Abstract
Background:
In 2019, the Society for Cardiovascular Angiography and Interventions (SCAI) proposed an algorithm to assess cardiogenic shock (CS) severity, known as SCAI classification (1). In 2022, the Cardiogenic Shock Working Group (CSWG) modified the original classification using specific parameters to better define hypotension and hypoperfusion with the goal to further refine stratification (2).
Methods:
Consecutive patients with CS managed at a quaternary academic medical center from May 2015 to December 2021 were evaluated (N=1,162). Our study cohort comprised patients who developed CS after admission. In-hospital mortality per SCAI stage at shock onset was assessed. The percentages of tMCS use per SCAI stage at shock onset was also reported. In our study, we included intra-aortic balloon pump (IABP), percutaneous ventricular assist device (pVAD), and veno-arterial extracorporeal membrane oxygenation (VA-ECMO).
Results:
Our study cohort comprised 742 patients with a median age of 62.0 years (interquartile range 51–72), mostly male (66%). Etiology of CS included 25% being acute myocardial infarction (AMI) and 75% being non-AMI. In-hospital mortality per CS severity stage via SCAI-CSWG classification was 14.30% for stage B, 20.90% for stage C, 32.70% for stage D, and 44.80% for stage E while tMCS utilization was greatest in SCAI Stage E.
Conclusion:
We identified a direct relationship of in-hospital mortality per CS severity stage via SCAI-CSWG classification in the overall cohort. These findings require further validation in prospective studies.
Introduction
Cardiogenic shock (CS) is a heterogeneous syndrome with numerous etiologies and clinical presentations where assessment of severity is imperative as it aids in prognostication and management decision making.
In 2019, the Society for Cardiovascular Angiography and Interventions (SCAI) aimed to create a uniform way to assess CS severity. A novel scoring system was developed, known as the SCAI CS classification, that categorizes patients into 5 different stages of CS severity: A through E, which range from “at risk” to “extremis” (1). The SCAI CS classification, developed by expert consensus, was then validated in both retrospective and prospective studies (3–7).
While SCAI classification allows for a uniform way to assess CS severity, the criteria set forth did not clearly define hypoperfusion and hypotension in all stages, which led to ambiguity among clinicians trying to assign stages. Therefore, in 2022, the Cardiogenic Shock Working Group (CSWG) modified the original classification incorporating objective clinical parameters and information. including lactate levels, alanine transaminase (ALT) levels, systolic blood pressure (SBP), mean arterial pressure (MAP), arterial pH, as well as out-of-hospital cardiac arrest and utilization of CS pharmacologic and mechanical circulatory support (MCS) therapies (2).
Application of the updated classification indicated a direct association between both baseline and maximum SCAI stage via CSWG criteria and in-hospital mortality. A direct relationship between increasing treatment intensity and in-hospital mortality was also observed. It was hypothesized that the more specific definitions of CS severity allowed for improved reliability of appropriate stage assignment and understanding of clinical trajectories (2). However, there has been limited validation of this novel CS severity classification after it was developed. Therefore, our objective was to further explore the relationship between SCAI-CSWG classification and in-hospital mortality. Specifically, our objectives were to assess in-hospital mortality using the SCAI-CSWG classification at shock onset in 1) all patients suffering CS; 2) patients suffering acute myocardial infarction (AMI)-related CS (AMI-CS); 3) patients suffering non-AMI-related CS (non-AMI-CS). Moreover, we assessed the association of SCAI-CSWG classification at shock onset with the use of temporary mechanical circulatory support (tMCS), which includes intra-aortic balloon pump (IABP), percutaneous ventricular assist device (pVAD), and veno-arterial membrane oxygenation (VA-ECMO). We further assessed those who received durable left ventricular assist device (LVAD) and heart transplantation per CS etiology and in-hospital mortality. Last, we compared in-hospital mortality stratification using the original SCAI classification and the SCAI-CSWG classification.
Methods
Study Population
Our study population was comprised of consecutive, prospectively enrolled, and retrospectively reviewed patients with CS, that were admitted and managed at the University of Utah Hospital in Salt Lake City, Utah, from May 2015 to December 2021, and followed up to December 2022. The University of Utah Hospital is a quaternary academic healthcare institution providing 24–7 advanced cardiovascular care including cardiovascular intensive care unit management, primary percutaneous coronary intervention and coronary artery bypass graft surgery, temporary mechanical circulatory support (MCS) with intra-aortic balloon pump (IABP), percutaneous ventricular assist devices (pVAD; i.e. Impella®; Abiomed, Danvers, MA, USA), surgically-implanted ventricular assist devices (CentriMag™; Abbott Laboratories, Abbott Park, IL, USA), percutaneous arterial or venous bypass systems (veno-arterial membrane oxygenation [VA-ECMO], TandemHeart® and ProtekDuo®; LivaNova PLC, London, United Kingdom), as well as advanced heart failure (HF) therapies including durable left ventricular assist device (LVAD) implantation and HTx. The Utah Cardiac Recovery (UCAR) Shock Team was established in April 2015 to evaluate patients suffering CS with a standardized comprehensive multidisciplinary assessment, as described previously (8,9).
Patients were originally included in the study if they met the diagnosis of CS, criteria described below, and were managed at the University of Utah Hospital at one point during their hospitalization. This encompassed both patients who were initially diagnosed with CS upon admission to the University of Utah Hospital, as well as those who received the diagnosis later during their hospital stay. Additionally, it included individuals who were transferred from an outside healthcare facility. The patients who were transferred from an outside healthcare facility were then excluded from the study if CS was initially diagnosed at that facility. These patients were excluded due to difficulties confirming the exact clinical data in the other institutions for the purpose of maintaining accuracy of SCAI stage assignment at shock onset. Patients with multifactorial shock etiologies or post-cardiotomy CS were excluded from the analysis as well. Last, patients with SCAI stage A were excluded from the analysis as they were deemed hemodynamically stable without the need of inotropes, vasopressors, or tMCS.
Data Collection and Definitions
Clinical data were collected via retrospective chart review of electronic medical records. Patient demographics and baseline laboratory, hemodynamic, and imaging data were collected at the timepoint closest to the shock onset. Shock onset was defined the first time the patient met the CS definition, as described below.
CS was defined using both clinical and hemodynamic criteria as previously described (8,10), including one of the following: 1) systolic blood pressure (SBP) <90 mmHg or mean arterial pressure (MAP) <50 mmHg for more than 30 minutes, 2) need for inotrope or vasopressor agent infusion to maintain an SBP >90 mmHg or MAP >50 mmHg, plus one of the following: i) pulmonary capillary wedge pressure (PCWP) or left ventricular end diastolic pressure (LVEDP) >15 mmHg and cardiac index (CI) <2.2 L/min/m2, ii) clinical or radiological signs of pulmonary edema, iii) impaired end-organ perfusion defined as altered mental status, cold clammy skin and extremities, or oliguria with urine output of less than 30 mL/hour.
Based on the underlying CS pathophysiology patients were stratified into suffering AMI-CS or non-AMI-CS. AMI-CS included ST-segment elevation MI, and non-ST-segment elevation MI, while non-AMI-CS included a multitude of underlying etiologies: ischemic cardiomyopathy (ICM), non-ischemic idiopathic cardiomyopathy (NICM), cardiac arrhythmias, valvular disease, myocarditis, infiltrative, peripartum, stress-induced cardiomyopathies etc. After a patient met the diagnosis of CS, pertinent data were obtained to assess CS severity using the original and updated SCAI CS classification criteria (1,2).
Study Endpoints
The endpoint was in-hospital mortality, defined as mortality during index CS hospitalization.
Statistical Methods
Continuous variables were reported as median with interquartile range (IQR). Categorical variables were reported as number with percentage. Due to distributions skew, non-normally distributed continuous variables were compared using the Wilcoxon-Mann-Whitney two-sample rank-sum test for underlying CS etiology and survival status at discharge. Categorical variables were compared using either the chi-squared test or Fisher’s exact test, as appropriate. Ordinal regression was used in comparisons between SCAI stages where proportional odds assumption was evaluated. Kappa statistic with linear weight was used to measure the agreement between original SCAI classification and SCAI-CSWG classification. We provided a 95% confidence interval (CI) of the kappa estimate based on 1000 bootstrap samples obtained from randomly selecting subjects with replacement. The interpretation using the lower limit of a 95% CI <0 indicates no agreement; 0.00–0.21, slight; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; 0.81–0.99, almost perfect; and 1.0 indicates perfect agreement (11). Statistical significance was accessed at the 0.05 level.
Results
Consecutive patients with CS treated at the University of Utah Hospital were evaluated (N=1,162). Patients who initially presented and developed CS at an outside healthcare facility prior to transfer to our center were excluded from the analysis (n=420). Moreover, patients assigned to stage A via SCAI-CSWG classification at shock onset were also excluded. Our final study cohort comprised 742 patients.
Patient Characteristics in the Total Study Cohort
Demographics, past medical history, hemodynamic, echocardiographic, and laboratory measurements at shock onset in the total study cohort are presented in Table 1. The majority of patients suffered non-AMI-CS (n=558, 75%) with the remaining patients suffering AMI-CS (n=184, 25%). The median age of patients was 62.0 (IQR: 51.0–71.8) years, 66% were male and 81% white. With regard to cardiac risk factors, 17% of patients were current smokers and 30% former smokers, 66% had a history of systemic hypertension, 30% a history of diabetes mellitus, 42% a history of hyperlipidemia, and 15% a history of chronic kidney disease (CKD). Twenty-two percent had a history of a prior myocardial infarction, 30% a coronary revascularization procedure, and 47% chronic HF. Of those with chronic HF, 83% had a reduced (defined as <50%) left ventricular ejection fraction (LVEF), 15% a preserved (defined as ≥50%) LVEF, 1% a reduced LVEF in the setting of prior heart transplantation, and 1% were unknown. At shock onset the median lactate was 2.3 (IQR: 1.4–4.8) mmol/L, serum creatinine was 1.5 (IQR: 1.0–2.1) mg/dL and B-type natriuretic peptide (BNP) was 1235.0 (IQR: 538.0–2447.0) ng/L. Regarding hemodynamic data at shock onset, the median pulmonary capillary wedge pressure (PCWP) was 23.0 (IQR: 17.0–29.0) and cardiac index (CI) via Fick method was 1.8 (IQR: 1.4–2.4) L/min/m2. Median echocardiographic data at shock onset included a LVEF of 30.0 (IQR: 19.0–51.0) % and median left ventricular end-diastolic diameter (LVEDD) of 5.2 (IQR: 4.5–6.0). Further details regarding laboratory, hemodynamic and echocardiographic data can be found in Table 1.
Table 1.
Baseline Clinical Characteristics in the Total Study Cohort
| All Patients (n = 742) | |
|---|---|
|
| |
| Cause of CS | |
| AMI CS (n, %) | 184 (25%) |
| Non-AMI CS (n, %) | 558 (75%) |
|
| |
| Demographics | |
| Age, y (median, IQR) | 62.0 (51.0–71.8) |
| BMI, kg/m2 (median, IQR) | 28.9 (24.6–34.0) |
|
| |
| Sex | |
| Male (n, %) | 493 (66%) |
| Female (n, %) | 249 (34%) |
|
| |
| Race | |
| Asian (n, %) | 13 (2%) |
| Black/African American (n, %) | 21 (3%) |
| Native Hawaiian/Other Pacific Islander (n, %) | 13 (2%) |
| White (n, %) | 598 (81%) |
| Other (n, %) | 62 (8%) |
| Unknown (n, %) | 17 (2%) |
|
| |
| Smoking History (n, %) | |
| Current Smoker | 126 (17%) |
| Former Smoker | 226 (30%) |
|
| |
| Past Medical History | |
| Hypertension (n, %) | 487 (66%) |
| Diabetes (n, %) | 225 (30%) |
| Hyperlipidemia (n, %) | 311 (42%) |
| CKD (n, %) | 109 (15%) |
| Prior MI (n, %) | 162 (22%) |
| Prior PCI (n, %) | 138 (19%) |
| Prior CABG (n, %) | 82 (11%) |
| Prior Chronic Heart Failure (n, %) | 351 (47%) |
| Reduced LVEF (n, % of HF) | 294 (83%) |
| Persevered LVEF (n, % of HF) | 52 (15%) |
|
| |
| Laboratory at Shock Onset | |
| Lactate, mmol/L (median, IQR) | 2.3 (1.4–4.8) |
| Sodium, mmol/L (median, IQR) | 137.0 (133.0–140.0) |
| Serum Creatinine, mg/dL (median, IQR) | 1.5 (1.0–2.1) |
| BNP, ng/L (median, IQR) | 1235.0 (538.0–2447.0) |
| ALT, U/L (median, IQR) | 47.0 (22.0–158.5) |
| pH (median, IQR) | 7.3 (7.3–7.4) |
|
| |
| Hemodynamics at Shock Onset | |
| RAP, mm Hg (median, IQR) | 12.0 (8.0–18.0) |
| PCWP, mm Hg (median, IQR) | 23.0 (17.0–29.0) |
| mPAP, mm Hg (median, IQR) | 35.0 (26.0–43.0) |
| CI via Fick, L/min/m2 (median, IQR) | 1.8 (1.4–2.5) |
| CI via TD, L/min/m2 (median, IQR) | 1.8 (1.4–2.5) |
|
| |
| Echocardiography at Shock Onset | |
| LVEF, % (median, IQR) | 36.0 (16.2) |
| LVEDD, mm Hg (median, IQR) | 5.2 (4.5–6.0) |
Abbreviations: ALT = alanine transaminase; AMI = acute myocardial infarction; BMI = body mass index; BNP = B-type natriuretic peptide; CABG = coronary artery bypass graft; CKD = chronic kidney disease; CI = cardiac index; CS = cardiogenic shock; IQR = interquartile range; LVEDD = left ventricular end-diastolic diameter; LVEF = left ventricular ejection fraction; MI = myocardial infarction; mPAP = mean pulmonary artery pressure; PCI = percutaneous intervention; PCWP = pulmonary capillary wedge pressure; RAP = right atrial pressure.
Patient Characteristics Stratified by Cardiogenic Shock Etiology
Clinical characteristics after stratifying patients based on the underlying CS etiology (AMI-CS versus non-AMI-CS) are presented in Table 2. Patients who suffered AMI-CS had higher mortality in comparison with non-AMI-CS (45.7 % vs. 29.9 % [p< 0.001]). Patients who suffered AMI-CS were older compared to non-AMI-CS (65.0 vs. 61.0 years [p < 0.001]), with no differences observed in terms of sex, race, or BMI. Regarding cardiac risk factors, patients with AMI-CS were more likely to be current smokers (22.3% vs.15.3% [p = 0.005]) and less likely to be former smokers (21.6% vs. 33.4% [p = 0.005], with no difference regarding history of systemic hypertension, diabetes mellitus, or hyperlipidemia. Rates of CKD, prior history of MI, PCI, or CABG were comparable between groups. A smaller proportion of patients had a history of chronic HF in AMI-CS at 12.0%, while 59.0% had prior chronic HF in non-AMI-CS (p < 0.001). Of those with chronic HF, 8.7% had preserved LVEF and 82.6% had reduced LVEF in AMI-CS, while 15.1% had preserved LVEF and 83.1% had reduced LVEF in non-AMI-CS (p = 0.01),
Table 2.
Baseline Clinical Characteristics in Patients Stratified Based on Cardiogenic Shock Etiology
| AMI-CS (n = 184) | Non-AMI-CS (n = 558) | P value | |
|---|---|---|---|
|
| |||
| In-Hospital Mortality (n, %) | 84 (45.7%) | 167 (29.9%) | <0.001 |
|
| |||
| Demographics | |||
| Age, y (median, IQR) | 65.0 (57.8–74.0) | 61.0 (48.0–70.0) | <0.001 |
| BMI, kg/m2 (median, IQR) | 28.9 (25.4–33.1) | 29.0 (24.5–34.4) | 0.70 |
|
| |||
| Sex | 0.44 | ||
| Male (n, %) | 118 (64.1%) | 375 (67.2%) | |
| Female (n, %) | 66 (35.9%) | 183 (32.8%) | |
|
| |||
| Race | 0.30 | ||
| Asian (n, %) | 4 (2.2%) | 9 (1.6%) | |
| Black/African American (n, %) | 1 (0.5%) | 20 (3.6%) | |
| Native Hawaiian/Other Pacific Islander (n, %) | 3 (1.6%) | 10 (1.8%) | |
| White (n, %) | 152 (82.6%) | 446 (79.9%) | |
| Other (n, %) | 15 (8.2%) | 47 (8.4%) | |
| Unknown (n, %) | 6 (3.3%) | 11 (2.0%) | |
|
| |||
| Smoking History (n, %) | 0.005 | ||
| Current Smoker | 41 (22.3%) | 85 (15.3%) | |
| Former Smoker | 40 (21.7%) | 186 (33.4%) | |
|
| |||
| Past Medical History | |||
| Hypertension (n, %) | 129 (70.1%) | 358 (64.2%) | 0.12 |
| Diabetes (n, %) | 58 (31.5%) | 167 (29.9%) | 0.20 |
| Hyperlipidemia (n, %) | 83 (45.1%) | 228 (40.9%) | 0.21 |
| CKD (n, %) | 157 (85.3%) | 88 (15.8%) | 0.27 |
| Prior MI (n, %) | 39 (21.2%) | 123 (22.0%) | 0.46 |
| Prior PCI (n, %) | 38 (20.7%) | 100 (17.9%) | 0.26 |
| Prior CABG (n, %) | 17 (9.2%) | 65 (11.6%) | 0.42 |
| Prior Chronic Heart Failure (n, %) | 22 (12%) | 329 (59%) | <0.001 |
| Reduced LVEF (n, % of HF) | 19 (82.6%) | 275 (83.1%) | 0.10 |
| Persevered LVEF (n, % of HF) | 2 (8.7%) | 50 (15.1%) | 0.10 |
|
| |||
| Laboratory at Shock Onset | |||
| Lactate, mmol/L (median, IQR) | 3.1 (1.7–6.8) | 2.2 (1.3–4.3) | <0.001 |
| Sodium, mmol/L (median, IQR) | 138.0 (134.0–141.0) | 136.0 (133.0–139.0) | <0.001 |
| Serum Creatinine, mg/dL (median, IQR) | 1.3 (1.0–1.8) | 1.5 (1.1–2.1) | 0.015 |
| BNP, ng/L (median, IQR) | 1011.0 (304.5–2043.5) | 1294.0 (576.0–2571.5) | 0.035 |
| ALT, U/L (median, IQR) | 58.0 (29.0–172.0) | 44.0 (21.0–156.0) | 0.09 |
| pH (median, IQR) | 7.3 (7.2–7.4) | 7.4 (7.3–7.4) | <0.001 |
|
| |||
| Hemodynamics at Shock Onset | |||
| RAP, mm Hg (median, IQR) | 13.0 (10.0–18.0) | 12.0 (7.0–18.0) | 0.39 |
| PCWP, mm Hg (median, IQR) | 23.0 (17.5–29.0) | 23.0 (17.0–29.0) | 0.96 |
| mPAP, mm Hg (median, IQR) | 32.0 (25.0–40.5) | 36.0 (26.5–44.0) | 0.027 |
| CI via Fick, L/min/m2 (median, IQR) | 2.1 (1.6–2.5) | 1.7 (1.4–2.3) | <0.001 |
| CI via TD, L/min/m2 (median, IQR) | 1.8 (1.3–2.4) | 1.8 (1.4–2.5) | 0.77 |
|
| |||
| Echocardiography at Shock Onset | |||
| LVEF, % (median, IQR) | 35.0 (25.0–47.5) | 28.0 (17.0–54.0) | 0.039 |
| LVEDD, cm (median, IQR) | 4.8 (4.3–5.4) | 5.3 (4.6–6.3) | <0.001 |
|
| |||
| tMCS | |||
| IABP (n, %) | 35 (19.0%) | 37 (6.6%) | <0.001 |
| pVAD (n, %) | 40 (21.7%) | 39 (7.0%) | <0.001 |
| VA-ECMO (n, %) | 29 (15.8%) | 39 (7.0%) | <0.011 |
|
| |||
| Durable LVAD (n, %) | 2 (1.1%) | 25 (4.5%) | 0.07 |
|
| |||
| Heart Transplant (n, %) | 0 (0%) | 20 (3.6%) | 0.006 |
Abbreviations: ALT = alanine transaminase; AMI = acute myocardial infarction; BMI = body mass index; BNP = B-type natriuretic peptide; CABG = coronary artery bypass graft; CKD = chronic kidney disease; CI = cardiac index; CS = cardiogenic shock; IQR = interquartile range; LVEDD = left ventricular end-diastolic diameter; LVEF = left ventricular ejection fraction; MI = myocardial infarction; mPAP = mean pulmonary artery pressure; PCI = percutaneous intervention; PCWP = pulmonary capillary wedge pressure; RAP = right atrial pressure. Additional abbreviations: IABP: Intra-aortic Balloon Pump; pVAD: Percutaneous Ventricular Assist Device; tMCS: Temporary Mechanical Circulatory Support; VAD: Ventricular Assist Device; VA-ECMO: Veno-atrial Extracorporeal Membrane Oxygenation.
Analysis of laboratory data at shock onset showed that AMI-CS patients had higher lactate (p < 0.001) and sodium levels (p < 0.0010), compared to non-AMI-CS. Patients with AMI-CS had lower serum creatinine (p = 0.015), BNP (p = 0.035), and pH (p < 0.001). Regarding those with AMI-CS, 68.3 % had a lactate <5, 0.6 % had a lactate between 5 and 10, and 31.3 % had a lactate >10 (p = 0.04). Regarding those with non-AMI-CS, 77.6 % had a lactate <5, 0.4 % had a lactate between 5 and 10, and 22.0 % had a lactate >10. Those with AMI-CS had a higher percentage of patients with a pH <7.2 than those with non-AMI-CS (p < 0.001).
Hemodynamic assessment at shock onset showed that AMI-CS patients had a significantly higher CI via Fick method (2.1 vs. 1.7 L/min/m2 [p < 0.001]), but similar PCWP compared to non-AMI-CS. Echocardiographic assessment revealed that the median LVEF was higher in AMI-CS compared to non-AMI-CS patients (35.0% vs. 28.0% [p = 0.039]) and LVEDD was lower in AMI-CS compared to non-AMI-CS (4.8 vs. 5.3 mm Hg [p < 0.001]).
There were no significant differences between the two groups in who received a durable LVAD. A higher percentage of non-AMI-CS patients received heart transplantation (3.6% vs. 0% [p = 0.006]).
We found no significant difference between AMI-CS and non-AMI-CS patients in those who received renal replacement therapy (RRT) during hospitalization.
Patient Characteristics Stratified by In-Hospital Mortality
The clinical characteristics of patients who were alive or not at discharge are presented in Table 3. Patients alive at discharge were younger compared to those who were not (61.0 [IQR: 48.5–70.0] vs. 64.0 [IQR: 56.0–74.0] years [p = 0.001]), with no statistical differences in terms of sex, race, and BMI. Regarding cardiac risk factors, patients who did survive versus patients who did not survive to discharge were more likely to be current (20.2 % vs. 10.8 % [p < 0.001]) or former smokers (32.7% vs. 26.3 % [p < 0.001]) and have a history of hyperlipidemia (59.1 % vs. 54.6 % [p = 0.014]). Patients surviving to discharge were less likely to have a history of hypertension (32.6% vs. 36.3% [p=0.01]), diabetes mellitus (29.5 % vs. 31.9 % [p = 0.043]), and CKD compared to those who did not (13.6 % vs. 16.7 % [p = 0.002]). Patients alive at discharge were less likely to have a history of MI (21.6% vs.22.3% [p = 0.023]), prior PCI (18.1% vs.19.5% [p = 0.02]), and CABG (10.0 % vs.13.1 % [p = 0.003]). Patients alive at discharge were more likely to have a history of chronic HF (52.1 % vs. 37.8 % [p <0.001]).
Table 3.
Baseline Clinical Characteristics in Patients Stratified Based on In-Hospital Mortality
| Alive at Discharge (n = 491) | Not Alive at Discharge (n = 251) | P value | |
|---|---|---|---|
|
| |||
| Demographics | |||
| Age, y (median, IQR) | 61.0 (48.5–70.0) | 64.0 (56.0–74.0) | 0.001 |
| BMI, kg/m2 (median, IQR) | 28.9 (24.8–34.0) | 28.8 (24.3–34.1) | 0.68 |
|
| |||
| Sex | |||
| Male (n, %) | 332 (67.6%) | 161 (64.1%) | 0.34 |
| Female (n, %) | 159 (32.4%) | 90 (35.9%) | |
|
| |||
| Race | |||
| Asian (n, %) | 10 (2%) | 3 (1.2%) | 0.14 |
| Black/African American (n, %) | 17 (3.5%) | 4 (1.6%) | |
| Native Hawaiian/Other Pacific Islander (n, %) | 9 (1.8%) | 4 (1.6%) | |
| White (n, %) | 394 (80.2%) | 204 (81.3%) | |
| Other (n, %) | 43 (8.8%) | 19 (7.6%) | |
| Unknown (n, %) | 6 (1.2%) | 11 (4.4%) | |
|
| |||
| Smoking History (n, %) | <0.001 | ||
| Current Smoker | 99 (20.2%) | 2 / (10.8%) | |
| Former Smoker | 160 (32.7%) | 66 (26.3%) | |
|
| |||
| Past Medical History | |||
| Hypertension (n, %) | 160 (32.6%) | 91 (36.3%) | 0.01 |
| Diabetes (n, %) | 145 (29.5%) | 80 (31.9%) | 0.043 |
| Hyperlipidemia (n, %) | 290 (59.1%) | 137 (54.6%) | 0.014 |
| CKD (n, %) | 67 (13.6%) | 42 (16.7%) | 0.002 |
| Prior MI (n, %) | 106 (21.6%) | 56 (22.3%) | 0.023 |
| Prior PCI (n, %) | 89 (18.1%) | 49 (19.5%) | 0.02 |
| Prior CABG (n, %) | 49 (10%) | 33 (13.1%) | 0.003 |
| Prior Chronic Heart Failure (n, %) | 256 (52.1%) | 95 (37.8%) | <0.001 |
| Reduced LVEF (n, % of HF) | 218 (85.2%) | 76 (77.6%) | 0.002 |
| Persevered LVEF (n, % of HF) | 37 (14.5%) | 5 (15.3%) | 0.002 |
|
| |||
| Laboratory at Shock Onset | |||
| Lactate, mmol/L (median, IQR) | 2.1 (1.3–3.8) | 3.4 (1.7–7.8) | <0.001 |
| Sodium, mmol/L (median, IQR) | 136.0 (133.0–139.0) | 137.0 (134.0–141.0) | 0.001 |
| Serum Creatinine, mg/dL (median, IQR) | 1.4 (1.0–2.0) | 1.6 (1.1–2.5) | 0.003 |
| BNP, ng/L (median, IQR) | 1138.0 (497.2–2430.2) | 1554.0 (693.0–2516.5) | 0.09 |
| ALT, U/L (median, IQR) | 42.0 (21.0–146.0) | 62.0 (27.0–196.0) | 0.007 |
| pH (median, IQR) | 7.4 (7.3–7.4) | 7.3 (7.2–7.4) | <0.001 |
|
| |||
| Hemodynamics at Shock Onset | |||
| RAP, mm Hg (median, IQR) | 12.0 (7.0–18.0) | 13.0 (8.0–18.0) | 0.24 |
| PCWP, mm Hg (median, IQR) | 22.0 (16.0–28.0) | 24.0 (18.0–31.0) | 0.016 |
| mPAP, mm Hg (median, IQR) | 34.0 (26.0–42.0) | 36.0 (27.0–43.8) | 0.26 |
| CI via Fick, L/min/m2 (median, IQR) | 1.8 (1.4–2.4) | 1.8 (1.5–2.5) | 0.22 |
| CI via TD, L/min/m2 (median, IQR) | 1.9 (1.4–2.6) | 1.8 (1.3–2.3) | 0.05 |
|
| |||
| Echocardiography at Shock Onset | |||
| LVEF, % (median, IQR) | 29.0 (18.0–49.8) | 33.0 (23.0–55.0) | 0.044 |
| LVEDD, cm (median, IQR) | 5.3 (4.6–6.2) | 4.9 (4.1–5.6) | <0.001 |
|
| |||
| tMCS | |||
| IABP (n, %) | 50 (10.2%) | 22 (8.8%) | 0.54 |
| pVAD (n, %) | 42 (8.6%) | 79 (11.0%) | 0.010 |
| VA-ECMO (n, %) | 31 (6.3%) | 68 (9%) | <0.001 |
|
| |||
| Durable LVAD (n, %) | 24 (4.9%) | 3 (1.2%) | 0.015 |
|
| |||
| Heart Transplant (n, %) | 20 (4.1%) | 0 (0%) | 0.001 |
Abbreviations: ALT = alanine transaminase; AMI = acute myocardial infarction; BMI = body mass index; BNP = B-type natriuretic peptide; CABG = coronary artery bypass graft; CKD = chronic kidney disease; CI = cardiac index; CS = cardiogenic shock; IQR = interquartile range; LVEDD = left ventricular end-diastolic diameter; LVEF = left ventricular ejection fraction; MI = myocardial infarction; mPAP = mean pulmonary artery pressure; PCI = percutaneous intervention; PCWP = pulmonary capillary wedge pressure; RAP = right atrial pressure. Additional abbreviations: IABP: Intra-aortic Balloon Pump; pVAD: Percutaneous Ventricular Assist Device; tMCS: Temporary Mechanical Circulatory Support; VAD: Ventricular Assist Device; VA-ECMO: Veno-atrial Extracorporeal Membrane Oxygenation.
Analysis of laboratory data at shock onset showed that patients who survived to discharge compared to those who did not, had lower lactate (p < 0.001), sodium (p = 0.001), serum creatinine (p = 0.003), and ALT levels (p = 0.007). There was no significant difference observed in BNP levels. Regarding those who survived to discharge, 81.9 % had a lactate <5, 0.2 % had a lactate between 5 and 10, and 17.9 % had a lactate >10 (p < 0.001). Regarding those who did not survive to discharge, 62.8 % had a lactate <5, 0.9 % had a lactate between 5 and 10, and 36.4 % had a lactate >10. Those who survived to discharge had a smaller percentage of patients with a pH <7.2 in comparison to those who did not survive to discharge (p < 0.001).
Analysis of hemodynamic data at shock onset showed that patients who survived to discharge compared to those who did not, had a lower PCWP (22.0 vs. 24.0 mmHg [p = 0.016]) and a higher CI via TD method (1.9 vs. 1.8 L/min/m2 [p = 0.05]), with no significant difference in CI via Fick method. Echocardiographic assessment revealed patients alive at discharge were more likely to have a lower LVEF (29.0% vs. 33.0% [p = 0.044]) and a higher LVEDD (5.3 vs. 4.9 mm Hg [p < 0.001]).
Those who survived to discharge compared to those who did not were more likely to receive a durable LVAD (4.9% vs. 1.2% [p = 0.015]) and undergo heart transplantation (4.1% vs. 0.0% [p = 0.001]).
Those who did not survive to discharge were more likely to receive RRT during hospitalization in comparison to those who did survive to discharge (18.7 % vs. 8.1 % [p < 0.001]).
In-Hospital Mortality
The all-cause in-hospital mortality in the total study cohort per CS severity stage via SCAI-CSWG classification, shown in Figure 1, was 14.30% for stage B, 20.9% for stage C, 32.70% for stage D, and 44.80% for stage E (p < 0.001). The all-cause in-hospital mortality in patients suffering AMI-CS per CS severity stage via SCAI-CSWG classification, shown in Figure 2, was 40% for stage B, 18.40% for stage C, 56.20% for stage D, and 52.90% for stage E (p = 0.005, proportional odds assumption was violated leading to non-significance between stages). Last, the all-cause in-hospital mortality for patients suffering non-AMI-CS per CS severity stage via SCAI-CSWG classification, shown in Figure 3, was 8.70% for stage B, 21.60% for stage C, 26.30% for stage D, and 41.20% for stage E (p < 0.001).
Figure 1.
Overall In-Hospital Mortality per Society for Cardiovascular Angiography and Intervention via Cardiogenic Shock Working Group Criteria
Abbreviations: SCAI = Society for Cardiovascular Angiography and Intervention; CSWG = Cardiogenic Shock Working Group
Figure 2.
In-Hospital Mortality of Acute Myocardial Infarction Cardiogenic Shock per Society for Cardiovascular Angiography and Intervention via Cardiogenic Shock Working Group Criteria
Abbreviations: SCAI = Society for Cardiovascular Angiography and Intervention; CSWG = Cardiogenic Shock Working Group.
Figure 3.
In-Hospital Mortality of Non-Acute Myocardial Infarction Cardiogenic Shock per Society for Cardiovascular Angiography and Intervention via Cardiogenic Shock Working Group Criteria
Abbreviations: SCAI = Society for Cardiovascular Angiography and Intervention; CSWG = Cardiogenic Shock Working Group.
Temporary Mechanical Circulatory Support
The use of tMCS was found to change per CS severity stage via SCAI-CSWG classification for all three devices (p<0.001, p=0.002 and p<0.001 respectively for IABP, pVAD, VA-ECMO), with stage E having the highest proportion of use for all three devices studied (Figure 4). In stage B, 10.7% received IABP, 3.6% received pVAD, and VA-ECMO was not used. In stage C, 5.2% received IABP, 0.5% received pVAD, and 0.5% received VA-ECMO. In stage D, 5.3% received IABP, 8.0% received pVAD, and 1.3% received VA-ECMO. In stage E, 14% received IABP, 18.6% received pVAD, and 18.9% received VA-ECMO. These findings illustrate patients who were administered one or a combination of these tMCS devices.
Figure 4.
Use of Temporary Mechanical Circulatory Support per Society for Cardiovascular Angiography and Intervention via Cardiogenic Shock Working Group Criteria
Abbreviations: SCAI = Society for Cardiovascular Angiography and Intervention; CSWG = Cardiogenic Shock Working Group.
Analysis of tMCS use showed that AMI-CS patients had a higher percentage of use in IABP (19.9 % in AMI-CS and 6.6 % use in non-AMI-CS [p < 0.001]), pVAD (21.7 % in AMI-CS and 7.0 % in non-AMI-CS [p < 0.001]), and VA-ECMO (15.8 % in AMI-CS and 7.0 % in non-AMI-CS [p < 0.001]). We also demonstrated that 45.9 % of those with AMI-CS who received tMCS survived hospital discharge while 61.6 % of those with AMI-CS who did not receive tMCS survived hospital discharge.
Analysis of tMCS implementation in patients who survived to discharge compared to those who did not, showcased a lower use of pVAD (8.6 % vs. 11.0 % [p = 0.010]) and VA-ECMO (6.3 % vs. 9.0 % in [p < 0.001]). There was no significant difference observed between groups in terms of IABP use.
Original SCAI Classification versus SCAI-CSWG Classification
The original SCAI classification was compared to the SCAI-CSWG classification to assess assignment agreement via physicians where a weighted kappa statistic estimate and its 95% bootstrapping CI was found to be 0.64 (0.61, 0.69). This finding indicates that there is a “substantial agreement” between the two scoring systems when assessing CS severity.
Discussion
To our knowledge, there have been no studies assessing the performance of the novel SCAI-CSWG classification using an external cohort of patients suffering CS. In this analysis, we were able to demonstrate the following: 1) a direct relationship of in-hospital mortality and CS severity stage via SCAI-CSWG classification at shock onset in non-AMI CS with no significant difference between stages in AMI-CS; 2) a >3-fold increase in mortality of stage E when compared to stage B; 3) stage E had the highest proportion of tMCS utilization; 4) the agreement of assignment between original SCAI classification and SCAI-CSWG classification was “substantial”.
Stratification of patients based on the etiology of CS, showcased a differential distribution of in-hospital mortality across CS severity stages via SCAI-CSWG classification. While patients suffering non-AMI-CS had a similar distribution of in-hospital mortality with the overall cohort, this was not the case for patients suffering AMI-CS. This may simply be due to the smaller sample size of our AMI-CS cohort. This may also suggest that patients presenting with AMI-CS have a clinical trajectory that is difficult to predict at shock onset using SCAI-CSWG classification, and other predictive scores may be beneficial to use in conjunction to assess severity (12). The difficulty in predicting clinical trajectory may also be further supported by our study showing a higher percentage of all tMCS devices in AMI-CS when compared to non-AMI-CS. Notably, the majority of patients in both our study and the original study assessing the CSWG-SCAI classification were non-AMI-CS, which may also indicate that the overall direct relationship of CS severity stage via SCAI-CSWG classification and in-hospital morality may be primarily driven by non-AMI-CS patients (2).
In our overall patient cohort, there was a 6.6% increase in mortality from stage B to stage C, while there was a 11.8% increase in mortality from stage C to stage D. The increase in mortality from stage C to stage D was nearly double when compared to the increase from stage B to stage C, which suggests that patients who present with CS that respond well to initial therapy (stage C), have a better prognosis compared to patients requiring multiple interventions or with persistent illness despite a single intervention. In addition, there was another large increase in mortality (12.1%) from stage D to stage E. This increase in mortality for stage E patients, alongside our observation that patients in stage E had the highest rates of tMCS utilization, might suggest that severity of CS in advanced stages may be excessively difficult to overcome despite the implementation of tMCS. It is also worth noting that while stage E patients had the highest rate of tMCS implementation, its usage was lower than expected. This is likely due to the high acuity of patients in this subgroup, which may result in patients expiring before tMCS can be implemented or patients opting for palliative care instead. Given the complexity of managing this critically ill population of SCAI stage E patients, decision-making in our institution was led by a dedicated cardiogenic shock team (8,9). This multidisciplinary team includes specialists from advanced heart failure cardiology, cardiothoracic surgery, interventional cardiology, cardiovascular ICU medicine, and nursing. The goal of our shock team is to implement a standardized, team-based approach to cardiogenic shock management. By leveraging multidisciplinary expertise, this team aims to optimize patient outcomes in this challenging clinical setting.
We used Cohen Kappa to assess agreement between the original and updated SCAI classification via CSWG, for which we concluded that there is “substantial agreement”. This suggests that both scoring systems result in comparable CS severity assessment. While there is substantial agreement between the two methods of CS classification, the novel SCAI-CSWG classification may be easier to implement due to less ambiguity among stages that could potentially lead to less intra-observer and inter-observer variability; however, this requires further analysis.
In our analysis, we found that patients who did survive versus patients who did not survive to discharge were more likely to be current smokers (20.2 % vs. 10.8 % [p < 0.001]) or former smokers (32.7% vs. 26.3 % [p < 0.001]). There have been studies that have shown lower mortality in current smokers than in nonsmokers at 30 days in patients undergoing mechanical reperfusion therapy in acute myocardial infarction (19). This study showed that smokers were more often younger while having lower rates of hypertension, diabetes, and prior myocardial infarction. Another study found that smoking status was not predictive of outcome in patients with myocardial infarction leading to cardiogenic shock (20). While our study did show lower mortality in current and former smokers, this is likely confounded by patient characteristics. This finding warrants further investigation prospective, larger scale studies.
The idea of a “hub-and-spoke” CS care model has been proposed where increasing evidence via observational studies suggest that patients with CS may benefit from a multidisciplinary care in tertiary care centers (hubs) with recourses necessary to provide the appropriate medical care (13–18). Prior analyses have shown that patients with CS transferred to a hub had a lower in-hospital mortality, and hospitals with the highest CS volume had a reduced 30-day mortality (15–17). Lastly, it has also been demonstrated that patients receive more tMCS, specifically pVAD and VA-ECMO, when transferred to a hub (15). Therefore, accurate assessment of CS severity is imperative as it could impact both treatment decisions and the need to transfer to a hub, particularly in patients with advanced CS severity classification.
There are many unmet needs within the CS field where future studies should focus. Our observation of no statistical difference in mortality rates across CS severity stages via SCAI-CSWG classification within the AMI-CS group, is interesting and welcomes future studies in this patient cohort. We demonstrated substantial agreement between the original SCAI classification and SCAI-CSWG classification, and thus future studies could focus on how differences in these classification systems may impact prognostication and therapeutics. Prior studies have suggested a benefit with a “hub-and-spoke” care model, with accurate assessment of CS severity being crucial in informing patient management, thus making the SCAI-CSWG classification a valuable clinical tool. Last, due to limited data informing therapeutic modality utilization in CS patients, accurate and reproducible CS severity assessment could aid in personalizing treatment approaches, with use of pharmacologic or tMCS treatment, timing of tMCS implementation, selection of tMCS type etc. Capitalizing on a uniform and reproducible CS severity assessment tool could positively impact clinical decision-making and future clinical research endeavors to improve outcomes for patients suffering CS.
Study Limitations
We would like to acknowledge several limitations of our study. First, this is a single-center analysis with findings impacted by institutional diagnostic and treatment practices. Second, we should acknowledge the retrospective nature of data collection and the biases associated with this type of investigation. Third, we might have introduced selection bias by electing to not include in our analysis patients developing CS at an outside healthcare facility. This selection, however, was made to mitigate biases associated with the inclusion of non-uniform diagnostic and therapeutic approaches implemented at outside facilities. Fourth, we should acknowledge the possibility for unmeasured confounding that might have influenced our findings. Fifth, we did not obtain the cause of death for the patients studied, which could provide important information on management. Last, we excluded SCAI Stage A patients as they did not exhibit signs of hypotension or hypoperfusion. The original 1,162 patient cohort did not include SCAI Stage A patients.
Conclusion
The updated SCAI-CSWG classification was designed in 2022 with the hope to create a more comprehensive way to assess CS severity. We found that our cohort validated a direct relationship of CS severity stage using the SCAI-CSWG classification at shock onset with in-hospital mortality; however, this finding was not consistent across all etiologies of CS, specifically AMI-CS. Therefore, this updated classification system may be better suited for non-AMI-CS patients and warrants further study. Nonetheless, SCAI-CSWG classification still provides a framework for addressing CS that will be imperative in triage, treatment, and appropriate transfer of patients with CS to centers with higher level of care. Further, the use of this CS severity score will ideally allow for consistent treatment and better clinical outcomes in the future. These findings require further validation in prospective studies.
Funding
Funding for this project included: 1) AHA Heart Failure Strategically Focused Research Network 16SFRN29020000 (Drs. Drakos and Selzman); 2) NHLBI R01 HL135121–01 (Dr. Drakos); 3) NHLBI R01 HL132067–01A1 (Dr. Drakos); 4) Nora Eccles Treadwell Foundation (Dr. Drakos); 5) U.S. Department of Veterans Affairs Merit Review Award I01 CX002291 (Dr. Drakos); 6) NHLBI 2T32HL007576–36 (Dr. Kyriakopoulos); 7) NHLBI 5T32HL007576–33 (Dr. Taleb); 8) The University of Utah Study Design and Biostatistics Center, with funding in part from the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, UL1TR002538 (Zhining Ou and Angela Presson).
Footnotes
Patient Consent Statement
The authors confirm that patient consent forms have been obtained for this article. The study was approved by the University of Utah Institutional Review Board (IRB). Informed consent was obtained under the IRB 80080 (Utah Cardiac Recovery Shock Registry), or a waiver of consent was granted under IRB 72747 (Clinical Analyses in Cardiovascular Medicine).
Disclosures
Dr. Drakos serves as a consultant for Abbott Laboratories and Pfizer and has received research support from Novartis and Merck. The remaining authors have nothing to disclose.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Baran DA, Grines CL, Bailey S et al. SCAI clinical expert consensus statement on the classification of cardiogenic shock: This document was endorsed by the American College of Cardiology (ACC), the American Heart Association (AHA), the Society of Critical Care Medicine (SCCM), and the Society of Thoracic Surgeons (STS) in April 2019. Catheter Cardiovasc Interv 2019;94:29–37. [DOI] [PubMed] [Google Scholar]
- 2.Kapur NK, Kanwar M, Sinha SS et al. Criteria for Defining Stages of Cardiogenic Shock Severity. J Am Coll Cardiol 2022;80:185–198. [DOI] [PubMed] [Google Scholar]
- 3.Jentzer JC, van Diepen S, Barsness GW et al. Cardiogenic Shock Classification to Predict Mortality in the Cardiac Intensive Care Unit. J Am Coll Cardiol 2019;74:2117–2128. [DOI] [PubMed] [Google Scholar]
- 4.Hanson ID, Tagami T, Mando R et al. SCAI shock classification in acute myocardial infarction: Insights from the National Cardiogenic Shock Initiative. Catheter Cardiovasc Interv 2020;96:1137–1142. [DOI] [PubMed] [Google Scholar]
- 5.Jentzer JC, Schrage B, Holmes DR et al. Influence of age and shock severity on short-term survival in patients with cardiogenic shock. Eur Heart J Acute Cardiovasc Care 2021;10:604–612. [DOI] [PubMed] [Google Scholar]
- 6.Thayer KL, Zweck E, Ayouty M et al. Invasive Hemodynamic Assessment and Classification of In-Hospital Mortality Risk Among Patients With Cardiogenic Shock. Circ Heart Fail 2020;13:e007099. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hernandez-Montfort J, Kanwar M, Sinha SS et al. Clinical Presentation and In-Hospital Trajectory of Heart Failure and Cardiogenic Shock. JACC Heart Fail 2022. [DOI] [PubMed] [Google Scholar]
- 8.Taleb I, Koliopoulou AG, Tandar A et al. Shock Team Approach in Refractory Cardiogenic Shock Requiring Short-Term Mechanical Circulatory Support: A Proof of Concept. Circulation 2019;140:98–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Senman B, Jentzer JC, Barnett CF, et al. Need for a Cardiogenic Shock Team Collaborative-Promoting a Team-Based Model of Care to Improve Outcomes and Identify Best Practices. J Am Heart Assoc. 2024;13(6):e031979. doi: 10.1161/JAHA.123.031979 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hochman JS, Sleeper LA, Webb JG et al. Early revascularization in acute myocardial infarction complicated by cardiogenic shock. SHOCK Investigators. Should We Emergently Revascularize Occluded Coronaries for Cardiogenic Shock. N Engl J Med 1999;341:625–34. [DOI] [PubMed] [Google Scholar]
- 11.Altman DG. Practical statistics for medical research. 1st ed. London: Chapman and Hall, 1991. [Google Scholar]
- 12.Marashly Q, Taleb I, Kyriakopoulos CP et al. Predicting mortality in cardiogenic shock secondary to ACS requiring short-term mechanical circulatory support: The ACS-MCS score. Catheter Cardiovasc Interv 2021;98:1275–1284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.van Diepen S, Katz JN, Albert NM et al. Contemporary Management of Cardiogenic Shock: A Scientific Statement From the American Heart Association. Circulation 2017;136:e232–e268. [DOI] [PubMed] [Google Scholar]
- 14.Moghaddam N, van Diepen S, So D, Lawler PR, Fordyce CB. Cardiogenic shock teams and centres: a contemporary review of multidisciplinary care for cardiogenic shock. ESC Heart Fail 2021;8:988–998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Papolos AI, Kenigsberg BB, Berg DD et al. Management and Outcomes of Cardiogenic Shock in Cardiac ICUs With Versus Without Shock Teams. J Am Coll Cardiol 2021;78:1309–1317. [DOI] [PubMed] [Google Scholar]
- 16.Tehrani BN, Sherwood MW, Rosner C et al. A Standardized and Regionalized Network of Care for Cardiogenic Shock. JACC Heart Fail 2022;10:768–781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bloom JE, Nehme Z, Andrew E et al. HOSPITAL CHARACTERISTICS ARE ASSOCIATED WITH CLINICAL OUTCOMES IN PATIENTS WITH CARDIOGENIC SHOCK. Shock 2022;58:204–210. [DOI] [PubMed] [Google Scholar]
- 18.Lu DY, Adelsheimer A, Chan K et al. Impact of hospital transfer to hubs on outcomes of cardiogenic shock in the real world. Eur J Heart Fail 2021;23:1927–1937. [DOI] [PubMed] [Google Scholar]
- 19.Weisz G, Cox DA, Garcia E, Tcheng JE, Griffin JJ, Guagliumi G, Stuckey TD, Rutherford BD, Mehran R, Aymong E, Lansky A, Grines CL, Stone GW. Impact of smoking status on outcomes of primary coronary intervention for acute myocardial infarction--the smoker's paradox revisited. Am Heart J. 2005. Aug;150(2):358–64 [DOI] [PubMed] [Google Scholar]
- 20.Saad M, Fuernau G, Desch S, Eitel I, Jung C, Pöss J, Schneider S, Schuler G, Werdan K, Zeymer U, Thiele H. “Smoker's paradox” in patients with cardiogenic shock complicating myocardial infarction - A substudy of the IABP-SHOCK II-trial and registry. Int J Cardiol. 2016. Nov 1;222:775–779. [DOI] [PubMed] [Google Scholar]