Suspected Fatal Arrhythmic Events in Japanese Patients With Coronary Artery Disease　― From the CREDO-Kyoto PCI/CABG Registries Cohorts-2 and -3 ―

Abstract

Background: Fatal arrhythmic events (FAEs), such as sudden cardiac death (SCD) and fatal ventricular arrhythmias, are a devastating complication in patients with coronary artery disease (CAD). Therefore, in this study we aimed to assess the incidence of FAEs in more recent Japanese patients with CAD and to examine whether risk stratification of FAEs can still be feasible using the left ventricular ejection fraction (LVEF).

Methods and Results: In the CREDO Kyoto PCI/CABG registry cohorts-2 and -3, there were 25,843 patients with LVEF data who received a first coronary revascularization (LVEF ≤35% group: N=1,671, 35%<LVEF≤40% group: N=1,075, 40%<LVEF≤45% group: N=1,594, and LVEF >45%: N=21,503). FAEs were defined as a composite of SCD or hospitalization for serious ventricular arrhythmias. The cumulative 5-year incidence of FAEs was 2.4% and it increased with decreasing LVEF (LVEF ≤35%: 8.84%, 35%<LVEF≤40%: 6.99%, 40%<LVEF≤45%: 4.49%, and LVEF >45%: 1.67%, log-rank P<0.0001). The adjusted risk of FAEs also increased with decreasing LVEF.

Conclusions: LVEF is still a strong independent factor for predicting FAEs in patients with CAD in the PCI era. There was no obvious decrease in the incidence of FAEs between the 2 cohorts. The risk factors for FAEs through the 2 cohorts, other than low LVEF, included age ≥75 years, diabetes, heart failure, hemodialysis, atrial fibrillation, and anemia.

Fatal arrhythmic events (FAEs), such as sudden cardiac death (SCD) and fatal ventricular arrhythmias, are devastating complications in patients with coronary artery disease (CAD).¹ Implantable cardioverter-defibrillators (ICDs) are an effective and well-established treatment for high-risk CAD patients to protect them from FAEs. As for the indication of a prophylactic ICD, more than 20 years ago several randomized trials^2–4 reported that prophylactic ICDs were effective in CAD patients, particularly those with a low left ventricular ejection fraction (LVEF). Since then, the prediction of SCD and the indication^5,6 for a prophylactic ICD in patients with CAD has been based on LVEF only, according to those trials.^2–4 However, the background of patients with CAD has markedly changed in the percutaneous coronary intervention (PCI) era and the rapidly aging societies, so it is uncertain if LVEF is still a useful predictor of FAEs in patients with CAD. Therefore, in the PCI era, it has become important to ask the following questions: Is the risk stratification for FAEs still possible with LVEF? Is the incidence of FAEs on the decline in Japanese CAD patients? Are there factors other than LVEF for predicting the incidence of FAEs? Thus, in this study we aimed to examine these 3 issues using 2 large cohorts of Japanese CAD patients: the Coronary Revascularization Demonstrating Outcome Study in Kyoto (CREDO-Kyoto) PCI/coronary artery bypass grafting (CABG) Registry cohort-2 and cohort-3.

Methods

Study Population

The CREDO-Kyoto PCI/CABG Registry cohort-2 and cohort-3 were physician-initiated, non-company-sponsored, multicenter registries enrolling consecutive patients who underwent their first coronary revascularization with PCI or isolated CABG: cohort-2 in 26 centers between January 2005 and December 2007 and cohort-3 in 22 centers between January 2011 and December 2013. In total, 30,257 patients were enrolled in the 2 cohorts (cohort-2: N=15,330, and cohort-3: N=14,927). In the present study, we compared the long-term clinical outcomes of patients in both cohorts with a confirmed LVEF, measured either by left ventriculography or echocardiography. The LVEF data we used were the latest data within the 3 months before the procedure. However, in cases of acute coronary syndrome (ACS), the LVEF data during hospitalization (before and after coronary revascularization) or within 3 months after the procedure were collected. After excluding 159 patients who refused to participate in the registries and 4,963 patients whose LVEF data were not available, the current study population consisted of 25,843 patients (cohort-2: N=12,779, and cohort-3: N=13,064). We divided the study population into 4 groups according to the LVEF (LVEF ≤35%: N=1,671, 35%<LVEF≤40%: N=1,075, 40%<LVEF≤45%: N=1,594, and LVEF >45%: N=21,503) (Figure 1). The duration of follow-up was a median of 5.1 (interquartile range [IQR]: 4.3–5.9 years in cohort-2 and 5.6 [IQR 4.5–6.7] years in cohort-3). Complete 1-, 3-, and 5-year follow-up information were obtained in 98.5%, 96.5%, and 69.7%, respectively, in cohort-2 and 96.5%, 94.2%, and 82.6%, respectively, in cohort-3. To standardize the follow-up duration, we censored the follow-up at 5 years after the index procedure.

Figure 1.

Study flowchart. PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting; LVEF, left ventricular ejection fraction.

The relevant ethics committees in all participating centers approved the research protocol. Because of the retrospective enrollment, written informed consent from the patients was waived; however, we excluded those patients who refused participation in the study when contacted for follow-up.

Clinical Outcome Measures and Definitions

The primary outcome measure in the current analysis was FAEs, defined as a composite of SCD, hospitalization for serious ventricular arrhythmia (ventricular tachycardia or ventricular fibrillation), and appropriate shocks by an ICD in a patient who already had an ICD implanted. SCD was regarded as death within 24 h of symptom onset, death during sleep, or death not witnessed in a patient whose condition had been stable. Cases of obvious non-cardiac cause of death were not regarded as SCD cases. Also, sudden death of an exogenous origin (mainly traffic accidents) was not regarded as SCD. The secondary outcome measures included the individual components of the primary outcome, all-cause death, cardiac death, non-cardiac death, cardiovascular death, and non-cardiovascular death. The definitions for the outcome measures were consistent across the 2 cohorts. Death was regarded as cardiac in origin unless obvious non-cardiac cause could be identified. Death of unknown cause and any death during the index hospitalization for coronary revascularization were regarded as cardiac death. Cardiovascular death included cardiac death and other vascular death related to stroke, renal disease, or vascular disease. Myocardial infarctions (MIs) were adjudicated according to the Arterial Revascularization Therapies Study (ARTS) definition.⁷ Strokes were defined as ischemic or hemorrhagic stroke with neurological symptoms lasting >24 h. Coronary revascularization was defined as either PCI or CABG for any reason. Scheduled staged coronary revascularization procedures performed within 3 months of the initial procedure were not regarded as follow-up events but were included in the index procedure.

Data collection for the patients’ baseline characteristics, and the clinical, angiographic, and procedural data were collected from hospital charts or hospital database according to the prespecified definitions by experienced clinical research coordinators in the Research Institute for Production Development (Kyoto, Japan). Follow-up data were collected from hospital charts and/or obtained by contacting the patients, their relatives, or referring physicians. The clinical event committee adjudicated events such as death, myocardial infarctions, definite stent thrombosis or symptomatic graft occlusions, strokes, and major bleeding (Supplementary Methods).

Statistical Analysis

Categorical variables are presented as numbers and percentages. Continuous variables are summarized as the mean±standard deviation or median and IQR (25th, 75th percentile) according to their distributions, and were compared using analysis of variance or Kruskal-Wallis test based on their distributions. Cumulative incidences of the outcome measures were estimated by the Kaplan-Meier method, and the differences were assessed with the log-rank test. Hazard ratios (HR) and their 95% confidence intervals (CI) were estimated as the risk of an FAE across each LVEF group for each outcome measure with univariate or multivariable Cox proportional hazard models, adjusting for the 28 clinically relevant clinical and procedural factors listed in Table 1. In the Cox proportional hazard model, we developed dummy code variables for LVEF ≤35%, 35%<LVEF≤40%, and 40%<LVEF≤45%, with LVEF >45% as the reference. Univariate Cox proportional hazards models were also constructed to assess the strength of the association between the 29 potentially relevant factors and FAEs stratified by cohorts-2 and -3. The 29 potentially relevant factors were the previous mentioned 28 factors and LVEF ≤35% (Table 1). Variables found to be associated at a P<0.10 were included in the multivariate models. Multivariable Cox proportional hazards models were finally constructed to select variables associated with the risk of FAEs using those variables with P<0.05 stratified by cohorts-2 and -3. All tests were two-sided, and P<0.05 was considered statistically significant. All statistical analyses were performed using JMP 13.0 software (SAS Institute Inc., Cary, NC, USA).

Table 1.

Baseline Characteristics of the Patients in the 4 LVEF Groups

Variables	LVEF ≤35% (N=1,671)	35%<LVEF≤40% (N=1,075)	40%<LVEF≤45% (N=1,594)	45% <LVEF (N=21,503)	P value
Clinical characteristics
Age (years)	69.50±11.8	70.0±11.4	68.7±11.7	68.6±10.7	<0.0001
Age ≥75*	620 (37.1)	416 (38.7)	550 (34.5)	6,777 (31.5)	<0.0001
Men*	1,271 (76.1)	773 (71.9)	1,247 (78.2)	15,623 (72.7)	0.02
BMI (kg/m2)	22.8±3.8	23.0±3.9	23.5±3.6	23.9±3.5	<0.0001
BMI <25.0*	1,232 (76.7)	766 (73.4)	1,083 (69.3)	14,030 (66.1)	<0.0001
Clinical presentation
Hypertension*	1,358 (81.3)	901 (83.8)	1,343 (84.3)	17,738 (82.5)	0.09
Diabetes mellitus*	815 (48.8)	499 (46.4)	673 (42.2)	8,213 (38.2)	<0.0001
On insulin therapy	235 (14.1)	139 (12.9)	162 (10.2)	1,822 (8.47)	<0.0001
Current smoker*	547 (32.7)	335 (31.2)	543 (34.1)	6,001 (27.9)	0.0002
Heart failure*	1,015 (60.7)	486 (45.2)	526 (33.0)	2,746 (12.8)	<0.0001
NYHA class ≥II†	536 (32.0)	225 (20.9)	234 (14.7)	1,240 (5.77)	<0.0001
Acute coronary syndrome*	704 (42.1)	539 (50.1)	836 (52.5)	6,704 (31.2)	<0.0001
Prior myocardial infarction*	502 (30.0)	294 (27.4)	360 (22.6)	2,062 (9.60)	<0.0001
Prior stroke*	290 (9.32)	137 (12.7)	197 (12.4)	2,487 (11.6)	<0.0001
Prior heart failure	379 (22.7)	172 (16.0)	147 (9.22)	765 (3.56)	<0.0001
Peripheral vascular disease*	150 (9.0)	87 (8.1)	128 (8.03)	1,830 (8.51)	0.759
eGFR (mL/min/1.73 m2)	55.0 (37.5–71.8)	58.5 (41.5–75.3)	62.5 (46.2–78)	65.5 (52.4–78.4)	<0.0001
Chronic renal failure (eGFR <30)	287 (17.2)	160 (14.9)	188 (11.8)	1,626 (7.56)	<0.0001
eGFR <30, without hemodialysis*	156 (9.34)	77 (7.16)	82 (5.14)	786 (3.66)	<0.0001
Hemodialysis*	131 (7.84)	83 (7.72)	106 (6.65)	840 (3.91)	<0.0001
Atrial fibrillation*	271 (16.2)	150 (14.0)	202 (12.7)	1,931 (8.98)	<0.0001
Anemia (Hb <11.0 g/dL)*	340 (20.4)	197 (18.3)	280 (17.6)	2,421 (11.3)	<0.0001
Thrombocytopenia (platelets <100×109/L)*	61 (3.65)	33 (3.07)	32 (2.01)	341 (1.59)	<0.0001
Chronic obstructive pulmonary disease*	81 (4.85)	55 (5.12)	70 (4.39)	768 (3.57)	0.003
Liver cirrhosis*	37 (2.21)	33 (3.07)	44 (2.76)	559 (2.60)	0.557
Malignancy*	164 (9.81)	118 (11.0)	167 (10.5)	2,331 (10.8)	0.587
Procedural characteristics
PCI*	1,376 (82.3)	897 (83.4)	1,389 (87.1)	18,626 (86.6)	<0.0001
Stent use‡	1,281 (93.1)	849 (94.7)	1,307 (94.1)	17,688 (95.0)	0.0206
DES use‡	883 (68.9)	540 (63.6)	820 (62.7)	12,619 (71.3)	<0.0001
Left main coronary artery	216 (12.9)	129 (12.0)	147 (9.22)	1,728 (8.04)	<0.0001
Proximal LAD*	1,312 (78.5)	836 (77.8)	1,149 (72.1)	14,241 (66.2)	<0.0001
Multivessel coronary artery disease	1,195 (71.5)	722 (67.2)	1,027 (64.4)	12,887 (59.9)	<0.0001
Chronic total occlusion*	552 (33.0)	324 (30.1)	410 (25.7)	3,770 (17.5)	<0.0001
Baseline medication
Aspirin	1,630 (97.6)	1,063 (98.9)	1,579 (99.1)	21,281 (99.0)	<0.0001
P2Y12 inhibitor	1,366 (81.8)	903 (84.0)	1,397 (87.6)	18,771 (87.3)	<0.0001
Statin*	868 (51.9)	605 (56.3)	1,020 (64.0)	13,812 (64.2)	<0.0001
ACE-I/ARB*	1,056 (63.2)	686 (63.8)	1,089 (68.3)	12,239 (56.9)	<0.0001
β-blocker*	879 (52.6)	557 (51.8)	818 (51.3)	7,136 (33.2)	<0.0001
Class III anti-arrhythmic drug	133 (7.96)	42 (3.91)	43 (2.70)	149 (0.69)	<0.0001
Nitrate	388 (23.2)	251 (23.4)	416 (26.1)	5,877 (27.3)	<0.0001
Calcium-channel blocker*	355 (21.2)	260 (24.2)	439 (27.5)	9,416 (43.8)	<0.0001
Oral anticoagulant*	463 (27.7)	257 (23.9)	357 (22.4)	2,647 (12.3)	<0.0001
PPI/H2 blocker*	1,158 (69.3)	766 (71.3)	1,143 (71.7)	14,334 (66.7)	<0.0001

Data expressed as the number (percentage). *Risk adjusting variables selected for the Cox proportional hazard models. ^†NYHA class scored only for complicated heart failure on admission. ^‡Stent use is the percentage of the overall PCI, and DES use is the percentage of overall stent use. Values were missing for BMI in 401 patients (LVEF ≤35%: 65 patients, 35%<LVEF≤40%: 31 patients, 40%<LVEF≤45%: 31 patients, and LVEF >45%: 274 patients), eGFR in 167 (LVEF ≤35%: 9 patients, 35%<LVEF≤40%: 9 patients, 40%<LVEF≤45%: 12 patients, and LVEF >45%: 137 patients), anemia in 180 (LVEF ≤35%: 12 patients, 35%<LVEF≤40%: 9 patients, 40%<LVEF≤45%: 13 patients, and LVEF >45%: 146 patients), and thrombocytopenia in 105 (LVEF ≤35%: 4 patients, 35%<LVEF≤40%: 5 patients, 40%<LVEF≤45%: 7 patients, and LVEF >45%: 89 patients). ACE-I, angiotensin converting enzyme inhibitor; ARB, angiotensin II receptor blocker; BMI, body mass index; DES, drug-eluting stent; eGFR, estimated glomerular filtration rate; H2, histamine type-2 receptor; Hb, hemoglobin; LAD, left anterior descending coronary artery; LVEF, left ventricular ejection fraction; PCI, percutaneous coronary intervention; PPI, proton pump inhibitor.

Results

Baseline Characteristics

Of the 25,843 patients in the current study, there were 1,671 (6.5%) with LVEF ≤35%, 1,075 (4.2%) with 35%<LVEF≤40%, 1,594 (6.2%) with 40%<LVEF≤45%, and 21,503 (83%) with LVEF >45% (Figure 1). The baseline characteristics according to the LVEF status are presented in Table 1. Patients in the lower LVEF groups were older and had a lower body mass index. The prevalence of diabetes, heart failure (HF), anemia, atrial fibrillation (AF), New York Heart Association class (NYHA) >II, and chronic kidney disease without hemodialysis (CKD) (eGFR <30 mL/min/1.73 m²) tended to increase with decreasing LVEF, while the prevalence of hypertension and malignancy did not differ among the 4 groups. The NYHA class data were evaluated only for patients with HF on admission. PCI was less often selected in the lower LVEF groups, and the prevalence of complex lesions such as multivessel disease, left main CAD, and chronic total occlusion tended to increase with decreasing LVEF (Table 1). Of the PCI cases, stents were used in >90% of cases, of which >60% were drug-eluting stents (DES) (Table 1). The baseline characteristics compared between cohort-2 and cohort-3 by LVEF group are presented in Supplementary Table 1. Compared with cohort-2, the clinical characteristics of the patients in cohort-3 showed a trend toward older age, and more prevalent HF and malignancy in all LVEF groups. PCI was more frequently selected in cohort-3. The proportion of DES use increased in cohort-3 (Supplementary Table 1). In cohort-3, baseline medication use, such as with P2Y₁₂ inhibitors, statins, β-blockers, and PPI/H2 blockers, increased significantly in all groups, while the use of nitrates significantly decreased in all groups (Supplementary Table 1).

Clinical Outcomes

The cumulative 5-year incidence of all-cause death and FAEs in the current study were 16.2% and 2.44%, respectively (Figure 2). The cumulative 5-year incidence of FAE incrementally increased with decreasing LVEF (LVEF ≤35%: 8.84%, 35<LVEF≤40%: 6.99%, 40≤LVEF<45%: 4.49%, and LVEF ≥45%: 1.67%, log-rank P<0.0001) (Figure 3). This trend remained the same for all-cause death, cardiac death, SCD, and cardiovascular death (Table 2). Even after adjusting confounders, the 5-year risk for FAE incrementally increased with decreasing LVEF (LVEF ≤35%: HR: 3.18, 95% CI: 2.47–4.09, 35<LVEF≤40%: HR: 2.81, 95% CI: 1.49–3.78, and 40≤LVEF<45%: 1.99, 95% CI: 1.49–2.65 with reference to LVEF ≥45%, all P<0.0001) (Table 3).

Figure 2.

Kaplan-Meier curves for (Left) all-cause death and (Right) fatal arrhythmic events. The Kaplan-Meier curves were truncated at 5 years.

Figure 3.

Kaplan-Meier curves for fatal arrhythmic events across the 4 LVEF groups. LVEF, left ventricular ejection fraction.

Table 2.

Clinical Outcomes at 5 Years in the 4 LVEF Groups

Variables	No. of patients with events (cumulative incidence)				P value
	LVEF ≤35% (N=1,671)	35%<LVEF≤40% (N=1,075)	40%<LVEF≤45% (N=1,594)	45% <LVEF (N=21,503)
All-cause death	586 (37.2)	330 (32.5)	372 (24.9)	2,672 (13.2)	<0.0001
Cardiac death	406 (26.5)	182 (18.9)	187 (13.0)	966 (4.93)	<0.0001
Fatal arrhythmic event	116 (8.84)	61 (6.69)	61 (4.49)	321 (1.67)	<0.0001
Sudden cardiac death	83 (6.5)	42 (4.95)	51 (3.75)	246 (1.29)	<0.0001
Hospitalization for ventricular arrhythmia	33 (2.45)	19 (2.12)	10 (0.76)	75 (0.38)	<0.0001
Non-cardiac death	180 (14.6)	148 (16.8)	185 (13.7)	1,706 (8.71)	<0.0001
Cardiovascular death	436 (28.3)	207 (21.4)	221 (15.2)	1,247 (6.33)	<0.0001
Non-cardiovascular death	150 (12.4)	123 (14.2)	151 (11.4)	1,425 (7.34)	<0.0001

Data expressed as the number (percentage). Cumulative incidence was estimated by the Kaplan-Meier method. LVEF, left ventricular ejection fraction.

Table 3.

Effects of the 3 Groups of Reduced LVEF Relative to the LVEF >45% Group for Fatal Arrhythmic Events

Variable	Crude									Adjusted
	Cohort-2			Cohort-3			Cohort-2+3			Cohort-2			Cohort-3			Cohort-2+3
	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value
LVEF ≤35%	5.81	4.36–7.73	<0.0001	5.76	4.20–7.92	<0.0001	5.82	4.70–7.19	<0.0001	2.78	1.97–3.92	<0.0001	3.65	2.52–5.29	<0.0001	3.18	2.47–4.09	<0.0001
35%<LVEF≤40%	4.82	3.39–6.86	<0.0001	3.87	2.49–6.00	<0.0001	4.44	3.37–5.84	<0.0001	2.99	2.05–4.37	<0.0001	2.59	1.61–4.16	<0.0001	2.81	1.49–3.78	<0.0001
40%<LVEF≤45%	2.65	1.81–3.89	<0.0001	2.89	1.95–4.28	<0.0001	2.76	2.10–3.63	<0.0001	1.82	1.22–2.73	0.0035	2.27	1.50–3.42	<0.0001	1.99	1.49–2.65	<0.0001
45% <LVEF	1 (Ref.)	–	–	1 (Ref.)	–	–	1 (Ref.)	–	–	1 (Ref.)	–	–	1 (Ref.)	–	–	1 (Ref.)	–	–

The HRs and their 95% CIs of the 3 groups of reduced LVEF relative to the LVEF >45% group were estimated by univariate and multivariable Cox proportional hazard models throughout the entire follow-up period after adjusting for the 28 clinically relevant factors listed in Table 1. CI, confidence interval; HR, hazard ratio; LVEF, left ventricular ejection fraction.

The incidence of ICD implantation increased with decreasing LVEF (LVEF ≤35%: 3.7%, 35<LVEF≤40%: 2.0%, 40≤LVEF<45%: 1.3%, and LVEF ≥45%: 0.2%) (Supplementary Table 2).

Clinical outcomes comparing between cohort-2 and cohort-3 were presented in Supplementary Table 3, and Supplementary Figure 1 and 2.

Discussion

The main findings of the present study from a large Japanese observational database of patients who underwent a first-time coronary revascularization are as follows. (1) The 5-year adjusted risk for FAEs incrementally increased with decreasing LVEF (Figure 3). LVEF was still a strong independent factor for predicting FAEs in patients with CAD in the PCI era. (2) The cumulative 5-year incidence of FAEs in patients with LVEF ≤35% was 8.84% (Figure 3). There was no obvious decrease in the incidence of FAEs between the 2 cohorts (Supplementary Figure 1). (3) The risk factors for FAEs other than low LVEF included age ≥75 years, diabetes, HF, hemodialysis, AF, and anemia in both cohort-2 and -3 (Table 4).

Table 4.

Univariate and Multivariate Analyses for Fatal Arrhythmic Events by Cohort

Variables	Cohort-2						Cohort-3
	Univariate			Multivariate			Univariate			Multivariate
	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value
Age ≥75	1.82	1.41– 2.36	<0.0001	1.82	1.42– 2.33	<0.0001	1.72	1.31– 2.26	0.0001	1.86	1.43– 2.41	<0.0001
Men	1.04	0.79– 1.37	0.791				1.35	1.00– 1.84	0.050	1.33	0.99– 1.80	0.0591
BMI <25.0	0.87	0.66– 1.15	0.343				0.91	0.68– 1.21	0.511
Hypertension	1.00	0.71– 1.42	0.981				1.14	0.78– 1.70	0.518
Diabetes mellitus	1.56	1.22– 1.98	0.0003	1.51	1.19– 1.91	0.0006	1.44	1.11– 1.86	0.006	1.45	1.13– 1.87	0.004
Current smoker	1.29	0.99– 1.68	0.063	1.26	0.98– 1.62	0.0765	0.82	0.59– 1.12	0.212
Heart failure	2.40	1.82– 3.15	<0.0001	2.41	1.85– 3.13	<0.0001	1.99	1.49– 2.64	<0.0001	2.01	1.52– 2.64	<0.0001
LVEF ≤35%	2.03	1.47– 2.81	<0.0001	2.33	1.71– 3.13	<0.0001	2.75	1.93– 3.86	<0.0001	2.74	1.95– 3.78	<0.0001
Acute coronary syndrome	1.29	0.96– 1.72	0.086	1.28	0.98– 1.67	0.066	1.53	1.13– 2.05	0.005	1.45	1.12– 1.89	0.0053
Prior myocardial infarction	1.59	1.17– 2.15	0.003	1.59	1.19– 2.13	0.0018	1.12	0.78– 1.61	0.523
Prior stroke	1.09	0.78– 1.51	0.626				1.30	0.95– 1.78	0.098	1.32	0.96– 1.79	0.082
Peripheral vascular disease	1.53	1.08– 2.18	0.018	1.61	1.15– 2.25	0.0056	1.26	0.86– 1.84	0.230
eGFR <30, without hemodialysis	1.96	1.28– 3.01	0.002	2.10	1.41– 3.13	0.0002	1.33	0.81– 2.20	0.257
Hemodialysis	2.11	1.29– 3.45	0.003	1.99	1.25– 3.17	0.0038	2.51	1.42– 4.41	0.001	3.30	2.24– 4.78	<0.0001
Atrial fibrillation	1.61	1.16– 2.24	0.005	1.49	1.10– 2.03	0.0101	1.59	1.09– 2.32	0.016	1.59	1.14– 2.19	0.0072
Anemia (Hb <11.0 g/dL)	1.57	1.16– 2.14	0.004	1.50	1.12– 2.01	0.0069	1.45	1.04– 2.01	0.027	1.53	1.11– 2.10	0.0108
Thrombocytopenia (platelet <100×109/L)	0.95	0.42– 2.14	0.893				0.71	0.31– 1.61	0.407
Chronic obstructive pulmonary disease	1.48	0.87– 2.50	0.145				1.52	0.93– 2.48	0.095	1.55	0.92– 2.44	0.098
Liver cirrhosis	0.65	0.30– 1.39	0.263				1.20	0.61– 2.36	0.589
Malignancy	1.61	1.16– 2.25	0.005	1.56	1.12– 2.15	0.0077	1.20	0.85– 1.70	0.291
PCI, n (%)	1.63	1.12– 2.38	0.011	1.59	1.12– 2.25	0.0092	1.24	0.80– 1.94	0.341
Proximal LAD	1.22	0.95– 1.56	0.126				0.99	0.74– 1.34	0.970
Chronic total occlusion	1.05	0.76– 1.45	0.780				1.40	1.06– 1.86	0.017	1.36	1.03– 1.78	0.0279
Statins	0.80	0.62– 1.03	0.078	0.80	0.63– 1.02	0.0708	0.73	0.55– 0.97	0.032	0.74	0.56– 0.98	0.0333
ACE-I/ARB	0.95	0.73– 1.22	0.669				1.01	0.76– 1.36	0.931
β-blockers	1.02	0.79– 1.32	0.869				0.96	0.74– 1.25	0.775
Calcium-channel blockers	0.92	0.72– 1.19	0.528				1.02	0.77– 1.35	0.878
Oral anticoagulants	0.88	0.62– 1.25	0.485				1.01	0.70– 1.46	0.943
PPI/H2 blocker	0.94	0.74– 1.20	0.636				0.97	0.70– 1.34	0.839

The HRs and their 95% CIs were estimated by the univariate and multivariable Cox proportional hazard models by cohort throughout the entire follow-up period after adjusting for the 28 clinically relevant factors listed in Table 1 in addition to an LVEF ≤35%. CI, confidence interval; HR, hazard ratio. Other abbreviations as in Table 1.

Is Risk Stratification of FAEs Still Possible Using LVEF in the PCI Era?

According to 3 previous Japanese reports,^8–10 an increased incidence of SCD has been reported with decreasing LVEF. In the present study, the 5-year adjusted risk for FAEs also incrementally increased with decreased LVEF (Table 3, Figure 3). Moreover, this trend did not change either cohort (Supplementary Figure 2). On the other hand, none of the previous Japanese reports,^8–10 including this present study, have reported the relationship between patients with an improved LVEF during the follow-up period and the incidence of FAEs. However, the DAPA trial¹¹ reported the benefit of early (<3 months) prophylactic ICD therapy in patients with a high-risk MI within the first 3 years. This survival benefit was also preserved during an extended 9-year follow-up, even though there were many patients whose LVEF improved during the follow-up period. Thus, low LVEF measured in the early phase is still a strong independent factor for predicting FAEs in Japanese patients with CAD in the PCI era. It would be useful to be able to assess the risk of FAEs in a simple way using LVEF, even in the PCI era.

Incidence and Timing of FAEs in the PCI Era

Regarding the reports on the incidence of FAEs according to LVEF in MADIT II,^3,12 cardiac deaths (14.5%) at an average follow-up of 20 months accounted for most of the all-cause deaths (19.8%) in the conventional group. Furthermore, SCD occurred in 9.8% of the control group during follow-up and accounted for the majority of cardiac deaths.

It is difficult to make general comparisons because MADIT II was conducted more than 20 years ago. The patients in the present study were older, the proportion of PCI cases was higher, and the usage rates of β-blockers and statins were lower than in MADIT II. Nevertheless, there was not much difference in the incidence of FAEs between the 2 studies. Moreover, the proportion of cardiac deaths to all-cause deaths also did not differ much between the 2 studies (Table 2). On the other hand, previous cohort studies in Japan^8,9 reported a lower incidence of FAEs than previous Western trials.^2–4 In the HIJAMI-II,⁸ a prospective observational study of 4,133 patients with acute MI, SCD during a mean follow-up period of 4.1 years was 1.2%, and the rate of SCD in patients with LVEF ≤30% was 5.1% at 5 years. In a large Japanese multicenter PCI registry,⁹ the 2-year cumulative incidence of all-cause death and SCD was 4.3% and 0.5%, respectively, and the rate of SCD in patients with LVEF ≤35% was 1.2% at 2 years. Those 2 registries^8,9 included data respectively on ACS patients and chronic CAD patients who underwent PCI. The patients’ backgrounds in the current study differed significantly from those in the 2 registries,^8,9 making it difficult to directly compare the incidence of FAEs. However, a possible reason for the higher incidence of FAEs in our study could be that it was a larger registry that also included patients with stable CAD and those who underwent CABG.

Despite the changes in medications toward guideline-directed medical therapy in cohort-3 compared with cohort-2 (Supplementary Table 1), a possible reason for the lack of a considerable change in the incidence of FAEs over the past decade could be the underuse of prophylactic ICD therapy in Japan.¹⁰ The number of new ICD implantations, including cardiac resynchronization therapy defibrillators (CRT-Ds), during the 5-year follow-up period was 3.7% (62 cases) in the LVEF ≤35% group in the present study (Supplementary Table 2). More frequent use of ICDs might be needed to reduce the incidence of FAEs in Japanese patients with CAD.

Regarding the high-risk period for FAEs, the large Japanese multicenter PCI registry⁹ reported that they were most likely to occur within the first 90 days after PCI. The STICH trial,¹³ a CABG patch trial that randomized patients with LVEF ≤35%, reported the highest risk of SCD within 30–90 days after the procedure. In the present study that included CAD patients after PCI or CABG, the incidence of FAEs also increased rapidly within the first 90 days after the procedure (Figure 3). Even though the incidence of FAEs was highest within the first 90 days after the procedure in reported studies, the guidelines^5,6,14 discourage treating patients with primary prevention ICD within 40 days of MI or 90 days of revascularization, because the significant reduction in SCD with early ICD implantation was offset by an equal and opposite increase in non-SCD death. Therefore, the data might suggest that more frequent use of wearable cardioverter defibrillators would be a practical option in patients at high risk for FAEs within the first 90 days after a procedure.¹⁵

Factors Other Than LVEF That Predict the Occurrence of SCD in the PCI Era

It is reported that only one-third of all SCD cases are associated with a severely reduced LVEF,¹⁶ making it difficult to predict the occurrence of FAEs from LVEF only. Thus, it is necessary to consider whether there are risk-stratifying factors other than LVEF to use to assess the risk of FAEs. The CHART2 study,¹⁰ a prospective observational study of HF patients, reported left ventricular end-diastolic dimension >65 mm and chronic AF as independent predictors of the incidence of FAEs. A Japanese large PCI registry⁹ also reported HF as the strongest predictor associated with the risk of SCD. In the present study, a multivariate analysis was performed stratified by the cohorts (Table 4). The risk factors for FAEs included age ≥75 years, diabetes, HF, hemodialysis, AF, and anemia in both cohort-2 and -3 (Table 4). The management of HF, which was intensified in cohort-3 compared with cohort-2 (Supplementary Table 1), is important to reduce the incidence of FAEs. In any case, the presence of these factors may tend to increase the incidence of FAEs. However, the presence of AF would likely lead to HF, and poor management of diabetes would be a risk of hemodialysis. The risk stratification for the incidence of FAEs complicated by factors such as comorbidities other than LVEF,^5,6 because of the interrelationship of the individual factors. On the other hand, compared with cohort-2, drugs such as β-blockers, statins, and PPI/H2 blockers were more often used in cohort-3 (Supplementary Table 1). It is interesting to note that the use of statins, but not β-blockers, significantly suppressed the incidence of FAEs (Table 4). Certainly, there are reports of decreased ventricular arrhythmias with statins, especially in CAD patients.^17,18 Active use of statins in CAD patients may also be important in reducing the incidence of FAEs. In any case, the challenge for the future is to establish an adequate method of estimating a high risk of FAEs for CAD patients using risk stratification with patient factors including LVEF.

Conclusions

LVEF is still a strong independent factor for predicting FAEs in patients with CAD in the PCI era. There was no obvious decrease in the incidence of FAEs between cohort-2 and cohort-3. The risk factors for FAEs other than a low LVEF included age ≥75 years, diabetes, HF, hemodialysis, AF, and anemia for both cohort-2 and -3.

Study Limitations

First, this study included patients with acute MI and chronic CAD. Moreover, the timing of LVEF measurement was variable (before and after coronary revascularization). There were no data on changes in LVEF over time. Therefore, we were unable to examine the appropriate timing for LVEF evaluation for accurate risk stratification and whether an improved LVEF improves the prognosis. Second, we did not obtain any information on the date of implantation of ICDs and CRT devices. We could not rule out the possibility that the implanted device might have affected the prognosis. Third, there were no data on medications during the follow-up that could have affected the clinical outcomes after coronary revascularization procedures. Fourth, the NYHA class data were evaluated only for complicated HF cases on admission not on discharge. Therefore, the indication for an ICD could not be assessed. Finally, we pooled the data for 2 different cohort studies conducted at separate times.

Acknowledgment

This study was supported by an educational grant from the Research Institute for Production Development (Kyoto, Japan).

Disclosure

K.A. reports honoraria from Abbott Medical Japan, Japan Lifeline, Medtronic Japan and Novartis Pharma. The remaining authors have nothing to disclose. K.M. is a member of the Circulation Journal’s Editorial Team.

IRB Information

The protocols for the CREDO-Kyoto PCI/CABG Registry Cohort-2 and Cohort-3 were approved by the human research ethics committees of the Kyoto University Graduate School of Medicine, reference numbers: E42, E2400.

Data Availability

The deidentified participant data will not be shared.

Supplementary Files

Please find supplementary file(s);

https://doi.org/10.1253/circj.CJ-24-0488

References

• 1.   Huikuri HV, Castellanos A, Myerburg RJ. Sudden death due to cardiac arrhythmias. N Engl J Med 2001; 345: 1473–1482.
• 2.   Moss AJ, Hall WJ, Cannom DS, Daubert JP, Higgins SL, Klein H, et al. Improved survival with an implanted defibrillator in patients with coronary disease at high risk for ventricular arrhythmia: Multicenter Automatic Defibrillator Implantation Trial Investigators. N Engl J Med 1996; 335: 1933–1940.
• 3.   Moss AJ, Zareba W, Jackson Hall W, 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.
• 4.   Bardy GH, Lee KL, Mark DB, Poole JE, Packer DL, Boineau R, et al. Amiodarone or an implantable cardioverter–defibrillator for congestive heart failure. N Engl J Med 2005; 352: 225–237.
• 5.   Stevenson WG, Vice F, Ackerman MJ, Bryant WJ, Callans DJ, Curtis AB, et al. 2017 AHA/ACC/HRS guideline for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death. Heart Rhythm 2018; 15: e73–e189.
• 6.   Zeppenfeld K, Tfelt-Hansen J, de Riva M, Winkel BG, Behr ER, Blom NA, et al. 2022 ESC Guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death. Eur Heart J 2022; 43: 3997–4126.
• 7.   Serruys PW, Unger F, Sousa JE, Jatene A, Bonnier HJ, Schönberger JP, et al. Comparison of coronary-artery bypass surgery and stenting for the treatment of multivessel disease. N Engl J Med 2001; 344: 1117–1124.
• 8.   Shiga T, Hagiwara N, Ogawa H, Takagi A, Nagashima M, Yamauchi T, et al. 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.
• 9.   Nakamaru R, Shiraishi Y, Niimi N, Ueda I, Ikemura N, Suzuki M, et al. Time trend in incidence of sudden cardiac death after percutaneous coronary intervention from 2009 to 2017 (from the Japanese Multicenter Registry). Am J Cardiol 2023; 188: 44–51.
• 10.   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.
• 11.   Haanschoten DM, Elvan A, Ramdat Misier AR, Delnoy PPHM, Smit JJJ, Adiyaman A, et al. Long-term outcome of the randomized DAPA trial. Circ Arrhythm Electrophysiol 2020; 13: e008484.
• 12.   Greenberg H, Case RB, Moss AJ, Brown MW, Carroll ER, Andrews ML. Analysis of mortality events in the multicenter automatic defibrillator implantation trial (MADIT-II). J Am Coll Cardiol 2004; 43: 1459–1465.
• 13.   Rao MP, Al-Khatib SM, Pokorney SD, She L, Romanov A, Nicolau JC, et al. Sudden cardiac death in patients with ischemic heart failure undergoing coronary artery bypass grafting: Results from the STICH randomized clinical trial (Surgical Treatment for Ischemic Heart Failure). Circulation 2017; 135: 1136–1144.
• 14.   Nogami A, Kurita T, Abe H, Ando K, Ishikawa T, Imai K, et al. JCS/JHRS 2019 Guideline on non-pharmacotherapy of cardiac arrhythmias. Circ J 2021; 85: 1104–1244.
• 15.   Klein HU, Goldenberg I, Moss AJ. Risk stratification for implantable cardioverter defibrillator therapy: The role of the wearable cardioverter-defibrillator. Eur Heart J 2013; 34: 2230–2242.
• 16.   Stecker EC, Vickers C, Waltz J, Socoteanu C, John BT, Mariani R, et al. Population-based analysis of sudden cardiac death with and without left ventricular systolic dysfunction: Two-year findings from the Oregon Sudden Unexpected Death Study. J Am Coll Cardiol 2006; 47: 1161–1166.
• 17.   Jacob S, Manickam P, Rathod A, Badheka A, Afonso L, Aravindhakshan R. Statin therapy significantly reduces risk of ventricular tachyarrhythmias in patients with an implantable cardioverter defibrillator. Am J Ther 2012; 19: 261–268.
• 18.   Vyas AK, Guo H, Moss AJ, Olshansky B, McNitt SA, Hall WJ, et al. Reduction in ventricular tachyarrhythmias with statins in the Multicenter Automatic Defibrillator Implantation Trial (MADIT)-II. J Am Coll Cardiol 2006; 47: 769–773.