Stepwise Provisional versus Planned Double Stenting Strategies in Treating Unprotected Left Main Distal Bifurcation Lesions: A Systematic Review and Meta-Analysis Comprising 11,672 Patients

Background: Provisional stenting is the preferred strategy for non-left main bifurcation lesions. However, its superiority over planned double stenting for unprotected left main distal bifurcation (UPLMB) lesions remains unclear. Previous studies have reported conflicting results. Methods: Randomised controlled trials (RCTs) and observational studies comparing the outcomes of provisional stenting to planned double stenting for UPLMB lesions were identified. The primary endpoint was major adverse cardiac events (MACE). The secondary endpoints were myocardial infarction (MI), target vessel revascularisation (TVR), target lesion revascularisation (TLR), all-cause death, cardiac death and stent thrombosis (ST). Aggregated odds ratios (OR) and 95% confidence intervals were calculated. A sensitivity analysis was conducted if I2 was >50% or p < 0.01. Publication bias analysis was considered if more than 10 studies were enrolled. Results: Two RCTs and 19 observational studies comprising 11,672 patients were enrolled. Provisional stenting had a significantly lower incidence of MACE, mainly driven by TLR and TVR. Double stenting had a significantly lower incidence of cardiac death. In addition, patients undergoing provisional stenting had a lower tendency towards the occurrence of MI, while patients undergoing double stenting had a lower tendency towards all-cause death and ST. Conclusions: A provisional stenting strategy was associated with lower MACE, TVR and TLR but higher cardiac death. Further investigation is needed through RCTs to assess which strategy performs better.


Introduction
An unprotected left main distal bifurcation (UPLMB) lesion is a lesion that involves the distal bifurcation of the left main (LM) coronary artery [1,2].It remains one of the most challenging lesions in the field of cardiac interventional therapy because of its unique anatomical location and geometry [3].LM lesions include protected and unprotected lesions based on the presence of blood supply from the vascular bridge or good collateral circulation from the right coronary artery.Among all types of coronary artery lesions, UPLMB has the worst prognosis.Currently, there are two percutaneous coronary intervention (PCI) strategies for UPLMB lesions: stepwise provisional stenting and planned double stenting.The stepwise provisional stenting strategy involves placing stents in the main vessel crossing over the side branch and another stent, if necessary, in the branch vessel.The planned double stenting strategy involves placing stents both in the main vessel and the branch vessels.The former has been proven to be the preferred strategy for non-LM bifurcation lesions [4].However, con-troversy still remains regarding which strategy is superior for UPLMB lesions.There have only been two multicentre randomised controlled trials (RCTs) addressing this issue, and they drew conflicting conclusions.In the DKCRUSH-V Registry, Chen et al. [5] concluded that provisional stenting increased the rate of target lesion revascularisation failure (TLF) and stent thrombosis (ST) over three years of follow-up.In contrast, the European Bifurcation Club Left Main (EBCLM) trial proved that provisional stenting had a lower rate of major adverse cardiac events (MACE) [6].Other observational cohort studies have also not come to consistent conclusions.Therefore, we performed this systematic review and meta-analysis to clarify which of the two interventional strategies was superior.We also compared the long-term outcomes in the drug-eluting stent (DES) era with the goal to provide convincing data-based medical evidence for selecting the best PCI plan for UPLMB patients.

Literature Searching
A comprehensive search was conducted using PubMed, Embase, Ovid Medline, Cochrane Database, Web of science, CNKI and ClinicalTrails.gov.RCTs and observational studies comparing provisional and planned double stenting for distal UPLMB disease published from library or database construction to 1 Jan. 2023, were searched.The key search terms included "left main", "provisional", "double", "one", "two", "simple" and "complex".The search terms were retrieved using a free combination method, and all relevant references were evaluated for additional studies that were not identified from the initial database searches.The search strategy is presented in Supplementary Table 1.This study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (Supplementary Table 2).

Literature Inclusion and Exclusion Criteria
Inclusion criteria were: (1) RCTs and observational studies comparing provisional stenting and planned double stenting strategies for distal UPLMB disease; (2) comparable general information between the two strategies; (3) DES stents used in both strategies; and (4) outcome indicators including at least one of MACE, all-cause death, cardiac death, myocardial infarction (MI), target vessel revascularisation (TVR), target lesion revascularisation (TLR), or ST.Exclusion criteria were: (1) incomplete or ambiguous data; (2) follow-up period of less than 6 months; and (3) studies that shared the same participants.DK-Crush 54 0,1,1 1/9 culotte 18 0,0,1 0/0 crush/mini crush 0 0,1,0 28/18 1,0,0 0/0 Post-PCI, 300 mg/day of aspirin was prescribed to all patients for one month, which was reduced to 75 mg/day to be continued indefinitely thereafter.In addition, they received clopidogrel 300 mg in divided doses for the first month, later reduced to 75 mg/day for at least one year after the PCI.

Data Extraction
Two reviewers in the research group (DL and HL) independently screened the retrieved literature and extracted information.In case of disagreement of the status of the study, it was resolved through discussion with a third reviewer (CG).The extracted data included: (1) basic information of the enrolled studies, including first author, publication year, follow-up period, and study type; (2) general data of participants, including sample size, mean age, gender ratio, ethnicity, clinical diagnosis, medication, and lesion characteristics; (3) PCI strategy, including provisional, T, V, Y, Crush, double kissing technique (DK)-Crush, culotte, etc.; (4) outcome indicators, including all-cause death, cardiac death, MI, TLR, TVR, ST, and MACE; and (5) other information such as stent type and number, intravascular ultrasound (IVUS), and proximal optimal technique (POT).

Outcomes and Definitions
The primary endpoint of this meta-analysis was MACE, defined as a composite of death, MI and TLR/TVR.The composition varied among the enrolled studies, and this review adopted the initial definition of the studies.In some articles, MACE is defined as TLF.The secondary endpoints were ST and the individual components of the primary endpoint, including all-cause death, cardiac death, MI, TLR, TVR, and ST.The definitions of every endpoint in each study are summarized in Supplementary Table 3.

Statistical Analysis
STATA/MP 17.0 (Stata Corporation, College Station, TX, USA) was used to calculate aggregated odds ratios (OR) at 95% confidence intervals.Heterogeneity between the studies was explored using the I 2 test and the fixedeffects model was used when p > 0.01 and I 2 < 50%, while the random-effects model was used if not.A heterogeneity test and sensitivity analysis were used to select the origin of heterogeneity.Contour-enhanced funnel plots, a regression-based Egger test, and non-parametric trim-andfill analysis were used to assess publication bias if the number of studies was more than 10. p-value < 5% was considered the difference was significant.

Discussion
A total of two RCTs and 19 observational studies were included in this study [5,6,[9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27].For the endpoints of MACE and TLR, the heterogeneity was relatively large, and it mainly came from the RCT subgroup.We only identified two RCTs, but they drew conflicting conclusions concerning MACE, TLR and MI, although the difference did not reach statistical significance.We believe that the heterogeneity of the two RCTs may be due to the different techniques of double stenting.In the study from Chen [5], only DK-Crush was performed for double stenting, while in the study from Hildick-Smith [6], a composition of Culotte, DK-minicrush, T or T stenting and small protrusion technique (TAP) was performed.This reminded us that DK-Crush was likely better than provisional stenting, while provisional stenting was better than other double stenting.
Subgroup analysis of RCT and non-RCT revealed that the two aggregated OR were opposing in MACE, TLR, allcause death and MI occurrences, and consistent in cardiac death and ST occurrences.Though RCTs have a higher level of evidence than in observational studies, their small size became the greatest limitation for this review.We identified publication bias only when analysing all-cause death occurrence.
We performed a nonparametric trim-and-fill analysis for the publication bias.After virtually imputing five studies, the funnel plot became symmetric, and the bias was reduced.The adjusted OR value was enlarged from 1.052 [0.872, 1.270] to 1.173 [0.984, 1.398].However, the results still favoured the double stenting strategy.
The aggregated OR values of all endpoints are displayed in Table 4.Our analysis revealed that provisional stenting had a significantly lower incidence of MACE, mainly driven by TLR and TVR and double stenting had a significantly lower incidence of cardiac death.Additionally, provisional stenting tended to have a lower occurrence of MI, while double stenting tended to have a lower occurrence of all-cause death and ST.From these results, it was hard for us to conclude which performed better.Considering the importance of survival, double stenting might be more recommended.
The latest systematic review and meta-analysis comparing the two strategies for LM was published by Abdelfattah et al. [28], in which 12 studies of 7105 patients were included.In that review, only the 2nd generation of DES was considered.However, in our pre-analysis we found that DES type didn't affect the OR value.So as to enlarge the sample size, we enrolled both the 1st and 2nd DES, and the sample size was nearly doubled.A recent large samplesized study conducted by Alasmari in 2022 [20] was added in our review.The differences in outcomes between the two meta-analyses mainly lie in the occurrences of cardiac death and MI.Vescovo et al. [29] published a network meta-analysis comparing different double stenting techniques and provisional stenting.Network meta-analysis was recommended to select a specific technique.However, detailed subdivisions reduced the sample size.As provisional stenting and double stenting were considered as two different strategies, rather than two different techniques, there was still a necessity to conduct this systematic review and meta-analysis to clarify which performed better.It could help operators make the optimal strategy when dealing with LM bifurcation lesions.

Limitations
The limitations of this study mainly lie in the definitions of endpoints that varied across studies, the double stenting techniques that varied across studies, the performance of IVUS, POT, and double balloon kissing (DBK) that varied across studies, and the long span of 2002 to 2019.At last, this review was not registered and a protocol was not prepared.

Conclusions
The provisional stenting strategy was associated with a significantly lower occurrence of MACE, mainly driven by TLR and TVR, but a higher occurrence of cardiac death.Further investigations are needed, especially those involving RCTs, to confirm which strategy performs better.

Fig. 3 .
Fig. 3. Forest plot of comparisons of major adverse cardiac events between provisional stenting and double stenting.RCT, randomized controlled trail.

Fig. 5 .
Fig. 5. Galbraith plot for Heterogeneity test of studies concerning major adverse cardiac events.

Fig. 6 .
Fig. 6.Sensitivity analysis of the heterogeneity of studies concerning major adverse cardiac events (A) and target lesion revascularization (B).

Fig. 7 .
Fig. 7. Forest plot of comparisons of target lesion revascularization between provisional stenting and double stenting.RCT, randomized controlled trail.

Fig. 8 .
Fig. 8. Forest plot of comparisons of target vessel revascularization between provisional stenting and double stenting.

Fig. 9 .
Fig. 9. Forest plot of comparisons of all-cause death between provisional stenting and double stenting.RCT, randomized controlled trail.

Fig. 10 .
Fig. 10.Forest plot of comparisons of cardiac death between provisional stenting and double stenting.RCT, randomized controlled trail.

Fig. 11 .
Fig. 11.Forest plot of comparisons of myocardial infarction between provisional stenting and double stenting.RCT, randomized controlled trail.

Fig. 12 .
Fig. 12. Forest plot of comparisons of stent thrombosis between provisional stenting and double stenting.RCT, randomized controlled trail.
---all patients were maintain aspirin lifelong, clopidogrel was prescribed for 6 months in both groups.
SYNTAX score, % Medina classification, % Double stenting type, % Duration of dual antiplatelet therapy -1,1,1 93.6/93.4-After the procedure, aspirin was continued indefinitely and P2Y12 inhibitors were prescribed for at least 12 months.100mg of aspirin was continued indefinitely, and the maintenance duration of clopidogrel (75 mg/day), prasugrel (10 mg/day), or ticagrelor (90 mg twice daily) were also at the operators' discretion.