Evaluation of Need for Implantable Cardioverter-Defibrillator by Thallium-201 Scintigraphy Among Japanese Patients With Prior Myocardial Infarction

Abstract

Background: Identifying who among current Japanese patients with prior myocardial infarction (MI) would benefit from an implantable cardioverter-defibrillator (ICD) is imperative. Accordingly, this study seeks to determine whether single-photon emission computed tomography (SPECT) can help identify such patients.

Methods and Results: This retrospective study enrolled 60 consecutive patients with prior MI who underwent stress thallium-201 SPECT and ICD implantation from February 2000 to October 2014. Occurrence of arrhythmic death and/or or appropriate ICD therapy, defined as shock or antitachycardia pacing for ventricular fibrillation or tachycardia, was identified until November 2016. During the median follow-up interval of 6.6 years, 18 (30%) patients experienced arrhythmic death and/or appropriate ICD therapy. Multivariate Cox proportional hazard regression analysis revealed that the summed stress score (SSS) [hazard ratio (HR)=1.14; P=0.005] and left ventricular ejection fraction (LVEF) at rest (HR=0.92; P=0.038) were significantly associated with the occurrence of arrhythmic events. Patients with SSS ≥21 and LVEF ≤30%, which were determined to be the best cutoff points, had significantly higher incidence of the arrhythmic events than the other patients (64% vs. 11%; HR=7.18; log-rank P=0.001).

Conclusions: SSS using stress thallium-201 SPECT in combination with LVEF can help determine the need for ICD therapy among current Japanese patients with prior MI.

The Multicenter Automatic Defibrillator Implantation Trial II (MADIT-II) proved that the prophylactic use of an implantable cardioverter-defibrillator (ICD) was associated with significantly improved survival in patients with prior myocardial infarction (MI) who had severe left ventricular (LV) dysfunction.¹ Based on this trial, LV ejection fraction (LVEF) became a criterion for prophylactic ICD implantation,² but the broad scope of this criterion places considerable financial burden on the medical economy.³

Recently, optimal medical therapy and revascularization procedures have greatly reduced post-MI cardiovascular events.⁴ Several studies in Japan have demonstrated a better prognosis among patients with prior MI than many other large clinical trials performed in Western countries.^5–10 Considering the high event-free survival rates reported in Japanese trials, applying the MADIT-II criteria to Japanese patients may be inappropriate.⁵ Hence, identifying those who would actually benefit from ICD implantation among current Japanese patients with prior MI is imperative.

Previous studies revealed that infarct size determined using single-photon emission computed tomography (SPECT) was not only associated with overall and cardiac mortality in patients with prior MI,^10–14 but also a predictor of life-threatening ventricular arrhythmias.^14–16 However, many of the studies of life-threatening ventricular arrhythmias were performed a decade ago when primary prevention was just established and the majority of the patients enrolled were from Western countries. Therefore, the present study sought to determine whether the extent of myocardial scarring and ischemia evaluated using stress thallium-201 SPECT can help predict the need for ICD therapy among current Japanese patients with prior MI.

Methods

Study Patients and Design

From February 2000 to October 2014, 95 patients with prior MI underwent transvenous ICD implantation at the Osaka Police Hospital, Osaka, Japan. Among them, 62 patients underwent stress thallium-201 SPECT; 2 patients who underwent coronary revascularization based on SPECT results were excluded, and the remaining 60 patients were enrolled in this study.

All patients were followed up at the outpatient clinic every 2–4 months until November 2016. Clinical evaluation and device testing were performed at each visit. Patients were questioned regarding the occurrence of prespecified symptoms including syncope, presyncope, palpitations, and lightheadedness. Additionally, device memory was interrogated for delivered therapy, and device programming adjustments were performed by trained electrophysiologists. Physicians were encouraged to follow the guidelines for revascularization and pharmacologic therapies.¹⁷ The endpoint of the present study was arrhythmic death and/or appropriate ICD therapy, including appropriate ICD shock therapy, for ventricular fibrillation and appropriate ICD shock therapy or antitachycardia pacing for ventricular tachycardia.

First, death, causes of death, and the incidence of arrhythmic death and of appropriate ICD therapy were determined. Patients’ characteristics and SPECT data were then compared between those who did and did not experience arrhythmic events. Second, multivariate Cox proportional hazards regression analysis was performed to identify the parameters associated with the occurrence of arrhythmic events. Finally, the best cutoff values for such parameters, as well as high-risk patients who need the appropriate ICD therapy, were determined. This study was approved by the Osaka Police Hospital Ethics Committee.

Device Implantation and Programming

Indications for ICD therapy were based on Japanese Circulation Society guidelines.² ICD implantation criteria for primary prevention included New York Heart Association (NYHA) class II or III heart failure, LVEF ≤35%, and detection of nonsustained ventricular tachycardia or induction of sustained ventricular tachycardia in an electrophysiology study. Criteria for secondary prevention included cardiac arrest survival and detection of sustained ventricular tachycardia. Commercially available transvenous ICD devices were implanted in the pectoral region in all patients.

Cardiac resynchronization therapy defibrillator (CRT-D) implantation was recommended for patients with NYHA class III or ambulatory class IV heart failure, LVEF ≤35%, and QRS duration ≥120 ms.^1,2 The following protocol was encouraged: ventricular arrhythmia >176 beats/min was initially treated using 2 bursts of antitachycardia pacing followed by shock therapy if unsuccessful and ventricular arrhythmia >222 beats/min was initially treated using high-output (i.e., 30–35 J) shock therapy.

Stress Thallium-201 SPECT

ECG-gated stress/rest myocardial perfusion imaging using thallium-201 and exercise or pharmacological stress was performed. Exercise thallium-201 myocardial perfusion SPECT was performed using a symptom-limited maximum exercise test with a graded bicycle ergometer protocol. At near maximum exercise, 111 MBq of thallium-201 was injected followed by exercise continuation for 1 min. Exercise was terminated if (1) the patient’s heart rate reached >85% of the maximum predictive heart rate (220−age), (2) the patient experienced chest discomfort, or (3) significant ST segment depression or elevation was observed.

Pharmacological (adenosine 120 µg/kg/min for 6 min) stress test was used for patients unable to perform the exercise stress test. After adenosine infusion, 111 MBq of thallium-201 was injected intravenously.

ECG-gated SPECT images were obtained 10 min after the stress test using a dual detector system (The Infinity; GE Healthcare Japan, Tokyo, Japan) with low-energy, high-resolution collimators. For computer acquisitions, a 64×64 matrix was used. The images were acquired in 64 projections with a 180° circular orbit from 45° right anterior oblique to 45° left posterior oblique using the step-and-shoot technique and 40 s of imaging per frame in a supine position. Redistribution imaging was performed 4 h later using the same acquisition measurements.

The gated SPECT image set was reconstructed on a dedicated workstation (Xeleris; GE Healthcare Japan), with X-ray-based attenuation correction into short-axis, vertical long-axis, and horizontal long-axis slices encompassing the entire left ventricle. Polar maps of perfusion were automatically produced using a commercially available software package (Ceders QPS; Ceders-Sinai Medical Center, Los Angeles, CA, USA). Myocardial perfusion SPECT was analyzed on a standard 17-segment model. The mean percentage uptake in each segment was automatically scored from normal (0) to absent (4) using a 5-point scale based on the guidelines of the American Society of Nuclear Cardiology.¹⁸ The values of the summed stress score (SSS), summed rest score (SRS), and summed difference score (SDS) on stress-rest thallium-201 gated SPECT were calculated using the QPS software with a normal database developed for Japanese patients¹⁹ (Figure 1). Reverse redistribution in a rest thallium-201 SPECT image was not reflected in the SDS, and it was considered to be 0. LV end-diastolic volume, LV endsystolic volume, and LVEF were also calculated using quantitative gated SPECT (QGS) software.

Figure 1.

Representative thallium-201 SPECT image of a patient with prior anterior septal myocardial infarction. (A) Stress and rest myocardial perfusion thallium-201 SPECT images in short-axis, vertical long-axis, and horizontal long-axis views. The French scale without white color (red-yellow-green-blue-black) was adopted. (B, Upper and Middle) Stress and rest reconstructed polar maps of perfusion, respectively. (Lower) Reversible perfusion image, which was developed using the stress and rest perfusion images and indicates myocardial viability in each region. In this case, no viable myocardium can be observed. (C) Diagrams showing SSS, SRS, and SDS from the same SPECT images. Based on myocardial segmentation, the semi-quantitative method for evaluating perfusion was utilized. Perfusion in each segment was scored using a 5-point system. SDS, summed difference score; SD%, summed difference percent; SPECT, single-photon emission computed tomography; SRS, summed rest score; SR%, summed rest percent; SSS, summed stress score; SS%, summed stress percent.

In the case of multiple SPECT examinations throughout the study period, the examination performed immediately before arrhythmic death and/or appropriate ICD therapy was adopted for those who had experienced arrhythmic events, while the most recent examination was adopted for those who did not.

Statistical Analysis

Baseline characteristics are presented as mean±standard deviation for continuous variables and numbers (%) for categorical variables. Patients who did or did not experience arrhythmic death and/or appropriate ICD therapy were compared according to their baseline characteristics and scintigraphy data using Student’s t-test or the Mann-Whitney U-test for continuous variables and chi-square test or Fisher’s exact test for categorical variables.

The clinical, echocardiographic, and scintigraphic variables associated with arrhythmic death and/or appropriate ICD therapy were examined using univariate and multivariate Cox proportional hazards regression analyses. Significant factors detected after univariate analysis were adopted during multivariate analysis using a forward stepwise selection method with entry and stay criteria of 0.05 and 0.10, respectively.

The best cutoff values for predictors of arrhythmic death and/or appropriate ICD therapy were determined using receiver-operating characteristic (ROC) curves. Diagnostic accuracy was evaluated using the area under the ROC curve (AUC). The arrhythmic death and/or appropriate ICD therapy-free survival rates were compared using Kaplan-Meier method and log-rank test. Prognostic abilities of the conventional criteria using LVEF ≤35% and additive criteria using scintigraphy data were evaluated by comparing the chi-square values for each model. Data analyses were performed using MedCalc (version 17.2 for Windows) software, with P<0.05 being statistically significant.

Results

Patients’ Characteristics, Mortality, and Occurrence of Arrhythmic Death and Appropriate ICD Therapy

The characteristics and stress thallium-201 SPECT results of the enrolled 60 patients are presented in Table 1 and Table 2, respectively. During the median follow-up period of 6.6 years [interquartile range (IQR), 3.8–10.5], 25 (42%) patients died (20 from cardiovascular causes, 5 from non-cardiovascular causes), leading to a mortality rate of 4.7%/year. Among the 20 patients who died of cardiovascular causes, 16 died during hospitalization and 4 died unwitnessed in a non-hospital setting. The majority of in-hospital deaths involved worsening heart failure (n=12), and none of the hospitalized patients had arrhythmic death. In contrast, among the 4 patients who died outside the hospital, 2 died from ventricular fibrillation despite ICD appropriate therapy and 2 from non-arrhythmic causes, which was later revealed in their ICD records.

Table 1. Baseline Characteristics of Japanese Patients With Prior MI

	Total (n=60)	No arrhythmic death/appropriate ICD therapy (n=42)	Arrhythmic death/appropriate ICD therapy (n=18)	P value
Age, years	70±10	70±10	72±11	0.39
Male, n (%)	53 (88)	37 (88)	16 (89)	0.93
Body mass index, kg/m2	23.8±5.6	24.0±6.2	23.3±4.3	0.66
Serum creatinine, mg/dL	1.6±1.1	1.5±1.1	1.6±1.3	0.93
eGFR, mL/min/1.73 m2	46±23	46±24	48±21	0.72
BNP level, pg/mL	261±215	250±215	284±220	0.59
Hemoglobin A1c, %	6.5±1.0	6.6±1.1	6.3±0.8	0.25
NYHA class, n (%)
1–2	43 (72)	31 (74)	12 (67)	0.66
3–4	17 (28)	11 (26)	6 (33)
Indication of ICD, n (%)
Primary prevention	24 (40)	18 (43)	6 (33)	0.49
Secondary prevention	36 (60)	24 (57)	12 (67)
Defibrillator device, n (%)
ICD	47 (78)	31 (79)	16 (89)	0.20
CRT-D	13 (22)	11 (26)	2 (11)
Intrinsic LBBB	19 (32)	15 (36)	4 (22)	0.31
Cum% Vp
<1	40 (66)	26 (62)	14 (78)	0.56
1–40	3 (5)	2 (5)	1 (5)
41–89	2 (3)	2 (5)	0 (0)
≥90	15 (25)	12 (29)	3 (17)
Any revascularization treatment, n (%)	53 (88)	40 (91)	13 (75)	0.021
Prior PCI	50 (72)	37 (88)	13 (72)	0.13
CABG	18 (30)	12 (29)	6 (33)	0.71
Reperfusion for acute MI, n (%)	44 (73)	32 (76)	12 (67)	0.45
No. of vessels associated with prior MI, n (%)
1	31 (52)	27 (64)	4 (22)	0.010
2	26 (43)	13 (31)	13 (72)
3	3 (5)	2 (5)	1 (6)
No. of vessels treated by any angioplasty, n (%)
1	16 (25)	8 (21)	8 (38)	0.043
2	23 (38)	20 (48)	3 (13)
3	21 (35)	14 (32)	7 (44)
Coexisting diseases, n (%)
Atrial fibrillation	24 (40)	18 (43)	6 (33)	0.81
Valvular heart disease	12 (20)	9 (22)	3 (17)	0.86
Hypertension	55 (92)	38 (91)	17 (94)	0.61
Dyslipidemia	50 (83)	35 (83)	15 (83)	1.00
Diabetes mellitus	38 (63)	26 (62)	12 (67)	0.73
Chronic kidney disease	16 (27)	12 (29)	4 (22)	0.61
Echocardiographic data
LAD, mm	49±9	48±9	50±10	0.49
LVDd, mm	63±9	61±7	67±11	0.052
LVDs, mm	51±11	49±9	56±12	0.030
LVEF, %	37±12	38±11	36±15	0.69
Medications, n (%)
β-blocker	55 (92)	37 (89)	18 (100)	0.13
ACEI/ARB	45 (75)	30 (71)	15 (83)	0.33
Amiodarone	23 (38)	16 (38)	7 (39)	0.95
Sodium-channel blocker	2 (3)	0 (0)	2 (11)	0.086
Antiplatelet	56 (93)	39 (93)	17 (94)	0.55
Anticoagulant	29 (48)	21 (50)	8 (44)	0.70

Valvular heart disease: presence of moderate-to-severe regurgitation or stenosis in the aortic or mitral valve. Patients with diabetes mellitus were defined as hemoglobin A1c >6.5% or already taking diabetic medication. Chronic kidney disease was defined as estimated glomerular filtration rate (eGFR) <30 mL/min/1.73 m². ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; BNP, B-type natriuretic peptide; CABG, coronary artery bypass grafting; CRT-D, cardiac resynchronization therapy defibrillator; Cum% Vp, cumulative percentage of ventricular pacing; ICD, implantable cardioverter defibrillator; LAD, left atrial diameter; LBBB, left bundle branch block; LVDd, left ventricular diastolic diameter; LVDs, left ventricular systolic diameter; LVEF, left ventricular ejection fraction; MI, myocardial infarction; NYHA, New York Heart Association; PCI, percutaneous coronary intervention.

Table 2. Results of Stress Thallium-201 SPECT in Japanese Patients With Prior MI

	Total (n=60)	No arrhythmic death/appropriate ICD therapy (n=42)	Arrhythmic death/appropriate ICD therapy (n=18)	P value
Stress test, n (%)
By exercise	10 (17)	8 (19)	2 (11)	0.71
By adenosine	50 (83)	34 (81)	16 (89)
Volumetric data
Stress LVEDV, mL	171±77	154±71	210±79	0.009
Stress LVESV, mL	125±73	109±66	163±77	0.008
Stress LVEF, %	31±14	34±15	25±10	0.026
Rest LVEDV, mL	169±80	151±72	211±84	0.006
Rest LVESV, mL	125±73	107±66	166±75	0.004
Rest LVEF, %	31±14	34±15	24±7	<0.001
Global perfusion data
SSS	20.3±9.6	17.5±8.7	27.1±8.1	<0.001
SRS	20.7±10.4	18.0±10.2	26.9±8.2	0.002
SDS	2.4±2.4	2.1±2.0	3.1±3.1	0.26

Stress and rest LVEDV were defined as LVEDV measured by stress and redistribution images, respectively. The same definition was applied to stress/rest LVESV and LVEF. LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular endsystolic volume; PDS, perfusion defect size; SDS, summed difference score; SPECT, single-photon emission computed tomography; SRS, summed rest score; SSS, summed stress score. Other abbreviations as in Table 1.

Appropriate ICD therapy was detected in 18 (30%) patients during the follow-up period; among these patients, 11 (61%) underwent shock therapy and 7 (39%) underwent antitachycardia pacing alone; 2 of the patients with arrhythmic death had a history of appropriate ICD shock therapy for ventricular fibrillation, but 12 (20%) patients experienced inappropriate ICD therapy because of atrial arrhythmia with rapid ventricular rate, including atrial fibrillation (n=8) and supraventricular tachycardia (n=3); 1 patient experienced inappropriate ICD therapy because of T-wave oversensing. The median frequency of ICD therapy was 2 (IQR, 2–4), and 14 (78%) patients underwent multiple ICD therapies.

Results of stress thallium-201 SPECT are shown in Table 2. The number of patients with moderate-to-severe ischemia, which was defined as SDS >7,²⁰ was 4 (7%). The other 56 (93%) patients had mild or no ischemia. Patients with arrhythmic death and/or appropriate ICD therapy showed a larger SSS and SRS and a more enlarged LV chamber size and more reduced LVEF than those without arrhythmic events. However, there was no difference in the SDS between the patient groups.

Predictors of Arrhythmic Death and Appropriate ICD Therapy

The results of Cox proportional hazards regression analyses are presented in Table 3. SRS, which showed high collinearity with SSS (r=0.90, P<0.001), was a possible predictor of arrhythmic death and/or appropriate ICD therapy in the univariate analysis but was excluded in the stepwise regression analysis. Accordingly, SSS and LVEF at rest were found to be significant independent predictors associated with future arrhythmic death and/or appropriate ICD therapy.

Table 3. Cox Regression Analysis of Factors Associated With Future Arrhythmic Death and/or Appropriate ICD Therapy

	Univariate		Multivariate*
	HR (95% CI)	P value	HR (95% CI)	P value
Baseline clinical variables
Age, years	1.02 (0.96–1.09)	0.46	–
Male, n (%)	1.08 (0.19–6.17)	0.93	–
Body mass index, kg/m2	0.98 (0.88–1.09)	0.65	–
Serum creatinine, mg/dL	1.02 (0.63–1.66)	0.93	–
eGFR, mL/min/1.73 m2	1.01 (0.98–1.03)	0.72	–
BNP level, pg/mL	1.00 (0.99–1.01)	0.58	–
Hemoglobin A1c, %	0.66 (0.33–1.35)	0.26	–
NYHA class ≥3	1.41 (0.43–4.67)	0.57	–
Secondary prevention	1.50 (0.47–4.76)	0.49	–
CRT-D	0.35 (0.07–1.79)	0.21	–
Intrinsic LBBB	0.51 (0.14–1.85)	0.31	–
Cum% Vp ≥90%	0.50 (0.12–2.05)	0.33	–
Any revascularization treatment	0.13 (0.02–0.75)	0.023	NA
Prior PCI	0.35 (0.09–1.41)	0.14	–
CABG	1.25 (0.38–4.10)	0.71	–
Reperfusion for acute MI, n (%)	0.63 (0.19–2.10)	0.45	–
≥2 vessels prior MI	6.30 (1.76–22.6)	0.005	NA
≥2 vessels treated by angioplasty	0.29 (0.09–0.98)	0.04	NA
Coexisting disease
Atrial fibrillation	0.74 (0.21–2.67)	0.65	–
Valvular heart disease	0.63 (0.12–3.39)	0.59	–
Hypertension	1.79 (0.19–17.2)	0.61	–
Dyslipidemia	1.00 (0.23–4.40)	1.00	–
Diabetes mellitus	1.23 (0.39–3.93)	0.73	–
Chronic kidney disease	0.63 (0.16–2.48)	0.51	–
Medication
β-blocker	2.30 (0.25–21.2)	0.66	–
ACEI/ARB	2.00 (0.49–8.18)	0.34	–
Amiodarone	1.03 (0.33–3.21)	0.95	–
Sodium-channel blocker use	8.46 (1.70–42.1)	0.009	NA
Antiplatelet	0.67 (0.18–2.53)	0.55	–
Anticoagulant	0.80 (0.26–2.43)	0.69	–
Echocardiography variables
LAD, mm	1.02 (0.96–1.09)	0.49	–
LVDd, mm	1.09 (1.01–1.17)	0.024	NA
LVDs, mm	1.07 (1.00–1.14)	0.037	NA
LVEF, %	0.99 (0.95–1.04)	0.69	–
SPECT variables
Volumetric data
Stress LVEDV, mL	1.01 (1.00–1.02)	0.017	NA
Stress LVESV, mL	1.01 (1.00–1.02)	0.015	NA
Stress LVEF, %	0.95 (0.90–0.99)	0.033	NA
Rest LVEDV, mL	1.01 (1.00–1.02)	0.012	NA
Rest LVESV, mL	1.01 (1.00–1.02)	0.008	NA
Rest LVEF, %	0.92 (0.86–0.98)	0.010	0.92 (0.85–0.99)	0.038
Global perfusion data
SSS	1.15 (1.05–1.24)	0.001	1.14 (1.04–1.24)	0.005
SRS	1.10 (1.03–1.18)	0.005	NA
SDS	1.17 (0.93–1.47)	0.18	–

*Multivariate Cox regression analysis was performed to assess SPECT data that reflected the occurrence of arrhythmic death and/or appropriate therapy. Covariates that were significant from the univariate analysis (P<0.05) were included in a multivariable model. The stepwise forward method with entry and stay criteria of 0.05 and 0.10 was used. Based on the results of the univariate analysis, many apparently non-significant factors were not included in the multivariate analysis. Those factors are indicated as “–”. Other factors were included in the multivariate stepwise analysis, but some were excluded during the analysis (“NA”). NA, not applicable. Other abbreviations as in Tables 1,2.

ROC Curve Analysis

We constructed ROC curves of SSS and LVEF for predicting future arrhythmic death and/or appropriate ICD therapy and determined the AUCs (Figure 2A). In the analysis, the best SSS and LVEF cutoff values for predicting the events were 21 and 30%, respectively. There was no significant difference in the diagnostic accuracy for both endpoints between SSS and LVEF. The scatter diagram and risk stratification of the patients according to the cutoff values are shown in Figure 3. The incidence of both arrhythmic death and/or appropriate ICD therapy was significantly higher in patients with SSS ≥21 and LVEF ≤30% than in the other patients.

Figure 2.

Receiver-operating characteristic (ROC) curves. (A) ROC curve analysis of SSS and LVEF at rest for the prediction of an arrhythmic death or appropriate ICD therapy. The comparison of AUCs using the Hanley and McNeil method revealed no significant difference, which suggested that the discrimination accuracy of SSS and LVEF was equivalent. (B) Comparison of the AUCs among perfusion data shows no significant difference between SSS and SRS. AUC, area under the curve; ICD, implantable cardioverter-defibrillator; LVEF, left ventricular ejection fraction; SDS, summed difference score; SRS, summed rest score; SSS, summed stress score.

Figure 3.

Scatter plot and classification of patients according to SSS (vertical axis) and LVEF (horizontal axis). A significant but weak correlation between SSS and LVEF (r=−0.42, P<0.001) can be seen. Patients were classified into the 4 quadrants according to the cutoff values of SSS=21 and LVEF=30%, which identify future arrhythmic death or appropriate ICD therapy. Group I included patients with SSS <21 and LVEF >30% (n=17); group II included patients with SSS ≥21 and LVEF >30% (n=8); group III included patients with SSS <21 and LVEF ≤30% (n=13); and group IV included patients with SSS ≥21 and LVEF ≤30% (n=22). Chi-square test to investigate differences in the incidence of arrhythmic death and/or appropriate ICD therapy revealed a statistically significant difference among the 4 groups (χ²=20.6, P<0.001). A residual analysis identified a difference between groups I and IV making the greatest contribution to the χ² test result (P<0.001). The incidence of an arrhythmic death and/or appropriate ICD therapy was highest (64%) in group IV and lowest (0%) in group I. The incidence in groups II and III was intermediate. Group I vs. group II, P=0.32; group I vs. group III, P=0.07; group I vs. group IV, P<0.001; group II vs. group III, P=1.00; group II vs. group IV, P=0.035; group III vs. group IV, P=0.035. ATP, antitachycardia pacing; ICD, implantable cardioverter-defibrillator; LVEF, left ventricular ejection fraction; SSS, summed stress score.

We also constructed ROC curves of SSS, SRS, and SDS and compared the diagnostic accuracy for the prediction of future arrhythmic death and/or appropriate ICD therapy (Figure 2B). On comparing the AUCs, SSS and SRS, but not the SDS, were found to be equally effective diagnostic tools for predicting the events.

Survival Analysis According to SSS and LVEF Cutoff Values

The arrhythmic death and/or appropriate ICD therapy-free survival curves according to the best SSS and LVEF cutoff values were compared. Among the 30 patients with SSS ≥21, 15 (50%) experienced arrhythmic death and/or appropriate ICD therapy, while among the 30 patients with SSS <21, only 3 (10%) experienced arrhythmic events (HR, 3.78; 95% confidence interval [CI], 1.49–9.60) (Figure 4A). Conversely, among 35 patients with LVEF ≤30%, 17 (49%) experienced arrhythmic death and/or appropriate ICD therapy, while among 25 patients with LVEF >30%, only 1 (4%) experienced arrhythmic events (HR, 17.4; 95% CI, 6.89–43.7) (Figure 4B). After dividing the patients into 4 groups using the cutoff values for SSS=21 and LVEF=30%, those with SSS ≥21 and LVEF ≤30% were found to have the highest incidence of arrhythmic death and/or appropriate ICD therapy (HR, 7.18; log-rank P=0.001) (Figure 4C). None of the patients with SSS <21 and LVEF >30% experienced arrhythmic death and/or appropriate ICD therapy during the follow-up period.

Figure 4.

Kaplan-Meier survival curves of arrhythmic death and/or appropriate ICD therapy-free survival (A–C) are presented from the day of ICD implantation. The median follow-up period was 6.6 (IQR, 3.8–10.5) years. Patients were stratified according to SSS (A), LVEF (B), and both factors (C). CI, confidence interval; HR, hazard ratio; ICD, implantable cardioverter-defibrillator; IQR, interquartile range; LVEF, left ventricular ejection fraction; SSS, summed stress score.

Analysis using a Cox proportional hazard model for the endpoint revealed that the chi-square values for the conventional criteria using LVEF ≤35% and additive criteria using both LVEF ≤35% and SSS ≥21 were 7.12 and 16.8, respectively (P=0.002).

Analysis for Lethal Arrhythmia

Additional analysis was performed to evaluate the risk of lethal arrhythmic events (n=11), defined as arrhythmic death and/or appropriate ICD shock therapy excluding antitachycardia pacing. Thus, SSS and LVEF at rest were also associated with future arrhythmic death and/or appropriate ICD shock therapy (Table 4). Successful risk stratification using the cutoff values for SSS=27 and LVEF=28% was also demonstrated. The incidence of arrhythmic death and/or appropriate ICD shock therapy was the lowest (0%) in patients with SSS <27 and LVEF >28% and the highest (60%) in patients with SSS ≥27 and LVEF ≤28% (P=0.017). Incidence was intermediate for the remaining patients (Figure S1). Patients who had arrhythmic death or experienced appropriate ICD shock therapy (n=11) and those who underwent antitachycardia pacing alone (n=7) showed a similar SSS (29.3±7.9 vs. 23.6±7.7; P=0.15) and LVEF at rest (22.5±4.7 vs. 25.0±9.3; P=0.46).

Table 4. Impact of SSS and LVEF on Different Clinical Outcomes

	Univariate		Multivariate*
	HR (95% CI)	P value	HR (95% CI)	P value
Appropriate ICD therapy including ATP therapy
SSS	1.15 (1.05–1.24)	0.001	1.14 (1.04–1.24)	0.005
Rest LVEF, %	0.92 (0.86–0.98)	0.010	0.92 (0.85–0.99)	0.038
Arrhythmic death and/or appropriate ICD shock therapy
SSS	1.18 (1.06–1.31)	0.002	1.19 (1.06–1.35)	0.004
Rest LVEF, %	0.92 (0.85–0.99)	0.029	0.90 (0.81–1.00)	0.048
Cardiovascular death
SSS	1.03 (0.97–1.09)	0.40	1.03 (0.97–1.10)	0.32
Rest LVEF, %	1.00 (0.96–1.04)	0.88	1.01 (0.97–1.06)	0.58
All-cause death
SSS	1.01 (0.95–1.06)	0.78	1.01 (0.95–1.07)	0.77
Rest LVEF, %	1.00 (0.96–1.04)	0.97	1.00 (0.96–1.04)	0.93

*Multivariate Cox regression analysis included SSS and LVEF evaluated by SPECT to evaluate the future occurrence of each clinical event. ATP, antitachycardia pacing. Other abbreviations as in Tables 1,2.

Discussion

This study demonstrated that SSS and LVEF at rest were independent predictors of arrhythmic death and/or appropriate ICD therapy among current Japanese patients with prior MI and that SSS=21 and LVEF=30% were the best cutoff values. The combined use of SSS and LVEF was found to be better than their individual use at identifying patients who would benefit from ICD therapy. Incidence of arrhythmic death and/or appropriate ICD therapy was highest among patients with SSS ≥21 and LVEF ≤30%.

Although the identification of high-risk patients for sudden cardiac death is important, the identification of very-low-risk patients should take precedence to avoid potentially negative consequences of ICD implantation, such as financial burden, device infection, and inappropriate ICD therapy.³ Therefore, we consider it remarkable that none of the patients with SSS <21 and LVEF >30% experienced arrhythmic death and/or appropriate ICD therapy. Accordingly, prophylactic ICD implantation to prevent sudden cardiac death may not be needed for patients with SSS <21 and LVEF >30%. Nonetheless, these findings should be confirmed using prospective, large-scale, randomized controlled trials.

Comparison With Previous Studies According to Baseline Characteristics

The survival rate of Japanese patients with MI was reported to be comparable with that of the MADIT-II defibrillator group and higher than that of the MADIT-II conventional therapy group.^5–10 The overall crude mortality rate determined in the present study (4.7%/year) was much lower than that determined in the MADIT-II conventional therapy group (11.9%) or defibrillator group (8.5%).¹ Possible reasons for the improved prognosis included different characteristics of the enrolled patients in the present study and the MADIT-II study. The patients in the present study had a higher mean age (70 vs. 65 years), more frequent percutaneous coronary interventions (PCI) (83% vs. 45%), and less frequent coronary artery bypass grafting (30% vs. 45%) than those in the MADIT-II study. Moreover, the participants in the present study used angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers (75% vs. 68%) and β-blockers (92% vs. 70%) more frequently than those in the MADIT-II study.

With the dawn of the PCI era in Japan, most patients with acute MI undergo primary PCI within a few hours of symptom onset, which may account for the improved survival in such patients. According to the MIYAGI-AMI Registry, primary PCI dramatically increased from 20% in 1992 to 80% in 2008. Patients who undergo primary PCI have significantly lower in-hospital mortality rates than those who do not (5% vs. 17%, P<0.01).⁷ Another prospective, multicenter cohort study conducted in Japan between July 2013 and May 2014 showed that primary PCI was performed in 85% of patients with acute MI.⁸ In the present study, primary PCI was performed in 57%, 87%, and 100% of patients with acute MI before 2000, between 2000 and 2010, and after 2010, respectively.

Towards Better Risk Stratification Using SSS

LV systolic function is the main determinant of increased ventricular arrhythmia, sudden cardiac death, and cardiovascular death among patients with prior MI.^1,2 Multiple factors resulting from a reduced LVEF, such as increased myocardial stretch, hormonal and electrolyte alterations, and repolarization abnormalities, have been advocated to cause electrical instability and facilitate the potential substrate for ventricular arrhythmias.²¹ However, limitations regarding the use of LVEF to identify patients who would have benefited from ICD implantation have already been established.²² Thus, we need an alternative parameter with sufficient predictive ability to guide decision-making regarding ICD implantation.

Stress thallium-201 SPECT is a reliable and established method for evaluating scarring and ischemia in patients with coronary artery disease. This method enables the identification of necrotic, fibrotic, or degenerative myocardia with delayed and incomplete electrical conduction, which is a fundamental component causing scar-related macro-reentrant ventricular tachyarrhythmia in patients with prior MI.²³ The present study, which evaluated patients by thallium-201, revealed the additive prognostic value of SSS on LVEF in identifying future arrhythmic death and/or appropriate ICD therapy.

The limited correlation between SSS and LVEF, which was consistent with previous studies,^16,24 further improved the risk stratification using both variables. Unlike LVEF, SSS reflects the extent of fibrosis and the peri-infarct ischemic zone and is less influenced by hemodynamic factors such as preload and afterload. Furthermore, the heterogeneous distribution of fibrosis²⁴ and the extent of the peri-infarct zone,²⁵ which can contribute to the susceptibility to ventricular arrhythmias, may not be indicated by the LVEF at rest.

We can infer that combining SSS and LVEF is probably the most appropriate approach for improving risk stratification and for identifying patients who would benefit from ICD implantation. Significantly higher chi-square values obtained by adding SSS ≤21 to conventional criteria using LVEF ≤35% demonstrated the additional value of SSS for predicting future arrhythmic death and/or appropriate ICD therapy. Using this approach in the present study enabled successful stratification of patients into high-, intermediate-, and low-risk groups (Figures 3,4C). Accordingly, identifying low-risk patients may contribute to early hospital discharge, improved utilization of healthcare resources, and considerable cost savings.

Contribution of Residual Ischemia to Ventricular Arrhythmia

Myocardial ischemia superimposed on a chronic scar potentially produces electrical instability, leading to lethal ventricular arrhythmia.^23,26 However, in the present study, residual ischemia, which was indicated by SDS on scintigraphy, did not prove to be a predictor of arrhythmic death and/or appropriate ICD therapy. Potential underestimation of ischemia (denoted by SDS) in patients with multivessel disease²⁷ might influence the result. However, the primary reason for these results was that ischemia responsible for lethal arrhythmic events was abolished by aggressive angioplasty. In fact, SDS in the enrolled patients was very low (2.4±2.4), and only 4 (7%) patients had moderate-to-severe ischemia, defined as SDS >7 (i.e., >10% ischemia), that would usually be treated by coronary intervention.²⁰

Considering the low SDS value in the present study, analysis using SRS instead of SSS was assumed to produce similar results. In the present study, SRS was a significant predictor of arrhythmic death and/or appropriate ICD therapy in the univariate analysis, and comparison of the AUCs for the prediction of events revealed no significant difference between SSS and SRS. After revascularization in the ischemic region, rest thallium-201 SPECT may be substituted for risk assessment of future lethal arrhythmic events, particularly in patients presenting with contraindications for the stress-rest protocol, such as NYHA class IV status and hemodynamic instability. The stepwise result that SSS, not SRS, was selected in the most appropriate model would indicate that even mild ischemia was associated with future arrhythmic death and/or appropriate ICD therapy when superimposed on a chronic scar, which requires examination in the future.

Use of Different Radiolabeled Tracers

The applicability of this study using different tracers, such as technetium-99 m and iodine-123-metaiodobenzyguanidine (¹²³I-MIBG), should be examined in the future owing to the decline in the use of thallium-201 over the past few decades because of high radiation exposure in patients. A previous study using technetium-99 m tetrofosmin revealed that myocardial defect volume, which describes scar and ischemia, is a useful indicator in predicting lethal arrhythmic events.²⁸ That result supported the use of technetium-99 m tetrofosmin as an alternative radioactive perfusion tracer.

On the other hand, ¹²³I-MIBG is a radiolabeled norepinephrine analog that is used for assessing cardiac sympathetic innervation. A prospective multicenter trial revealed that the ¹²³I-MIBG defect size, which described denervated myocardium, was associated with inducibility of lethal arrhythmias.²⁹ Positive correlation between the defect size identified by ¹²³I-MIBG and infarct size by perfusion SPECT supported the expected similar result using ¹²³I-MIBG instead of thallium-201.³⁰ However, the ¹²³I-MIBG defect size is known to exceed the infarct size identified by perfusion SPECT, possibly because of enhanced sensitivity of neural tissue to ischemia.³⁰ This innervation-perfusion mismatch may provide potential additive value over perfusion SPECT in identifying appropriate candidates for ICD treatment.

Study Limitations

Several limitations of the present study are worth noting. This was a single-center, small-scale retrospective study. SPECT was based on clinical need and not in a series of consecutive patients, which might have introduced selection bias. Moreover, the ICDs in this study were not consistently programed (i.e., heart rate ranges for ventricular tachycardia and ventricular fibrillation were not the same), which might have influenced the occurrence of appropriate ICD therapy. Given that progression of coronary artery disease would likely have been present during the follow-up period, SPECT findings might have altered accordingly. Furthermore, a false-negative result of stress thallium-201 SPECT was reported in patients with prior MI and multivessel disease.²⁷ On the other hand, left bundle branch block and right ventricular pacing are known to have a high incidence of false-positive SPECT findings.³¹ The present study showed no difference in SSS between the 10 patients under ventricular pacing during SPECT and the rest (17±9 vs. 21±10; P=0.29).

Conclusions

SSS and LVEF evaluated using stress thallium-201 SPECT were significantly associated with future occurrence of arrhythmic death and/or appropriate ICD therapy among current Japanese patients with prior MI. SSS in combination with LVEF can help determine the need for ICD therapy but the results presented herein need to be confirmed in a prospective randomized controlled trial.

Conflict of Interest

None.

Supplementary Files

Supplementary File 1

Figure S1. Kaplan-Meier survival curves for lethal arrhythmia.

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-17-1436

References

• 1.   Moss AJ, Zareba W, Hall WJ, Klein H, Wilber DJ, Cannom DS, et al. Prophylactic implantation of a defibrillator in patients with myocardial infarction and reduced ejection fraction. N Engl J Med 2002; 346: 877–883.
• 2.   JCS Joint Working Group. Guidelines for non-pharmacotherapy of cardiac arrhythmias (JCS 2011): Digest version. Circ J 2013; 77: 249–274.
• 3.   Zwanziger J, Hall WJ, Dick AW, Zhao H, Mushlin AI, Hahn RM, et al. The cost effectiveness of implantable cardioverter-defibrillators: Results from the Multicenter Automatic Defibrillator Implantation Trial (MADIT)-II. J Am Coll Cardiol 2006; 47: 2310–2318.
• 4.   Ishikawa K, Kimura A, Taniwa T, Takenaka T, Hayashi T, Kanamasa K. Modification of treatment strategies over a period of 14 years has markedly reduced cardiac events among post-myocardial infarction patients. Circ J 2002; 66: 881–885.
• 5.   Tanno K, Miyoshi F, Watanabe N, Minoura Y, Kawamura M, Ryu S, et al. Are the MADIT II criteria for ICD implantation appropriate for Japanese patients? Circ J 2005; 69: 19–22.
• 6.   Shiga T, Hagiwara N, Ogawa H, Takagi A, Nagashima M, Yamauchi T, et al. Heart Institute of Japan Acute Myocardial Infarction-II (HIJAMI-II) Investigators. Sudden cardiac death and left ventricular ejection fraction during long-term follow-up after acute myocardial infarction in the primary percutaneous coronary intervention era: Results from the HIJAMI-II registry. Heart 2009; 95: 216–220.
• 7.   Takii T, Yasuda S, Takahashi J, Ito K, Shiba N, Shirato K, et al. Trends in acute myocardial infarction incidence and mortality over 30 years in Japan: Report from the MIYAGI-AMI Registry Study. Circ J 2010; 74: 93–100.
• 8.   Ishihara M, Fujino M, Ogawa H, Yasuda S, Noguchi T, Nakao K, et al. Clinical presentation, management and outcome of Japanese patients with acute myocardial infarction in the troponin era: Japanese registry of acute myocardial infarction diagnosed by universal definition (J-MINUET). Circ J 2015; 79: 1255–1262.
• 9.   Satake H, Fukuda K, Sakata Y, Miyata S, Nakano M, Kondo M, et al. Current status of primary prevention of sudden cardiac death with implantable cardioverter defibrillator in patients with chronic heart failure: A report from the CHART-2 Study. Circ J 2015; 79: 381–390.
• 10.   Nakajima K, Nishimura T. Cardiovascular events in Japan: Lessons from the J-ACCESS multicenter prognostic study using myocardial perfusion imaging. Circ J 2012; 76: 1313–1321.
• 11.   Silverman KJ, Becker LC, Bulkley BH, Burow RD, Mellits ED, Kallman CH, et al. Value of early thallium-201 scintigraphy for predicting mortality in patients with acute myocardial infarction. Circulation 1980; 61: 996–1003.
• 12.   Miller TD, Christian TF, Hopfenspringer MR, Hodge DO, Gersh BJ, Gibbons RJ. Infarct size after acute myocardial infarction measured by quantitative tomographic 99 mTc sestamibi imaging predicts subsequent mortality. Circulation 1995; 92: 334–341.
• 13.   Adelstein EC, Tanaka H, Soman P, Miske G, Haberman SC, Saba SF, et al. Impact of scar burden by single-photon emission computed tomography myocardial perfusion imaging on patient outcomes following cardiac resynchronization therapy. Eur Heart J 2011; 32: 93–103.
• 14.   Gioia G, Bagheri B, Gottlieb CD, Schwartzman DS, Callans DJ, Marchlinski FE, et al. Prediction of outcome of patients with life-threatening ventricular arrhythmias treated with automatic implantable cardioverter-defibrillators using SPECT perfusion imaging. Circulation 1997; 95: 390–394.
• 15.   Gradel C, Jain D, Batsford W, Finkel J, Wackers FJT, Zaret BL. Relationship of scar and ischemia to the results of programmed electrophysiological stimulation in patients with coronary artery disease. J Nucl Cardiol 1997; 4: 379–386.
• 16.   De Sutter J, Tavernier R, Van de Wiele C, Kazmierckzak J, De Buyzere M, Jordaens L, et al. Infarct size and recurrence of ventricular arrhythmias after defibrillator implantation. Eur J Nucl Med 2000; 27: 807–815.
• 17.   JCS Joint Working Group. Guidelines for secondary prevention of myocardial infarction (JCS 2011): Digest version. Circ J 2013; 77: 231–248.
• 18.   Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul S, Laskey WK, et al. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: A statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. J Nucl Cardiol 2002; 9: 240–245.
• 19.   Nakajima K. Normal values for nuclear cardiology: Japanese databases for myocardial perfusion, fatty acid and sympathetic imaging and left ventricular function. Ann Nucl Med 2010; 24: 125–135.
• 20.   Hachamovitch R, Rozanski A, Shaw LJ, Stone GW, Thomson LE, Friedman JD, et al. Impact of ischaemia and scar on the therapeutic benefit derived from myocardial revascularization vs. medical therapy among patients undergoing stress-rest myocardial perfusion scintigraphy. Eur Heart J 2011; 32: 1012–1024.
• 21.   Di Diego JM, Antzelevitch C. Ischemic ventricular arrhythmias: Experimental models and their clinical relevance. Heart Rhythm 2011; 8: 1963–1968.
• 22.   Buxton AE, Lee KL, Hafley GE, Pires LA, Fisher JD, Gold MR, et al. Limitations of ejection fraction for prediction of sudden death risk in patients with coronary artery disease: Lessons from the MUSTT study. J Am Coll Cardiol 2007; 50: 1150–1157.
• 23.   de Bakker JM, van Capelle FJ, Janse MJ, Wilde AA, Coronel R, Becker AE, et al. Reentry as a cause of ventricular tachycardia in patients with chronic ischemic heart disease: Electrophysiologic and anatomic correlation. Circulation 1988; 77: 589–606.
• 24.   Klem I, Weinsaft JW, Bahnson TD, Hegland D, Kim HW, Hayes B, et al. Assessment of myocardial scarring improves risk stratification in patients evaluated for cardiac defibrillator implantation. J Am Coll Cardiol 2012; 60: 408–420.
• 25.   Heidary S, Patel H, Chung J, Yokota H, Gupta SN, Bennett MV, et al. Quantitative tissue characterization of infarct core and border zone in patients with ischemic cardiomyopathy by magnetic resonance is associated with future cardiovascular events. J Am Coll Cardiol 2010; 55: 2762–2768.
• 26.   Elhendy A, Chapman S, Porter TR, Windle J. Association of myocardial ischemia with mortality and implantable cardioverter-defibrillator therapy in patients with coronary artery disease at risk of arrhythmic death. J Am Coll Cardiol 2005; 46: 1721–1726.
• 27.   Dunn RF, Freedman B, Bailey IK, Uren R, Kelly DT. Noninvasive prediction of multivessel disease after myocardial infarction. Circulation 1980; 62: 726–734.
• 28.   Morishima I, Sone T, Tsuboi H, Mukawa H, Uesugi M, Morikawa S, et al. Risk stratification of patients with prior myocardial infarction and advanced left ventricular dysfunction by gated myocardial perfusion SPECT imaging. J Nucl Cardiol 2008; 15: 631–637.
• 29.   Bax JJ, Kraft O, Buxton AE, Fjeld JG, Parízek P, Agostini D, et al. 123 I-mIBG scintigraphy to predict inducibility of ventricular arrhythmias on cardiac electrophysiology testing: A prospective multicenter pilot study. Circ Cardiovasc Imaging 2008; 1: 131–140.
• 30.   Matsunari I, Schricke U, Bengel FM, Haase HU, Barthel P, Schmidt G, et al. Extent of cardiac sympathetic neuronal damage is determined by the area of ischemia in patients with acute coronary syndromes. Circulation 2000; 101: 2579–2585.
• 31.   Skalidis EI, Kochiadakis GE, Koukouraki SI, Chrysostomakis SI, Igoumenidis NE, Karkavitsas NS, et al. Myocardial perfusion in patients with permanent ventricular pacing and normal coronary arteries. J Am Coll Cardiol 2001; 37: 124–129.