Five-year major clinical outcomes between first-generation and second-generation drug-eluting stents in acute myocardial infarction patients underwent percutaneous coronary intervention.

BACKGROUND: There were limited data comparing the major clinical outcomes between first-generation (1G)-drug eluting stents (DES) and second-generation (2G)-DES in patients with acute myocardial infarction (AMI) after percutaneous coronary intervention (PCI) during very long follow-up periods. We thought to investigate the comparative efficacy and safety of 2G-DES compared with 1G-DES in AMI patients during 5-year follow-up periods.
METHOD: A total of 1016 eligible AMI patients who underwent PCI with 1G-DES [paclitaxel-, sirolimus-, 1G-zotarolimus-eluting stent (endeavor® or endeavor sprint®), n = 554] or 2G-DES [2G-zotarolimus (endeavor resolute®)- or everolimus-eluting stent, n = 462] were enrolled. The primary endpoint was the occurrence of major adverse cardiac events (MACE) defined as total death, non-fatal myocardial infarction (MI), target lesion revascularization (TLR), target vessel revascularization (TVR), non-target vessel revascularization (Non-TVR) and the secondary endpoint was stent thrombosis (ST) at 5 years.
RESULTS: Two propensity score-matched (PSM) groups (232 pairs, n = 464, C-statistic = 0.802) were generated. During the 5-year follow-up period, the cumulative incidence of TLR [hazard ratio (HR): 3.133; 95% confidence interval (CI): 1.539-6.376; P = 0.002], TVR (HR: 3.144; 95% CI: 1.596-6.192; P = 0.001) and total revascularization rate (HR: 1.874; 95% CI: 1.086-3.140; P = 0.023) were significantly higher in 1G-DES compared with 2G-DES after PSM. However, the incidence of total death, non-fatal MI and ST were similar between the two groups.
CONCLUSION: In this single-center and all-comers registry, 2G-DES's superiorities for TLR, TVR and total revascularization in AMI patients suggested during 5-year clinical follow-up periods.

Introduction

At present, second-generation (2G)-drug-eluting stents (DES) have nearly replaced first-generation (1G)-DES during percutaneous coronary intervention (PCI) in our routine daily clinical practice. Although acute myocardial infarction (AMI) milieu tends to higher thrombotic condition compared to stable coronary artery disease, DES implantation during primary PCI or staged PCI commonly done from the beginning of DES era up to now. During the last few years, several studies demonstrated that 1G-DES such as sirolimus-eluting stent (SES, Cypher®, Cordis Corp., Miami Lakes, Florida)[1] or paclitaxel-eluting stent (PES, Taxus®, Boston Scientific, Natick, Massachusetts)[2] were associated with reductions in angiographic target vessel revascularization (TVR) and major adverse cardiac events (MACE) compared with bare-metal stents (BMS). Chen, et al.[3] reported everolimus-eluting stent (EES, Xience V®, Abbott Vascular, Santa Clara, CA) in the setting of AMI appears to be superior to PESs in reducing target lesion failure, and stent thrombosis (ST). Kang, et al.[4] also reported Zotarolimus-eluting stent (ZES, Resolute®, Medtronic Inc, Santa Rosa, California) showed similar rates of MACE, cardiac death and recurrent myocardial infarction (MI) compared with SES and PES at 12 months and 18 months of follow-up periods in patients with ST-segment elevation MI (STEMI) who undergoing primary PCI. Most 2G-DES were showed non-inferior clinical outcomes compared with 1G-DES.[5],[6] However, there is limited very long-term clinical outcome data comparing the safety and efficacy between 1G-DES and 2G-DES in patients with AMI who underwent successful PCI. Recently one all-comer, randomized, multicenter AMI trials[7] showed that cardiac death (EES: 2.5% vs. SES: 2.7%, P = 0.86) and ST (EES: 2.3% vs. SES: 3.2%, P = 0.60) rates were comparable in both groups during 3–year follow-up periods. The aim of this study was to compare the efficacy and safety of 2G-DES with 1G-DES in AMI patients during long-term clinical follow-up periods.

Methods

This study is a single-center, prospective, all-comers registry designed to reflect the “real world” practice since 2004. Data were collected by a trained study-coordinator with a standardized case report form. This study examined and approved by the institutional review committee and the subjects gave informed written consent. This study performed in accordance with the ethical standards laid down in the 1964 declaration of Helsinki. Informed consent was obtained from all individual participants included in the study prior to enrollment.

Study design and population

A total 1161 AMI patients were underwent coronary angiography (CAG) from January 2004 to October 2012. Among them, these patients were excluded if they had: (1) BMS (n = 26), (2) other types of DES [except for SES, PES, 1G-ZES (endeavor, endeavor sprint), 2G-ZES (endeavor resolute) and EES, n = 40] implantation, (3) stent size required to treat lesion > 3.5 mm (maximum diameter of SES, n = 61), (4) not participated or follow-up loss (n = 18). Finally, a total 1016 eligible AMI patients who treated with 1G-DES (SES, PES, or 1G-ZES, total n = 554) or 2G-DES (2G-ZES, or EES, total n = 462) were enrolled. After a propensity score matched (PSM) analysis, two propensity-matched groups (232 pairs, n = 464) were generated (Figure 1).

Figure 1.

Flow chart of study number of patients.

AMI: acute myocardial infarction; DES: durg eluting balloon; PCI: percutaneous coronary intervention; SES: sirolimus-eluting stent; 1G: first-generation; 2G: second-generation.

PCI procedure and medical treatment

A diagnostic CAG and PCI done through either trans-femoral or trans-radial approaches after an administration of unfractionated heparin (70–100 IU/kg). Patients' activated clotting time maintained above 250 seconds during the procedure. All patients received a loading dose of 200 to 300 mg aspirin and 300 to 600 mg of clopidogrel as the dual antiplatelet therapy (DAPT) and maintained with 100 mg of aspirin and 75 mg of clopidogrel. The use of cilostazol (Pletaal®, Otsuka Pharmaceutical Co., Tokyo, Japan) or platelet glycoprotein IIb/IIIa receptor blockers was left to the discretion of the individual operators. After stent implantation, DAPT (100 mg daily aspirin and 75 mg daily clopidogrel) prescribed at least 12 months. During hospitalization, enrolled patients had taken cardiovascular beneficial medications, including beta-blockers (BB), angiotensin converting enzyme inhibitors (ACEI), or angiotensin receptor blockers (ARB), calcium channel blockers (CCB), and lipid lowering agents. After discharge, the patients were encouraged to stay on the same medications they received during hospitalization.

Study definitions and clinical follow-up

The recording of cardiovascular risk factors and past medical histories based on patients' self-report. The primary endpoint was the occurrence of MACE defined as total death, recurrent non-fatal MI, target lesion revascularization (TLR), TVR, Non-TVR. The secondary endpoint was ST. All deaths classified as cardiac or non-cardiac death. Re-AMI was defined as the presence of clinical symptoms, electrocardiographic changes, or abnormal imaging findings of MI, combined with an increase in the creatine kinase myocardial band fraction above the upper normal limits or an increase in troponin-T/troponin-I to greater than the 99th percentile of the upper normal limit. TLR was defined as a revascularization of the target lesion due to restenosis or re-occlusion within the stent or 5 mm in and adjacent of the distal or proximal segment. TVR was defined as a revascularization of the target vessel or any segment of the coronary artery containing the target lesion. Non-TVR defined as a revascularization of any segment of the non-target coronary artery. Multi-vessel disease was defined as the presence of a lesion with > 50% diameter stenosis in a non–infarct related coronary artery by visual estimation. ST defined as acute (0–24 h), subacute (24 h–30 d), late (30 d–1 year) and very late (> 1 year) according to the onset time of stent thrombosis.[8] The participants were required to visit the outpatient department of cardiology at the end of the first month and then every 3 to 6 months after the index PCI procedure and we could follow up on the clinical data of all enrolled patients through face-to-face interviews at regular outpatient clinic, medical chart reviews, and telephone contacts.

Statistical analysis

For continuous variables, differences between the two groups evaluated with the unpaired t-test or Mann-Whitney rank test. Data expressed as mean ± SD. For discrete variables, differences expressed as counts and percentages and analyzed with χ2 or Fisher's exact test between the groups as appropriate. To adjust for any potential confounders, PSM analysis performed using the logistic regression model. We tested all available variables which could be of potential relevance; gender (men), age, left ventricular ejection fraction (LVEF), STEMI, known cardiovascular diseases (CVD) risk factors, chronic kidney disease, routine angiographic follow-up (RAF), laboratory findings and post-PCI medications (aspirin, clopidogrel, cilostazol, BB, CCB, ACEI, ARB, diuretics, lipid lowering agents). Angiographic and procedural characteristics also considered as covariate. The propensity score (PS) was estimated with the use of C-statistic for the logistic regression model and the C-statistics for the two groups was 0.802. Subjects matched with a caliper width equal to 0.01. Various clinical outcomes estimated with the Kaplan-Meier method, and differences between the two groups compared with the log-rank test. Proportional hazard models used to assess the hazard ratio of the 1G-DESs compared with 2G-DESs adjusted PS. For all analyses, a 2-sided P < 0.05 considered statistically significant. All data processed with SPSS (version 20.0, SPSS-PC, Inc. Chicago, Illinois).

Results

Baseline clinical and angiographic characteristics

Baseline clinical and angiographic characteristics shown in Table 1. Before PSM adjustment, the mean age of the 1G-DES group was 61.9 ± 11.8 years and 2G-DES was 63.0 ± 12.9 years (P = 0.183). Gender distribution was also similar between the two groups (69.3% vs. 74.0%, P = 0.098). The LVEFs were significantly higher in the 2G-DES compared with 1G-DES (49.9 ± 11.2 % vs. 45.7 ± 10.5%, P < 0.001). The histories of previous CVA, MI and PCI were significantly frequent in the 1G-DES. Routine follow-up angiography was more frequently done in the 1G-DES (67.9% vs. 41.3%, P < 0.001). In the aspect of angiographic and procedural characteristics, American College of Cardiology/American Heart Association (ACC/AHA) type B2 and C lesion were more common in 2G-DES and the use of intra-aortic balloon pump (IABP) was more common in 1G-DES. Mean total stent length (25.6 ± 6.4 vs. 23.1 ± 6.4 mm, P < 0.001) and total procedure time (minutes, 48.4 ± 35.5 vs. 36.8 ± 24.1, P < 0.001) were much longer in 1G-DES compared with 2G-DES. However, all of these differences were disappeared after PSM analysis.

Table 1.

Baseline and angiographic characteristics.

Variables	Entire patients			Propensity-matched patients
	1G-DESs (n = 554)	2G-DESs (n = 462)	P-value	1G-DESs (n = 232)	2G-DESs (n = 232)	P-value
Men	384 (69.3%)	342 (74.0%)	0.098	164 (70.7%)	161 (69.4%)	0.706
Age, yrs	61.9 ± 11.8	63.0 ± 12.9	0.183	62.9 ± 11.8	62.8 ± 13.6	0.868
LVEF, %	45.7 ± 10.5	49.9 ± 11.2	< 0.001	48.0 ± 10.3	47.3 ± 11.9	0.471
ST segment elevation MI	282 (50.9%)	252 (54.5%)	0.247	121 (52.2%)	128 (55.2%)	0.515
Hypertension	335 (60.5%)	262 (56.7%)	0.225	137 (59.1%)	123 (53.0%)	0.190
Diabetes mellitus	174 (31.4%)	174 (37.7%)	0.036	89 (38.4%)	88 (37.9%)	0.924
Dyslipidemia	127 (22.9%)	70 (15.2%)	0.002	41 (17.7%)	39 (16.8%)	0.806
Previous cerebrovascular accident	39 (7.0%)	13 (2.8%)	0.002	16 (6.9%)	10 (4.3%)	0.226
Previous MI	14 (2.5%)	1 (0.2%)	0.002	1 (0.4%)	1 (0.4%)	1.000
Previous PCI	34 (6.1%)	4 (0.9%)	< 0.001	5 (2.2%)	4 (1.7%)	0.736
Peripheral vascular disease	13 (2.3%)	6 (1.3%)	0.220	6 (2.6%)	2 (0.9%)	0.154
Chronic kidney disease	33 (6.0%)	25 (5.4%)	0.709	15 (6.5%)	16 (6.9%)	0.853
Routine angiographic follow-up	376 (67.9%)	191 (41.3%)	< 0.001	127 (54.7%)	128 (55.2%)	0.926
CK-MB, mg/dL, initial	100.1 ± 149.3	109.2 ± 146.0	0.330	84.1 ± 139.3	104.9 ± 141.1	0.112
CK-MB, mg/dL, peak	115.9 ± 153.3	121.3 ± 143.9	0.236	104.9 ± 147.2	117.9 ± 138.9	0.330
Troponin T, ng/dL, initial	1.19 ± 2.70	1.41 ± 3.02	0.236	1.21 ± 2.67	1.39 ± 3.42	0.312
Troponin T, ng/dL, peak	1.64 ± 3.19	1.69 ± 3.21	0.836	1.47 ± 3.07	2.00 ± 3.67	0.104
High sensitivity CRP, mg/dL	20.5 ± 38.2	14.7 ± 28.3	0.008	16.1 ± 28.7	16.7 ± 31.2	0.803
Total cholesterol, mg/L	181.9 ± 42.8	184.0 ± 43.4	0.441	181.2 ± 41.8	181.7 ± 44.9	0.448
Triglyceride, mg/L	128.5 ± 70.9	130.5 ± 116.6	0.810	139.3 ± 79.2	122.3 ± 104.5	0.130
HDL cholesterol, mg/L	43.9 ± 11.5	43.3 ± 10.1	0.522	43.4 ± 11.5	42.5 ± 10.2	0.492
LDL cholesterol, mg/L	118.5 ± 38.5	119.6 ± 37.7	0.762	122.9 ± 41.9	117.7 ± 39.2	0.282
Serum creatinine, mg/L	1.05 ± 0.80	1.04 ± 1.00	0.825	1.03 ± 0.46	1.10 ± 1.20	0.385
Serum glucose, mg/dL	134.2 ± 58.4	139.0 ± 70.1	0.260	138.1 ± 63.7	141.7 ± 72.4	0.584
Hemoglobin A1C, %	6.5 ± 1.4	6.6 ± 1.5	0.665	6.9 ± 1.4	6.5 ± 1.2	0.065
Angiographic characteristics
Target vessel
Left anterior descending	309 (55.8%)	276 (59.7%)	0.203	131 (56.5%)	144 (62.1%)	0.219
Left circumflex	157 (28.3%)	139 (30.1%)	0.542	79 (34.1%)	69 (29.7%)	0.319
Right coronary artery	214 (38.6%)	167 (36.1%)	0.416	86 (37.1%)	87 (37.5%)	0.924
Left main	25 (4.5%)	6 (1.3%)	0.003	12 (5.2%)	5 (2.2%)	0.084
Ramus	6 (1.1%)	4 (0.9%)	0.727	2 (0.9%)	3 (1.3%)	0.653
Number of multivessel disease (≥ 2 vessels)	133 (24.0%)	108 (23.4%)	0.814	64 (27.6%)	62 (26.7%)	0.835
ACC/AHA Lesion type
Type B1	9 (1.6%)	9 (1.9%)	0.697	3 (1.3%)	4(1.7%)	0.703
Type B2	53 (9.6%)	112 (24.2%)	< 0.001	36 (15.5%)	33 (14.2%)	0.695
Type C	492 (88.8%)	341 (73.8%)	< 0.001	193 (83.2%)	195 (84.1%)	0.802
Extent of coronary artery disease
1-vessel	419 (75.6%)	354 (76.6%)	0.712	168 (72.4%)	170 (73.3%)	0.835
2-vessel	114 (20.6%)	87 (18.8%)	0.487	51 (22.0%)	49 (21.1%)	0.821
3-vessel	21 (3.8%)	21 (4.5%)	0.547	13 (5.6%)	13 (5.6%)	1.000
Ostial lesion (≤ 5 mm)	118 (21.3%)	81 (17.5%)	0.132	13 (5.6%)	13 (5.6%)	1.000
Bifurcation	205 (37.0%)	187 (40.5%)	0.258	85 (36.6%)	93 (40.1%)	0.445
Heavy Calcification	80 (14.4%)	68 (14.7%)	0.900	38 (16.4%)	39 (16.8%)	0.901
IABP	106 (19.1%)	15 (3.2%)	< 0.001	15 (6.5%)	14 (6.0%)	0.848
Mean total stent length, mm	25.6 ± 6.4	23.1 ± 6.4	< 0.001	24.1 ± 6.7	24.8 ± 6.7	0.189
Mean stent diameter, mm	2.98 ± 0.38	2.92 ± 0.33	0.008	2.93 ± 0.39	2.94 ±0.32	0.543
Number of stents/patient	1.18 ± 0.56	1.18 ± 0.57	0.991	1.21 ± 0.62	1.22 ± 0.61	0.764
Total procedure time, min	48.4 ± 35.5	36.8 ± 24.1	< 0.001	47.2 ± 39.6	45.2 ± 26.5	0.102

Values are mean ± SD or n (%). The P value for continuous data from analysis of variance; the P value for categorical data from chi-square test. ACC/AHA: American college of cardiology/American heart association; CK-MB: creatine kinase-muscle and brain; CRP: c-reactive protein; DESs: drug-eluting stents; HDL: high-density lipoprotein; Hemoglobin A1C: glycated hemoglobin; IABP: intra-aortic balloon pump; LDL: low-density lipoprotein; LVEF: left ventricular ejection fraction; MI: myocardial infarction; PCI: percutaneous coronary intervention; 1G: first-generation; 2G: second-generation.

Post-percutaneous coronary intervention medications

Table 2 shows post-PCI medications between the two groups. Before PSM analysis 1G-DES group was more likely to have received BB, CCB and ACEI after PCI than 2G-DES. But these differences were also disappeared after PSM analysis. The prescription rates of other kinds of medications including aspirin, clopidogrel, cilostazole (Pletaal®, Otsuka Pharmaceutical Co., Tokyo, Japan), ARB, diuretics and lipid lowering agents were similar between two groups before and after PSM analysis.

Table 2.

Post-percutaneous coronary intervention medications.

Variables	Entire patients			Propensity-matched patients
	1G-DESs (n = 554)	2G-DESs (n = 462)	P-value	1G-DESs (n = 232)	2G-DESs (n = 232)	P-value
Aspirin	504 (91.0%)	427 (92.4%)	0.406	214 (92.2%)	213 (91.8%)	0.864
Clopidogrel	505 (91.2%)	416 (90.0%)	0.544	217 (93.5%)	210 (90.5%)	0.230
Cilostazol	138 (24.9%)	102 (22.1%)	0.290	57 (24.6%)	45 (19.4%)	0.179
Beta blockers	291 (52.5%)	311 (67.3%)	< 0.001	134 (57.8%)	144 (62.1%)	0.343
Calcium channel blockers	203 (36.3%)	119 (25.8%)	< 0.001	70 (30.2%)	65 (28.0%)	0.609
ACEIs	179 (32.3%)	211 (45.7%)	< 0.001	88 (37.9%)	94 (40.5%)	0.588
ARBs	178 (32.1%)	126 (27.3%)	0.092	70 (30.2%)	73 (31.5%)	0.763
Diuretics	126 (22.7%)	99 (21.4%)	0.615	54 (23.3%)	57 (24.6%)	0.744
Lipid lowering agents	469 (84.7%)	401 (86.8%)	0.333	199 (85.8%)	201 (86.6%)	0.788

Values are mean ± SD or n (%). The P value for categorical data from chi-square test. ACEIs: angiotensin converting enzyme inhibitors; ARBs: angiotensin receptor blockers; DESs: drug-eluting stents; 1G: first-generation; 2G: second-generation.

Clinical outcomes

Clinical outcomes at 30 d, one year, 3 years and 5 years shown in Table 3. During one month, the incidence of MACE was not significantly different between the two groups. At 1 year after the index PCI, and the incidence of TLR and TVR was significantly higher in the 1G-DESs group compared with the 2G-DESs group before PSM (TLR: 8.8% vs. 3.2%, P < 0.001; TVR: 10.1% vs. 3.7%, P < 0.001) and after PSM analysis (TLR: 9.1% vs. 3.4%, P = 0.020; TVR: 9.5% vs. 3.9%, P = 0.024). At 5 years, the cumulative incidences of TLR [hazard ratio (HR): 3.133; 95% confidence intervals (CI): 1.596–6.376; P = 0.002] and TVR (HR: 3.144; 95% CI: 1.596–6.192; P = 0.001) were significantly higher in the 1G-DESs group compared with the 2G-DESs group after PSM analysis during 5-year follow-up period (Table 4). Figure 2 shows Kaplan-Meier curved analysis of MACE Free survival, TLR, TVR and ST at 5 years according to the generation of DESs (1G vs. 2G) and the types of DESs (SES vs. PES vs. 1G-ZES vs. 2G-ZES vs. EES). Figure 3 shows subgroup analysis for MACE and ST up to 5 years. Although, in cases of male, STEMI, non-small vessel disease (≥ 2.5 mm), the choice of 2G-DESs may be prefer rather than 1G-DESs to reduce MACE during PCI for AMI patients.

Table 3.

Clinical outcomes at 30 days, 1 year, 3 years and 5 years.

Outcomes	Entire patients				Propensity-matched patients
	Total (n = 1016)	1G-DESs (n = 554)	2G-DESs (n = 462)	P-value	1G-DESs (n = 232)	2G-DESs (n = 232)	P-value
30 d follow-up
All death	40 (3.9%)	19 (3.4%)	21 (4.5%)	0.419	7 (3.0%)	12 (5.2%)	0.349
Cardiac death	39 (3.8%)	18 (3.2%)	21 (4.5%)	0.326	6 (2.6%)	12 (5.2%)	0.229
Non-fatal MI	40 (3.9%)	21 (3.8%)	19 (4.1%)	0.872	6 (2.6%)	11 (4.7%)	0.323
Total revascularization	13 (1.3%)	7 (1.3%)	6 (1.3%)	0.960	2 (0.9%)	2 (0.9%)	1.000
TLR	13 (1.3%)	7 (1.3%)	6 (1.3%)	0.960	2 (0.9%)	2 (0.9%)	1.000
TVR	13 (1.3%)	7 (1.3%)	6 (1.3%)	0.960	2 (0.9%)	2 (0.9%)	1.000
Non-TVR	0 (0.0%)	0 (0.0%)	0 (0.0%)	-	0 (0.0%)	0 (0.0%)	-
MACEs	54 (5.3%)	28 (5.1%)	26 (5.6%)	0.779	10 (4.3%)	14 (6.0%)	0.530
Stent thrombosis (definite, probable)
Acute	7 (0.7%)	5 (0.9%)	2 (0.4%)	0.465	1 (0.4%)	0 (0.0%)	0.317
Subacute	7 (0.7%)	3 (0.5%)	4 (0.9%)	0.708	1 (0.4%)	2 (0.9%)	0.562
Total	14 (1.4%)	8 (1.4%)	6 (1.3%)	0.843	2 (0.9%)	2 (0.9%)	1.000
1-yr follow-up
All death	64 (6.3%)	38 (6.9%)	26 (5.6%)	0.440	14 (6.0%)	16 (6.9%)	0.851
Cardiac death)	52 (5.1%)	28 (5.2%)	24 (5.1%)	0.919	10 (4.3%)	15 (6.5%)	0.411
Non-fatal MI	55 (5.4%)	31 (5.6%)	24 (5.2%)	0.889	9 (3.9%)	12 (5.2%)	0.656
Total revascularization	100 (9.8%)	73 (13.2%)	27 (5.8%)	< 0.001	25 (10.8%)	17 (7.3%)	0.257
TLR	64 (6.3%)	49 (8.8%)	15 (3.2%)	< 0.001	21 (9.1%)	8 (3.4%)	0.020
TVR	73 (7.2%)	56 (10.1%)	17 (3.7%)	< 0.001	22 (9.5%)	9 (3.9%)	0.024
Non-TVR	10 (1.0%)	6 (1.1%)	4 (0.9%)	0.767	1 (0.4%)	3 (1.3%)	0.623
MACEs	163 (16.0%)	110 (19.9%)	53 (11.5%)	< 0.001	38 (16.4%)	33 (14.2%)	0.606
Stent thrombosis (definite, probable)
Late (31–365 d)	6 (0.6%)	6 (1.1%)	0 (0.0%)	0.035	2 (0.9%)	0 (0.0%)	0.156
Total (1–365 d)	20 (2.0%)	14 (2.5%)	6 (1.3%)	0.180	4 (1.7%)	2 (0.9%)	0.685
3-yr follow-up
All death	83 (8.2%)	53 (9.6%)	30 (6.5%)	0.085	21 (9.1%)	19 (8.2%)	0.869
Cardiac death	63 (6.2%)	37 (6.7%)	26 (5.6%)	0.516	12 (5.2%)	16 (6.9%)	0.559
Non-fatal MI	73 (7.2%)	42 (7.6%)	31 (6.7%)	0.627	14 (6.0%)	14 (6.0%)	1.000
Total revascularization	125 (12.3%)	85 (15.3%)	40 (8.7%)	0.001	30 (12.9%)	21 (9.1%)	0.235
TLR	79 (7.8%)	58 (10.5%)	21 (4.5%)	< 0.001	25 (10.8%)	10 (4.3%)	0.013
TVR	93 (9.2%)	68 (12.3%)	25 (5.4%)	< 0.001	27 (11.6%)	11 (4.7%)	0.010
Non-TVR	14 (1.4%)	9 (1.6%)	5 (1.1%)	0.592	2 (0.9%)	3 (1.3%)	0.653
MACEs	205 (20.2%)	134 (24.2%)	71 (15.4%)	< 0.001	50 (21.6%)	41 (17.7%)	0.350
Stent thrombosis (definite, probable)
Very late (366–1095 d)	8 (0.8%)	6 (1.1%)	2 (0.4%)	0.303	2 (0.9%)	0 (0.0%)	0.156
Total (1–1095 d)	28 (2.8%)	20 (3.6%)	8 (1.7%)	0.083	6 (2.6%)	2 (0.9%)	0.285
5-yr follow-up
All death	90 (8.9%)	57 (10.3%)	33 (7.1%)	0.096	23 (9.9%)	21 (9.1%)	0.874
Cardiac death	67 (6.6%)	38 (6.9%)	29 (6.3%)	0.800	12 (5.2%)	18 (7.8%)	0.345
Non-fatal MI	91 (9.0%)	55 (9.9%)	36 (7.8%)	0.270	20 (8.6%)	16 (6.9%)	0.603
Total revascularization	148 (14.6%)	102 (18.4%)	46 (10.0%)	< 0.001	39 (16.8%)	21 (9.1%)	0.018
TLR	97 (9.5%)	74 (13.4%)	23 (5.0%)	< 0.001	32 (13.8%)	10 (4.3%)	0.001
TVR	112 (11.0%)	84 (15.2%)	28 (6.1%)	< 0.001	35 (15.1%)	11 (4.7%)	< 0.001
Non-TVR	17 (1.7%)	10 (1.8%)	7 (1.5%)	0.809	3 (1.3%)	3 (1.3%)	1.000
MACEs	236 (23.2%)	155 (28.0%)	81 (17.5%)	< 0.001	61 (26.3%)	43 (18.8%)	0.058
Stent thrombosis (definite, probable)
Very late (1096–1825 d)	7 (0.7%)	6 (1.1%)	1 (0.2%)	0.134	1 (0.4%)	0 (0.0%)	0.317
Very late (366–1825 d)	15 (1.5%)	12 (2.2%)	3 (0.6%)	0.065	3 (1.3%)	0 (0.0%)	0.248
Total (1–1825 d)	35 (3.4%)	26 (4.7%)	9 (1.9%)	0.023	7 (3.0%)	2 (0.9%)	0.175

Values are presented as n (%). The P value for categorical data from chi-square test. DESs: drug-eluting stents; MI: myocardial infarction; MACE: major adverse cardiac events; TLR: target lesion revascularization; TVR: target vessel revascularization; 1G: first-generation; 2G: second-generation.

Table 4.

Five-year clinical outcomes by Kaplan-Meier curved analysis and cox-proportional hazard ratio model analysis.

Outcomes	Cumulative Events at 5 years, %			Hazard Ratio (95% CI)	P-value
	1G-DESs	2G-DESs	Log Rank
Entire patients
Primary outcomes
All death	57 (10.3%)	33 (7.1%)	0.089	1.447 (0.942–2.222)	0.091
Cardiac death	38 (6.9%)	29 (6.3%)	0.715	1.094 (0.675–1.774)	0.715
Non-fatal myocardial infarction	55 (10.2%)	36 (8.1%)	0.317	1.239 (0.813–1.886)	0.319
Total revascularization	102 (19.6%)	46 (10.8%)	< 0.001	1.908 (1.347–2.703)	< 0.001
Target lesion revascularization	74 (14.7%)	23 (5.4%)	< 0.001	2.744 (1.718–4.382)	< 0.001
Target vessel revascularization	84 (16.5%)	28 (6.8%)	< 0.001	2.566 (1.673–3.937)	< 0.001
Non-target vessel revascularization	10 (1.9%)	7 (1.8%)	0.744	1.175 (0.447–3.090)	0.744
MACEs	155 (28.0%)	81 (18.1%)	< 0.001	1.639 (1.253–2.145)	< 0.001
Secondary outcome
Stent thrombosis	26 (4.7%)	9 (1.9%)	0.018	2.432 (1.140–5.190)	0.022
Propensity-matched patients
Primary outcomes
All death	23 (9.9%)	21 (9.1%)	0.783	1.087 (0.601–1.964)	0.783
Cardiac death	12 (5.2%)	18 (7.8%)	0.261	0.660 (0.318–1.370)	0.265
Non-fatal myocardial infarction	20 (8.9%)	16 (7.3%)	0.608	1.187 (0.615–2.294)	0.609
Total revascularization	39 (17.8%)	21 (9.7%)	0.021	1.874 (1.086–3.140)	0.023
Target lesion revascularization	32 (14.9%)	10 (4.7%)	0.001	3.133 (1.539–6.376)	0.002
Target vessel revascularization	35 (16.2%)	11 (5.2%)	< 0.001	3.144 (1.596–6.192)	0.001
Non-target vessel revascularization	3 (1.4%)	3 (1.4%)	0.990	0.990 (0.200–4.906)	0.990
MACEs	61 (26.3%)	43 (18.8%)	0.077	1.419 (0.960–2.096)	0.079
Secondary outcome
Stent thrombosis	7 (3.0%)	2 (0.9%)	0.094	3.524 (0.732–16.96)	0.116

Values are presented as n (%). DESs: drug-eluting stents; MACEs: major adverse cardiac events;1G: first-generation; 2G: second-generation.

Figure 2.

Kaplan-Meier curved analysis of MACEs-free survival (A, B), TLR (C, D), TVR (E, F) and stent thrombosis (G, H) at 5-year according to the generation of DES (1G vs. 2G) and types of DES (SES vs. PES vs. 1G-ZES vs. 2G-ZES vs. EES).

DES: drug eluting balloon; EES: everolimus-eluting stent; MACEs: major adverse cardiac events; PES: paclitaxel-eluting stent; SES: sirolimus-eluting stent; TLR: target lesion revascularization; TVR: target vessel revascularization; ZES: zotarolimus-eluting stent; 1G: first-generation; 2G: second-generation.

Figure 3.

Subgroup analyses for MACE.

ACC/AHA: American college of cardiology/American heart association; DESs: drug-eluting stents; LVEF: left ventricular ejection fraction; MACE: major adverse cardiac events; RAF: routine angiographic; STEMI: ST-segment elevation myocardial infarction; 1G: first-generation; 2G: second-generation.

Discussion

The main finding of this “real-world” PSM analysis was that the cumulative incidence of TLR, TVR and total revascularization rates were significantly higher in the 1G-DES group compared with 2G-DES group in AMI vessels after PSM analysis during 5-year follow-up periods. However, the incidence of total death, non-fatal MI and ST were similar between the two groups.

1G-DES (SES) could significantly reduce revascularization rates compared with BMS among patients with STEMI undergoing primary PCI.[9] However, 1G-DES can cause late ST due to delayed re-endothelization and poor strut coverage.[10] To overcome these limitations stent platforms and polymers have rapidly evolved during a short period. Newer antiproliferative drugs and more biocompatible polymers have been adapted in reducing the rate of late ST in stable coronary artery disease.[11]

Comparison with randomized controlled study

Until recently, five randomized control trials (RCTs) comparing clinical outcomes of the 1G-DES versus 2G-DES in patients with AMI were conducted.[4],[12]–[15] However, their follow-up periods were less than 5 years (e.g., 7 months, 12 months, 18 months and 3 years). Among these 5 RCTs, 3 trials had compared ZES versus 1G-DES (two for ZES vs. SES vs. PES[4],[12]); one for ZES vs. SES[14]) and other 2 trials compared EES and SES.[13],[15] Four trials had evaluated only STEMI patients and one trial included 96% subjects with STEMI and 4% with non-ST segment elevation MI (NSTEMI)[13] and the number of included patients of these 5 RCTs ranged from 35 patients to 625 patients. Compared with these 5 RCTs, our study included relatively large number of NSTEMI patients (n = 482), and the total number of STEMI patients were up to 534. This means that our study might have included meaningful number of STEMI vessels. A meta-analysis report of above 5 RCTs[16] showed that 2G-DES could not showed a significant advantage over the 1G-DES in lowering the incidence of TLR, MACE, or all-cause death, only EES seemed to lower the occurrence of MACE than the 1G-DES. Although above 5 RCTs demonstrated useful clinical outcomes between 1G-DES and 2G-DES, more large scaled, randomized long-term follow-up studies are needed due to their relatively small study populations and short-term follow-up periods. In our study, 5-year cumulative incidence of MACE-free survival was significantly lower in the 1G-DES compared with 2G-DES in the entire patients (81.9% vs. 72.0%, Log-Rank: P < 0.001, Figure 2A). The main causes of this differences were caused by the difference between SES and 2G-ZES (Log-Rank: P = 0.007), SES and EES (Log-Rank: P = 0.001), PES and 2G-ZES (Log-Rank: P = 0.020), and PES and EES (Log-Rank: P = 0.002). Although the total cumulative incidence of MACE-free survival was not significantly different in PSM patients, EES's beneficial effect for MACE-free survival was sustained only between PES and EES (Log-Rank: P = 0.033, Figure 2B).

Comparison with non-randomized controlled study

In non-randomized studies, Chen, et al.[3] analyzed 2911 AMI patients receiving PESs (n = 1210) or EESs (n = 1701). In his multicenter registry (Korea Acute Myocardial Infarction Registry, KAMIR) data, EESs group had significantly lower incidence of Re-MI (2.8% vs. 1.4%, P = 0.002), TLR (3.1% vs. 1.8%, P < 0.001) and probable or definite ST (1.8% vs. 0.3%, P < 0.001) than the PESs group during 1 year follow-up period. In our study, as shown in Table 3, Figure 2C and Figure 2D, the incidences of TLR were significantly higher in 1G-DES than 2G-DES after PSM (9.1% vs. 3.4%, P = 0.020) but Re-MI (3.9% vs. 5.2%, P = 0.656) and ST were 1.7% vs. 0.9%, P = 0.685) not significantly different during 1 year follow-up period. At 5 years, the cumulative incidence of TLR was significantly different between PES and EES (19.1% vs. 6.6%, Log-Rank: P = 0.013), PES and 2G-ZES (19.1% vs. 3.1%, Log-Rank: P = 0.001), SES and 2G-ZES (14.4% vs. 3.1%, Log-Rank: P = 0.014), and 1G-ZES and 2G-ZES (14.4% vs. 3.1%, P = 0.009). The 5-year cumulative incidence of TVR was significantly higher in 1G-DES than 2G-DES in entire patients (16.5% vs. 6.8%, Log-Rank: P < 0.001, Table 4) and PSM patients (16.2% vs. 5.2%, Log-Rank: P = 0.001, Table 4, Figures 2E and 2F). In this study, the comparison between the 2G-DESs (2G-ZES and EES) showed that 2G-ZES was non-inferior to EES in 5-year long-term clinical outcomes (MACE-free survival, ST, TLR, TVR, Figures 2A to 2G) and as similar as previous studies.[6],[17]–[20] By contrast, there were several other different results on this comparison. Chen, et al.[21] also had reported another study comparing ZES and EES. They analyzed 3309 AMI patients treated with ZES (n = 1608) or EES (n = 1701) from KAMIR. After PSM analysis, EES significantly lowered the incidence of target lesion failure (TLF, 6.5% vs. 8.7%, P = 0.029) and probable or definite ST (0.3% vs. 1.6%, P < 0.001) compared with ZES. We think one of important factors determine this different result is the difference in the follow-up period and Chen, et al. compared the EES with mainly with original ZES, not the 2G-ZES.

Comparison between 1G-ZES and 2G-ZES

Characteristically ZES group divided separately as 1G-ZES and 2G-ZES and enrolled as 1G-DESs group or 2G-DESs group in this study. We think that even though endeavor® or endeavor sprint® (early-generation ZES) have relatively thinner stent strut and biocompatible polymer compared with SES or PES, in real-world practice they may be considered as 1G-DES and later-generation ZES (endeavor resolute®, resolute integrity® and resolute onyx®) are really considered as 2G-DESs. There were also limited comparisons of long-term clinical outcomes in patients with AMI after PCI between 1G-ZES and 2G-ZES. The total MACE was not significantly different between these two stents (Log-Rank: P = 0.039 in entire patents, Figure 2A and Log-Rank: P = 0.466 in PSM patients, Figure 2B) during 5-year follow-up period in our study. However, in addition to increased TLR rate in 1G-ZES compared with 2G-ZES (as shown previously), TVR rate was also higher in 1G-ZES group compared with 2G-ZES (14.1% vs. 4.1%, Log-Rank: P = 0.009, Figure 2F). However, the rate of stent thrombosis was not significantly different between the two groups (5.9% vs. 0.9%, Log-Rank: P = 0.054). Considering these several clinical outcomes, 1G-ZES showed similar clinical outcome patterns of other 1G-DESs (SES and PES).

Stent thrombosis

Stent thrombosis is another debatable issue in the DES era. Several meta-analyses comparing SESs with PESs demonstrated no significant difference in ST between these two types of stents.[22],[23] But Schömig, et al.[24] showed that SESs was better than PESs in terms of reducing ST in his meta-analysis report. Previously mentioned 5 RCTs studies demonstrated that 2G-DES showed decreased incidence of definite or probable stent thrombosis compared with 1G-DES (RR: 0.53; 95% CI: 0.25–1.13; P = 0.10).[16] Hofma, et al.[7] reported the 3-year results of their all-comer, randomized, multicenter AMI trial (XAMI trial, Xience V stent vs. Cypher stent in Primary PCI for AMI trial). According to this report, the incidence of definite/probable ST was similar between the two groups (EES: 2.3% vs. SES: 3.2%, P = 0.60). In our study, the total and individual cumulative incidence of ST were not significantly different between 1G-DES and 2G-DES during 5-year follow-up periods in PSM patients (Figure 2G and Figure 2H). If the sample size were fully larger than our study, there may be statistically significant differences may be exist between 1G-ZES and 2G-ZES or EES.

Others

The RAF can increase revascularization due to possible ‘oculo-stenotic reflex’, a term describing revascularization with PCI according to anatomic lesion severity regardless of clinical or physiologic evidence of ischemia.[25] In our study, RAF was considered as an important bias and included covariate of PSM. After PSM this bias was abolished.

Until today, very long-term clinical follow-up data comparing 1G-DES vs. 2G-DES are rare and debatable especially in patients with AMI. The purpose of this study was to investigate the efficacy and safety of 2G-DES over 1G-DES regardless of specific types of DES in AMI patients during very long-term clinical follow-up periods. Most of the previous studies reported comparative results of only two-types or three-types of DESs, not considered each generation group as a whole.[3]–[5],[8],[11]–[15],[22]–[24] In addition to comparative analysis of 1G-DES vs. 2G-DES as a whole, we could obtain comparative subgroup results between different types of DESs through Kaplan-Meier curved analysis or Cox-regression analysis. During 5-year follow-up periods, we suggested 2G-DES's superiority for reducing TLR, TVR and total revascularization rates in AMI patients. However, this result maybe more precisely be defined by any other large and long-term follow-up randomized and controlled trials in the future. Under the circumstances where very long-term major clinical outcomes between 1G-DES and 2G-DES are still debatable, our results can be provide useful clinical outcome information and trends between 1G-DES and 2G-DES to some extent during very long-term follow-up periods in the DES era.

This study has some limitations. First, because it is a non-randomized registry design and single center study, several confounding factors such as under-reporting and/or missing value and selection bias may have affect the end results. Second, although PSM analysis and subgroup analysis done, the proportions of each stents in both groups were not evenly distributed. Third, in our study AMI patients were consisted of STEMI and NSTEMI. This heterogeneity can affect each other and may act as bias. Fourth, the strategy of antiplatelet therapy [e.g., DAPT or triple antiplatelet therapy (TAPT)] was left to the physician's discretion, which may have influenced the major clinical outcomes. Lastly, several important factors that can determine the end results such as total ischemia time could not be precisely obtained in our study patients.

In conclusion, in our single-center, all-comer registry, 2G-DES's superiority in reducing TLR, TVR and total revascularization rate over 1G-DES in AMI patients suggested during 5-year follow-up periods. Special cautions with careful clinical follow-up would be necessary for the AMI patients who are treating with 1G DESs in this aspect. However, long-term follow-up data from large scale, randomized studies are necessary to confirm these results.