Evaluation of factors associated with selection for coronary angiography and in‐hospital mortality among patients presenting with out‐of‐hospital cardiac arrest without ST‐segment elevation

Abstract

Background

Clinical factors favouring coronary angiography (CA) selection and variables associated with in‐hospital mortality among patients presenting with out‐of‐hospital cardiac arrest (OHCA) without ST‐segment elevation (STE) remain unclear.

Methods

We evaluated clinical characteristics associated with CA selection and in‐hospital mortality in patients with OHCA, shockable rhythm and no STE.

Results

Between 2014 and 2018, 118 patients with OHCA and shockable rhythm without STE (mean age 59; males 75%) were stratified by whether CA was performed. Of 86 (73%) patients undergoing CA, 30 (35%) received percutaneous coronary intervention (PCI). CA patients had shorter return of spontaneous circulation (ROSC) time (17 vs. 25 min) and were more frequently between 50 and 60 years (29% vs. 6.5%), with initial Glasgow Coma Scale (GCS) score >8 (24% vs. 6%) (all p < 0.05). In‐hospital mortality was 33% (n = 39) for overall cohort (CA 27% vs. no‐CA 50%, p = 0.02). Compared to late CA, early CA ( ≤ 2 h) was not associated with lower in‐hospital mortality (32% vs. 34%, p = 0.82). Predictors of in‐hospital mortality included longer defibrillation time (odds ratio 3.07, 95% confidence interval 1.44‒6.53 per 5‐min increase), lower pH (2.02, 1.33‒3.09 per 0.1 decrease), hypoalbuminemia (2.02, 1.03‒3.95 per 5 g/L decrease), and baseline renal dysfunction (1.33, 1.02‒1.72 per 10 ml/min/1.73 m² decrease), while PCI to lesion (0.11, 0.01‒0.79) and bystander defibrillation (0.06, 0.004‒0.80) were protective factors (all p < 0.05).

Conclusions

Among patients with OHCA and shockable rhythm without STE, younger age, shorter time to ROSC and GCS >8 were associated with CA selection, while less effective resuscitation, greater burden of comorbidities and absence of treatable coronary lesion were key adverse prognostic predictors.

Abbreviations

APACHE II

Acute Physiology and Chronic Health Evaluation II

BARC

Bleeding Academic Research Consortium

CA

coronary angiography

CABG

coronary artery bypass graft

CAD

coronary artery disease

CKD

chronic kidney disease

CPR

cardiopulmonary resuscitation

CS

cardiogenic shock

GCS

Glasgow Coma Scale

ICU

intensive care unit

MACCE

major adverse cardiovascular and cerebrovascular events

OHCA

out‐of‐hospital cardiac arrest

PCI

percutaneous coronary intervention

RCT

randomized controlled trial

ROSC

return of spontaneous circulation

STE

ST‐segment elevation

1. INTRODUCTION

Despite advances in resuscitation and intensive care medicine, outcomes following out‐of‐hospital cardiac arrest (OHCA) have remained poor, with survival rates of <40% even among those successfully resuscitated. ¹ The most common cause of OHCA overall is coronary artery disease (CAD). ² Up to 70% of patients with OHCA and shockable rhythm are found to have a culprit lesion on coronary angiography (CA). ³ In our prior analysis of patients with OHCA and shockable rhythm, 67% of those with ST‐segment elevation (STE) on initial electrocardiogram (ECG) who underwent CA proceeded with ad hoc percutaneous coronary intervention (PCI) to a culprit lesion. ⁴ Current international consensus recommends immediate CA and revascularization in patients with resuscitated OHCA and STE on initial ECG (class of recommendation I, level of evidence B). ⁵ , ⁶ For OHCA patients without STE, the benefit of CA remains disputed in the contemporary literature. Observational evidence is conflicting in this cohort, with some studies indicating early CA to be associated with reduced mortality and others suggesting a lack of benefit. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² Four recent randomized controlled trials (RCTs) completed among patients without STE following OHCA have demonstrated no difference in survival whether CA was performed early or delayed. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ One RCT excluded nonshockable rhythms ¹⁶ and two RCTs were noted to be underpowered. ¹⁴ , ¹⁵ Adverse prognostic indicators have been previously described for the general cohort with OHCA and shockable rhythm, ² , ⁴ however, these variables are poorly defined among those without STE. ⁷ , ⁸ While high‐level evidence is limited, real‐world practice of patient selection for CA remains largely dependent on individual physician judgment. Therefore, we aimed to identify characteristics associated with patient selection for CA, and to determine adverse predictors of in‐hospital mortality, including the effect and timing of undertaking CA, in patients with OHCA, shockable rhythm and no STE.

2. PATIENT AND METHODS

2.1. Study design and patient population

We conducted a retrospective observational study involving two of the largest tertiary and quaternary health services—Alfred Health and Western Health—in Melbourne (Victoria, Australia). Ethics approval from the Institutional Review Board was obtained (Project no. 55974) with a waiver of consent due to the observational design of this study. ⁴ Between the two centers, annual attendance for emergency medical services totals more than 200,000 episodes, with a combined catchment area of approximately 2 million people.

Consecutive patients identified by International Classification of Diseases Clinical Modification (ICD‐10‐CM) codes for cardiac arrest (I46, I46.2, I46.8, and I46.9) and shockable rhythms of ventricular fibrillation (VF) (I49.0 and I49.01) and ventricular tachycardia (VT) (I49.02) between January 1, 2014 and December 31, 2018 were included. ⁴ Patients were stratified into two groups according to whether they received CA (CA and no‐CA groups) (Figure 1). Those with ≥1 mm STE on initial ECG (assessed on arrival in emergency department) were omitted. Patients were subsequently separated into survivor and nonsurvivor groups for comparison to evaluate independent predictors of in‐hospital mortality. OHCA with shockable rhythm was defined as a sudden pulseless event with lack of effective cardiac output where VF or VT was recorded as the initial rhythm. Those with nonshockable rhythms (pulseless electrical activity or asystole) and in‐hospital cardiac arrest (occurring after arrival in emergency department) were excluded. Patients who died before arrival at hospital or before a decision to activate the catheterization laboratory could be made were also excluded.

Figure 1.

Flow diagram of study patients. CA, coronary angiogram; ECG, electrocardiogram; PCI, percutaneous coronary intervention; PEA, pulseless electrical activity; pVT, pulseless ventricular tachycardia; VF, ventricular fibrillation. ^aPre‐CA electrocardiograms were missing for seven patients in the overall cohort with shockable rhythms.

Clinical data including demographic characteristics, past medical conditions, periarrest variables, clinical examination findings, investigation results (including initial biochemistry, admission ECG, and bedside echocardiography), angiographic results, PCI procedural data, and in‐hospital outcomes were collected from medical charts. ⁴ Data were cross‐verified by study authors (W.Z., S.N., R.B., and L.H.) when deemed unclear or ambiguous.

2.2. Study definitions

CA was considered “early” when performed immediately (≤2 h) following arrival at the PCI‐capable centre. ¹⁵ , ¹⁶ “Late” CA was defined as CA performed at any other time during the index or subsequent hospital admission. This study was not randomized and the indication for performing early CA was based on the clinical judgment of the treating cardiologist. Both study sites have catheterization laboratories that were available at all times. The decision to delay CA or to not perform CA was usually made after initial assessment of neurological status and temperature management. We defined sustained return of spontaneous circulation (ROSC) as the resumption of spontaneous perfusing rhythm resulting in palpated pulse, arterial waveform or measurable blood pressure for >20 min. Assisted circulation (e.g., extracorporeal membrane oxygenation) was not regarded as ROSC until “patient‐generated” circulation was detected. ⁴ Cardiogenic shock (CS) was defined as a continuous event of >30 min where either systolic blood pressure was <90 mmHg or vasoactive/inotropic medications were utilized to support blood pressure >90 mmHg, with clinical signs of elevated filling pressures (e.g., pulmonary congestion on chest radiograph or clinical examination) and signs of end‐organ hypoperfusion (e.g., cold and diaphoretic peripheries, altered conscious state, or urine output ≤30 ml/h). ¹⁷

Significant coronary artery stenosis was defined as stenosis >70% in any major coronary artery branch (left anterior descending artery, left circumflex artery or right coronary artery) or graft vessel. ¹⁸ Stenosis >50% in the left main coronary artery was also considered to be significant. ¹⁸ Significant CAD was defined as having ≥1 significant stenosis. Acute coronary occlusion was defined as a presumably new obstructive lesion with Thrombolysis in Myocardial Infarction (TIMI) 0 or 1 flow with the ability to pass a guidewire easily through the occlusion if PCI was attempted. ¹⁹ , ²⁰ Complete occlusion to a main coronary artery branch considered >3 months old was deemed to be a chronic total occlusion (CTO). ²¹ Coronary lesion complexity was defined as A, B1, B2, or C, according to the American College of Cardiology/American Heart Association classification guidelines. ²² A lesion was considered to be successfully treated if the residual stenosis is <10% following coronary stenting, or <50% after balloon angioplasty/atherectomy alone, in addition to restoration of coronary perfusion with TIMI 3 flow. ¹⁹ , ²⁰ , ²³ Major bleeding was specified according to the Bleeding Academic Research Consortium (BARC) classification: BARC 3 (overt hemorrhage with hemoglobin drop of >3 g/dl, cardiac tamponade, intracranial bleeding or transfusion requirement) or BARC 5 (probable or confirmed fatal hemorrhage). ²⁴ , ²⁵

The Glasgow Coma Scale (GCS) is a frequently recorded index by emergency physicians that objectively describes the extent of impaired consciousness and therefore, the global status of the central nervous system. ²⁶ Patients were stratified into severely impaired conscious state (GCS ≤ 8) and compared against those with mild/moderate impairment (GCS > 8). ²⁶ The Acute Physiology and Chronic Health Evaluation II (APACHE II) score is a commonly used intensive care unit (ICU) mortality prediction model. Patient scores were retrospectively computed using past medical conditions available from clinical charts, initial examination findings and blood investigations sampled on admission to emergency department. ²⁷ To calculate APACHE II scores, acute kidney injury was specified as an absolute increase in serum creatinine of ≥44.2 μmol/L or ≥25% from baseline levels. In‐hospital new renal impairment was defined similarly, where the aforementioned change in serum creatinine can occur up to 5 days after index cardiac arrest.

The primary outcome was in‐hospital mortality. Secondary in‐hospital outcomes included a composite endpoint of major adverse cardiac and cerebrovascular events (MACCE)—comprising all‐cause mortality, new or recurrent myocardial infarction or stent thrombosis, target vessel revascularization, or nonfatal stroke—and its individual components. Other short‐term outcomes (e.g., hypoxic brain injury and ischemic hepatitis) were only recorded as positive if they were detailed in medical notes or imaging reports.

2.3. Statistical analysis

Continuous variables are expressed as mean and standard deviation or median and interquartile range, and categorical variables are expressed as counts and percentages, as appropriate. Continuous variables were compared using Student's t tests or Mann–Whitney U test. Categorical variables were compared using Pearson's chi‐square test. All tests were two‐tailed and differences between comparison groups were considered statistically significant if the p < 0.05. Univariate comparisons of baseline, angiographic and procedural characteristics against in‐hospital mortality were undertaken for the overall cohort. Univariate variables with p < 0.10 were subsequently included in multivariable logistic regression analysis with stepwise backward selection to identify independent predictors of in‐hospital mortality. Missing values in multivariable analyses were replaced with the series mean and median for normally and non‐normally distributed continuous variables, respectively. Variables with >20% missing values were omitted from multivariable analyses. All statistical analyses were performed using IBM SPSS Statistics 26.0 (IBM Corp).

3. RESULTS

3.1. Patient baseline characteristics

From 761 patients presenting with OHCA to Alfred and Western Health between 2014 and 2018, 118 patients had an initial shockable rhythm without STE on postarrest ECG (Figure 1). Baseline characteristics and investigation results of 86 patients (73%) selected for CA and 32 patients (27%) not selected for CA are presented in Table 1. Compared to no‐CA group, fewer patients receiving CA had atrial fibrillation, chronic kidney disease (CKD) and previous coronary artery bypass graft (CABG) surgery. Those undergoing CA were more likely to be between 50 and 60 years of age, have shorter cardiopulmonary resuscitation (CPR) time and time to ROSC, initial GCS >8, and higher albumin. Clinical characteristics were generally similar between the early CA group and the late or no CA group. Those undergoing early CA were more likely to achieve prehospital ROSC.

Table 1.

Clinical characteristics of patients with OHCA, no ST‐elevation and shockable rhythm.

Variable	CA (n = 86)	No CA (n = 32)	p Value	Early CA (n = 38)	Late/no CA (n = 80)	p Value
Age, years, median [IQR]	60 [53‒71]	67 [42‒74]	0.64	57 [49‒71]	54 [66‒74]	0.13
Age ≥65 years	39 (45)	17 (53)	0.45	15 (40)	41 (51)	0.23
Male, n (%)	68 (79)	21 (66)	0.13	30 (79)	59 (74)	0.54
BMI, kg/m2, median [IQR]	26 [24‒30]	25 [23‒30]	0.15	26 [24‒28]	27 [24‒30]	0.56
History of smoking, n (%)	33 (38)	10 (31)	0.48	16 (42)	27 (34)	0.38
Pre‐existing medical conditions
Diabetes, n (%)	21 (24)	10 (31)	0.45	9 (24)	22 (28)	0.66
Coronary artery disease, n (%)	30 (35)	10 (31)	0.71	11 (29)	29 (36)	0.43
Previous MI, n (%)	19 (22)	6 (19)	0.69	8 (21)	17 (21)	0.98
Previous PCI, n (%)	13 (15)	3 (9.4)	0.42	6 (16)	10 (13)	0.63
Previous CABG, n (%)	11 (13)	9 (28)	0.05	6 (16)	14 (18)	0.82
Peripheral vascular disease, n (%)	3 (3.5)	2 (6.3)	0.51	1 (2.6)	4 (5.0)	0.55
Previous cerebrovascular disease, n (%)	4 (4.7)	3 (9.4)	0.33	2 (5.3)	5 (6.3)	0.83
Hypertension, n (%)	46 (54)	21 (66)	0.24	19 (50)	48 (60)	0.31
Dyslipidaemia, n (%)	43 (50)	15 (47)	0.76	15 (40)	43 (54)	0.15
Atrial fibrillation or flutter, n (%)	15 (17)	11 (34)	0.05	6 (16)	20 (25)	0.26
VT without hemodynamic compromise, n (%)	0 (0.0)	2 (6.3)	0.02	0 (0.0)	2 (2.5)	0.33
Existing PPM or ICD, n (%)	8 (9.3)	1 (3.1)	0.26	2 (5.3)	7 (8.8)	0.51
Heart failure, n (%)	10 (12)	8 (25)	0.07	5 (13)	13 (16)	0.66
Chronic kidney disease, n (%)	5 (5.8)	6 (19)	0.03	2 (5.3)	9 (11)	0.30
Chronic obstructive pulmonary disease, n (%)	13 (15)	3 (9.4)	0.42	7 (18)	9 (11)	0.29
Medications and treatments before CA
Anticoagulation, n (%)	12 (14)	10 (32)	0.03	4 (11)	18 (23)	0.11
Antiplatelet therapy, n (%)	29 (34)	14 (45)	0.26	12 (32)	31 (39)	0.42
Dialysis, n (%)	0 (0.0)	2 (6.5)	0.02	0 (0.0)	2 (2.5)	0.32
Antihypertensive, n (%)	41 (48)	19 (61)	0.19	16 (42)	44 (56)	0.17
Before hospital admission
Arrest on weekend (between 00:00 Saturday and 00:00 Monday), n (%)	25 (29)	11 (34)	0.58	12 (32)	24 (30)	0.86
Arrest at night (between 22:00 and 06:00), n (%)	14 (17)	8 (25)	0.29	4 (11)	18 (23)	0.11
Bystander CPR, n (%)	63 (73)	22 (69)	0.63	26 (68)	59 (74)	0.55
Total CPR duration, min, median [IQR]	16 [9.5‒21]	23 [14‒34]	0.02	15 [9.5‒21]	19 [10‒29]	0.25
Long low‐flow period (>30 min before ROSC), n (%)	21 (24)	16 (52)	0.01	10 (26)	27 (34)	0.39
Bystander AED use, n (%)	17 (20)	4 (13)	0.36	6 (16)	15 (19)	0.69
Time to defibrillation, min, median [IQR]	10 [5.0‒12]	11 [6.5‒14]	0.32	10 [6.0‒12]	10 [5.0‒13]	0.56
Arrest witnessed by emergency services, n (%)	11 (13)	2 (6.3)	0.31	5 (13)	8 (10)	0.63
BLS to ROSC duration, min, median [IQR]	17 [10‒22]	25 [15‒36]	0.01	17 [11‒22]	19 [10‒30]	0.40
Collapse to ROSC, min, median [IQR]	17 [10‒28]	28 [15‒37]	0.04	17 [13‒28]	19 [12‒30]	0.70
Prehospital ROSC, n (%)	80 (93)	27 (87)	0.31	38 (100)	69 (87)	0.02
Public location of arrest, n (%)	51 (59)	21 (66)	0.53	25 (66)	47 (59)	0.46
Transfer from outside hospital, n (%)	14 (16)	4 (13)	0.61	3 (7.9)	15 (19)	0.13
Automated mechanical chest compression, n (%)	9 (11)	5 (16)	0.47	2 (5.4)	12 (15)	0.13
Mechanical ventilation, n (%)	69 (81)	29 (91)	0.22	34 (90)	64 (81)	0.25
During hospital admission
Cardiogenic shock at presentation, n (%)	49 (57)	23 (72)	0.14	22 (58)	50 (63)	0.63
Automated mechanical chest compression, n (%)	6 (7.0)	3 (9.4)	0.66	2 (5.3)	7 (8.8)	0.51
Vasopressor or inotrope support, n (%)	66 (77)	27 (84)	0.37	31 (82)	62 (78)	0.61
Mechanical ventilation, n (%)	64 (76)	28 (88)	0.18	31 (84)	61 (77)	0.42
Intravenous antiarrhythmic medication, n (%)	62 (72)	17 (53)	0.05	29 (76)	50 (63)	0.14
Therapeutic hypothermia, n (%)	43 (52)	8 (26)	0.01	22 (61)	29 (38)	0.02
Intensive care unit admission, n (%)	65 (76)	28 (88)	0.16	32 (84)	61 (76)	0.32
Examination (first assessment in ED)
Mean arterial pressure, mmHg, median [IQR]	90 [78‒110]	92 [80‒112]	0.89	97 [82‒117]	90 [75‒106]	0.09
Systolic blood pressure, mmHg, median [IQR]	130 [110‒150]	133 [100‒150]	0.82	139 [116‒160]	129 [100‒150]	0.10
Diastolic blood pressure, mmHg, median [IQR]	72 [60‒90]	70 [60‒91]	0.66	80 [62‒98]	81 [60‒86]	0.15
Heart rate, per min, median [IQR]	90 [78‒110]	100 [75‒116]	0.57	91 [78‒110]	98 [76‒111]	0.47
GCS ≤ 8, n (%)	65 (76)	29 (94)	0.03	33 (87)	61 (77)	0.22
Investigations
Initial blood gas (on arrival)
FiO2, %, median [IQR]	100 [50‒100]	100 [50‒100]	0.71	100 [50‒100]	90 [50‒100]	0.79
pO2, mmHg, median [IQR]	140 [97‒282]	124 [89‒301]	0.60	141 [90‒259]	138 [95‒302]	0.96
pCO2, mmHg, median [IQR]	49 [41‒59]	47 [42‒62]	0.85	49 [41‒57]	47 [42‒62]	0.93
Arterial pH, median [IQR]	7.24 [7.14‒7.32]	7.22 [7.06‒7.32]	0.38	7.24 [7.14‒7.31]	7.22 [7.10‒7.33]	0.91
Standard HCO3 −, mmol/L, median [IQR]	20 [17‒22]	18 [12‒22]	0.10	21 [17‒22]	19 [16‒22]	0.38
Base excess, mmol/L, median [IQR]	−7.0 [−11 to −4.0]	‒8.7 [−15.9 to −4.2]	0.25	−7.0 [−11 to −4.5]	−8.0 [−12 to −4.0]	0.90
O2 saturation, %, median [IQR]	98 [95‒99]	98 [93‒99]	0.62	98 [95‒99]	98 [94‒99]	0.91
Glucose, mmol/L, median [IQR]	12 [8.8‒16]	13 [9.8‒17]	0.76	12 [9.0‒15]	13 [9.0‒16]	0.91
Lactate, mmol/L, median [IQR]	3.1 [1.8‒5.5]	3.7 [2.4‒9.3]	0.10	3.7 [1.7‒5.9]	3.0 [2.0‒6.5]	0.98
Hemoglobin, g/L, median [IQR]	141 [130‒153]	134 [115‒149]	0.09	139 [134‒155]	139 [123‒150]	0.14
Haematocrit, %, median [IQR]	0.42 [0.40‒0.46]	0.42 [0.35‒0.46]	0.46	0.42 [0.40‒0.46]	0.42 [0.37‒0.46]	0.57
White cell count, ×103/mm3, median [IQR]	16 [10‒21]	15 [12‒20]	0.92	17 [10‒22]	15 [10‒21]	0.43
Serum urea, mmol/L, median [IQR]	6.9 [6.0‒8.6]	7.3 [5.4‒11]	0.36	6.9 [5.8‒8.3]	7.1 [5.8‒9.8]	0.25
eGFR, ml/min/1.73 m2, median [IQR]	64 [52‒78]	69 [46‒90]	0.65	61 [52‒77]	68 [51‒88]	0.46
Baseline creatinine, μmol/L, median [IQR]	104 [81‒127]	100 [79‒140]	0.50	105 [84‒132]	102 [80‒125]	0.32
Bilirubin, µmol/L, median [IQR]	9.0 [7.0‒12]	8.0 [4.5‒15]	0.20	8.0 [6.0‒13]	9.0 [6.0‒12]	0.79
Albumin, g/L, median [IQR]	33 [30‒36]	31 [27‒35]	0.02	34 [30‒36]	32 [28‒35]	0.23
Atrial fibrillation on initial ECG, n (%)	14 (16)	11 (34)	0.03	3 (7.9)	22 (28)	0.02
New left bundle branch block, n (%)	15 (18)	3 (9.7)	0.28	5 (14)	13 (17)	0.66
New right bundle branch block, n (%)	10 (12)	5 (16)	0.55	4 (11)	11 (14)	0.62
Moderate/severe LV systolic dysfunction on echocardiogram (LVEF < 45%), n (%)	50 (64)	11 (79)	0.29	20 (61)	41 (70)	0.39
Prognostic scores
APACHE II score, median [IQR]	24 [18‒28]	26 [22‒28]	0.32	26 [19‒28]	24 [18‒28]	0.76

Note: Values are median [interquartile range] or count (%).

Abbreviations: AED, automated external defibrillator; APACHE II, Acute Physiology And Chronic Health Evaluation II; BLS, basic life support; BMI, body mass index; BSA, body surface area; CA, coronary angiography; CABG, coronary artery bypass graft; CPR, cardiopulmonary resuscitation; ECG, electrocardiogram; ED, emergency department; eGFR, estimated glomerular filtration rate; FiO₂, fraction of inspired oxygen; GCS, Glasgow Coma Score; ICD, implantable cardioverter‐defibrillator; LVEF, left ventricular ejection fraction; MI, myocardial infarction; NFR, not for resuscitation; PCI, percutaneous coronary intervention; pCO₂, partial pressure of carbon dioxide; pO₂, partial pressure of oxygen; PPM, permanent pacemaker; ROSC, return of spontaneous circulation; VT, ventricular tachycardia.

3.2. Procedural characteristics

Table 2 summarizes procedural and coronary angiographic data for those receiving CA. Of 86 patients in CA group, 38 (44%) underwent early CA and 48 (56%) underwent late CA, with median time to catheterization of 86 and 861 min, respectively. Significant CAD was identified in 50 patients (58%), of whom 44% had single‐vessel disease and 56% had multivessel disease. An acute culprit coronary lesion was found in 30 patients (35%). Thirty patients underwent PCI, of whom 21 (70%) were treated for an acute culprit lesion and 9 (30%) received PCI to nonculprit obstructive lesions. The lesion of interest was successfully treated for 90% of patients undergoing PCI. The distribution of significant CAD, acute coronary occlusion and CTO were similar among those undergoing early CA versus late CA.

Table 2.

Angiographic and procedural characteristics in the CA group.

Variable	Early CA (n = 38)	Late CA (n = 48)	p Value	PCI (n = 30)	No‐PCI (n = 56)	p Value
Time to catheterization from presentation, min, median [IQR]	86 [65‒99]	861 [184‒4238]	<0.01	180 [91‒1042]	128 [88‒1198]	0.78
LV function on catheterization
Normal or mild dysfunction, n (%)	14 (64)	13 (48)	0.28	9 (64)	18 (51)	0.41
Moderate or severe dysfunction, n (%)	8 (36)	14 (52)		5 (36)	17 (49)
LV end diastolic pressure, mmHg, median [IQR]	19 [12‒24]	20 [15‒28]	0.34	20 [12‒28]	20 [15‒24]	0.94
Femoral percutaneous entry, n (%)	18 (47)	24 (50)	0.81	14 (47)	28 (50)	0.77
Radial/brachial percutaneous entry, n (%)	20 (53)	23 (48)	0.66	16 (53)	27 (48)	0.65
Procedural Intubation required, n (%)	7 (18)	5 (11)	0.31	3 (10)	9 (16)	0.42
Mechanical circulatory support, n (%)	3 (7.9)	5 (10)	0.69	2 (6.7)	6 (11)	0.54
IABP, n (%)	2 (5.3)	2 (4.2)	0.81	1 (3.3)	3 (5.4)	0.67
ECMO, n (%)	2 (5.3)	3 (6.3)	0.85	2 (6.7)	3 (5.4)	0.81
GP IIb/IIIa inhibitors, n (%)	4 (11)	7 (15)	0.61	10 (33)	1 (1.8)	<0.01
Heparin, n (%)	25 (66)	33 (69)	0.77	28 (93)	30 (54)	<0.01
Significant coronary artery disease, n (%)	20 (56)	30 (64)	0.45	28 (93)	22 (42)	<0.01
Left main coronary artery, n (%)	6 (17)	2 (4.3)	0.06	2 (6.7)	6 (11)	0.49
Left anterior descending artery, n (%)	16 (44)	19 (40)	0.71	19 (63)	16 (30)	<0.01
Left circumflex artery, n (%)	11 (31)	16 (34)	0.74	12 (40)	15 (28)	0.27
Right coronary artery, n (%)	13 (36)	18 (38)	0.84	17 (57)	14 (26)	<0.01
Graft vessel, n (%)	2 (5.6)	2 (4.3)	0.78	2 (6.7)	2 (3.8)	0.55
Single vessel disease, n (%)	8 (21)	14 (29)	0.39	14 (47)	8 (14)	<0.01
Multivessel disease, n (%)	12 (32)	16 (33)	0.86	14 (47)	14 (25)	0.04
Acute coronary occlusion, n (%)	14 (39)	16 (34)	0.65	21 (70)	9 (17)	<0.01
Left main coronary artery, n (%)	1 (2.8)	0 (0.0)	0.25	1 (3.3)	0 (0.0)	0.18
Left anterior descending artery, n (%)	5 (14)	7 (15)	0.90	9 (30)	3 (5.7)	<0.01
Left circumflex artery, n (%)	3 (8.3)	5 (11)	0.72	5 (17)	3 (5.7)	0.10
Right coronary artery, n (%)	3 (8.3)	5 (11)	0.72	7 (23)	1 (1.9)	<0.01
Graft vessel, n (%)	2 (5.6)	1 (2.2)	0.42	1 (3.4)	2 (3.8)	0.94
Chronic total occlusion, n (%)	10 (28)	15 (32)	0.68	10 (33)	15 (28)	0.63
Left main coronary artery, n (%)	1 (2.8)	0 (0.0)	0.25	0 (0.0)	1 (1.9)	0.45
Left anterior descending artery, n (%)	6 (17)	6 (13)	0.62	5 (17)	7 (13)	0.67
Left circumflex artery, n (%)	5 (14)	5 (11)	0.65	3 (10)	7 (13)	0.67
Right coronary artery, n (%)	8 (22)	11 (23)	0.90	8 (27)	11 (21)	0.54
Coronary lesion type B2/C, n (%)	14 (40)	14 (30)	0.37	9 (32)	19 (36)	0.74
PCI to lesion(s), n (%)	12 (32)	18 (38)	0.57	‒	‒	‒
Left main coronary artery, n (%)	0 (0.0)	1 (2.9)	0.35	1 (3.3)	‒	‒
Left anterior descending artery, n (%)	7 (24)	10 (29)	0.64	17 (57)	‒	‒
Left circumflex artery, n (%)	4 (14)	4 (12)	0.81	8 (27)	‒	‒
Right coronary artery, n (%)	3 (10)	6 (18)	0.41	9 (30)	‒	‒
Graft vessel, n (%)	0 (0.0)	2 (5.9)	0.18	2 (6.7)	‒	‒
Lesion(s) successfully treated, n (%)	12 (100)	15 (83)	0.14	27 (90)	‒	‒
In‐stent restenosis, n (%)	2 (5.6)	1 (2.1)	0.41	2 (6.9)	1 (1.9)	0.24
Stent thrombosis, n (%)	1 (2.6)	0 (0.0)	0.26	1 (3.3)	0 (0.0)	0.17

Note: Values are median [interquartile range] or count (%).

Abbreviations: ECMO, extracorporeal membrane oxygenation; IABP, intra‐aortic balloon pump; LV, left ventricular; PCI, percutaneous coronary intervention.

3.3. Clinical outcomes

Compared to the no‐CA group, patients in the CA group had lower rates of in‐hospital mortality (27% vs. 50%) and MACCE (29% vs. 50%) (all p < 0.05) (Table 3). In‐hospital mortality and MACCE was 33% and 35% for the overall cohort, respectively. Patients treated with PCI had lower in‐hospital mortality compared to those not treated with PCI (7.7% vs. 92%, p < 0.01). Compared to late or no CA, early CA was not associated with decreased in‐hospital mortality (32% vs. 34%, p = 0.82). Recurrent in‐hospital CS was 32% for overall cohort, 27% in CA group and 47% in no‐CA group (p = 0.04). Hypoxic brain injury was observed in 37% of the overall cohort, with no difference between CA and no‐CA groups. Patients in the CA group were more likely to be discharged home compared to the no‐CA group (51% vs. 16%, p = 0.01). Other in‐hospital outcomes are summarized in Table 3.

Table 3.

In‐hospital outcomes.

Variable	CA (n = 86)	No CA (n = 32)	p Value	Early CA (n = 38)	Late/no CA (n = 80)	p Value
Peak CK, units/L, median [IQR]	906 [260‒2190]	629 [247‒1870]	0.58	908 [176‒2751]	792 [252‒1744]	0.67
Length of overall hospital stay, h, median [IQR]	208 [103‒363]	103 [31‒215]	<0.01	192 [96‒299]	168 [83‒323]	0.76
Length of stay in ICU, days, median [IQR]	4.0 [3.0‒8.0]	3.0 [1.0‒7.5]	0.10	4.0 [2.0‒5.3]	4.0 [2.0‒9.0]	0.51
New renal impairment, n (%)	25 (30)	8 (25)	0.61	8 (22)	26 (32)	0.27
New requirement for dialysis, n (%)	11 (13)	3 (9.4)	0.57	5 (14)	9 (11)	0.70
Ischemic hepatitis, n (%)	8 (9.4)	5 (16)	0.34	5 (14)	8 (10)	0.57
Cardiogenic shock, n (%)	23 (27)	15 (47)	0.04	11 (30)	27 (34)	0.67
Recurrent myocardial infarction, n (%)	2 (2.4)	0 (0.0)	0.38	2 (5.4)	0 (0.0)	0.04
In‐hospital PCI (subsequent PCI and distinct from index PCI), n (%)	5 (6.0)	0 (0.0)	0.16	2 (5.6)	3 (3.8)	0.66
CABG, n (%)	7 (8.1)	0 (0.0)	0.10	4 (11)	3 (3.8)	0.15
Clinically significant bleeding (Type 3 and Type 5), n (%)	6 (7.0)	2 (6.3)	0.89	1 (2.6)	7 (8.8)	0.22
Hypoxic brain injury, n (%)	28 (33)	16 (50)	0.08	12 (32)	32 (40)	0.38
Ischemic stroke, n (%)	2 (2.4)	1 (3.1)	0.81	0 (0.0)	3 (3.8)	0.23
In‐hospital mortality, n (%)	23 (27)	16 (50)	0.02	12 (32)	27 (34)	0.82
In‐hospital MACCE, n (%)	25 (29)	16 (50)	0.03	13 (34)	28 (35)	0.93
Discharge status
Home, n (%)	44 (51)	5 (16)	0.01	19 (50)	30 (38)	0.51
Rehabilitation unit or hospital, n (%)	10 (12)	7 (22)		5 (13)	12 (15)
Local or referring hospital, n (%)	7 (8.1)	3 (9.4)		1 (2.6)	9 (11)
Tertiary referral center, n (%)	2 (2.3)	1 (3.1)		1 (2.6)	2 (2.5)

Note: Values are median [interquartile range] or count (%).

Abbreviations: CA, coronary angiography; CABG, coronary artery bypass graft; CK, creatine kinase; ICU, intensive care unit; MACCE, major adverse cardiac and cerebrovascular events; PCI, percutaneous coronary intervention.

3.4. Predictors of in‐hospital mortality

Following multivariable adjustment, predictors of in‐hospital mortality included longer defibrillation time (odds ratio [OR] 3.07, 95% confidence interval [CI] 1.44‒6.53 per 5‐min increase), lower pH (OR 2.02, 95% CI 1.33‒3.09 per 0.1 decrease), lower albumin (OR 2.02, 95% CI 1.03‒3.95 per 5 g/L decrease), and worse baseline renal function (OR 1.33, 95% CI 1.02‒1.72 per 10 ml/min/1.73 m² decrease), while predictors of survival to hospital discharge included PCI to lesion (OR 0.11, 95% CI 0.01‒0.79) and bystander defibrillation (OR 0.06, 95% CI 0.004‒0.80) (all p < 0.05) (Figure 2). Neither selection for CA (OR 0.40, 95% CI 0.09‒1.75, p = 0.22) nor early CA (OR 1.47, 95% CI 0.34‒6.31, p = 0.60) predicted survival until hospital discharge.

Figure 2.

Independent predictors of in‐hospital mortality. Variables entered into multivariable logistic regression included: age ≥65 years, bystander AED use, time to defibrillation, arrest at home, prehospital automated mechanical chest compression, cardiogenic shock on admission requiring inotropic support, pCO₂, arterial pH, O₂ saturation, glucose, lactate, eGFR, albumin, selection for cardiac catheterization, early CA (≤2 h from arrival) and PCI to lesion.

4. DISCUSSION

In this observational cohort study of 118 patients presenting with OHCA, shockable rhythm and no STE, we evaluated clinical characteristics favouring selection for CA and defined adverse prognostic variables associated with in‐hospital mortality. Patients with downtime >30 min, longer CPR duration, lower albumin, previous CABG, past history of CKD or atrial fibrillation, GCS ≤8 and advanced age >60 years were more likely to be excluded from cardiac catheterization. Key adverse prognostic predictors of in‐hospital mortality comprised longer time to defibrillation, lower pH, poorer baseline renal function, lower albumin, the absence of a treatable coronary lesion and the lack of bystander defibrillation. No differences in survival until hospital discharge or other in‐hospital clinical outcomes were observed between those undergoing an early CA strategy versus those undergoing late CA or nonselection for CA. In the absence of high‐level evidence, these real‐world data highlight the interplay of physician decision‐making, family preference, comorbidity burden and perceived futility in the current clinical practice of patient selection for cardiac catheterization. Identifying adverse predictors for in‐hospital mortality allows clinicians to evaluate the balance of positive and negative prognostic variables and provide guidance to the unclear and oftentimes subjective process of prognostication in this cohort.

The 2015 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science suggests considering emergency CA over delayed CA or no CA in select patients with ROSC after OHCA of suspected cardiac origin without STE. ⁶ However, selecting patients who are often comatose on arrival, with an unclear past medical history and unreliable account of preceding symptoms and arrest time, remains challenging with a paucity of evidence existing in this population. Similar to our study findings, a Swedish nationwide observational study of 799 patients with OHCA and no STE found that those receiving early CA were younger, more likely to have ECG patterns of ST‐depression and previous known left bundle branch block, and less likely to have a history of heart failure, ischemic heart disease, previous myocardial infarction, atrial fibrillation, and diabetes. ⁸ , ²⁸ In addition, in our cohort, previous CABG appeared to deter selection for CA and such patients were more likely to have a type B2 or C coronary lesion or CTO. These data suggest that patients with a complex coronary history were considered to have lesions less amenable to PCI and were therefore more likely to be excluded from emergent CA. Another Danish review of 244 patients with OHCA and no‐STE reported that those undergoing emergent CA more frequently had VF leading to arrest, and lower lactate levels, whereas a history of chronic obstructive pulmonary disease was less prevalent. ⁷ Other observational studies have shown that male sex and CS on admission to be associated with urgent selection for CA. ⁹ , ¹¹ Overall, these findings suggest that patients were more frequently selected for CA based on factors associated with adequate initial resuscitation, the absence of significant comorbidities and complex coronary disease history, and perceived favorable postarrest neurological status.

In the presence of this substantial selection bias, it is most pertinent to identify prognostic factors and the extent to which they drive a potential survival benefit. We analysed a select population of adult OHCA survivors with an initial shockable rhythm but no STE where the prevalence of CAD was still expected to be considerable, although lower than those with OHCA and STE. ³ , ⁴ , ¹⁰ In the current cohort with OHCA, shockable rhythm and no STE, 73% of patients received CA, of whom 35% underwent PCI. Following multivariable adjustment, selection for CA was not an independent predictor of survival to hospital discharge. Contrary to previous observational reports, our study did not find early CA to be associated with a survival benefit over late CA or no CA. ⁸ , ⁹ , ¹⁰ , ²⁹ Several considerations may explain the lack of benefit of an early CA strategy in our cohort. First, the strategy of CA improving survival in patients with OHCA without STE is likely contingent on the presence of acute coronary lesion(s) amenable to PCI in the study group. In the absence of STE on postresuscitation ECG, previous reports have demonstrated the prevalence of significant CAD and acute coronary lesion(s) to be 25%‒50% and 25%‒35%, respectively. ³ , ¹⁰ , ³⁰ In keeping with these reports, 35% of patients in our cohort had at least one culprit coronary lesion treated with PCI and thus, the benefit from CA can only be expected for this subgroup if revascularization was successfully achieved. A strategy of unselected CA for all other patients with OHCA, and no STE may delay the diagnosis of alternative aetiologies for cardiac arrest (e.g., pulmonary embolism, aortic dissection, alcohol/drug poisoning, traumatic brain/medullar injury) and their timely treatment, and the added unnecessary risk of procedural complications related to CA. Furthermore, 37% of patients in our study successfully resuscitated following OHCA suffered from hypoxic brain injury, consistent with rates reported by other contemporary reports. ⁹ , ¹³ , ³⁰ It is recognized that poor neurological recovery rather than ongoing cardiac injury has a greater effect on short‐term mortality, thus counterbalancing the potential survival benefit for those with successful early revascularization. ³¹ Four RCTs evaluating the benefit of an early CA strategy in those presenting with OHCA and no STE did not demonstrate superiority over delayed or selective CA with respect to short‐ and midterm all‐cause mortality. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Notably, patients with CS were excluded from TOMAHAWK and COACT but included in EMERGE and PEARL RCTs. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Taken together, the evaluation of patients presenting with OHCA, shockable rhythm and no STE on post‐resuscitation ECG warrants a more cautious approach with consideration of noncoronary causes of cardiac arrest and delaying CA to permit adequate assessment of patients' comorbidities, CAD history, the success of their initial and postarrest resuscitation efforts, as well as their postarrest neurological status.

Timely prognostication of in‐hospital mortality is vital in this group of gravely ill patients. As such, we identified several simple and readily available clinical and biochemical variables to help prognosticate those at risk of an unfavorable outcome after their OHCA presentation. Corroborating previous observational reports, ineffective resuscitation attempt (lack of bystander defibrillation and longer time to defibrillation, leading to worse acidemia) and a greater burden of comorbidities (poor baseline renal function and chronic hypoalbuminemia) were major drivers of in‐hospital mortality in our study. ² , ⁴ , ⁷ Hypoalbuminemia has been consistently validated as an independent predictor of poor neurological outcome and higher short‐ and long‐term mortality in patients with OHCA and patients admitted to cardiac ICU. ³² , ³³ For those with postcardiac arrest syndrome, the systemic inflammatory response to ischemia and reperfusion remains unchecked in patients with hypoalbuminemia. Reduced oncotic pressure further leads to insufficient blood flow to vital organs. ³² Low arterial pH and high lactate also appear to be surrogate markers for prolonged downtime and inadequate resuscitation attempt, with several single institution studies reporting these parameters to predict poorer survival and functional outcomes. ³⁴ , ³⁵ Our data reiterate the critical elements in the “chain of survival,” from early recognition of cardiac arrest and contact with emergency services to prompt initiation of adequate quality CPR and timely defibrillation. Multiple large nationwide registry studies have convincingly demonstrated that those with arrest in public locations and those receiving early defibrillation have greater 30‐day survival and more favorable neurological outcomes. ³⁶ , ³⁷ , ³⁸ The importance of promoting high‐quality CPR education and awareness in the community, and advocacy for public‐access defibrillation programs and automated external defibrillator availability cannot be understated.

5. LIMITATIONS

Several limitations should be considered when interpreting these results. First, while multivariable analysis was performed to adjust for baseline characteristics, unmeasured confounders (such as family preference, physician risk aversion and overall perceived futility, among others) not captured by chart review that affect the decision to perform CA and ultimately, clinical outcomes were likely present. Secondly, we did not capture the severity of some pre‐existing medical conditions (e.g., hypertension, diabetes, dyslipidaemia), which were generally treated as binary variables. This may lead to over‐ or underestimation of overall disease burden and skew the association between disease severity and selection for CA or in‐hospital mortality. Thirdly, for variables with a greater amount of missing data, missing values imputation has a larger centralizing effect. These variables are less likely to appear significant on multivariable analysis and their odds ratios may be underestimated. Finally, the study sample size was relatively small, which could impact on the power to detect a significant effect of performing CA on clinical outcomes.

6. CONCLUSION

Among patients with OHCA, shockable rhythm and no STE, younger age, shorter time to ROSC and GCS >8 were associated with selection for CA while less effective resuscitation, greater burden of comorbidities and the absence of a treatable coronary lesion were key adverse prognostic predictors. A strategy of early CA was not better than a strategy of delaying CA with respect to survival until hospital discharge. Future work should focus on optimizing early resuscitation efforts and postarrest care to improve survival in this cohort of critically ill patients.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

Zheng WC, Noaman S, Batchelor RJ, et al. Evaluation of factors associated with selection for coronary angiography and in‐hospital mortality among patients presenting with out‐of‐hospital cardiac arrest without ST‐segment elevation. Catheter Cardiovasc Interv. 2022;100:1159‐1170. 10.1002/ccd.30442

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.