Electrocardiographic Predictors of Out-of-Hospital Sudden Cardiac Arrest in Patients With Coronary Artery Disease

Electrocardiographic Predictors of Out-of-Hospital Sudden Cardiac Arrest in Patients With Coronary Artery Disease Miguel E. Lemmert, MDa,*, Jacqueline J.M. de Vreede-Swagemakers, PhDb, Luc W.M. Eurlings, MDa, Luc Kalb, MDa, Harry J.G.M. Crijns, MD, PhDa, Hein J.J. Wellens, MD, PhDc, and Anton P.M. Gorgels, MD, PhDa Sudden cardiac arrest (SCA), due mainly to coronary artery disease (CAD), is a leading cause of death. To identify electrocardiographic and clinical differences between patients with CAD with and without SCA, 87 victims of SCA with CAD were compared with 131 patients with CAD without SCA. Patients’ latest routine electrocardiograms and clinical variables were compared. Patients with CAD with and without previous myocardial infarctions (MIs) were included. Patients with SCA had a higher incidence of echocardiographic evidence of left ventricular hypertrophy and/or heart failure than controls. The median left ventricular ejection fractions for patients with SCA with and without previous MIs were 0.30 (interquartile range 0.24 to 0.41) and 0.41 (interquartile range 0.25 to 0.56). The median time between the last electrocardiographic assessment and SCA was 59 days (interquartile range 29 to 137). Regarding electrocardiographic characteristics, in patients with and without previous MIs, QRS width (odds ratio 1.032, 95% confidence interval 1.012 to 1.053, p ⴝ 0.002, and odds ratio 1.035, 95% confidence interval 1.015 to 1.056, p ⴝ 0.001) was the only significant predictor of SCA. In conclusion, in patients with CAD, regardless of a previous MI, a longer QRS width and echocardiographic parameters consistent with heart failure are associated with SCA, even in patients with ischemic cardiomyopathy currently not eligible for an implantable cardioverter-defibrillator. © 2012 Elsevier Inc. All rights reserved. (Am J Cardiol 2012;109:1278 –1282) Sudden cardiac arrest (SCA) is a leading mode of death in industrialized countries, frequently with ventricular fibrillation (VF) as the underlying arrhythmogenic mechanism. VF may occur in the acute and chronic phases of myocardial infarction (MI) or in ischemic conditions without apparent infarction.1,2 Using 12-lead electrocardiography, we recently showed longer conduction intervals, independent of the amount of ischemia, in patients with first ST-segment elevation MI (STEMI) developing ischemic VF.3 Inhomogeneity of intramyocardial conduction velocity plays a role as a substrate for reentrant ventricular arrhythmias and sudden death during acute ischemia,4 and we found QRS widening related to the infarct location as a possible expression of such inhomogeneity of conduction velocity.3 To develop strategies to prevent SCA, it is important to correctly select patients at highest risk for SCA. In the present study, we studied victims of SCA with known coronary artery disease (CAD) and evaluated possible differences in clinical and electrocardiographic (ECG) characteristics, the latter possibly pointing to inhomogeneity of

a Department of Cardiology, Maastricht University Medical Center; Public Health Service (GGD); and cCardiovascular Research Institute Maastricht, Maastricht, The Netherlands. Manuscript received September 12, 2011; revised manuscript received and accepted December 19, 2011. Dr. Lemmert is supported as a PhD fellow by an unrestricted grant from Philips Healthcare, Seattle, Washington. *Corresponding author: Tel: 31-43-3875072; fax: 31-43-3875104. E-mail address: [email protected] (M.E. Lemmert). b

0002-9149/12/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2011.12.020

conduction velocity, already present before the event in nonischemic conditions. Methods A registry of victims of SCA, described previously,2 was used. Briefly, all victims of out-of-hospital SCA aged 20 to 75 years living in the Maastricht region were included. The Maastricht region is 203 km2 in size and has approximately 181,500 inhabitants, of whom approximately 133,000 (73%) are aged 20 to 75 years. Served by 1 hospital, 1 emergency medical service, and a network of cooperative general practitioners, the region is suitable for populationbased studies. All witnessed and unwitnessed (when circumstances were suggestive of unexpected SCA) events of out-of-hospital SCA were included in the registry. Patients with circulatory arrest caused by trauma or intoxication or SCA occurring in the terminal phase of a chronic disease were not included. For all victims of SCA included in the registry, the last available routine ECG and clinical characteristics before the events were collected from the hospital, general practitioners, family members, ambulance staff members, or a combination of these. For the present study, 87 victims of SCA from the SCA registry with known CAD were compared to 1.5 times as many randomly derived patients with CAD without SCA (controls), providing a total of 131 control patients. These control patients were random patients with CAD drawn from the patient population from the Department of Cardiology, applying the same age restriction, but not matched in another way with the patients with SCA. www.ajconline.org

Coronary Artery Disease/SCA in Patients With CAD

1279

Table 1 Clinical characteristics of the study population Variable

Infarction on Electrocardiogram Yes

Hypertension No sports Body mass index (kg/m2) Heart failure Medications Digoxin Diuretics Nitrates Antiplatelet agents Echocardiography Left atrial dilation Left ventricular hypertrophy Left ventricular dilation Tricuspid regurgitation Left ventricular ejection fraction (%) Left ventricular end-diastolic diameter (mm) Left ventricular end-systolic diameter (mm)

No

SCA Patients (n ⫽ 59)

Controls (n ⫽ 61)

p Value*

SCA Patients (n ⫽ 28)

Controls (n ⫽ 70)

p Value*

50% 93% 25 (23–27) 37%

27% 79% 27 (24–28) 10%

0.036 0.007 ⬍0.001

39% 86% 24 (22–27) 32%

20% 60% 26 (24–28) 3%

0.073 0.017 0.032 ⬍0.001

27% 54% 58% 20%

5% 15% 44% 51%

0.001 ⬍0.001 0.150 0.001

25% 39% 68% 39%

4% 13% 44% 63%

48% 31% 18% 41% 30 (24–41) 55 (52–64) 42 (38–51)

29% 11% 6% 24% 45 (38–59) 50 (46–57) 34 (29–38)

0.051 0.018 0.074 0.068 ⬍0.001 ⬍0.001 ⬍0.001

46% 39% 27% 46% 41 (25–56) 56 (48–63) 38 (32–51)

18% 18% 2% 17% 60 (45–64) 49 (46–52) 32 (28–34)

0.005 0.006 0.045 0.043 0.015 0.058 0.001 0.007 ⬍0.001 0.002 ⬍0.001

Data are expressed as percentages or as median (interquartile range). * Student’s t test, Mann-Whitney U test, or Fisher’s exact test as appropriate. Table 2 Electrocardiographic characteristics of the study population Variable

Infarction on Electrocardiogram Yes

Heart rate (beats/min) QRS width (ms) Fridericia corrected QT interval (ms) Left bundle branch block Right bundle branch block T axis intermediate T axis left T axis right T axis extreme right Left atrial hypertrophy Left ventricular hypertrophy Sinus rhythm Atrial fibrillation Ventricular ectopic beats

No

SCA Patients (n ⫽ 59)

Controls (n ⫽ 61)

p Value*

SCA Patients (n ⫽ 28)

Controls (n ⫽ 70)

p Value*

76 (69–96) 112 (100–124) 424 (393–461) 10% 9% 37% 20% 27% 15% 52% 17% 88% 10% 25%

74 (67–86) 92 (85–105) 396 (382–424) 2% 7% 72% 7% 21% 0% 30% 5% 100% 0% 7%

0.128 ⬍0.001 0.085 0.059 0.741 ⬍0.001 0.033 0.525 0.001 0.021 0.042 0.006 0.012 0.006

75 (68–87) 100 (89–146) 412 (383–428) 29% 14% 50% 0% 25% 25% 39% 21% 89% 11% 25%

68 (58–76) 94 (86–102) 399 (388–413) 4% 6% 88% 0% 10% 1% 17% 16% 96% 4% 0%

0.020 0.013 0.735 0.002 0.220 ⬍0.001 NA 0.106 0.001 0.054 0.561 0.344 0.344 ⬍0.001

Data are expressed as percentages or as median (interquartile range). * Student’s t test, Mann-Whitney U test, or Fisher’s exact test as appropriate. NA ⫽ not applicable.

To study the characteristics of patients with CAD without previous MIs, patients were stratified according to the ECG presence of a previous MI. Clinical characteristics, including baseline characteristics, cardiovascular risk factors and history, use of medication, and echocardiographic characteristics including ejection fraction, left ventricular dimensions, valvular disease, and left atrial dilatation, were analyzed. For SCA and control patients, the most recent standard 12-lead 10-second paper electrocardiogram, recorded dur-

ing the last routine visit, was analyzed. The electrocardiograms were recorded using a Marquette MAC VU recorder, providing heart rate, PR interval, QRS width, QRS axis, and axis of the T wave in the frontal plane. Additional ECG parameters collected on the single electrocardiograms included rhythm, QT interval measured in lead V5, corrected QT interval (using Fridericia’s formula), signs of a previous MI according to the Minnesota criteria,5 bundle branch block, left, right, and biatrial hypertrophy, left ventricular hypertrophy, and presence of ventricular ectopic beats.

1280

The American Journal of Cardiology (www.ajconline.org)

Data analysis was performed using SPSS for Windows release 16.0.1 (SPSS, Inc., Chicago, Illinois). Continuous variables are expressed as medians and interquartile ranges and categorical variables as percentages. For comparison of continuous variables, Student’s t tests for normally distributed data and Mann-Whitney U tests for non-normally distributed data were used. For comparison of categorical variables, Fisher’s exact tests were used. All statistical tests were 2 tailed, and p values ⬍0.05 were considered statistically significant. ECG characteristics showing significant univariate relations with the occurrence of SCA, but lacking multicollinearity, were included in multivariate logistic regression. Variables were removed stepwise from the model when the p value exceeded 0.10. Variables with p values ⬍0.05 in the final model were considered main effect predictors. The predictive accuracy of the final model is reported as the area under the receiver-operating characteristic curve. Cut-off values for ECG characteristics by which most patients with SCA can be correctly classified are identified by applying the Pythagorean theorem to receiver-operating characteristic curves, which is a mathematical determination of the cut-off value closest to sensitivity and specificity of 1. Results The median time between the last available electrocardiogram and SCA was 59 days (interquartile range 29 to 137). In 25 of the 87 patients with SCA (29%), the initial rhythm documented by the emergency medical services was VF. In 22 patients with SCA, other rhythms were documented, whereas in 40 patients with SCA, no initial rhythm was documented. A total of 120 patients had electrocardiographically documented previous MIs, and of these patients, 34 showed signs of previous anterior MIs, 66 of nonanterior MIs, and 20 of anterior and nonanterior MIs. Significantly different clinical characteristics are listed in Table 1. SCA and control patients were not significantly different regarding baseline characteristics. Echocardiographic data were available in 111 of 120 patients with previous MIs and in 86 of 98 patients without previous MIs. Given the large number of significantly different variables relative to the number of patients with SCA, no multivariate analysis was applied for the clinical characteristics. Table 2 lists the significantly different ECG variables, as well as the nonsignificant difference in corrected QT intervals. For the 2 groups, logistic regression was applied, with SCA as the dependent variable and ECG characteristics showing univariate significance as the independent variables, including bundle branch block. In view of multicollinearity and/or the number of SCA cases relative to the independent variables, the selected independent variables in the infarction group were QRS width, presence of left and right bundle branch block, left atrial hypertrophy, left ventricular hypertrophy, sinus rhythm, and ventricular ectopic beats. For similar reasons, the independent variables in the no-infarction group were heart rate, QRS width, presence of left and right bundle branch block, and ventricular ectopic beats.

In the infarction and no-infarction groups, QRS width was the only independent predictor of SCA (odds ratio 1.032, 95% confidence interval [CI] 1.012 to 1.053, p ⫽ 0.002, and odds ratio 1.035, 95% CI 1.015 to 1.056, p ⫽ 0.001, with predictive accuracy of 0.792 and 0.774, respectively). Discussion Our study shows that QRS width, measured on the electrocardiograms of nonacutely ischemic patients with CAD, plays a significant role preceding SCA. In general, patients with SCA with or without previous MIs showed more frequent lack of sports, hypertension, and/or heart failure and also use of medications, reflecting the higher prevalence of heart failure and/or hypertension. In the infarction and noinfarction groups, patients with SCA had significantly lower body mass indexes. This may be explained by the higher prevalence of heart failure in the patients with SCA: the lower body mass index could be related to more advanced stages of heart failure. In the 2 groups, left ventricular hypertrophy or left atrial and ventricular dilatation were significantly more frequent in patients with SCA, as well as a decreased left ventricular ejection fraction and increased left ventricular end-diastolic and end-systolic diameters. Heart failure has previously been shown to be an important determinant of the risk for sudden arrhythmic death.6,7 Furthermore, left ventricular hypertrophy in particular could lead to the genesis of ventricular tachyarrhythmias and SCA by a number of mechanisms, including increased ventricular mass, interstitial fibrosis (especially in the presence of fibrosis due to previous MI), and oxygen supplydemand mismatch. Our findings suggest that apart from measurement of the left ventricular ejection fraction, the aforementioned echocardiographic parameters may be useful in recognizing patients at risk for SCA. In the infarction and no-infarction groups, QRS width was an independent predictor of SCA, importantly independent of bundle branch block. The significantly broader QRS widths in the patients with SCA coincide with our previous findings in 12-lead STEMI electrocardiograms.3,8 Briefly, in those studies, we found longer conduction intervals, independent of the amount of ischemia, in patients with STEMI who developed ischemic VF during STEMI. We proposed that these findings could indicate inhomogeneity in conduction velocity induced by local ischemia, providing a substrate for VF. In contrast with those studies, the present study evaluates the nonischemic period before the occurrence of SCA. Unfortunately, in many of the patients with SCA, no documentation of the initial rhythm was available. Given the facts that in patients aged ⬎35 years, SCA is caused by CAD in 80% of cases,1 with ischemic VF as a leading cause, and that our study population consisted only of patients with CAD, it could be speculated that many of the unknown initial rhythms were actually tachyarrhythmic. Assuming this, the broader QRS widths in the patients with SCA may indicate already present slowing of conduction velocity and/or increased wall thickness, which may be a substrate for reentry, especially when triggered by local ischemia. In univariate analysis, the absence of an intermediate T axis was more frequent in patients with SCA in the infarc-

Coronary Artery Disease/SCA in Patients With CAD

tion and no-infarction groups. These findings are in accordance with previous research,9 in which an abnormal T axis was associated with sudden death in patients aged ⬎55 years. The role of the direction of the T axis was not analyzed further in multivariate analysis, in view of multicollinearity and the number of SCA cases relative to the independent variables. In univariate analysis, patients with SCA more frequently showed ventricular ectopic beats. Among patients with previous MIs, those with SCA also more frequently had atrial fibrillation. The larger number of ventricular ectopic beats and atrial fibrillation indicate more overall RRinterval irregularity. Such RR-interval irregularity may induce beat-to-beat changes in refractoriness, promoting intraventricular reentry. We recently demonstrated a larger overall RR-interval irregularity to be an independent predictor of ischemic VF in patients with STEMI.8 The present findings indicate that even in nonischemic conditions, patients with SCA may already be more prone to dispersion of refractoriness. During ischemia, this dispersion may become more pronounced and set the stage for VF. The SCA registry used was characterized by an age limit of 75 years. Therefore, we do not know whether our findings are applicable to elderly patients aged ⬎75 years. The study was aimed at patients with known CAD, while within the registry used for this study, 53% of female and 44% of male victims of SCA were patients without cardiac histories.2 However, studying patients with known CAD who are by definition known to cardiologists offers the unique possibility to assess the electrocardiogram before the event of SCA. The median period between the recording of the assessed electrocardiogram and the event of SCA was relatively short (59 days [interquartile range 29 to 137]). It is therefore not known whether the identified predicting ECG parameters also apply to electrocardiograms recorded longer before SCA. On the basis of the optimal (mathematical) balance between sensitivity and specificity, we identified cut-off values for QRS width to correctly identify most patients with SCA. For patients with previous MIs, the cut-off value for the QRS width is 99 ms, with sensitivity of 76% (95% CI 63% to 86%), specificity of 66% (95% CI 52% to 77%) and a positive predictive value of 68% (95% CI 55% to 79%). For patients without previous MIs, the cut-off value for QRS width is also 99 ms, but with sensitivity of 61% (95% CI 41% to 78%), specificity of 69% (95% CI 56% to 79%), and a positive predictive value of 44% (95% CI 28% to 60%). The combination of a QRS width ⬎99 ms and ⱖ1 echocardiographic parameter could be helpful in identifying those patients at highest risk for SCA. Besides tailoring therapy and the implantation of implantable cardioverterdefibrillators to diminish patients’ risk, patients considered at highest risk for SCA could be eligible for continuous monitoring outside the hospital, as has been advocated by our group and others.10 –12 Such devices will have to be developed incorporating a warning predictor of ischemic VF, such as QRS width or RR-interval irregularity, which could possibly lead to improved early identification of patients at risk.

1281

Table 3 Characteristics of patients with previous myocardial infarctions and left ventricular ejection fractions ⬎40% Variable

Baseline Body mass index (kg/m2) Co-morbidities Hypertension Echocardiography Left ventricular hypertrophy Left ventricular end-systolic diameter (mm) Medications Diuretics Electrocardiography Ventricular ectopic beats QRS width (ms)

Infarction On Electrocardiogram SCA Patients (n ⫽ 12)

Controls (n ⫽ 27)

p Value*

23 (21–27)

27 (24–28)

75%

15%

50% 35 (31–46)

15% 30 (26–34)

0.043 0.004

33%

4%

0.025

42% 100 (95–111)

0% 88 (80–96)

0.001 0.002

0.031 ⬍0.001

Data are expressed as percentages or as median (interquartile range). * Student’s t test, Mann-Whitney U test, or Fisher’s exact test as appropriate.

In patients with previous MIs, class I indications for cardioverter-defibrillator implantation include a left ventricular ejection fraction ⱕ30%, an ejection fraction ⱕ35% with New York Heart Association functional class II or III, and an ejection fraction ⱕ40% but with (non-)sustained VT induced during electrophysiologic study.13 There is currently no recommendation for cardioverterdefibrillator implantation in patients with previous MIs and relatively preserved left ventricular ejection fractions (⬎40%). As listed in Table 3, a subanalysis in patients with SCA with previous MI and ejection fractions ⬎40% showed a broader QRS width and an increased frequency of ventricular ectopic beats. This suggests that these simple ECG variables may help identify patients at risk for SCA. Our study provides information only on patients known with CAD, while, as shown in the registry used for this study, many victims of SCA are patients without cardiac histories.2 It is unknown whether the present findings can be extrapolated to victims of SCA without cardiac histories. Studies of possible victims of SCA without known CAD require a broader prospective population-based study design. 1. Zipes DP, Wellens HJJ. Sudden cardiac death. Circulation 1998;98: 2334 –2351. 2. de Vreede-Swagemakers JJM, Gorgels APM, Dubois-Arbouw WI, van Ree JW, Daemen MJAP, Houben LGE, Wellens HJJ. Out-of-hospital cardiac arrest in the 1990s: a population-based study in the Maastricht area on incidence, characteristics and survival. J Am Coll Cardiol 1997;30:1500 –1505. 3. Lemmert ME, de Jong JS, van Stipdonk AM, Crijns HJ, Wellens HJ, Krucoff MW, Dekker LR, Wilde AA, Gorgels AP. Electrocardiographic factors playing a role in ischemic ventricular fibrillation in ST elevation myocardial infarction are related to the culprit artery. Heart Rhythm 2008;5:71–78. 4. Kleber AG, Janse MJ, Wilms-Schopmann FJ, Wilde AA, Coronel R. Changes in conduction velocity during acute ischemia in ventricular myocardium of the isolated porcine heart. Circulation 1986;73:189 – 198. 5. Uusitupa M, Pyorala K, Raunio H, Rissanen V, Lampainen E. Sensitivity and specificity of Minnesota Code Q-QS abnormalities in the

1282

6. 7.

8.

9. 10. 11.

The American Journal of Cardiology (www.ajconline.org)

diagnosis of myocardial infarction verified at autopsy. Am Heart J 1983;106:753–757. Huikuri HV, Castellanos A, Myerburg RJ. Sudden death due to cardiac arrhythmias. N Engl J Med 2001;345:1473–1482. Moss AJ, Zareba W, Hall WJ, Klein H, Wilber DJ, Cannom DS, Daubert JP, Higgins SL, Brown MW, Andrews ML. Prophylactic implantation of a defibrillator in patients with myocardial infarction and reduced ejection fraction. N Engl J Med 2002;346:877– 883. Lemmert ME, Majidi M, Krucoff MW, Bekkers SC, Crijns HJ, Wellens HJ, Kosinski AS, Gorgels AP. RR-interval irregularity precedes ventricular fibrillation in ST elevation acute myocardial infarction. Heart Rhythm 2010;7:65–71. Kors JA, de Bruyne MC, Hoes AW, van Herpen G, Hofman A, van Bemmel JH, Grobbee DE. T axis as an indicator of risk of cardiac events in elderly people. Lancet 1998;352:601– 605. Wellens HJ, Gorgels AP, de Munter H. Cardiac arrest outside of a hospital: how can we improve results of resuscitation? Circulation 2003;107:1948 –1950. Arzbaecher R, Jenkins J, Burke M, Song Z, Garrett M. Database testing of a subcutaneous monitor with wireless alarm. J Electrocardiol 2006;39:S50 –S53.

12. Fischell TA, Fischell DR, Fischell RE, Virmani R, DeVries JJ, Krucoff MW. Real-time detection and alerting for acute ST-segment elevation myocardial ischemia using an implantable, high-fidelity, intracardiac electrogram monitoring system with long-range telemetry in an ambulatory porcine model. J Am Coll Cardiol 2006;48:2306 –2314. 13. Zipes DP, Camm AJ, Borggrefe M, Buxton AE, Chaitman B, Fromer M, Gregoratos G, Klein G, Moss AJ, Myerburg RJ, Priori SG, Quinones MA, Roden DM, ilka MJ, Tracy C, Priori SG, Blanc JJ, Budaj A, Camm AJ, Dean V, Deckers JW, Despres C, Dickstein K, Lekakis J, McGregor K, Metra M, Morais J, Osterspey A, Tamargo JL, Zamorano JL, Smith SC Jr, Jacobs AK, Adams CD, Antman EM, Anderson JL, Hunt SA, Halperin JL, Nishimura R, Ornato JP, Page RL, Riegel B. ACC/AHA/ESC 2006 guidelines for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death: a report of the American College of Cardiology/American Heart Association Task Force and the European Society of Cardiology Committee for Practice Guidelines (Writing Committee to Develop Guidelines for Management of Patients With Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death) developed in collaboration with the European Heart Rhythm Association and the Heart Rhythm Society. Europace 2006;8:746 – 837.