Data were extracted following an intention-to-treat approach in Covidence. The investigators resolved disagreements through consensus or by third review (T.Z.). Data obtained from the studies were study characteristics, study design, baseline demographics of the enrolled population, infarct characteristics, procedural or treatment characteristics, and details of the interventions performed in each arm.

The primary outcome was MACE, defined as a composite endpoint of all-cause mortality, MI and unplanned revascularization, at a minimum one-year follow-up. This endpoint was estimated by manually combining the reported number of events for the individual components of this composite endpoint. To address the potential of double counting patients experiencing multiple components of MACE, we conducted additional analysis using the primary outcome reported by the individual RCTs. Our prespecified secondary outcomes were the individual components of MACE, death from cardiovascular cause, all-cause revascularization, coronary artery bypass grafting (CABG) and stent thrombosis at one-year follow-up or later. Safety outcomes included cerebrovascular events, major bleeding events, rehospitalization due to heart failure and unstable angina, contrast-associated events, and periprocedural MI.

Two reviewers (T.S. and A.D.A.) independently assessed the quality of the included RCTs using the second version of the Cochrane Collaboration’s tool for assessing the risk of bias in randomized trials (RoB 2). Any disagreements were resolved through consensus or via a third reviewer (T.Z.). We included all eligible studies in the manuscript regardless of their quality.

We used DerSimonian and Laird random effects meta-analytic models with inverse variance weighing and Jackson and modified Knapp-Hartung method extensions to pool crude count data across all included RCTs and obtain risk ratios (RRs) and corresponding 95% confidence intervals (CIs). The primary comparison was the risk of MACE at a minimum one-year follow-up in patients with MI and multivessel CAD randomized to FFR-guided complete versus culprit-only revascularization. Heterogeneity was assessed using the I ² and Tau ² statistics. All data were analyzed using Review Manager Cochrane software.

Details on the characteristics of included RCTs are reported in Table 1 . The 3 included RCTs were multicenter trials conducted exclusively in European countries, except for the FULL-REVASC trial which also included centers in Australia and New Zealand. ^,, Participants were randomized to receive either FFR-guided complete revascularization ( n = 1,373) or culprit-only revascularization ( n = 1,681). Of note, COMPARE-ACUTE used a 1:2 randomization scheme while the other 2 trials randomized participants in a 1:1 manner. Across the 3 trials, participants randomized to FFR-guided complete revascularization required flow limitation of FFR ≤ 0.8 in noninfarct related arteries (IRA) to receive stenting. The median follow-up period ranged from 1.0 to 4.7 years across the 3 RCTs.

The baseline characteristics of included participants are presented in Table 2 . The majority of participants (77.5%) were male, and the mean age ranged from 61 to 66 years. Overall, baseline characteristics across the 3 trials were well-balanced. Less than 10% of participants in both treatment arms had a history of previous MI. Only 2 of the 3 included trials reported previous PCI among participants, which showed no difference between treatment groups. ^, The FULL-REVASC trial was the only RCT to also enroll high-risk NSTEMI patients (8.6%). The most common comorbidity across included RCTs was hypertension (48.5%), and 40.4% of enrolled participants were actively smoking. Infarct-related baseline characteristics are provided in Supplemental Table 4 . In all 3 trials, the inferior wall was the most observed infarct location, with the FULL-REVASC trial specifically identifying the right coronary artery as the culprit vessel.

Treatment characteristics, stent characteristics, median length of hospital stay, mean volume of contrast used during index PCI and medical management at discharge, are presented in Table 3 . Stent diameter and total length and number used were evenly distributed across both treatment groups in all 3 trials. In all 3 RCTs, the ratio of number of stents used per patient during index hospitalization was 2:1 for FFR-guided complete versus culprit-only revascularization. Of note, FFR-guided revascularization increased contrast use by 20mL in 2 trials and by over 100 mL in the DANAMI-3-PRIMULTI trial, compared to the culprit-only arms. ^,, Additionally, FFR-guided complete revascularization was associated with longer index procedure duration than culprit-only revascularization. Participants in all 3 trials had a median length of stay of 4 days.