Author + information
- Received September 21, 2016
- Revision received November 10, 2016
- Accepted November 30, 2016
- Published online March 6, 2017.
- Youn-Jung Kim, MDa,
- Sun-Yang Min, MD, PhDb,
- Dong Hun Lee, MD, PhDc,
- Byung Kook Lee, MD, PhDc,
- Kyung Woon Jeung, MD, PhDc,
- Hui Jai Lee, MD, PhDd,
- Jonghwan Shin, MD, PhDd,
- Byuk Sung Ko, MDa,
- Shin Ahn, MD, PhDa,
- Gi-Byoung Nam, MD, PhDe,
- Kyoung Soo Lim, MD, PhDa and
- Won Young Kim, MD, PhDa,∗ ()
- aDepartment of Emergency Medicine, Ulsan University College of Medicine, Asan Medical Center, Seoul, Korea
- bHealth Screening and Promotion Center, Asan Medical Center, Seoul, Korea
- cDepartment of Emergency Medicine, Chonnam National University Hospital, Gwangju
- dDepartment of Emergency Medicine, Seoul Metropolitan Government-Seoul National University Boramae Medical Center, Seoul, Korea
- eDepartment of Internal Medicine, Ulsan University College of Medicine, Asan Medical Center, Seoul, Korea
- ↵∗Address for correspondence:
Dr. Won Young Kim, Department of Emergency Medicine, Ulsan University College of Medicine, Asan Medical Center, 88, Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Korea.
Objectives The authors aimed to evaluate the role of post-resuscitation electrocardiogram (ECG) in patients showing significant ST-segment changes on the initial ECG and to provide useful diagnostic indicators for physicians to determine in which out-of-hospital cardiac arrest (OHCA) patients brain computed tomography (CT) should be performed before emergency coronary angiography.
Background The usefulness of immediate brain CT and ECG for all resuscitated patients with nontraumatic OHCA remains controversial.
Methods Between January 2010 and December 2014, 1,088 consecutive adult nontraumatic patients with return of spontaneous circulation who visited the emergency department of 3 tertiary care hospitals were enrolled. After excluding 245 patients with obvious extracardiac causes, 200 patients were finally included.
Results The patients were categorized into 2 groups: those with ST-segment changes with spontaneous subarachnoid hemorrhage (SAH) (n = 50) and those with OHCA of suspected cardiac origin group (n = 150). The combination of 4 ECG characteristics including narrow QRS (<120 ms), atrial fibrillation, prolonged QTc interval (≥460 ms), and ≥4 ST-segment depressions had a 66.0% sensitivity, 80.0% specificity, 52.4% positive predictive value, and 87.6% negative predictive value for predicting SAH. The area under the receiver-operating characteristic curves in the post-resuscitation ECG findings was 0.816 for SAH.
Conclusions SAH was observed in a substantial number of OHCA survivors (25.0%) with significant ST-segment changes on post-resuscitation ECG. Resuscitated patients with narrow QRS complex and any 2 ECG findings of atrial fibrillation, QTc interval prolongation, or ≥4 ST-segment depressions may help identify patients who need brain CT as the next diagnostic work-up.
The current advanced cardiac life support guidelines suggest performing a 12-lead electrocardiogram (ECG) as soon as possible after return of spontaneous circulation (ROSC) in order to identify patients with post-cardiac arrest who needed urgent coronary angiography (CAG) (1,2). They also recommended appropriate percutaneous intervention (PCI) in emergency situations for out-of-hospital cardiac arrest (OHCA) survivors with ST-segment elevation and suspected cardiac origin without ST-segment elevation (1,2). However, spontaneous subarachnoid hemorrhage (SAH) is a well-known cause of cardiac arrest. In 3% to 16% of the cases of nontraumatic cardiac arrests, ECG changes such as ST-segment elevation or ST-segment depression and abnormal T-wave morphologies are seen, which may mimic myocardial infarction or ischemia (3–8). Furthermore, patients with SAH have elevated troponin levels. These nonspecific findings can lead to misdiagnosis and incorrect therapeutic decisions such as administration of fibrinolytics and PCI in these cardiac arrest survivors (9).
The need and timing for brain computed tomography (CT) in nontraumatic OHCA remain controversial (10). Although brain CT is necessary for early identification of SAH, it is not routinely recommended because of its potential disadvantages such as increased cardiovascular instability in OHCA patients with a risk of fatal dysrhythmia and delay in emergency CAG/PCI during CT (1,2). Therefore, it is reasonable to perform an immediate CT scan only in patients who are likely to develop brain hemorrhage. Because history taking is impossible in many OHCA survivors, more specific knowledge of ECG changes in OHCA patients with SAH, which lead to cardiac arrest, may help accelerate the selection of appropriate therapies (11–13).
The present study aimed at determining the role of the post-resuscitation ECG in patients with significant ST-segment changes on initial ECG to investigate the difference in post-resuscitation ECG characteristics between OHCA patients with SAH and those with suspected cardiac origin of OHCA. Such information may provide useful diagnostic clues for physicians to decide whether to perform a CT before urgent CAG in patients with OHCA.
Study design and patients
This multicenter retrospective, observational, registry-based study was conducted at 3 emergency departments, which were affiliated with a university-affiliated teaching hospital in Korea. Data were extracted from the OHCA registry, which prospectively collected data of consecutive patients with OHCA between January 2010 and December 2014. The institutional review board of the University of Ulsan College of Medicine reviewed the study protocol and approved the study. Informed consent was waived due to the retrospective nature of this study. Patients who met the following criteria were included in this study: nontraumatic OHCA; age ≥18 years; sustained ROSC (defined as return of evident signs of circulation for more than 20 consecutive minutes); no obvious extracardiac cause of OHCA such as hanging, drowning, asphyxia, or poisoning; post-resuscitation ECG demonstrating significant ST-segment changes; and brain CT examined within 3 h after resuscitation. Significant ST-segment changes were defined as ST-segment elevation or ST-segment depression present in at least 2 contiguous leads in a 12-lead ECG performed after ROSC. Patients were excluded if their post-resuscitation ECG was lost, if they presented with no/insignificant ST-segment changes, or if they did not undergo brain CT within 3 h after ROSC.
Management and data collection
During the study period, OHCA patients were administered cardiopulmonary resuscitation (CPR) and post-resuscitation care, including CAG/PCI and targeted temperature management, in accordance with the then-current advanced cardiac life support guidelines of 2010 (14,15). An acute cardiac care physician and an interventional cardiologist made the decision for CAG. Brain CT was performed for all the study patients as soon as possible. Brain CT images obtained within 3 h after ROSC were evaluated by board-certified radiologists of each center. Hyperattenuation of the sulci and/or basal cisterns on non-contrast–enhanced brain CT was defined as acute SAH. The radiologists were careful to differentiate SAH from the pseudo-SAH sign, which is defined as hypoxia-induced high attenuation areas along the basal cisterns or cortical sulci, mimicking SAH. We reviewed all patient medical records to evaluate clinical characteristics.
Post-resuscitation ECG used for analysis was the first interpretable 12-lead ECG performed immediately after ROSC. The ECG was recorded at a paper speed of 25 mm/s and amplification of 10 mm/mV. To compare the J-wave distribution, the ECG lead areas were grouped as the right precordial (V1 to V3), left precordial (V4 to V6), high lateral (I, aVL), and inferior (II, III, aVF). One experienced cardiologist blinded to the clinical history of the patients and angiographic and brain CT findings retrospectively analyzed these ECGs. Recorded ECG changes were ST-segment elevation, ST-segment depression, QRS widening, QTc prolongation, and nonspecific ST-T changes. An ST-segment elevation was defined as an elevation of ≥1 mm in all leads other than leads V2 to V3, where the following cutoff points were applied: an elevation of ≥2.5 mm in men age <40 years; ≥2 mm in men age ≥40 years, or ≥1.5 mV in women (16). An ST-segment depression was defined as depression of ≥0.5 mm in 2 contiguous leads. The number of leads showing ST-segment depression and ST-segment elevation was calculated, and the magnitude of ST-segment depression and ST-segment elevation was measured. The QTc interval was calculated after correction for heart rate using Bazett's formula (17), and QT prolongation was defined as QTc ≥460 ms.
On the basis of the post-resuscitation ECG and brain CT findings, patients were categorized into 2 groups: ST-segment changes with SAH, and OHCA of suspected cardiac origin (ST-segment change without SAH). The clinical and demographic characteristics of all patients, including age, sex, patient condition at the scene, bystander administration of CPR, causes of cardiac arrest, and post-resuscitation 12-lead ECG were compared. To describe the differences in clinical characteristics, patients in the suspected cardiac origin group were subsequently categorized into 2 subgroups: a group with ST-segment elevation on ECG (STE-ECG), and a group with other ECG patterns (NSTE-ECG).
Continuous variables are expressed as mean ± SD or median (interquartile range), depending on the data distribution. Categorical data are presented as absolute numbers and percent frequencies. Differences between means were analyzed by the Student t test and Mann-Whitney U test. Differences between categorical variables were analyzed by the chi-square test or Fisher exact test, as appropriate. At baseline, ECG findings of potential predictors of SAH were first examined using univariate logistic analysis with <0.05 as the p value cutoff. Of the significant variables, we selected the variables on the basis of clinical judgment. Atrial fibrillation, narrow QRS complex, QTc prolongation, and ST-segment depression in ≥4 of 12 leads were candidates for the multivariable model and were examined using multiple logistic regression analysis. The results of the multivariate logistic regression analyses were summarized by estimating the odds ratios and 95% confidence intervals (CIs). The Hosmer-Lemeshow test for logistic regression model was performed. To assess the ability of post-resuscitation ECG findings to predict SAH, area under the receiver-operating characteristic curves (AUCs) were calculated. Sensitivity, specificity, positive predictive value, negative predictive value (NPV), positive likelihood ratio, and negative likelihood ratio for SAH were calculated using a binomial 95% CI. A 2-sided p value of <0.05 was considered statistically significant. All statistical analyses were performed using SPSS version 21.0 (SPSS, Chicago, Illinois) and MEDCALC software version 18.104.22.168 (Medcalc Software, Mariakerke, Belgium).
During the study period, 2,087 nontraumatic OHCA patients presented to the emergency department, and 1,088 patients achieved sustained ROSC (Figure 1). A total of 245 patients were excluded from the study because they had obvious extracardiac causes such as asphyxia (n = 106), hanging (n = 65), drowning (n = 27), and poisoning (n = 14), or their post-resuscitation ECG was lost (n = 33). A total of 843 post-resuscitation ECGs were analyzed, and 591 were excluded because of no or insignificant ST-segment changes. Fifty-two of the 252 patients with ST-segment changes in their post-resuscitation ECG were excluded because they did not undergo brain CT within 3 h after ROSC. Thus, 200 patients were finally included in the study. Patients were categorized into an SAH group (n = 50) and OHCA of suspected cardiac origin group (n = 150). The suspected cardiac origin group was further categorized into the STE-ECG group (n = 54) and the NSTE-ECG group (n = 96).
Comparison of clinical characteristics
Demographic and baseline characteristics of our study patients between SAH and suspected cardiac origin group are summarized in Table 1. The median age of our patients was 59 years, with male predominance (60.0%). The median duration of arrest was 32.5 min. The patients in the SAH group were younger than those in the suspected cardiac origin group (median 52 vs. 61 years; p < 0.001) and less frequently had a witnessed cardiac arrest (50.0% vs. 70.0%; p = 0.01). Asystole was documented as an initial rhythm in one-half of the patients in the SAH group (50.0%), and shockable rhythm was noted in 53.3% of patients in the suspected cardiac origin group (53.3%). Although the potassium and lactic acid levels did not differ between the 2 groups, troponin I levels were lower in patients with SAH than in those with OHCA of suspected cardiac origin (median 0.05 vs. 0.34 ng/ml; p < 0.001).
Comparison of 12-lead ECG findings
Post-resuscitation ECG findings such as rate, rhythm, QRS interval, and QTc interval were significantly different between the 2 groups (Table 1). The mean heart rate of patients in the SAH group was 20 beats more than those in the suspected cardiac origin group in 1 min (mean 124 vs. 104 beats/min; p < 0.001), and the predominant post-resuscitation rhythm in the SAH group was atrial fibrillation (56.0%) in contrast to sinus rhythm in the suspected cardiac origin group (56.0%). In addition, the interval of the QRS complex was significantly shorter in the SAH group than in the suspected cardiac origin group (median 97.0 vs. 108.0 ms; p < 0.001), whereas the QTc interval was significantly prolonged in the SAH group (median 493.0 vs. 471.5 ms; p = 0.01). These differences were also observed when comparing the SAH group, STE-ECG group, and NSTE-ECG group; detailed information on this is presented in Online Table 1 in the Appendix.
Table 2 showed the results of post-resuscitation ECG analysis for the SAH and suspected cardiac origin groups. In contrast to ST-segment elevation, the number of ST-segment depressions and the sum of the measurement of ST-segment depression were significantly different between the SAH and suspected cardiac origin groups: multiple ST-segment depressions (4.0 vs. 3.0; p = 0.004) and ST-segment elevation in aVR lead (44.0% vs. 26.0%; p = 0.02) were more common in the SAH group. Changes in the ST-segment elevation in the suspected cardiac origin group were distinctive in V3, V4 in contrast to the changes observed in the SAH group. In addition, in the SAH group, ST-segment depression was dominant in the left precordial leads (V4 to V6), and about two-thirds of the patients in the SAH group showed ST-segment depression. Additionally, patients in the SAH group had ST-segment depression in leads II and aVF more frequently than those in the suspected cardiac origin group (40.0% vs. 23.3% [lead II]; p = 0.02; 42.0% vs. 20.0% [lead aVF]; p = 0.002). The patterns of ST-segment changes among the SAH, STE-ECG, and NSTE-ECG groups are presented in Online Figure 1: The patterns of ST-segment changes were similar between the SAH group and NSTE-ECG group.
Predictive value of ECG findings
After including factors that showed significant differences in ECG findings in the univariate analysis, multivariate analysis showed that atrial fibrillation, QRS and QTc intervals, and the number of ST-segment depressions were significant factors. Although QRS and QTc intervals and the number of ST-segment depressions were continuous variables, to simplify the calculation, they were converted into binary variables (QRS interval <120 vs. ≥120 ms; QTc interval ≥460 vs. <460 ms; and ST-segment depressions ≥4 vs. <4) (Table 3). On the basis of the multivariate analysis, we set a limit to predict SAH. Predictors for SAH were narrow QRS complex (2 points), atrial fibrillation (1 point), QTc prolongation (1 point), and ST-segment depressions ≥4 in all leads (1 point). Through this model, the sum of scores ranged between 0 and 5, and a higher global score indicated a higher likelihood of SAH in brain CT. The diagnostic value of our post-resuscitation ECG findings to predict SAH was analyzed using Youden’s J statistic. We focused on specificity more than sensitivity to analyze SAH and selected a threshold score of 4 for the highest sum of sensitivity and specificity (Table 4). A score ≥4 had 66.0% sensitivity, 80.0% specificity, 52.4% positive predictive value, 87.6% NPV, 3.3 positive likelihood ratio, and 0.4 negative likelihood ratio. The AUC in our post-resuscitation ECG findings was 0.816 (95% CI: 0.751 to 0.880) for SAH (Figure 2).
Owing to the problems associated with history-taking and false-positive elevated cardiac enzyme, the standard 12-lead ECG is a key test for resuscitated OHCA patients suspected of having acute coronary occlusion as a cause of cardiac arrest. However, previous studies have examined the value of post-resuscitation ECG for diagnosis of acute coronary disease in OHCA survivors. They found that ECG findings, except ST-segment elevation, have poor diagnostic accuracy (11–13,18). Some researchers recommend that all comatose OHCA survivors without obvious extracardiac arrest cause should undergo emergency CAG, irrespective of the ECG findings (12,19,20). We conducted this study to evaluate the role of post-resuscitation ECG in patients with significant ST-segment changes observed on initial ECG to decide whether to perform a CT scan before emergency CAG in patients with OHCA. Our main findings were as follows: 1) SAH was present in 25.0% (50 of 200) of OHCA survivors with significant ST-segment changes on ECG performed immediately after ROSC; and 2) the combination of 4 ECG characteristics including narrow QRS, atrial fibrillation, QTc interval prolongation, and ≥4 ST-segment depressions had 80.0% specificity and 87.6% NPV for identification of SAH. Considering that SAH was observed in a substantial number of OHCA survivors, immediate brain CT before emergency CAG in selective OHCA patients could reduce unnecessary diagnostic work-up.
ECG changes following SAH have been well documented, and these changes are associated with the severity of SAH (21–23). The largest study of SAH patients included 406 pre-operative SAH patients, of which 36% showed supraventricular arrhythmias (atrial fibrillation in 8% cases), 15% showed ST-segment depression, 32% showed T-wave abnormalities, and 32% showed prolonged QTc, which seem to indicate the presence of a repolarization abnormality (24). However, the post-resuscitation ECG characteristics of SAH were not well known. Mitsuma et al. (5) reported that all OHCA survivors with SAH (n = 10) showed QTc interval prolongation and multiple ST-segment depressions, and Yamashina et al. (25) showed that 80% of the OHCA survivors with SAH (16 of 20) had at least 1 ST-segment depression on their post-resuscitation ECG. Despite the small number of patients, these studies highlighted the presence of prolonged QTc interval and multiple ST-segment depressions, which are consistent with our results. However, compared with our previous research, wherein we found that ST-segment elevation in the aVR lead was seen in 51.4% of ECGs of intracranial hemorrhage as a cause of OHCA, ST-segment elevation in aVR provided no significant value to differentiate OHCA survivors with SAH from those with suspected cardiac origin (7). This could be due to the presence of left main coronary lesions, which cause ST-segment elevation in aVR, in the cardiac origin group (26). It is interesting to note that a narrow QRS interval and atrial fibrillation were distinct findings on the post-resuscitation ECG of OHCA survivors with SAH as compared with those without SAH. To our knowledge, no report has thus far described these findings. We believe that these manifestations may have been caused by excessive catecholamine surge with hyperactivity of the sympathetic nervous system in OHCA patients with SAH (27).
Our study determined the frequency of SAH in ROSC patients with significant ST-segment change on post-resuscitated ECG, for the first time. Although several studies have determined the frequency of SAH in OHCA patients, our results may add information about the timing and utility of brain CT examination in resuscitated patients, which are still under debate in the current guidelines on advanced cardiac life support and are determined as per the physician’s discretion. In addition, we found that SAH patients have lower initial troponin levels than those experiencing an OHCA of suspected cardiac origin (median 0.05 vs. 0.34 ng/ml; p < 0.001). However, the diagnostic value of troponin was not significant in the multivariate analysis, which implies that the post-resuscitated ECG remains a key test to determine whether the OHCA is of cardiac origin.
In the present study, we suggested that a simple combination of post-resuscitation ECG indicators would differentiate between OHCA survivors with SAH and those with suspected cardiac origin, irrespective of the history and other clinical variables. Resuscitated patients with narrow QRS complex and any 2 ECG findings of atrial fibrillation, QTc interval prolongation, or ≥4 ST-segment depressions may help identify patients who need brain CT among resuscitated patients with significant ST-segment changes.
In our study, 52 patients did not undergo brain CT within 3 h after ROSC, and because we could not confirm that there was no possibility of brain lesion, we excluded them. However, they were suspected of having a cardiac origin of arrest on the basis of coronary angiography (n = 34, 65.4%) and bedside echocardiography (n = 18, 34.6%). When we included them as suspected cardiogenic group, the differences in ECG patterns including post-resuscitation rhythm, QRS and QTc intervals, and changes in the ST-segment tended to be more prominent between the 2 groups. In multivariate analysis, the predictors for SAH including atrial fibrillation, QRS and QTc intervals, and the number of ST-segment depressions were also significant with the AUC of 0. 834 (95% CI: 0.773 to 0.834) for SAH.
Certain issues should be considered when interpreting the findings of the present study. Due to the retrospective nature of the study, patient eligibility may have been subject to selection bias; for example, 52 of the 252 patients with ST-segment changes observed on post-resuscitation ECG were excluded because they did not undergo brain CT within 3 h after ROSC. On the other hand, the consecutive inclusion of patients reflects a real-world situation, free of typical inclusion bias. Additionally, the epidemiology of OHCA is different in southeast Asia compared with other parts of the world and might limit the generalizability of the findings. Another issue is that some patients did not undergo CAG. Therefore, the etiology of cardiac arrest was uncertain. In addition, one board-certified cardiologist interpreted the post-resuscitation ECG in a blinded manner to other clinical history and results. Finally, the norm in South Korea is that all patients found to have OHCA are transported to the hospital with ongoing CPR. Therefore, the prevalence of SAH may be higher than that in other parts of the world. Further prospective multicenter studies are required to validate the predictive value of the ECG findings proposed in this study.
The interpretation of post-resuscitation ECG may be used as a predictive tool to determine the next diagnostic work-up in OHCA survivors. Although the timing of brain CT and CAG has been controversial, immediate brain CT could be considered in resuscitated patients with a narrow QRS complex, atrial fibrillation, QTc interval prolongation, and ≥4 ST-segment depressions on their post-resuscitation ECG before CAG.
WHAT IS KNOWN? The current advanced cardiac life support guidelines recommended emergent percutaneous intervention for OHCA survivors with ST-segment elevation and suspected cardiac origin without ST-segment elevation. However, spontaneous SAH is a well-known cause of cardiac arrest, and its electrocardiogram may mimic myocardial infarction or ischemia. The need and timing for brain computed tomography in nontraumatic OHCA remain controversial.
WHAT IS NEW? In OHCA survivors with significant ST-segment changes on their post-resuscitation electrocardiogram, the combination of 4 ECG characteristics including narrow QRS (<120 ms), atrial fibrillation, prolonged QTc interval (≥460 ms), and ≥4 ST-segment depressions could be a predictive tool of SAH.
WHAT IS NEXT? Prospective multicenter studies are needed to validate the predictive value of the ECG findings proposed in this study, especially in the western area.
For a supplemental figure and tables, please see the online version of this article.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- area under the receiver-operating characteristic curve
- coronary angiography
- confidence interval
- cardiopulmonary resuscitation
- computed tomography
- negative predictive value
- non–ST-segment elevation on electrocardiogram
- out-of-hospital cardiac arrest
- percutaneous intervention
- return of spontaneous circulation
- subarachnoid hemorrhage
- ST-segment elevation on electrocardiogram
- Received September 21, 2016.
- Revision received November 10, 2016.
- Accepted November 30, 2016.
- American College of Cardiology Foundation
- Callaway C.W.,
- Donnino M.W.,
- Fink E.L.,
- et al.
- Sommargren C.E.
- Mitsuma W.,
- Ito M.,
- Kodama M.,
- et al.
- Arnaout M.,
- Mongardon N.,
- Deye N.,
- et al.
- Kim Y.J.,
- Kim W.Y.,
- Nam G.B.
- Park I.,
- Kim Y.J.,
- Ahn S.,
- Sohn C.H.,
- Seo D.W.,
- Kim W.Y.
- Prout R.,
- Nolan J.
- Staer-Jensen H.,
- Nakstad E.R.,
- Fossum E.,
- et al.
- Neumar R.W.,
- Otto C.W.,
- Link M.S.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- Jaffe A.S.,
- et al.
- Bazett H.C.
- Abe T.,
- Watanabe S.,
- Mizuno A.,
- Toyama M.,
- Totten V.Y.,
- Tokuda Y.
- Dumas F.,
- Cariou A.,
- Manzo-Silberman S.,
- et al.
- Katsanos A.H.,
- Korantzopoulos P.,
- Tsivgoulis G.,
- Kyritsis A.P.,
- Kosmidou M.,
- Giannopoulos S.
- Qaqa A.Y.,
- Suleiman A.,
- Alsumrain M.,
- Debari V.A.,
- Kirmani J.,
- Shamoon F.E.
- Rudehill A.,
- Olsson G.L.,
- Sundqvist K.,
- Gordon E.
- Yamashina Y.,
- Yagi T.,
- Ishida A.,
- et al.
- Yamaji H.,
- Iwasaki K.,
- Kusachi S.,
- et al.
- Samuels M.A.