Comparison of Clinical Outcomes Between Incomplete and Complete Revascularization in NSTEMI Patients With Multivessel Disease

Abstract

Background

The optimal revascularization strategy for patients with non–ST-segment elevation myocardial infarction (NSTEMI) and multivessel disease remains controversial.

Objectives

This study aimed to compare the 3-year clinical outcomes of incomplete revascularization (IR) vs complete revascularization (CR) in NSTEMI patients with multivessel disease.

Methods

Among 29,625 patients enrolled in the nationwide pooled registry of the KAMIR-NIH (Korea Acute Myocardial Infarction Registry–National Institutes of Health) and KAMIR-V, 6,987 NSTEMI patients with multivessel disease who underwent percutaneous coronary intervention were analyzed. CR was defined as residual stenosis <30% in all vessels ≥2.25 mm with TIMI flow grade 3. The primary endpoint was major adverse cardiac events at 3 years, a composite of all-cause death, recurrent myocardial infarction, unplanned repeat revascularization, and hospitalization for heart failure.

Results

Of 6,987 patients, 3,072 underwent CR and 3,915 IR. The majority of patients underwent percutaneous coronary intervention for non–infarct-related artery during index procedure (6,144 of 6,987; 88.0%). CR was associated with a lower 3-year risk of major adverse cardiac events compared with IR (18.6% vs 26.9%; adjusted HR: 0.75; 95% CI: 0.66-0.85; P < 0.001), driven by reductions in cardiac death or recurrent myocardial infarction (6.9% vs 10.4%; adjusted HR: 0.76; 95% CI: 0.60-0.95; P = 0.015) and unplanned repeat revascularization (8.6% vs 13.1%; adjusted HR: 0.68; 95% CI: 0.56-0.82; P < 0.001). Findings were consistent after adjustment using propensity score matching and inverse probability weighting.

Conclusions

In NSTEMI patients with multivessel disease, CR was associated with significantly improved 3-year clinical outcomes compared with IR.

Central Illustration

Non–ST-segment elevation myocardial infarction (NSTEMI) represents a substantial proportion of acute coronary syndrome and is frequently associated with multivessel coronary artery disease (MVD), observed in approximately half of patients undergoing coronary angiography.^1,2 The presence of MVD in NSTEMI is associated with increased risk of recurrent myocardial infarction (MI), heart failure, and mortality.3, 4, 5

Although percutaneous coronary intervention (PCI) with drug-eluting stent remains a cornerstone of treatment for NSTEMI patients, the optimal revascularization strategy in the setting of MVD— complete revascularization (CR) vs incomplete revascularization (IR)—remains controversial.6, 7, 8 In patients with ST-segment elevation myocardial infarction (STEMI) and MVD, multiple randomized controlled trials and observational studies have demonstrated that CR significantly reduces adverse cardiovascular events compared with infarct related artery (IRA)–only PCI.9, 10, 11, 12 In contrast, there is no dedicated trial comparing CR vs IRA-only PCI in patients with NSTEMI and current European Society of Cardiology guidelines recommend CR in patients with NSTEMI and MVD as a Class IIa (Level of Evidence: C) recommendation based on observational data and expert opinion in the absence of robust evidence from a randomized controlled trial.⁹

Several observational studies have addressed this issue in NSTEMI and MVD; however, findings have been inconsistent, and long-term follow-up data remain limited.^5,13, 14, 15 As such, large-scale, real-world evidence with extended follow-up is needed to inform optimal revascularization strategies in this population. Therefore, the current study aimed to compare 3-year clinical outcomes between IR and CR in NSTEMI patients with MVD using data from a nationwide registry.

Methods

Study protocols and population

The current study analyzed pooled data from the KAMIR-NIH (Korean Acute Myocardial Infarction Registry–National Institutes of Health) and the KAMIR-V registries, which are nationwide, multicenter, prospective registries that consecutively enrolled patients with acute MI from 35 tertiary hospitals in Korea. The enrollment periods were from April 2012 to December 2015 for KAMIR-NIH and from January 2016 to June 2020 for KAMIR-V. Detailed study protocols have been previously described.¹⁶ The protocols of both registries were approved by the institutional review boards of all participating hospitals and were conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all patients before enrollment.

Among the 29,625 patients enrolled, those diagnosed with NSTEMI and angiographically confirmed MVD, and who underwent PCI during the index hospitalization were included. NSTEMI was defined as acute myocardial infarction without newly detected ST-segment elevation ≥0.1 mV in ≥2 contiguous leads or new left bundle branch block on a 12-lead electrocardiogram, accompanied by elevation of at least 1 cardiac biomarker above the upper reference limit.¹⁷ The presence of MVD was defined as the presence of at least 1 additional angiographically significant lesion (≥50% diameter stenosis) in non–infarct-related artery (non-IRA) with a diameter ≥2.25 mm, or in the left main coronary artery.10, 11, 12 CR was defined anatomically as a visually estimated residual stenosis <30% with TIMI flow grade 3 in all coronary arteries ≥2.25 mm in diameter on post-PCI angiography.⁷ IR was defined as the failure to achieve CR in 1 or more coronary arteries ≥2.25 mm in diameter, as evidenced by the presence of residual lesions with ≥30% stenosis and/or TIMI flow grade <3 after PCI, irrespective of whether non-IRA PCI was attempted. All non-IRA PCI procedures contributing to the determination of the revascularization strategy were performed during the index hospitalization, either during the index procedure or as a staged PCI before discharge. Elective staged PCI of non-IRA lesions performed after hospital discharge was not considered in defining CR. Patients were excluded if they had a diagnosis other than NSTEMI (eg, STEMI, unstable angina, or other; n = 14,333), single-vessel disease (n = 7,072), no PCI performed for the IRA (n = 960), suboptimal or incomplete PCI of the IRA (n = 86), or lacked 3-year follow-up data (n = 187). As a result, a total of 6,987 patients were analyzed and classified into either the CR or IR group based on revascularization strategy (Figure 1).

Figure 1.

Study Flow

The study population was derived from pooled data from the nationwide, multicenter, prospective KAMIR-NIH and KAMIR-V registries. IRA = infarct-related artery; PCI = percutaneous coronary intervention; STEMI = ST-segment elevation myocardial infarction; UA = unstable angina.

Patient management, data collection, and follow-up

Patients were treated according to current guidelines and coronary interventions were performed using standard techniques. The choice of treatment strategy such as the type of stents, use of intravascular imaging, thrombus aspiration, mechanical circulatory support, and pharmacologic therapies were at the operator’s discretion. PCI was considered successful if the final residual stenosis was <30% with TIMI flow grade 3. After the index procedure, all patients were prescribed aspirin indefinitely and P2Y₁₂ inhibitor (clopidogrel, prasugrel, or ticagrelor) for at least 12 months, unless contraindicated. The choice of P2Y₁₂ inhibitor was guided by clinical judgment and patient-specific bleeding risk. Additional medications including renin-angiotensin-aldosterone system blockades, beta-blockers, and statins were also prescribed based on practice guidelines.^6,9

Demographic features and cardiovascular risk factors were obtained through structured interviews. Angiographic and procedural data, complications, and discharge medications were recorded during hospitalization. After discharge, patients were followed at 6, 12, 24, and 36 months through outpatient clinic visits or telephone contact, and whenever any clinical event occurred. Data was collected by independent clinical research coordinators using a web-based case report form within the Internet-based Clinical Research and Trial management system (iCReaT), a data management system established by the Centers for Disease Control and Prevention, Ministry of Health and Welfare, Republic of Korea (iCReaT Study No. C110016).

Study outcomes

The primary endpoint was a major adverse cardiac event (MACE) at 3 years, a composite of all-cause death, recurrent MI, unplanned repeat revascularization, and admission for heart failure. Secondary endpoints included the composite of cardiac death or recurrent MI and each component of the primary endpoint. All-cause death was considered cardiac in origin unless an unequivocal noncardiac cause was identified.¹⁸ Recurrent MI was defined by recurrence of ischemic symptoms or electrocardiographic changes with an increase in cardiac biomarkers above the upper reference limit; periprocedural MI was not considered a clinical event.¹⁹ Unplanned repeat revascularization was defined as any clinically driven revascularization after discharge from the index hospitalization, according to the Academic Research Consortium definitions.¹⁸ Staged PCI for non-IRA lesions performed during the index hospitalization was not classified as an unplanned repeat revascularization. Unplanned repeat revascularization was classified as infarct-related lesion revascularization, IRA revascularization, and non-IRA revascularization. Admission for heart failure was defined as first hospitalization for heart failure. Hospitalization for heart failure should include all the following criteria: 1) hospitalization with primary diagnosis of heart failure; 2) duration of hospitalization of at least 24 hours; 3) new or worsening symptoms of heart failure; 4) objective evidence of new or worsening heart failure on physical examination or laboratory findings; and 5) initiation or intensification of heart failure treatment. Cardiogenic shock was defined as Killip class IV at initial presentation and/or the use of mechanical circulatory support, including extracorporeal membrane oxygenation or intra-aortic balloon pump, during the index hospitalization.

Statistical analysis

Categorical variables were presented as numbers and relative frequencies (percentages) and were compared using the chi-square test. Continuous variables were expressed as mean ± SD or median (IQR), depending on distribution, and compared using the independent samples Student’s t test or Mann-Whitney U test, as appropriate. Kaplan-Meier survival curves were used to estimate event rates, and the 2 groups were compared using the log-rank test.

To adjust for baseline differences as much as possible, multiple sensitivity analyses were conducted. First, multivariable Cox proportional hazards models were constructed using covariates that were either significantly imbalanced or clinically relevant. Covariates included age, sex, body mass index, Killip class at initial presentation, systolic blood pressure, hypertension, diabetes, previous MI, previous cerebrovascular accident, previous congestive heart failure, hemoglobin, estimated glomerular filtration rate, peak creatine kinase–myocardial band, left ventricular ejection fraction, use of clopidogrel, potent P2Y₁₂ inhibitor, angiographic disease extent, left main artery disease, culprit lesion location, American College of Cardiology/American Heart Association type B2/C lesion, timing of non-IRA PCI, post-PCI TIMI flow of the IRA, treatment modalities for the IRA, thrombus aspiration, total stent length of the IRA, transradial approach, intravascular imaging, and mechanical circulatory support. The assumption of proportionality was assessed graphically by the log-minus-log plot and by a 2-sided score test of the scaled Schoenfeld residuals over time at the 0.05 level. Cox proportional hazard models for all clinical outcomes satisfied the proportional hazards assumption.

Second, analyses using propensity score matching and inverse probability weighting (IPW) were performed.²⁰ A multivariable logistic regression model was used to generate propensity scores that indicate the probability of receiving CR. All the available covariates were included in this model, precisely following the recommendations of analysis using propensity score.²¹ Propensity score matching was conducted using 1:1 matching process without replacement and a caliper width of 0.2 SDs, resulting in 2,604 matched pairs. For the IPW adjustment, inverse of propensity score was adjusted proportional hazard regression model. Percent standardized mean differences after propensity score matching or IPW adjustment were within ±10% across all matched covariates, demonstrating successful balance achievement between comparative groups (Supplemental Table 1). To identify independent association of CR with the risk of MACE and cardiac death or recurrent MI, we used multivariable Cox proportional hazard model. Model discrimination was assessed using Harrell C statistics with 95% CIs. Subgroup analyses were performed to evaluate treatment effects across key clinical and procedural characteristics, with interaction terms tested in the Cox model. For nonfatal outcomes, competing risk analyses were additionally performed using Fine-Gray subdistribution hazard models, treating all-cause death as a competing event. The proportion of missing data among variables included in the statistical analyses was minimal, and analyses were performed using a complete-case approach without imputation.

All probability values were 2-sided, and P values <0.05 were considered statistically significant. Statistical analyses were performed using R version 4.4.0 (R Foundation for Statistical Computing).

Results

Baseline characteristics

Baseline clinical characteristics are summarized in Table 1. Among the overall population, 3,915 patients underwent IR, whereas 3,072 patients underwent CR. The mean age of the entire cohort was 66.4 ± 11.8 years, 25.0% of patients were aged >75 years, and 72.6% were male. Patients in the IR group were older, had more comorbidities, including diabetes mellitus, and a previous history of MI, cerebrovascular accident, or congestive heart failure, and were more frequently presented with Killip class III or IV than the CR group. Patients in the IR group had higher N-terminal pro–B-type natriuretic peptide levels and lower left ventricular ejection fraction than the CR group.

Table 1.

Baseline Clinical Characteristics

	Total Population (N = 6,987)	IR (n = 3,915)	CR (n = 3,072)	P Value
Age, y	66.4 ± 11.8	66.9 ± 11.9	65.7 ± 11.6	<0.001
Age >75 y	1,745 (25.0)	1,054 (26.9)	691 (22.5)	<0.001
Male	5,073 (72.6)	2,822 (72.1)	2,251 (73.3)	0.279
BMI, kg/m2	24.1 ± 3.3	24.1 ± 3.3	24.2 ± 3.3	0.039
Killip class at initial presentation				<0.001
Class I or II	6,106 (89.2)	3,367 (87.4)	2,739 (91.5)
Class III or IV	740 (10.8)	487 (12.6)	253 (8.5)
Hemodynamic data
Systolic blood pressure, mm Hg	135.7 ± 27.5	136.6 ± 27.8	134.6 ± 27.2	0.003
Diastolic blood pressure, mm Hg	80.3 ± 16.3	80.5 ± 16.5	80.1 ± 16.0	0.346
Heart rate, beats/min	81.0 ± 18.9	81.3 ± 19.1	80.7 ± 18.5	0.150
Symptom-to-door time, h	5.4 (2.0-18.5)	5.2 (1.9-16.9)	5.6 (2.0-20.6)	0.076
Door-to-balloon time, h	13.6 (3.7-25.2)	13.5 (3.5-25.6)	13.8 (4.0-24.9)	0.417
Cardiovascular risk factors
Hypertension	4,031 (57.7)	2,339 (59.7)	1,692 (55.1)	<0.001
Diabetes	2,483 (35.5)	1,459 (37.3)	1,024 (33.3)	0.001
Dyslipidemia	1,032 (14.8)	597 (15.2)	435 (14.2)	0.215
Current smoking	2,214 (32.7)	1,242 (32.7)	972 (32.6)	0.929
End-stage renal disease	291 (4.2)	179 (4.6)	112 (3.7)	0.061
Previous MI	621 (8.9)	373 (9.5)	248 (8.1)	0.038
Previous CVA	602 (8.7)	389 (10.0)	213 (7.0)	<0.001
Previous CHF	145 (2.1)	103 (2.6)	42 (1.4)	<0.001
Laboratory data
Hemoglobin, g/dL	13.4 ± 2.2	13.3 ± 2.2	13.5 ± 2.2	<0.001
Estimated GFR, mL/min/1.73 m2	85.3 ± 30.8	82.9 ± 31.6	88.3 ± 29.6	<0.001
LDL-cholesterol, mg/dL	78.3 ± 45.4	78.9 ± 45.3	77.5 ± 45.5	0.203
Glycated hemoglobin, %	6.6 ± 1.5	6.6 ± 1.5	6.6 ± 1.5	0.616
Peak CK-MB, ng/mL	61.9 ± 98.0	64.8 ± 106.4	58.2 ± 85.9	0.004
Peak troponin-I, ng/mL	26.0 ± 125.9	27.4 ± 164.2	24.3 ± 52.7	0.328
NT-proBNP, pg/mL	3,802.2 ± 8,681.2	4,207.0 ± 8,805.1	3,255.5 ± 8,482.9	0.001
Atrial fibrillation	348 (5.0)	200 (5.2)	148 (4.8)	0.588
Left ventricular ejection fraction, %	52.7 ± 11.7	51.9 ± 11.9	53.6 ± 11.4	<0.001
Medications at discharge
Aspirin	6,956 (99.6)	3,899 (99.6)	3,057 (99.5)	0.752
P2Y12 inhibitor	6,948 (99.4)	3,892 (99.4)	3,056 (99.5)	0.834
Clopidogrel	5,013 (71.7)	2,854 (72.9)	2,159 (70.3)	0.017
Potent P2Y12 inhibitor	2,929 (41.9)	1,548 (39.5)	1,381 (45.0)	<0.001
RAAS blocker	5,297 (75.8)	2,942 (75.1)	2,355 (76.7)	0.150
Beta-blocker	5,501 (78.7)	3,065 (78.3)	2,436 (79.3)	0.321
Statin	6,498 (93.0)	3,641 (93.0)	2,857 (93.0)	1.000
Oral anticoagulant	215 (3.1)	134 (3.4)	81 (2.6)	0.069

Values are mean ± SD, n (%), or median (Q1-Q3).

BMI = body mass index; CHF = congestive heart failure; CK-MB = creatine kinase-myocardial band; CR = complete revascularization; CVA = cerebrovascular accident; GFR = glomerular filtration rate; IR = incomplete revascularization; IRA = infarct-related artery; LDL = low-density lipoprotein; MI = myocardial infarction; NT-proBNP = N-terminal pro–B-type natriuretic peptide; PCI = percutaneous coronary intervention; RAAS = renin-angiotensin-aldosterone system.

Regarding discharge medications, potent P2Y₁₂ inhibitors were more frequently used in the CR group (IR vs CR; 39.5% vs 45.0%; P < 0.001), whereas clopidogrel was more commonly prescribed in the IR group (72.9% vs 70.3%; P = 0.017). The use of other medications was similar between the groups.

Baseline lesion- and procedure-related characteristics

Angiographic and procedural characteristics are summarized in Table 2. Two-vessel disease was more common in the CR group, whereas 3-vessel disease was more prevalent in the IR group. The prevalence of left main coronary artery disease was significantly higher in the CR group (10.3% vs 13.5%; P < 0.001). In both groups, the most common culprit lesion was in the left anterior descending artery, followed by the right coronary artery, the left circumflex artery, and the left main artery. PCI for non-IRA lesions was more frequently performed as a staged procedure in the CR group, whereas immediate PCI was more common in the IR group. Post-PCI TIMI flow grade 3 in the IRA was achieved in most patients in both groups. Drug-eluting stents were predominantly used in both groups and total number of implanted stents was higher in the CR group (1.53 ± 0.90 vs 2.27 ± 1.07; P < 0.001). Although mean stent diameter of the IRA did not differ significantly between groups (3.05 ± 0.46 mm vs 3.06 ± 0.49 mm; P = 0.273), the total stent length of the IRA was longer in the IR group (33.32 ± 17.21 mm vs 32.21 ± 16.99 mm; P = 0.009). The use of intravascular imaging was more frequent in the IR group (31.2% vs 28.0%; P = 0.004), as was the use of mechanical circulatory support (2.6% vs 1.8%; P = 0.030).

Table 2.

Baseline Procedural Profiles

	Total Population (N = 6,987)	IR (n = 3,915)	CR (n = 3,072)	P Value
Angiographic disease extent				<0.001
2-vessel disease	4,225 (60.5)	1,924 (49.1)	2,301 (74.9)
3-vessel disease	2,762 (39.5)	1,991 (50.9)	771 (25.1)
Left main artery disease	818 (11.7)	403 (10.3)	415 (13.5)	<0.001
Culprit lesion location				<0.001
Unprotected left main artery	422 (6.0)	173 (4.4)	249 (8.1)
LAD	2,580 (36.9)	1,507 (38.5)	1,073 (34.9)
LCX	1,839 (26.3)	995 (25.4)	844 (27.5)
RCA	2,146 (30.7)	1,240 (31.7)	906 (29.5)
ACC/AHA type B2/C lesiona	5,800 (85.6)	3,230 (85.5)	2,570 (85.8)	0.811
Timing of non-IRA PCI				<0.001
Immediate PCI	6,144 (88.0)	3,585 (91.6)	2,559 (83.3)
Staged PCI at index hospitalization	840 (12.0)	328 (8.4)	512 (16.7)
Pre-PCI TIMI flow of IRA				0.514
TIMI flow grade <3	4,157 (60.4)	2,338 (60.8)	1,819 (60.0)
TIMI flow grade 3	2,725 (39.6)	1,510 (39.2)	1,215 (40.0)
Post-TIMI flow of IRA				0.005
TIMI flow grade <3	114 (1.6)	79 (2.0)	35 (1.1)
TIMI flow grade 3	6,873 (98.4)	3,836 (98.0)	3,037 (98.9)
Treatment modalities for IRA				<0.001
Drug-eluting stent	6,533 (93.5)	3,626 (92.6)	2,907 (94.6)
Bare metal stent	10 (0.1)	2 (0.1)	8 (0.3)
Plain ballon angioplasty	337 (4.8)	216 (5.5)	121 (3.9)
Others	107 (1.5)	71 (1.8)	36 (1.2)
Glycoprotein IIb/IIIa inhibitor	483 (6.9)	254 (6.5)	229 (7.5)	0.125
Thrombus aspiration	589 (8.4)	354 (9.0)	235 (7.6)	0.042
Mean stent diameter of IRA, mm	3.06 ± 0.47	3.05 ± 0.46	3.06 ± 0.49	0.273
Total stent length of IRA, mm	32.83 ± 17.12	33.32 ± 17.21	32.21 ± 16.99	0.009
Total number of implanted stents in IRA	1.23 ± 0.54	1.23 ± 0.56	1.23 ± 0.53	0.731
Total number of implanted stents	1.85 ± 1.04	1.53 ± 0.90	2.27 ± 1.07	<0.001
Trans-radial approach	3,228 (46.2)	1,855 (47.4)	1,373 (44.7)	0.028
Intravascular imaging	2,082 (29.8)	1,222 (31.2)	860 (28.0)	0.004
IVUS	1,916 (27.4)	1,129 (28.8)	787 (25.6)	0.003
OCT	189 (2.7)	107 (2.7)	82 (2.7)	0.929
Mechanical circulatory support	154 (2.2)	100 (2.6)	54 (1.8)	0.030
IABP	106 (1.5)	68 (1.7)	38 (1.2)	0.110
ECMO	60 (0.9)	38 (1.0)	22 (0.7)	0.311

Values are mean ± SD or n (%).

ACC = American College of Cardiology; AHA = American Heart Association; ECMO = extracorporeal membrane oxygenation; IABP = intra-aortic balloon pump; IVUS = intravascular ultrasound; LAD = left anterior descending artery; LCX = left circumflex artery; OCT = optical coherence tomography; RCA = right coronary artery; other abbreviations as in Table 1.

^aType B2 or C lesions according to ACC/AHA classification.

Clinical outcomes according to treatment strategy

The median follow-up duration of the study was 3.0 years (IQR: 2.8-3.1) years. Clinical outcomes during the follow-up period according to revascularization strategy are summarized in Table 3 and Figure 2. At 3 years, the risk of MACE was significantly lower in the CR group than in the IR group (26.9% vs 18.6%; adjusted HR: 0.75; 95% CI: 0.66-0.85; P < 0.001) (Figure 2). The lower risk of MACE in the CR group was mainly driven by lower risk of cardiac death or recurrent MI (10.4% vs 6.9%; adjusted HR: 0.76; 95% CI: 0.60-0.95; P = 0.015) and unplanned repeat revascularization (13.1% vs 8.6%; adjusted HR: 0.68; 95% CI: 0.56-0.82; P < 0.001) (Figure 2). The reduction in unplanned repeat revascularization was mainly attributed to significantly lower risk of non-IRA revascularizations in the CR group (7.7% vs 4.6%; adjusted HR: 0.60; 95% CI: 0.47-0.78; P < 0.001), whereas the risks of infarct-related lesion and IRA revascularization were comparable between groups. Sensitivity analysis using multivariable Cox regression, propensity score matching, and IPW adjustment yielded consistent results, which showed significantly lower risk of MACE, cardiac death, and unplanned repeat revascularization in the CR group compared with the IR group (Table 3). Supplemental Figure 1 illustrates the 3-year cumulative incidence of secondary outcomes, including all-cause death, cardiac death, recurrent MI, unplanned repeat revascularization, and hospitalization for heart failure. In landmark analyses at 6 months and 1 year, CR was associated with a lower cumulative incidence of both MACE and the composite of cardiac death or recurrent MI during the early post-PCI period up to each landmark, as well as during the subsequent follow-up among patients who were event-free at the landmark time points. The between-group differences remained statistically significant across all intervals (Supplemental Figures 2 and 3). In multivariable Cox proportional hazard models, CR was independently associated with a reduced risk of MACE (adjusted HR: 0.74; 95% CI: 0.65-0.84; P < 0.001) and cardiac death or recurrent MI (adjusted HR: 0.75; 95% CI: 0.60-0.94; P = 0.012) at 3 years (Supplemental Table 2). In competing risk analyses treating all-cause death as a competing event, complete revascularization was consistently associated with a lower risk of recurrent MI, unplanned repeat revascularization, and hospitalization for heart failure. These findings were concordant with the results of the cause-specific Cox regression, propensity score matching, and inverse probability weighting analyses (Supplemental Table 3).

Table 3.

Comparison of Clinical Outcomes at 3 Years Between IR and CR

	IR (n = 3,915)	CR (n = 3,072)	Univariable Analysis			Multivariable Analysisb			Propensity Score–Matchedc			IPW-Adjustedc
			HR	95% CI	P Value	HR	95% CI	P Value	HR	95% CI	P Value	HR	95% CI	P Value
MACEa	998 (26.9)	533 (18.6)	0.65	0.58-0.72	<0.001	0.75	0.66-0.85	<0.001	0.76	0.67-0.86	<0.001	0.74	0.66-0.82	<0.001
All-cause death	475 (12.8)	237 (8.1)	0.62	0.53-0.73	<0.001	0.73	0.60-0.90	0.003	0.76	0.63-0.91	0.003	0.73	0.62-0.87	<0.001
Cardiac death	269 (7.3)	133 (4.6)	0.62	0.50-0.76	<0.001	0.75	0.57-0.99	0.042	0.71	0.56-0.91	0.007	0.73	0.58-0.91	0.005
Recurrent MI	138 (4.1)	74 (2.8)	0.66	0.50-0.88	0.005	0.72	0.50-1.02	0.066	0.76	0.55-1.06	0.103	0.73	0.54-0.99	0.040
Unplanned repeat revascularization	450 (13.1)	233 (8.6)	0.63	0.53-0.73	<0.001	0.68	0.56-0.82	<0.001	0.69	0.58-0.83	<0.001	0.68	0.58-0.81	<0.001
Infarct-related lesion revascularization	112 (3.4)	70 (2.6)	0.76	0.56-1.02	0.066	0.84	0.58-1.21	0.339	0.86	0.61-1.21	0.384	0.82	0.60-1.12	0.216
IRA revascularization	133 (4.1)	83 (3.1)	0.75	0.57-0.99	0.041	0.82	0.59-1.14	0.236	0.86	0.63-1.18	0.357	0.84	0.63-1.12	0.242
Non-IRA revascularization	260 (7.7)	120 (4.6)	0.56	0.45-0.69	<0.001	0.60	0.47-0.78	<0.001	0.60	0.47-0.76	<0.001	0.60	0.48-0.76	<0.001
Admission for heart failure	183 (5.6)	96 (3.6)	0.65	0.51-0.83	0.001	0.87	0.64-1.18	0.365	0.96	0.72-1.28	0.793	0.84	0.64-1.10	0.198
Cardiac death or recurrent MI	378 (10.4)	195 (6.9)	0.64	0.54-0.76	<0.001	0.76	0.60-0.95	0.015	0.73	0.60-0.90	0.003	0.73	0.61-0.88	<0.001

Values are n (%) unless otherwise indicated. The cumulative incidences of clinical outcomes are presented as Kaplan-Meier estimates during median follow-up of 1,091 days. Multivariable Cox proportional hazard regression model, propensity score–matched cohort, and inverse probability of treatment weighting method were used to adjust for baseline differences between comparative groups.

IPW = inverse probability weighted; MACE = major adverse cardiac events; other abbreviations as in Table 1 and Table 2.

^aMACEs were defined as a composite of all-cause death, recurrent MI, unplanned repeat revascularization, or admission for heart failure.

^bThe adjusted variables included in the multivariable analysis are age, sex, BMI, Killip class at initial presentation, systolic blood pressure, hypertension, diabetes, previous MI, previous CVA, previous CHF, hemoglobin, estimated GFR, peak CK-MB, left ventricular ejection fraction, clopidogrel, potent P2Y₁₂ inhibitor. angiographic disease extent, left main artery disease, culprit lesion location, ACC/AHA type B2/C lesion, timing of non-IRA PCI, post-PCI TIMI flow of IRA, treatment modalities for IRA, thrombus aspiration, total stent length of IRA, trans-radial approach, intravascular imaging, and mechanical circulatory support.

^cThe adjusted covariables were age, sex, BMI, Killip class at initial presentation, hypertension, diabetes mellitus, dyslipidemia, current smoking, end-stage renal disease, previous MI, previous CVA, previous CHF, hemoglobin, estimated GFR, peak CK-MB, atrial fibrillation, left ventricular ejection fraction, aspirin, clopidogrel, potent P2Y₁₂ inhibitor, RAAS blocker, beta-blocker, statin, oral anticoagulant, angiographic disease extent, left main artery disease, culprit lesion location, ACC/AHA type B2/C lesion, pre-PCI TIMI flow of IRA, intravascular imaging, and mechanical circulatory support.

Figure 2.

Cumulative Incidence of MACE and Cardiac Death or Recurrent MI

Kaplan-Meier curves with cumulative hazards of (A) MACE and (B) cardiac death or recurrent MI compared according to the PCI strategy are presented. MACE = major adverse cardiac events; MI = myocardial infarction.

Subgroup analysis

Figure 3 illustrates the prognostic impact of CR compared with IR across various clinically relevant subgroups. The significantly lower risk of MACE in the CR group was similarly observed across all subgroups, with no significant interaction P values, except for angiographic disease extent (P for interaction = 0.015).

Figure 3.

Comparison of MACE Between IR and CR Across Subgroups

Cumulative incidence and HR with a 95% CI of MACE at 3 years are presented for the IR and CR groups across various subgroups. The interaction P value indicates the likelihood of interaction between the variable and the relative prognostic value. ACC = American College of Cardiology; AHA = American Heart Association; CR = complete revascularization; GFR = glomerular filtration rate; IR = incomplete revascularization.

Patients without cardiogenic shock

In patients without cardiogenic shock, defined as the absence of Killip class IV at initial presentation and mechanical circulatory support, baseline characteristics and procedural profiles are summarized in Supplemental Tables 4 and 5. In this cohort, as in the overall study population, CR was associated with a significantly lower risk of MACE at 3 years compared with IR. Consistent findings were observed for other clinical outcomes, including cardiac death or recurrent MI and unplanned repeat revascularization (Supplemental Table 6, Supplemental Figure 4).

Additional analyses using propensity score matching and inverse probability weighting demonstrated adequate balance between the 2 groups (Supplemental Table 7). Sensitivity analyses using multivariable Cox regression, propensity score matching, and inverse probability weighting yielded consistent results, supporting the association between CR and improved clinical outcomes even after exclusion of hemodynamically unstable patients.

Discussion

In the current study, we compared 3-year clinical outcomes between CR and IR strategies in patients with NSTEMI and MVD who underwent successful PCI for IRA, using pooled data from a nationwide, multicenter, prospective acute myocardial infarction registry. The main findings were as follows. First, CR was associated with a significantly lower risk of MACE compared with IR, and this association remained consistent across multiple sensitivity analyses adjusting for baseline differences. Second, CR was also associated with a significantly lower risk of the composite outcome of cardiac death or recurrent MI, and unplanned repeat revascularization for non-IRA. Third, in multivariable Cox proportional hazards models, CR was as an independent predictor of reduced risk for both MACE and the composite of cardiac death or recurrent MI (Central Illustration).

Central Illustration.

Prognostic Impact of CR in Patients With NSTEMI With MVD

Among 6,987 patients with NSTEMI and MVD from the pooled KAMIR-NIH and KAMIR-V registries (2012-2020), 56.0% (n = 3,915) underwent IR and 44.0% (n = 3,072) underwent CR. At 3 years, CR was associated with a significantly lower incidence of MACE compared with IR, primarily driven by reductions in cardiac death or recurrent MI and unplanned repeat revascularization. These findings remained consistent after rigorous adjustment for baseline differences, supporting the prognostic benefit of CR in NSTEMI patients with MVD. CR = complete revascularization; IPTW = inverse probability of treatment weighting IR = incomplete revascularization; IRA = infarct-related artery; MACE = major adverse cardiac events; MI = myocardial infarction; MVD = multivessel disease; NSTEMI = non–ST-segment elevation myocardial infarction; PCI = percutaneous coronary intervention; PS = propensity score.

MVD is frequently encountered in patients presenting with NSTEMI and is associated with adverse clinical outcomes, including increased mortality.1, 2, 3^,22 Despite its clinical relevance, no randomized trial to date has directly compared IR and CR strategies in NSTEMI patients with MVD, resulting in limited evidence to guide optimal management. For example, the FULL-REVASC (FFR-Guidance for Complete Nonculprit Revascularization) trial compared physiology-guided CR with culprit-only PCI in patients with STEMI or very-high-risk NSTEMI and MVD.²³ Over a median follow-up of 4.8 years, there was no significant difference in major cardiovascular outcomes, including death, MI, or unplanned revascularization. However, NSTEMI patients comprised <10% of the study population, and only those with high-risk features were included. Therefore, the findings cannot be readily applied to the broader NSTEMI population. These findings support the rationale for the current study, which focuses exclusively on patients with NSTEMI and MVD to clarify the prognostic implications of IR vs CR in this understudied population.

Although there has been no dedicated trial comparing CR vs IRA-only PCI in patients with NSTEMI and MVD, previous observational studies consistently demonstrated the prognostic benefit of CR than IR in these patients. Rathod et al,⁵ in an analysis of more than 21,000 NSTEMI patients with MVD from a nationwide British registry, demonstrated that CR was independently associated with a lower risk of all-cause mortality at 4.6 years, supporting its role in long-term survival benefit. Likewise, a meta-analysis by Siebert et al,¹⁵ which included more than 60,000 patients from 15 observational studies, showed that CR significantly reduced all-cause mortality and repeat revascularization. More recently, the FIRE (Functional Assessment in Elderly MI Patients with Multivessel Disease) trial, the randomized clinical trial to evaluate CR in older patients with acute coronary syndrome and MVD, demonstrated that CR significantly lowered the risk of major adverse cardiac and cerebrovascular events, including cardiovascular death and MI, without increasing procedural complications.²⁴ Notably, 64% of the FIRE trial population presented with NSTEMI. In line with these studies, the current study consistently observed that the CR group was significantly associated with the reduced risk of MACE, mainly driven by lower risk of hard clinical endpoints including cardiovascular death or MI. These results suggest that CR may be associated with favorable long-term outcomes in patients with NSTEMI and MVD.

Despite an increased number of implanted coronary stents in the CR group, the incidences of infarct-related lesion and IRA revascularization were comparable with the IR group. In contrast, the risk of unplanned non-IRA revascularization was significantly higher in the IR group, and this remained significant even after rigorous adjustment for baseline characteristics. These findings suggest that IR for angiographically significant non-IRA lesions may have contributed to persistent reversible myocardial ischemia, recurrent angina, and progressive left ventricular dysfunction over time, leading to unplanned non-IRA revascularization. Untreated flow-limiting stenoses result in an imbalance between myocardial oxygen supply and demand, which can lead to reversible myocardial dysfunction, such as hibernating myocardium. Revascularization of these territories has been shown to restore perfusion and improve contractile function, particularly in viable but chronically underperfused myocardial segments.²⁵ Moreover, watershed zones (areas between perfusion territories) are especially vulnerable to ischemic injury when adjacent vessels remain stenotic, which may accelerate adverse remodeling and the progression of heart failure.²⁶ These pathophysiological mechanisms support the rationale for CR in patients with NSTEMI and MVD. Furthermore, in multivariable Cox proportional hazards models, CR remained an independent predictor of reduced risk for both MACE and the composite outcome of cardiac death or recurrent MI, consistent with previous research.⁵ Taken together, these findings suggest the potential prognostic relevance of achieving CR in patients with NSTEMI and MVD.

In the current study, the benefit of CR over IR was consistently observed across a wide range of clinically relevant subgroups. Subgroup analyses showed no significant interaction for most variables, indicating that the advantage of CR was preserved regardless of baseline characteristics such as age, sex, diabetes, left ventricular ejection fraction, or Killip class. The only significant interaction was observed for angiographic disease extent, where patients with 3-vessel disease experienced greater benefit from CR compared with those with 2-vessel disease. Angiographic disease extent is known to correlate strongly with adverse cardiovascular outcomes.27, 28, 29 Previous research has demonstrated that the number of diseased vessels and overall atherosclerotic burden are independently associated with increased risk of cardiovascular death or MI.³⁰ These findings suggest that the clinical benefit of CR may be amplified in patients with more extensive coronary disease.

Study Limitations

First, observational design of the current study precludes causal inference, and despite comprehensive statistical adjustments including multivariable analysis, propensity score matching, and inverse probability weighting, the possibility of residual confounding from unmeasured variables cannot be excluded. Second, the decision to perform PCI for non-IRA was made at the discretion of the treating physician, potentially introducing selection bias related to clinical status, lesion complexity, or patient frailty, which were not fully captured in the registry. Additionally, angiographic assessment of CR vs IR was based on operator visual estimation at each participating center, without independent core laboratory analysis. Third, the registry lacked detailed procedural information regarding the timing, location, and technical aspects of staged non-IRA PCI. In addition, key anatomical characteristics that define complex coronary lesions—including chronic total occlusion, bifurcation, and severe calcification—were not systematically collected, limiting our ability to evaluate the impact of lesion complexity on revascularization strategy and clinical outcomes. Fourth, information regarding the use of physiology-guided assessment, such as fractional flow reserve, instantaneous wave-free ratio, or angiography-derived fractional flow reserve was not systematically collected in the registry. Therefore, the role of physiology-guided decision making for non-IRA could not be evaluated. Fifth, IR should not be interpreted as synonymous with a culprit-only PCI strategy, as non-IRA PCI was attempted in a substantial proportion of patients classified as having IR, but residual angiographically significant or stenosis or flow-limiting lesions remained. In addition, detailed central adjudication of MI was not available in the current study, limiting exploration of the underlying causes of the observed reduction in recurrent MI in the CR group.

Conclusions

In patients with NSTEMI and MVD, CR was associated with a lower risk of MACE at 3 years, mainly driven by fewer cardiac death or MI events and unplanned repeat revascularization, compared with IR. Further studies are warranted to better define the optimal revascularization strategy, including the timing and extent of non-IRA intervention.

Funding Support and Author Disclosures

This work was supported by the Korea Centers for Disease Control and Prevention (Grant No. 2016-ER6304-01) and the National Institutes of Health (Grant No. 2016-ER6304-02). Dr Seung Hun Lee received an Institutional Research Grant from Abbott Vascular and Boston Scientific. Prof Joo-Yong Hahn received an Institutional Research Grant from National Evidence-based Healthcare Collaborating Agency, Ministry of Health & Welfare, Korea, Abbott Vascular, Biosensors, Boston Scientific, Daiichi-Sankyo, Donga-ST, Hanmi Pharmaceutical, and Medtronic Inc. Prof Hyeon-Cheol Gwon received an Institutional Research Grant from Boston Scientific, Genoss, and Medtronic Inc. Dr Joo Myung Lee received an Institutional Research Grant from Abbott Vascular, Boston Scientific, Terumo Corporation, Pulse Medical, MicroPort, Donga-ST, and Yuhan Pharmaceutical. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.