Percutaneous treatment of de novo unprotected left main stenosis in unselected consecutive patients : Experience of a high volume center

Background: Percutaneous coronary intervention (PCI) has been increasingly employed to treat unprotected left main (ULM) stenosis, with hard endpoints similar to by-pass surgery, in patients selected by a Heart Team. Methods: From January 2008 to December 2011, 317 unselected and consecutive patients with de novo ULM stenosis underwent PCI with both bare metal (BMS) and drug-eluting (DES) stents. Major adverse cardiovascular events, target lesion (TLR) and vessel (TVR) revascularization were evaluated over a mean period of 590 ± 371 days. Results: Our population was characterized by a mean age 72 ± 10 years, high rate of acute coronary syndrome (ACS) (either with ST or non-ST elevation myocardial infarction, 15.5% and 35% respectively), severe comorbidity 16%, mean Euroscore 7 ± 3, mean Syntax Score 25 ± 9. In-hospital mortality was 6%. During the follow-up period, all-cause mortality was 16.7%, falling to 7% at the end of the follow-up, excluding patients presenting with ACS. TLR was observed and treated in 15% of patients. BMS utilization, age >75 years, ACS indication, Syntax Score >32 and associated peripheral artery disease were independent predictors of mortality at multivariate analysis. Conclusions: Stenting of ULM stenosis appears to be associated with a favorable mid-term outcome, even in an unselected population.


INTRODUCTION
Unprotected left main (ULM) percutaneous coronary intervention (PCI) has recently become a valid alternative to coronary artery bypass graft (CABG) showing similar mid-term results for hard endpoints (death and myocardial infarction) even if still penalized by a higher rate of repeated revascularization in the drug-eluting stent (DES) era.
A potentially higher procedural risk due to the frequent involvement of the bifurcation with the majority of myocardium being jeopardized has been balanced by the relative easiness of ULM PCI due to the vessel shortness and big dimensions.DES development and diffusion, the experience gained by intravascular ultrasound (IVUS) use, the benefits of post-dilation and kissing techniques and the protection given by a prolonged dual antiplatelet therapy have all led to the growth and diffusion of this procedure.
The purpose of the present paper is to report the ULM PCI results of a single centre without on-site cardiac surgery, in consecutive "all comers" patients from 2008 until 2011.In all patients with de novo ULM lesions, PCI option was evaluated by clinical and interventional car-diologists, and preferred over CABG, whenever a complete or at least a functionally complete revascularization was feasible or in presence of a high surgical risk determined by comorbidities.

Clinical Procedure
All consecutive unselected patients treated with PCI for de novo lesions of ULM from 1.1.2008to 31.12.2011 in our catheterization laboratory (CL) were collected in this registry.
Our laboratory is a province based CL in a region with almost 400.000inhabitants and 3 hospitals (1 hub and 2 spoke centres), with a yearly volume of more than 2000 total procedures and 1300 -1400 PCI, resulting the highest volume CL in the Emilia-Romagna Region and one of the most productive in Italy.
All clinical, angiographic and procedural data, hospital discharge and clinical follow-up data were recorded in a retrospective/prospective database.Data relative to the clinical conditions of the last available follow-up were collected by means of clinical examinations, telephone interviews or from civil registries.
All patients gave written informed consent to the procedure after full discussion of the risk/benefit profile of PCI and of the surgical alternative.
Indication and timing of the procedure (one stage or subsequent steps in case of multivessel coronary disease) were evaluated by an expert team (at least 2 expert interventional and one clinical cardiologists), while all procedural techniques (arterial access, provisional/intended two stents, type of stent, intra-aortic balloon pump (IABP), IVUS, etc.) were left to the discretion of the first operator with a strong recommendation for second generation DES utilization.
All patients were pre-treated with aspirin and clopidogrel (at least 300 mg in the previous 24 hours and 75 mg the morning of the procedure).They also received aspirin 100 mg daily indefinitely and clopidogrel 75 mg daily for at least 12 months post-procedure.
Angiographic follow-up 7 ± 1 months after PCI was strongly recommended and it was performed whenever possible.

Definitions
Cardiac death: all deaths due to cardiac causes (ischemic and non ischemic), or all deaths without a known cause.
Vascular death: all deaths due to cerebral or peripheral artery disease.
Spontaneous MI: every post-discharge elevation of CKMB or troponin above UNL associated with clinical or ECG signs of ischemia.
Target Lesion Revascularization (TLR): every new target lesion (left main) revascularization comprehensive of the 5 mm pre and post target (also left anterior descendent artery and circumflex artery ostia).
Target Vessel Revascularization (TVR): every new index target vessel/s revascularization also outside the target lesion.
Major vascular complications: every ischemic or hemorrhagic in-hospital complications requiring surgery or transfusions.
Euroscore: 30 days death risk was stratified according to the additive Euroscore, evaluating clinical, cardiac and surgical parameters [17].
Syntax Score: anatomic complexity of the patients was evaluated with the Syntax Score algorithm, which is available on the SYNTAX Score Website (www.syntaxscore.com)[18].
Complete revascularization: left main (LM) PCI and of every >70% stenosis of major epicardial coronary arteries.
Functionally complete revascularization: LM PCI and of every >70% stenosis of major epicardial vessels with a vital/ischemic bed.
Stent thrombosis (ST): defined on the basis of Academic Research Consortium criteria [19], as definite in case of angiographic or pathologic confirmation of ST, probable in case of any unexplained death within 30 days or target vessel myocardial infarction without angiographic confirmation, and possible ST defined as unexplained death after 30 days.

Statistical Analysis
Statistical analysis was performed with SAS statistical package.All continuous variables were expressed as mean ± SD and categorical variables as percentages.
Comparison between groups was performed using chi-square test for categorical variables and the unpaired Student's T test for continuous data.A two-side p value < 0.05 was considered statistically significant.
Logistic regression analysis was performed to identify independent predictors of mortality, adjusted for baseline confounding characteristics that were identified on the basis of univariate analysis.
The odds ratios (OR) and 95% confidence intervals (CI) were reported with two tailed probability values.

Clinical Characteristics
Clinical characteristics of treated patients are described in Table 1.
Mean age was 72.6 years, with 49.5% of patients aged ≥75 years and 27% aged ≥80 years.

Angiographic Characteristics
Angiographic characteristics of the population are sum-marized in Table 2.
LM disease was associated with multivessel coronary artery disease in more than 87% of patients, and in 28% of cases there was a chronic total occlusion (CTO), in 54 patients (17%) of the right coronary artery.
Syntax Score was calculated in 311 patients; it was ≤22 in 125, between 23 and 32 in 120, and >32 in 66 patients.LM stenosis was distal in 75% of cases, with bifurcation involvement in 227 patients (72%).
Procedural characteristics are summarized in Table 3. LM PCI was "ad hoc" in 54% of cases, with femoral access in 81% and with stent implantation in 99%, bare metal stents (BMS) 16%, DES 84%, first generation in 15 patients and second generation (mainly Xience and Biomatrix) in 249 patients.Isolated LM PCI was performed in 80 patients, LM and other vessel/s PCI in 237 in a unique or staged procedure; in 21 patients also an associated CTO was tempted.Mean number of implanted stents was 2.6 ± 1.5 per patient (range 0 -10).Revascularization was complete in 177 patients (56%).
Bifurcation LM lesions were treated as follows: provisional stenting strategy was adopted in 183 patients, a two stent strategy in 41 (18%), with a minicrush technique in 20, T-stenting in 15, and V-stenting in 6 patients.Final kissing-balloon was performed in 92% of cases.
One hundred and ninety patients (64%) had an angiographic follow-up, routinely performed in 138 and clinically driven in 52 patients.
Table 4 depicts clinical events at follow-up.Total mortality was 16.7%, cardiac mortality was 10% (7.9% at 1 year); excluding patients with ACS (either with ST or non ST elevation myocardial infarction) as indication for the index procedure, total mortality falls to 3.8% at 1 year and 7% at the end of follow-up.Among cardiac deaths, 4 were undetermined possibly due to ST: 1 at 185 days in a patient with moderate to severe mitral regurgitation, 1 at 417 days, 1 at 487 days after a red blood cells transfusion for acute anemia and 1 at 231 days for a documented ventricular fibrillation, while no cases of definite ST were documented angiographically or anatomically.
A repeated revascularization procedure for a LM lesion was necessary in 15% of cases (47/48 cases were treated with a repeated PCI and only one case with CABG), in 13 patients this was due to a LM edge restenosis, in 16 to a distal LM or LAD and circumflex artery ostial lesion, in 17 to an isolated circumflex ostial lesion and in 2 to an isolated LAD ostial lesion.TLR was clinically driven only in 25 patients, with an angiographic follow-up due to symptoms or inducible ischemia.Clinically driven TLR was performed at a mean time of 208 days from the index procedure, a mean of 30 days before TLR performed during a routine angiographic follow-up (233 days).
Forty-two patients (13.2%) had a repeated procedure on an already treated vessel and 41 (12.9%) had a PCI of another vessel/de novo lesion.
Comparisons between patients treated with DES or BMS and between the three Syntax Score groups are shown in Table 6 and 7.

DISCUSSION
The population studied represents a real world scenario, being composed by unselected, consecutive patients ad-mitted to a high volume hub centre without on-site cardiac surgery.Most of the patients were admitted with an ACS, who had frequent comorbidities and a complex anatomic situation frequently associated with multivessel disease (Tables 1 and 2).
The peculiarity of our LM PCI experience lies in the absence of a proper "heart team" and in the presence of experienced operators leading to the preferred choice of the percutaneous revascularization option if possible (at least a functionally complete revascularization) without employing more than 4 -5 stents, and in the absence of absolute contraindications to DES.The main intent was to avoid adverse events related to CABG, even those less known or not routinely reported (i.e., sequelae of surgical wounds, psycho-intellective deterioration) that significantly impact on morbidity and on the quality of life and costs [20].
Our results show a satisfactory short-term outcome, with an acceptable in-hospital mortality considering the type of patients (15% STEMI, comorbidity, age, mean Euroscore 7 ± 3); moreover, also at 12 months, excluding patients with acute MI (STEMI and NSTEMI) total mortality results 3.8% (Table 4), similar to the 12-month mortality of the LM subgroup of the Syntax trial that did not enroll patients with acute MI (2).Our results also do well in comparison with the experience of the large DELTA Registry (6), with cardiac mortality of 10% versus 6.8%, if we considered that our patients were older (mean age 72 ± 10 vs 65.8 ± 1.1 years), and presented more frequently an ACS (STEMI 15% vs 2.9%, NSTEMI  35% vs 11.6%) and a higher Euroscore (7 ± 3 vs 4.9 ± 3.6).Medium term TLR rate is comparable to that reported by another study with clinical and/or angiographic follow-up [21] and seems to be linked to the routine angiographic follow-up, more often with a significant involvement of circumflex artery ostium.Almost half of (23/48) TLR were not clinically driven.Although angiographic follow-up is largely used to monitor the results of LM PCI, there are many doubts about its appropriateness [22].All but one case of LM in-stent resteno-sis were treated percutaneously in our experience.
In this population of anatomically complex patients, mostly with multivessel disease, the need for a repeated revascularization of already treated segments of other vessels, or of de novo lesions or progression of disease has all the same importances of total TLR (13.2% of TVR and 12.9% of PCI of de novo lesions, Table 4); progression of atherosclerosis is the main cause of the PCI inferiority versus CABG as long as the long term need for new revascularizations is concerned.
Although IVUS was rarely used, a post-dilatation with a non-compliant balloon (254/317, 80%) and final kissing-balloon in almost all bifurcation lesions were frequently performed (Table 3).This technical feature, derived from the results of IVUS-guided stenting studies [23], may account for the low prevalence of ST (no case of definite ST, 4 unknown cause death cases, possibly due to ST).
Multivariate analysis shows a strong correlation of BMS utilization with total mortality.This result deserves a comment because it is likely due to the presence of a higher risk profile (older age, more often STEMI presentation, severe comorbidities, necessity for oral anticoagulation, lower LVEF and compromised hemodynamic status, Table 6), which leads to the choice of BMS utilization in order to avoid prolonged dual antiplatelet therapy or is based on the supposedly worse cost/efficacy ratio in very compromised clinical conditions.Probably BMS utilization alone is not really an independent predictor of mortality but rather an indicator of a worse clinical setting.This hypothesis is supported by the very high acute in-hospital mortality of this subset of patients (6/50, 12% vs 13/264, 4.5% in the BMS and DES group respectively), without a concomitant higher TVR versus DES group in patients who survived the in-hospital phase (Table 6).
We cannot derive a difference between the first and second generations DES from our data, even if recent studies showed no significant advantage of the latter one [24].
The presence of a Syntax Score >32 is also a predictor of mortality and this result is in line with the increased major adverse event rate at 3rd and 5th years in higher Syntax Score subgroup in Syntax trial [25,26].Notably, our patients with Syntax Score >32 had lower LVEF, higher Euroscore, more frequent Killip class 3-4 (Table 6).All these features made us to prefer percutaneous coronary treatment over a surgical approach.
In our experience, besides anatomical situation, clinical features (age, ACS presentation, PAD) are important determinants of prognosis.An integrated clinical judgment may be represented by the new risk scores (global and clinical Syntax Score) [27,28], and is mandatory to choose the right revascularization strategy for each individual patient in order to guarantee the best benefit-torisk ratio and quality of life.

Table 1 .
Clinical characteristics of 317 patients treated with PCI for de novo unprotected left main lesions.

Table 2 .
Angiographic characteristics of the studied population.

Table 4 .
Clinical events at follow up.

Table 5 .
Independent predictors of total mortality at multivariate analysis.

Table 6 .
Clinical characteristics of patients treated with BMS or DES.

Table 7 .
Clinical characteristics and events in the three Syntax score groups.