Long-Term Outcomes After Stenting Versus Coronary Artery Bypass Grafting for Unprotected Left Main Coronary Artery Disease

10-Year Results of Bare-Metal Stents and 5-Year Results of Drug-Eluting Stents From the ASAN–MAIN (ASAN Medical Center–Left MAIN Revascularization) Registry

Study population and revascularization procedures

The study population consisted of 2 follow-up cohorts: 1) a 10-year follow-up cohort of patients with bare-metal stents (BMS) and concurrent CABG; and 2) a 5-year follow-up cohort of patients with DES and concurrent CABG, as a part of the ASAN-MAIN (ASAN medical center–left MAIN revascularization) registry; a single-center, retrospective study designed to evaluate the treatment effects of stenting and CABG for LMCA disease in the “real world.”

For the 10-year follow-up cohort, consecutive patients with unprotected LMCA disease (defined as stenosis of more than 50%) who underwent PCI with BMS or isolated CABG at the Asan Medical Center (Seoul, Korea) between January 1, 1995, and April 30, 1999, were included, and the follow-up period was extended through May 31, 2009, to ensure that all patients had an opportunity to provide at least 10 years of follow-up information. For the 5-year follow-up cohort, consecutive patients with unprotected LMCA disease who underwent DES implantation or isolated CABG between January 1, 2003, and May 31, 2004, were included, and the follow-up period was extended through June 30, 2009, to ensure that all patients had an opportunity for at least 5 years of follow-up information if they were alive. Patients who had prior CABG, those who underwent concomitant valvular or aortic surgery, and those who had myocardial infarction (MI) with ST-segment elevation or presented with cardiogenic shock were excluded.

Patients underwent PCI, instead of CABG, because of either the patient's or physician's preference or the high risk associated with CABG. Methods of stent implantation for patients with LMCA disease have been described previously (9,15). The choice of the specific type of DES (i.e., sirolimus-eluting stent [Cypher, Cordis, Johnson & Johnson, Miami Lakes, Florida] or paclitaxel-eluting stent [Taxus, Boston Scientific, Natick, Massachusetts]) was left to the operator's discretion. All patients undergoing PCI were prescribed aspirin plus ticlopidine (loading dose, 500 mg) or clopidogrel (loading dose, 300 or 600 mg) before or during the coronary intervention. After the procedure, aspirin was continued indefinitely. Patients treated with BMS were prescribed ticlopidine (250 mg twice a day) for at least 1 month, and patients treated with DES were prescribed clopidogrel (75 mg once a day) for at least 6 months, regardless of DES type (16). Treatment beyond this duration was at the discretion of the physician. Surgical revascularization was performed with the use of standard bypass techniques (1). Whenever possible, the internal thoracic artery was used preferentially for revascularization of the left anterior descending artery. Complete revascularization was performed when possible using arterial conduits or saphenous vein grafts.

This study was approved by the local institutional review board.

Outcome variables and definitions

The primary safety outcomes were death and the composite of death, Q-wave MI, or stroke. The primary efficacy outcome was target vessel revascularization (TVR).

All deaths were considered cardiac unless an unequivocal noncardiac cause could be established. The diagnosis of MI was assessed by the universal definition of MI (17). Q-wave MI was defined as the documentation of a new pathologic Q-wave after index treatment. Stroke, as indicated by neurologic deficits, was confirmed by a neurologist on the basis of imaging studies. TVR was defined as repeat revascularization of the treated vessel, including any segments of the left anterior descending or left circumflex coronary artery; TLR was defined as any revascularization performed on the treated segment. The occurrence of stent thrombosis in the DES group was assessed by the Academic Research Consortium definitions (18).

All events were carefully verified and adjudicated by independent clinicians.

Data collection and follow-up

Clinical, angiographic, procedural or operative, and outcome data were recorded in the dedicated PCI and surgical databases by independent research personnel (16). Clinical follow-up after PCI and after CABG was recommended at 1 month, 6 months, and 1 year and then annually thereafter. Routine angiographic follow-up for all patients treated with PCI was recommended 6 to 10 months after the procedure. However, patients who were at high risk for procedural complications of angiography and had no symptoms or signs of ischemia, as well as patients who declined the recommendation, did not undergo routine follow-up angiography. For patients who underwent CABG, angiographic follow-up was recommended only if there were ischemic symptoms or signs during follow-up.

For validation of complete follow-up data, information about vital status was obtained through May 31, 2009, for the 10-year follow-up cohort or through June 30, 2009, for the 5-year follow-up cohort, from the National Population Registry of the Korea National Statistical Office with the use of a unique personal identification number. To ensure accurate assessment of clinical end points, additional information was obtained from visits or telephone contacts with living patients or family members and from medical records obtained from other hospitals, as necessary.

Statistical analysis

Continuous variables were compared with the ttest or Wilcoxon rank sum tests, and categorical variables were compared with the chi-square statistics or Fisher exact test, as appropriate. Survival curves were constructed using Kaplan-Meier estimates and compared with the log-rank test.

To compensate for the nonrandomized design of our observational study and to test for selection bias, we used propensity-score methods (19). The propensity scores were estimated without regard to outcomes, using multiple logistic regression analysis. A full nonparsimonious model was developed that included all variables shown in Table 1.Model discrimination was assessed with c-statistics, and model calibration was assessed with Hosmer-Lemeshow statistics. The individual propensity score, as well as the type of revascularization, was incorporated into the Cox regression model as a covariate to calculate the propensity-adjusted hazard ratio (HR). The propensity score was also subdivided into quartiles (20). Treatment effect was separately estimated within each quartile, and quartile estimates were combined to measure an overall estimate of the treatment effect. All statistical analyses were performed using SAS version 9.1 (SAS Institute, Cary, North Carolina). A 2-tailed p value of <0.05 was considered statistically significant.

Baseline characteristics and revascularization procedures

Between January 1995 and April 1999, a total of 350 patients with unprotected LMCA disease met the criteria for inclusion. Of these, 100 were treated with PCI with BMS and 250 with CABG, and were eligible for the 10-year follow-up cohort. Between January 2003 and May 2004, a total of 395 patients with unprotected LMCA disease met the study inclusion criteria and met none of the criteria for exclusion. Of these, 176 who were treated with PCI with DES and 219 with CABG were eligible for the 5-year follow-up cohort.

The baseline characteristics of study patients according to revascularization procedure are shown in Table 1. CABG patients had a significantly higher-risk clinical and angiographic profile than PCI patients. Procedural characteristics are presented in Table 2.For the 10-year follow-up cohort, complete revascularization was achieved in 67 BMS patients (67.0%) and in 209 CABG patients (83.6%; p = 0.001). For the 5-year follow-up cohort, complete revascularization was achieved in 126 DES patients (71.6%) and in 176 CABG patients (80.4%; p = 0.04). In the DES group, sirolimus-eluting stents were used exclusively, and intravascular ultrasound (IVUS)-guided stenting was performed in almost 90% of patients.

Observed in-hospital and long-term event rates

The observed (unadjusted) in-hospital clinical events are shown in Table 3.Individual rates of in-hospital death, MI, stroke, and TVR were not different between the 2 groups. However, the rate of the composite of death, Q-wave MI, or stroke in the CABG group was significantly higher than in the stenting group, irrespective of stent type.

For the 10-year follow-up cohort, complete follow-up data for major clinical events at 10 years were obtained in 96.9% of the overall cohort (96.0% for the BMS group and 97.2% for the CABG cohort). For the 5-year follow-up cohort, complete follow-up data for major clinical events at 5 years after index treatment were obtained in 97.2% of the overall cohort (97.7% for the DES group and 96.8% for the CABG group). Crude event rates and unadjusted long-term event-free survival curves according to treatment approach are presented in Table 3and Figure 1.The observed (unadjusted) long-term rates of death and the composite of death, Q-wave MI, or stroke were significantly lower in the PCI group than in the CABG group, whereas the rate of TVR was higher in the PCI group.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Unadjusted Survival Curves

(A and B)show long-term survival, (C and D)show survival free from Q-wave myocardial infarction or stroke, and (E and F)show survival free from target vessel revascularization. A, C, and Dshow 10-year unadjusted survival curves between bare-metal stents (BMS) and concurrent coronary artery bypass grafting (CABG), and B, D, and Fshow 5-year unadjusted survival curves between drug-eluting stents (DES) and concurrent CABG.

In the DES group, 3 patients had definite thrombosis, none had probable thrombosis, and 2 had possible thrombosis. At 5-year follow-up, the cumulative incidence of definite or probable stent thrombosis was 1.8%. Among patients with definite or probable stent thrombosis, 1 patient had acute, 1 had subacute, and 1 had very late (4.3 years after the procedure) thrombosis. Acute and subacute cases were on dual antiplatelet therapy at the time of thrombotic events, and a very late case arose 2 months later, after all antiplatelet therapy was interrupted.

Adjusted primary safety and efficacy outcomes

Table 4and Figure 2represent the adjusted primary safety and efficacy outcomes based on revascularization procedure. The c-statistic for the propensity score model was 0.87 (Hosmer-Lemeshow goodness-of-fit p = 0.56) in the 10-year cohort and 0.88 (Hosmer-Lemeshow goodness-of-fit p = 0.97) in the 5-year cohort. All covariates differed nonsignificantly after propensity score adjustment (Table 1). In the propensity score– and propensity score quartile–adjusted analyses, long-term risks of death and the composite of death, Q-wave MI, or stroke in the 10-year follow-up cohort of BMS and concurrent CABG were similar in the 2 groups. However, the adjusted risk of TVR was significantly higher in the PCI group. These findings were consistent with those of the 5-year follow-up cohort of DES and concurrent CABG.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Adjusted Survival Curves

(A and B)show long-term survival, (C and D)show survival free from Q-wave myocardial infarction (MI) or stroke, and (E and F)show survival free from target vessel revascularization (TVR). A, C, and Dshow 10-year adjusted survival curves between bare-metal stents (BMS) and concurrent coronary artery bypass grafting (CABG), and B, D, and Fshow 5-year adjusted survival curves between drug-eluting stents (DES) and concurrent CABG. The curves represent estimations from the Cox regression model at the mean level of the propensity score.

Study limitations

First, our study is an observational study, and therefore, the choice of treatment was at the discretion of the physician or the patient. Although we used propensity analysis to enable a rigorous adjustment for selection bias and confounding, there is no way to eliminate bias caused by the influence of unmeasured confounders, or the presence of patients deemed to be ineligible for 1 of the procedures. Thus, rather than subgroup analysis, LMCA-specific randomized trials with a larger number of patients and longer follow-up, such as the PRE-COMBAT (Randomized Comparison of Bypass Surgery Versus Angioplasty Using Sirolimus-Eluting Stent in Patients With Left Main Coronary Artery Disease) or EXCEL (Evaluation of Xience Prime versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization) clinical trial, are needed to confirm whether PCI treatment for LMCA disease is equivalent to CABG (14). Second, our analysis was underpowered to detect significant differences in mortality and serious composite outcomes among the treatment groups and to assess the long-term risk of stent thrombosis, due to the limited number of events. Third, we did not use a detailed scoring system (i.e., the SYNTAX score) to more accurately reflect atherosclerotic disease burden and anatomic complexity. Fourth, the average surgical risk of our study population, as measured by EuroSCORE and Parsonnet score, was relatively low. Finally, the particulars of clinical practice as well as the specific expertise of the interventional cardiologists and cardiac surgeons in our institution may differ from those of other institutions and practitioners, potentially limiting the reproducibility of these results in other settings.