Early Utilization of Mechanical Circulatory Support in Acute Myocardial Infarction Complicated by Cardiogenic Shock: The National Cardiogenic Shock Initiative

Abstract

Background

Acute myocardial infarction complicated by cardiogenic shock (AMI‐CS) is associated with significant morbidity and mortality. Mechanical circulatory support (MCS) devices increase systemic blood pressure and end organ perfusion while reducing cardiac filling pressures.

Methods and Results

The National Cardiogenic Shock Initiative (NCT03677180) is a single‐arm, multicenter study. The purpose of this study was to assess the feasibility and effectiveness of utilizing early MCS with Impella in patients presenting with AMI‐CS. The primary end point was in‐hospital mortality. A total of 406 patients were enrolled at 80 sites between 2016 and 2020. Average age was 64±12 years, 24% were female, 17% had a witnessed out‐of‐hospital cardiac arrest, 27% had in‐hospital cardiac arrest, and 9% were under active cardiopulmonary resuscitation during MCS implantation. Patients presented with a mean systolic blood pressure of 77.2±19.2 mm Hg, 85% of patients were on vasopressors or inotropes, mean lactate was 4.8±3.9 mmol/L and cardiac power output was 0.67±0.29 watts. At 24 hours, mean systolic blood pressure improved to 103.9±17.8 mm Hg, lactate to 2.7±2.8 mmol/L, and cardiac power output to 1.0±1.3 watts. Procedural survival, survival to discharge, survival to 30 days, and survival to 1 year were 99%, 71%, 68%, and 53%, respectively.

Conclusions

Early use of MCS in AMI‐CS is feasible across varying health care settings and resulted in improvements to early hemodynamics and perfusion. Survival rates to hospital discharge were high. Given the encouraging results from our analysis, randomized clinical trials are warranted to assess the role of utilizing early MCS, using a standardized, multidisciplinary approach.

Nonstandard Abbreviations and Acronyms

CS

cardiogenic shock

IABP

intra‐aortic balloon pump

MCS

mechanical circulatory support

NCSI

National Cardiogenic Shock Initiative

SBP

systolic blood pressure

Clinical Perspective.

What Is New?

• Early use of mechanical circulatory support in acute myocardial infarction and cardiogenic shock is feasible across varying health care settings and resulted in improvements to early hemodynamics and perfusion.

What Are the Clinical Implications?

• Use of a shock protocol and multidisciplinary team facilitated early delivery of mechanical circulatory support as an adjunct to coronary revascularization, which may lead to improved clinical outcomes.

Cardiogenic shock (CS) is a state of circulatory failure characterized by hemodynamic collapse and end organ hypoperfusion. CS complicates ~5% to 10% of patients who present with an acute myocardial infarction (AMI). ¹ The cornerstone of therapy for AMI‐CS is early revascularization, which improves survival. ² , ³ Despite continued development and growth of ST‐elevation myocardial infarction systems of care, CS remains a challenging condition for clinicians to manage and is associated with high morbidity and mortality. ² , ³ , ⁴

Mechanical circulatory support (MCS) devices may provide therapeutic benefits in patients with AMI‐CS. MCS devices serve as an adjunctive therapy to primary percutaneous coronary intervention (PCI) by providing hemodynamic stability, reducing intracardiac filling pressures, and improving end organ perfusion while alleviating the need for increasing doses of vasopressors and inotropes. ⁵ , ⁶

Intra‐aortic balloon pump (IABP) counter‐pulsation is the only form of MCS to be adequately studied in a well‐powered randomized clinical trial (RCT). Unfortunately, its use has not been demonstrated to improve survival. ⁷ IABPs primarily function by decreasing afterload, providing 0.5 to 1 L/min of cardiac output and thereby limiting their efficacy in AMI‐CS. ⁸

We hypothesize that use of a more robust form of MCS, such as Impella, may lead to improved clinical outcomes and survival. Previous work from the catheter‐based ventricular assist devices (cVAD) registry demonstrated that early use of Impella, particularly pre‐PCI and before escalating the number of vasoactive agents, was associated with improved survival. ⁹ We therefore created a shock protocol that emphasized the early use of MCS with Impella, guided by invasive hemodynamics, and assessed the feasibility of using such an approach and its effect on early hemodynamics, perfusion, and in‐hospital survival.

METHODS

Study Sites

The data that support the findings of this study are available from the corresponding author upon reasonable request. The NCSI (National Cardiogenic Shock Initiative) is a single‐arm, multicenter study designed to assess outcomes associated with the early use of MCS in patients with AMI‐CS. The study was completed in 2 phases; a phase 1 pilot study was termed the “Detroit Cardiogenic Shock Initiative,” which involved 4 hospitals in metropolitan Detroit. ¹⁰ After local feasibility was demonstrated, the study was expanded to a phase 2 study that included 80 sites throughout the United States and was referred to as the “National Cardiogenic Shock Initiative” (NCSI; NCT03677180). The primary outcome of the study was to assess feasibility of using a shock protocol and assessing in‐hospital survival. Secondary outcomes included assessing the effect of MCS on early hemodynamics, perfusion, and 30‐day and 1‐year survival.

Of the 80 participating sites, 48 hospitals identified as community‐based and 32 as academic (Figure 1). Given the variability of adoption in the use of Impella, site principal investigators were requested to ensure that (1) the institution placed at least 10 devices per year; (2) there was broad adoption of the protocol, defined as acceptance of the protocol by 80% of interventional cardiologists; and (3) there was approval of the protocol by a multidisciplinary care team within the hospital.

Figure 1. Study trial sites, including 32 academic centers and 48 community centers.

NCSI indicates National Cardiogenic Shock Initiative.

Adherence to the protocol was voluntary. Deviation from the treatment algorithm could occur at the discretion of the primary clinicians without consultation of the primary investigators. All patients, including those with deviations from the suggested treatment protocol, were included in the study if they met the inclusion criteria and did not meet the exclusion criteria (Table S1).

Definitions

Patients were included if they presented with an AMI‐CS and underwent PCI and MCS with Impella during the index procedure.

AMI (ST‐elevation myocardial infarction or non–ST‐elevation myocardial infarction) was defined as electrocardiographic changes indicative of new ischemia (ST‐T changes), detection of elevated cardiac biomarkers, and angiographic findings of an infarct‐related artery on coronary angiogram in the presence of ischemic symptoms.

CS was defined as the presence of at least 2 of the following: (1) prolonged hypotension (ie, systolic blood pressure [SBP] <90 mm Hg, or the use of vasopressors or inotropes to maintain an SBP >90 mm Hg); (2) signs of end organ hypoperfusion (ie, altered mentation, cool extremities, oliguria or anuria, or elevated lactate levels); or (3) poor invasive hemodynamics (ie, cardiac index <2.2 L/min per m² or cardiac power output <0.6 watts). ¹¹ Although the Society for Cardiovascular Angiography and Intervention shock classification system was not created at the start of our study, it provides an important standard definition for CS and was adopted by our group retrospectively. ¹² All patients met the definitions of Society for Cardiovascular Angiography and Intervention shock stage C, D, or E. Stage E patients included 3 distinct patient cohorts: (1) active cardiopulmonary resuscitation at the time of MCS insertion, (2) lactate >10 mmol/L, and (3) lactate between 5 and 10 mmol/L who also required multiple vasopressors and inotropes, or those with recent preprocedural cardiac arrest.

Inclusion and exclusion criteria are listed in Table S1 and were used to allow for meaningful comparison to prior RCTs. The protocol is shared in Figure 2 and detailed further in Data S1.

Figure 2. Suggested treatment protocol for patients presenting with acute myocardial infarction and cardiogenic shock.

AMICS indicates acute myocardial infarction and cardiogenic shock; CO, cardiac output; CPO, cardiac power output; Cr, creatinine; CTO, chronic total occlusion; dPAP, diastolic pulmonary artery pressure; ECG, electrocardiogram; LVAD, left ventricular assist device; MAP, mean arterial pressure; MCS, mechanical circulatory support; MV, multivessel; NSTEMI, non–ST‐elevation myocardial infarction; PA, pulmonary artery; PAPI, pulmonary artery pulsatility index; PCI, percutaneous coronary intervention; RA, right atrial; RHC, right heart catheterization; RV, right ventricular; SBP, systolic blood pressure; sPAP, systolic pulmonary artery pressure; STEMI, ST‐elevation myocardial infarction; and TIMI, thrombolysis in myocardial infarction.

Data Collection and Statistical Analysis

Data were collected retrospectively from medical records of patients, and screening was conducted according to local practice. Missing data were not analyzed, and imputation was not utilized. Where appropriate, 30‐day and 1‐year survival status was collected by contact with the patient, patient surrogate, or through medical records review. Continuous variables were described using the mean and standard deviation. Categorical variables were described with frequency and percentage. T test was used for 2‐group comparison, and ANOVA was used for multiple groups comparison. The Kruskal–Wallis test was used to compare medians between 2 groups. The χ² test or Fisher exact test was used for categorical variables, as appropriate. All statistical tests or CIs were performed with a significance level of α<0.05. Multivariable logistic regression was used to assess the effect of pre‐MCS variables on in‐hospital mortality, and a generalized estimating equations model was used to analyze repeat measures. Both models utilized a stepwise backward selection and a selection criterion using a P value <0.1

Regulatory Approval

Institutional review board approval was obtained at the participating sites according to institutional requirements (participating sites are listed in Data S2). Most sites conducted the study under a waiver of consent, with HIPAA authorization for data collection and submission under the guidelines of 45 CFR 46.116(f) and 45 CFR 64.512(i) (2)(ii). In a minority of sites, consent was obtained from patients, patient surrogates, or capturing of deidentified data for patients who did not survive and would not require follow‐up, according to their local institutional review board requirements. The National Cardiogenic Shock Initiative was funded by research grants from Abiomed (Danvers, Massachusetts) and Chiesi Pharmaceuticals Inc (Cary, North Carolina). Neither company had direct involvement in the study design or the present analysis.

RESULTS

Patient Cohort

From 2016 to 2020, 406 patients with AMI‐CS were included in the study. Two hundred ninety‐five (72.7%) patients were classified as experiencing stage C/D shock at the time of MCS implantation, while 111 (27.3) were classified as stage E shock. The average age of the cohort was 63.7±12.3 years, 24% were female, 31% were non‐White, and 40% had diabetes. Sixty‐seven percent presented with CS on admission and 82% with ST‐elevation myocardial infarction. Twenty‐six percent were transferred from another hospital, and 16% of these patients had MCS placed before transfer. Forty‐six percent experienced cardiac arrest. Patients in stage E shock were more likely to experience out‐of‐hospital and in‐hospital cardiac arrest. Patient demographics and admission characteristics are listed in Table 1.

Table 1.

Patient Demographics and Admission Characteristics

	All (N=406)	Stage C/D (N=295)	Stage E (N=111)	P value
Demographics
Age, y	63.7±12.3	63.5±12.5	64.2±11.8	0.66
Sex—Female (%)	24% (96)	25% (73)	21% (23)	0.40
Race
White	68% (277)	71% (209)	61% (68)	0.06
Black	8% (34)	9% (26)	7% (8)
Other	23% (95)	20% (60)	32% (35)
Diabetes	40% (160)	37% (106)	51% (54)	0.01
Cerebrovascular disease	9% (37)	8% (24)	12% (13)	0.25
Renal insufficiency	13% (50)	12% (35)	14% (15)	0.60
ESRD	4% (15)	4% (11)	4% (4)	0.98
Congestive heart failure	23% (88)	24% (67)	20% (21)	0.41
Prior myocardial infarction	20% (79)	19% (55)	23% (24)	0.43
Prior PCI	24% (97)	25% (73)	22% (24)	0.52
Prior CABG	6% (24)	5% (15)	8% (9)	0.25
Admission characteristics
Patient transferred from another hospital	26% (106)	27% (79)	24% (27)	0.61
Support before transfer	16% (17)	15% (12)	19% (5)	0.74
Shock present on admission	67% (270)	64% (189)	74% (81)	0.07
Cardiac arrest	46% (188)	25% (103)	77% (85)	<0.01
Out‐of‐hospital arrest	17% (68)	14% (42)	23% (26)	0.03
In‐hospital arrest	30% (121)	22% (66)	50% (55)	<0.01
CPR at the time of MCS insertion	9% (37)	0% (0)	33% (37)	<0.01
STEMI	82% (333)	81% (238)	86% (95)	0.25

CABG indicates coronary artery bypass grafting; CPR, cardiopulmonary resuscitation; ESRD, end‐stage renal disease; MCS, mechanical circulatory support; PCI, percutaneous coronary intervention; and STEMI, ST‐elevation myocardial infarction.

MCS was implanted pre‐PCI in 70%, 9% received MCS during PCI, and 21% post‐PCI. A pulmonary artery catheter was utilized in 91% during the index procedure. PCI was performed using femoral access in 78%. The culprit vessel was most commonly the left anterior descending coronary artery (47%). Sixty‐six percent had multivessel coronary artery disease. Single vessel PCI was performed in the majority (61%), and 39% underwent multivessel PCI. The median door to support time in ST‐elevation myocardial infarction was 78 minutes, and median door‐to‐balloon time was 81 minutes. Ninety percent of patients presented with less than thrombolysis in myocardial infarction 3 flow, and 91% achieved thrombolysis in myocardial infarction 3 flow at the end of the procedure. Procedural characteristics are summarized in Table 2.

Table 2.

Procedural Characteristics

Characteristics	All (N=406)	Stage C/D (N=295)	Stage E (N=111)	P value
Impella insertion
Pre‐PCI	70% (285)	69% (201)	76% (84)	0.38
Intraprocedural	9% (35)	9% (27)	8% (7)
Post‐PCI	21% (84)	22% (65)	17% (19)
PAC insertion
Pre‐impella	23% (93)	24% (71)	20% (22)	0.46
Post‐impella	67% (270)	66% (190)	72% (80)
PAC not performed	9% (38)	10% (29)	8% (9)
Initial device used
Impella 2.5	5% (21)	5% (16)	5% (5)	0.71
Impella CP	92% (375)	91% (268)	96% (107)	0.06
Impella access
Femoral	98% (399)	98% (290)	98% (109)	0.71
Axillary	1% (5)	1% (4)	1% (1)
PCI access
Radial	20% (83)	25% (73)	9% (10)	<0.01
Femoral	78% (317)	74% (217)	90% (100)	<0.01
Thrombectomy used	27% (107)	28% (81)	24% (26)	0.47
Atherectomy used	5% (20)	4% (12)	7% (8)	0.19
Culprit vessel (n=343 culprit vessels)
Left main	13% (44)	11% (27)	17% (17)	0.08
Left anterior descending	47% (162)	49% (120)	43% (42)	0.60
Left circumflex	20% (68)	20% (49)	19% (19)	0.90
Right coronary artery	19% (64)	19% (46)	18% (18)	0.88
Ramus	1% (5)	1% (3)	2% (2)	0.52
Number of diseased vessels (>70% stenosis)
1 vessel	35% (140)	37% (107)	30% (33)	0.37
2 vessels	30% (118)	28% (81)	34% (37)
3 vessels	36% (142)	35% (102)	36% (40)
Number of vessels treated
1 vessel treated	61% (247)	63% (184)	63% (57)	0.60
2 vessels treated	30% (120)	29% (84)	36% (33)
3 vessels treated	9% (36)	9% (25)	11% (10)
Number of stents placed*	2 [1,2]	1 [1,2]	2 [1,3]	0.07
Door‐to‐balloon time in STEMI, min*	81 [34–247]	79 [46–246]	85 [52–,244]	0.40
Door‐to‐support time in STEMI, min*	78 [41–237]	80 [41–238]	70 [43–226]	0.60
TIMI flow pre‐PCI
0	72% (277)	74% (204)	68% (73)	0.38
1	11% (41)	9% (25)	15% (16)
2	9% (35)	9% (24)	10% (11)
3	8% (30)	8% (22)	7% (8)
TIMI flow post PCI
0	1% (4)	1% (3)	1% (1)	0.66
1	2% (6)	2% (5)	1% (1)
2	7% (27)	6% (17)	10% (10)
3	91% (356)	91% (259)	89% (97)

PAC indicates pulmonary artery catheter; PCI, percutaneous coronary intervention; STEMI, ST‐elevation myocardial infarction; and TIMI, thrombolysis in myocardial infarction.

^*Median [interquartile range].

Hemodynamics and Laboratory Trends

Mean SBP was 77.2±19.2 mm Hg, with stage E patients having lower SBP. Patients frequently presented with multiorgan dysfunction, characterized by elevated creatinine and liver function tests. Mean admission lactate was 4.8±3.9 mmol/L but was significantly higher in stage E shock. Initial hemodynamics demonstrated congestion with elevated central venous pressures (15.9±6.5 mm Hg) and left ventricular end diastolic pressures (31.5±24.8 mm Hg) and decreased cardiac power output (0.67±0.29 watts). At 24 hours, mean SBP improved to 103.9±17.8 mm Hg, lactate to 2.7±2.8 mmol/L, and cardiac power output to 1.0±1.3 watts. Patients in stage E shock had persistently lower BP, higher lactate levels, and a lower pulmonary artery pulsatility index. Hemodynamic trends within the first 24 hours are presented in Table S2 and Table 3.

Table 3.

Hemodynamic and Laboratory Trends Within the First 24 Hours

	Pre‐MCS	Post‐MCS	12 h	24 h	P value
HR, bpm	95.5±30.3	93.6±23.1	85.7± 17.9	88.5±19.0	<0.0001
SBP, mm Hg	95.3±26.5	113.8±24.0	105.4±19.8	103.9±17.8	<0.0001
DBP, mm Hg	61.6±19.8	77.9±19.4	72.4±14.1	66.9±12.2	<0.0001
MAP, mm Hg	72.8±21.4	90.2±19.4	82.9±14.8	78.6±12.3	<0.0001
LVEDP, mm Hg	31.5±24.8	n/a	n/a	n/a	n/a
dPA, mm Hg	24.9±7.5	23.5±8.8	19.9±7.1	18.1±5.5	<0.0001
CVP, mm Hg	15.9±6.5	14.4±6.3	11.2±5.2	10.5±4.4	<0.0001
CPO, W	0.67±0.29	0.88±0.44	0.83±0.34	0.98±1.32	<0.0001
PA saturation	59.5±15.8	60.4±13.8	62.9±16.2	63.9±15.8	<0.0001
CO	3.9±1.3	4.4±1.6	4.5 ± 1.7	5.2±1.6	<0.0001
CI	2.0±0.7	2.2±0.8	2.2±0.8	2.6±0.8	<0.0001
Number of inotropes/vasopressors	1.0±0.9	0.9±1.0	1.1±1.1	1.0±1.0	0.0142
Lactate, mmol/L	4.8±3.9	n/a	3.8±3.8	2.7±2.8	<0.0001

CI indicates cardiac index; CO, cardiac output; CPO, cardiac power output; CVP, central venous pressure; DBP, diastolic blood pressure; dPA, diastolic pulmonary artery pressure; HR, heart rate; LVEDP, left ventricular end diastolic pressure; MAP, mean arterial pressure; MCS, mechanical circulatory support; PA, pulmonary artery; and SBP, systolic blood pressure.

Procedural Outcomes

Eighty‐nine patients (24.4%) had a reported complication related to MCS. The most common complication reported was the need for blood transfusion (22.8%), with approximately half of these patients requiring >4 units of blood. Other common complications included hemolysis (7.4%), bleeding or hematoma at the access site (6.3%), other bleeding (4.1%), limb ischemia (3.0%), need for percutaneous or surgical intervention (2.7%), and stroke (1.9%). Other reported complications included deep venous thrombosis, infection, device migration, and kinking of the delivery sheath. It is important to highlight that complications were self‐reported, without adjudication, and may be underestimated.

The most common cause of death was worsening cardiogenic shock and multi‐organ failure. Of those who died, only a minority received MCS escalation (Figure 3). Fifty‐five patients received additional MCS after the index procedure, including extracorporeal membrane oxygenation (70%), reimplantation of an Impella CP after weaning and removal of the first device (14%), Impella 5.0 (9%), Impella RP (5%), Protek Duo (2%), or an unknown device (22%).

Figure 3. Cause of death and rates of mechanical circulatory support escalation among patients who died.

Procedural survival, survival to discharge, survival to 30 days, and survival to 1 year were 99%, 71%, 68%, and 53%, respectively. When stratified by shock stage, patients with stage C/D shock had 99%, 79%, 77%, and 62% survival, respectively. Those in stage E shock had 98%, 54%, 49%, and 31% survival, respectively (Table 4).

Table 4.

Survival Rates According to Society for Cardiovascular Angiography and Intervention Shock Stage at the Time of the Index Procedure

	All	Stage C/D	Stage E	P value
Procedural survival	99%	99%	98%	0.74
Survival to discharge	71%	79%	54%	<0.01
Survival at 30 d	68%	77%	49%	<0.01
Survival at 1 y	53%	62%	31%	<0.01

Predictors of Mortality

Multivariable logistic analysis identified numerous independent predictors of in‐hospital mortality (Table 5). Age, a history of diabetes, and transient ischemic attack/cerebrovascular accident were associated with an increased risk of in‐hospital mortality. Similarly, pre‐MCS heart rate, SBP, lactate, and door‐to‐support times were also associated with mortality. Being transferred was the only variable associated with decreased mortality. Post‐MCS heart rate, diastolic blood pressure, mean arterial pressure, lactate, and cardiac power output were also identified as independent predictors of in‐hospital mortality (Figure 4A through 4C).

Table 5.

Predictors of Mortality

Covariates	Odds ratio (95% CI)	Odds ratio P value
Multivariable logistical regression analysis*
Age	1.04 (1.01–1.07)	<0.01
Diabetes	2.67 (1.35–5.29)	<0.01
TIA/CVA	4.34 (1.67–11.31)	<0.01
Transfer	0.40 (0.18–0.92)	0.03
Pre‐MCS HR	1.02 (1.01–1.03)	<0.01
Pre‐MCS SBP	0.99 (0.98–1.00)	0.05
Pre‐MCS lactate	1.08 (1.00–1.17)	0.05
Door to support	1.00 (1.00–1.00)	0.04
Generalized estimating equations model†
HR	1.00 (1.00–1.00)	0.03
DBP	1.00 (1.00–1.00)	0.11
MAP	1.00 (1.00–1.00)	0.03
Lactate	1.00 (1.00–1.00)	0.03
CPO	1.00 (1.00–1.00)	<0.01

CPO indicates cardiac power output; CVA, cerebral vascular attack; DBP, diastolic blood pressure; HR, heart rate; MAP, mean arterial blood pressure; MCS, mechanical circulatory support; SBP, systolic blood pressure; and TIA, transient ischemic attack.

^*Number of observations in the original data set = 406. Number of observations used = 231.

^† Number of observations in the original data set = 812. Number of observations used = 400.

Figure 4. Lactate trends.

A, In‐hospital survival according to admission lactate level. B, In‐hospital survival according to lactate clearance at 12 to 24 hours. C, Trends in mean lactate levels (with 95% CIs) among survivors and nonsurvivors.

DISCUSSION

Our analysis highlights several important findings: (1) Use of MCS as an adjunctive therapy to PCI is feasible in varying practice models, including academic and community programs. (2) Numerous factors, including age, poor hemodynamics, and lactate levels, are associated with in‐hospital mortality. (3) Early use of MCS rapidly improves hemodynamics and systemic perfusion without the need for increasing doses of vasopressors and inotropes, providing a rational and physiological mechanism for their potential benefit. (4) Use of MCS as part of a shock protocol or team‐based system was associated with high survival when compared with prior clinical trials with similar inclusion criteria and shock stages. ² , ⁷ , ¹² , ¹³ , ¹⁴ (5) Of the patients who died, the majority died from worsening CS or multi‐organ failure, and only a minority (<20%) received MCS escalation, a possible avenue to enhance future survival.

Investigators implemented the NCSI to provide clinicians a more uniform mechanism for the use of MCS and to help fill the void of inexperience and unfamiliarity regarding MCS. Investigators began by identifying practices associated with improved survival, including the following: (1) rapid identification and activation of the catheterization laboratory for patients with AMI‐CS, (2) utilization of early MCS (ie, before escalating doses of vasopressors and inotropes and before PCI), and (3) the routine use of invasive hemodynamics to guide clinical decision‐making, including MCS weaning and escalation (Figure 5). ¹⁵ Impella was utilized because it provides greater hemodynamic support compared with IABP and is readily available in the cardiac catheterization laboratory, whereas extracorporeal membrane oxygenation was infrequently available at many participating hospitals or required a significant delay due to the need for surgical consultation and operating room availability.

Figure 5. National Cardiogenic Shock Initiative best practices.

CS indicates cardiogenic shock; CTO, chronic total occlusion; GDMT, guideline‐directed medical therapy; MCS, mechanical circulatory support; PCI, percutaneous coronary intervention; and TIMI, thrombolysis in myocardial infarction.

Within this context, the 71% in‐hospital survival in our cohort (which included 27% of patients in Society for Cardiovascular Angiography and Intervention stage E shock, who are unlikely to be included in the RCT) is promising. Though direct comparisons with prior clinical trials should be done cautiously, it is apparent that patients included into the NCSI may in fact represent a sicker cohort compared with previous trials, with a >10 mm Hg lower blood pressure than the IABP SHOCK trial and higher rates of cardiac arrest when compared with the SHOCK trial. In fact, patients included into the NCSI seem to be more on par with patients included in the recently completed DanGer Trial (Danish‐German Cardiogenic Shock Trial) Table 6. ¹⁴

Table 6.

Comparison of Characteristics in Acute Myocardial Infarction and Cardiogenic Shock Trials

	Sample size	Age	Inotropes (%)	Cardiac arrest (%)	HR, bpm	BP, mm Hg	Lactate, mmol/L	Lactate ≥2, mmol/L	30‐day survival %
SHOCK	302	66	99	28	102	89/54	N/A	N/A	53
IABP SHOCK	600	70	90	45	92	90/55	4.1	74	60
Culprit SHOCK	686	70	90	54	91	100/60	5.1	66	49
DanGer	100	68	94	0	N/A	76/50	5.5	100	N/A
NCSI	406	64	85	46	95	77/50	4.8	77	68

BP indicates blood pressure; HR, heart rate; IABP, intra‐aortic balloon pump; and NCSI, National Cardiogenic Shock Initiative.

Since the publication of the IABP SHOCK II trial in 2012, there has been a gradual abandonment of IABP across Europe and to a lesser degree in the United States as well. ¹⁵ The European guidelines downgraded use of IABP from a previous class I recommendation to a class III B recommendation. ¹⁶ This has led some clinicians to increase use of more robust forms of MCS; however, these devices remain infrequently utilized. In an analysis conducted from 2005 to 2014, <5% of MCS‐assisted PCIs were conducted with Impella and <1% with extracorporeal membrane oxygenation. ¹⁷ There is also considerable institutional variability in both utilization and outcomes associated with the use of MCS. ¹⁸ , ¹⁹ Shock teams and institutional protocols help to standardize and streamline the use and escalation of MCS.

At most centers, at the time NCSI was initiated, there was little to no utilization of extracorporeal membrane oxygenation in patients with AMI‐CS. Similarly, though we provided hemodynamic and clinical context for when MCS escalation should be considered, there were significant differences in the ability to provide MCS escalation within hospitals. Of the patients who died in our study, the majority died from CS or multi‐organ failure, and only a minority (<20%) received MCS escalation, a possible avenue to enhance future survival.

The results of our analysis should be considered hypothesis generating, given the inherent limitations of a single‐arm study. It is, however, important to emphasize the practical reasons for choosing this study design. While the gold standard for testing a hypothesis involves performing a well‐powered RCT, clinicians, hospitals, and health care systems often require a system to utilize novel technologies. It is also important to emphasize that conducting AMI‐CS trials in the United States is challenging. Challenges include a low incidence of CS and limited experience with the implantation, management, weaning, and explantation of MCS devices. Even more daunting challenges include difficulties in obtaining informed consent, varying shock phenotypes, existing US Food and Drug Administration approval for use, varying levels of equipoise, and the ability to allow crossover. ²⁰ Over the past 2 decades, only 2154 patients have been recruited into 14 RCTs, the majority of which were conducted in Europe, and dozens of trials have been halted due to slow or poor recruitment. ²¹ Patients enrolled into RCTs also experience less severe shock, received more aggressive therapies, and have higher survival when compared with those in registries. ²¹ , ²² Therefore, there is value in evaluating whether a strategy or therapy is associated with clinical improvement and reproducible in different settings and across a spectrum of patients with AMI‐CS. In fact, these are the very efforts that push the field toward conducting large, well‐funded RCTs. Thus, the results of our study are promising and should be the basis of future RCTs, including the planned Recover IV trial.

Limitations

Given the observational nature of this study, selection bias may have contributed to the exclusion of certain patients despite aggressive measures to perform screening of consecutive patients. Similarly, reports of adverse events were self‐reported without adjudication or monitoring, and the impact and timing of these events is unknown. Adverse event rates are generally lower than previous reports, and it is unclear whether this relates to improved operator experience or is due to underreporting. Angiographic data were also self‐reported without the presence of an angiographic core laboratory.

CONCLUSIONS

Early use of MCS in AMI‐CS is feasible across varying health care settings and resulted in improvements to early hemodynamics and perfusion. Survival rates to hospital discharge were high. Given the encouraging results from our analysis, RCTs are warranted to assess the early use of MCS using a standardized, multidisciplinary approach.

Sources of Funding

The National Cardiogenic Shock Initiative was funded by research grants from Abiomed (Danvers, Massachusetts) and Chiesi Pharmaceuticals Inc (Cary, North Carolina).

Disclosures

Dr Basir is a consultant for Abiomed, Boston Scientific, Chiesi, Saranas, and Zoll; and receives research funds with Abiomed, Chiesi, Shockwave, Saranas, and Zoll. Dr Kolski is a consultant for Abiomed, Penumbra, Cardiovascular Systems, Medtronic, Shockwave Medical, Edwards, Abbott Vascular, and Biotronik. Dr Bharadwaj is a consultant and on the speaker's bureau for Abiomed, Cardiovascular Solutions Inc., and Shockwave. Dr Todd is a consultant for Abiomed. Dr Tehrani reports research grants from Boston Scientific and Inari. Dr Truesdell is a consultant for Abiomed and Shockwave Medical. Dr Lasorda is a speaker for Abiomed, Shockwave Medical, Boston Scientific, and Edwards. Dr Kaki is a consultant for Abiomed, Cardiovascular Systems, Shockwave Medical, Amgen, AstraZeneca, Abbott Vascular, CathWorks, Inari, and Philips. Dr Rahman is a consultant for Abbott Vascular, Medtronic, Abiomed, and Edwards. Dr Federici is a speaker for Abiomed, medical advisor and proctor for Boston Scientific. Dr Khuddus is a consultant and on the speaker's bureau for Abbott Vascular and Terumo, a member of the medical advisory board for Procyrion, and a shareholder of Saranas. Dr Goldsweig reports consulting income from Inari Medical and Philips. Dr Lim is a consultant for Abiomed. Dr Wohns reports receiving honoraria from Abiomed. Dr Kapur receives institutional research support from Abbott, Abiomed, Boston Scientific, Getinge, LivaNova, and consulting/speaker honoraria from Abbott, Abiomed, Boston Scientific, Getinge, LivaNova, Medtronic, MDStart, Precardia, and Zoll. Dr O'Neill reports receiving fees from Abiomed. The remaining authors have no disclosures to report.

Supporting information

Data S1

Table S1–S2

Reference [23]