We searched PubMed, Embase, the Cochrane Library, and CNKI up to July 2024 using keywords related to “STEMI”, “FFR”, “revascularization” and “randomized controlled trials”. Exact search strategies were listed in the Supplementary Table S1. Additional studies were identified by reviewing the reference lists of included articles.

Two reviewers(XY and FL) independently screened titles and abstracts against inclusion criteria, resolving disagreements by consensus or a third reviewer. Inclusion criteria were: (1) randomized controlled trials, (2) comparison of FFR-guided CR with culprit-only revascularization(COR) percutaneous coronary intervention, and (3) STEMI patients with multivessel disease.

Two authors independently collected information from each included eligible study. The recorded information covering the first author, the total number of features and subjects, follow-up duration, mean age, gender, smoking, diabetes mellitus, hypertension, three-vessel disease and Killip class II–III. A third researcher made the ultimate determination if there were any controversies between the two researchers.

Primary outcome included major adverse cardiac events, while secondary outcomes covered all-cause mortality, ischemia-driven revascularization, repeat percutaneous coronary intervention, and safety events such as cardiac death and stroke.

Risk of bias was evaluated using the Cochrane risk of bias tool, categorizing studies into low risk, some concerns, or high risk based on five key domains (12).

Evidence quality was assessed using the Grading of Recommendations, Assessment, Development, and Evaluations approach, initially rated high but downgraded based on limitations, indirectness, inconsistency, imprecision, and other factors (Supplementary Table S4) (13).

The information was examined with Review Manager (Version 5.4) and STATA. Relative risks (RR) and 95% confidence intervals (CI) were calculated, with heterogeneity assessed by Cochran's Q test and Higgins I2 statistics. Prediction intervals (PI) were calculated to describe the distribution of true effects around the summary effect (14). We used the DerSimonian-Laird method and random-effects models for all meta-analyses, which are conservative as they consider both within- and between-study variability (15). A sensitivity analysis using the “leave-one-out” method was conducted to evaluate the impact of individual studies on the overall outcome. Publication bias was evaluated through funnel plots and Egger's regression. Trial sequential analysis (TSA) (16, 17). was conducted to ensure robustness and to assess the cumulative evidence.

The characteristics of the six articles are summarized in Table 1. These studies were published from 2012–2024. These studies included a total of 3,482 patients, and the number of patients in each study varied from 110–1,542. The summary of six studies is shown in Supplementary Table S2 and the main results extracted from the six articles are shown in Supplementary Table S3. The inverted funnel plots for the primary outcome of major adverse cardiac events and all-cause mortality or ischemia-driven revascularization did not suggest publication bias (Figure 2). The risk of bias assessment for each included trial by domain appear in Figure 3.

Four studies reported major adverse cardiac events (5, 6, 9, 20). Among these studies, 195 events (13.4%) occurred in the 1,452 patients undergoing FFR-guided CR compared with 331 events (19.2%) in 1,721 patients undergoing COR (RR: 0.65, 95% CI: 0.45–0.94, 95% PI: 0.20–2.19; P = 0.02) (Figure 4). The 95% CI (0.46–0.94) demonstrates a statistically significant reduction in major adverse cardiac events with FFR-guided CR. However, the 95% PI (0.20–2.19), which includes 1.0, suggests potential variability in the magnitude of this benefit across different populations or settings, consistent with the observed heterogeneity (I² = 74%).

Among the trials, moderate significant heterogeneity (I² = 74%) was observed, so a leave-one-out sensitivity analysis was conducted to identify potential sources of heterogeneity. Regardless of which study was excluded, there was always a significantly higher risk of major adverse cardiac events in the COR group (Supplementary Table S5). Trial sequential analysis confirmed that the required sample size (2,845) to reach definitive conclusions was achieved (Supplementary Figure S1).

Among the six studies (5, 6, 9, 18–20) reporting all-cause death, 107 deaths (6.9%) among 1,553 patients occurred with FFR-guided CR vs. 101 deaths (5.5%) among 1,832 patients with COR, which did not reveal any noteworthy variance in statistical terms between the two groups(RR: 1.12, 95% CI: 0.86–1.46, 95% PI: 0.79–1.58; P = 0.42; I² = 0%) (Figure 5A). Trial sequential analysis revealed that the required sample size (2,714) to reach safe conclusions was reached (Figure 6).

Five trials (6, 9, 18–20) provided data on ischemia-driven revascularization. For revascularization, a consistent benefit with CR was found compared with COR when an FFR-guided non-culprit lesion percutaneous coronary intervention strategy was used (RR: 0.35, 95% CI: 0.19–0.63; P = 0.00005), although there was considerable heterogeneity (I² = 79%). A leave-one-out sensitivity analysis was conducted, revealing that heterogeneity was absent when Böhm et al.'s study (9) was excluded (Supplementary Table S6). A potential explanation for these findings is the variation in guiding strategies for revascularization. In contrast to other studies included in our analysis, this particular study focused on any planned or unplanned revascularization, which is more lenient on revascularization metrics. This difference may have contributed to the observed heterogeneity. However, when we removing this study, the results showed that FFR-guided CR led to a significant reduction in ischemia-driven revascularization (RR: 0.27, 95% CI: 0.19 −0.40, 95% PI: 0.16–0.46; P < 0.00001; I² = 14%) (Figure 5B).

Repeat percutaneous coronary intervention was reported in four studies (5, 6, 19, 20). The occurrence of repeat percutaneous coronary intervention was 6.2% within FFR-guided CR group compared to 17.0% in COR group, demonstrating a marked decrease in the FFR-guided group (RR: 0.35, 95% CI: 0.22–0.56, 95% PI: 0.16–0.78; P < 0.0001; I² = 43%) (Figure 5C).

Cardiac death was assessed in four trials (6, 9, 19, 20). The incidence of cardiac death showed a borderline statistical significance favoring FFR-guided CR (RR: 0.69, 95% CI: 0.48–0.99; P = 0.05), though the 95% PI (0.38–1.25) included 1.0, indicating uncertainty in the consistency of this effect across populations. Thus, while the point estimate suggests a potential reduction in cardiac death, the variability across studies precludes definitive conclusions. No heterogeneity was detected for this outcome measure (I² = 0%) (Supplementary Figure S2A).

Similarly, no differential association of treatment was found between FFR-guided CR strategy compared with COR strategy on repeat myocardial infarction (RR: 0.91, 95% CI: 0.58–1.42, 95% PI: 0.58–1.55; P = 0.68; I² = 32%), major bleeding (RR: 1.24, 95% CI: 0.65–2.35, 95% PI: 0.28–3.27; P = 0.51; I² = 0%), coronary artery bypass grafting (RR: 0.97, 95% CI: 0.30 −3.13, 95% PI: 0.04–21.13; P = 0.95; I² = 48%), stroke (RR: 1.32, 95% CI: 0.46–3.82, 95% PI: 0.00–16,643.00; P = 0.60; I² = 29%), contrast-induced nephropathy (RR: 0.98, 95% CI: 0.75 −1.28, 95% PI: 0.18–5.52; P = 0.90; I² = 0%) and re-hospitalization (RR: 0.72, 95% CI: 0.44 −1.15, 95% PI: 0.02–33.70; P = 0.17; I² = 28%) (Supplementary Figures S2B–G).

This meta-analysis had certain drawbacks. First, the scheduling of the FFR measurement differed among the trials. In the Compare-Acute investigation (18) and the research by Joshi et al. (18), FFR was executed concurrently with the initial percutaneous coronary intervention. Conversely, in the DANAMI 3-PRIMULTI trial (6) and Lønborg et al.'s examination (19), the FFR evaluation was carried out two days post the primary procedure. The time of the FFR measurement was not strictly regulated in the following two trials (5, 9); it could be done at the operator's discretion either during the index surgery or later during the index hospitalization. Therefore, research on the instantaneous wave-free ratio (iFR) in this context is warranted (29). Secondly, different eras might have different techniques (like evolving interventional procedures), devices (like stent types), concepts, and the introduction of new drugs, which could affect the results. Thirdly, the cut-off value of FFR was different, the Ghani trial (5) used 0.75, while the rest used 0.80. Finally, since all included studies were conducted in developed countries, our findings are only applicable to those regions, and further research in other countries is needed to enhance generalizability.

Our findings suggest that FFR-guided CR reduces major adverse cardiac events in multivessel revascularization for STEMI patients. Physicians should consider patient-specific factors and monitor kidney function, given the potential for complications such as major bleeding and stroke with both FFR-guided CR and COR. Due to the higher complication risks in these patients (30), careful monitoring during and after FFR-guided revascularization is recommended. For patients with borderline lesions, FFR measurements may help guide treatment strategies based on lesion functional significance.

We suggest that trials for chronic complete occlusions be carried out in the future to determine the actual impact of FFR-guided percutaneous coronary intervention on distal and multivessel revascularization. FFR-guided percutaneous coronary intervention may still be clinically beneficial for patients with coronary artery disease, according to a single-center retrospective investigation alone (31). As a result, nothing is known about how FFR-guided percutaneous coronary intervention actually works in individuals who have persistent complete occlusions. Since stroke and cerebrovascular events were recorded in relatively few trials, we therefore suggest including them as outcomes of interest. Consequently, we were unable to look at how FFR-guided CR affected the progression of cerebrovascular events. We suggest more research to support this claim. A network indirect meta-analysis could provide valuable insights by enabling a comprehensive comparison of FFR-guided CR, COR, and standard CR. However, despite conducting an extensive literature search, the number of available randomized controlled trials was insufficient to perform a robust network meta-analysis (Two for FFR-guided CR VS conservative treatment (32, 33); six for COR VS CR (21, 34–38). This limitation underscores the need for further high-quality studies directly comparing these revascularization strategies to guide clinical practice.