Objective In patients with multivessel disease and proximal left anterior descending artery (LAD) involvement, the best revascularisation strategy is still unclear. We assess outcomes after coronary artery bypass graft surgery (CABG) and percutaneous coronary intervention (PCI) with drug-eluting stents in a pooled analysis of individual patient-level data of the SYNTAX and BEST randomised trials.
Design Proximal LAD involvement was defined by any lesion ≥ 50% diameter stenosis in the arterial segment starting from the left-main bifurcation up to (and including) the origin of the first major septal branch. The primary endpoint was the composite of all-cause death, myocardial infarction (MI) or stroke at 5 years of follow-up.
Results The present study population comprises 1166 patients of which 577 were randomised to PCI and 589 to CABG. Baseline characteristics were well balanced across study arms. The primary endpoint occurred in 94 (16.3%) patients in the PCI arm and in 68 (11.5%) patients in the CABG arm (HR 1.43; 95%CI 1.05 to 1.95; p=0.026). CABG was also associated with a significantly lower rate of cardiac death (p=0.007), MI (p<0.001), all-cause revascularisation (p<0.001) and major adverse cardiovascular and cerebrovascular events (all-cause death, MI, stroke, revascularisation) (p<0.001). The rates of all-cause mortality (p=0.06) and stroke (p=0.09) were not statistically different between the two groups. The overall study results for the primary outcome were consistent across several subgroups.
Conclusions In patients with multivessel disease with proximal LAD involvement, CABG is associated with lower rates of the safety composite endpoint of death, MI or stroke as compared with PCI with drug-eluting stents at 5 years of follow-up (number needed to treat=21).
Statistics from Altmetric.com
The proximal segment of the left anterior descending artery (LAD) is an important part of coronary tree responsible for a major portion of the blood supply to the left ventricle myocardium. Coronary artery disease involving the proximal part of the LAD is associated with poor outcomes if left untreated.1 ,2 For that reason, such patients receive a class IA recommendation for revascularisation to improve prognosis in latest guidelines.3
Currently available data show equivalence in safety outcomes (death, myocardial infarction (MI) and stroke) for coronary artery bypass graft surgery (CABG) and percutaneous intervention (PCI) in patients with single-vessel disease involving the proximal LAD.4 ,5 In the case of patients with multivessel disease (MVD) with proximal LAD involvement, the literature harbours conflicting data.
In registry-based studies, CABG was associated with better outcomes than PCI.6 ,7 Alternatively, the Arterial Revascularisation Therapies Study Part II (ARTS II) trial ruled out very large differences in safety in terms of death, MI and stroke between the two revascularisation modalities.8 ,9 Nevertheless, the ARTS II was underpowered to detect differences in these outcomes.
The use of bare metal or exclusively first-generation drug-eluting stents (DESs) along with the observational nature of these studies prevents a definite conclusion about the best revascularisation strategy for this subset of patients. Therefore, the present study had the objective of comparing CABG and PCI outcomes in patients with MVD with proximal LAD involvement in a pooled database of two large randomised trials, including the use of new-generation DESs.
The BEST (Randomised Comparison of Coronary Artery Bypass Surgery and Everolimus-Eluting Stent Implantation in the Treatment of Patients with Multivessel Coronary Artery Disease) multicentre, randomised controlled trial compared PCI with the use of everolimus-eluting stents and CABG for treatment of patients with multivessel coronary artery disease (CAD).10
The SYNTAX (Synergy Between PCI With TAXUS and Cardiac Surgery) multicentre, randomised controlled trial compared PCI with the use of paclitaxel-eluting stents and CABG in patients with multivessel CAD.11 ,12 In the present study, we performed a merging of the individual patient-level data of both trials. The institutional review board at each site approved the protocol, and all patients provided written informed consent. The protocol and consent forms were consistent with the Declaration of Helsinki, the International Conference on Harmonisation and all local regulations, as appropriate.
Patients with three-vessel CAD without left-main involvement randomised in the BEST and SYNTAX trial that had at least one lesion (≥50% diameter stenosis) in the proximal LAD segment were included. Proximal LAD involvement was defined as any lesion present on the segment originating just after the left-main stem bifurcation up to the origin of (and including) the first major septal branch.13 ,14
Merging of the databases
A merged database with several important variables including demographics, risk factors, angiographic and echocardiographic findings, revascularisation strategies and clinical outcomes was created. After agreement by both studies’ principle investigators (PIs: PWS and S-JP), the merged database was filled in with all patients' characteristics. All analyses and final results were approved by both PIs.
Study endpoints and definitions
The primary endpoint of the present study was the composite of all-cause death, non-fatal MI or stroke. Secondary endpoints included the individual components of the primary endpoint as well as cardiac death and all-cause revascularisation and major adverse cardiovascular and cerebrovascular events (MACCE) (a composite of all-cause death, stroke, MI and all-cause revascularisation). All endpoints were defined using the original trials' definitions and described in detail elsewhere.10 ,11
In brief, in the SYNTAX trial, MI was defined as development of new pathological Q waves and elevation of creatine kinase myocardial band (CK-MB) ≥5× upper limit of normal (ULN) if occurring within 7 days post intervention or development of new pathological Q waves or elevation of CKMB ≥5× ULN if occurring more than 7 days post intervention. Stroke was defined as a focal central neurological deficit lasting >72 hours resulting in irreversible brain damage or body impairment.
In the BEST trial, MI was defined as new Q waves and increase in the creatine kinase myocardial band (MB) concentration to >5× ULN if occurring within 48 hours after the procedure, or as typical rise and fall of troponin or creatine kinase myocardial band (CK-MB) with symptoms of ischaemia or development of pathological Q waves on the ECG or new left bundle branch block (LBBB), if occurring more than 48 hours after the procedure. Stroke was defined as a focal neurological deficit of central origin lasting more than 72 hours. In both studies, independent clinical events committees blinded to group allocation adjudicated all events.
Baseline characteristics are summarised per study arm as means±SD for continuous variables and percentages (counts) for categorical variables. Continuous data are compared with t tests and binary data with Pearson's χ2 test or Fischer's exact test as appropriate. The present study consists of an independent patient-level pooled analysis using a ‘one-stage’ approach. Outcomes in the overall population and stratified according to SYNTAX score terciles are presented as Kaplan-Meier estimates. HRs and p values derive from Cox proportional hazards models stratified by trial. The proportional hazards assumption was verified for all outcomes with time-dependent Cox models using time as a continuous variable. A Forest plot is used to illustrate the primary endpoint in different subgroups with interaction for treatment effect assessed. All analyses follow the intention-to-treat principle. Two-sided p values <0.05 are considered statistically significant. SPSS software V.21 (IBM) was used for all analyses.
The present study population comprised 1166 patients of which 577 were randomised to PCI and 589 randomised to CABG (figure 1). Overall, mean age was 64.7±9.5 years, 75.9% of patients were male and 37.3% were diabetics. Three-vessel disease was present in 87.5% of cases and two-vessel disease in the remainder 12.5%. Baseline characteristics were well balanced across the two study arms (table 1).
Medical treatment at discharge and at final follow-up is depicted on table 2. At discharge, patients in the PCI group had higher rates of treatment with aspirin, thienopyridines, statins, β-blockers and ACE inhibitors. At the final follow-up, only thienopyridines and ACE inhibitors were more frequently used in the PCI arm, as compared with the CABG arm.
Outcomes at 5 years
The primary endpoint of all-cause death, MI or stroke occurred in 94 (16.3%) patients in the PCI arm and in 68 (11.5%) patients in the CABG arm (HR 1.43; 95% CI 1.05 to 1.95; p=0.026). PCI was also associated with significantly higher rates of cardiac death (6.6% vs 3.1%, HR 2.17, 95% CI 1.24 to 3.81; p=0.007), MI (7.6% vs 2.2%, HR 3.50, 95% CI 1.89 to 6.50; p<0.001), all-cause revascularisation (16.5% vs 7.5%, HR 2.35, 95% CI 1.64 to 3.36; p<0.001) and MACCE (27.4% vs 17.3%, HR 1.68, 95% CI 1.31 to 2.15; p<0.001) (PCI vs CABG, respectively, for all). The rates of all-cause mortality were 10.1% (n=58) in the PCI arm and 7.0% (n=41) in CABG arm (HR 1.46, 95% CI 0.98 to 2.18; p=0.06). There were 11 strokes (1.9%) in the PCI arm and 21 (3.6%) in the CABG arm (HR 0.54, 95% CI 0.26 to 1.11; p=0.09) (figure 2; table 3).
A total of 349 (29.9%) patients had low SYNTAX scores (≤22) of which 173 were randomised to PCI and 176 to CABG. In this subgroup, the rate of MI was significantly lower in the CABG arm (1.7% vs 6.9%, p=0.028). For the other outcomes, no significant difference was observed between the two treatment strategies (table 3), For the 817 (70.1%) patients with intermediate and high SYNTAX scores (≥23), with the exception of stroke, which was similar between the two study arms, CABG was associated with significantly lower rates of all outcomes, including all-cause death (7.3% vs 11.9% for PCI, p=0.027) (table 3).
The overall study results for the primary outcome of all-cause death, MI or stroke were consistent across several subgroups (figure 3). No interaction with treatment effect was present in any of the analysed subgroups neither when comparing patients with and without proximal LAD involvement (pINT=0.41). The benefit of CABG was statistically significant in male patients, patients with creatinine clearance ≥60 mL/min, patients without peripheral vascular disease, SYNTAX scores ≥23, presence of three-vessel disease and patients who underwent complete revascularisation (figure 3). In the BEST trial (which used new generation everolimus-eluting stents), no significant difference was observed between CABG and PCI for the occurrence of the primary endpoint. Nevertheless, the p value for interaction for study type (SYNTAX vs BEST) was not significant (PINT=0.28) suggesting that the benefit of CABG was consistent with both early and new generation DESs.
The results of the present study can be summarised as follows: (1) in patients with MVD with proximal LAD involvement, CABG was associated with a significantly lower rate of the composite endpoint of all-cause death, MI or stroke; (2) the rates of cardiac death, MI, all-cause revascularisation and MACCE were also lower after CABG compared with PCI; (3) in the subset of patients studied, low SYNTAX scores (≤22) were present in 29.9% of cases; (4) in this subgroup with low SYNTAX scores, CABG was associated with a significantly lower rate of MI; (5) in the subgroup with intermediate or high SYNTAX scores (≥23), CABG was associated with lower rates of all-cause death, cardiac death, MI, all-cause revascularisation and MACCE.
The proximal part of the left anterior descending artery is a noble segment of the coronary tree. Its disease may put in risk up to 50% of the left ventricle myocardium mass and is definitely associated with poor outcomes.2 While CABG or PCI are currently recommended as revascularisation strategies to improve prognosis in patients with proximal LAD disease,3 ,15 the best revascularisation modality for this specific subset of patients has long been subject of debate amidst the world cardiology community.
For patients with single-vessel disease and proximal LAD involvement, equivalent safety between CABG and PCI has been observed in multiple studies.4 ,5 In patients with MVD and proximal LAD involvement, a potential benefit of CABG over PCI is still unclear. In the large New York State registry that assessed close to 5000 matched pairs with two-vessel or three-vessel disease with proximal LAD involvement, a survival benefit for CABG was observed at 5 years of follow-up.6 However, even though adjusting for confounders was undertaken with propensity matching, this is still observational data subject to bias by several potential unmeasured confounders.
The ARTS II trial enrolled patients with multivessel CAD to treatment with PCI using first-generation DES and compared the historical control of CABG and PCI with bare metal stents (BMS) from the ARTS I trial.16 In the subgroup of patients with proximal LAD involvement, the rates of the safety endpoint of death, MI or stroke were similar among the three groups of BMS, DES and CABG.8 ,9 Nevertheless, the lack of difference could partially be explained to the relatively small number of patients included (n=682).
In the present study, we were able to overcome the major limitations of previous studies. This is the first report on randomised data of patients with MVD with LAD involvement that used exclusively DES in the PCI arm, including newer generation devices in the BEST trial. By merging individual patient-level data of two large randomised trials, we were able to detect a significant difference in safety outcomes in favour of CABG. Furthermore, we could also assess important subgroups, including the low SYNTAX score subset. Interestingly, even in patients with low SYNTAX scores, a significant difference in MI rates was observed in favour of CABG. Of note, in the BEST trial, where new generation everolimus-eluting stents were used, no differences in the safety primary endpoint were observed between CABG and PCI. Again, the number of patients might have played a role. Nevertheless, no interaction was observed for study type (SYNTAX vs BEST) suggesting that the benefit of CABG was consistent with both early and new-generation DESs.
Our findings support the recommendation for surgery in patients with MVD with proximal LAD involvement with an acceptable surgical risk. Nevertheless, we acknowledge that the choice of the best revascularisation strategy should still be made on the basis of an individualised and thorough analysis of aspects including the coronary anatomy and several clinical factors known to impact prognosis and interact with either or both revascularisation options.
The present study has some limitations. First, due to the intrinsic differences in patient populations and procedural techniques between the two trials included, bias could have been introduced. Moreover, the stent used in the SYNTAX trial is associated with poorer outcomes than new generation devices and is no longer used in clinical practice in the vast majority of centres. The BEST trial was prematurely discontinued due to slow enrolment and this could have impacted our results. Lastly, due to the number of patients included, subgroup analyses are underpowered.
In patients with MVD with proximal LAD involvement, CABG is associated with lower rates of the safety composite endpoint of death, MI or stroke as compared with PCI with DESs at 5 years of follow-up. Twenty-one patients would need to be treated by CABG to prevent one event (number needed to treat (NNT)=21).
What is already known on this subject?
While registry studies show a safety benefit of coronary artery bypass graft surgery (CABG) over percutaneous coronary intervention (PCI) in patients with multivessel coronary disease and proximal LAD involvement, randomised data show equivalent safety between the two strategies.
What might this study add?
The present study is the first report comparing only drug-eluting stent PCI (including second-generation devices) with CABG in patients with multivessel disease and proximal LAD involvement in a randomised fashion with a large number of subjects.
How might this impact on clinical practice?
Due to the sample size and randomised nature of the data, this study represents the most robust available data on this specific patient subset. Its results might impact practice guidelines to recommend CABG over PCI in patients with multivessel coronary disease and proximal LAD involvement.
RC and YS contributed equally to this study. An abstract of this study has been presented at the American College of Cardiology Meeting in Chicago on 2 April 2016.
Contributors PWS, S-JP, YO, CWL, J-MA, RC and YS: acquired the pooled data. RC: analysed the data. RC and YS: wrote the manuscript. All authors critically reviewed the manuscript and made substantial contribution to its improved final version.
Competing interests None declared.
Ethics approval Ethics committee/institutional review board at each participating centre.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.