STICH-3.0 International Trial Consortium

• Study rationale The evidence from observational studies comparing long-term outcomes between percutaneous coronary intervention (PCI), coronary artery bypass graft (CABG), and medical therapy in patients with iLVSD is inconsistent and likely affected by selection bias, with lower risk patients being accepted for CABG than PCI. In a large observational study including 4,794 propensity-matched patients with a LVEF <35% and CAD involving either the left anterior descending artery or multiple vessels, PCI was associated with a significantly higher risk of long-term mortality and of major adverse cardiovascular events (MACE) compared with CABG, but the risk of stroke was lower with PCI. A network meta-analysis from 2020 comparing CABG, PCI, and medical therapy (23 studies, 23,633 patients) suggested that CABG was the preferred strategy, with a reduced risk of mortality, cardiac death, MI, and urgent repeat revascularization (RR) compared to both medical therapy and PCI. A population level cohort study from Sweden including 2,509 patients from 2000-2018 with a left ventricular ejection fraction <50% also reported a significantly lower risk of death with CABG vs. PCI (odds ratio 0.62; 95%CI 0.41-0.96, p= 0.03).

There is currently a lack of evidence from randomized trials powered to compare all-cause mortality between contemporary state-of-the-art PCI vs. CABG in patients with multivessel CAD and iLVSD. Given the current clinical equipoise between both revascularization modalities in this high-risk population, a definitive answer is critically required to guide clinical practice.

The International STICH 3.0 International Trial Consortium has been established to address this ongoing conundrum in the STICH 3.0 International Collaborative Study.

• Study objective The primary objective of the STICH 3.0 Study is to determine whether CABG is superior to PCI in terms of all-cause mortality at 7 years in patients with severe CAD and iLVSD.
• Study design The STICH 3.0 study is composed of several individual national trials that were all prospectively designed by a multi-national steering committee to be combined in an international prospective individual patient data meta-analysis powered to examine whether CABG compared to PCI reduces the risk of death in patients with iLVSD. Individual patient data from these national RCTs, which are independently powered for different composite efficacy endpoints, will be pooled to compare mortality risk with PCI vs. CABG in patients with iLVSD. Each national study will be powered for a specific primary composite endpoint and all will collect all-cause mortality as a secondary endpoint. Harmonizing the study design elements across the national trials will enable data pooling to compare both revascularization modalities in terms of mortality with enough power. Patients with similar inclusion and exclusion criteria across trials will be merged in a large dataset powered for the objective of the international STICH 3.0 study.

3.1 National RCTs involved in the STICH 3.0 study Sweden: The primary objective is to determine whether PCI is non-inferior to CABG for revascularization in 470 patients with ischemic heart failure and LVEF≤40 in terms of the composite of death, stroke, non-procedural myocardial infarction or heart failure hospitalization at 3 years.

Canada: The primary objective is to determine whether CABG compared to PCI is superior in reducing all-cause death, stroke, spontaneous myocardial infarction (MI), urgent repeat revascularization (RR), or heart failure (HF) readmission over a median follow-up of 5 years in 754 patients with multivessel/LM CAD and iLVSD.

Australia and New Zealand: The primary objective is to determine whether PCI is non-inferior in to CABG in terms of the number of days alive out of hospital (DAOH) over an average of 3 years in 192 patients with iLVSD (LVEF≤40%) and multi-vessel CAD who are eligible for revascularization.

Denmark: The primary objective is to determine whether PCI is non-inferior to CABG for the composite of all-cause mortality, stroke, MI, or hospitalization for HF at 5 years in 1,550 patients with CAD and at least one of the following: 1) severe renal disease; 2) HF (irrespective of LVEF); 3) severe CAD (proximal LAD disease or 3-vessel disease or with a SYNTAX score <23; or LM disease with a SYNTAX score <33).

Germany: The primary objective is to determine whether CABG is superior to PCI in 440 patients with relevant CAD and LVEF≤40% in terms of death or cardiovascular re-hospitalization at 5 years.

UK: The primary objective is to determine whether CABG is superior to PCI in terms of all-cause mortality or cardiovascular hospitalisation at a median follow-up of 5 years in 630 patients with significant CAD and LVEF≤40%.

USA: The primary objective is to compare the rates of mortality, stroke, non-procedural MI, or CV hospitalization in patients with CAD and LVEF≤40%.

Netherlands: The primary objective is to determine whether CABG is superior to PCI in reducing the composite of all-cause mortality, stroke, non-procedural MI, cardiovascular hospitalization, and Days Alive and Out of the Hospital (composite endpoint tested in an hierarchical way - win-ratio) in 400 patients with CAD and LVEF≤40%.

3.2 Study endpoints and definitions The primary endpoint is all-cause mortality. Secondary endpoints that are collected in each national trial include: Death, stroke or non-procedural myocardial infarction; Death or heart failure hospitalization; Heart failure hospitalization; Coronary revascularization; Death or myocardial infarction; Death or stroke. In-hospital complications after index procedure will also be collected as safety events. The definitions used by each national study will be used for secondary endpoint analysis of the STICH 3.0 International Collaborative Study.

3.3 Study interventions All study participants will be treated with optimal medical therapy based on national contemporary clinical practice guidelines, and at the discretion of the Heart Teams, including the use of beta-blockers; ACEi, ARB, or ARNi; MRAs; ivabradine; diuretics; SGLT2i; and devices (ICD or CRT-D). Medical therapy for secondary prevention of CAD will also be individualized by Heart Teams (antiplatelet therapy, lipid-lowering agents, etc.). Complete revascularization is the goal for all study participants.

Revascularization by PCI: For patients randomized to PCI, contemporary best practices will be encouraged, including the use of intracoronary physiology (FFR, iFR, dPR, or other technologies) and of intracoronary imaging (IVUS and/or OCTs). CTOs will be treated by dedicated operators. Staged procedures are allowed if necessary.

Revascularization by CABG: For patients randomized to CABG, the choice of performing on-pump or off-pump surgery will be left at the discretion of the operator based on the local expertise. The left internal thoracic artery is the preferred graft for the LAD. Minimally invasive surgery is not allowed. The use of multi-arterial graft, intraoperative echocardiography, epiaortic ultrasound, and anesthetic technique will also be tailored to the patient and the center, with best contemporary practices encouraged.

• Participants Patients with multi-vessel disease for which revascularization is deemed appropriate and suitable with both PCI and CABG by the local heart team are eligible for randomization if they meet all the inclusion criteria, and no exclusion criteria.
• Data collection and pooling Baseline and procedural data as well as endpoints will be collected separately for each national trial. The different national case report forms will contain all key datapoints for the international prospective individual patient data meta-analysis. These key elements of the case report forms will include: age, LVEF, LVEF imaging modality, coronary anatomy, myocardial viability, sex, height, weight, race, ethnicity, co-morbidities, past medical history, ACS status, baseline medication, index procedural details and adverse events, and endpoints.

Denominalized and cleaned databases will be transferred to the data coordinating center and merged for analyses.

• Statistical analyses The superiority of CABG over PCI will be evaluated using a Bayesian Poisson regression model to analyses time-to-event data with baseline hazards stratified by trial and a random slope to account for variation in treatment effect between trials, using non-informative priors.

6.1 Sample size This is a superiority analysis. Assuming low heterogeneity in the log rate ratio of all-cause death between national trials and a cumulative incidence of death of 50% at 7 years in the control group, the first analysis, conducted after after all randomized patients have reached at least 60 days of follow-up and the median follow-up across all patients is at least 4 years, will have at least 80% power to detect a rate ratio of 0.85 in favour of CABG if 2,000 patients can be analysed, and at least 90% power to detect a rate ratio of 0.85 if at least 2800 patients can be analysed.

• Study organization The executive committee of the international prospective individual patient data meta-analysis will include principal investigators of each participating national RCT and 2 methodologists from the CTSU at the University of Oxford, and will be responsible for the operations of the analysis. The data coordinating center located at CTSU of the University of Oxford will be responsible for collecting, cleaning, and harmonizing the data from all national RCTs required for this study. The statistical analyses will be performed at CTSU of the University of Oxford.
• Ethics Participants of each national RCT will provide their informed consent for their data to be included in the analysis described in this protocol at the time of providing their consent to the trial.

Data will be handled in compliance with the national regulations in place, including the the EU General Data Protection Regulation where applicable. The content of the informed consent form complies with relevant integrity and data protection legislation. In the subject information and the informed consent form, the subject will be given complete information about how collection, use and publication of their study data will take place. The subject information and the informed consent form will explain how study data are stored to maintain confidentiality in accordance with national data legislation. All information processed by the sponsor will be pseudonymized and identified with study code. National patient information leaflets and consent forms will contain clear wording that allows pseudo-anonymised data to leave each country and to be received by the University of Oxford. The informed consent form will also explain that for verification of the data, authorized representatives of the sponsor, as well as relevant authority, may require access to parts of medical records or study records that are relevant to the study, including the subject's medical history.

• Publication policy The individual national trials will all be registered at clinicaltrials.gov. before recruitment of their first subject. A detailed study protocol will be published before start of the pooled analysis in a peer reviewed. The Principal Investigator will co-ordinate dissemination of data for this study. Authorship will be determined according to the degree of engagement in the study as decided by the steering committee. A detailed separate publication plan shall be prepared by the steering committee for all secondary papers.

Inclusion criteria

• Age >18 years
• LVEF ≤40%
• Extensive coronary artery disease (typically defined as a British Cardiovascular Intervention Society jeopardy score of ≥6, on a scale from 0 to 12, with higher scores indicating greater extent of disease).

Exclusion criteria

• Decompensated heart failure requiring inotropic/adrenergic support, invasive or non-invasive ventilation or intra-aortic balloon pump/ ventricular assist device therapy less than 48 hours prior to randomization;
• Valvular heart disease or any other cardiac conditions (e.g. LV aneurysm) indicating the need for surgical repair/replacement;
• Prohibitive bleeding risk or clinical scenario mandating avoidance of long-term dual antiplatelet therapy;
• Pregnancy;
• Circumstances likely to lead to poor treatment adherence;
• STEMI within 72 hours;
• PCI within 3 months.