Predictors of coronary artery disease in cardiac arrest survivors: coronary angiography for everyone? A single-center retrospective analysis

Objective To identify predictors of coronary artery disease in survivors of cardiac arrest, to define the best timing for coronary angiography and to establish the relationship between coronary artery disease and mortality. Methods This was a single-center retrospective study including consecutive patients who underwent coronary angiography after cardiac arrest. Results A total of 117 patients (63 ± 13 years, 77% men) were included. Most cardiac arrest incidents occurred with shockable rhythms (70.1%), and the median duration until the return of spontaneous circulation was 10 minutes. Significant coronary artery disease was found in 68.4% of patients, of whom 75% underwent percutaneous coronary intervention. ST-segment elevation (OR 6.5, 95%CI 2.2 - 19.6; p = 0.001), the presence of wall motion abnormalities (OR 22.0, 95%CI 5.7 - 84.6; p < 0.001), an left ventricular ejection fraction ≤ 40% (OR 6.2, 95%CI 1.8 - 21.8; p = 0.005) and elevated high sensitivity troponin T (OR 3.04, 95%CI 1.3 - 6.9; p = 0.008) were predictors of coronary artery disease; the latter had poor accuracy (area under the curve 0.64; p = 0.004), with an optimal cutoff of 170ng/L. Only ST-segment elevation and the presence of wall motion abnormalities were independent predictors of coronary artery disease. The duration of cardiac arrest (OR 1.015, 95%CI 1.0 - 1.05; p = 0.048) was an independent predictor of death, and shockable rhythm (OR 0.4, 95%CI 0.4 - 0.9; p = 0.031) was an independent predictor of survival. The presence of coronary artery disease and the performance of percutaneous coronary intervention had no impact on survival; it was not possible to establish the best cutoff for coronary angiography timing. Conclusion In patients with cardiac arrest, ST-segment elevation, wall motion abnormalities, left ventricular dysfunction and elevated high sensitivity troponin T were predictive of coronary artery disease. Neither coronary artery disease nor percutaneous coronary intervention significantly impacted survival.


INTRODUCTION
Sudden cardiac arrest (CA) is one of the leading causes of death in Europe and the United States. (1,2) Although successful resuscitation is currently achieved in 40% -60% of patients (3) in whom advanced cardiac life support is attempted, the long-term survival rates following out-of-hospital CA remain dismal. (4) The etiology of CA is diverse and includes cardiac and noncardiac causes. Evidence suggests that coronary artery disease (CAD) is a main cause of CA, with significant CAD documented by coronary angiography or autopsy in more than 70% of patients with resuscitated CA, (3,5,6) but even in those patients, it is often difficult to establish whether the patient had an acute coronary event or simply bystander chronic CAD.
If acute coronary occlusion is the cause of CA, percutaneous coronary intervention (PCI) might reduce infarct size, improve hemodynamic status and reduce the recurrence of life-threatening arrhythmias and recurrent CA. (3,7) Considering this evidence, international guidelines recommend coronary angiography and PCI when indicated in patients following CA with ST-segment elevation (class of recommendation I, Level of Evidence B) and when there is a high index of suspicion of ongoing infarction (class of recommendation IIa, Level of Evidence C in European guidelines and B in American guidelines). (8,9) However, since patients after CA have been excluded from the main randomized trials that demonstrated the benefits of primary PCI in acute coronary syndrome (ACS) with or without ST-segment elevation, the real benefit of systematic coronary angiography in these patients is still under debate. (3) Although some studies support immediate coronary angiography and possible PCI in the setting of out-ofhospital CA, as it appears to be associated with better survival at discharge, (10,11) this remains controversial, and some authors highlight that routine emergency coronary angiography is probably not cost-effective or entirely riskfree. (12,13) Indeed, the recently published COACT (7) trial found that immediate angiography was not better than delayed angiography with respect to overall survival in these patients.
The main objective of this study was to identify independent predictors of CAD in CA survivors, with the ultimate goal of helping to determine which patients should undergo coronary angiography. The secondary objectives were to establish the best timing for coronary angiography in these patients and to determine the relationship of the presence of CAD, PCI, and mortality.

METHODS
A retrospective single-center observational study was performed in a tertiary university hospital (Department of Cardiology, Hospital Universitário de Santa Maria, Lisboa, Portugal).
We included all consecutive adult patients who underwent coronary angiography in the setting of CA after the return of spontaneous circulation (ROSC) from January 2015 to July 2018.
Demographic, clinical, laboratory, electrocardiographic, echocardiographic and angiographic data were reviewed based on clinical records, which were fully electronic.
Comorbidities were acknowledged by being referenced in the patient clinical file or by laboratory results available prior to the index hospital admission (for diabetes, chronic kidney disease and dyslipidemia).
The assumed cause for CA was adjudicated by reviewing the electronic clinical records of the patient.
All electrocardiograms (ECGs) available in the electronic clinical records were reviewed by two authors to identify criteria of ST-segment elevation or equivalents according to the current recommendations. (8) Other signs of ischemia (i.e., de novo ST-segment depression and T-wave inversion) were also recorded.
Left ventricular (LV) systolic dysfunction, based on the LV ejection fraction (LVEF), was classified according to the current European recommendations. (14) Significant CAD was defined by a lesion resulting in a lumen reduction of at least 50% in the left main artery and of at least 70% in the remaining vessels. The first available laboratory results after ROSC were considered.

Statistical analysis
The statistical analysis was performed using IBM Statistical Package for Social Disease (SPSS, Chicago, IL, United States), version 23. Categorical variables are reported in the absolute number and percentage, and continuous variables are reported as the mean and standard deviation or median and interquartile range (IQR), according to the normality of the distribution, which was tested with the Shapiro-Wilk test. Factors associated with CAD were identified by the chi-square test or Fisher's exact test for categorical variables and Student's t test or a nonparametric test for continuous variables with normal and nonnormal distribution, respectively. Independent predictors of CAD were identified by logistic regression analysis. Independent predictors of death were identified with Cox regression analysis. Receiver operating characteristic (ROC) curve analysis was used to determine the best cutoff of troponin level to predict CAD and the best timing for coronary angiography. A p-value below 0.05 was considered significant.
All data were analyzed anonymously. Informed Consent was waived because no change was applied to the patients' management. The present study conformed to the ethical guidelines of the Declaration of Helsinki.

RESULTS
The patient demographics are summarized in table 1. One hundred seventeen patients (n=117) were included, mostly male (77%, n = 90), with a mean age of 63 ± 13 years (range 30 -97). median time until ROSC was 10 minutes (IQR = 4 -20). Seventy-six percent of the patients underwent coronary angiography in the first 24 hours after ROSC and 38% in the first 2 hours, with a median time of ROSC to coronary angiography of 3 hours. Table 3 summarizes the characteristics of CA.
In coronary angiography, significant CAD was found in 68.4% (n = 80) of the patients, mainly involving the left anterior descending artery (44%; n = 52), and 14% (n = 16) had left main artery involvement. Considering patients with significant CAD, 75.0% (n = 60) underwent PCI, 12.5% (n = 10) were considered for surgical intervention, and the remaining 12.5% (n = 10) did not undergo revascularization. Among these, four patients were considered to have chronic CAD, since all of these patients had previously known lesions or occluded bypass grafts; three patients had disease in vessels of small diameter only; and the remaining three patients had chronic total occlusions. Considering patients without significant CAD, eight underwent left ventriculography, of whom 71% (n = 5) had wall motion abnormalities (WMAs); one patient had WMAs compatible with Takotsubo syndrome, and 50% (n = 4) had a severely reduced LVEF.
The results of the post-CA ECG were available in 86 patients. Of these, most did not have ST-segment elevation (54.7%, n = 47), and only 12.3% (n = 10) had left bundle branch block. However, only 7% of the patients (n = 6) had an unremarkable ECG. In the subgroup of patients without significant CAD, 17.9% (n = 5) had ST-segment elevation, 44.4% (n = 12) had other electrocardiographic CAD -coronary artery disease. *p indicates the difference between patients with and those without significant coronary artery disease. Results expressed as mean ± standard deviation or n (%).
Cardiovascular risk factors were common: 63% of the patients had arterial hypertension, 43% dyslipidemia, 27% diabetes and 32% a history of smoking. Thirty-seven patients had a previous history of CAD, and 13% had cerebrovascular disease.
In 24.8% of the patients, there was no information regarding the presumed cause of CA before coronary angiography. In the remainder, based on clinical judgment, type 1 myocardial infarction was the presumed cause in 79.5%, mainly with ST-segment elevation (60%). The presumed causes of CA are detailed in table 2.
Most cases (58.2%) were out-of-hospital CA, which usually occurred in shockable rhythm (70.1%). The Considering the laboratory results, only 7% of the patients (n = 7) had troponin in the normal range. In the subgroup without significant CAD, only 3.0% (n = 1) had troponin levels in the normal range. Table 4 details the results of the exams performed after ROSC.
ST-segment elevation was also associated with CAD (88% of patients with ST-segment elevation had significant CAD versus 51% of patients without STsegment elevation; p < 0.001), as were the presence of WMAs in echocardiograms (93% of patients with WMAs had significant CAD versus 36% of patients without; p < 0.001) and an LVEF ≤ 40% (85% of patients with an LVEF ≤ 40% had significant CAD versus 47% of patients with an LVEF > 40%; p = 0.003). None of the other ECG or echocardiogram characteristics were associated with significant CAD.
Considering the comorbidities, only arterial hypertension was found to be marginally associated with significant CAD (75% of patients with hypertension had significant CAD versus 57% of patients without hypertension, p = 0.048).
In the subgroup of patients without ST-segment elevation, 24 (51.1%) had significant CAD, and 54.2% (n = 13) underwent primary PCI. In this subgroup, the presence of WMAs and an LVEF ≤ 40% were also associated with significant CAD: 88% of patients with WMAs had significant CAD versus 25% of patients without WMAs (p < 0.001), and 69% of patients with an LVEF ≤ 40% had significant CAD versus 21%

The impact of coronary artery disease on mortality
Among patients with significant CAD, 56.3% (n = 45) died, and 43.8% (n = 35) survived. The presence of significant CAD was not associated with mortality (p = 0.19). Additionally, in the subgroup of patients with significant CAD who underwent PCI, 53.3% died (n = 32), and 46.7% (n = 28) survived. Percutaneous coronary intervention in patients with significant CAD was not associated with reduced mortality (p = 0.79). Additionally, there was no significant increase in mean survival time in patients with versus those without significant CAD (3.0 ± 9.7 versus 4.7 ± 8.2 months, p = 0.703).
Patients with indications for surgical revascularization had higher 30-day and overall mortality (p = 0.016 and p = 0.012, respectively).
There was no significant difference in the time from ROSC to coronary angiography between patients who survived and those who died (3.0 hours, IQR = 1.5-36 versus 2.9 hours, IQR = 1.9 -21.5). Performing coronary angiography in the first 24 hours after ROSC versus more than 24 hours was not associated with reduced mortality (p = 0.134 for 30-day mortality and p = 0.67 for overall mortality), and the same was true for the 2-hour cutoff suggested in the guidelines (8) (p = 0.35 for 30-day mortality and p = 0.27 for overall mortality). ROC curve analysis further strengthened these results (AUC = 0.5, p = 0.99).
The results of the univariate and multivariate analyses are detailed in table 6. In the multivariate analysis, the time until ROSC was an independent mild predictor of overall mortality (OR 1.015, 95%CI 1.0 -1.03; p = 0.048) and mortality in the first 24 hours after ROSC (OR 1.02, 95%CI 1.0 -1.05; p = 0.022). Conversely, the presence of shockable rhythm was a strong independent predictor of survival (OR 0.4, 95%CI 0.2 -0.9; p = 0.031).

DISCUSSION
We evaluated the role of previous medical history, laboratory values, and electrocardiographic and echocardiographic data, usually available in the emergency department, to predict the presence of CAD in survivors of CA. More than 100 patients were included, and similar to previous studies, almost 70% had significant CAD. (5,7,11,15,16) Our PCI rate (75%) was similar to previous reports, where rates ranged from 53% to 95%. (5,11,12,15,(17)(18)(19)(20)(21) This wide range of PCIs might be justified by the different definitions of significant CAD between studies. Importantly, our study also reported a sizeable proportion of patients considered for surgical intervention (13%), an outcome seldom reported in the literature.
Herein, we found a significant difference in time from CA to coronary angiography between patients with CAD and those without CAD (Table 3), which might be explained by selection bias, as patients with high pretest probability of CAD were probably underwent coronary angiography earlier, and patients with low pretest probability followed a different way of investigation The etiology of CA is often difficult to ascertain in an emergency setting, as medical history is often unavailable and complementary exams results are difficult to interpret. (12) Therefore, the indications and best timing for coronary angiography in survivors of CA are very much a matter of debate. (10,11) In our study, only direct evidence of cardiac abnormalities -ST-segment elevation, WMAs, at least moderate LV systolic dysfunction and an elevation of hs-TnT above 170pg/mL -were predictors of significant CAD.
The ECG role in the post-CA setting as a predictor of ACS is still debatable. (3,19) ST-segment deviation might be caused by electrolyte abnormalities, defibrillation, ischemia-reperfusion injury and intravenous drugs administered during resuscitation, (12,17,19) rendering the ECG difficult to interpret, (22) with poor predictive values reported. (5,11,17,19) However, some studies support our findings. Anyfantakis et al. studied 72 consecutive out-of-hospital CA survivors who underwent systematic emergency coronary angiography and found that ST-segment elevation after ROSC was independently correlated with the presence of CAD and strongly correlated with a final diagnosis of acute myocardial infarction. (12) Additionally, Lee et al. found that ST-segment elevation and new-onset left bundle branch block are highly associated with acute coronary lesions. (17) Emergency echocardiography in the post-CA setting may have limited value, as regional WMAs are not specific for acute myocardial ischemia and may be related to stunning caused by hypoperfusion or by resuscitation maneuvers. (20) Notwithstanding, our findings suggest that emergency echocardiography might be of use in selecting patients for coronary angiography, especially if there are WMAs or an LVEF ≤ 40%.
In our study, 93% of the patients had elevated troponin levels irrespective of the presence of CAD. This finding was also reported in previous studies and reflects the very high sensitivity of the assay, suggesting that myocardial injury biomarkers are not reliable for the diagnosis of ACS in this setting because of their low specificity. (12,17) We found that an elevated hs-TnT value, even with a cutoff value of 170pg/mL (almost 12 times the upper limit of normal), while predictive of CAD, has poor accuracy in CA survivors, in keeping with previous results. (15) These findings render the usefulness of troponin level elevations limited from a clinical perspective.
Overall, our results support the current recommendations, (8) where patients who survived to CA and present with ST-segment elevation or with high suspicion of ACS (here represented by LV systolic dysfunction, WMAs and elevated hs-TnT) should undergo coronary angiography. However, emergency doctors and interventional cardiologists must be aware of the limitations of these findings in this particular setting, which might lead to false-positive and false-negative results.
When managing these patients, one should consider that the treatment of the underlying cause of CA is paramount and that ACS is the most common cause of malignant ventricular arrhythmias leading to sudden cardiac death. (6) Interestingly, we did not find an association between CA with shockable rhythm and CAD, which might have been the result of selection bias; the occurrence of ventricular arrhythmias rendered patients more likely to undergo coronary angiography. This finding is evidenced by the high prevalence of shockable rhythm in our cohort.
The results in the literature on the role of coronary angiography in this setting, performed as soon as possible after ROSC, with PCI as needed, are controversial. Several studies document the survival benefit of early coronary angiography and PCI in CA survivors. (4,5,11,13,16,23) Here, as previously described, (20,21,24) the only predictor of mortality was the time to ROSC, with shockable rhythm being predictive of survival. We found no association between the presence of significant CAD and PCI with survival. Our results are in keeping with those of Garot et al., who studied predictors of survival in CA complicated acute myocardial infarction in 186 patients and found that the rates of PCI success had no impact on survival. (21) The authors concluded that although successful PCI could have played an important role in improving survival, their data indicated that it was not sufficient in itself to show any differences in both 30-day and 6-month mortality rates. (21) Similarly, Anyfantakis et al. found that although PCI was warranted in one-third of patients, it was not associated with hospital survival. (12) These findings might be the result of the small sample size of these studies, including ours, but also because in post-CA survival patients, other therapeutic interventions probably influence the prognosis.
As a result of these findings, we were unable to find a threshold for the best timing for coronary angiography in our patients, nor did the time until coronary angiography have any influence on survival. These findings are in line with the results of the COACT trial, a multicentric prospective study with 522 patients without ST-segment elevation after CA, which found that immediate angiography was not better than delayed angiography with respect to overall survival at 90 days. (7) According to the literature, following successful resuscitation, only one-third of patients die because of cardiovascular causes; approximately one-third of deaths are secondary to central nervous system injury, while the remaining deaths are due to a variety of other reasons. (4) Some authors even report that most patients die from neurological complications. (13) Therefore, the absence of survival impact of early coronary angiography and PCI might be justified by the fact that immediate angiography leads to a delay in the implementation of interventions that might contribute to brain injury prevention, since the current practice in the center is to implement these measures only when the patient is admitted to the intensive care unit. In addition, performing coronary angiography in this setting requires the mobilization of an unstable patient, exposure to contrast agents, and a risk of vascular and bleeding complications, (13) which can contribute to further mortality. Furthermore, even in patients who have genuine ischemic heart disease-related CA, some may have chronic ischemic heart disease that provides an arrhythmogenic substrate for CA. Although these patients are expected to have significant CAD, it is unclear how angiography and PCI would improve their outcomes. (13) Therefore, while the identification of an acute culprit coronary lesion amenable to urgent PCI may be important for the prognosis of CA survivors, this is not the only factor involved, and careful, individual evaluation is critical. Attempts to reduce the time to ROSC and to improve post-ROSC care are probably at least as important as coronary angiography. Further randomized studies with more patients and a prospective design are essential to define the appropriate pathway for therapeutic interventions in survivors of CA -especially for selecting who should undergo coronary angiography and when.
This study has several limitations. It is a single-center retrospective study based on the information available in the electronic clinical records, and as a result, missing data are inevitable; specifically, there was a high number of patients without information regarding the post-ROSC ECG, since many of these patients are referred from other hospitals or the ECG is performed in the prehospital setting. Additionally, there was selection bias, as all patients included underwent coronary angiography based on the clinical decision of the emergency physician and the interventional cardiologist, rendering a high pretest probability, so the prevalence of significant CAD might be overrepresented. As in previous studies, the sample size was small. Another limitation was the use of two different troponin assays (TnI and hs-TnT) because we included the period of transition to a high sensitivity assay in our institution; thus, only a few patients had TnI values available. Finally, CAD was classified as significant based on the visual appreciation of the interventional cardiologist and not a core laboratory; therefore, operator bias cannot be ruled out.

CONCLUSION
Our study suggests that in survivors of cardiac arrest, only direct evidence of cardiac abnormalities is strongly associated with coronary artery disease. STsegment elevation in the post-return of spontaneous circulation electrocardiogram, the presence of wall motion abnormalities, an left ventricular ejection fraction ≤ 40% in transthoracic echocardiography, and an high sensitivity troponin T level above 170pg/mL were predictors of significant coronary artery disease. These results may help in the selection of patients for coronary angiography, thereby reducing unnecessary procedures, and may support current guidelines, which recommend performing coronary angiography in patients with a high index of suspicion of ongoing infarction.
However, in our population, the time until coronary angiography, the presence of significant coronary artery disease and the performance of percutaneous coronary intervention did not influence survival, and it was not possible to establish the best cutoff for coronary angiography timing. This outcome was likely the result of the small sample size and confounding factors, especially noncardiovascular issues affecting prognosis.