Impact of revascularization strategies versus optimized medical treatment on coronary events related only to revascularized arteries: a target clinical trial emulation approach

Abstract

Background

Traditional assessments of coronary revascularization strategies often rely on long-term composite endpoints that reflect the natural progression of coronary artery disease (CAD) rather than the direct efficacy of the intervention itself. This study sought to evaluate the specific impact of surgical and percutaneous revascularization by analyzing nonfatal myocardial infarction (MI) or death occurring exclusively in the territory of the revascularized coronary artery.

Methods

This retrospective, single-center cohort study utilized data from the MASS (Medicine, Angioplasty, or Surgery Study) registry. Patients with multivessel CAD and documented ischemia were included if they were eligible for all three therapeutic strategies: coronary artery bypass grafting (CABG), percutaneous coronary intervention (PCI), or optimal medical therapy (OMT). The primary endpoint was a composite of nonfatal MI or death attributable solely to the target revascularized vessel. The follow-up period was five years.

Results

Of the 1,695 eligible patients, 572 (33.7%) underwent CABG, 549 (32.4%) PCI, and 574 (33.9%) received OMT. Over the five-year follow-up, 155 revascularized-vessel-related events were recorded. The global log-rank test demonstrated significant differences between the groups (p < 0.001). OMT was associated with significantly lower event-free survival compared to both CABG (p < 0.001) and PCI (p < 0.001), while no significant difference was observed between CABG and PCI (p = 0.346). Revascularization resulted in a relative risk reduction of 46–33% compared to OMT.

Conclusion

Both surgical and percutaneous revascularization were associated with a lower incidence of vessel-specific clinical events compared to optimal medical therapy over five years. These findings highlight the direct, vessel-specific benefits of revascularization strategies in multivessel CAD.

Clinical trial registration

https://clinicaltrials.gov/study/NCT07362108.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13019-026-04544-5.

Background

Myocardial revascularization—whether surgical (CABG) or percutaneous (PCI)—remains a cornerstone in the treatment of coronary artery disease (CAD), particularly for patients with complex lesions or refractory symptoms. Despite substantial technical advances in both modalities, the incidence of post-procedural adverse cardiovascular events remains clinically significant, necessitating ongoing research into long-term outcomes.

Traditionally, therapeutic efficacy is assessed using composite endpoints such as Major Adverse Cardiovascular Events (MACE), which typically include all-cause mortality, non-fatal myocardial infarction (MI), stroke, and repeat revascularization [1]. However, these indicators often incorporate events unrelated to the treated vessel, which can dilute the perceived direct impact of the intervention [2]. In contrast, events strictly linked to the target vessel—such as target-vessel MI, repeat revascularization of the same segment, or vessel occlusion—constitute more precise markers of clinical and technical success.

Seminal clinical trials, such as CASS [3] and RITA-2 [4], compared clinical and interventional strategies but prioritized overall mortality or combined outcomes of death and MI, often failing to establish a direct causal relationship with the revascularized vessel [5]. While the SYNTAX study [6] adopted a more comprehensive composite endpoint (MACCE) by including stroke, the ISCHEMIA trial [7] expanded definitions further to include hospitalizations for unstable angina, heart failure, or resuscitated cardiac arrest. Although such expansions increase statistical power, they often compromise clinical specificity.

Despite facilitating statistical analysis, the components of composite endpoints frequently reflect heterogeneous pathophysiological mechanisms dissociated from the intervention. For instance, stroke may result from cerebral atherothrombosis; all-cause mortality includes non-cardiovascular deaths; and unplanned revascularization often stems from disease progression in untreated segments. Even MI episodes can occur outside the target territory, complicating causal attribution. Consequently, overly broad indicators can obscure the mechanisms intrinsic to the revascularization procedure itself.However, that traditional MACE outcomes remain the cornerstone of clinical decision-making.

In this context, the present study proposes an evaluation focused on non-fatal MI and cardiac death directly attributable to vessels revascularized by PCI or CABG. These outcomes are compared to a control group of CAD patients eligible for intervention but maintained exclusively on optimal medical therapy. This approach aims to increase accuracy in assessing the true impact of revascularization, fill existing gaps in the literature, and support individualized clinical decision-making based on specific pathophysiological mechanisms.

Methods

Study design

The study design and baseline patient characteristics have been described previously [8]. Briefly, this is a retrospective single-center cohort study based on the MASS (Medicine, Angioplasty, or Surgery Study) database at the Heart Institute of the University of São Paulo (InCor-HCFMUSP). The MASS database includes patients with CAD and proven myocardial ischemia who were indicated for surgical, percutaneous, or exclusively medical treatment. All eligible patients provided informed consent according to the protocol established by the MASS Research Unit. Written informed consent was obtained from all individual participants prior to data collection or intervention, in accordance with the Declaration of Helsinki.

Sample selection

Eligible patients presented with angina pectoris and multivessel CAD, defined as stenosis > 70% of the luminal diameter in at least two main epicardial coronary arteries, confirmed by invasive coronary angiography. Myocardial ischemia was objectively documented via exercise stress testing, pharmacological stress echocardiography, or myocardial perfusion scintigraphy (physical or pharmacological stress). Angina intensity was classified according to the Canadian Cardiovascular Society (CCS) criteria.

Inclusion and exclusion criteria

Inclusion Criteria: Individuals with multivessel CAD and ischemia documented by functional methods. Patients must have had a simultaneous formal indication for any of the three strategies: CABG, PCI, or Optimal Medical Therapy (OMT).

Exclusion Criteria: Reduced life expectancy or inability to undergo prolonged clinical follow-up. Presence of implantable cardiac devices. Significant left ventricular dysfunction (LVEF < 55%). Acute coronary syndrome within the preceding six months. Previous percutaneous revascularization. Clinically relevant valvular disease. Chronic kidney disease (creatinine ≥ 2.0 mg/dL). Active rheumatological diseases or sepsis. Pulmonary thromboembolism or deep vein thrombosis within the previous six months. Malignant neoplasm under active treatment.

Methodological definitions

CAD was characterized by stenosis > 70% of the lumen in major epicardial coronary arteries. Left ventricular systolic function was considered preserved when the left ventricular ejection fraction (LVEF) was ≥ 55%, estimated by echocardiography or another validated method.

Therapeutic strategies

All participants received Optimal Medical Therapy (OMT) in accordance with contemporary guidelines for symptomatic angina control and secondary prevention. The regimen included, as individually indicated: aspirin, P2Y12 inhibitors, beta-blockers, calcium channel blockers, spironolactone, ACE inhibitors, angiotensin receptor blockers, diuretics, statins, nitrates, and antiarrhythmic drugs. Glycemic control was individualized using oral hypoglycemics or insulin. All patients received dietary guidance focusing on reduced fat and simple carbohydrate intake.

Surgical Group (CABG)

Complete revascularization was prioritized, with a preference for using the internal thoracic artery. Procedures followed standardized techniques, utilizing either mild hypothermia and cardioplegia under extracorporeal circulation or off-pump techniques, at the surgeon’s discretion.

Angioplasty Group (PCI)

Patients underwent elective PCI according to international guidelines. Procedures used standard radial or femoral access, appropriate guide catheters, balloon predilation when necessary, and the implantation of drug-eluting or bare-metal stents at the operator’s discretion. Patients received dual antiplatelet therapy (DAPT) and anticoagulation per institutional protocols.

Medical Group (OMT)

This group comprised patients with obstructive stenoses and a formal indication for revascularization who were instead managed medically. This therapeutic choice was based on the attending physician’s discretion, patient preference, or documented symptom stability. Invasive strategies remained available if symptoms progressed or at the request of the patient or physician. Follow-up was conducted semi-annually.

Follow-up and outcomes

Follow-up was conducted on an outpatient basis with semi-annual assessments over an average period of five years. Each visit included a detailed clinical evaluation and laboratory/functional tests to monitor metabolic control and cardiac function.

Primary outcomes

Non-fatal acute myocardial infarction (AMI) related to the revascularized artery.

Cardiac death following an MI related to the revascularized artery. AMI was defined by international criteria: compatible chest pain, elevation of myocardial necrosis biomarkers, and/or new Q waves in at least two contiguous ECG leads. Perioperative myocardial infarctions have been excluded from the definition of the primary outcome. The treated artery was identified via angiographic review. To ensure methodological rigor and reduce bias, all diagnoses were validated by three experienced, independent physicians not involved in the study. Final definitions were based on consensus among these evaluators. Cardiac deaths or MIs unrelated to the target vessel, as well as all-cause mortality, were recorded and analyzed separately.

Statistical analysis

Continuous variables were expressed as mean and standard deviation or median and interquartile range, according to the distribution (evaluated by graphical inspection and normality tests). Categorical variables were reported as absolute and percentage frequencies.

Survival analysis and concurrent risks

The primary outcome was defined as the occurrence of clinical events (infarction or death) directly attributable to the treated coronary artery or eligible for clinical treatment. For the analysis of time to the first event, the Kaplan-Meier method and the log-rank test were used. However, in long-term follow-up studies, the occurrence of death from causes unrelated to the treated artery (e.g., non-cardiovascular causes) acts as a concurrent risk. Unlike conventional censoring, concurrent risk not only prevents observation of the primary outcome but also alters the probability of its occurrence. To adjust this probability, the Fine-Gray Model (Risk of Subdistribution) was applied. While the Cox model estimates the event rate among those who are still “at risk,” the Fine-Gray model keeps patients who experienced the concurrent event in the denominator, allowing for a more precise estimate of the cumulative incidence in clinical practice. Results were reported as subdistribution hazard ratios (subHR) with 95% CI.

Multivariate modeling and predictors

Cox and Fine-Gray proportional hazards regression models were constructed to identify independent predictors. The following covariates were selected for clinical relevance: Demographic and Clinical Data: Age, sex, BMI, diabetes, hypertension, smoking, and renal function (creatinine). Metabolic and Cardiac Profile: Ejection fraction, extent of coronary artery disease, lipid profile (LDL, HDL, TG), and HbA1c. Primary Exposure: Treatment group and completeness of revascularization (categorized as complete, incomplete, or absent/medical therapy).

In addition to analyzing the first event, the cumulative event burden (average number of events per patient) was assessed to capture the frequency of recurrent events during follow-up. All analyses were performed using R software (version 4.x), utilizing the survival (Cox analyses), cmprsk (Fine-Gray analysis), survminer, and tidyverse packages. A significance level of p < 0.05 was adopted with two-tailed tests and Benjamini-Hochberg adjustment for multiple comparisons.

Results

Population characteristics

Between January 2012 and December 2019, 1,695 patients met the eligibility criteria. Of these, 572 (33.7%) underwent coronary artery bypass grafting (CABG), 549 (32.4%) underwent percutaneous coronary intervention (PCI), and 574 (33.9%) received optimal medical therapy (OMT). Detailed demographic, clinical, laboratory, and angiographic data are presented in Table 1. Baseline characteristics varied significantly across the three groups. The median age differed by treatment strategy (p = 0.004): 62 [55–68] years for CABG, 60 [53.7–66] years for PCI, and 61 [54–67] years for OMT. Sex distribution was similar across groups (p = 0.455), with a consistent male predominance: 69.4% (n = 397) in the CABG group, 66.1% (n = 363) in the PCI group, and 66.7% (n = 383) in the OMT group.Regarding comorbidities, the prevalence of hypertension was slightly higher in the PCI (84.2%) and CABG (82.0%) groups. Conversely, the prevalence of diabetes mellitus was significantly higher in the OMT group (58.2%; n = 334) compared with the PCI (52.6%; n = 289) and CABG (51.4%; n = 294) groups (p = 0.049). Consequently, glycated hemoglobin was also higher in the OMT cohort (6.70% [6.00–7.70]) than in the PCI (6.20% [5.70–7.50]) and CABG (6.10% [5.70–7.20]) groups (p = 0.004). Other laboratory variables, such as cholesterol, glucose, and creatinine, did not show statistically significant differences.

Table 1.

Clinical, laboratory, and angiographic characteristics of patients according to the initial therapeutic strategy

Variable	CABG N = 572	PCI N = 549	OMT N = 574		p-valor
				p-valor (global)	CABG/ PCI	CABG/ OMT	PCI/ OMT
Age (Years)	62 [55–68]	60 [53.7–66]	61 [54–67]	0.004	0.003	0.129	0.340
BMI (kg/m²)	27.1 [24.9–30.1]	27.3 [24.6–29.9]	27.5 [24.9–30.3]	0.567	0.984	0.662	0.602
Gender Male n (%)	397 (69.4)	363 (66.1)	383 (66.7)	0.455	0.265	0.362	0.879
Hypertension n (%)	469 (82.0)	462 (84.2)	453 (78.9)	0.074	0.376	0.216	0.029
Diabetes n (%)	294 (51.4)	289 (52.6)	334 (58.2)	0.049	0.721	0.024	0.070
TC (mg/dL)	196 [160–236]	196 [158–230]	192 [160–225]	0.359	0.997	0.408	0.458
LDL (mg/dL)	121 [90–152]	118 [91–149]	111 [90–150]	0.636	0.997	0.662	0.720
HDL (mg/dL)	38 [33.8–44]	37 [32–44]	38 [32–45.8]	0.068	0.153	0.925	0.084
TG (mg/dL)	150.5 [108–209]	153 [111–219]	144 [103–203]	0.068	0.591	0.359	0.057
Glucose (mg/dL)	112 [97–149]	112 [98–144]	115 [99–151]	0.155	0.975	0.188	0.259
HbA1c (%)	6.10 [5.70–7.20]	6.20 [5.70–7.50]	6.70 [6.00–7.70]	0.004	0.258	0.003	0.201
Creatinine (mg/dL)	1.02 [0.90–1.20]	1.00 [0.90–1.20]	1.00 [0.90–1.20]	0.179	0.231	0.260	0.994
Angina Grade n (%)				< 0.001	< 0.001	< 0.001	< 0.001
1	60 (10.5)	62 (11.3)	124 (21.6)
2	367 (64.2)	234 (42.6)	314 (54.7)
3	121 (21.2)	173 (31.5)	127 (22.1)
4	24 (4.2)	80 (14.6)	9 (1.6)
AMI previous n (%)	242 (42.3)	246 (44.8)	222 (38.7)	0.111	0.433	0.233	0.043
LVEF (%)	62 [56–68]	62 [56–68]	61 [55–68]	0.400	0.621	0.876	0.402
Number of diseased vessels				< 0.001	< 0.001	0.209	< 0.001
1	0 (0.0)	4 (0.7)	1 (0.2)
2	135 (23.6)	221 (40.3)	154 (26.8)
3	437 (76.4)	324 (59.0)	419 (73.0)
AD	547 (95.6)	515 (93.8)	557 (97.0)	0.032	0.218	0.266	0.014
LCX	494 (86.4)	437 (79.6)	492 (85.7)	0.003	0.003	0.817	0.009
RCA	518 (90.6)	450 (82.0)	513 (89.4)	< 0.001	< 0.001	0.569	< 0.001
AMI previous	55 (9.6)	61 (11.1)	69 (12.0)	0.42	0.469	0.224	0.702

Data is presented as number (%) and median [IQR]. TC= Total Cholesterol; LDL = Low-Density Lipoprotein; HDL=High-Density Lipoprotein. TG= Triglycerides. Comparisons were performed using the chi-square test and the Kruskal-Wallis’s test

Angiographic profile and clinical characteristics

Angina symptoms differed significantly among the groups (p < 0.001). The PCI group had a higher proportion of severe symptoms (grades 3 and 4: 46.1%), whereas the CABG group primarily presented with milder symptoms (grades 1 and 2: 74.7%). In the OMT group, grade 2 predominated (54.7%).The extent of coronary disease also varied (p < 0.001): most patients in the CABG (76.4%) and OMT (73.0%) groups had triple-vessel disease, compared to 59.0% in the PCI group. Additionally, lesions in the LAD, LCX, and RCA were more frequent in the CABG group (p < 0.05). After reviewing the original angiographic records of five patients, we confirmed that two of the three main coronary arteries presented moderate obstructive lesions that did not justify percutaneous or surgical intervention. A sensitivity analysis excluding these patients was performed and did not alter the study conclusions.

Event-free survival analysis

The survival analysis (Fig. 1) illustrates cumulative incidence functions for the primary outcome. In the unadjusted analysis, OMT showed the highest cumulative incidence, while CABG and PCI were associated with lower event accumulation. The gap between revascularization strategies (CABG/PCI) and OMT widened over time, whereas PCI and CABG showed similar incidence patterns. Because treatment allocation was non-random, propensity score (PS) models were used to balance covariates. Post-adjustment, standardized mean differences remained below 0.10 across all comparisons, indicating adequate balance (Supplementary Table S1).In Fine-Gray competing risk models, CABG was associated with a lower cumulative incidence of the primary outcome than OMT (unadjusted sHR 0.410, 95% CI 0.272–0.617; p < 0.0001). This association remained robust after adjusting for the PS logit (sHR 0.439, 95% CI 0.275–0.700; p = 0.00055) and baseline covariates (sHR 0.444, 95% CI 0.283–0.698; p = 0.00042). Similarly, PCI showed a significantly lower incidence compared to OMT (unadjusted sHR 0.523, 95% CI 0.358–0.766; p = 0.00088). These findings remained consistent after adjusting for the PS logit (sHR 0.467, 95% CI 0.314–0.695; p = 0.00017) and baseline covariates (sHR 0.448, 95% CI 0.296–0.678; p = 0.00015).

Fig. 1.

Cumulative incidence according to therapeutic strategy

Impact of interventions

When pooling revascularization patients (CABG + PCI) against OMT, the intervention group showed a significantly higher probability of remaining event-free (log-rank p < 0.001), with curves diverging after approximately three years. Competing risks analysis confirmed this benefit (Gray’s test p < 0.001; Supplementary Figs. 1 S and 2 S). Preliminary analysis revealed significant differences (p < 0.05) between treatments for myocardial infarction and deaths related to the treated artery; in these instances, OMT had significantly higher event rates. Conversely, mortality related to untreated arteries was similar across groups (p = 0.746), suggesting that the initial intervention type did not impact this specific outcome (Table 2). Additionally, the average number of events per patient was higher in the OMT group (0.179) than in the PCI (0.073) and CABG (0.058) cohorts (Fig. 2).

Table 2.

Distribution of Clinical Events According to Treatment Strategy

Event	CABG (n = 572)	PCI (n = 549)	OMT (n = 574)	P-value
Myocardial infarction – artery treated %	4.9 (28)	5.8 (32)	12.0 (69)	p < 0.001 (5.14 × 10− 6)
Death – artery treated %	0.9 (5)	1.5 (8)	5.9 (34)	p < 0.001 (9.65 × 10− 8)
Total events related to treated artery %	5.8 (33)	7.3 (40)	17.9 (103)	p < 0.001 (1.87 × 10− 12)
Myocardial infarction –untrated artery %	4.7 (27)	5.3 (29)	12.0 (69)	p < 0,001 (1.04 × 10− 6)
Death - untrated artery %	13.3 (76)	12.0 (66)	13.4 (77)	0.746
Death from other causes %	11.4 (65)	9.7 (53)	7.5 (43)	0.081

Percentages were calculated based on the total number of patients in each treatment group (CABG n = 572; PCI n = 549; OMT n = 574). All events occurring during follow-up were included. Therefore, percentages represent total event burden relative to the total number of patients and do not correspond to time-to-first-event incidence estimates. In the OMT group, artery-related events refer to myocardial infarction or ischemic cardiac death occurring in coronary territories deemed eligible for revascularization at baseline. * Death from other causes excludes myocardial infarction and sudden cardiac death

Fig. 2.

Mean cumulative treated-artery–related event burden per patient according to treatment strategy

Completeness of revascularization

Regarding completeness of revascularization (Fig. 4S), the probability of event-free survival differed according to the degree of completeness (overall log-rank $p < 0.001$). Patients who underwent complete revascularization achieved the highest event-free survival, followed by those with incomplete revascularization, whereas patients without revascularization showed the highest cumulative incidence of events. In pairwise comparisons, the following results were observed: Complete revascularization vs. none (p < 0.001) Incomplete revascularization vs. none (p = 0.024) Complete vs. incomplete revascularization (p = 0.432) Although no statistically significant difference was found between complete and incomplete revascularization, the curve for the complete group remained consistently superior. After adjustment using the Cox model, complete revascularization was associated with a significantly lower risk of events (HR 0.57; 95% CI 0.39–0.84; p = 0.004), while incomplete revascularization showed a non-significant trend toward benefit (HR 0.62; 95% CI 0.38–1.03; p = 0.066). In the Fine–Gray model, both incomplete (subHR 0.59; 95% CI 0.36–0.98; p = 0.041) and complete revascularization (subHR 0.54; 95% CI 0.37–0.80; p = 0.002) were associated with a reduced risk of events.

Multivariable predictors of events

After multivariable adjustment for relevant clinical, laboratory, and anatomical variables, both revascularization strategies remained independently associated with a lower risk of events related to the treated artery compared with medical therapy alone. Coronary artery bypass grafting (CABG) was associated with a 46% relative risk reduction (HR 0.54; 95% CI 0.33–0.86; p = 0.010), while percutaneous coronary intervention (PCI) was associated with a 33% relative reduction (HR 0.67; 95% CI 0.45–1.00; p = 0.047) (Fig. 3 and Table 1S). Among clinical factors, diabetes mellitus was independently associated with a 51% increase in the risk of events related to the treated artery (HR 1.51; 95% CI 1.04–2.18; p = 0.029), identifying it as an important determinant of residual risk. In contrast, age, male sex, body mass index, hypertension, and smoking were not independently associated with the outcome after multivariable adjustment.

Fig. 3.

Primary Composite Outcome, according to treatment strategy

Regarding ventricular function and coronary anatomy, left ventricular ejection fraction was not significantly associated with the outcome (HR 0.99; 95% CI 0.97–1.01; p = 0.252). A trend toward higher risk was observed in patients with triple-vessel disease, showing a 46% relative increase in risk, although this did not reach statistical significance (HR 1.46; 95% CI 0.98–2.18; p = 0.066). Renal function, assessed by serum creatinine, was also not significantly associated with the outcome (HR 1.07; 95% CI 0.66–1.73; p = 0.786). Among metabolic and laboratory parameters, higher triglyceride levels were associated with an increased risk of events related to the treated artery (HR 1.00; 95% CI 1.00–1.00; p = 0.038), suggesting that even small elevations may contribute to cumulative risk over time. In contrast, LDL cholesterol, glycated hemoglobin, and HDL cholesterol were not independently associated with the outcome in the adjusted model.

Competing risks analysis

To account for competing events, particularly death before the occurrence of the event of interest, a Fine–Gray competing risks analysis was performed (Fig. 5S and Table 2S). The results robustly confirmed those of the multivariable Cox analysis, demonstrating an independent association between treatment strategies—particularly when revascularization was complete—and a lower risk of events related to the treated arteries, even after adjusting for clinical, laboratory, and anatomical factors.

Impact of completeness and clinical variables

To further assess the impact of revascularization completeness on events related to the treated artery, a second multivariable analysis was conducted, categorizing patients into three groups: no revascularization, incomplete revascularization, or complete revascularization (Fig. 5S; Table 3 S). After adjustment, complete revascularization was associated with a 43% relative risk reduction compared with no revascularization (HR 0.57; 95% CI 0.39–0.84; p = 0.004). Incomplete revascularization was also associated with a lower risk (38% relative reduction), although this trend did not reach statistical significance (HR 0.62; 95% CI 0.38–1.03; p = 0.066). Among the analyzed covariates, diabetes mellitus emerged as the sole independent predictor of increased risk (HR 1.55; 95% CI 1.07–2.24; p = 0.019).

Extended competing risks analysis

A competing risks analysis using the Fine–Gray model was subsequently performed to evaluate the cumulative incidence of events, treating death from other causes as a competing event (Fig. 6S). Consistent with the Cox model, both revascularization strategies demonstrated a protective effect compared with no intervention. Incomplete revascularization was associated with a 41% relative reduction in the sub-distribution hazard (subHR 0.59; 95% CI 0.36–0.98; p = 0.041), whereas complete revascularization showed a more pronounced 46% risk reduction (subHR 0.54; 95% CI 0.37–0.80; p = 0.002). These findings reinforce and extend the primary findings, demonstrating that the clinical benefit of complete revascularization remains robust even after rigorous control for competing risks. In this model, diabetes mellitus remained an independent predictor of adverse outcomes (subHR 1.51; 95% CI 1.04–2.18; p = 0.029). The remaining clinical, laboratory, and anatomical variables—including age, ventricular function, renal function, lipid profile, and the number of diseased vessels—showed no statistically significant association with the primary outcome.

Discussion

The routine use of MACE as the primary endpoint in myocardial revascularization trials has shaped decades of comparative research between coronary interventions. However, the aggregation of heterogeneous systemic events may obscure relevant mechanistic differences between strategies, resulting in statistical neutrality. Moreover, the widespread adoption of this endpoint should not be mistaken for its conceptual appropriateness.

When the goal is to evaluate the technical and anatomical efficacy of revascularization strategies, the use of MACE requires critical reappraisal: by combining heterogeneous events, it fails to answer which strategy best protects the territory supplied by the treated vessel. Methodological rigor therefore demands vessel-specific analysis. This is because all-cause death, myocardial infarction, stroke, and unplanned revascularization involve distinct biological pathways and varying degrees of causal relationship with the index procedure.

The present study prospectively evaluated long-term outcomes in patients with stable (CAD), comparing three main strategies: (CABG), (PCI), and (OMT). In an analysis of a cohort of 1,695 participants followed for 5 years, the interventional approaches (CABG and PCI) indicating that CABG and PCI were associated with lower rates of vessel-specific coronary events compared with (OMT) in reducing events related to the treated arteries.

This superiority manifested as a lower cumulative incidence of adverse events in the interventional groups, as observed in Kaplan-Meier curves and confirmed by proportional hazards analyses. The pooled analysis (CABG + PCI versus OMT) revealed progressive divergence in event-free survival curves starting from the third year, with p < 0.001 (log-rank test), reinforcing the sustained benefit of invasive strategies.

The robust application of statistical methods — including Kaplan-Meier models, log-rank test, and the Fine-Gray competing risks model — ensured the validity and precision of the findings, detecting significant differences among strategies (global p < 0.001). The Fine-Gray model, adopted to account for competing events (such as non-cardiovascular death), confirmed higher cumulative incidence in the OMT group (p < 0.001), with overlapping trajectories between CABG and PCI (p = 0.32).

Multivariable Cox regression analyses adjusted for clinical, laboratory, and anatomical variables showed relative risk reductions of 46% for CABG (HR 0.54; 95% CI 0.33–0.86; p = 0.010) and 33% for PCI (HR 0.67; 95% CI 0.45–1.00; p = 0.047) compared with OMT. These associations remained robust in the Fine-Gray model even after confounder adjustment, with diabetes mellitus emerging as an independent predictor of residual risk (HR 1.51; 95% CI 1.04–2.18; p = 0.029).

An innovative and clinically relevant aspect was the analysis focused exclusively on events related to the treated vessel, dispensing with the composite MACE endpoint. This vessel-specific approach allowed a more precise assessment of the local impact of interventions, avoiding dilution by unrelated systemic events. The results showed a substantially higher mean event burden per patient in OMT (0.179) compared with PCI (0.073) and CABG (0.058), highlighting the relative ineffectiveness of conservative treatment in preventing local disease progression.

When stratified by completeness of revascularization, greater protection was observed with complete revascularization (HR 0.57; 95% CI 0.39–0.84; p = 0.004 in Cox; subHR 0.54; 95% CI 0.37–0.80; p = 0.002 in Fine-Gray), followed by incomplete (subHR 0.59; 95% CI 0.36–0.98; p = 0.041), relative to no revascularization. Although no statistically significant difference was found between complete and incomplete (p = 0.432), the curves suggest a trend favoring completeness.

Baseline differences between groups — such as higher prevalence of diabetes in OMT (58.2% vs. 52.6% in PCI and 51.4% in CABG; p = 0.049) and greater proportion of three-vessel disease in CABG (76.4% vs. 59.0% in PCI; p < 0.001) — were adequately controlled in the adjusted analyses, minimizing selection bias.

These findings contrast with those of the ISCHEMIA trial [9], which compared an initial invasive versus conservative strategy in 5,179 patients with moderate to severe ischemia, showing no significant difference in the primary composite endpoint over a median of 3.2 years (13.3% vs. 15.5%; adjusted HR ≈ 0.93; p = 0.34). The inclusion of broad MACE in ISCHEMIA, encompassing systemic events and periprocedural infarctions of questionable clinical relevance, may have diluted vessel-specific differences.

In our study, the analysis restricted to events related to the treated arteries revealed significantly lower cumulative rates in the CABG (0.058) and PCI (0.073) groups versus OMT (0.179), with p < 0.001 (log-rank and Fine-Gray). Complete revascularization was associated with the greatest benefit, reinforcing its role in mitigating local atherosclerosis progression — particularly in patients with diabetes or multivessel disease. Additionally, in patients who have previously undergone coronary artery bypass grafting, commissural alignment during percutaneous aortic valve implantation is particularly relevant, as it can facilitate future coronary access and preserve options for possible subsequent coronary interventions [10].

Clinical implications

In the pre-specified endpoint of events related to the treated coronary arteries, revascularization strategies were associated with a sustained risk reduction compared with OMT alone. The magnitude and consistency of the effect across statistical models suggest a robust association. These results support the value of anatomical intervention in preventing recurrent territorial ischemic events in multivessel CAD, especially when complete revascularization is achieved.

Study limitations

We acknowledge that the retrospective, non-randomized design is an important limitation of this study. Because treatment was not randomly assigned, the possibility of selection bias and residual confounding cannot be entirely excluded. Clinical judgment, anatomical complexity, patient preferences, and physician experience may all have influenced the choice among medical therapy, percutaneous coronary intervention (PCI), and coronary artery bypass grafting (CABG). However, several factors mitigate these limitations. First, the study population was derived from a period when all three therapeutic strategies were concurrently available and routinely offered for stable multivessel coronary artery disease. Consequently, decisions were made within a clinical environment where medical therapy, PCI, and CABG were considered legitimate alternatives rather than sequential options reserved for patients with markedly different risk profiles. Notably, the sample is balanced, and all patients were concurrently eligible for all three therapeutic strategies.This concurrent availability helped create clinically comparable groups and reduced—though did not eliminate—selection bias.

Second, baseline demographic, clinical, and angiographic characteristics were substantially similar across groups, reflecting the relative therapeutic equipoise during patient enrollment. Importantly, our objective was not to establish the superiority of one modality over another, but to determine whether subsequent coronary events originated predominantly from previously treated vessels or from disease progression in untreated territories. This vessel-oriented approach minimizes some of the challenges traditionally associated with comparing global treatment strategies.

Another limitation is the absence of contemporary physiological lesion assessment tools, such as fractional flow reserve (FFR) or instantaneous wave-free ratio (iFR), which were not routinely available during the study period. Instead, treatment decisions relied on clinical evaluation, angiographic findings, and noninvasive ischemia assessments according to the prevailing standards of care. While physiological assessment would have enhanced lesion-specific characterization, its absence reflects the historical context of the cohort rather than a unique methodological deficiency.

Furthermore, while retrospective analyses are inherently dependent on the quality of recorded data, all patients were enrolled in a prospectively maintained institutional registry with systematic follow-up and standardized event adjudication. This robust framework enhanced data reliability and minimized information bias.

Finally, because this study focused specifically on events related to the treated vessels, myocardial infarction and cardiovascular death related to the target artery reveal the direct outcomes of the interventions. Therefore, these endpoints were analyzed at the vessel level rather than the patient level.

Conclusion

In this cohort of patients with multivessel CAD, both CABG and PCI were independently associated with lower incidence of treated-artery events compared with OMT over 5 years. The association remained consistent after multivariable adjustment and competing risks analysis. Completeness of revascularization further modified risk, conferring greater protection when complete.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CAD

Coronary Artery Disease

MI

Myocardial Infarction

MASS

Medicine, Angioplasty, or Surgery Study

CABG

Coronary Artery Bypass Grafting

PCI

Percutaneous Coronary Intervention

OMT

Optimal Medical Therapy

MACE

Major Adverse Cardiovascular Events

CASS

Coronary Artery Surgery Study

RITA-2

Randomized Intervention Treatment of Angina – 2

SYNTAX

Synergy Between PCI With TAXUS and Cardiac Surgery

MACCE

Major Adverse Cardiac and Cerebrovascular Events

subHR

subdistribution hazard ratios

LDL

Low Density Lipoproteins

HDL

High Density Lipoproteins

TG

Triglycerides

HbA1c

Glycated hemoglobin

PS

Propensity score

HR

Hazard ratio

Author contributions

TH, researched data and wrote the manuscript. EAB, supervised the research and contributed to interpretation of results and statistical analysis. PCR, supervised the research and contributed to interpretation of results and statistical analysis. TLS, conceived this study, contributed to discussion and reviewed the manuscript. EGL, and PRS, contributed to discussion and reviewed the manuscript. FAV, performed all statistical analysis. MST and GO, organized spreadsheets, updated data, filtered inconsistencies, and corrected results in the manuscript, IJM and EDMN, supervised the research and contributed to interpretation of results and statistical analysis.CVS, JAFR and R.K.F. contributed to review the manuscript. All authors approved the final version of the manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. The study protocol was reviewed and approved by the Institutional Review Board/Ethics Committee of the participating institution. All patients provided free and informed written consent to participate in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.