Abstract
Background
Cardiogenic shock (CS) develops in up to 10% of patients with ST‐segment–elevation myocardial infarction and is associated with high mortality and morbidity rates. The objective of the current study was to generate a clinical scoring system that can be easily applied in the prehospital setting to predict the development of in‐hospital CS among patients undergoing primary percutaneous coronary intervention for ST‐segment–elevation myocardial infarction.
Methods
The authors conducted a retrospective cohort study using prospective data from a dual hub‐and‐spoke health system. Logistic regression was used to assess the relationship between prespecified clinical predictors and the occurrence of in‐hospital CS. Internal validation was conducted to assess the C statistic and calibration curve of the prediction model. The prediction model was converted to a risk score by scaling of the regression coefficients.
Results
From April 1, 2012, to December 31, 2020, there were 2736 consecutive patients with ST‐segment–elevation myocardial infarction undergoing primary percutaneous coronary intervention. Of these, 415 (15.2%) developed CS. Eight strong predictors were independently associated with CS by multivariable analysis and used to develop a prediction model. The model achieved a C statistic of 0.87. The EARLY SHOCK risk scoring algorithm incorporates Emergency Medical Services Heart Rate and Systolic Blood Pressure, Age, Renal Replacement, Location of Infarction, Sugar (diabetes), Heart Failure, and Cardiac Arrest.
Conclusions
The authors identified 8 clinical variables that strongly predict CS among patients with ST‐segment–elevation myocardial infarction undergoing primary percutaneous coronary intervention. This has been developed into the EARLY SHOCK score, which can be easily applied in the prehospital setting to rapidly identify CS and enable shock team activation. External validation for the scoring system is pending for broader application.
Keywords: cardiogenic shock, prehospital, shock team, STEMI
Subject Categories: Revascularization, Cardiopulmonary Resuscitation and Emergency Cardiac Care, Heart Failure, Myocardial Infarction
Nonstandard Abbreviations and Acronyms
- CS
cardiogenic shock
- DanGer Shock
Danish‐German Cardiogenic Shock
- FITT‐STEMI
Feedback Intervention and Treatment Times in ST‐Elevation Myocardial Infarction
- FMC
first medical contact
- IABP‐SHOCK II
Intraaortic Balloon Pump in Cardiogenic Shock II
- MCS
mechanical circulatory support
- ORBI
Observatoire Régional Breton sur L'Infarctus
- SCAI
Society for Cardiovascular Angiography and Interventions
- SHOCK
Should We Revascularize Occluded Coronaries for Cardiogenic Shock
- SYNTAX
Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery
- TIMI
Thrombolysis in Myocardial Infarction
- VCHA
Vancouver Coastal Health Authority
Clinical Perspective.
What Is New?
EARLY SHOCK is a novel risk scoring algorithm using only clinical variables readily available in the prehospital setting to identify patients with ST‐segment–elevation myocardial infarction at risk for developing cardiogenic shock.
What Are the Clinical Implications?
Early identification of patients with cardiogenic shock in the setting of ST‐segment–elevation myocardial infarction can facilitate activation of cardiogenic shock teams and appropriate triaging of patients to centers with mechanical circulatory support and reperfusion capabilities.
The optimal timing of shock team activation and institution of mechanical circulatory support in patients with concurrent ST‐segment–elevation myocardial infarction remain to be studied.
Cardiogenic shock (CS) is a complication of acute coronary syndrome associated with high mortality and morbidity rates. 1 , 2 , 3 , 4 Large contemporary series have reported the incidence of CS to be between 7.9% and 10% among patients with ST‐segment–elevation myocardial infarction (STEMI). 4 , 5 , 6 , 7 , 8 Mortality is high and occurs early following the onset of CS in the setting of STEMI (STEMI‐CS).ˡ−4 In‐hospital and 30‐day mortality rates range from 29% to 47% in large contemporary series. 5 , 6 , 7 , 8 , 9 , 10 Early culprit vessel revascularization in the setting of STEMI‐CS improves outcomes as shown in the landmark SHOCK (Should We Revascularize Occluded Coronaries for Cardiogenic Shock) trial in 1999. 11 Since then, the Society for Cardiovascular Angiography and Interventions (SCAI) SHOCK staging system has improved the systematic classification of patients with CS and improved ability to study their outcomes. 12 The recent DanGer Shock (Danish‐German Cardiogenic Shock) study highlights a promising, evidence‐based intervention (Impella CP) in addition to protocol‐driven care in STEMI‐CS, significantly reducing mortality at 180 days, with an absolute risk reduction of 12.7% and a relative risk reduction of 26%. 13 Despite the encouraging results, mortality from CS remains high in the DanGer shock study, at 45.8% in the microaxial‐flow‐pump group. The spoke‐and‐hub CS system of care centers on tertiary hospitals capable of providing such interventions, and timely triage of patients is fundamental to improving patient outcome.
Early identification of CS and treatment of the underlying pathology is crucial to the management of these patients. Several North American observational studies have examined the outcomes following implementation of protocolized shock teams within a single health system. 14 In‐hospital and 30‐day survival rates were significantly improved following implementation of a protocol‐driven shock team. The impact of the timing of shock team involvement on its efficacy remains to be systematically evaluated. Common features among these protocols were the early identification of patients with CS, early activation of cardiac catheterization laboratories, and routine use of invasive hemodynamic monitoring. 15 The Critical Care Cardiology Trials Network compared management and mortality outcomes between 10 centers across North America with shock teams and 14 centers without shock teams. Centers with shock teams were more likely to have invasive hemodynamics, use advanced mechanical circulatory support (MCS), and achieve lower mortality rates.
While the strict definition of CS is based on hemodynamic variables such as cardiac index, these parameters are rarely available on initial presentation, including first medical contact (FMC) in the prehospital setting. 4 Rapid identification of patients at risk of STEMI‐CS is crucial to the early activation of shock teams and timely implementation of downstream processes. We propose a clinical scoring system (EARLY SHOCK score), which can be easily applied in the prehospital setting to predict the development of in‐hospital CS among patients with STEMI treated with a regional protocol, with primary percutaneous coronary intervention (PCI) as the initial revascularization strategy.
METHODS
The data that support the findings of this study are available from the corresponding author on reasonable request.
Patients and Data Collection
All patients 18 years or older with STEMI referred to the Vancouver Coastal Health Authority (VCHA) for primary PCI between April 1, 2012, and December 31, 2020, were included, as previously described. 1 , 3 The study period was chosen as prehospital cardiac arrest data became routinely available after April 1, 2012. Patients with STEMI were stratified into whether they developed CS at any point during the index admission. This included those who were diagnosed with shock on initial admission and those who subsequently developed CS in‐hospital. Data were obtained from the VCHA STEMI database, which provides continuous and prospective collection of detailed prehospital and in‐hospital information on consecutive patients presenting to all hospitals within this health system, which serves 25% of the urban and rural population in British Columbia, Canada. One tertiary center within VCHA serves as the PCI‐capable hospital in this hub‐and‐spoke model. Prehospital care in the VCHA region includes advanced life support–trained paramedic units, who are dispatched to 9‐1‐1 calls involving chest pain, and who have the capacity to perform prehospital ECGs that are transmitted to the 2 STEMI centers in the region for emergency physician interpretation and possible activation of the STEMI protocol while the patient is en route. Institutional review board approval was obtained through the University of British Columbia's Behavioral Research Ethics Board. Informed consent was not required from patients.
Definitions
The primary outcome was the occurrence of CS at any point during the index admission. CS was defined as the presence of sustained hypotension with systolic blood pressure (SBP) <90 mm Hg for >30 minutes, with EITHER (a) signs of poor organ perfusion such as any of cool and clammy skin, altered mentation, elevated lactate, low urine output, narrow pulse pressure, or low mixed venous oxygen saturation thought to be resulting from myocardial dysfunction and in the absence of hypovolemia OR (b) the requirement for parenteral inotropic or vasopressor agents or MCS. As invasive hemodynamic data are rarely available on initial presentation, a clinical diagnosis of CS was used rather than the strict definition by cardiac index and intracardiac filling pressures. Of note, transient episodes of hypotension reversible with intravenous fluids or atropine do not constitute CS. Successful PCI was defined by the primary operator per 2011 SCAI guideline. 16 Other definitions of relevant clinical end points are detailed in Table S1.
Statistical Analysis
Potential clinical predictors were prespecified for analysis based on their availability to prehospital emergency healthcare providers. These included age, sex, cardiac arrest on presentation, prior heart failure hospitalization or documented left ventricular systolic dysfunction (left ventricular ejection fraction <50%), prior atherosclerotic cardiovascular disease defined as coronary artery disease, prior myocardial infarction, PCI or coronary artery bypass grafting, anterior location of ST‐segment elevation on initial 12‐lead ECG, diabetes, current renal replacement therapy (dialysis), and initial heart rate and blood pressure measured by emergency medical services. Renal replacement therapy and diabetes were defined as binary variables. The location of ST‐segment elevation was defined based on the first high‐quality 12‐lead ECG obtained by a prehospital healthcare provider or in the emergency department.
Univariable and multivariable logistic regression were used to assess the relationship between candidate variables and the occurrence of CS. Potential predictors based on prior clinical knowledge and availability to emergency medical services were included in the univariable analysis. Multivariable analysis was then performed to discern interactions between variables. Patients with missing data were excluded from the analysis. To determine which variables to include in the final prediction model, a stepwise selection procedure based on Akaike information criterion was used. The discrimination of the final logistic prediction model was assessed using area under the receiver operating characteristic curve (ie, C statistic) and the goodness of fit of the model was assessed using Hosmer‐Lemeshow statistic. Internal validation based on bootstrapping (1000 samples) was conducted to assess the optimism‐corrected C statistic and calibration curve of the prediction model. The prediction model was converted into a risk score by dividing all of the regression coefficients by the smallest regression coefficient in the model to make the coefficients closer to integer value before rounding them off. Youden index, which maximized the sum of sensitivity and specificity, was used to determine the optimal cutoff for identifying patients at high risk for developing CS at any point during their hospital admission. The diagnostic accuracy of the EARLY SHOCK risk score was assessed using sensitivity and specificity. To increase the ease of use of the scoring algorithm, continuous variables were categorized based on quintiles (age, initial HR, initial SBP), and/or prior clinical knowledge.
RESULTS
Patient Characteristics
From April 1, 2012, to December 31, 2020, there were 3138 consecutive patients with STEMI included in the regional STEMI protocol. Of these, 357 did not receive PCI as the initial revascularization strategy and 45 had unknown CS status. Of the 357 patients, 70 underwent medical management, 286 received lysis, and 1 had unknown management. These patients were excluded from the analysis and the final study sample included 2736 patients who received successful primary PCI. Among the 2736 patients, 415 had CS (15.2%). Of these, 61 were diagnosed with CS on initial admission that resolved following admission and PCI, 111 did not present with initial CS but later developed CS during hospitalization, and 243 had CS both on admission before PCI and during the hospitalization after revascularization. The patient flow is summarized in Figure 1.
Figure 1. Patient flow chart.
All patients with STEMI referred for PCI at the VCHA between April 2012 and December 2020 were included for analysis. Forty‐five patients with unknown CS status were excluded from the analysis. Seventy patients were medically managed at the discretion of the primary treating cardiologist and interventional cardiologist due to contraindications to angiography based on comorbidities. CS indicates cardiogenic shock; PCI, percutaneous coronary intervention; STEMI, ST‐segment–elevation myocardial infarction; and VCHA, Vancouver Coastal Health Authority.
Baseline characteristics are shown in Table 1. Compared with those with STEMI, patients with STEMI‐CS were more likely to be older, to be currently undergoing dialysis, and to have a history of heart failure, stroke, and atherosclerotic cardiovascular disease. Patients with STEMI‐CS had a higher heart rate and lower SBP on hospital arrival, as well as lower median hemoglobin. Patients with STEMI‐CS were also more likely to have prehospital cardiac arrest and anterior location of infarct. Patients with shock were more likely to be transported by ambulance or via interhospital transfer, with only 6.3% self‐presenting. The median time of symptom onset to FMC was shorter in those with CS (32.5 minutes [interquartile range (IQR), 13.0–112.0 minutes]) compared with those without shock (64.5 minutes [IQR, 30.0–164.0 minutes], P<0.001).
Table 1.
Baseline Characteristics of the Study Population
| Variable | All (N=2736) | Cardiogenic shock | P value | |
|---|---|---|---|---|
| No (n=2321) | Yes (n=415) | |||
| Age, y | <0.001 | |||
| Mean ±SD | 66.0±12.6 | 65.5±12.4 | 68.6±13.1 | |
| Range | (25.3–105.6) | (27.3–105.6) | (25.3–99.1) | |
| Age range, y, n (%) * | <0.001 | |||
| <55 | 541 (19.8) | 472 (20.3) | 69 (16.6) | |
| 55–59 | 368 (13.5) | 324 (14.0) | 44 (10.6) | |
| 60–69 | 856 (31.3) | 739 (31.8) | 117 (28.2) | |
| 70–79 | 566 (20.7) | 469 (20.2) | 97 (23.4) | |
| ≥80 | 405 (14.8) | 317 (13.7) | 88 (21.2) | |
| BMI | 0.035 | |||
| Missing, n | 27 | 12 | 15 | |
| Mean ±SD | 26.7±5.7 | 26.8±5.9 | 26.1±4.5 | |
| Range | (12.6–183.0) | (12.6–183.0) | (14.9–45.5) | |
| BMI, n (%) | 0.028 | |||
| Unknown | 27 | 12 | 15 | |
| <18.5 | 51 (1.9) | 39 (1.7) | 12 (3.0) | |
| 18.5–24.9 | 1002 (37.0) | 856 (37.1) | 146 (36.5) | |
| 25–29.9 | 1163 (42.9) | 977 (42.3) | 186 (46.5) | |
| ≥30 | 493 (18.2) | 437 (18.9) | 56 (14.0) | |
| Men, n (%) | 2187/2736 (79.9) | 1861/2321 (80.2) | 326/415 (78.6) | 0.446 |
| Current/recent smoker, n (%) | 656/2714 (24.2) | 559/2316 (24.1) | 97/398 (24.4) | 0.919 |
| Recent cocaine use, n (%) | 57/2711 (2.1) | 43/2304 (1.9) | 14/407 (3.4) | 0.041 |
| Dyslipidemia, n (%) | 1225/2721 (45.0) | 1053/2320 (45.4) | 172/401 (42.9) | 0.354 |
| Hypertension, n (%) | 1583/2721 (58.2) | 1337/2319 (57.7) | 246/402 (61.2) | 0.184 |
| Currently undergoing dialysis, n (%) | 23/2724 (0.8) | 12/2320 (0.5) | 11/404 (2.7) | <0.001 |
| Diabetes, n (%) | 646/2720 (23.8) | 529/2319 (22.8) | 117/401 (29.2) | 0.006 |
| Prior heart failure, n (%) | 92/2716 (3.4) | 53/2317 (2.3) | 39/399 (9.8) | <0.001 |
| Prior TIA/CVA, n (%) | 237/2720 (8.7) | 182/2318 (7.9) | 55/402 (13.7) | <0.001 |
| Prior ASCVD (any of the below), n (%) | 491/2719 (18.1) | 390/2318 (16.8) | 101/401 (25.2) | <0.001 |
| Prior MI, n (%) | 422/2718 (15.5) | 341/2319 (14.7) | 81/399 (20.3) | 0.004 |
| Prior PCI, n (%) | 336/2722 (12.3) | 279/2320 (12.0) | 57/402 (14.2) | 0.226 |
| Prior CABG, n (%) | 71/2724 (2.6) | 58/2320 (2.5) | 13/404 (3.2) | 0.403 |
| History of peripheral arterial disease, n (%) | 95/2719 (3.5) | 67/2318 (2.9) | 28/401 (7.0) | <0.001 |
| HR on presentation, beats per min | <0.001 | |||
| Missing, n | 7 | 3 | 4 | |
| Median (IQR) | 77.0 (63.0–92.0) | 76.0 (63.0–90.0) | 86.0 (63.0–105.0) | |
| Range | (20.0–220.0) | (20.0–220.0) | (24.0–190.0) | |
| HR on presentation, beats per min, n (%)a | <0.001 | |||
| Unknown | 7 | 3 | 4 | |
| <90 | 483 (17.7) | 401 (17.3) | 82 (20.0) | |
| 60–69 | 505 (18.5) | 460 (19.8) | 45 (10.9) | |
| 70–79 | 493 (18.1) | 449 (19.4) | 44 (10.7) | |
| 80–99 | 771 (28.3) | 669 (28.9) | 102 (24.8) | |
| ≥100 | 477 (17.5) | 339 (14.6) | 138 (33.6) | |
| First measured SBP, mm Hg | <0.001 | |||
| Missing, n | 11 | 5 | 6 | |
| Median (IQR) | 140.0 (118.0–162.0) | 142.0 (123.0–164.0) | 110.0 (89.0–138.0) | |
| Range | (40.0–282.0) | (47.0–282.0) | (40.0–252.0) | |
| First measured SBP, mm Hg, n (%)a | <0.001 | |||
| Unknown | 11 | 5 | 6 | |
| <90 | 197 (7.2) | 91 (3.9) | 106 (25.9) | |
| 90–109 | 293 (10.8) | 197 (8.5) | 96 (23.5) | |
| 110–129 | 515 (18.9) | 446 (19.3) | 69 (16.9) | |
| 130–149 | 688 (25.2) | 620 (26.8) | 68 (16.6) | |
| 150–169 | 516 (18.9) | 478 (20.6) | 38 (9.3) | |
| ≥170 | 516 (18.9) | 484 (20.9) | 32 (7.8) | |
| Admission hemoglobin, g/L | <0.001 | |||
| Missing, n | 21 | 4 | 17 | |
| Median (IQR) | 143.0 (131.0–153.0) | 144.0 (133.0–154.0) | 136.0 (123.0–150.0) | |
| Range | (42.0–1581.0) | (42.0–1581.0) | (64.0–205.0) | |
| Initial creatinine, mmol/L | <0.001 | |||
| Missing, n | 20 | 4 | 16 | |
| Median (IQR) | 95.0 (80.0–112.0) | 92.0 (78.0–107.0) | 116.0 (93.0–139.0) | |
| Range | (26.0–1392.0) | (26.0–1227.0) | (50.0–1392.0) | |
| Heart failure on presentation, n (%) | 189/2729 (6.9) | 63/2319 (2.7) | 126/410 (30.7) | <0.001 |
| Prehospital cardiac arrest, n (%) | 294/2709 (10.9) | 104/2300 (4.5) | 190/409 (46.5) | <0.001 |
| Infarct type, n (%) | <0.001 | |||
| Anterior | 1340 (49.0) | 1097 (47.3) | 243 (58.6) | |
| Nonanterior | 1396 (51.0) | 1224 (52.7) | 172 (41.4) | |
| Hospital type: initial presentation, n (%) | <0.001 | |||
| PCI capable | 1943 (71.0) | 1590 (68.5) | 353 (85.1) | |
| PCI noncapable | 793 (29.0) | 731 (31.5) | 62 (14.9) | |
| Hospital type and means of transport to first facility, n (%) | <0.001 | |||
| PCI capable and ambulance | 1528 (55.8) | 1201 (51.7) | 327 (78.8) | |
| PCI capable and self/family | 415 (15.2) | 389 (16.8) | 26 (6.3) | |
| PCI noncapable and ambulance | 199 (7.3) | 168 (7.2) | 31 (7.5) | |
| PCI noncapable and self/family | 594 (21.7) | 563 (24.3) | 31 (7.5) | |
| Means of transport to PCI‐capable facility, n (%) | <0.001 | |||
| Ambulance: direct | 1528 (55.8) | 1201 (51.7) | 327 (78.8) | |
| Self | 415 (15.2) | 389 (16.8) | 26 (6.3) | |
| Transfer | 793 (29.0) | 731 (31.5) | 62 (14.9) | |
| Symptom onset to FMC, min | <0.001 | |||
| Missing, n | 8 | 3 | 5 | |
| Median (IQR) | 60.0 (27.0–155.5) | 63.5 (30.0–164.0) | 32.5 (13.0–112.0) | |
| Range | (0.0–1512.0) | (0.0–1512.0) | (0.0–1430.0) | |
Baseline characteristics of consecutive patients with STEMI undergoing primary PCI at VCHA between April 1, 2012, and December 31, 2020. Patients are stratified by presence of cardiogenic shock. ASCVD indicates atherosclerotic cardiovascular disease; BMI, body mass index; CABG, coronary artery bypass grafting; CVA, cerebrovascular accident; FMC, first medical contact; HR, heart rate; IQR, interquartile range; MI, myocardial infarction; PCI, percutaneous coronary intervention; SBP, systolic blood pressure; STEMI, ST‐segment–elevation myocardial infarction; TIA, transient ischemic attack; and VCHA, Vancouver Coastal Health Authority.
Highlighted variables were prespecified for including into the risk scoring algorithm for ease of collection in the prehospital setting. Categories were based on prior clinical knowledge and quintiles rounded to the nearest 5 (for age) or 10 (for HR and SBP). P value was based on χ2 test, Fisher exact test, t test, or Wilcoxon rank sum test as appropriate. Patients with missing data (eg, BMI, creatinine) were excluded from the analysis. Missingness was low (<1%).
Compared with patients without CS, those with CS had worse outcomes (Table 2). Patients with STEMI‐CS had significantly higher rates of death, in‐hospital cardiac arrest, major bleeding, reinfarction, intracranial hemorrhage and stroke, heart failure, and left ventricular ejection fraction ≤40% at the time of discharge. The median hospital length of stay was 9.4 days (IQR, 5.5–19.7 days) in patients with STEMI‐CS compared with 3.0 days (IQR, 2.4–3.9 days) in those with STEMI only. Despite presenting early after symptom onset, patients with CS were more likely to experience delays in FMC‐to‐device times above recommended guidelines of within 90 minutes at PCI‐capable hospitals and 120 minutes at PCI‐noncapable hospitals. 17 Among patients with STEMI‐CS presenting to PCI‐capable hospitals, only 25.7% achieved an FMC‐to‐device time of ≤90 minutes as compared with 44.9% with STEMI only (odds ratio [OR], 0.43 [95% CI, 0.31–0.57], P<0.001).
Table 2.
Patient Outcomes Stratified by Development of Cardiogenic Shock
| Outcomes | All (N=2736) | Cardiogenic shock | P value | |
|---|---|---|---|---|
| No (n=2321) | Yes (n=415) | |||
| In‐hospital cardiac arrest, n (%) | 383 of 2710 (14.1) | 117 of 2300 (5.1) | 266 of 410 (64.9) | <0.001 |
| Major bleeding, n (%)† | 416 of 2736 (15.2) | 260 of 2321 (11.2) | 156 of 415 (37.6) | <0.001 |
| Reinfarction, n (%) | 15 of 2731 (0.5) | 5 of 2320 (0.2) | 10 of 411 (2.4) | <0.001 |
| ICH/CVA/stroke, n (%) | 54 of 2735 (2.0) | 21 of 2320 (0.9) | 33 of 415 (8.0) | <0.001 |
| Heart failure, n (%) | 557 of 2736 (20.4) | 241 of 2321 (10.4) | 316 of 415 (76.1) | <0.001 |
| Deceased, n (%) | 219 of 2736 (8.0) | 12 of 2321 (0.5) | 207 of 415 (49.9) | <0.001 |
| LVEF closest to discharge ≤40, n (%) | 820 of 2655 (30.9) | 600 of 2288 (26.2) | 220 of 367 (59.9) | <0.001 |
| LVEF closest to discharge | <0.001 | |||
| Missing, n | 81 | 33 | 48 | |
| Mean±SD | 46.6±11.0 | 47.9±9.9 | 38.1±13.3 | |
| Range | (10.0–74.0) | (15.0–74.0) | (10.0–65.0) | |
| Hospital length of stay, d * | <0.001 | |||
| Missing, n | 4 | 4 | 0 | |
| Median (IQR) | 3.1 (2.4–4.6) | 3.0 (2.4–3.9) | 9.4 (5.5–19.7) | |
| Range | (0.3–373.2) | (0.3–373.2) | (1.8–202.8) | |
Outcome of patients with STEMI undergoing primary PCI stratified by the development of cardiogenic shock. P value was based on χ2 test, Fisher exact test, t test, or Wilcoxon rank sum test as appropriate. CVA indicates cerebrovascular accident; ICH, intracranial hemorrhage; IQR, interquartile range; LVEF, left ventricular ejection fraction; PCI, percutaneous coronary intervention; and STEMI, ST‐segment–elevation myocardial infarction.
Among patients who were discharged alive.
Major bleeding was defined as hemoglobin drop of at least 3 g/dL, hematocrit drop of at least 10%, transfusion of whole blood or packed red blood cells or if surgical/procedural intervention was required for bleeding. Other definitions of end points are listed in Table S2.
Early Predictors of In‐Hospital CS
Eight predictors were independently associated with development of CS by multivariable analysis. These are shown in Table 3 and Figure 2. Age, heart rate, and SBP were represented as categorical variables as shown. Three separate prediction models were generated and are shown in Table 4. Prediction model 1 included all predictors in the multivariable model as candidate predictors and the stepwise algorithm retained 8 predictors in the final model. Model 2 excluded prehospital cardiac arrest as the candidate predictor. Model 3 included only numerical variables of age, heart rate, and blood pressure on initial presentation as candidate predictors. The C statistic of the logistic regression models are as shown in Table 4, with the most comprehensive model 1 having a C statistic of 0.87 (95% CI, 0.86–0.89). The P value of the Hosmer‐Lemeshow goodness‐of‐fit test was 0.11, with nonsignificant P values indicating appropriate calibration. Cardiac arrest is a strong predictor of CS following STEMI. Excluding cardiac arrest, model 2 still demonstrated a C statistic of high discriminatory ability at 0.80 (95% CI, 0.78–0.83).
Table 3.
Univariable and Multivariable Predictors of Cardiogenic Shock
| Univariable | P value | Multivariable | P value | ||
|---|---|---|---|---|---|
| OR (95% CI) | OR (95% CI) | ||||
| Predictors | |||||
| Age (per 10‐y increase) | 1.21 (1.12–1.32) | <0.001 | … | ||
| Age, y | |||||
| <55 | 1 | 1 | |||
| 55–59 | 0.93 (0.62–1.39) | 0.721 | 0.91 (0.53–1.54) | 0.718 | |
| 60–69 | 1.08 (0.79–1.49) | 0.624 | 1.45 (0.95–2.22) | 0.086 | |
| 70–79 | 1.41 (1.01–1.98) | 0.042 | 1.85 (1.17–2.91) | 0.009 | |
| ≥80 | 1.90 (1.34–2.68) | <0.001 | 2.48 (1.53–4.00) | <0.001 | |
| Female vs male | 1.10 (0.86–1.43) | 0.446 | 1.16 (0.83–1.64) | 0.385 | |
| Currently undergoing dialysis | 5.39 (2.36–12.29) | <0.001 | 4.53 (1.47–13.98) | 0.009 | |
| Diabetes | 1.39 (1.10–1.77) | 0.006 | 1.34 (0.98–1.82) | 0.065 | |
| Prior heart failure | 4.63 (3.02–7.10) | <0.001 | 2.56 (1.41–4.64) | 0.002 | |
| Prior ASCVD | 1.66 (1.30–2.14) | <0.001 | 1.13 (0.80–1.59) | 0.496 | |
| HR on presentation (per 10 beats per min increase) | 1.13 (1.09–1.18) | <0.001 | … | ||
| HR on presentation, beats per min | |||||
| <60 | 2.09 (1.41–3.08) | <0.001 | 1.66 (1.00–2.78) | 0.052 | |
| 60–69 | 1.00 (0.65–1.54) | 0.994 | 1.19 (0.68–2.06) | 0.542 | |
| 70–79 | 1 | 1 | |||
| 80–99 | 1.56 (1.07–2.26) | 0.020 | 1.88 (1.16–3.05) | 0.010 | |
| ≥100 | 4.15 (2.88–6.00) | <0.001 | 3.25 (2.00–5.29) | <0.001 | |
| First measured SBP, per 10‐mm Hg decrease | 1.32 (1.28–1.37) | <0.001 | … | ||
| First measured SBP, mm Hg | |||||
| <90 | 17.62 (11.19–27.75) | <0.001 | 22.41 (12.86–39.05) | <0.001 | |
| 90–109 | 7.37 (4.78–11.36) | <0.001 | 8.65 (5.10–14.68) | <0.001 | |
| 110–129 | 2.34 (1.51–3.63) | <0.001 | 2.59 (1.54–4.35) | <0.001 | |
| 130–149 | 1.66 (1.07–2.57) | 0.023 | 2.01 (1.21–3.34) | 0.007 | |
| 150–169 | 1.20 (0.74–1.96) | 0.458 | 1.46 (0.83–2.56) | 0.184 | |
| ≥170 | 1 | 1 | |||
| Prehospital cardiac arrest | 18.32 (13.89–24.15) | <0.001 | 17.56 (12.53–24.62) | <0.001 | |
| Infarct type: anterior vs nonanterior | 1.58 (1.28–1.95) | <0.001 | 1.60 (1.21–2.12) | 0.001 | |
| Goodness of fit | |||||
| Area under the ROC curve (95% CI) | … | 0.87 (0.86–0.90) | |||
| Area under the ROC curve (optimism corrected) | … | 0.87 | |||
| Hosmer and Lemeshow goodness‐of‐fit test (P value) * | … | 0.18 | |||
Univariable and multivariable predictors of cardiogenic shock. There may be confounding between variables, and, in the multivariable model, some factors may lose significance (eg, prior ASCVD). After multivariable analysis, 8 clinical variables were concluded to be independently associated with the development of cardiogenic shock. These were initial HR and SBP on presentation (emergency medical services), age, dialysis, diabetes, history of heart failure, prehospital cardiac arrest, and anterior location of infarction. ASCVD indicates atherosclerotic cardiovascular disease; HR, heart rate; OR, odds ratio; ROC, receiver operating characteristic curve; and SBP, systolic blood pressure.
A small P value (P<0.05) suggests a poor fit. Nonsignificant P values indicate appropriate calibration in the Hosmer and Lemeshow goodness‐of‐fit test.
Figure 2. Forest plot of univariable and multivariable predictors of CS.
CS indicates cardiogenic shock.
Table 4.
Prediction Models of Cardiogenic Shock Using Early Predictors
| Model 1 (n=2675) | Model 2 (n=2701) | Model 3 (n=2725) | ||||
|---|---|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | |
| Predictors | ||||||
| Age, y | ||||||
| <55 | 1 | 1 | 1 | |||
| 55–59 | 0.91 (0.54–1.55) | 0.739 | 0.79 (0.50–1.26) | 0.322 | 0.83 (0.53–1.30) | 0.423 |
| 60–69 | 1.49 (0.98–2.28) | 0.063 | 1.03 (0.72–1.48) | 0.855 | 1.07 (0.75–1.52) | 0.711 |
| 70–79 | 1.91 (1.22–2.99) | 0.005 | 1.35 (0.92–1.98) | 0.124 | 1.48 (1.02–2.14) | 0.039 |
| ≥80 | 2.64 (1.66–4.18) | <0.001 | 1.51 (1.01–2.25) | 0.044 | 1.68 (1.14–2.47) | 0.008 |
| Female vs male | … | … | … | |||
| Currently undergoing dialysis | 4.58 (1.48–14.14) | 0.008 | 4.25 (1.50–12.10) | 0.007 | … | |
| Diabetes | 1.38 (1.01–1.87) | 0.041 | 1.29 (0.98–1.70) | 0.066 | … | |
| Prior heart failure | 2.67 (1.49–4.79) | <0.001 | 2.80 (1.67–4.68) | <0.001 | … | |
| HR on presentation, beats per min | ||||||
| <60 | 1.60 (0.96–2.66) | 0.070 | 1.09 (0.69–1.71) | 0.713 | 0.97 (0.62–1.50) | 0.885 |
| 60–69 | 1.15 (0.66–1.98) | 0.621 | 0.92 (0.57–1.49) | 0.734 | 0.93 (0.58–1.48) | 0.750 |
| 70–79 | 1 | 1 | 1 | |||
| 80–99 | 1.83 (1.13–2.95) | 0.014 | 1.70 (1.12–2.57) | 0.012 | 1.79 (1.20–2.68) | 0.005 |
| ≥100 | 3.20 (1.98–5.18) | <0.001 | 3.81 (2.52–5.77) | <0.001 | 4.52 (3.02–6.77) | <0.001 |
| First measured SBP, mm Hg | ||||||
| <90 | 22.82 (13.11–39.72) | <0.001 | 24.26 (14.66, 40.12) | <0.001 | 23.93 (14.67–39.06) | <0.001 |
| 90–109 | 8.90 (5.25–15.09) | <0.001 | 10.09 (6.30–16.16) | <0.001 | 9.68 (6.14–15.25) | <0.001 |
| 110–129 | 2.60 (1.55–4.38) | <0.001 | 2.79 (1.75–4.46) | <0.001 | 2.93 (1.87–4.60) | <0.001 |
| 130–149 | 2.02 (1.22–3.36) | 0.006 | 2.02 (1.27–3.19) | 0.003 | 1.97 (1.26–3.08) | 0.003 |
| 150–169 | 1.47 (0.84–2.58) | 0.174 | 1.30 (0.78–2.17) | 0.307 | 1.30 (0.79–2.13) | 0.298 |
| ≥170 | 1 | 1 | 1 | |||
| Prehospital cardiac arrest | 17.20 (12.30–24.05) | <0.001 | … | … | ||
| Infarct type: anterior vs nonanterior | 1.57 (1.19–2.08) | 0.002 | 1.73 (1.34–2.23) | <0.001 | … | |
| Goodness of fit | ||||||
| Area under the ROC curve (95% CI) | 0.87 (0.86–0.89) | 0.80 (0.78–0.83) | 0.79 (0.76–0.81) | |||
| Area under the ROC curve (optimism corrected) | 0.87 | 0.79 | 0.78 | |||
| Hosmer and Lemeshow goodness‐of‐fit test (P value)* | 0.11 | 0.28 | 0.88 | |||
Three separate prediction models were generated. Model 1 includes all clinical variables identified in the multivariable analysis as candidate predictors, and the stepwise algorithm retained all 8 predictors in the final model. Model 2 excludes prehospital cardiac arrest as a candidate predictor. Model 3 includes only the numerical variables of age, HR, and SBP on presentation. HR indicates heart rate; OR, odds ratio; ROC, receiver operating characteristic; and SBP, systolic blood pressure.
A small P value (P<0.05) suggests a poor fit. Nonsignificant P values indicate appropriate calibration in the Hosmer and Lemeshow goodness‐of‐fit test.
Risk Score Generation
A risk scoring algorithm to determine early prediction of STEMI‐CS was generated using weight coefficients from the multivariable logistic regression for all models. The score is titled EARLY SHOCK, which incorporates Emergency Medical Services Vital Signs, Age, Renal Replacement, Location of Infarction, Sugar (diabetes), Heart Failure, and Cardiac Arrest. This is shown in Figure 3. The total possible scores range from 105 in model 1 to 61 in model 2 and 60 in model 3. The number of patients in each score category is shown in Table S2 in the Supplement. Table 5 illustrates the comparison of predicted versus observed incidence of in‐hospital CS among the categories of risk score. In model 1, scores >29 to 30 confers a >50% predicted probability of shock, and corresponds to observed ≥53.2% probability of CS. At an EARLY SHOCK score of ≥42, the predicted probability of CS is ≥90% and corresponds to an observed probability of 92.7%. Sensitivity analysis excluding those with initial SBP <90 mm Hg (first blood pressure measured by emergency medical services) is shown in the last column of Table 5. We found that this did not significantly impact model performance. Figure 4 illustrates the relationship between the EARLY SHOCK score and the predicted incidence of CS (Y: predicted probability; X: score). In model 1, scores ≥18, the optimal cutoff by Youden index, were associated with a sensitivity of 0.79 and a specificity of 0.83. In model 2, scores ≥10 were associated with a sensitivity of 0.72 and a specificity of 0.75. In model 3, scores >12 were associated with a sensitivity of 0.67 and a specificity of 0.79.
Figure 3. The EARLY SHOCK Scoring Algorithm.
The EARLY SHOCK scoring algorithm was generated using weight coefficients from the multivariable logistic regression for all models. The score incorporates Emergency Medical Services Vital Signs, Age, Renal Replacement, Location of Infarction, Sugar (diabetes), Heart Failure, and Cardiac Arrest. HR indicates heart rate; and SBP, systolic blood pressure.
Table 5.
Predicted vs Observed Probability of Cardiogenic Shock
| Total score | Predicted probability of shock | Observed probability of shock (95% CI) | |
|---|---|---|---|
| All patients | Excluded hypotensive patients | ||
| Model 1 | |||
| 0–11 | <5% | 27 of 1225=2.2% (1.5–3.2%) | 27 of 1225=2.2% (1.5–3.2%) |
| 12–16 | 5%–9% | 46 of 681=6.8% (5.0–8.9%) | 46 of 681=6.8% (5.0–8.9%) |
| 17–20 | 10%–19% | 37 of 222=16.7% (12.0–22.2%) | 36 of 219=16.4% (11.8–22%) |
| 21–23 | 20%–29% | 38 of 123=30.9% (22.9–39.9%) | 32 of 103=31.1% (22.3–40.9%) |
| 24–26 | 30%–39% | 43 of 106=40.6% (31.1–50.5%) | 24 of 66=36.4% (24.9–49.1%) |
| 27–28 | 40%–49% | 39 of 95=41.1% (31.1–51.6%) | 23 of 48=47.9% (33.3–62.8%) |
| 29–30 | 50%–59% | 25 of 47=53.2% (38.1–67.9%) | 19 of 36=52.8% (35.5–69.6%) |
| 31–33 | 60%–69% | 37 of 62=59.7% (46.4–71.9%) | 28 of 45=62.2% (46.5–76.2%) |
| 34–36 | 70%–79% | 21 of 29=72.4% (52.8–87.3%) | 16 of 22=72.7% (49.8–89.3%) |
| 37–41 | 80%–89% | 36 of 44=81.8% (67.3–91.8%) | 26 of 30=86.7% (69.3–96.2%) |
| ≥42 | ≥90% | 38 of 41=92.7% (80.1–98.5%) | 9 of 11=81.8% (48.2–97.7%) |
| Model 2 | |||
| 0–5 | <5% | 30 of 864=3.5% (2.4–4.9%) | 30 of 864=3.5% (2.4–4.9%) |
| 6–8 | 5%–9% | 59 of 821=7.2% (5.5–9.2%) | 59 of 821=7.2% (5.5–9.2%) |
| 9–11 | 10%–19% | 63 of 406=15.5% (12.1–19.4%) | 63 of 406=15.5% (12.1–19.4%) |
| 12–13 | 20%–29% | 48 of 212=22.6% (17.2–28.9%) | 48 of 212=22.6% (17.2–28.9%) |
| 14–15 | 30%–39% | 39 of 118=33.1% (24.7–42.3%) | 37 of 109=33.9% (25.1–43.6%) |
| 16–17 | 40%–49% | 43 of 92=46.7% (36.3–57.4%) | 14 of 33=42.4% (25.5–60.8%) |
| 18–19 | 50%–59% | 53 of 99=53.5% (43.2–63.6%) | 22 of 38=57.9% (40.8–73.7%) |
| 20–21 | 60%–69% | 25 of 42=59.5% (43.3–74.4%) | 7 of 15=46.7% (21.3–73.4%) |
| 22–23 | 70%–79% | 22 of 32=68.8% (50–83.9%) | 6 of 6=100% (−) |
| 24–27 | 80%–89% | 7 of 11=63.6% (30.8–89.1%) | 2 of 2=100% (−) |
| ≥28 | ≥90% | 3 of 4=75% (19.4–99.4%) | 1 of 1=100% (−) |
| Model 3 | |||
| 0–4 | <5% | 25 of 624=4% (2.6–5.9%) | 25 of 624=4% (2.6–5.9%) |
| 5–8 | 5%–9% | 64 of 942=6.8% (5.3–8.6%) | 64 of 942=6.8% (5.3–8.6%) |
| 9–12 | 10%–19% | 66 of 510=12.9% (10.2–16.2%) | 66 of 510=12.9% (10.2–16.2%) |
| 13–15 | 20%–29% | 68 of 281=24.2% (19.3–29.6%) | 68 of 281=24.2% (19.3–29.6%) |
| 16–17 | 30%–39% | 32 of 98=32.7% (23.5–42.9%) | 27 of 84=32.1% (22.4–43.2%) |
| 18–20 | 40%–49% | 66 of 119=55.5% (46.1–64.6%) | 23 of 38=60.5% (43.4–76%) |
| 21 | 50%–59% | 43 of 76=56.6% (44.7–67.9%) | 17 of 29=58.6% (38.9–76.5%) |
| 23–24 | 60%–69% | 26 of 38=68.4% (51.3–82.5%) | 13 of 20=65% (40.8–84.6%) |
| 25–26 | 70%–79% | 12 of 24=50% (29.1–70.9%) | 0 of 0=− |
| 28–29 | ≥80% | 7 of 13=53.8% (25.1–80.8%) | 0 of 0=− |
Figure 4. Relationship between EARLY SHOCK predicted and observed incidence of cardiogenic shock.
Comparison of predicted vs observed incidence of in‐hospital cardiogenic shock across the range of risk score. 95% CI for the observed incidence is presented as error bars.
DISCUSSION
In the current analysis, nearly 1 in 6 patients with STEMI treated with primary PCI developed CS. Compared with patients without CS, patients with STEMI‐CS have a significantly higher proportion of death, in‐hospital cardiac arrest, major bleeding, reinfarction, intracranial hemorrhage, stroke, and heart failure. The development of CS was associated with delays in FMC‐to‐device time above national guidelines. Eight easily identifiable prehospital clinical predictors were independently associated with the development of CS. Using these variables, 3 prediction models were generated with C statistics ranging from 0.80 to 0.87. A risk scoring algorithm was then generated using weight coefficients from all models, which may be useful in the prehospital setting for early activation of CS teams. Following appropriate external validation, these 8 prehospital variables could be easily provided in a pocket card reference or phone application for ease of adoption.
CS is a syndrome whereby myocardial dysfunction leads to reduced cardiac output and organ hypoperfusion. Traditionally, CS is defined as hypotension and organ hypoperfusion with low cardiac output (defined as a cardiac index <2.2 L/min per m2 on inotropes and a cardiac index <1.8 L/min per m2 without inotropes) in the presence of adequate filling pressures. 18 However, different phenotypes of CS are increasingly recognized. Patients can present with normotensive shock, whereby a compensatory increase in systemic vascular resistance in response to low cardiac output preserves blood pressure leading to organ hypoperfusion. 19 , 20 Patients can also present in a preshock state with hypotension but preserved organ perfusion. 20 A recent study examining mortality in the cardiac intensive care unit of patients with CS found that isolated hypotension and hypoperfusion are associated with increased mortality. The combination of both hypotension and hypoperfusion is associated with the highest mortality, followed by isolated hypoperfusion. 21
While hypotension on initial presentation may be an important marker of CS, it is not always present in patients who eventually develop CS. Whereas we found that the strongest predictor of CS was an initial SBP <90 mm Hg, an initial SBP of 130 to 149 mm Hg, although in the range conventionally considered to be normal, was also predictive of the eventual development of CS (multivariable OR, 2.01 [95% CI, 1.21–3.34], P=0.007). As SBP falls, the strength of the association with CS becomes stronger, such that an initial SBP of 90 to 109 mm Hg is associated with an OR of 8.65 (95% CI, 5.10–14.68; P<0.001) for the occurrence of CS. Our study supports findings from other large studies that end‐organ hypoperfusion occurs without hypotension in these patients with normotensive shock. This also presents an opportunity for improvement in the early recognition of CS. Use of the EARLY SHOCK score could allow prediction of eventual CS, even when the initial SBP is in the conventionally normal range. Such patients could be triaged appropriately to receive early subspecialty care and MCS if needed.
The strength of the EARLY SHOCK score lies in its ability to identify patients who are at risk for developing CS despite normotension on initial presentation and triaging them appropriately. In settings where the CS team and MCS are not readily available, early identification of patients in the prehospital setting is crucial to help triaging physicians appropriately direct these patients to centers with primary PCI, a CS team, and MCS capability.
This work extends that of previous analyses. In the United States, <40% of patients with STEMI‐CS achieved FMC‐to‐device time targets. 22 The FITT‐STEMI (Feedback Intervention and Treatment Times in ST‐Elevation Myocardial Infarction) study demonstrated that for every 10‐minute delay in FMC‐balloon time, an additional 3.3 deaths occurred per 100 PCI‐treated patients. 23 In the Canadian setting, Kochan et al showed that for every 10‐minute delay in reperfusion between 60 to 90 minutes of symptom onset, mortality increases by 4% to 7% in those with STEMI‐CS, compared with <0.5% in those with STEMI alone. 17 This increase in mortality is compounded by the fact that patients with STEMI‐CS are more likely to experience delays in reperfusion above nationally recommended FMC to device times as compared with patients with STEMI alone. Similar to previous studies, we found that among patients with STEMI‐CS presenting to a PCI‐capable hospital, only 25.7% achieved the target FMC‐to‐device time of ≤90 minutes as compared with 44.9% in patients without CS, P<0.001.
Several clinical prediction scores have been studied and validated in the identification of patients at risk for CS following STEMI, as well as quantifying mortality in the setting of STEMI‐CS. These include the Observato ire Régional Breton sur L'Infarctus (ORBI) score and the Intraaortic Balloon Pump in Cardiogenic Shock II (IABP‐SHOCK II) score. 7 , 24 Both scoring systems utilize a combination of clinical variables as well as angiographic characteristics including culprit lesion location, postprimary PCI Thrombolysis in Myocardial Infarction (TIMI) grade, Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery (SYNTAX) score, and cardiac output and index. While these scores are useful, their applications are limited to the time of coronary angiography.
Our study has some important clinical implications. To our knowledge, we report the first scoring system using only clinical variables readily obtainable by prehospital emergency services. This can enable even earlier recognition and management of CS, as well as set in place the necessary mechanisms for treating the underlying pathology. In the case of STEMI, this can facilitate early catheterization laboratory activation, as well as activation of shock teams to facilitate timely decisions regarding advanced circulatory supports for this group of patients with high mortality and morbidity. In light of the recent evidence supporting microaxial flow pump use in patients with STEMI‐CS, 13 early triage of patients to a CS center with MCS capabilities is crucial. Putatively, prehospital activation could then be separated into both STEMI and STEMI‐SHOCK, the latter of which could facilitate more rapid mobilization of shock teams in addition to cardiac catheterization personal. Future research about the feasibility of a 2‐tiered regional approach to STEMI is warranted.
Study Limitations
The current study is limited by its retrospective nature, which has inherent limitations such as the potential for confounders. Invasive hemodynamic parameters and markers of end‐organ perfusion were not routinely collected and hence we were not able to centrally adjudicate whether patients met criteria for CS or had features of mixed shock. A future prospective study incorporating hemodynamic variables either at the time of index angiography or via a Swan‐Ganz catheter in the cardiac intensive care unit would provide useful risk stratification of patients with STEMI‐CS. That being said, our registry included a standardized definition of CS using clinical variables that would also be easily identifiable in the prehospital setting.
Classification of shock using established definitions such as the SCAI classification would add to the analysis; however, as the SCAI CS classification was published in 2019, the majority of patients did not have this documented during their presentation. The results of the study can only be applied to patients presenting with CS in the setting of STEMI with intent to undergo primary PCI. It cannot be extrapolated to those with CS from chronic heart failure or those with non‐STEMI with CS. Last, we recognize that in our cohort only 25.7% of patients with STEMI‐CS achieved the PCI target time of ≤90 minutes, which may have contributed to worse outcomes. However, regardless of the setting and adherence to guidelines, patients with STEMI‐CS are most likely to benefit from an early shock team approach. Although not available at this time, future research will include external validation in other STEMI subsets for calibration of the scoring system.
CONCLUSIONS
CS complicated up to 15% of patients undergoing primary PCI for STEMI at a large regional health system and was associated with significantly worse outcomes both in terms of cardiovascular events as well as prolonged length of hospitalization and resource utilization. We developed the EARLY SHOCK risk scoring algorithm using only clinical variables readily available in the prehospital setting. Following external validation, this score could be easily applied in the STEMI‐primary PCI population to enable prompt prehospital activation of both STEMI and STEMI‐SHOCK teams, potentially improving outcomes in this group of patients with high mortality and morbidity.
Sources of Funding
Funding is provided by Vancouver Coastal Health Research Institute (VCHRI), Vancouver, British Columbia, Canada.
Disclosures
None.
Supporting information
Tables S1–S2
This article was sent to Krishnaraj S. Rathod, MBBS, BMedSci, MRCP, Associate Editor, for review by expert referees, editorial decision, and final disposition.
Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.124.040681
For Sources of Funding and Disclosures, see page 14.
REFERENCES
- 1. Fordyce CB, Cairns JA, Singer J, Lee T, Park JE, Vandegriend RA, Perry M, Largy W, Gao M, Ramanathan K, et al. Evolution and impact of a regional reperfusion system for ST‐elevation myocardial infarction. Can J Cardiol. 2016;32:1222–1230. doi: 10.1016/j.cjca.2015.11.026 [DOI] [PubMed] [Google Scholar]
- 2. Wenner JB, Wong GC, Cairns JA, Perry‐Arnesen M, Tocher W, Mackay M, Singer J, Lee T, Fordyce CB. Impact of patient‐ and system‐level delays on reperfusion among patients with ST‐elevation myocardial infarction. CJC Open. 2020;2:94–103. doi: 10.1016/j.cjco.2020.01.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Williams C, Fordyce CB, Cairns JA, Turgeon RD, Lee T, Singer J, Mackay M, Wong GC. Temporal trends in reperfusion delivery and clinical outcomes following implementation of a regional STEMI protocol: a 12‐year perspective. CJC Open. 2023;5:181–190. doi: 10.1016/j.cjco.2022.11.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. van Diepen S, Katz JN, Albert NM, Henry TD, Jacobs AK, Kapur NK, Kilic A, Menon V, Ohman EM, Sweitzer NK, et al. Contemporary Management of Cardiogenic Shock: a scientific statement from the American Heart Association. Circulation. 2017;136:e232–e268. doi: 10.1161/CIR.0000000000000525 [DOI] [PubMed] [Google Scholar]
- 5. Kolte D, Khera S, Aronow WS, Mujib M, Palaniswamy C, Sule S, Jain D, Gotsis W, Ahmed A, Frishman WH, et al. Trends in incidence, management, and outcomes of cardiogenic shock complicating ST‐elevation myocardial infarction in the United States. J Am Heart Assoc. 2014;3:e000590. doi: 10.1161/JAHA.113.000590 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Omer MA, Tyler JM, Henry TD, Garberich R, Sharkey SW, Schmidt CW, Henry JT, Eckman P, Megaly M, Brilakis ES, et al. Clinical characteristics and outcomes of STEMI patients with cardiogenic shock and cardiac arrest. JACC Cardiovasc Interv. 2020;13:1211–1219. doi: 10.1016/j.jcin.2020.04.004 [DOI] [PubMed] [Google Scholar]
- 7. Auffret V, Cottin Y, Leurent G, Gilard M, Beer J‐C, Zabalawi A, Chagué F, Filippi E, Brunet D, Hacot J‐P, et al. Predicting the development of in‐hospital cardiogenic shock in patients with ST‐segment elevation myocardial infarction treated by primary percutaneous coronary intervention: the ORBI risk score. Eur Heart J. 2018;39:2090–2102. doi: 10.1093/eurheartj/ehy127 [DOI] [PubMed] [Google Scholar]
- 8. Obling L, Frydland M, Hansen R, Møller‐Helgestad OK, Lindholm MG, Holmvang L, Ravn HB, Wiberg S, Thomsen JH, Jensen LO, et al. Risk factors of late cardiogenic shock and mortality in ST‐segment elevation myocardial infarction patients. Eur Heart J Acute Cardiovasc Care. 2018;7:7–15. doi: 10.1177/2048872617706503 [DOI] [PubMed] [Google Scholar]
- 9. Shroff GR, Garcia S, Schmidt C, Okeson B, Tannenbaum E, Pacheco R, Smith TD, Garberich R, Sharkey S, Aguirre F, et al. Renal impairment and mortality in patients with STEMI and cardiogenic shock/cardiac arrest. Catheter Cardiovasc Interv. 2023;102:179–190. doi: 10.1002/ccd.30753 [DOI] [PubMed] [Google Scholar]
- 10. Kunadian V, Qiu W, Ludman P, Redwood S, Curzen N, Stables R, Gunn J, Gershlick A; National Institute for cardiovascular outcomes research . Outcomes in patients with cardiogenic shock following percutaneous coronary intervention in the contemporary era: an analysis from the BCIS database (British cardiovascular intervention society). J Am Coll Cardiol Intv. 2014;7:1374–1385. doi: 10.1016/j.jcin.2014.06.017 [DOI] [PubMed] [Google Scholar]
- 11. Hochman JS, Sleeper LA, Webb JG, Sanborn TA, White HD, Talley JD, Buller CE, Jacobs AK, Slater JN, Col J, 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;341:625–634. doi: 10.1056/NEJM199908263410901 [DOI] [PubMed] [Google Scholar]
- 12. Naidu SS, Baran DA, Jentzer JC, Hollenberg SM, van Diepen S, Basir MB, Grines CL, Diercks DB, Hall S, Kapur NK, et al. SCAI SHOCK stage classification expert consensus update: a review and incorporation of validation studies: this statement was endorsed by the American College of Cardiology (ACC), American College of Emergency Physicians (ACEP), American Heart Association (AHA), European Society of Cardiology (ESC) Association for Acute Cardiovascular Care (ACVC), International Society for Heart and Lung Transplantation (ISHLT), Society of Critical Care Medicine (SCCM), and Society of Thoracic Surgeons (STS) in December 2021. J Am Coll Cardiol. 2022;79:933–946. doi: 10.1016/j.jacc.2022.01.018 [DOI] [PubMed] [Google Scholar]
- 13. Møller JE, Engstrøm T, Jensen LO, Eiskjær H, Mangner N, Polzin A, Schulze PC, Skurk C, Nordbeck P, Clemmensen P, et al. Microaxial flow pump or standard Care in Infarct‐Related Cardiogenic Shock. N Engl J Med. 2024;390:1382–1393. doi: 10.1056/NEJMoa2312572 [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: 10.1002/ehf2.13180 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Papolos AI, Kenigsberg BB, Berg DD, Alviar CL, Bohula E, Burke JA, Carnicelli AP, Chaudhry S‐P, Drakos S, Gerber DA, 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: 10.1016/j.jacc.2021.07.044 [DOI] [PubMed] [Google Scholar]
- 16. Levine GN, Bates ER, Blankenship JC, Bailey SR, Bittl JA, Cercek B, Chambers CE, Ellis SG, Guyton RA, Hollenberg SM, et al. 2011 ACCF/AHA/SCAI guideline for percutaneous coronary intervention: a report of the American College of Cardiology Foundation/American Heart Association task force on practice guidelines and the Society for Cardiovascular Angiography and Interventions. Circulation. 2011;124:e574–e651. doi: 10.1161/CIR.0b013e31823ba622 [DOI] [PubMed] [Google Scholar]
- 17. Kochan A, Lee T, Moghaddam N, Milley G, Singer J, Cairns JA, Wong GC, Jentzer JC, van Diepen S, Alviar C, et al. Reperfusion delays and outcomes among patients with ST‐segment‐elevation myocardial infarction with and without cardiogenic shock. Circ Cardiovasc Interv. 2023;16:e012810. doi: 10.1161/CIRCINTERVENTIONS.122.012810 [DOI] [PubMed] [Google Scholar]
- 18. Jentzer JC, Pöss J, Schaubroeck H, Morrow DA, Hollenberg SM, Mebazaa A. Advances in the Management of Cardiogenic Shock. Crit Care Med. 2023;51:1222–1233. doi: 10.1097/CCM.0000000000005919 [DOI] [PubMed] [Google Scholar]
- 19. Menon V, Slater JN, White HD, Sleeper LA, Cocke T, Hochman JS. Acute myocardial infarction complicated by systemic hypoperfusion without hypotension: report of the SHOCK trial registry. Am J Med. 2000;108:374–380. doi: 10.1016/S0002-9343(00)00310-7 [DOI] [PubMed] [Google Scholar]
- 20. Saxena A, Garan AR, Kapur NK, O'Neill WW, Lindenfeld J, Pinney SP, Uriel N, Burkhoff D, Kern M. Value of hemodynamic monitoring in patients with cardiogenic shock undergoing mechanical circulatory support. Circulation. 2020;141:1184–1197. doi: 10.1161/CIRCULATIONAHA.119.043080 [DOI] [PubMed] [Google Scholar]
- 21. Jentzer JC, Burstein B, Van Diepen S, Murphy J, Holmes DR, Bell MR, Barsness GW, Henry TD, Menon V, Rihal CS, et al. Defining shock and Preshock for mortality risk stratification in cardiac intensive care unit patients. Circ Heart Fail. 2021;14:e007678. doi: 10.1161/CIRCHEARTFAILURE.120.007678 [DOI] [PubMed] [Google Scholar]
- 22. Kochar A, Al‐Khalidi HR, Hansen SM, Shavadia JS, Roettig ML, Fordyce CB, Doerfler S, Gersh BJ, Henry TD, Berger PB, et al. Delays in primary percutaneous coronary intervention in ST‐segment elevation myocardial infarction patients presenting with cardiogenic shock. JACC Cardiovasc Interv. 2018;11:1824–1833. doi: 10.1016/j.jcin.2018.06.030 [DOI] [PubMed] [Google Scholar]
- 23. Scholz KH, Maier SKG, Maier LS, Lengenfelder B, Jacobshagen C, Jung J, Fleischmann C, Werner GS, Olbrich HG, Ott R, et al. Impact of treatment delay on mortality in ST‐segment elevation myocardial infarction (STEMI) patients presenting with and without haemodynamic instability: results from the German prospective, multicentre FITT‐STEMI trial. Eur Heart J. 2018;39:1065–1074. doi: 10.1093/eurheartj/ehy004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Pöss J, Köster J, Fuernau G, Eitel I, de Waha S, Ouarrak T, Lassus J, Harjola V‐P, Zeymer U, Thiele H, et al. Risk stratification for patients in cardiogenic shock after acute myocardial infarction. J Am Coll Cardiol. 2017;69:1913–1920. doi: 10.1016/j.jacc.2017.02.027 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Tables S1–S2