Coronary artery bypass grafting (CABG) is still the standard recommended treatment for unprotected left main coronary artery (ULMCA) lesions.1 Technical improvements in percutaneous intervention (PCI) and stent technology mean that it is increasingly considered as a treatment option for this high-risk patient subset.2 Several registries and randomized clinical trials have shown that percutaneous revascularization of ULMCA lesions is safe and effective, yielding results that are comparable to CABG, particularly when hard clinical endpoints are considered. In the PRECOMBAT study, PCI with sirolimus-eluting stents was non-inferior to CABG with respect to the primary composite endpoint of major adverse cardiac or cerebrovascular events at one year.3 In the subset of 705 patients with left main (LM) lesions (13% isolated LM) included in the randomized SYNTAX trial, all-cause death, cardiac death, myocardial infarction (MI) and the composite endpoint of all-cause death, stroke or MI were similar with both treatment strategies (PCI with first-generation paclitaxel-eluting stents vs. CABG) at four years, despite a significantly higher rate of repeat revascularization in the PCI group (23.5% vs. 14.6%; p=0.003), and a higher rate of stroke in patients randomized to CABG (14.6% vs. 4.3%; p=0.03). The overall results of the trial show that the majority of events occurred in patients with higher SYNTAX scores.4 Likewise, in the five-year outcome analysis of the MAIN-COMPARE registry, stenting resulted in similar rates of death and of the composite endpoint of death, Q-wave MI or stroke, but higher rates of target vessel revascularization were observed compared to CABG.5 The evidence thus suggests that PCI is a reasonable treatment alternative for this subset of coronary artery disease patients, although at the cost of a higher rate of target lesion revascularization.

Our aim was to evaluate and report the long-term (five-year) outcome of patients undergoing PCI of ULMCA disease in a high-volume PCI center.

In this observational study, the long-term prognosis (five years) of patients undergoing PCI of a ULMCA was reported. Both acute procedural and clinical success rates were high, indicating that the strategy is safe and effective when performed by experienced interventional cardiologists, in this case mostly using drug-eluting stents and a single-stent technique.

Although the choice of PCI strategy was at the operator's discretion, a single stent was used in nearly 70% of cases. In our study, more than half of the patients had distal LM involvement, with Medina classification 1.1.0 and 1.1.1 being the most prevalent disease patterns (Figure 1). This finding is consistent with data from previous trials and registries, including the SYNTAX trial.4,7 Evidence from early trials suggested that the outcome of patients treated for distal LM disease was worse than for those with aorto-ostial or mid-shaft lesions.8 However, subsequent studies have failed to demonstrate the superiority of a systematic double-stent approach to distal LM PCI.9,10

Total MACE rates in our cohort were within the range of those previously reported in randomized trials and registries.4,5 It should be noted that total mortality was higher and TLR was relatively low, especially considering the long follow-up period. The high mortality rate was most likely related to the overall worse clinical profile of the patients included in our analysis. Compared to the MAIN-COMPARE registry,5 our population was older and had higher rates of previous MI and chronic renal disease and more extensive coronary artery disease and, importantly, patients with ST-elevation MI and cardiogenic shock were not excluded. In our cohort, the proportion of troponin-positive acute coronary syndromes was ∼71%, compared to ∼8–9% of non-ST-elevation MI patients in the MAIN-COMPARE registry. The finding that early mortality was comparable to the EuroSCORE-predicted fatality rate at 30 days reinforces the observation that long-term mortality was most likely related to patient characteristics, rather than to procedure-related issues. When patients with cardiogenic shock were excluded from the analysis, mortality was 19.8% and when only troponin-negative (i.e., stable coronary artery disease and unstable angina) patients were included, mortality was 13.6%, which is comparable to similar published trials and registries.

On the other hand, in comparison to the PCI arm of the SYNTAX trial (23.5% at four years)4 and to the MAIN-COMPARE registry (16% at five years),5 TLR in our study was less frequent and occurred mostly during the first year after the index procedure, despite the low use of intravascular ultrasound-guided PCI. This is consistent with the fact that the mean SYNTAX score observed in our cohort was lower than that reported in the left main subset of PCI-treated patients of the SYNTAX trial (24.2±11.8 vs. 28.1±12.4),4 as well as with the high proportion of DES used. In addition, the low rate of systematic elective angiography during follow-up could provide further explanation for these observations: not only could cases of angiographic restenosis have been missed, but undiagnosed restenosis could actually have been responsible for some cases of cardiac death. Previous studies have suggested that cardiac death can be a manifestation of restenosis after LM PCI11,12 and indeed, in our study, total MACE and mortality rates were higher in BMS- than in DES-treated patients. Of the 11 TLR events, at least five were clinically driven (two of them non-ST-elevation MI), while three followed elective control angiography in asymptomatic patients. For the remaining three, revascularization details were not available, but vital status was known. In our view, these findings may argue in favor of routine follow-up angiography, as significant restenosis presented as an acute coronary syndrome in at least 18% of cases, at best.

The independent predictors of MACE at five years in our cohort – hypertension, previous MI and EuroSCORE I – have all been shown to be associated with worse prognosis in the setting of coronary disease. Elderly patients receiving ULMCA PCI are at high risk for major adverse events following the procedure.13 However, this was not evident in our study, possibly because important unmeasured comorbidity and the EuroSCORE offset the effect of age in the prediction model.

The implantation of DES was statistically associated with a highly significant lower MACE rate. Several studies have reported the survival benefit of ULMCA revascularization with DES, but the mechanism by which stent type affects the risk of death remains unclear.13,14 Palmerini et al. showed that DES use was associated with a significant decrease in adjusted two-year cardiac mortality rates, compared to patients treated with BMS.9 In our population the association between DES and lower mortality alone was consistently stronger than with reduced TLR. Notwithstanding the possibility, discussed above, that some cases of restenosis (more frequent in BMS-treated patients) may have presented clinically as fatal events, it cannot be firmly excluded that this observation is a spurious statistical association driven by the fact that fatal events (n=20) occurred more often than TLR events (n=11) and that a treatment bias is still present, despite adequate corrections and the good fitness of the regression model for the propensity variable (c-statistic=0.83).

LM disease is present in 3–10% of patients undergoing coronary angiography and is associated with high mortality. Despite the upgrade in the recommendation for PCI of ULMCA lesions, this strategy is usually reserved for patients with high burden of comorbidities and frequently for severely ill patients (under resuscitation and/or in cardiogenic shock), in whom surgical revascularization cannot be carried out with an acceptable risk.

In patients presenting with acute coronary syndromes, ULMCA lesions are associated with high mortality, especially in the setting of ST-elevation MI and/or hemodynamic or arrhythmic instability. For these, PCI may be the preferred revascularization strategy, being feasible and yielding reasonable outcomes.15,16

Technological improvements in interventional techniques, particularly the widespread use and availability of DES (which decrease stent restenosis and need for repeat revascularization), the development of both percutaneous and pharmacological adjunctive therapies such as thrombus aspiration and antiplatelet or anticoagulant therapy is contributing to increasingly consistent and satisfactory results.15,16 Nevertheless, and despite significant geographical variability, fewer than 5% of patients with significant ULMCA lesions are percutaneously treated, as shown by the CathPCI Registry.3 In the multinational all-comers DELTA registry,17 no difference was observed in the occurrence of death, cerebrovascular accidents, and MI between PCI with first-generation drug-eluting stents and CABG at a median follow-up of 3.5 years.

Another feasible alternative for the treatment of ULMCA lesions is hybrid coronary revascularization. Combining the advantages of the long-term durability of a left internal mammary artery-left anterior descending bypass with the less invasive option of PCI for lesions in other coronary arteries may reduce the possible complications of the two interventions in isolation.9

In view of the complexity of LM lesions, certain strategies can help improve both acute angiographic and long-term clinical success: performance of PCI in high-volume centers by operators with expertise in the management of complex anatomic lesions; routine use of intravascular ultrasound for procedural guidance, as it provides accurate assessment of lesion severity and can confirm optimal stent expansion and apposition; hemodynamic support, when appropriate; implantation of newer DES, which reduce stent thrombosis and restenosis; implantation of a single stent whenever possible; emphasis on compliance with dual antiplatelet therapy, especially after DES implantation; application of clinical and angiographic risk models such as the SYNTAX score to facilitate an individualized approach for each patient; and encouragement of collaborative and multidisciplinary discussion and patient selection within the “heart team”, which is crucial for clinical decision-making.18 Ongoing randomized trials such as NOBLE (Nordic-Baltic-British Left Main Revascularization) and EXCEL (Evaluation of Xience Prime Versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization) will bring further insight into treatment options for ULMCA stenosis using contemporary devices and surgical techniques.

The present study has some limitations, among which we highlight those inherent to a retrospective non-randomized design and to the small number of patients included. Unknown confounders might have affected the results (despite the appropriate statistical adjustments) and the small population size renders the study underpowered to detect significant differences in outcome according to important baseline characteristics. Finally, some of the procedural characteristics, mainly those related to the PCI technique itself, were at the operator's discretion, making standardization difficult and assessment of the potential impact of the PCI strategy on outcome virtually impossible.

In conclusion, in this high-risk population, ULMCA PCI was safe and effective. Although the five-year event rate (especially mortality) was high (reflecting poor patient clinical status), early mortality compared favorably to the EuroSCORE I-predicted 30-day mortality. Overall, long-term clinical results are acceptable, and thus this strategy may be considered a valid alternative to CABG. The high rate of MACE in the first year, especially in BMS-treated patients, suggests the need for routine control angiography in all patients in whom a BMS is implanted, and whenever possible, that a DES should be used as first choice.

The authors have no conflicts of interest to declare.