Standardized Definitions for Cardiogenic Shock Research and Mechanical Circulatory Support Devices: Scientific Expert Panel From the Shock Academic Research Consortium (SHARC)

Abstract

The Shock Academic Research Consortium is a multi-stakeholder group, including representatives from the US Food and Drug Administration and other government agencies, industry, and payers, convened to develop pragmatic consensus definitions useful for the evaluation of clinical trials enrolling patients with cardiogenic shock, including trials evaluating mechanical circulatory support devices. Several in-person and virtual meetings were convened between 2020 and 2022 to discuss the need for developing the standardized definitions required for evaluation of mechanical circulatory support devices in clinical trials for cardiogenic shock patients. The expert panel identified key concepts and topics by performing literature reviews, including previous clinical trials, while recognizing current challenges and the need to advance evidence-based practice and statistical analysis to support future clinical trials. For each category, a lead (primary) author was assigned to perform a literature search and draft a proposed definition, which was presented to the subgroup. These definitions were further modified after feedback from the expert panel meetings until a consensus was reached. This manuscript summarizes the expert panel recommendations focused on outcome definitions, including efficacy and safety.

Cardiogenic shock (CS) is associated with a high mortality risk. Despite advances in temporary mechanical circulatory support (MCS) technology in the past 2 decades, studies have reported 30% to 50% mortality at 6 to 12 months, though this may exceed 70% depending on severity of shock and individual patient factors.^1-3 Among patients diagnosed with CS, therapeutic options may vary depending on the individual patient’s comorbidities, underlying pathology, severity of illness, comorbidities, and clinical shock phenotype.^2,4-6 Contemporaneous management includes medical therapies, revascularization with percutaneous coronary intervention or coronary artery bypass grafting in patients with infarct-related CS, MCS devices and surgical or percutaneous treatment of structural abnormalities.⁷ Given limited high-quality evidence to inform best therapeutic practices, the design and execution of future clinical trials with comparative evaluation of these therapies is necessary. However, the complexity of CS has made it challenging to perform clinical trials in part because of a lack of standardized definitions describing stages of CS and clinical safety and effectiveness outcomes. This lack of consistent definitions makes it difficult to pool and compare data across studies and registries. A separate but related barrier to clinical research in patients with CS is lack of consensus on who can or cannot provide informed consent.⁸

The Academic Research Consortium (ARC) is a multi-stakeholder group, including representatives from the US Food and Drug Administration (FDA) and other government agencies, industry, and payers, convened to develop pragmatic consensus definitions with the aim of improving the consistency and execution of clinical trials supporting regulatory decisions for cardiovascular devices.⁹ Previous ARC activities include establishing consensus outcome definitions for the evaluation of heart failure (HF) therapies and aortic valve, mitral valve, and coronary interventions, among others.^10-13 The Shock Academic Research Consortium (SHARC) Writing Committee was initiated in conjunction with the Cardiac Safety Research Consortium cardiogenic shock program in 2020 to develop pragmatic consensus definitions useful for the evaluation of MCS devices and other therapies in CS.⁸

METHODOLOGY

The SHARC Writing Committee was composed of diverse stakeholders, including experts and academic leaders from heart failure, interventional cardiology, cardiac surgery, critical care cardiology, emergency medicine, industry, payers, statistics, and the FDA. Several in-person and virtual (during the pandemic) meetings were convened between 2020 and 2022 to discuss the need for developing the standardized definitions required for evaluation and performance of therapies including MCS devices and drugs in clinical trials for CS patients. The Writing Committee identified key concepts and topics by performing literature reviews, including previous clinical trials, while recognizing current challenges and the need to advance evidence-based practice and statistical analysis to support future clinical trials. For each category, a lead (primary) author was assigned to perform a literature search and draft a proposed definition, which was presented to the subgroup. After agreement from each member, the definition was presented to the full expert panel. These definitions were further modified after feedback from the expert panel meetings until a consensus was reached.

DEFINING CARDIOGENIC SHOCK FOR CLINICAL TRIALS AND REGISTRIES

Cardiogenic Shock Definition

In the development of a consensus CS definition, the SHARC Writing Committee recognized that CS is fundamentally a clinical diagnosis⁹ but that there is also a need for a more robust definition, including phenotyping CS, in order to maintain the internal validity of clinical research evidence. Accordingly, we recommend the adoption of 2 definitions (Table 1): a pragmatic definition best suited to clinical routine, which defines CS as a cardiac disorder that results in both clinical and biochemical evidence of sustained tissue hypoperfusion irrespective of underlying blood pressure; and a comprehensive definition for application in clinical trials or registries, which defines CS as a cardiac disorder that results in a systolic blood pressure <90 mm Hg for ≥30 min (or the need for vasopressors, inotropes or MCS to maintain systolic blood pressure ≥90 mm Hg) with evidence of tissue hypoperfusion.

Table 1.

Proposed Definitions of Cardiogenic Shock

Proposed definition	Cardiac disorder
Definition for clinical practice	Cardiac disorder that results in both clinical and biochemical evidence of sustained tissue hypoperfusion
Definition for clinical trials	Cardiac disorder that results in a systolic blood pressure <90 mm Hg for ≥30 min (or the need for vasopressors, inotropes or mechanical circulatory support to maintain systolic blood pressure ≥90 mm Hg) with evidence of hypoperfusion
Hemodynamic criteria (optional)
Cardiac index	≤2.2 L/(min·m2)*
Hypoperfusion criteria (≥1)	Elevated arterial lactate (>2 mmol/L)
	Acute kidney injury (creatinine ≥2× upper limit of normal) or oliguria (eg, urine output <0.5 mL/(kg·h)
	Acute hepatic injury (eg, ALT >3× upper limit of normal)
	Cool or mottled extremities
	Altered mental status not explained by an alternative cause
Hemodynamic criteria (optional)
Cardiac index	≤2.2 L/(min·m2)
Systemic vascular resistance index	>2200 dynes/(cm·sec−5)*
Normotensive cardiogenic shock subtype cardiac disorder	Systolic blood pressure ≥90 mm Hg without the need for vasopressors, inotropes, or mechanical circulatory support with evidence of hypoperfusion and other potential causes of markers of hypoperfusion have been excluded
Hemodynamic criteria (optional)
Cardiac index	≤2.2 L/(min·m2)
Systemic vascular resistance index	>2200 dynes/(cm·sec−5)*

ALT indicates alanine transaminase.

^*Measured via pulmonary arterial catheterization, noninvasive cardiac output monitors, or echocardiography.

Noninvasive testing such as x-ray, electrocardiogram, and echocardiogram can contribute to determining the underlying mechanism of hemodynamic instability. Further additional testing such as transesophageal echocardiogram or cardiac computed tomography scan can help rule out other etiologies such as aortic dissection or pulmonary embolism. Noninvasive cardiac output monitoring, including but not limited to bioreactance technique, can be used to estimate stroke volume.¹⁴ However, many of these approaches to noninvasive testing have very limited studies available to validate their application in patients with CS.¹⁵ The use of invasive hemodynamic parameters is included as an optional diagnostic criterion given their variable utilization, lack of high-quality data^1,16,17 to support the routine use of pulmonary arterial catheters, and uncertainty regarding the optimal cardiac index cutoff to define CS.^18-20 In addition, a limitation of commonly used historical clinical trial definitions is the reliance on an absolute systolic blood pressure inclusion criterion that would exclude patients with a normotensive CS subtype that occurs in approximately 5% of cases. Recently, studies have shown that patients with isolated hypoperfusion may carry higher mortality than patients with isolated hypotension.²¹ To address this diagnostic gap, the Writing Committee has incorporated a novel CS definition for patients with an unsupported systolic blood pressure ≥90 mm Hg and hypoperfusion in whom other causes of hypoperfusion have been reasonably excluded. Evidence of tissue hypoperfusion is essential for the diagnosis of shock and supported by the presence of one or more of the clinical hypoperfusion criteria in the absence of a clear alternative cause for the abnormality. We have described this category separately as the normotensive CS subtype, which should be defined as hypoperfusion related to presumed cardiac etiology in the absence of a clear alternative cause.

Acuity Staging

The Society for Cardiovascular Angiography and Interventions (SCAI) classification of CS was originally proposed in 2019 and endorsed by international cardiovascular and critical care professional societies and organizations as a means of providing a common simplified language that can be dynamically applied across phases of care (Table S1).^4,6 The recently updated SCAI consensus recommends using specific parameters such as lactate level, ALT (alanine transaminase), systolic blood pressure or mean arterial pressure, systemic pH, and the occurrence of out-of-hospital cardiac arrest to define severity in the staging of CS. We suggest that SCAI staging be applied using both specific clinical variables and temporary trajectory, as well as separate classification based on clinical judgment. At a minimum, clinical studies should be explicit about the methods of capture and classification of SCAI shock stage. This staging system, as well as its clinical modifiers, has relatively good risk discrimination for inpatient mortality, with wide heterogeneity in multiple validation datasets, and has been adopted into clinical care and registries.^3,5,22-24 Recently, incorporating phenotypic variables into the SCAI classification was proposed to improve predictability.²⁵ The SCAI shock stages may help effectively communicate the time sensitivity and appropriateness of medical or MCS interventions. The Writing Committee recommends incorporating the SCAI CS staging framework into clinical research as a key element in data collection in clinical CS registries and clinical trials, as well as a potential eligibility criterion for such clinical studies. Serial assessment of SCAI stage may be valuable given the dynamic nature of shock severity and relevance to prognosis. Standardizing the phenotyping of study populations will potentially allow for better interstudy comparisons and interpretation of treatment effects (Figure 1).

Figure 1. Society for Cardiovascular Angiography and Interventions (SCAI) classification of cardiogenic shock.

Cardiac arrest (A) modifier: includes patients who require or receive cardiopulmonary resuscitation (CPR) or defibrillation for pulseless arrhythmia. CS indicates cardiogenic shock; ECMO, extracorporeal membrane oxygenation; and MCS, mechanical circulatory support.

CLASSIFICATION OF CS POPULATIONS FOR CLINICAL RESEARCH

As noted, the population of patients presenting with CS is highly heterogeneous with respect to the primary underlying cause of CS, its subsequent clinical manifestations, and outcomes.²³ This variation may relate to whether the presentation is acute (de novo), acute-on-chronic, or occurring in specific settings such as postcardiotomy or after cardiac arrest.^18,26 In addition, pathophysiologic and hemodynamic phenotypes, such as whether myocardial dysfunction is left, right, or bi-ventricular, may influence the natural history, management strategy, and therapeutic response.^2,25,27 In addition, initial vasoconstriction may evolve to vasodilation over time, leading to a different phenotype of CS. In selecting and describing patients with CS enrolled in epidemiologic studies or clinical trials, the Writing Committee believes it essential to characterize the nature of the CS being studied (Table 2). The Writing Committee developed a framework (Figures 2 and 3) which identifies 4 main etiologic categories of CS: (1) acute myocardial infarction–related CS (AMI-CS); (2) heart failure–related CS (HF-CS); (3) postcardiotomy CS; and (4) nonmyocardial (secondary) CS. Finally, special consideration needs to be given for whether patients with cardiac arrest should be included in or excluded from a clinical trial. Stratification for resuscitation at randomization should be considered. The Writing Committee recognizes that, in some cases ≥1 contributor to CS may be present. In general, we recommend capturing the suspected predominant driving etiology of shock, as well as other contributors.

Table 2.

Factors to Be Considered in the Classification of Cardiogenic Shock Populations

Factor	Clinical scenario
Underlying Cause	Acute myocardial infarction Chronic heart failure Primary nonmyocardial cause (eg, valvular, arrhythmia)
Setting	Postcardiotomy
Timing	Acute (de novo) Acute-on-chronic Unknown chronicity
Other considerations	Left ventricular, right ventricular, or biventricular predominance Hemodynamic subtypes: “mixed” vasodilatory, normotensive Transient shock Postcardiac arrest syndrome

eCPR indicates extracorporeal life support for cardiopulmonary resuscitation.

Figure 2. Framework for classification of cardiogenic shock.

Figure 3. Summary of classifications of cardiogenic shock for clinical research.

Acute Myocardial Infarction–Related CS

AMI-CS includes patients with the onset of CS related to acute myocardial infarction (MI) with or without ST-segment elevation (ST-segment elevation MI [STEMI] and non–ST-elevation MI [NSTEMI], respectively). Left or right ventricular dysfunction from ongoing myocardial ischemia, ischemic injury, or mechanical complications of MI must be the primary cause of shock. CS resulting from acute bradyarrhythmia, heart block or tachyarrhythmia, post–cardiac arrest, or any other complications in the setting of AMI are also classified as AMI-CS.

Heart Failure–Related CS

HF-CS includes patients with CS related to primary myocardial dysfunction who do not meet the criteria for AMI-CS. HF-CS may be related to chronic progressive left, right, or bi-ventricular dysfunction and may be non-ischemic or ischemic in etiology. This category does not include patients with CS related to a primary nonmyocardial cause, such as severe valvular disease, pericardial disease, pulmonary embolism, or ongoing arrhythmia. In addition, patients with CS in the immediate postcardiotomy setting are classified elsewhere. Patients with HF-CS may be further characterized by the specific etiology of the underlying myocardial dysfunction (Table 3), such as acute myocarditis, takotsubo cardiomyopathy, peripartum cardiomyopathy, alcohol-related cardiomyopathy, tachycardia-related cardiomyopathy, excess use of beta blockers or other negative inotropes, septic cardiomyopathy, or infiltrative diseases such as cardiac amyloidosis. Patients in this category may have reduced or preserved left ventricular ejection fraction depending on the etiology. For example, patients with hypertrophic cardiomyopathy with left ventricular outflow tract obstruction are also included in this category. Patients who develop myocardial dysfunction related to an arrhythmia that is no longer present (eg, myocardial stunning from ventricular tachycardia or ventricular fibrillation arrest or tachycardia-induced cardiomyopathy) are also included in this category.

Table 3.

Relatedness and Reintervention

Relatedness	Attribution of events
Device-related*	Events directly attributable to the device itself Events that occur as a result of improper use of a device are still device-related
Procedure-related*	Events that occur from the procedure, irrespective of the device.
Shock-related*	Events to be ascertained prospectively
Reintervention	Secondary procedure performed to address problems that arise as a result of the index procedure. Diagnostic procedures are not included in this definition.

^*For each category of adverse events, additional classification to be considered as not related, possible, probable, or causal (or definite).

Patients with HF-CS should be further subcategorized into those with de novo HF-CS related to acute myocardial dysfunction that is known or suspected to be new in onset versus those with acute-on-chronic HF-CS who present with an acute decompensation of chronic HF.²⁸ For completeness of categorization, patients with newly discovered myocardial dysfunction without sufficient data to determine duration can be classified as being of “unknown chronicity.”

Postcardiotomy CS

Postcardiotomy CS includes patients who develop CS in the perioperative period with low cardiac output syndrome as the primary driver of shock. Postoperative patients with specific alternative causes of CS (eg, pericardial tamponade, patients developing perioperative right ventricular dysfunction) would still be best classified under their primary category, with the caveat that it might be difficult to define the primary cause in this patient population because CS in this setting could be from a combination of causes.

Nonmyocardial (Secondary) CS

Secondary CS includes patients with CS related to a primary nonmyocardial cardiac cause. Potential primary nonmyocardial cardiac causes include severe valvular heart disease, pericardial disease, and hemodynamically significant pulmonary embolism. Shock related to an ongoing refractory arrhythmia is included in this category, but not myocardial dysfunction related to an arrhythmia that is no longer present (eg, stunning after sustained ventricular tachycardia arrest or tachycardia-induced cardiomyopathy), which is included in the HF-CS category. By convention, this category also includes shock predominantly related to pulmonary hypertension with resultant right ventricular failure. This category does not include postcardiotomy CS as the primary driver of shock or toxic causes of myocardial dysfunction, such as beta-blocker overdose or septic cardiomyopathy. Rather, cases of toxic myocardial dysfunction are included in the HF-CS category. Similarly, patients with pulmonary hypertension from left-sided disease are classified as HF-CS.

The SHARC Writing Committee recognizes the deliberate heterogeneity of the natural history and treatment of patients within the category of secondary CS. This grouping is by consensus, recognizing the value of parsimony in the overall framework, as well as preserving the relative specificity of the 3 other categories.

Additional Modifiers to the Primary Category

In addition to classification of patients into one of the four primary categories of CS, the SHARC Writing Committee recommends capture of additional potential modifiers of prognosis or therapeutic response.

First, CS should be classified as being associated with the presence of predominantly left, right, or bi-ventricular dysfunction based on the clinical, hemodynamic, and echocardiographic data determining whether the patient has elevated pulmonary capillary wedge pressure, central venous pressure, or both.²⁵ In addition, when available, clinical and hemodynamic data should capture left, right, or biventricular congestion profile.

Second, when sufficient hemodynamic data are available to do so, CS may be classified into key hemodynamic subtypes: (1) classic, which also infers congestion and vasoconstriction; (2) vasodilatory or mixed; (3) normotensive; and (4) euvolemic CS. Patients with “mixed shock” are now recognized as a high-risk, understudied subpopulation of patients with CS. This sub-phenotype includes patients categorized as having initially normotensive CS before developing vasoplegia or those with combined CS and septic shock. Although inconsistently defined, mixed shock has been described as “vasodilatory CS,” with a hemodynamic profile characterized by low cardiac output, elevated filling pressures, and low systemic vascular resistance (SVR) interpreted in the context of administered vasoactive medications.² Although CS is usually marked by a compensatory increase in SVR to maintain mean arterial pressure in the setting of a low-cardiac-output state, SVR can vary widely in patients with CS and is inappropriately low in a substantial proportion of patients, likely driven by cytokine cascades triggered in the setting of severe CS (ie, SCAI stage D or E). Temporary MCS devices can trigger the inflammatory response and can worsen the vasodilatory CS. In addition, other factors may contribute to distributive physiology in some patients with CS, including vasodilatory medications (eg, sedatives) or an infection complicated by sepsis. The Writing Committee recognizes a need to identify standardized clinical, hemodynamic, and possibly laboratory biomarkers to identify this group; currently, a combination of SVR (when available) and clinical assessment is reasonable.^29-31

It may also be appropriate to separately identify shock that is transient (eg, in the setting of hemodynamic instability during a procedure). Finally, the development of CS in the setting of cardiac arrest, or in its extreme manifestation as resuscitation with extracorporeal cardiac life support, identifies a subpopulation of patients with a discrete natural history. This additional modifier would apply to all categories of CS, including AMI-CS.

Considering that patients with cardiac arrest have poor prognoses and high mortality risk,³² clinical trials focusing on patients with out-of-hospital cardiac arrest with the use of extracorporeal cardiopulmonary resuscitation should be performed separately.³³ Additionally, the presence of shockable rhythms, duration of cardiopulmonary resuscitation required for obtaining return of spontaneous circulation, age, and other comorbidities must be taken into account while recruiting these patients into clinical trials. Excluding patients with previous resuscitation would cut enrollment by approximately one-half, and exclusion of these patients would lead to a lower risk of the patient group included. Exclusion of these patients would also limit generalizability of the results. Furthermore, even those with resuscitation before randomization have a roughly 50% survival rate at 30 days.³⁴ In addition, if the decision is made to include resuscitated patients, pre-defined subanalyses should be performed on those with and without resuscitation.

HUMAN SUBJECT ISSUES IN TRIALS SEEKING TO ENROLL PATIENTS WITH STEMI AND CS

CS can be associated with altered mental status, sedation, and endotracheal intubation, which can restrict patients’ ability to provide consent for enrollment in clinical research. In addition, treatment of stage C or higher is likely time-sensitive and should be initiated as soon as technically feasible, so there may not be time to obtain consent from the subject’s legally authorized representative before initiating treatment, which may preclude clinical trial enrollment.

Previous trials of interventions in patients with CS in the US have required consent before the onset of symptoms³⁵ or have restricted enrollment to those who can provide written consent before enrollment.¹⁶ Because the prognosis of CS depends on the severity of illness at presentation,^36,37 trial results obtained in consenting subjects likely do not generalize to those when consent cannot be obtained before treatment. Therefore, research in shock outcomes restricted to consenting subjects alone would create a highly biased cohort from which to generate meaningful benefit–risk safety profiles of MCS devices. To reflect clinical practice and provide device benefit–risk data that are generalizable, shock trials enrolling patients with STEMI-CS or NSTEMI-CS must enroll both patients who can provide consent and those who cannot. Acute CS studies that include patients unable to provide informed consent could qualify for conduct under exception from informed consent for emergency research.³⁸ Before initiation in the United States, such a trial must meet specific FDA criteria for justification of exception from informed consent as described in Table S2.³⁸

After justification of exceptions from informed consent, investigators must consult with representatives of, and publicly disclose to, communities in which the trial will be conducted and from which subjects will be drawn. Approval from the institutional review board and FDA must also be obtained.

The consensus of the expert panel was that the following is a reasonable approach to implementation of a trial enrolling patients with STEMI and CS under exception from informed consent (Figure 4). Patients who are otherwise eligible for enrollment, or their immediately available legally authorized representative, should be briefly interviewed to assess their capacity to provide prospective written consent. If consent can be reliably and expeditiously obtained, the patient should be enrolled. If the patient lacks capacity, they should be given an opportunity to object to enrollment in response to a brief, structured verbal script. If consent is declined or the patient objects, they should not be enrolled. If the patient lacks capacity and their legally authorized representative is not available, they can be enrolled under exception from informed consent. After enrollment under exception from informed consent, a patient or their legally authorized representative must be notified of enrollment as soon as feasible and given an opportunity to opt out of ongoing enrollment.

Figure 4. Framework for enrollment with written consent or under exception from informed consent.

The Writing Committee is mindful that other approaches to consent are utilized in countries with different regulations.^17,39

OUTCOME DEFINITIONS

Previous ARC initiatives have dealt with outcomes that are typically assessed in shock research; consequently, the SHARC Writing Committee leveraged as much as possible from existing ARC definitions. When necessary, shock-related adaptations or additions are proposed and highlighted as modifications. If not otherwise stated, the existing ARC definitions are recommended in SHARC without change. Further, SHARC has prioritized outcomes that will most likely be utilized in shock research either as primary or secondary outcomes. Death, HF, neurological events, bleeding events, and MI are all considered high-priority outcomes for shock interventions. Death further classified, and preferably adjudicated, as cardiovascular and non-cardiovascular is recommended, following the recommendations of the Standardized Data Collection for Cardiovascular Trials Initiative (Table S3).⁴⁰ These recommendations have similarly been adopted for coronary and valvular research.^11,41 Importantly, in device trials, the relatedness of death to a procedure, device, or shock should be prospectively ascertained.⁴² SHARC MI definitions as (1) enrollment criterion and (2) postprocedural MI follow the recommendations of Valve Academic Research Consortium 3 (VARC-3), which leverages from the fourth universal definition of MI, as well as the SCAI and ARC-2 definitions (Table S4).^41,43,44 Type-5 MI, or non-percutaneous coronary intervention–related procedural MI, should likewise be classified based on VARC-3 recommendations for any non-percutaneous coronary intervention–related cardiac intervention. Neurologic events were recently standardized in the Neurological Academic Research Consortium.⁴⁵ Nevertheless, the granularity described in the Neurological Academic Research Consortium may exceed the scope necessary in shock research, and SHARC recommends following the VARC-3 classification (Table S5)⁴¹ because it is more pragmatic for shock evidence. SHARC classifies stroke severity according to the National Institutes of Health Stroke Scale and subsequent degree of disability according to the modified Rankin scale.

Bleeding events are paramount in high-risk interventions or in interventions in high-risk patients, both of which are present in shock research. We recommend using the classification proposed by the Mechanical Circulatory Support Academic Research Consortium because this classification (Table S6) is most applicable to this field and is modified from the Bleeding Academic Research Consortium.⁴⁶ Limited text adaptations have been applied to type-4 bleeding events to clearly signal that they are associated with any cardiovascular procedure and not limited to ventricular assist device implantation–related bleeding. Notably, hemorrhagic stroke is not classified as bleeding to avoid duplicate reporting. If fit for purpose to a specific study, hemorrhagic stroke could be included with other bleeding events for selected statistical (safety) analyses.

Hospitalization

Hospitalization is a marker of disease severity and progression,⁴⁷ as well as an indicator of quality of care. Thirty-day readmission rates are ≈20% for CS^48,49 and 28% for CS supported with extracorporeal life support,⁵⁰ with cardiovascular causes predominating.^48,49 Two-thirds of CS admissions are caused by decompensated HF,^18,51 and around one-fifth of readmissions in a de novo CS cohort are caused by HF.⁴⁸ Reflecting the burden of HF in the CS population, the SHARC Writing Committee felt that hospitalization should be broadly classified to align with HF ARC definitions, including the classification of death within 24 hours of hospitalization as death without associated hospitalization, because a hospitalization event requiring a hospital stay ≥24 hours is measured by a change in calendar date.¹⁰ Further, because readmission may be influenced by the duration of the index hospitalization, the SHARC Writing Committee recommended that reporting days alive out of hospital could also be a valuable metric.Table S7 defines the suggested subclassification of hospitalization events. As previously described in the HF ARC document, there is a need to determine adequate medical and device therapies in patients with heart failure for comparing the adequacy of different therapies across different trials.¹⁰ Additionally, it is also important to determine the patients who recover with native heart survival. Transition to heart transplantation and durable MCS in refractory CS are key therapeutic outcomes that vary according to the etiology of shock,^5,52,53 and should be collected and analyzed separately from all-cause mortality. For SHARC, they are defined as follows:

1. Heart transplant: Patient undergoes a surgical heart transplant procedure, with timing specified by the date of surgical anastomosis.

2. Durable MCS: Patient undergoes implantation of a durable MCS device including left ventricular assist device, otherwise described as a mechanical circulatory device, and can be cared for outside of the hospital setting.

Timing Classification and Reporting of Time Points

Figure S1 summarizes the recommended timing classification for SHARC evidence. Clinical outcomes within 30 days or up to discharge are defined as periprocedural (for research focused on an initial procedure), outcomes between 30 days and 1 year as intermediate, and anything beyond 1 year is considered late.

Patient-Centered Outcomes

More than in other areas of cardiovascular research, patient-centered outcomes play a pivotal role in shock research. Indeed, there are a growing number of studies suggesting that patient-reported outcomes may conflict with clinician-reported outcomes,^54-56 with patient-reported outcomes potentially being superior in reporting clinically meaningful changes in health status through time.⁵⁷ Because of these potential discrepancies, future trial design panels should consider inclusion of patients who have survived CS.⁵⁸

Leveraging work on core outcomes from critical care and extracorporeal membrane oxygenation (ECMO),^59-61 the SHARC Writing Committee recommends that at least four domains of patient-reported outcomes be strongly considered for incorporation into shock research: health-related quality of life, activities of daily living, disability, and days alive out of hospital. Simple, reliable, and validated instruments are available for each of these domains, specifically the Euro-QOL-5D-5L for health-related quality of life, Lawton’s instrumental activities of daily living questionnaire for activities of daily living (Table S8), and modified Rankin scale for disability (Table S9). The Euro-QOL-5D-5L is a self-reported health-related quality of life questionnaire that has 5 health-related dimensions, each with 5 levels of response (Figure S2).^62,63 It takes approximately 2 minutes to complete and has been translated into multiple languages.

Results are displayed as Euro-QOL-5D-5L index value, with 1 indicating best quality of life, and the Euro-QOL Visual Analog Scale, with 100 indicating the best subjective health status.

The Lawton Instrumental Activities of Daily Living questionnaire is a self-report of physical function that has 8 items and takes about 2.5 minutes to complete.⁶⁴ The modified Rankin scale is a single-item global scale of function to evaluate disability that takes 5 to 15 minutes to complete.^65,66 Days alive and out of hospital, also known as hospital-free days, has been used in a number of cardiovascular trials in the past.^67,68 This outcome provides a summary measure of 2 patient-important health events and is easily calculated from administrative or electronic heath records, helping to balance feasibility of measurement with patient-centeredness in future shock trials.⁶⁹

SAFETY OUTCOMES

As previously mentioned, to ensure consistency of definitions across different ARC and guideline documents, we relied upon most of the previously available definitions.^41,46,70,71 The SHARC Writing Committee recognizes that adverse events might be related to the natural history of CS phenotype and stages, candidacy for advanced therapies, MCS devices utilized, or procedures performed for MCS insertion or concomitantly for shock (eg, percutaneous coronary intervention, coronary artery bypass grafting, right heart catheterization, and pacemaker wires, among others). Device escalation or concurrent use of more than one device is also relatively common, resulting in a substantial challenge of attributing an adverse event to a specific device. Furthermore, additional procedures are occasionally performed to address or prevent complications. For example, atrial septostomy might be performed to prevent refractory pulmonary edema or pulmonary hemorrhage in patients on venoarterial ECMO support, and the procedure might then result in additional complications. Thus, the SHARC Writing Committee strongly recommends independent, structured adjudication for appropriate classification of adverse events (device-related, procedure-related, or disease-specific), as well as for attribution of device relatedness. SHARC recognizes the challenge of adjudication of device relatedness when concurrent devices or serial devices are used. Support for such adjudication through capture of data and supporting source documentation by carefully designed clinical trial and registry data collection systems should be recognized as an essential feature for shock research.

The 2020 International Organization for Standardization 14155 defines “an adverse event” as an untoward medical occurrence, unintended disease or injury, or untoward clinical signs (including abnormal laboratory findings) in subjects, users or other persons, whether or not related to the investigational medical device.⁷² This definition has been adopted by international regulatory agencies, and as such, it is also recommended by SHARC. In addition, the SHARC Writing Committee recommends considering (when feasible), assessment of relatedness of procedure, device, and shock adverse events according to the definitions listed in Table 3.⁴² Major domains of safety outcomes relevant to patients with cardiogenic shock are listed in Table S10.

Access-Related Complications

Vascular access complications remain important procedure-related adverse events in patients undergoing MCS for CS. The reported incidence of adverse events such as limb ischemia, arterial thrombosis, and vascular injury requiring surgery has been variable across different case series.⁷³ Differences in device characteristics and access site, operator experience, sample size in reported case series, and definitions have all likely contributed to the notable variability in overall incidence reported in the current literature. In line with the recently released VARC-3 outcome definitions, the SHARC Writing Committee recommends including in the classification both major and minor vascular complications (Table S11),^41,70 and the category of access-related nonvascular complications (Table S12). In addition, because procedures such as atrial septostomy to vent the left ventricle or the insertion of percutaneous left atrial cannula can result in adverse events including cardiac perforation and pericardial tamponade, structural cardiac complications have been included in this broad category of access-related complications.

Hematologic Events

Mild hemolysis from shear stress forces on red blood cells is a relatively common occurrence with some mechanical support devices.^74-76 Hemolysis can also be severe and sudden in the setting of device thrombus formation^{74 77} and in the setting of “chugging,” which is characterized by intermittent occlusion of a suctioning cannula occurring from transient venous or atrial collapse. Definitions of major and minor hemolysis mirroring the definitions of the Mechanical Circulatory Support Academic Research Consortium are listed in Table S13.⁴⁶ It is recognized that there are numerous possible causes other than hemolysis for serum lactate dehydrogenase elevation. Therefore, this abnormality alone is not very specific, and should be followed by an additional confirmatory assessment.

Infection

Infections are an inherent risk of prolonged use of MCS devices; they include access-site infections, infective endocarditis, bloodstream infection, and sepsis (Table S14). It cannot be emphasized enough that there are challenges in the appropriate diagnosis of sepsis in the setting of CS and limitations in the use of available definitions. For example, according to the Third International Consensus Definitions for Sepsis, sepsis can be defined as a “life-threatening organ dysfunction caused by a dysregulated host response to infection. For clinical operationalization, organ dysfunction can be represented by an increase in the Sequential [Sepsis-related] Organ Failure Assessment (SOFA) score of 2 points or more.”⁷⁸ The definition is based on changes in SOFA score, which is a measurement of respiratory, coagulation, liver, cardiovascular, renal, and neurologic organ dysfunction. Given that the SOFA score can be affected by CS independent of whether sepsis is present, the use of this definition might be problematic. The same concerns apply to the definition adopted by SHARC and listed in Table S6, which mirrors the Interagency Registry for Mechanically Assisted Circulatory Support and Mechanical Circulatory Support Academic Research Consortium definitions.

Organ Dysfunction

Tables S2, S5, and S14 list definitions of neurologic, hepatic, renal, pulmonary, gastrointestinal, and infection-related organ systems dysfunction.^46,79,80 SHARC encourages collection of the outcomes listed in these tables according to the recommended definitions, with the following considerations: (1) organ dysfunction is usually a manifestation of the natural course of CS, and therefore, the development of organ dysfunction should not necessarily be attributed to a device- or treatment-related adverse event; (2) temporary MCS devices are often used as a bridge to allow organ function improvement before implantation of a durable left ventricular assist device or as a bridge to full recovery in patients with CS and multisystem organ failure.^81-83 Therefore, serial collection of parameters related to organ dysfunction is recommended as a potential outcome to assess effectiveness of pharmacologic and nonpharmacologic interventions in CS (Table 4).

Table 4.

Organ Dysfunction

Injury	Criteria
Hepatic dysfunction/liver injury46,80	Increase to a level >3× the upper limit of normal in any 2 hepatic laboratory values (ie, total bilirubin, AST, and ALT) Impaired liver function (ie, jaundice and INR >1.5)
AKI41
Stage 1	AKI with ≥1: Increase in serum creatinine >150–200% (≥1.5- to 2.0-fold increase) within 7 days compared with baseline Increase ≥0.3mg/dL (≥26.4 mmol/L) within 48 h of index procedure
Stage 2	AKI with increase in serum creatinine >200–300% (>2.0- to 3.0-fold increase) within 7 days compared with baseline
Stage 3	AKI with ≥1: Increase in serum creatinine >300% (>3.0-fold increase) within 7 days compared with baseline Serum creatinine ≥4.0 mg/dL (≥354 mmol/L) with an acute increase ≥0.5 mg/dL (≥44 mmol/L)
Stage 4	AKI requiring new temporary or permanent renal replacement
Pulmonary hemorrhage	Development of hemoptysis or of blood collection in the endotracheal tube, and with diffuse alveolar infiltrates or diagnosis by bronchoscopy
Refractory pulmonary edema	Persistent, diffuse alveolar infiltrates with hypoxemia and evidence of elevated pulmonary capillary wedge pressure
Mesenteric ischemia79	Development of unexplained abdominal distension, gastrointestinal bleeding, and abdominal pain, with abdominal computed tomography findings demonstrating bowel ischemia and free fluid

AKI indicates acute kidney injury; ALT, alanine transaminase; AST, aspartate transaminase; and INR, international normalized ratio.

Device Malfunction

Table S15 lists a recommended classification of device malfunctions and corresponding definitions within the context of MCS. The classification and definitions follow the general classification and definitions of the Mechanical Circulatory Support Academic Research Consortium,⁴⁶ but with adaptations to focus on acute management of CS and without definitions of device malfunction events that are specific to durable ventricular-assist devices. Accidental decannulation and device cannula migrations are rare occurrences and have been included in the list of potential device malfunctions.

CONCLUSIONS

The SHARC expert consensus definitions are intended to develop more informative consistency across bodies of evidence on therapeutic benefit–risk and safety of interventions, particularly involving cardiovascular devices such as MCS for CS. The academic research community is encouraged to use these definitions and outcomes in registries and when performing clinical trials as a pathway in particular to the most robust determination of best practice guidelines. For regulatory evidence supporting better and safer innovative devices reaching patient bedsides more quickly, 15 years of ARC definitions have unequivocally proven the critical value of utilization of consistent, pragmatic definitions across many cardiovascular areas of need, and they are almost certain to have the same impact on shock device outcomes. The goal of this multistakeholder collaboration involving academia, industry representatives, regulators, and payers was to help provide consensus definitions, including outcomes and safety, with the goal of performing better trials.

Nonstandard Abbreviations and Acronyms

ALT

alanine transaminase

AMI

acute myocardial infarction

ARC

Academic Research Consortium

CS

cardiogenic shock

ECMO

extracorporeal membrane oxygenation

FDA

Food and Drug Administration

HF

heart failure

INTERMACS

Interagency Registry for Mechanically Assisted Circulatory Support

MCS

mechanical circulatory support

MI

myocardial infarction

NSTEMI

non–ST-elevation myocardial infarction

SCAI

Society for Cardiovascular Angiography and Interventions

SHARC

Shock Academic Research Consortium

SOFA

Sequential Organ Failure Assessment

STEMI

ST-segment elevation myocardial infarction

VARC-3

Valve Academic Research Consortium 3