Myocardial Scar but not Ischemia is Associated with Defibrillator Shocks and Sudden Cardiac Death in Stable Patients with Reduced Left Ventricular Ejection Fraction

Abstract

Background:

Although myocardial scar is a known substrate for ventricular arrhythmias, the association of myocardial ischemia with ventricular arrhythmias in stable patients with left ventricular dysfunction is less clear.

Objectives:

To investigate the association of myocardial scar and ischemia with major arrhythmic events (MAEs) in patients with left ventricular ejection fraction (LVEF) ≤35%.

Methods:

A total of 439 consecutive patients (median age 70 years, 78% male, 55% with implantable cardioverter defibrillator [ICD]) referred for stress/rest positron emission tomography (PET) and resting LVEF ≤35% were included. Primary outcome was time-to-first MAE defined as sudden cardiac death, resuscitated sudden cardiac death, or appropriate ICD shocks for ventricular tachyarrhythmias ascertained by blinded adjudication of hospital records, Social Security Administration’s Death Masterfile, National Death Index, and ICD vendor databases.

Results:

Ninety-one MAEs including 20 sudden cardiac deaths occurred in 75 (17%) patients over a median follow-up of 3.2 years. Transmural myocardial scar was strongly associated with MAEs beyond age, sex, cardiovascular risk factors, beta-blocker therapy, and resting LVEF (adjusted hazard ratio [95% confidence interval] per 10% increase in scar: 1.48[1.22-1.80], p<0.001). However, non-transmural scar/hibernation or markers of myocardial ischemia on PET including global or peri-infarct ischemia, coronary flow reserve, and resting or hyperemic myocardial blood flows were not associated with MAEs in univariable or multivariable analysis. These findings remained robust in subgroup analyses of patients with ICD (n=223), with ischemic cardiomyopathy (n=287), and in patients without revascularization after the PET scan (n=365).

Conclusions:

Myocardial scar but not ischemia was associated with appropriate ICD shocks and sudden cardiac death in patients with LVEF ≤35%. These findings have implications for risk-stratification of patients with left ventricular dysfunction who may benefit from ICD therapy.

CONDENSED ABSTRACT

We investigated the association of myocardial scar and ischemia, assessed using positron emission tomography, with major arrhythmic events (MAEs) - sudden cardiac death and implantable cardioverter defibrillator shocks - in 439 stable patients with left ventricular ejection fraction ≤35% over a median follow-up of 3.2 years. We found that myocardial scar had a strong independent association with MAEs beyond age, sex, cardiovascular risk factors, beta-blocker therapy, and resting left ventricular ejection fraction. However, non-transmural scar/hibernation and multiple highly sensitive markers of myocardial ischemia including global or peri-infarct ischemia, coronary flow reserve, and resting or hyperemic myocardial blood flows were not associated with MAEs.

Graphical Abstract

Myocardial Scar, Ischemia, and Major Arrhythmic Events Myocardial scar (arrow with green check) but not hibernation or ischemia (arrows with red crosses) was associated with appropriate implantable cardioverter defibrillator shocks or sudden cardiac death in stable patients with LVEF ≤ 35%.

INTRODUCTION

Arrhythmic cardiovascular death and pump failure are the two leading causes of death in patients with heart failure and reduced left ventricular ejection fraction (LVEF) (1,2). Current guidelines for the primary prevention of sudden cardiac death in patients with chronic left ventricular (LV) dysfunction recommend implantable cardioverter-defibrillator (ICD) therapy based largely on the presence of depressed left ventricular function, typically LVEF ≤ 35%, consistent with the inclusion criteria from large clinical trials (3-5). However, only a minority of these patients receive appropriate ICD shocks. In the SCD-HeFT (Sudden Cardiac Death in Heart Failure) trial of primary prevention ICD, only 21% of patients received appropriate ICD shocks for ventricular tachyarrhythmias over 5 years of follow-up (5). Further, ICD implantation is associated with significant morbidity (6). Therefore, identification of patients with LV dysfunction who are most or least likely to benefit from the ICD therapy can reduce the associated morbidity from it and improve its cost-effectiveness (7). Myocardial scar and ischemia are routinely assessed in many of these patients and could potentially serve as markers for risk of ventricular tachyarrhythmias to enhance risk-stratification beyond LVEF. Acute myocardial ischemia as well as myocardial scar, especially tissue heterogeneity in peri-infarct zone, are known to be associated with ventricular arrhythmias (8-10). However, the association of myocardial ischemia with ventricular arrhythmias and arrhythmic death in patients without acute coronary syndrome is less clear.

In this study, we aimed to evaluate the independent associations between the myocardial scar as well as comprehensively assessed myocardial ischemia - including global and peri-infarct ischemia, coronary flow reserve, hyperemic and resting myocardial blood flows - with ventricular tachyarrhythmias and sudden cardiac death in stable patients with LVEF ≤ 35%.

METHODS

Study Population

The study population was comprised of consecutive patients referred for rest/stress cardiac positron emission tomographic (PET) scan at Brigham & Women’s Hospital, Boston, MA between January 1, 2006 and December 31, 2013 with resting LVEF ≤ 35%. Patients with prior heart transplantation, and those whose images were missing or uninterpretable owing to poor image quality were excluded. If a patient had repeat PET evaluations during the study period, the earliest evaluable scan was included. A total of 439 patients met the study criteria of which 223 (51%) had ICD. A total of 227 patients (52%) were referred for the rest/stress PET for evaluation of symptoms of chest pain (n=57), dyspnea (n=139), or both (n=31). Other reasons for the scan included assessment of myocardial viability, pre-operative evaluation, or pre-transplant evaluation. Demographic variables, cardiovascular risk factors, and medication use were ascertained at time of the study by patient interview and review of medical records. The Partners Healthcare Institutional Review Board approved the study with waiver of informed consent. The study was conducted in accordance with the institutional guidelines.

PET Imaging

Patients were imaged using positron emission tomography (PET)–computed tomography scanner (Discovery RX or STE LightSpeed 64, GE Healthcare, Milwaukee, WI). Resting and hyperemic myocardial perfusion was assessed with rubidium-82 (1480–2200 MBq) or N-13 ammonia (700–900 MBq) as flow tracer, as described previously (11,12), along with a standard intravenous infusion of dipyridamole, adenosine, regadenoson, or dobutamine as stress agent. Rest LVEF was calculated from gated myocardial perfusion images with commercially available software (Corridor4DM; Ann Arbor, MI).

Assessment of Myocardial Scar and Ischemia

Using Cedars-Sinai QPS software, semi-automated quantitative assessment of the myocardial perfusion images was performed using the standardized 17-segment model (13,14). A standard 5-point scoring system (0-4) was used to assign scores to each segment based on the percentage reduction in counts. Summed rest and stress scores for the entire left ventricular myocardium were calculated as the sum of individual segmental scores on the respective rest and stress images. The difference between summed stress and rest scores was recorded as summed difference score. Summed resting score corresponds to a fixed perfusion defect and summed difference score corresponds to a reversible perfusion defect. As there are 17 segments and scores in each segment could range from 0-4, the minimum score possible is 0 and maximum score possible is 68. These scores were converted to percent myocardium using the formula: (Actual score/Maximum possible score of 68)*100.

Myocardial scar was defined as severe fixed perfusion defect with more than 50% reduction in counts. This definition of myocardial scar is supported by an extensive literature that compared myocardial perfusion imaging with pathology (15-18), metabolic imaging (19,20), cardiac MRI (21,22), and with functional recovery after revascularization (17,18,23). Segments with ≤ 50% reduction in counts were assumed to contain an admixture of non-transmural scar and/or hibernating myocardium, and labeled as non-transmural scar/hibernation. Global ischemia was defined as reversible perfusion defect. Peri-infarct ischemia was defined as the presence of at least 5% of reversible perfusion defect in a coronary vascular territory that also had at least 5% of fixed perfusion defect. Hyperemic and resting myocardial blood flows (in ml ∙ g⁻¹ ∙ min⁻¹) were computed from the dynamic stress and rest imaging series respectively, using compartmental tracer kinetic modeling with commercially available software (Corridor4DM; Ann Arbor, MI) (11,12,24). Coronary flow reserve for each patient was calculated as the ratio of hyperemic to resting flows. Coronary flow reserve is a highly sensitive marker of ischemia that provides integrated assessment of both epicardial and microvascular coronary circulatory function, and impaired coronary flow reserve has been shown to be a strong predictor of cardiovascular mortality in patients with known or suspected stable coronary artery disease (25).

Outcome Assessment

The primary outcome of interest was time to first major arrhythmic event (MAE). MAEs comprised of sudden cardiac death, resuscitated sudden cardiac death, and/or appropriate ICD shocks for ventricular tachyarrhythmias. The vital status of all study patients was ascertained by integrating data from the Social Security Administration’s Death Master File, the National Death Index, and the Partners Healthcare Research Patient Data Registry from January 1, 2006 to December 31, 2015. Two independent reviewers blindly adjudicated electronic medical records, any available autopsy certificates, and death certificates to determine the cause of death. In case of disagreement on the cause of death, consensus adjudication was done. Sudden cardiac death was defined as death that occurred unexpectedly and not within 30 days of acute myocardial infarction. The various scenarios included in the definition of sudden cardiac death, as adapted from the 2014 American College of Cardiology/American Heart Association consensus definitions for cardiovascular endpoints (26), are detailed in the Supplementary Appendix. All deaths that were not adjudicated as sudden cardiac death were considered non-arrhythmic deaths and were censored. Censoring was done at the time of first MAE, death, or December 31, 2015 whichever occurred earlier.

ICD shocks were blindly adjudicated using electronic medical records and intracardiac electrocardiograms from ICD vendor databases. ICD shocks were considered appropriate for ventricular tachyarrhythmias at threshold rates programmed clinically by cardiac electrophysiology experts. We did not have information on ventricular arrhythmic events terminated by anti-tachycardia pacing. All study patients with ICD had ICD placed for primary prevention except for 3 patients in whom the ICD was placed after the initial event that led to PET scan.

Statistical Analysis

Univariable and multivariable Cox proportional hazards models were used to assess the association of myocardial scar and ischemia with MAEs. Each model contained only one exposure of interest. The ties in failure times were handled using Efron’s approximation. Inference testing was based on Wald Chi-square statistic. Proportional hazards assumption was tested by inclusion of a time-varying covariate term and was found to be valid. In multivariable Cox models, adjustment was made for age, sex, diabetes, end-stage renal disease on dialysis, ischemic cardiomyopathy, beta-blocker therapy at baseline, revascularization post-PET, and rest LVEF. The covariate selection was based on clinical knowledge. Revascularization post-PET scan was ascertained from the Partners Healthcare Research Patient Data Registry, electronic medical records and billing claims. Ischemic cardiomyopathy (n=287) was defined as presence of scar or ischemia burden of at least 5% of the left ventricular myocardium in the setting of obstructive coronary artery disease (history of myocardial infarction, prior percutaneous coronary intervention, prior coronary artery bypass grafting, or coronary revascularization within 90 days after the PET scan). Poisson regression was used to obtain annualized event rate for MAEs by tertiles of myocardial scar.

The following sub-group analysis were conducted: 1) in patients with ICD (n=223); 2) in patients with ischemic cardiomyopathy (n=287).

The following sensitivity analyses were conducted: 1) in all study patients without revascularization post-PET (n=365), 2) in patients with ischemic cardiomyopathy without revascularization post-PET (n=215); 3) after excluding events at time 0 that triggered the PET scan (n-excluded = 38); 4) after accounting for non-arrhythmic death as competing risk for MAEs using Fine and Gray competing risk model (27).

All statistical analyses were performed with SAS 9.4 (SAS Institute Inc, Cary, NC). A 2-sided p-value < 0.05 was considered statistically significant.

RESULTS

Patient and Imaging Characteristics

Table 1 shows baseline patient and imaging characteristics in all study patients (n = 439) as well as in patients with ICD (n=223, 51%), stratified by whether MAEs occurred during follow-up. Median age of the study population was 70 years with majority being men (78%). Ischemic cardiomyopathy was present in 287 (65%) patients. Median (interquartile range, IQR) resting LVEF was 28% (IQR, 23%-32%). Median amounts of myocardial scar, non-transmural scar/hibernation, and myocardial ischemia were 4.4%, 11.8%, and 7.4% respectively. Median CFR was severely depressed (1.43), and median hyperemic and resting myocardial blood flows were reduced at 1.12 and 0.78 ml ∙ g⁻¹∙ min⁻¹, respectively. These baseline characteristics were similar in the subset of patients with ICD (Table 1).

Table 1.

Baseline Patient and Imaging Characteristics

	All Patients				Patients with ICD
Variable	Total (n = 439)	Patients with MAE (n = 75)	Patients without MAE (n = 364)	p-value*	Total (n = 223)	Patients with MAE (n = 67)	Patients without MAE (n = 156)	p-value†
Demographics
Age, y	70(62-78)	72(65-78)	69(61-78)	0.16	70(62-77)	72(65-78)	69(61-77)	0.17
Males	344 (78.4)	62 (82.7)	282 (77.5)	0.32	186 (83.4)	56 (83.6)	130 (83.3)	0.96
Race				0.04				0.13
White	302 (68.8)	61 (81.3)	241 (66.2)		167 (74.9)	56 (83.6)	111 (71.2)
Black	62 (14.1)	7 (9.3)	55 (15.1)		26 (11.7)	6 (9.0)	20 (12.8)
Cardiovascular Risk Factors/History
Diabetes	195 (44.4)	30 (40.0)	165 (45.3)	0.40	89 (39.9)	26 (38.8)	63 (40.4)	0.83
Dialysis	24 (5.5)	3 (4.0)	21 (5.8)	0.54	5 (2.2)	2 (3.0)	3 (1.9)	0.62
Ischemic Cardiomyopathy	287 (65.4)	62 (82.7)	225 (61.8)	<0.001	169 (75.8)	57 (85.1)	112 (71.8)	0.03
Prior PCI	129 (29.4)	28 (37.3)	101 (27.7)	0.10	80 (35.9)	24 (35.8)	56 (35.9)	0.99
Prior CABG	133 (30.3)	38 (50.7)	95 (26.1)	<0.001	87 (39.0)	36 (53.7)	51 (32.7)	0.003
Revascularization Post-PET	74 (16.9)	6 (8.0)	68 (18.7)	0.02	29 (13.0)	5 (7.5)	24 (15.4)	0.11
Early Revascularization‡	63 (14.4)	6 (8.0)	57 (15.7)	0.09	20 (9.0)	5 (7.5)	15 (9.6)	0.61
ACE Inhibitors	237 (54.0)	37 (49.3)	200 (54.9)	0.37	131 (58.7)	33 (49.3)	98 (62.8)	0.06
Beta-blockers Therapy	354 (80.6)	62 (82.7)	292 (80.2)	0.63	186 (83.4)	55 (82.1)	131 (84.0)	0.73
Digoxin	70 (15.9)	16 (21.3)	54 (14.8)	0.16	51 (22.9)	16 (23.9)	35 (22.4)	0.81
Diuretics	267 (60.8)	50 (66.7)	217 (59.6)	0.25	159 (71.3)	48 (71.6)	111 (71.2)	0.94
Nitrates	92 (21.0)	12 (16.0)	80 (22.0)	0.25	44 (19.7)	10 (14.9)	34 (21.8)	0.24
Imaging Parameters
Rest LVEF, %	28 (23-32)	26 (22-32)	28 (23-32)	0.13	26 (22-32)	26 (22-30)	26 (22-32)	0.92
Scar (Transmural), %	4.4 (0-13.2)	8.8 (4.4-19.1)	0 (0-13.2)	<0.001	4.4 (0-17.6)	8.8 (4.4-22.1)	4.4 (0-14.0)	0.001
Non-Transmural Scar/Hibernation, %	11.8 (8.8-16.2)	13.2 (8.8-17.6)	11.8 (7.4-16.2)	0.07	13.2 (8.8-16.2)	13.2 (8.8-17.6)	13.2 (8.8-16.2)	0.67
Ischemia, %	7.4 (2.9-11.8)	7.4 (4.4-10.3)	7.4 (2.9-11.8)	0.58	7.4 (4.4-11.8)	7.4 (4.4-10.3)	6.6 (3.7-11.8)	0.71
Peri-Infarct Ischemia	120 (27.3)	20 (26.7)	100 (27.5)	0.89	55 (24.7)	18 (26.9)	37 (23.7)	0.62
CFR	1.43 (1.13-1.82)	1.48 (1.19-1.82)	1.42 (1.11-1.82)	0.50	1.47 (1.17-1.88)	1.56 (1.22-1.88)	1.42 (1.14-1.88)	0.42
Hyperemic MBF, ml ∙ g−1∙ min−1	1.12 (0.87-1.46)	1.07 (0.86-1.30)	1.14 (0.87-1.51)	0.23	1.12 (0.89-1.43)	1.10 (0.90-1.33)	1.14 (0.88-1.45)	0.69
Resting MBF, ml ∙ g−1∙ min−1	0.78 (0.63-0.96)	0.74 (0.58-0.87)	0.79 (0.63-0.98)	0.06	0.75 (0.61-0.96)	0.74 (0.59-0.91)	0.75 (0.61-0.96)	0.41

Continuous variables are presented as median [25^th–75^th percentile]; Categorical variables are presented as n (%).

^*p-value for comparison between patients with and without MAE in all Study Patients

^†p-value for comparison between patients with and without MAE in patients with ICD.

^‡Early revascularization was considered as revascularization within 90 days of positron emission tomography scan.

Abbreviations: ACE, angiotensin converting enzyme; CABG, coronary artery bypass grafting; CFR, coronary flow reserve; LVEF, left ventricular ejection fraction; MAE, major arrhythmic event; MBF, myocardial blood flow; PCI, percutaneous coronary intervention

A total of 74 (17%) of the patients underwent revascularization after the PET scan. Twenty-nine of these revascularizations occurred in patients with ICD (Table 1). After excluding patients who had revascularization after the PET scan, the remaining patients (n=365) had a median myocardial ischemia of 5.9%, similar to the overall study population (Online Table 1). ICD was present in 56% (194/365) of these patients.

Outcomes

Over a median follow-up of 3.2 years (IQR, 1.0-5.8), there were 91 MAEs in 75 (17.1%) patients (Table 2). Annualized rate of MAEs was 4.8% per year (95% confidence interval, CI: 3.9-6.1%). A total of 20 patients (4.6%) experienced sudden cardiac death, of which 12 sudden cardiac deaths occurred in patients with ICD. Thirty-three appropriate ICD shocks occurred on follow-up after the PET scan. Thirty-eight patients were referred for PET scan due to resuscitated SCD or appropriate ICD shocks. The distribution of rest LVEF, non-transmural scar/hibernation, and markers of myocardial ischemia - global ischemia, peri-infarct ischemia, hyperemic and resting myocardial blood flows, and coronary flow reserve - were similar between the two groups with or without MAEs (Table 1). However, patients with MAEs had a greater burden of myocardial scar than those without MAEs (median myocardial scar: 9% vs 0% in patients with vs without MAEs).

Table 2.

Major Arrhythmic Events

	All Patients (n=439)	Patients with ICD (n=223)	Patients with Ischemic Cardiomyopathy (n=287)
Number of patients with MAEs	75 (17.1)	67 (30.0)	62 (21.6)
Number of MAEs	91	83	77
Type of MAEs
Sudden Cardiac Death	20	12	15
Appropriate ICD Shocks	33	33	29
Resuscitated SCD or ICD Shocks  that triggered the PET scan	38	38	33

Abbreviations: ICD, implantable cardioverter defibrillator; MAE, major arrhythmic event; PET, positron emission tomography; SCD, sudden cardiac death

Prognostic Value of Myocardial Ischemia and Scar

The amount of non-transmural scar/hibernation, global LV ischemia, presence of peri-infarct ischemia, coronary flow reserve, hyperemic or resting myocardial blood flows were not associated with MAEs in univariable or multivariable analysis (Graphical Abstract and Figure 1). In contrast, myocardial scar was strongly associated with the risk of MAEs (unadjusted hazard ratio, HR [95% CI] for every 10% increase in myocardial scar: 1.52 [1.27-1.81], p<0.001), Figure 1. Annualized event rates by tertiles of myocardial scar showed that the rate of MAE was significantly lower when there was no scar compared with presence of scar - annual MAE rates were 2.1% (95% CI: 1.3-3.3%) in the lower tertile (0% scar), 8.1% (95% CI: 5.5-12.0%) in the middle tertile (0-9% scar), and 8.4% (95% CI: 5.9-11.9%) in the upper tertile (>9% scar), Figure 2.

Figure 1. Association of Myocardial Scar and Ischemia with Major Arrhythmic Events in All Study Patients.

Figure 1A shows univariable analyses. Figure 1B shows multivariable analyses after adjustment for age, sex, diabetes, end-stage renal disease on dialysis, ischemic cardiomyopathy, beta-blocker therapy, left ventricular ejection fraction, and revascularization post-positron emission tomography scan. Hazard Ratios are presented for every 10% increase in myocardial scar, non-transmural scar/hibernation, or global ischemia, for present versus absent of peri-infarct ischemia, and for every 0.25 unit decrease in coronary flow reserve, hyperemic or resting myocardial blod flows.

Myocardial scar was defined as severe fixed perfusion defect with > 50% reduction in counts. Non-transmural scar/hibernation was defined as mild fixed perfusion defect with up to 50% reduction in counts. Relative ischemia was measured as percentage of left ventricular myocardium with reversible perfusion defect. Peri-infarct ischemia was defined as the presence of at least 5% ischemia in a coronary vascular territory that also had at least 5% fixed defect. Coronary flow reserve was defined as ratio of hyperemic and resting myocardial blood flows which were measured in ml ∙ g⁻¹∙ min⁻¹. Abbreviations: CI, confidence interval; HR, hazard ratio

Figure 2. Annualized Major Arrhythmic Event Rates by Tertiles of Myocardial Scar.

Annualized event rates by tertiles of myocardial scar showing the rate of major arrhythmic events was significantly lower when the scar was absent compared with presence of scar. Abbreviation: MAE, major arrhythmic events

In multivariable analysis, after adjustment for age, sex, diabetes, end-stage renal disease on dialysis, ischemic cardiomyopathy, beta-blocker therapy at baseline, revascularization post-PET, and rest LVEF, myocardial scar remained strongly associated with MAEs (adjusted HR for every 10% increase in myocardial scar: 1.48 [95% CI: 1.22-1.80], p<0.001), Figure 1. There was no significant interaction between scar and ischemia for association with MAEs (p-value for interaction = 0.13).

Subgroup and Sensitivity Analyses

The lack of association of non-transmural scar/hibernation and myocardial ischemia with MAEs and strong independent association of myocardial scar with MAEs remained unchanged when analyzed in the subset of patients with ICD (Figure 3) as well as in the subset of patients with ischemic cardiomyopathy (Figure 4). As the amount of myocardial ischemia on PET scan may be modified by subsequent revascularization, we also performed a sensitivity analysis in the subset of patients without post-PET revascularization. This subset of patients had ischemia burden (median, 5.9%) similar to the overall study population (median, 7.4%) (Online Table 1). A total of 83 MAEs occurred in 69 patients (18.9%) out of 365 patients without revascularization after the PET scan. However, even in this subset of patients without post-PET revascularization, none of the markers of myocardial ischemia were associated with MAEs (Online Table 2). The findings were similar in patients with ischemic cardiomyopathy without revascularization post-PET (Online Table 3). We also performed an analysis after excluding events at time 0 where PET was done because of the appropriate ICD shocks. A total of 38 events were excluded with remaining 53 events occurring in 49 patients over a median follow-up of 3.8 years. In this analysis as well, myocardial scar but not ischemia remained associated with MAEs (Online Table 4). Finally, a sensitivity analysis accounting for the competing risk of non-arrhythmic death showed robustness of our findings (Online Table 5).

Figure 3. Association of Myocardial Scar and Ischemia with Major Arrhythmic Events in Patients with ICD.

Figure 3A shows univariable analyses. Figure 3B shows multivariable analyses after adjustment for age, sex, diabetes, end-stage renal disease on dialysis, ischemic cardiomyopathy, beta-blocker therapy, left ventricular ejection fraction, and revascularization post-positron emission tomography scan. Hazard Ratios are presented for every 10% increase in myocardial scar, non-transmural scar/hibernation, or global ischemia, for present versus absent of peri-infarct ischemia, and for every 0.25 unit decrease in coronary flow reserve, hyperemic or resting myocardial blood flows. Abbreviations: CI, confidence interval; HR, hazard ratio; ICD, implantable cardioverter defibrillator

Figure 4. Association of Myocardial Scar and Ischemia with Major Arrhythmic Events in Patients with Ischemic Cardiomyopathy.

Figure 4A shows univariable analyses. Figure 4B shows multivariable analyses after adjustment for age, sex, diabetes, end-stage renal disease on dialysis, beta-blocker therapy, left ventricular ejection fraction, and revascularization post-positron emission tomography scan. Hazard Ratios are presented for every 10% increase in myocardial scar, non-transmural scar/hibernation, or global ischemia, for present versus absent of peri-infarct ischemia, and for every 0.25 unit decrease in coronary flow reserve, hyperemic or resting myocardial blood flows. Abbreviations: CI, confidence interval; HR, hazard ratio.

DISCUSSION

Acute myocardial ischemia is a known trigger for ventricular arrhythmias (10,28,29); however, whether chronic myocardial ischemia increases propensity for ventricular tachyarrhythmias and sudden cardiac death in patients with stable coronary disease with LV dysfunction is less clear. In our study, we investigated the association of comprehensively assessed myocardial ischemia with MAEs in patients with severely reduced LVEF. In a cohort of 439 patients with LVEF ≤ 35% followed over a median period of 3.2 years, we showed that neither global or peri-infarct ischemia, absolute coronary blood flows, or coronary flow reserve was associated with appropriate ICD shocks for ventricular tachyarrhythmias and/or sudden cardiac death. We further confirmed the role of myocardial scar as a powerful non-invasive imaging biomarker for assessing the risk of MAEs in patients with LVEF ≤ 35% (30,31).

Our study contradicts the conclusions from previous small studies examining the association of myocardial ischemia with inducible ventricular arrhythmias. In a study of 30 patients with LVEF ≤ 35%, Rijnierse et al. found univariable association of impaired hyperemic myocardial blood flow and impaired coronary flow reserve with inducible ventricular arrhythmias on programmed electrical stimulation (32). This study, however, did not account for any confounders due to a small sample size. In another study of 90 patients with history of myocardial infarction, Paganelli et al. showed increased prevalence of residual peri-infarct ischemia in patients with inducible ventricular tachyarrhythmias compared with those without inducibility (33). In contrast to these findings, we found no association between hyperemic myocardial blood flow, coronary flow reserve, or peri-infarct ischemia with MAEs on clinical ventricular tachyarrhythmia endpoints. A post hoc analysis of MADIT II study (Multicenter Automatic Defibrillator Implantation Trial) has shown that the electrophysiological testing has limited predictive value for future ventricular tachycardia or ventricular fibrillation events (34). A recent study involving 110 patients with LVEF < 35% found a marginally significant association between hyperemic myocardial blood flow, but not other markers of myocardial ischemia, and ventricular tachycardia or fibrillation (35). However, confidence intervals were wide and risk-adjustment was extremely limited and no analysis was presented after excluding patients with revascularization after the PET scan (35).

There are some potential explanations for a lack of association of myocardial ischemia with MAEs. First, in patients with LVEF ≤ 35%, myocardial ischemia has been shown to be an infrequent cause of arrhythmic sudden cardiac death (36). In a study of patients with LVEF ≤ 35% and ICD implantation after myocardial infarction, ischemia defined by ST-segment change of ≥ 0.1 mV prior to arrhythmia, occurred in only 15.4% of patients prior to ventricular tachycardia or ventricular fibrillation over a 6-year follow-up period (36). Second, myocardial scar may be a major driver of MAEs in patients with LVEF ≤ 35%; whereas in patients with low burden of scar and relatively preserved LVEF, ischemia may make play a greater role as a trigger of ventricular arrhythmias (37). However, even in this patient population, left ventricular scar is an important determinant of risk. In a large series of patients with relatively preserved left ventricular function (82% of study patients with LVEF ≥ 40%), and angiographically documented coronary artery disease, the summed stress score (reflecting combined myocardial scar and ischemia) on nuclear myocardial perfusion imaging but not summed difference score (reflecting myocardial ischemia) was significantly associated with sudden cardiac death (38).

Our study also adds to the accumulating evidence in support of the prognostic significance of myocardial scar, beyond LVEF, for sudden cardiac death risk even after accounting for impact of competing causes of death. Patients with LV dysfunction are at increased risk of death due to pump failure and sudden cardiac death (39). Therefore, it is important to account for the competing risk of pump failure deaths when evaluating markers of sudden cardiac death in this patient population (40). In the present study, we showed that myocardial scar was strongly associated with MAEs even after accounting for competing risk of non-arrhythmic death. Randomized clinical trials are needed to prospectively investigate the role of substrate-based risk-stratification combining amount of myocardial scar with LVEF to ascertain the group of patients most likely to benefit from ICD therapy. As structural and functional changes in cardiac autonomic innervation have been shown to be associated with sudden cardiac death (41), their interaction with myocardial scar and ischemia in stable patients merits further investigation.

Limitations

Our study is a single-center observational study with its inherent limitations. We adjusted the analyses for many risk factors for MAEs but there is likely residual and unmeasured confounding. The study population is heterogeneous that includes patients with or without ICD as well as patients with ischemic or non-ischemic cardiomyopathy. However, this allows for increased generalizability of our study results. Further, sub-group analyses in patients with ICD and in patients with ischemic cardiomyopathy showed robustness of our study findings. Although there is an extensive literature validating myocardial scarring in areas with severe fixed perfusion defect (42), however, we could not separate non-transmural scar from hibernating myocardium in areas with mild to moderate fixed perfusion defect. The viability assessment using 18F-fluorodeoxyglucose could not be used for this study as it was not obtained in majority of our study patients. Coronary revascularization after the PET scan may modify the ischemic burden, mitigating its association with MAEs. However, in the sensitivity analysis after excluding patients with post-PET revascularization, myocardial ischemia was not associated with MAEs despite similar amount of ischemia in the remaining cohort. Improvement in LVEF due to medical therapy, revascularization or cardiac resynchronization therapy could lead to improvement in prognosis and may modify the association of myocardial scar and ischemia with outcomes. As this was a clinical cohort, there was no systematic follow-up PET scan to identify change in LVEF and hence, this effect could not be quantified and adjusted for. In addition, we were unable to adjust for antiarrhythmic therapy at baseline, and/or changes in medications and ablative therapies during follow-up. However, as anti-arrhythmic and ablative therapies usually follow the arrhythmic event, this is unlikely to affect the analysis of time to first MAE. Given the retrospective nature of our study, ICD programming parameters lacked uniformity across study patients. Lastly, the adjudication of sudden cardiac death is inherently challenging in a retrospective clinical cohort. However, this is a closely followed clinical cohort in a tertiary care academic medical center with detailed electronic medical records. Further, sensitivity analysis in the subset of patients who had ICD (and hence, more definitive adjudication of MAEs) showed robustness of our findings.

CONCLUSION

In conclusion, myocardial scar but not ischemia on PET imaging was associated with appropriate ICD shocks and sudden cardiac death in patients with LVEF ≤ 35%. The role of myocardial scar in risk-stratification of patients with LV dysfunction who may benefit from primary prevention ICD needs investigation in randomized trials.

PERSPECTIVES

Clinical Competencies:

The presence of myocardial scar has strong independent association with appropriate implantable cardioverter defibrillator shocks and sudden cardiac death in patients with left ventricular ejection fraction ≤ 35%. Although acute myocardial ischemia is a known trigger for ventricular arrhythmias, markers of myocardial ischemia in patients with stable coronary artery disease is not associated with clinical ventricular tachyarrhythmia endpoints.

Translational Outlook:

As only a minority of patients with severe left ventricular dysfunction and primary prevention implantable cardioverter defibrillator receive appropriate shock therapy, there is unmet need for better risk-stratification for arrhythmic death in these patients. The role of myocardial scar in risk-stratification of patients with left ventricular dysfunction who may benefit from primary prevention ICD needs investigation in randomized trials.

ABBREVIATIONS

ICD

implantable cardioverter defibrillator

LV

left ventricular

LVEF

left ventricular ejection fraction

MAE

major arrhythmic events

PET

positron emission tomography