CS management has evolved towards centre-based protocols integrating early recognition, multidisciplinary activation, invasive haemodynamic assessment and selective MCS use. Evidence supporting these approaches is largely observational, with important limitations including selection bias, incomplete cohort data capture and limited external validity. Nevertheless, although derived from observational data, several dedicated CS protocols delivered at Level 1 centres have reported improved outcomes compared with historical care (Table 7), despite marked heterogeneity in physiology-driven strategies and device usage.
The Detroit Cardiogenic Shock Initiative used standardised early MCS approaches. In a pilot cohort of 41 consecutive patients with AMI-CS in a Level 1 centre, 19.5% of patients were transferred from another site.21 This protocol included a single-call activation with rapid invasive haemodynamic assessment and near-universal Impella use, with pre-PCI MCS application in 66% of patients. Escalation and weaning were guided by serial physiological metrics, including CPO <0.6 W and PAPi <0.9.
In this small, select observational cohort, survival to discharge was 76%, exceeding historical benchmarks.21
The INOVA/Virginia programme evaluated outcomes before and after implementation of a multidisciplinary shock team, activated through a dedicated call, in 82 patients treated at a Level 1 centre.63 The protocol incorporated early transthoracic echocardiography, coronary angiography and invasive haemodynamics in most patients. Shock was defined using haemodynamic criteria (cardiac index ≤1.8 l/min/m² [≤2.2 on inotropes] and pulmonary capillary wedge pressure ≥15 mmHg), with management and MCS selection guided by serial reassessment including CPO <0.6 W and PAPi. Daily joint critical care and shock team meetings facilitated management plans, including MCS escalation.
Following protocol implementation, 30-day survival improved from 47% to 77% over two consecutive years.63
In Utah, an HF-led CS team managed a cohort of 123 patients with refractory CS requiring MCS, compared with 121 historical controls.23 Management incorporated early invasive haemodynamic assessment and revascularisation where appropriate. Escalation to MCS was guided by haemodynamic criteria (cardiac index <2.2 l/min/m2 or pulmonary capillary wedge pressure >15 mmHg), alongside regular shock team consultation. Implementation of the shock team was associated with higher in-hospital survival (61% versus 47.9%) and reduced 30-day mortality (HR 0.61; 95% CI [0.41–0.93]).23
The Ottawa programme reported outcomes in 100 patients with CS (64 managed with a code-shock protocol and 36 historical controls), representing a heterogeneous cohort with predominantly non-ischaemic aetiologies.64 Approximately three-quarters of patients were transferred from external institutions, although this differed between groups (63% in the code-shock group versus 94% in the historical controls). CS was defined by reduced cardiac output (hypotension or low cardiac index) with evidence of end-organ hypoperfusion. Management was coordinated through a digitally enabled activation pathway with daily shock team review and included the selective use of MCS (45%).
Although short-term outcomes were unchanged, with similar 30-day survival (~70%), programme implementation was associated with improved longer-term survival (67% versus 42%; HR 0.53).64 The Bordeaux experience evaluated 250 patients with refractory CS before and after implementation of a multidisciplinary shock team.65 Selection for advanced support was guided by validated risk scores (including CARDSHOCK ≥7) alongside markers of hypoperfusion and endorgan dysfunction. One-year survival improved from 45% to 59%, with shock team management, which was independently associated with improved survival (HR 0.59; 95% CI [0.40–0.88]). A substantial proportion of patients received short-term MCS (n = 166), reflecting a highly selected population and introducing important selection bias.65
The Ancona programme evaluated protocolised CS care in 167 patients within a CICU using the Shock Academic Research Consortium definition of CS.66 Ninety per cent of patients received MCS, primarily IABP. There were daily shock team consultations, without protocolised escalation pathways. This organisational model was associated with improvement in in-hospital survival from 43% to 71%, although the findings are limited by the observational study design and high device use.66
Finally, the ACT-SHOCK pathway (Bristol, UK) was a pragmatic, physiology-triggered shock team model embedded within a highvolume UK Level 1 centre.67 Activation of a CS team was based on predefined markers of hypoperfusion (e.g. SBP <90 mmHg, heart rate <50 or >150 BPM, SpO₂ <90%, lactate >4 mmol/l, base excess <-5 mmol/l or left ventricular end-diastolic pressure >25 mmHg), and could be triggered prehospital, prior to PCI or during PCI (19.6%, 17.6% and 64.8% of patients, respectively). The revascularisation strategy was not protocolised and left to operator discretion.
Unlike device-driven protocols, ACT-SHOCK prioritised rapid multispecialty mobilisation during PCI without mandated routine invasive haemodynamic monitoring or MCS. Critical care management was not protocolised, including decisions regarding MCS escalation. Overall, MCS use was low (14% Impella, 6% ECMO).
In 165 patients, ACT-SHOCK was associated with lower 1-year mortality (35% versus 48.1%) and an independent reduction in adjusted 1-year death (HR 0.53; 95% CI [0.30–0.92]) compared with a historical control cohort with AMI-CS in the preceding 12 months.67
Collectively, these centre-based experiences demonstrate two important observations. First, that bespoke patient care pathways are associated with higher observed survival compared with conventional care. Second, that improved outcomes in contemporary AMI-CS may be driven less by any single device and more by structured systems of care that integrate early activation, haemodynamic assessment and protocolised escalation aligned with institutional expertise.
Nevertheless, RCTs are now required to assess the value of the components of this specialised approach before they can be recommended in routine clinical care. Notably, many of these programmes included patients transferred from referring centres, highlighting the importance of coordinated regional systems to enable timely access to specialist care.
These findings should be interpreted with caution because most of the available evidence is observational and prone to bias. Differences in patient selection, survival to treatment and underlying illness severity, particularly in relation to interhospital transfer and the use of MCS, may influence outcomes. In addition, comparisons with historical cohorts and the lack of data on patients not selected for transfer or advanced therapies may further distort estimates of benefit.
Regionalised Cardiogenic Shock Pathways of Care
Early recognition, rapid triage and MDT-supported transfer between centres of differing capability may be associated with improved outcomes, with decisions guided by expert teams and clear protocols.
In a large observational analysis of 412,000 CS hospitalisations between 2010 and 2014 in the US, in-hospital mortality was lower at hub centres than in spoke hospitals both for direct admissions (39.3% versus 47.8%) and for patients transferred to hubs (33.4% versus 47.8%).19 On multivariable analysis, both direct admission to a hub (OR 0.86) and interhospital transfer (OR 0.72) were associated with lower mortality after adjustment. These findings suggest that transfer to a hub centre may be associated with improved survival, although this likely reflects the selection of patients sufficiently stable for transfer and causality cannot be inferred.
Operational regionalisation is particularly important for AMI-CS, where timely coronary reperfusion and access to advanced therapies require coordinated delivery across specialised centres. However, high-quality evidence defining optimal transfer criteria, the timing of transfer and pretransfer management remains limited. Existing data are derived predominantly from observational studies focused on Level 1 centres and are subject to important biases, including incomplete capture of network activity and lack of outcome data for patients not selected for transfer, which may overestimate the apparent benefit of centralised care.
Multidisciplinary Cardiogenic Shock Hub Multidisciplinary Teams
A central component of these systems is the provision of a cardiogenic shock MDT, typically a team of specialists coordinating care from Level 1 centres to referring Level 2 and Level 3 hospitals. Activation may occur prehospital, upon emergency department arrival or during PCI, enabling rapid MDT decisions on local management, transfer or escalation.
The composition of MDTs, transfer criteria and protocols vary between systems, particularly for AMI-CS. Although MDT-led care may support more coordinated decision-making across a network, the absence of transparent evidence-based transfer criteria may introduce variability and potential delays. Regionalisation within the three principal reported North American programmes is summarised in Table 8.
The National Cardiogenic Shock Initiative represents a multicentre standardisation model of care centred around Level 1 centres rather than a tightly governed regionalised network.20 Although no emergency medical services bypass or formal PCI prior to transfer mandate was reported, centres used a 24/7 single-call activation with cath lab preactivation. Transfer decisions were guided by conventional CS criteria, including hypotension or inotrope/vasopressor requirement, hypoperfusion and a cardiac index <2.2 l/min/m² or cardiac power output ≤0.6 W. Overall, 27% of patients were transferred from external institutions (including 16% receiving MCS prior to transfer).
On multivariable analysis, interhospital transfer was independently associated with higher inhospital survival, although this may reflect the selection of patients suitably stable for transfer. MDT activation included interventional cardiology, critical care and perfusion teams. However, no formal spoke–hub triage algorithm or screening log of non-transferred patients was used.
The University of Utah Cardiac Recovery Program reflects a quaternary hub-and-spoke model centred on interhospital transfer. In the model, shock identification occurs at the spoke hospital, with subsequent transfer to a dedicated HF-led cardiac recovery team integrating interventional cardiology, cardiac surgery and MCS expertise.23 Activation operates 24/7, and CS was defined using conventional criteria, including hypotension or vasopressor requirement, cardiac index <2.2 l/min/m², elevated filling pressures and evidence of end-organ hypoperfusion. PCI before transfer was not mandated but recommended if at a PCI-capable spoke hospital.
Earlier transfer and shorter spoke dwell time were associated with improved survival.23 Although over 700 patients were transferred, no comprehensive screening log of non-transferred patients was reported.
The Inova Heart and Vascular Institute network provides the most explicit regional governance structure, evolving from earlier single-centre shock team experience.22,63 Regionalisation begins at first hospital presentation, without a formal emergency medical services bypass pathway. Thus, Level 2 and Level 3 centres transfer AMI-related shock patients directly to a Level 1 hub via a structured single-call pathway, generally without prior PCI, although initial stabilisation, including with MCS, may be undertaken at the spoke centre. Conventional CS criteria are used, with most patients presenting to Level 1 centres having an SCAI Grade C–E.
Invasive haemodynamic phenotyping at the hub guides escalation, with decisions taken by an MDT integrating HF, interventional cardiology, critical care and cardiothoracic surgery specialists. A uniform algorithm for transfer and activation is applied across all centres with biweekly hub–spoke quality review. Clinical and administrative champions at spoke centres reinforce network fidelity, although a screening log of non-transferred patients does not exist.
Across reported regional models, several common features emerge, including single-call activation, predefined shock criteria and hub-based MDT review. However, approaches differ in key areas, including whether PCI is undertaken before transfer, whether MCS is initiated at the spoke centre and how transfer decisions are operationalised. In particular, no model provides robust comparative evidence to determine whether AMI-CS patients should undergo culprit PCI locally before transfer or be transferred directly to a Level 1 hub.
The absence of screening logs and outcomes for non-transferred patients further limits comparison between systems and prevents firm conclusions regarding optimal transfer strategy.
Collectively, these models demonstrate that single-call activation, predefined shock criteria and rapid MDT mobilisation underpin effective regional systems.
Out-of-hospital Cardiac Arrest and Cardiogenic Shock
OHCA represents abrupt circulatory collapse in the community, with CS complicating up to half these presentations and contributing to high mortality.32 Recent observational data highlight the impact of cardiac arrest on outcomes in AMI-CS beyond clinical trial populations, with both OHCA and in-hospital cardiac arrest associated with significantly worse survival.68 In a simulated cohort study of 2,331 resuscitated OHCA patients from EUCAR and GLOBAL-MIRACLE, eligibility for landmark AMI-CS trials was low and varied substantially by trial criteria, with ECLS-SHOCK-like criteria selecting a higher-risk phenotype and DanGer Shock-like criteria selecting a more favourable cardiac phenotype.69
Postarrest CS exhibits a distinct pathophysiology characterised by global ischaemia–reperfusion injury, myocardial stunning and systemic inflammation, resulting in profound haemodynamic instability. Within regionalised systems of care, OHCA represents a particular challenge, because triage and transfer decisions must balance potential haemodynamic recovery against the risk of severe neurological injury.
Neurological outcome is a key determinant of prognosis and frequently limits the benefit of advanced therapies.70 Although MCS may improve haemodynamics, evidence for clinical benefit in OHCA remains inconsistent and inconclusive across observational studies and early randomised trials.71,72 Within regional systems, extracorporeal life support/ECMO may provide haemodynamic support in select patients with refractory arrest or postarrest shock, but its benefit remains uncertain and is highly dependent on early patient selection, neurological prognosis, timing and centre expertise.73
Risk stratification models, including the MIRACLE2 score, have been shown to have equivalent performance in those with concomitant CS and may support decision-making by identifying patient subgroups most likely to benefit from escalation of care.74,75 Integration of such approaches into regionalised pathways has the potential to refine triage, optimise resource use and improve patient selection. An increasing proportion of patients present with advanced age, comorbidity and limited physiological reserve. In such patients, early identification of poor prognosis using clinical assessment and risk stratification tools may support decisionmaking regarding escalation of care.
The incorporation of palliative care principles should be considered within MDT discussion and regional transfer pathways.
Survivorship, Rehabilitation and Research
Long-term survival following AMI-CS is favourable among patients discharged alive, but is frequently accompanied by substantial morbidity.46 Survivors, particularly those with concomitant OHCA, often experience persistent physical, cognitive and psychological impairment, in addition to complications related to critical illness, MCS and cardiac surgery.76 Data on late mortality, functional recovery and rehabilitation after AMI-CS remain limited, particularly among patients with concomitant OHCA.
Within regionalised systems of care, survivorship should be considered an integral component of specialist centre delivery rather than a downstream consequence of acute management. Although cardiac rehabilitation improves outcomes following AMI, evidence in AMI-CS remains limited and access is variable.77 Specialist centres may be uniquely positioned to coordinate the rehabilitation and management of CS-specific complications. The integration of these processes within regional networks will be essential to optimise long-term outcomes.
Future Directions
Advancing care in AMI-CS requires improvements in diagnosis, treatment and systems of care. Key evidence gaps and proposed research priorities are summarised in Figure 3. Standardisation of definitions and classification, with integration of haemodynamic, clinical and biochemical markers, is needed to better characterise disease severity and phenotype.25 Advances in machine learning and large-scale registries may enable more precise risk stratification.
Uncertainty remains regarding the optimal timing and selection of PCI, MCS and post-procedural care, particularly across differing centre capabilities. Expansion of regional shock networks, including improved prehospital stratification and streamlined MDT decision making, alongside large-scale, adaptive randomised trials and collaborative research, will be essential to guide future care pathways.