Outcomes in provisional one-stent versus dedicated two-stent coronary bifurcation stenting techniques: a systematic review and meta-analysis Authors: Harsh Agrawal, Pooja Dubey, Jay Tewari, Shubhechha Neupane, Vanshika Singh, Kendrick Shunk, Muditha Perera, Alok Kumar Dwivedi, Debabrata Mukherjee Published: 2026 ABSTRACT AbstractBackgroundThe optimal long-term strategy for coronary bifurcation PCI remains debated. Earlier meta-analyses limited to long-term data suggested better outcomes with a provisional approach versus routine two-stent techniques.MethodsWe conducted an updated, comprehensive meta-analysis of randomized controlled trials comparing provisional versus two-stent strategies. MEDLINE, Embase, and the Cochrane Library were searched through September 2025. Fifteen RCTs (n = 6978) met inclusion criteria. Using Stata 16.1, random-effects (DerSimonian–Laird) risk ratios (RRs) with 95% confidence intervals (CIs) were calculated for prespecified outcomes.ResultsRelative risks (95% CIs) for provisional vs two-stent were: all-cause mortality 0.97 (0.72–1.30); cardiovascular mortality 0.98 (0.68–1.40); myocardial infarction 0.86 (0.62–1.19); target lesion revascularization 1.07 (0.80–1.44); stent thrombosis 1.36 (0.81–2.29); and MACE 1.27 (0.81–1.99). Across endpoints, pooled estimates did not show statistically significant differences between strategies.ConclusionIn this updated synthesis of randomized trials, the available evidence did not demonstrate clear superiority of either provisional or two-stent strategies for major clinical outcomes. Unlike prior long-term-only analyses, we did not observe higher mortality or myocardial infarction with two-stent approaches. Technique selection should remain individualized according to lesion anatomy, procedural complexity, and operator expertise. These findings should not be interpreted as proof of equivalence or non-inferiority.Trial registration CRD420251167534.Supplementary InformationThe online version contains supplementary material available at 10.1186/s43044-026-00757-0.Keywords: Percutaneous coronary intervention, Bifurcation stenting, Complex PCI Background Clinically significant coronary bifurcation disease is not uncommon and presents unique challenges. Observational and trial data suggest that 15–20% of all lesions undergoing percutaneous coronary intervention (PCI) involve bifurcation coronary disease [1, 2] and are associated with poorer outcomes when compared to interventions on non-bifurcation lesions [3]. Early randomized controlled trials compared one-stent provisional and dedicated two-stent interventional techniques; initially presenting short-term data that favored provisional stenting or were equivocal [4–7]. Later trials were larger, examined different techniques such as double crush and T-and-protrusion, and included left main interventions [8–15]. Prior meta-analyses have compared one- and two-stent strategies; however, each included limited or sub-optimal data, such as observational data [16]. Two recent meta-analyses by Nairooz et al. and Ford et al. only used long-term data and found that provisional stenting was associated with lower all-cause mortality [17, 18]. Contemporary bifurcation PCI practice generally supports a provisional-first approach for most bifurcation lesions, while recognizing that selected complex bifurcations may require an upfront planned two-stent strategy. European Bifurcation Club consensus documents emphasize the importance of lesion anatomy, side-branch size and disease length, proximal optimization, final kissing balloon inflation when appropriate, and operator expertise in selecting and executing bifurcation PCI strategies. Therefore, pooled comparisons of provisional and two-stent approaches should be interpreted in the context of substantial clinical and procedural heterogeneity across trials. We aimed to perform an updated systematic review and meta-analysis of randomized controlled trials comparing provisional one-stent versus dedicated two-stent strategies for coronary bifurcation PCI. Methods The study protocol was registered with PROSPERO under registration number CRD420251167534. This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. MEDLINE, Embase, and the Cochrane Library were searched from inception through September 2025 for English-language, peer-reviewed randomized controlled trials comparing provisional one-stent and two-stent strategies for coronary bifurcation PCI. Search terms included combinations of “provisional,” “bifurcation,” “complex,” “one-stent,” “single-stent,” “two-stent,” “double-stent,” “DK crush,” “culotte,” “T-stenting,” “TAP,” “randomized,” and “randomized controlled trial.” Eligible studies were randomized controlled trials comparing a provisional one-stent strategy with an upfront or dedicated two-stent strategy in patients undergoing PCI for coronary bifurcation lesions. Trials comparing two different two-stent techniques without a provisional comparator, observational studies, conference abstracts without peer-reviewed full-text publication, and non-English articles were excluded. All-cause mortality was the primary outcome of interest. Additional outcomes were major adverse cardiac events (MACE), myocardial infarction, target lesion revascularization, and stent thrombosis. Because only five studies directly reported all-cause mortality, we performed an exploratory mortality analysis in which cardiac death was used when all-cause mortality was unavailable. This approach was used only to maximize study-level information and should not be interpreted as a validated surrogate endpoint analysis, because cardiac mortality is not interchangeable with all-cause mortality and may underestimate non-cardiovascular deaths. Outcome definitions were accepted as reported by each individual trial. Definitions of myocardial infarction and MACE were not fully uniform across studies. MACE variably included combinations of death, cardiac death, myocardial infarction, target lesion revascularization, target vessel revascularization, stent thrombosis, or target lesion failure depending on the trial. Because individual patient-level data were unavailable, outcomes were pooled according to the definitions used in the original publications. This variability was considered an important source of clinical and methodological heterogeneity. Two authors (HA and PD) independently extracted data on study population size, patient demographics, stenting strategies, and outcomes. Relative risks (RR) with 95% confidence intervals were calculated using the Mantel–Haenszel method for each study outcome. Pooled effects with a 95% confidence interval for the outcomes were obtained using a random effects model, which used the DerSimonian and Laird method. In the case where a study contained a zero cell, 0.5 was added to derive finite variance estimators. A random effects model was used because the observed estimates of treatment effect can vary across studies due to real differences in treatment effect as well as differences in study populations. Higgins I2 statistic and the test of heterogeneity were used to assess heterogeneity between studies. An I2 > 50% was considered indicative of significant heterogeneity. Subgroup analyses were performed by follow-up duration, study size, and publication era when feasible. A formal technique-specific subgroup analysis by individual two-stent technique was considered but not performed because the number of trials per technique was small, several trials used mixed or protocol-dependent two-stent approaches, and procedural optimization practices differed substantially across studies. Therefore, pooled comparisons should not be interpreted as evidence that all two-stent techniques have equivalent efficacy or safety. Statistical significance level of ps was set at 0.05. Publication bias was assessed using funnel plots and Egger’s regression test when feasible. These assessments were interpreted cautiously because the number of studies was small for several endpoints, particularly MACE and directly reported all-cause mortality. Statistical analyses were conducted in Stata 16.1 (Stata Corp., College Station, TX, USA). Results Fifteen randomized controlled trials met our inclusion criteria. The PRISMA study flow diagram has been shown in Fig. 1. A total of 6978 patients were enrolled in these studies: 4053 patients in the provisional strategy and 2925 patients in the two-stent strategy. The clinical follow-up time of these studies ranged from 6 to 60 months. Study and population characteristics are summarized in Table 1. Egger’s regression test did not show statistically significant evidence of small-study effects; however, this finding should be interpreted cautiously because the number of studies was limited for several endpoints. PRISMA study flow diagram Table 1Study characteristicsStudyDKCRUSH-VDKCRUSH-IIEBC twoFerencBBS oneCACTUSColomboPublication year2017201720162008200920082003StrategyProvisional vs. DK crushProvisional vs. DK crushProvisional vs. culotteProvisional vs. routine T-stentingSimple vs. complexProvisional vs. crushStent/PTCA vs. stent/stentN242|240183|183103|97101|101250|250173|17723|63Follow up time (months)12601212966DemographicsMean age (years) ± SD64 ± 10|65 ± 964.7 ± 10|63.9 ± 1162.9 ± 10.8|63.5 ± 12.166.7 ± 9.2|66.9 ± 10.564 ± 10|64 ± 1167 ± 10|65 ± 1062 ± 9|63 ± 10Male (%)78|8379|7985|7879|7877|7776|8091|76Risk factorsDiabetes mellitus (%)25.6|28.823.1|19.625|3125.7|18.813|1122|23.726|21Hyperlipidemia (%)47.5|47.529.1|33.770|7076|7670.5|63.8Hypertension (%)64.5|72.960.9|65.263|6892.1|89.157|6279.8|70.6Smoking (%)32.2|34.256|509.9|13.917|1716.8|20.3Renal dysfunction (%)14.4|17.1––20.8|23.7–––Medical historyMI %21.1|21.714.2|17.439|4118.8|20.823|2535.3|44.6–PCI %17.8|13.820.9|21.240|4144.6|51.517|1626.6|31.1–CABG %0.8|0.80.5|0–4|35.8|4.5–LVEF %60 ± 9|59 ± 9––59 ± 12|61 ± 12–57|55–Clinical presentationSilent ischemia (%)4.1|2.93.8|1.6–––13.3|17.5–Stable angina (%)10.4|14.211|15.369|68–36.4|31.1–Unstable angina (%)74.4|7068.7|66.8––11|1347.4|4417|17Procedural characteristicsTransradial access (%)74.8|77.9–63|57–34|29––GP IIb/IIIa (%)19|21.7–4|7–28|4417.3|22.636.4|42.9Kissing balloon inflation performed (%)78.9|99.6–94|96100|10029|7690.2|92.181.8|90.5Stent diameter mm ± SD3.29 ± 0.38|3.32 ± 0.372.79 ± 0.48|2.87 ± 0.493.06 ± 0.32|3.03 ± 0.33–3|3.2––LM segment length mm ± SD28.8 ± 10.4|27.9 ± 9.928.8 ± 13.5|28.6 ± 12.423.4 ± 4.8|22.9 ± 5.121.7 ± 7.5|20.9 ± 8.2–––Side Branch diameter mm ± SD2.97 ± 0.38|2.92 ± 0.35–2.61 ± 0.29|2.72 ± 0.25––––Side branch/second-vessel length mm ± SD21.25 ± 7.44|21 ± 7.32–19.9 ± 6.8|20.7 ± 5.510.4 ± 4.1|9.9 ± 4.2–––Number of vessels affected (%)1–34.6|30.476|6534.7|25.869|73––2––17|2927.7|26.725|22––3––6|537.6|47.56|5––Location of bifurcation (%)–––––––LAD/D159.9|58.858.8|60.978|7775.2|73.381|84––LCx/OM48.8|50.416.5|12.515|1915.8|20.814|11––RCA/PDA/PLV64.5|62.58.8|9.26|48.9|5.94|5––Medina classification (%)–––––––1,1,178.5|8578.7|84.281|6835.6|30.760|60––1,1,0–––11.9|9.910|8––1,0,1––6|77.9|5.98|10––1,0,0–––3|24|5––0,1,121.5|1521.3|15.812|2424.8|31.713|14––StudyCrossPerfectNordicPanSmart strategyDefinition IIEBC mainCrossover x SB openingPublication year20152015201320042016202020212021StrategyControl vs. experimentalControl vs. experimentalMain vessel vs. Main vessel + side branchSimple vs. complexConservative vs. aggressiveProvisional vs. two stentProvisional vs. dual stentSimple crossover vs. SB openingN155|151206|213202|20247|44128|130325|328230|2371685|509Follow up time (months)8860636121260DemographicsMean age (years) ± SD61 ± 7.9|61 ± 9.261.1 ± 8.8|60.9 ± 8.963 ± 10|63 ± 1061 ± 10|58 ± 1161.8 ± 10.1|61.5 ± 10.264 ± 10|63 ± 1170.8 ± 10.1|71.4 ± 9.863.4 ± 11.1|63.7 ± 10.6Male (%)67|7175|7576|7872|8677|7879|7476|79Risk factorsDiabetes mellitus (%)29|30.529.1|25.813|1242|3928.9|25.4116 (35.7)|112 (34.1)66 (29%)|62 (27%)562 (33.4)|167 (32.8)Hyperlipidemia (%)49.7|4757.3|6278|7253|4112.5|13.1223 (68.6)|227 (69.2)158 (70%)|166 (72%)628 (37.3)|207 (40.7)Hypertension (%)58.7|55.655.3|55.454|5859|5754.7|57.7230 (70.1)|215 (66.2)180 (79%)|190 (82%)961 (57.0)|289 (56.8)Smoking (%)25.2|33.132.5|25.438|5225.8|17.798 (30.2)|93 (28.4)36 (16%)|30 (13%)Renal dysfunction (%)0|2.60.5|0.5––1.6|3.148 (14.8)|59 (18)12 (5%)|9 (4%)527 (31.3)|152 (29.9)Medical historyMI %––––7|7.712.9|11.926|28PCI %–25|25–10.9|6.916.6|19.841|4311.1|13.8CABG %–4|3–0|0.80.6|0LVEF %62.2 ± 5.7|60.9 ± 7.059.5 ± 7.2|60.4 ± 6.8–60 ± 11|55 ± 1160.5 ± 7.3|59.3 ± 10.760 ± 10|59 ± 1058.43 ± 10.17|59.41 ± 8.77Clinical presentationSilent ischemia (%)––2|1–10.2|5.45.2|5.210.9|13.5Stable angina (%)54.2|4962|61.366|65–62.5|63.121.8|24.113.5|837.7|42.0Unstable angina (%)40|43.731.7|34.932|34–20.3|23.850.5|48.83.5|3.062.3|58.0Procedural characteristicsTransradial access (%)35.5|37.112.1|11.7–––80.6|78.771|70GP IIb/IIIa (%)––51|51––14.8|18.35|4Kissing balloon inflation performed (%)––32|7460|77–89|60|76.8Stent diameter mm ± SD3.3 ± 0.3|3.5 ± 2.23.3 ± 0.3|3.3 ± 0.32.9 ± 0.3|2.9 ± 0.33.3 ± 0.4|3.3 ± 0.43.02 ± 0.32|3.05 ± 0.323.8 ± 0.5|3.6 ± 0.63.34 ± 0.62|3.34 ± 0.62LM segment length mm ± SD33.0 ± 14.8|33.2 ± 13.136.9 ± 15.3|37.3 ± 14.723 ± 10|23 ± 1025 ± 12|26 ± 924.9 ± 5.6|25.1 ± 5.322.1 ± 7|22.1 ± 7Side Branch diameter mm ± SD2.8|2.6 ± 0.12.7 ± 0.2|2.7 ± 0.2––2.8 ± 0.2|2.9 ± 0.42.76 ± 0.38|2.76 ± 0.383.5 ± 0.6|3.5 ± 0.6Side branch/second-vessel length mm ± SD30|24.7 ± 2.921.5 ± 6.9|21.4 ± 6.7––18.4 ± 7.8|17.7 ± 5.617.6 ± 6.9|19.3 ± 6.7Number of vessels affected (%)170.3|73.570.4|74.6–––227.7|23.225.2|21.6–––31.9|3.34.4|3.8––44.5|43.8Location of bifurcation (%)–––––LAD/D188.4|90.792.2|93.90.7|0.777|807|3.160.6|62.577|7747.2|47.3LCx/OM7.7|7.37.3|4.70.2|0.217|13–7.7|5.223|2316.4|9.8RCA/PDA/PLV3.9|20.5|1.40.1|0.16|7–2.8|3.77.4|4.9Medina classification (%)–––––1,1,115.7|18.962.4|65.9––6.3|3.182.5|86.390|8923.1|42.61,1,048.4|35.110.9|2.4––1.6|1.520.2|12.81,0,12.6|5.48.9|8.7––4.7|6.95.8|7.51,0,09.8|12.22|1––20.3|14.615.6|5.10,1,12.6|4.112.4|18.8––0.8|0.814.5|12.510|116.7|8.3*Values reported as provisional | two stent; SD, standard deviation; Kissing balloon inflation timing varied across trials and may refer to first-stent, final, or post-second-stent kissing balloon inflation depending on the original study report Study characteristics *Values reported as provisional | two stent; SD, standard deviation; Kissing balloon inflation timing varied across trials and may refer to first-stent, final, or post-second-stent kissing balloon inflation depending on the original study report Length variables were reported inconsistently across trials and may refer to lesion length or stent length depending on the original study. The risk of bias assessment was done using the Revised Risk of Bias Tool for Randomized Trials (RoB2). The ROB traffic light plot (Fig. 2) and summary plot (Fig. 3) were generated using the Robvis tool. Risk of bias traffic light plot Risk of bias summary plot All-cause mortality Figure 4 shows the estimated risks for each study for all-cause mortality. Forest plot for the estimated risks for all-cause mortality Primary analyses Only five studies directly reported all-cause mortality. For the exploratory mortality endpoint, in which cardiac death was used when all-cause mortality was unavailable, pooled estimates did not show a statistically significant difference between provisional and two-stent strategies (RR, 0.97; 95% CI 0.72–1.30; p = 0.82; I2 = 0.0%). Pooled estimates also did not show statistically significant differences for cardiac death (RR, 0.98; 95% CI 0.68–1.40; p = 0.89), myocardial infarction (RR, 0.86; 95% CI 0.62–1.19; p = 0.36), target lesion revascularization (RR, 1.07; 95% CI 0.80–1.44; p = 0.64), stent thrombosis (RR, 1.36; 95% CI 0.81–2.29; p = 0.25), or MACE (RR, 1.27; 95% CI 0.81–1.99; p = 0.29). These non-significant findings should not be interpreted as evidence of equivalence or non-inferiority. Heterogeneity varied by outcome. No meaningful heterogeneity was observed for the exploratory mortality endpoint, cardiac death, or stent thrombosis. In contrast, heterogeneity was moderate to substantial for myocardial infarction (I2 = 49.3%), target lesion revascularization (I2 = 44.0%), and MACE (I2 = 71.4%). This likely reflects differences across trials in lesion complexity, left main involvement, bifurcation anatomy, side-branch size and disease length, stenting technique, generation of drug-eluting stents, follow-up duration, and use of procedural optimization techniques. Subgroup analyses: Subgroup analyses by follow-up duration and study size are summarized in Table 2. These analyses did not show statistically significant differences between provisional and two-stent strategies across the evaluated outcomes. Because stent platforms and bifurcation PCI techniques evolved over time, we also explored publication era as a pragmatic proxy for procedural and technological evolution. Studies published before 2015 showed a lower risk of myocardial infarction with provisional stenting compared with two-stent strategies (RR, 0.57; 95% CI 0.38–0.84; p = 0.004; I2 = 2.2%), whereas later studies did not show a statistically significant difference. This suggests that older trials and earlier procedural techniques may have contributed to heterogeneity in myocardial infarction outcomes. Sensitivity analyses also suggested that individual studies contributed to heterogeneity in MACE estimates; therefore, these findings should be interpreted cautiously. Table 2Subgroup analyses for the exploratory mortality endpoint and secondary outcomesCharacteristic# of RCTsRR (95% CI)All-cause mortalityCardiac deathsMITLRSTMACEFollow-up duration (months)Less than or equal to 1240.9260.9850.9391.1411.3611.168(0.572, 1.498)(0.582, 1.667)(0.631, 1.396)(0.743, 1.751)(0.724, 2.560)(0.641, 2.129)Greater than 12110.9110.9670.6090.9971.3591.517(0.479, 1.731)(0.593, 1.577)(0.354, 1.049)(0.638, 1.559)(0.546, 3.386)(0.995, 2.313)Study sizeLess than or equal to 30051.0470.9030.6220.8660.6821.172(0.333, 3.293)(0.264, 3.086)(0.311, 1.244)(0.535, 1.400)(0.188, 2.479)(0.481, 2.856)Greater than 300100.960.9820.9121.1221.5551.247(0.703, 1.310)(0.675, 1.428)(0.623, 1.333)(0.781, 1.611)(0.881, 2.742)(0.857, 1.816)Overall150.9660.9750.8591.0731.361.277(0.715, 1.304)(0.682, 1.395)(0.618, 1.193)(0.798, 1.443)(0.809, 2.287)(0.829, 1.966)RR, relative risk; CI, confidence interval; MI, myocardial infarction; TLR, target lesion revascularization; ST, stent thrombosis; MACE, major adverse cardiac event Subgroup analyses for the exploratory mortality endpoint and secondary outcomes RR, relative risk; CI, confidence interval; MI, myocardial infarction; TLR, target lesion revascularization; ST, stent thrombosis; MACE, major adverse cardiac event Cardiac deaths Figure 5 shows the estimated risks for cardiac deaths. Data on cardiac deaths were available from fourteen studies. There was no significant difference in risk for cardiac deaths among the two treatment groups (pooled RR, 0.98, 95% CI 0.68–1.40, p = 0.891). I2 was 0.0% and the corresponding p for the test of heterogeneity was 0.923, providing no significant evidence of heterogeneity among the studies. Forest plot for the estimated risks for cardiac death Myocardial infarction Figure 6 shows the estimated risks for myocardial infarction. Thirteen studies reported data on myocardial infarction. The pooled relative risk for MI was 0.86 (95% CI 0.62–1.19) with a p = 0.363, indicating that there is no significant difference in MI between the two strategies. I2 was 49.3% with a p = 0.022 for the test of heterogeneity, providing significant evidence of heterogeneity among the studies. Forest plot for the estimated risks for myocardial infarction Target lesion revascularization Figure 7 shows the estimated risks for target lesion revascularization. Data on TLR were available from thirteen studies. There was no significant difference in risk for TLR among the two treatment groups (pooled RR, 1.07, 95% CI 0.80–1.44, p = 0.642). The p for the test of heterogeneity was 0.044, and the corresponding I2 was 44.0% providing borderline evidence of heterogeneity among the studies. Forest plot for the estimated risks for target lesion revascularization Stent thrombosis Figure 8 shows the estimated risks for stent thrombosis. Ten studies out of the fifteen trials reported ST data. The pooled relative risk for ST was 1.36 (95% CI 0.81–2.29) with a p = 0.246. This indicates that there is no statistically significant difference between the provisional and the two-stent techniques for ST in each of the studies. There was no significant evidence of heterogeneity among the studies (I2 = 0.0%, p = 0.627). Forest plot for the estimated risks for stent thrombosis Major adverse cardiac events Figure 9 shows the estimated risks for major adverse cardiac events. Data from only five studies were available on MACE. The pooled relative risk for MACE was 1.27 (95% CI 0.81 to 1.99) with a p = 0.294, indicating that there is no significant difference in MACE between the two strategies. Significant evidence of heterogeneity among the studies regarding MACE was observed (I2 = 71.4%, p = 0.007). Forest plot for the estimated risks for major adverse cardiac events Summary of the subgroup analysis is shown in Table 2. There were no significant differences between the provisional strategy and the two-stent strategy in any of the outcomes across the subgroups of follow-up time and study size. Visual inspection of the funnel plot for the exploratory mortality endpoint did not suggest marked asymmetry, and Egger’s regression test was not statistically significant (p = 0.641). However, this assessment was limited by the small number of available studies and should be interpreted cautiously (Fig. 10). Funnel plot for all-cause mortality Discussion In this updated meta-analysis of randomized controlled trials, pooled estimates did not demonstrate a statistically significant difference between provisional and two-stent strategies for the exploratory mortality endpoint, cardiac death, myocardial infarction, target lesion revascularization, stent thrombosis, or MACE. These findings should not be interpreted as proof of equivalence or non-inferiority, because the included trials differed in lesion complexity, bifurcation anatomy, stenting technique, follow-up duration, procedural optimization, and endpoint definitions. Publication bias assessment did not suggest small-study effects for the mortality endpoint, but this analysis was limited by the relatively small number of studies available for several outcomes. The mortality analysis requires cautious interpretation. Because only a minority of trials directly reported all-cause mortality, we used cardiac death when all-cause mortality was unavailable to create an exploratory mortality endpoint. This approach increased the number of analyzable trials but introduced conceptual limitations, as cardiac death is not interchangeable with all-cause mortality and may underestimate non-cardiovascular deaths. Differences in event adjudication and reporting across studies may therefore have influenced the pooled mortality estimate. Moderate heterogeneity was observed for MI, TLR, and especially MACE. This variability is clinically plausible because the included trials differed in lesion complexity, left main involvement, Medina anatomy, side-branch size and disease length, use of DK-crush, culotte, crush, TAP, or T-stenting techniques, intravascular imaging use, DES generation, final kissing balloon inflation, and proximal optimization practices. In addition, MACE definitions varied across trials, further contributing to methodological heterogeneity. Therefore, pooled results should be interpreted as an overall randomized-trial summary rather than as a definitive comparison of any single two-stent technique against provisional PCI. Importantly, the present analysis should not be interpreted as showing that all two-stent techniques are equivalent. DK-crush, culotte, crush, TAP, and T-stenting differ procedurally and may have different clinical performance depending on anatomy and operator expertise. Some contemporary trials enriched for complex bifurcation or distal left main disease suggest that selected planned two-stent approaches may be beneficial in anatomically complex lesions. However, the available aggregate data did not permit a sufficiently powered comparison of individual two-stent techniques. Therefore, conclusions regarding the relative performance of specific two-stent methods remain limited. These findings are broadly consistent with contemporary consensus and guideline recommendations that support provisional stenting as the default strategy for most bifurcation lesions, with planned two-stent PCI reserved for selected complex anatomies, large diseased side branches, or situations in which an acceptable side-branch result is unlikely with a provisional approach [19–21]. A recent systematic review and meta-analysis of ostial LAD bifurcation PCI also highlighted that procedural strategy selection remains anatomy-dependent, reinforcing the need to interpret pooled bifurcation PCI outcomes in the context of lesion location, side-branch involvement, and technique-specific procedural optimization [22]. The present meta-analysis supports an individualized framework but does not directly test procedural optimization steps such as proximal optimization, final kissing balloon inflation, imaging guidance, or physiology-guided side-branch assessment. In addition to including all relevant randomized controlled trials, this study is most distinguished from its predecessors by the inclusion of the DKCRUSH-V trial, which found that a planned double-kissing crush two-stent technique, compared with a provisional one-stent technique, had decreased target lesion failure at one year but no significant difference in cardiac death between the groups. The weight of this large trial (n = 482) may account for the negative result of this meta-analysis. Our findings stand at odds with previous meta-analyses, such as Nairooz et al. whose analysis only used data from studies published after 2010 with long-term data and found that a one-stent strategy was associated with reduced all-cause mortality (RR, 0.66, 95% CI 0.45–0.98). These findings are in accord with Behan et al.’s pooled analysis of BBC1 and Nordic 1 studies [12], which similarly found an association between one-stent technique and reduced mortality. Behan et al. noted a separation of all-cause mortality Kaplan–Meier curves at approximately 2 years; therefore, it may be that isolating any advantageous effects of one- over two-stent strategies are apparent only in data that reflect modern stent technology and long-term follow up. A recent observational sub-study of the EXCEL trial that has been presented but not yet published by Kandzari et al., found that patients who underwent provisional stenting compared to an up-front two-stent strategy had no difference in outcomes at 30 days but reduced composite end point of mortality, myocardial infarction, stroke, or target lesion revascularization at 3 years [23]. While this may represent a long term benefit to provisional stenting, the lack of randomization makes it more likely that the worse outcomes associated with the patients who underwent planned two-stent interventions was caused by the more severe and complex coronary artery disease which led the concerned operators to choose a two-stent strategy. While the optimal interventional strategy for significant coronary bifurcation disease remains debated, the present synthesis of randomized trial data does not support a generalized preference for one- or two-stent technique across all bifurcation lesions. Further investigation using modern stent platforms, standardized procedural optimization, and long-term follow-up is needed and may clarify whether specific anatomical subgroups benefit from specific two-stent techniques. Limitations This study has several limitations. First, all-cause mortality was not directly reported in all trials, requiring an exploratory mortality endpoint that substituted cardiac death when all-cause mortality was unavailable. The exploratory mortality endpoint may introduce misclassification bias because cardiac death and all-cause mortality are conceptually distinct outcomes. Therefore, the pooled mortality estimate should be interpreted as hypothesis-generating rather than definitive. Second, myocardial infarction and MACE definitions varied across trials, limiting direct comparability. Third, moderate to substantial heterogeneity was observed for myocardial infarction, target lesion revascularization, and MACE, likely reflecting differences in lesion complexity, left main involvement, bifurcation anatomy, stent platforms, procedural optimization, and follow-up duration. Fourth, the two-stent group included multiple techniques, including DK-crush, Culotte, crush, TAP, T-stenting, and mixed approaches, which should not be assumed to have equivalent outcomes. A formal technique-specific subgroup analysis by individual two-stent technique was considered but not performed because the number of trials per technique was small, several trials used mixed or protocol-dependent two-stent approaches, and procedural optimization practices differed substantially across studies. Fifth, publication bias testing was limited by the small number of studies available for several endpoints. Finally, this was an aggregate-data meta-analysis, and individual patient-level analyses would be needed to identify which anatomical or procedural subgroups benefit most from each strategy (Additional file 1). Conclusion The available randomized evidence does not demonstrate clear superiority of either provisional or planned two-stent strategies for major clinical outcomes. However, because included trials differed in lesion complexity, left main involvement, two-stent technique, stent generation, and optimization practices, these findings should not be interpreted as proof of equivalence among all bifurcation PCI strategies. Strategy selection should remain individualized according to anatomy, side-branch disease, procedural complexity, and operator expertise, in line with contemporary consensus recommendations. Future studies should use standardized endpoint definitions, longer follow-up, and adequately powered comparisons of specific two-stent techniques, particularly in complex bifurcation and left main disease. Supplementary Information Additional file 1: Table S1. Trial-reported definitions of myocardial infarction and major adverse cardiac events. Abbreviations Drug-eluting stent Double-kissing crush European Bifurcation Club Left main Proximal optimization technique T-and-protrusion Percutaneous coronary intervention Randomized controlled trial Confidence interval Preferred Reporting Items for Systematic reviews and Meta-Analyses Major adverse cardiac events Relative risk Risk of bias Myocardial infarction Target lesion revascularization Stent thrombosis Author contributions Conceptualization: JT, HA, DM; protocol/methodology: JT, HA, AD, MP; search strategy: JT, SN, VS, PD (oversight: HA); study selection (title/abstract and full text): JT, SN, VS, PD; data extraction and verification: JT, SN, VS, PD; risk of bias assessment (RoB 2): JT, PD, AD, HA; statistical analysis and software (meta-analysis): JT, AD, MP; visualization (figures, RoB plots, forest): JT, VS, PD; interpretation of results: JT, HA, AD, MP, SN, VS, PD, KS, DM; writing (original draft): JT, VS, HA, PD; writing (review and editing): HA, AD, MP, PD, SN, VS, KS, DM; supervision: HA, DM; project administration: JT, HA. Funding None. Data availability All data relevant to the study are included in the article or uploaded as supplementary information. Corresponding author can also be contacted on email. Declarations Ethics approval and consent to participate No human/animal subjects involved. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Harsh Agrawal, Email: harshagrawal@hotmail.com. Jay Tewari, Email: jayt.2910@gmail.com. References Associated Data Supplementary Materials Additional file 1: Table S1. Trial-reported definitions of myocardial infarction and major adverse cardiac events. Data Availability Statement All data relevant to the study are included in the article or uploaded as supplementary information. Corresponding author can also be contacted on email.