Comparison of Percutaneous Coronary Intervention with Coronary Artery Bypass Graft
Registries
Evidence from the early registries comparing PCI versus CABG suggested that PCI and CABG had a similar MACE rate.9,10 The two observations that emerged were a higher risk of a peri-procedural cerebrovascular accident (CVA) in the CABG group and a higher incidence of target lesion revascularisation (TLR) in the PCI group. The Revascularization for Unprotected Left Main Coronary Artery Stenosis: Comparison of Percutaneous Coronary Angioplasty Versus Surgical Revascularization (MAIN-COMPARE) registry, with data on a total of 2,240 patients, showed comparable outcomes at 5 years though the need for repeat revascularisation with PCI was again highlighted.11 Given the limitations of registry data, more robust evidence was needed to confirm or refute this observation. This spurred the development of randomised controlled trials (RCTs) to gain further insight into the validity of PCI as a viable therapeutic option.
Data from Randomised Controlled Trials
Four RCTs have specifically looked at the cohort of patients with LMS disease to this day.
The Study of Unprotected Left Main Stenting versus Bypass Surgery (LE MANS) was the first RCT and enrolled 105 patients with significant LMS disease (defined as >50 % stenosis angiographically).12 The primary endpoint was the change in left ventricular ejection fraction (LVEF) at 12 months, while the secondary endpoint was a major adverse cardiac and cerebrovascular event (MACCE) at 30 days and 1 year. Surprisingly, there was a statistically significant improvment in LVEF with patients treated with PCI versus CABG (58 % versus 54 %). PCI was also associated with a lower MACE rate at 30 days (2 % versus 13 %) with a MACE being equivalent at 1 year in the two groups. The study did have a number of limitations, including a small sample size, high use of BMS and a lower than contemporary use of left internal mammary artery (LIMA) grafts.
The Synergy Between PCI With Taxus and Cardiac Surgery (SYNTAX) trial remains to this day the largest RCT to date to compare PCI to CABG in LMS disease.13 The LMS subset consisted of 705 patients randomised to receiving either the first-generation TAXUS (Boston Scientific Corporation, US) DES or CABG. The primary endpoint of MACE at 1 year was comparable with 15.8 % for PCI versus 13.7 % for CABG. However, when the cohorts were further subdivided into categories based on lesion complexity, it became apparent that CABG was more favourable for the more complex lesions. The SYNTAX score was developed to objectively quantify lesion complexity and has been fully described elsewhere.14 Stratification into tertiles of syntax score (0–22, 23–32 and >32) revealed equivalent MACCE in the lower two tertiles, but a clear superiority of CABG in the highest tertile. The 5-year follow-up results were published in 2014.15 The results bear a similar trend to the 1-year follow-up with similar outcomes across the whole cohort (36.9 % PCI versus 31 % CABG; P=0.12) but a statistically greater benefit of CABG in the highest risk group based on lesion complexity.
While the SYNTAX trial remains the best evidence-based RCT on revascularisation strategy, it is important to bear in mind that 45 % of patients were excluded from randomisation at the outset due to the complexity of coronary disease. Out of that group, 85 % went on to have CABG. So the evidence we have is from a pre-selected population and this has to be taken into consideration when evaluating whether PCI is an appropriate therapeutic option.
The Premier of Randomized Comparison of Bypass Surgery versus Angioplasty Using Sirolimus-Eluting Stent in Patients with Left Main Coronary Artery Disease (PreCOMBAT) trial was a non-inferiority trial randomising 600 patients with LMS disease to either a first-generation Cypher (Cordis Coporation, US) DES or CABG.16 Once again the primary endpoint was MACCE at 1 year. The trial was designed to test whether PCI was non-inferior and indeed proved the non-inferiority of the percutaneous option at both 1-year and 2-years follow-up.
Boudriot et al. studied a smaller sample group of 201 patients and unlike the previous studies, PCI failed to achieve non-inferiority of MACE at 1 year.17 Excess MACE was driven by target vessel revascularisation. Unusually for this type of comparison, stroke was not included as a clinical endpoint, which may well have contributed to the end result.
The results from the RCTs would support PCI as a reasonable alternative in the treatment of LMS disease but the studies have been hampered by limitations. These include relatively small sample sizes, different clinical endpoints among studies and the use of earlier generation DES. This has made it difficult to precisely define where PCI stands.
Meta-analysis
Athappan et al. published a meta-analysis of first-generation DES versus CABG in 2013.18 The authors included 21 observational studies and three RCTs involving a total of 14,203 patients. The MACE rates did not differ between the PCI and CABG group up to 5 years though as has been previously noted, PCI was associated with a lower rate of stroke and higher rate of target vessel revascularisation (TVR). When the cases were stratified according to the SYNTAX score, the results were consistent with those of the main SYNTAX trial with comparable results for the lower two tertiles but a clear superiority of CABG for the highest group.