Epidemiology of cardiogenic shock using the Shock Academic Research Consortium (SHARC) consensus definitions

Abstract

Introduction

Cardiogenic shock (CS) is a heterogeneous syndrome with important variation in aetiology,¹ chronicity,² end-organ injury,³ haemodynamic profile,^4,5 and severity.^6,7 One hypothesized explanation for the challenges with establishing a benefit of many studied interventions in CS is that trials have inadequately accounted for this heterogeneity.⁸ To this end, there have been efforts to standardize grading of CS severity [e.g. Society for Cardiovascular Angiography and Intervention (SCAI) shock stage] and to develop a common taxonomy for describing CS subtypes and outcomes.

The Shock Academic Research Consortium (SHARC) is a multi-stakeholder group that developed consensus definitions for classifying CS populations in clinical trials and registries.⁹ The SHARC framework has not yet been applied in a contemporary registry. The aim of this study was to describe contemporary CS epidemiology using the SHARC classifications.

Methods

Data collection

The Critical Care Cardiology Trials Network (CCCTN) is a multinational network of cardiac intensive care units (CICUs) coordinated by the TIMI Study Group (Boston, MA). The CCCTN Registry methods have been reported.¹⁰ From 2017–23, participating centres contributed consecutive medical CICU admissions during annual 2-month collection periods. The CCCTN Registry protocol and waiver of informed consent were approved by the IRB at each centre.

Shock classification

Patients were classified according to their initial presenting shock state. Consistent with the SHARC definition, CS was defined clinically in the CCCTN Registry as a cardiac disorder that results in haemodynamic impairment [i.e. SBP < 90 mmHg for ≥30 min or the need for vasopressors, inotropes, or mechanical circulatory support (MCS) to maintain SBP ≥ 90 mmHg] with evidence of hypoperfusion (see Supplemental Methods). Primary aetiologic categories included acute myocardial infarction-related CS (AMI-CS), heart failure-related CS (HF-CS), and non-myocardial (secondary) CS (e.g. cardiac tamponade and severe valvular disease). Heart failure-related CS was subcategorized as de novo (i.e. related to acute myocardial dysfunction known or suspected to be new) vs. acute-on-chronic (i.e. decompensation of chronic HF). Mixed CS classification is described in the Supplemental Methods.

Data analysis

Standardized data elements (e.g. demographics, medical history, and presenting clinical characteristics), shock-specific variables [e.g. whether CS occurred in the setting of cardiac arrest (SHARC modifier) and SCAI stage], and in-hospital mortality were summarized according to profile and aetiology. Results are reported as counts and percentages for categorical variables and medians (25th–75th percentiles) for continuous variables. Absolute 95% confidence intervals (CI) were calculated using the binomial method.

Results

Shock classification

Among the 8974 patients meeting shock criteria, 65% had CS and 17% had mixed shock (Figure 1). Among patients with CS (n = 5869), 27% had AMI-CS (65% STEMI), 59% had HF-CS (72% acute-on-chronic, 28% de novo), and 14% had secondary CS. Leading causes of secondary CS (n = 799) included arrhythmia (33%), severe valvular disease (27%), pericardial disease (9%), and pulmonary vascular disease (8%). Among patients with confirmed mixed CS (n = 947; 2019–23), comprising 93% of all mixed shock, 19% of cases were related to AMI (53% STEMI). Five per cent of isolated CS cases were known to have evolved to mixed CS, and 3% of mixed CS cases to isolated CS.

Figure 1

Shock profile and aetiology in a cardiac intensive care unit (CICU) population. Shock cases were classified according to primary haemodynamic profile. Cardiogenic shock (CS) cases were sub-classified according to primary aetiology. Mixed shock indicates that more than one shock category substantially contributed to the haemodynamic profile, and mixed CS refers to mixed shock cases with a known cardiogenic component (∼93% of all mixed shock). Cases for which the shock category was uncertain (n = 188; 2% of all shock cases) are not included in the figure. AMI-CS, acute myocardial infarction-related cardiogenic shock; CS, cardiogenic shock; HF-CS, heart failure-related cardiogenic shock.

Open in new tabDownload slide

Clinical characteristics and cardiac intensive care unit resource utilization by shock aetiology

Clinical characteristics and CICU resource utilization are summarized by shock aetiology in Table 1. Patients with AMI-CS and de novo HF-CS were most likely to have had CS in the setting of cardiac arrest (33% and 31%, respectively).

The majority of isolated and mixed CS was SCAI stage C (56% and 51%, respectively) with variation by aetiology (P < 0.001; Table 1 and Supplementary material online, Figure S1). Among CS subtypes, patients with AMI-CS and mixed CS were most likely to present in SCAI stage D or E. Patients with AMI-CS and mixed CS also had the highest SOFA scores and vasoactive medication requirements (Table 1). Biventricular failure was most common in acute-on-chronic HF-CS, and this group tended to have the lowest presenting LVEF.

The use of advanced CICU therapies was common (Table 1). Temporary MCS use was highest in AMI-CS (59%). Rates of mechanical ventilation were highest in AMI-CS and mixed CS (65% and 67%, respectively). Rates of acute renal replacement therapy were similar across isolated CS categories (11–15%), but higher in patients with mixed CS (24%).

In sensitivity analyses pooling isolated and mixed CS, the distribution of CS aetiologies was unchanged (see Supplementary material online, Figure S2 and Supplementary material online, Table S1). Mixed shock syndrome (as defined in SHARC) was most common in de novo HF-CS (see Supplementary material online, Table S2).

Mortality by shock aetiology

In-hospital mortality was higher in patients with mixed CS (48%; 95% CI, 45–51%) as compared to those with isolated CS (P < 0.001). Among patients with isolated CS, mortality was highest in patients with AMI-CS (41%; 95% CI, 39–44%), similar in patients with de novo HF-CS (31%; 95% CI, 28–34%) and secondary CS (31%; 95% CI, 28–34%), and lowest in patients with acute-on-chronic HF-CS (25%; 95% CI, 23–26%; Figure 2). SCAI staging identified stepwise gradients of in-hospital mortality rates across each SHARC CS category (see Supplementary material online, Figure S3).

Figure 2

In-hospital mortality by shock profile and aetiology. Error bars indicate the 95% confidence intervals around the in-hospital mortality estimates for each shock category. AMI-CS, acute myocardial infarction-related cardiogenic shock; CS, cardiogenic shock; HF-CS, heart failure-related cardiogenic shock.

Open in new tabDownload slide

Discussion

In this analysis from the CCCTN Registry, we applied the SHARC definitions for CS classification, depicting the contemporary landscape of CS and the utility of the SHARC framework for identifying CS subtypes with distinct clinical profiles and outcomes. Building on our previous work^1,2 but with an updated much larger sample size of shock patients (n = 8974), we found that HF-CS is the most common category of CS in contemporary CICUs, and that HF disease chronicity (acute-on-chronic vs. de novo HF-CS) is associated with distinct patterns of severity, end-organ injury, and outcomes. In addition, we showed that presenting SCAI shock stage varies across CS categories, with AMI-CS and mixed CS patients presenting with more advanced SCAI shock stages than HF-CS and secondary CS patients. Owing to the large size of our dataset, we were able to estimate in-hospital mortality rates for CS categories more precisely than has been done in previous studies.^1,11 We found that patients with mixed CS had the highest in-hospital mortality, followed by those with AMI-CS, who in turn had significantly higher mortality than those with HF-CS or secondary CS.

The SHARC definitions thus provide a standardized framework for classifying clinically important CS phenotypes with distinct clinical features and prognoses. These foundational epidemiologic data are not only relevant to clinical practice (e.g. when counselling patients and families) but also provide benchmarks that may be pertinent to quality improvement efforts and to future clinical trial design and interpretation. For instance, the differing results of recently completed trials of temporary MCS highlight the potential relevance of cardiogenic shock phenotype selection in testing specific therapies.^12,13

Limitations of these analyses should be acknowledged. First, since CCCTN is a registry of medical CICU admissions, we could not include post-cardiotomy CS. Moreover, they reflect the epidemiology of CS in medical CICUs, which may be different from other practice settings. Second, since full details on mixed shock presentations were not collected in early annual cycles, descriptions of our mixed CS population were restricted to 2019–23. Finally, clinical assessment of systemic hypoperfusion can be challenging, particularly in the setting of a missing or normal serum lactate; although it is possible that some patients without shock were included, shock assessments were performed by trained investigators applying criteria aligned with the SHARC definitions.

Conclusions

SHARC consensus definitions for CS classification can be pragmatically applied in contemporary registries and reveal discrete subpopulations of CS with distinct phenotypes and outcomes.

Supplementary material

Supplementary material is available at European Heart Journal: Acute Cardiovascular Care online.

Funding

A.D.T. is supported by National Institutes of Health-National Heart, Lung, and Blood Institute (K08HL163328).

Data availability

We encourage parties interested in collaboration and data sharing to contact the corresponding author.

References

Author notes