CABG versus PCI
Angioplasty of the LM was first attempted over 30 years ago. Initial outcomes of PCI for LMD were associated with unacceptable rates of restenosis, acute thrombosis, and repeat revascularization. With the advent of advanced stent design, improved technique, and aggressive adjunctive medical therapy, PCI has evolved over the last 2 decades.
Observational Registries
Observational studies ( Table 1 ) suggest that PCI for LMD may be a reasonable alternative to CABG in terms of myocardial infarction (MI) and mortality. These registries are extremely heterogeneous in terms of clinical characteristics, spectrum of risk of patients, stents used, and PCI technique. CABG has been associated with a higher incidence of adverse in-hospital/early outcomes because of significant peri-procedural MI, stroke, and death. Long term, the endpoints of MI and mortality were similar, but repeat revascularization rates were significantly higher for PCI. Clinically asymptomatic occlusions were found in the PCI arms due to routine angiographic follow-up. Graft occlusions probably were under-recognized in the CABG arms as follow-up angiography was only clinically driven. Most of these registries noted a higher incidence of repeat revascularization in the distal LM/bifurcation location. The CABG arms had longer hospital stays and were accompanied with significant postoperative complications. Although a 25% relative risk reduction in major adverse cardiovascular and cerebrovascular events (MACCE) and a lower hazard ratio was noted in the drug-eluting stent (DES) group compared to the bare-metal stent (BMS) group, the event-free survival was superior in the CABG-treated arms.
Meta-analyses
Three meta-analyses comparing CABG and PCI (2905, 3773, and 5479 patients) of ULMD showed that MI, stroke, and death were similar at 1, 3, and 5 years while target vessel revascularization (TVR) was higher for PCI. There was no difference in mortality between the 2 groups. To address the issue of limited duration of follow-up, Park et al reported long-term results for BMS and DES use. At 10 years, the adjusted risks of MI, stroke, and death were similar in the BMS and concurrent CABG groups. At 5 years, there was no difference in the adjusted risk of death in the DES and CABG groups. However, the TVR rates were higher in the PCI group.
Randomized Controlled Trials
Currently, four RCTs have compared PCI with CABG for the treatment of LMD ( Table 2 ). The main limitation of the LE MANS study was the non-specific primary endpoint (change in LV function at 12 months), with both groups demonstrating similar improvement in angina and functional capacity at 1 year. The absolute change in LV function was greater after PCI (P=.04) than after CABG (P=.85). There was a significant difference in LV function between the 2 groups after 12 months (P=.01). The MACCE-free 1-year survival rate was not significantly lower in the PCI group compared with the CABG group (71.2% vs 75.5%; P=.29), with the difference mainly related to repeat revascularization (P=.01). Similarly, MACCE-free survival did not differ significantly between the 2 groups (53.9% vs 56.6%; P=.47). No in-stent thrombosis was noted, which was probably due to selective use of DES/BMS and the technique of LM stenting (provisional stenting of the side branch).
Boudriot et al randomized 201 patients with ULMD to CABG and PCI. The primary endpoint of MACCE was 13.9% and 19%, respectively (P=.19 for non-inferiority). Stenting was inferior to surgery for repeat revascularization (5.9% vs 14%; P=.35). There was a higher incidence of MACCE in distal LM lesions compared to ostial/shaft lesions in both the treatment arms (PCI, 18% vs 1%; CABG, 8.9% vs 5%). The PRECOMBAT trial is the largest RCT comparing CABG and PCI to date. A combined primary endpoint of MACCE at 1 year was 8.7% and 6.7% (P=.01 for non-inferiority) and at 2 years was 12.2% and 8.1% (P=.12) in the PCI and CABG groups, respectively. Although PCI was shown to be non-inferior to CABG, the study was underpowered due to low event rates. The apparent safety of PCI compared to CABG at 2 years is consistent with that observed in SYNTAX.
In the SYNTAX trial, patients were categorized according to angiographic risk, which took into account lesion characteristics and the amount of myocardium supplied by the affected vessels. Patients were divided into 3 terciles according to their angiographic risk score and clinical outcomes were reported for each tercile. In patients with low and intermediate SYNTAX score (0–32), the combined endpoint of MI, stroke, and death was better in the PCI arm and the rates of repeat revascularization were similar in the 2 groups. Finally, in the patients with the most complex anatomical disease (≥33), MACE rates were significantly higher in PCI-treated patients ( Table 3 ).
Unfortunately, the issue of restenosis remained the same in the RCTs as in the observational registries. Stent thrombosis was not noted in two RCTs. A substantial portion of restenosis rates occurred in patients with distal LMD. Based on these findings, PCI of LMD now has a class IIB indication.
DES versus BMS
Studies comparing DES and BMS for LMD showed lower rates of restenosis and TVR with DES ( Table 4 ). One RCT showed a significant reduction in restenosis rate with paclitaxel-eluting stents (PES, 6% vs BMS, 22%) and significant improvement in major adverse cardiovascular event (MACE)-free survival due to reduction in TVR. The DELFT registry showed a high clinical and procedural success rate with DES use at 3 years. Recent meta-analyses of >10,000 patients showed that PCI with DES was superior to BMS with respect to MACE and mortality. Recent studies have shown comparable clinical and angiographic outcomes for SES and PES. Furthermore, in an observational study, Valenti et al showed PCI with everolimus-eluting stent (EES) to have a reduced incidence of MACE (10.2% vs 21.9%), TVR (7.8% vs 20.5%), and restenosis (5.2% vs 15.6%) at 1 year compared to PES.
Ostial versus Bifurcation Lesions
Stenting of ostial and mid-shaft LM lesions appears to be safe and effective with low rates of MACE and restenosis. The outcome is significantly worse for patients undergoing PCI of distal LMD with higher rates of MI, death and revascularization.