Author + information
- Received July 27, 2005
- Revision received October 25, 2005
- Accepted October 31, 2005
- Published online February 21, 2006.
- Matthew J. Price, MD, FACC⁎,
- Ecaterina Cristea, MD†,
- Neil Sawhney, MD⁎,
- John A. Kao, MD, FACC⁎,
- Jeffrey W. Moses, MD, FACC†,‡,
- Martin B. Leon, MD, FACC†,‡,
- Ricardo A. Costa, MD†,
- Alexandra J. Lansky, MD, FACC†,‡ and
- Paul S. Teirstein, MD, FACC⁎,‡,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Paul S. Teirstein, Scripps Clinic, 10666 North Torrey Pines Road, Maildrop S1056, La Jolla, California 92037.
Objectives This study was performed to evaluate the clinical and serial angiographic outcomes of patients undergoing sirolimus-eluting stent (SES) implantation for unprotected left main coronary artery (LMCA) stenosis.
Background The efficacy of SES has led to their expanded use for off-label indications, including LMCA disease.
Methods Unprotected LMCA intervention with SES was attempted in 50 patients. Surveillance angiography was performed at three and nine months’ follow-up.
Results The target lesion involved the distal LMCA in 47 patients (94%). In-lesion restenosis occurred in 21 patients (42%), was focal in 85% of cases, and in 82% involved the branch ostia, sparing the LMCA itself. Target lesion revascularization (TLR) occurred in 19 patients (38%) over a mean follow-up of 276 ± 57 days; TLR was ischemia-driven in 7 patients (14%). Late loss was significantly greater within the left circumflex (LCX) ostium compared to the parent vessel (PV) of the LMCA bifurcation (0.83 ± 0.89 mm vs. 0.49 ± 0.72 mm, p = 0.04). Late loss continued to increase between three- and nine-month follow-up. Final minimal luminal diameter and maximal balloon pressure were independent predictors of restenosis of the PV.
Conclusions Restenosis is a frequent finding when serial angiographic follow-up is performed after SES implantation for unprotected distal LMCA lesions. Restenosis is usually focal, most often involves the LCX ostium, and often occurs without symptoms.
The accepted treatment of unprotected left main coronary artery (LMCA) disease is coronary artery bypass surgery (CABG). The approval of sirolimus-eluting stents (SES) has increasingly led to their off-label use in this patient population despite the absence of randomized data supporting its efficacy (1). The goal of this study was to evaluate the serial angiographic outcomes of consecutive patients at our institution who underwent SES implantation for unprotected LMCA stenosis.
Patient selection and study population
All patients who underwent elective, urgent, or emergent placement of SES for lesions within the unprotected LMCA at Scripps Clinic between February 2003 and July 2004 were included in this analysis. During this time period, implantation of SES followed by serial angiography was the default approach for percutaneous LMCA intervention. Patients were considered eligible for percutaneous LMCA intervention if they were deemed to be a high surgical risk by a cardiac surgeon or if they refused CABG despite the recommendations of a cardiac surgeon, and agreed to return for surveillance angiography at three and nine months to exclude restenosis.
Patients not previously taking clopidogrel were loaded with 300 mg or 600 mg before the procedure or at the conclusion of the procedure before leaving the catheterization laboratory. Either intraprocedural heparin (with a goal activated clotting time of ≥250 s) or bivalirudin was administered during the procedure. All patients were instructed to take daily aspirin 325 mg and clopidogrel 75 mg indefinitely.
One of three treatment techniques were used for lesions involving the LMCA bifurcation: 1) “kissing stents” where two stents were simultaneously deployed across the ostia of the left anterior descending (LAD) and left circumflex (LCX) arteries with proximal overlap within the body of the LMCA; 2) “crush” stenting, in which the stent within the side branch (usually, the LCX) was deployed first, and then the proximal portion of this stent projecting within the LMCA was “crushed” by deployment of the stent within the parent vessel; and 3) “rescue” stenting where a single stent was deployed from the left main across the ostium of either the LAD or LCX with subsequent balloon angioplasty of the side branch if needed. Generally, operators performed “kissing stents” when there was disease within the ostium of an LCX of significant caliber, “crush” stenting when there was disease within the ostium of an LCX of smaller caliber, and “rescue” stenting when there was no disease within the LCX or if the LCX was diminutive. Intra-aortic balloon counterpulsation, glycoprotein IIb/IIIa use, and intravascular ultrasound (IVUS) were used at the discretion of the operator.
Serial cardiac enzymes (creatine kinase and CK-MB) were drawn every 12 h post-intervention until discharge. Follow-up angiography was performed at three and nine months, or earlier for symptoms of ischemia. Routine stress testing was not performed.
Quantitative coronary angiography was performed by an off-site, independent core laboratory (Cardiovascular Research Foundation, New York, New York) using a validated edge-detection program (Cardiovascular Measurement System, Medis Medical Imaging Systems, Nuenen, the Netherlands). For purposes of bifurcation analysis, the LMCA of every patient was divided into the parent vessel and side branch. By convention, the parent vessel was defined as the LMCA and proximal LAD, and the side branch as the LCX. The target lesion was defined as involving the distal LMCA if it was within 3 mm of the branch ostia. For the parent vessel and side branch, the reference vessel diameter (RVD), minimal luminal diameter (MLD), acute luminal gain (MLD immediately after the procedure minus the MLD before the procedure), binary restenosis (stenosis of ≥50% of the luminal diameter), percent diameter stenosis, and late luminal loss (MLD immediately after the procedure minus the MLD at follow-up) were measured separately. In-stent restenosis was defined as within the stented segment and in-lesion restenosis as spanning the stented segment plus the 5-mm proximal and distal peristent area (edges).
End points and definitions
The LMCA was defined as “unprotected” if there was no history of CABG, if CABG had been performed and no grafts to the left coronary system were patent, or if CABG had been performed to the right coronary artery only. Non–Q-wave myocardial infarction (MI) was defined as creatine kinase greater than two times the upper limit of normal with an abnormal CK-MB in the absence of pathological Q waves. Stent thrombosis was defined as any of the following: angiographic demonstration of stent closure or intrastent thrombus, unexplained sudden death, or MI without concomitant documentation of a patent stent. Acute stent thrombosis was defined as thrombosis occurring within 24 h, subacute thrombosis between 24 h and 30 days, and late thrombosis more than 30 days after the index procedure. Target lesion revascularization (TLR) was defined as any repeat intervention (surgical or percutaneous) to treat a stenosis anywhere within the LMCA or within 10 mm distal to the LAD and LCX ostia. Target lesion revascularization was further characterized as “ischemia-driven” if signs or symptoms of ischemia were present. Major adverse cardiac events (MACE) were defined as any MI, any TLR, thrombosis, or death. Technical success was defined as a final diameter stenosis ≤30% and Thrombolysis In Myocardial Infarction (TIMI) flow grade 3.
Each patient’s pre-procedure cardiac surgical operative mortality risk at 30 days was estimated post-hoc by using a risk scoring system (the EuroSCORE) (2,3). “High-risk” was defined as an estimated operative mortality of ≥5%.
The computer-based analysis program SPSS (Statistical Package for the Social Sciences, 10.1 for PC, SPSS Inc., Chicago, Illinois) was used for statistical calculations. The chi-square test, or Fisher exact test when any expected cell count was <5 for a 2 × 2 table, was used to detect differences in categorical variables; p < 0.05 was considered significant. Comparison of continuous variables was performed using Student ttests. Multiple logistic regression analysis (forward conditional) was applied to estimate the association between the proposed risk factors and each of the outcomes (all-cause death, non-cardiac death, TLR, MI, parent vessel restenosis, side branch restenosis, and in- and out-of-hospital major adverse coronary events). Risk factors included age, gender, hypertension, diabetes, peripheral vascular disease, prior MI, left ventricular ejection fraction <40%, creatinine clearance ≤45 ml/min (according to the Cockroft-Gault formula), high cardiac surgery risk, non-elective procedure, in-stent restenosis target lesion, angiographic calcification, stenting of both limbs of the LMCA bifurcation, kissing stents, “crush” stenting, use of IVUS post-stent deployment, use of intra-aortic balloon counterpulsation, RVD, final MLD, and the maximal balloon dilation pressure. Kaplan-Meier survival functions were analyzed and graphed using Prophet (version 5.0, Bolt, Beranek, and Newman, Inc., sponsored by the National Center for Research Resources of the National Institutes of Health).
The study protocol was approved by the Scripps Clinic Institutional Review Board, and appropriate informed consent was obtained.
A total of 50 patients underwent attempted percutaneous intervention of the unprotected LMCA with SES. Patient and procedural characteristics are listed in Table 1.Patients were predominantly at high risk for cardiac surgery, with 58% having an estimated EuroSCORE operative mortality of ≥5%. The procedure was elective for stable angina or ischemia in 33 patients (66%), and was urgent or emergent in 17 patients (34%) (13 patients with unstable angina or non–ST-segment elevation MI, one patient with ST-segment elevation MI, two patients for “bailout” of iatrogenic left main dissection, and one patient with acute closure after diagnostic IVUS). Cardiogenic shock was present in four patients (8%). The target lesion involved the distal LMCA in 47 patients (94%). The lesion type was de novo atherosclerosis in 43 patients (86%), in-stent restenosis within bare metal stents in 5 patients (10%) (2 of whom had failed prior intracoronary radiation therapy), and iatrogenic LMCA dissection in 2 patients (4%). The average lesion length was 14.9 ± 6.3 mm, pre-procedural TIMI flow grade was <3 in 3 patients (6%), and moderate-to-severe calcification was present in 14 patients (28%). Procedural characteristics are shown in Table 2.No patient received rotational or directional coronary atherectomy.
Technical success was achieved in all patients. There were no in-hospital deaths. Acute stent thrombosis occurred in two patients (4%), resulting in a non–Q-wave MI in one patient and a Q-wave MI in the second. Both episodes of thrombosis occurred within 12 h of “kissing stents” within the distal LMCA. One of these patients failed to receive periprocedural clopidogrel loading due to a miscommunication regarding his medication status. Neither patient received intraprocedural glycoprotein IIb/IIIa inhibitors. Repeat revascularization with balloon angioplasty was successful in both patients. A third patient underwent TLR (total in-hospital TLR = 6%) for residual LMCA stenosis outside the previously stented segment that was noted during a planned staged percutaneous coronary intervention (PCI) of the right coronary artery before hospital discharge. The combined end point of death, any MI, TLR, or thrombosis occurred more frequently in non-elective compared to elective cases (4 of 17 [24%] vs. 1 of 33 [3%], p = 0.04 by Fisher exact test.) The average hospital stay post-procedure was 2.7 ± 4.0 days (median 1 day, range 1 to 24 days).
Clinical events at follow-up are shown in Table 3.Clinical follow-up was 100% over a median of 281 days (range, 144 to 391 days). There was one episode of out-of-hospital non–Q-wave MI. Out-of-hospital TLR occurred in a total of 19 patients (38%); of these, only 7 patients (14%) had ischemia-driven TLR. The target lesion had been treated with bifurcation stenting in 18 of the 19 patients undergoing repeat revascularization. Of the eight patients treated with crush stenting, TLR occurred in three of the five that had final kissing-balloon post-dilation and two of the three that did not. There were no significant differences in TLR between elective and non-elective cases (overall TLR, 12 of 33 [36%] vs. 7 of 17 [41%], p = NS by chi-square test; ischemic TLR, 5 of 33 [14%] vs. 2 of 17 [15%], p = NS by Fisher exact test), or between de novo and non-de novo lesions (overall TLR, 17 of 43 [39%] vs. 2 of 7 [29%], p = NS by Fisher exact test; ischemic TLR, 6 of 43 [14%] vs. 1 of 7 [14%], p = NS by Fisher exact test).
There were five deaths, four non-cardiac and one cardiac. The combined end point of out-of-hospital MACE occurred in 22 patients (44%). Kaplan-Meier survival free of MACE is shown in Figure 1.
Quantitative coronary angiography
Baseline and follow-up angiographic results are shown in Table 4.Overall, follow-up angiography was performed on 49 patients (98%), and was of suitable quality for quantitative coronary angiography in 48 patients. Three-month follow-up angiography was complete in 47 patients (94%) at a median of 98 days (range, 44 to 199 days). Nine-month angiography was complete in 31 patients (90% of patients alive and without repeat revascularization at the 3-month follow-up) at a median of 287 days (range, 165 to 447 days). In-lesion restenosis was present in 21 patients (44%) and was focal (Mehran classification I) (4) in 85% of cases, with an average lesion length of 6.1 ± 2.1 mm. The distribution of restenosis is shown in Table 5.Restenosis was most frequent within the left circumflex ostium, and occurred within the LMCA itself in only four patients (8%). In-stent late loss was significantly greater within the LCX side branch compared to the parent vessel of the LMCA bifurcation (0.83 ± 0.89 mm vs. 0.49 ± 0.72 mm, p = 0.04). Restenosis developed in five patients (10%) between the three- and nine-month follow-up; the in-stent late loss between three- and nine-month surveillance angiography in patients without TLR at the three-month follow-up was 0.22 ± 0.66 mm in the parent vessel and 0.41 ± 0.71 mm in the LCX ostium.
Multivariate predictors of outcomes are listed in Table 6.There were no significant predictors of in-hospital MACE. Creatinine clearance ≤45 ml/h was associated with out-of-hospital TLR, death, and MACE. Diabetes mellitus tended to be associated with parent vessel restenosis (p = 0.052). Maximal balloon pressure within the parent vessel and final MLD of the parent vessel were independent predictors of parent vessel restenosis, while there were no independent predictors of LCX restenosis.
The major finding of our study is that PCI of the unprotected LMCA with SES is feasible and relatively safe, but is limited by frequent restenosis, most often involving the left circumflex ostium. In our series of consecutive patients with predominantly distal left main disease and mandated, serial angiographic surveillance, the rate of TLR was 38% and the rate of restenosis was 44% over a median follow-up of 9.2 months.
We observed higher revascularization and restenosis rates than previously reported for drug-eluting stent (DES) intervention within the LMCA (5–8). This may be due to our high rate of angiographic follow-up in combination with the large proportion of patients undergoing LMCA bifurcation stenting in our study. The rate of mandated follow-up angiography in previously reported series was quite low, generally under 50%. While one study reported a follow-up angiography rate of 84%, in that study, only 40% of patients had bifurcation stenting (5,7,8). In comparison, our angiographic follow-up was nearly complete (98%), and a very large proportion of patients (84%) had stenting of both limbs of the LMCA bifurcation. In our study, TLR was performed for signs or symptoms of ischemia in only seven patients, resulting in a “clinical” ischemia-driven TLR rate of only 14%. Therefore, our high revascularization rate may have been due to the “oculostenotic reflex” in response to asymptomatic left main or ostial major branch vessel disease uncovered during surveillance angiography. Although such lesions may jeopardize a large amount of myocardium, the benefit of revascularization in this setting has not been proven. While the clinical appropriateness of our approach remains unanswered by this study, we believe that the low revascularization rates observed in previous studies with clinical follow-up alone (7) underestimate the actual number of patients with DES failure after LMCA intervention. The anatomic pattern of restenosis we observed—focal in 85% of cases and involving only the branch ostia in 81%—is consistent with previous reports (6,9). Therefore, while failure is frequent, it occurs uncommonly within the LMCA itself, and is usually amenable to repeat revascularization using a percutaneous approach.
Our disappointing results may also be due to the large proportion of lesions involving the distal left main that were treated with bifurcation stenting. Distal left main stenosis has been shown to be associated with adverse events after DES implantation (7). Indeed, every case of restenosis in the three largest published case series have involved target lesions within the distal LMCA (5–7), the majority of which had been treated with stents in both branches. Previous studies report restenosis rates ranging from 2% to 19%, but the proportion of patients in these studies that were treated with bifurcation stenting varied from 8% to 74%, respectively (8,10). For example, Chieffo et al. (10) performed LMCA intervention with DES in 85 patients; all 12 episodes of TLR that occurred were in the 51 patients that had undergone stenting of both limbs of the LMCA, resulting in a restenosis rate (24%) that approaches our findings.
We found that SES failure was usually due to restenosis within the left circumflex ostium. The maximal balloon pressure used and the final MLD were independently associated with restenosis of the parent vessel but not for the left circumflex branch. This may be due to unique characteristics of the left circumflex ostium. For example, the left circumflex ostium often contains an acute bend that may predispose to non-apposition of stent struts. Conclusions cannot be drawn about the relative merit of the stenting techniques used because the bifurcation stent approach was not randomized, the number of patients receiving the “crush” was small, and in some patients post-crush kissing balloon post-dilation (which may improve outcomes) (11) was not performed. Regardless of the technique used, our findings that the maximal balloon pressure and final MLD are independently associated with restenosis of the parent vessel emphasize the need for high-pressure stent deployment and adequate post-dilation with non-compliant balloons to ensure the long-term success of the parent vessel when undertaking distal LMCA intervention with SES.
The temporal pattern of restenosis after DES for unprotected LMCA disease has not been previously reported. We found that while the majority of cases of restenosis occurred within three months of the procedure, significant late loss accrued between three- and nine-month follow-up, leading to restenosis and TLR in an additional five patients (10%). We chose to perform initial surveillance angiography at three months post-procedure based on earlier experiences of unprotected LMCA PCI that found an early risk of mortality within the first two to four months post-procedure, possibly due to restenosis (12,13). Given our findings of continued incremental late loss, surveillance angiography at a single early time point, while disclosing early restenosis (e.g., three months), is not sufficient to exclude the development of later restenosis. Whether further late loss may occur after nine months is unknown, although no late “catch-up” has been observed in previous studies evaluating DES implanted for non-bifurcation lesions outside the LMCA (14,15).
Severe renal insufficiency (creatinine clearance ≤45 ml/min) was an independent predictor of death, TLR, and MACE in our study. Renal dysfunction is known to be associated with in- and out-of-hospital mortality after PCI, independent of restenosis (16,17). While a decreased anti-proliferative effect of sirolimus in patients with renal insufficiency patients can not be excluded, other studies have found that SES provide substantial reductions in restenosis rates outside of the LMCA compared to historical bare metal controls in such patients (17).
We found that unprotected left main coronary intervention with SES was relatively safe. The majority of the deaths (four of five patients) were non-cardiac and unrelated to restenosis. No deaths occurred in patients at low-risk for cardiac surgery, although by multivariate analysis, cardiac surgical risk was not an independent predictor of outcome. Acute stent thrombosis occurred in two patients (4%), both of whom had kissing stents, and in neither case was a glycoprotein IIb/IIIa inhibitor used. Our findings are consistent with a randomized study of SES treatment for coronary bifurcation lesions that showed a relatively high rate of early thrombosis in patients receiving stents in both branches (3%) (18), reinforcing the importance of this significant complication during complex intervention. While it is unknown whether more aggressive use of glycoprotein IIb/IIIa inhibitors may have reduced the rate of stent thrombosis in our study, it is now our policy to use glycoprotein IIb/IIIa inhibitors in all unprotected LMCA interventions in the absence of contraindications. Pre-procedure loading with clopidogrel may also be prudent.
This was a single-center, non-randomized trial with a relatively small number of patients. We included “all-comers” undergoing implantation of SES within the unprotected LMCA, and therefore any selection bias should be minimized. Routine non-invasive stress testing was not performed, so in asymptomatic patients who underwent TLR the presence of ischemia cannot be determined. A 600-mg loading dose of clopidogrel was used before PCI in some patients; this dose has not been approved by the U.S. Food and Drug Administration, and is therefore considered “off-label.”
Distal left main coronary intervention with SES is safe. When serial surveillance angiography is performed, restenosis occurs frequently, is typically focal, and involves the left circumflex ostium. Restenosis is usually silent, and late loss continues to accrue after three months, suggesting that longer-term angiographic follow-up is necessary. While aggressive balloon dilation and stent expansion can optimize the longer-term results within the main vessel of the left main coronary bifurcation, techniques to improve success within the left circumflex ostium are required. In the DES era, the longer-term success of unprotected LMCA intervention, particularly in the presence of distal disease, remains a challenge.
The authors thank David Cloutier for assistance with the statistical analysis.
↵‡ Dr. Moses is a consultant for and stock shareholder of Johnson & Johnson. Dr. Leon is a stock shareholder of Johnson & Johnson. Dr. Lansky receives research grants from Johnson & Johnson. Dr. Teirstein receives royalties and research grants from Johnson & Johnson.
- Abbreviations and Acronyms
- coronary bypass surgery
- drug-eluting stents
- intravascular ultrasound
- left anterior descending coronary artery
- left circumflex artery
- left main coronary artery
- major adverse cardiac events
- myocardial infarction
- minimal luminal diameter
- percutaneous coronary intervention
- reference vessel diameter
- sirolimus-eluting stent
- target lesion revascularization
- Received July 27, 2005.
- Revision received October 25, 2005.
- Accepted October 31, 2005.
- American College of Cardiology Foundation
- Nashef S.A.,
- Roques F.,
- Michel P.,
- Gauducheau E.,
- Lemeshow S.,
- Salamon R.
- Nashef S.A.,
- Roques F.,
- Hammill B.G.,
- et al.
- Mehran R.,
- Dangas G.,
- Abizaid A.S.,
- et al.
- Park S.J.,
- Kim Y.H.,
- Lee B.K.,
- et al.
- Chieffo A.,
- Stankovic G.,
- Bonizzoni E.,
- et al.
- Valgimigli M.,
- van Mieghem C.A.,
- Ong A.T.,
- et al.
- Chieffo A.,
- Tsagalou E.,
- Iakovou I.,
- et al.
- Takagi T.,
- Stankovic G.,
- Finci L.,
- et al.
- Tan W.A.,
- Tamai H.,
- Park S.J.,
- et al.
- Fajadet J.,
- Morice M.C.,
- Bode C.,
- et al.
- Sousa J.E.,
- Costa M.A.,
- Abizaid A.,
- et al.
- Colombo A.,
- Moses J.W.,
- Morice M.C.,
- et al.