Article Text

PDF

Original article
Long-term outcomes of percutaneous coronary intervention versus coronary artery bypass grafting for unprotected left main coronary bifurcation disease in the drug-eluting stent era
  1. Kiyuk Chang1,2,
  2. Yoon-Seok Koh1,3,
  3. Seung Hee Jeong4,5,
  4. Jong-Min Lee1,3,
  5. Sung-Ho Her1,6,
  6. Hun-Jun Park1,2,
  7. Pum-Joon Kim1,2,
  8. Young-Hak Kim7,
  9. Wook-Sung Chung1,2,
  10. Hyeon-Woo Yim4,5,
  11. Seung-Jung Park7,
  12. Ki Bae Seung1,2
  1. 1Department of Cardiovascular Medicine, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea
  2. 2Cardiovascular Center, Seoul St Mary's Hospital, Seoul, Republic of Korea
  3. 3Cardiovascular Center, Uijeongbu St Mary's Hospital, Uijeongbu, Republic of Korea
  4. 4Department of Preventive Medicine, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea
  5. 5Clinical Research Coordinating Center of Catholic Medical Center, Seoul, Republic of Korea
  6. 6Cardiovascular Center, Daejeon St. Mary's Hospital, Daejeon, Republic of Korea
  7. 7Department of Cardiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea
  1. Correspondence to Professor Ki Bae Seung, Cardiovascular Center and Cardiology Division, Seoul St Mary's Hospital, 505 Banpodong, Seochogu, Seoul 137-701, Republic of Korea; kbseung{at}catholic.ac.kr

Abstract

Objectives There are limited data on long-term outcomes (ie, beyond 4 years) for patients with unprotected left main bifurcation disease who underwent percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG) in the drug-eluting stent (DES) era. This study therefore compared the treatment effects of PCI and CABG in unprotected left main bifurcation disease.

Methods 865 patients with unprotected left main bifurcation disease treated with either PCI using DES (n=556) or CABG (n=309) were evaluated between May 2003 and December 2009. PCI-treated patients were further categorised into simple stenting (n=360) or complex stenting (n=196).

Results Median follow-up was 4.2 years (IQR 2.9–5.2 years). After adjusting covariates with multivariate Cox hazard regression model and inverse probability of treatment weighting, the long-term cumulative rates of death (HR 0.95; 95% CI 0.62 to 1.45) or composite of death, Q-wave myocardial infarction, or stroke (HR 0.97, 95% CI 0.64 to 1.48) were not significantly different for patients undergoing PCI or CABG except for target-vessel revascularisation (TVR) (HR 4.42, 95% CI 2.39 to 8.18). The complex stenting group had similar long-term clinical outcomes compared with the simple stenting group except for TVR (HR 1.94, 95% CI 1.22 to 3.10). In further analysis with propensity score matching, overall findings were consistent.

Conclusions In patients with unprotected left main bifurcation disease, PCI using DES provides similar long-term (up to 5.2 years) clinical outcomes except for TVR compared with CABG. Complex and simple stenting yielded similar outcomes except for a higher TVR rate in complex stenting.

  • Atherosclerosis
  • chest pain clinic
  • coronary artery disease
  • coronary artery bypass grafting
  • coronary intervention (PCI)
  • coronary stenting
  • fractional flow reserve
  • hypertension
  • inflammation
  • interventional cardiology
  • intravascular ultrasound
  • left main bifurcation disease
  • percutaneous coronary intervention
  • platelets
  • radionuclide imaging
  • spasm
View Full Text

Statistics from Altmetric.com

While coronary artery bypass grafting (CABG) remains the standard treatment for unprotected left main coronary artery disease,1–3 percutaneous coronary intervention (PCI) has shown similar long-term clinical outcomes up to 5 years in terms of death and the composite of death, myocardial infarction (MI) or stroke, as compared with CABG.4 5 Furthermore, the introduction of drug-eluting stents (DES) along with advances in interventional techniques has significantly improved the outcomes of patients with unprotected left main disease. Despite these advances in the DES-based therapeutic approach, PCI for unprotected left main bifurcation disease remains challenging because of procedural risks, technical expertise, and lack of concrete and/or substantial evidence of long-term safety and efficacy. In addition, the high restenosis rate seen in non-left main bifurcation disease treated with DES and the potential risks of acute and late stent thrombosis in the left main area make treatment selection difficult between PCI and CABG in patients with unprotected left main bifurcation disease.6 7 There is thus a need to study CABG and PCI further comparatively for unprotected left main bifurcation disease in the DES era.

The wide spectrum of anatomical complexity of left main bifurcation disease has given rise to diverse stenting techniques ranging from simple crossover stenting to complex stenting. The Coronary Bifurcations: Application of the Crushing Technique Using Sirolimus-Eluting Stents (CACTUS) study, a randomised trial in which a simple stenting technique was compared with a complex stenting technique, specifically the crush technique, for non-left main true coronary bifurcation disease, demonstrated that the provisional one-stent technique is as effective as the complex stenting technique.8 Notably, the complex stenting technique was not associated with a higher incidence of adverse cardiovascular events at 6 months in that study. As a result of there being few relevant studies, it is not yet clear whether complex stenting techniques as a treatment for unprotected left main bifurcation disease increase the risk of major adverse cardiovascular events compared with simple stenting techniques.

To address these issues, we first compared the treatment effects of PCI using DES and CABG in patients with unprotected left main bifurcation disease. We then evaluated whether simple crossover stenting techniques are superior to complex stenting techniques in terms of long-term safety and efficacy in PCI-treated patients. Compared with previous reports dealing with unprotected left main coronary artery disease, this study is unique in recruiting only patients with true left main bifurcation disease from 18 cardiac centres throughout the country, and having a longer term of follow-up of up to 5.2 years.

Methods

Study population

This study was designed to evaluate the real-world outcomes of consecutive all comers with left main bifurcation disease undergoing PCI or CABG at 18 major cardiac centres in Korea from May 2003 and December 2009. As this registry was planned and conducted before the PRECOMBAT trial,9 some patients were enrolled in advance. In addition, PRECOMBAT-non-participating cardiac centres enrolled more patients with unprotected left main bifurcation disease. All hospitals perform high-volume PCI (>500 PCI/year) and are located throughout the country. Left main coronary artery disease was considered unprotected if there were no patent grafts to the left anterior descending (LAD) artery or left circumflex (LCX) artery. This large observational registry included demographic, clinical and procedural data, and short-term and long-term clinical outcome data of patients with unprotected left main bifurcation disease. After excluding patients with left main ostium or shaft diseases, and those receiving bare-metal stents, we enrolled only those with unprotected left main bifurcation disease treated with either PCI using DES or CABG. A total of 865 patients was finally analysed.

Left main bifurcation disease was defined as a stenosis of more than 50% involving both distal left main and at least one of ostia of LAD or LCx arteries, which matched with Medina classification 1,1,1; 1,1,0 or 1,0,1. Patients who had undergone previous CABG or concomitant valvular or aortic surgery, and those who had an ST-segment elevation MI or who presented with cardiogenic shock were excluded. The institutional review boards at each hospital approved the use of clinical data for this study, and all patients provided written informed consent.

Revascularisation procedures

Patients underwent PCI, instead of CABG, because of either patient or physician preference or the high risk associated with CABG. The principle of revascularisation procedures was complete revascularisation. In addition, we did not include hybrid revascularisation in this registry. Therefore, the patients who underwent PCI at the left main bifurcation received another PCI at the diseased segment distant from the left main bifurcation. The CABG group also followed the same strategy. However, it is possible on rare occasions to have included patients undergoing incomplete revascularisation because of technical or lesion-related problems. The technique of stent implantation and the kinds of DES were at the discretion of the operator. In the simple stenting group, a stent was implanted from the left main across the LCx artery and final kissing ballooning was left up to the operator. In the complex stenting group, several techniques including T-stenting, kissing stenting, crush technique or culotte technique were employed. The use of predilatation, intra-aortic balloon pump or intravascular ultrasound was at the discretion of the operator. All patients undergoing PCI were prescribed aspirin plus clopidogrel (loading dose 300 mg or 600 mg) before or during the procedure. After the procedure, aspirin was continued indefinitely, and clopidogrel was prescribed for more than 6 months regardless of the DES type implanted. Surgical revascularisation was performed using standard bypass techniques. Whenever possible, complete revascularisation was performed and the internal thoracic artery was used preferentially for revascularisation of the LAD artery.

Study endpoints and follow-up

The endpoints of the study were death; the composite of death, Q-wave MI, or stroke and target-vessel revascularisation (TVR). Death was defined as death from any cause. Q-wave MI was defined as documentation of a new abnormal Q wave after the index treatment. Stroke, as indicated by neurological deficits, was confirmed by a neurologist on the basis of imaging studies. TVR was defined as repeat revascularisation of the treated vessel, including any segments of the LAD and LCx arteries. All clinical outcomes of interest were confirmed by source document and were centrally adjudicated at the Cardiovascular Center of Seoul St Mary's Hospital, Seoul, Korea, by an independent group of clinicians whose members were unaware of patient status. Clinical, angiographic, procedural or operative and outcome data were collected in the dedicated PCI and surgical databases by independent research personnel. For validation of complete follow-up data, information on censored survival data was obtained to 30 September 2010 from the database of the National Health Insurance Corporation, Korea, with the use of a unique personal identification number. Information on the events of Q-wave MI, stroke and TVR were confirmed by telephone interview and a review of the whole medical records.

Statistical analysis

All statistical tests, performed with the use of SAS software, V.9.1, and R programming language, were two-sided, and p values of less than 0.05 were considered to indicate statistical significance. We first compared long-term outcomes between patients undergoing PCI and CABG. Cumulative event rates of clinical outcomes were next estimated and compared between simple stenting and complex stenting groups among PCI-treated patients. Continuous variables were compared with the t test or Mann–Whitney U test, and categorical variables were compared using the χ2 test or Fisher's exact test, as appropriate. Survival curves was obtained by Kaplan–Meier analysis and compared with the log-rank test.

To reduce the impact of treatment selection bias and potential confounding in an observational study, we performed rigorous adjustment for differences in baseline characteristics of patients by the use of weighted Cox proportional hazards regression models with the inverse probability of treatment weighting (IPTW).10 The propensity scores were estimated without regard to outcome variables, using multiple logistic regression analysis.11 Adjusted covariates including patient's age, sex, the presence or absence of a variety of clinical and coexisting conditions, the clinical diagnosis at the time of PCI, left ventricular function, the extent of diseased vessels, and the synergy between percutaneous coronary intervention with taxus and cardiac surgery (SYNTAX) score were used for the generation of propensity scores. For the IPTW, the weights for patients who underwent CABG were inverse of (1—propensity score), and weights for patients who underwent PCI were the inverse of propensity score. All model discrimination was assessed with c-statistics, and calibration was assessed with Hosmer–Lemeshow statistics. Subgroup analysis between simple stenting and complex stenting groups in PCI-treated patients was done in the same way.

With the Greedy 5→1 digit match algorithm,12 we created propensity score-matched pairs without replacement (a 1:1 match). After all of the propensity score matches were performed, we assessed the balance in baseline covariates between the two intervention groups with the paired t test or the Wilcoxon signed rank test for continuous variables, and McNemar's test or the marginal homogeneity test for categorical variables. Comparisons were completed with Cox regression models with robust standard errors that accounted for the clustering of matched pairs. As for the patients receiving PCI, the same methods mentioned above were applied to compare the clinical outcomes between the simple stenting and complex stenting groups.

Results

Characteristics of the overall study population

From May 2003 to December 2009, 865 patients with unprotected left main bifurcation disease were finally enrolled for the analyses; 556 patients were treated with PCI using DES (PCI group) and 309 patients were treated with CABG (CABG group). PCI-treated patients were further categorised into two groups consisting of those treated with a simple crossover stenting technique (simple stenting group, n=360) and with a complex stenting technique (complex stenting group, n=196). The baseline characteristics of the overall populations according to revascularisation procedures are shown in table 1. After propensity score matching, there was no longer any significant difference for any covariate including clinical manifestation, extent of diseased vessel, right coronary artery (RCA) involvement, left ventricular ejection fraction, and specifically EuroSCORE and SYNTAX in the matched cohorts (table 1).

Table 1

Baseline characteristics of the overall population and propensity score matching population according to treatment modalities

Characteristics of the PCI population

Compared with the simple stenting group, the complex stenting group had more stent implanted in a patient (1.7±1.0 vs 2.7±1.2; p<0.0001), and showed greater mean total stent length at the left main bifurcation (30.0±15.3 vs 48.6±24.3; p<0.0001). In the complex stenting group, crush stenting was most frequently employed (48.5%), followed by kissing stenting (35.7%) and T-stenting (14.8%). The baseline characteristics of the PCI population are shown in table 2. After propensity score matching, there was no longer any significant difference for any covariate in the matched cohorts (table 2).

Table 2

Baseline characteristics of the overall population and propensity score matching population according to stent techniques

Follow-up and clinical outcomes between the PCI and CABG groups

The median follow-up was 4.2 years (IQR 2.9–5.2 years) for the overall patient population. Complete follow-up data for major clinical events was obtained in 98.9%. During follow-up, 102 patients died (57 patients (10.2%) in the PCI group and 45 patients (14.6%) in the CABG group), and 115 patients had a primary clinical event including death, Q-wave MI or stroke (63 patients (11.5%) in the PCI group and 51 patients (16.5%) in the CABG group). In addition, five patients experienced stroke (one patient in the PCI group and four patients in the CABG group), and eight patients had Q-wave MI (five patients in the PCI group and three patients in the CABG group). TVR was performed in 97 patients (88 patients (15.8%) in the PCI group and nine patients (2.9%) in the CABG group), of whom 89 patients underwent PCI and eight patients CABG for TVR. TVR was driven by restenosis at the left main bifurcation in 29 patients (29.9%) and progression of distal diseases other than left main stenosis in 68 patients (70.1%). In addition, we analysed the clinical outcomes according to disease extent such as left main only, left main with one vessel disease, left main with two vessel disease and left main with three vessel disease. The more coronary arteries are diseased, the more often death or composite of death, Q-wave MI, or stroke occurred in both the PCI and CABG groups (see supplementary table 1, available online only). Notably, whereas the rate of TVR was significantly higher as the disease extent was greater in the PCI group, this were not evident in the CABG group. Observed (unadjusted) event-free survival and crude RR according to the treatment approach for the overall cohort are presented in figure 1 and table 3. There were no significant differences in the long-term outcomes of death and the composite of death, Q-wave MI, or stroke between the PCI group and the CABG group, whereas the rate of TVR was significantly higher in the PCI group. After thorough adjustment of baseline covariates with multivariate Cox proportional hazards regression model and IPTW, the long-term risks of death and the composite of death, Q-wave MI, or stroke were also similar between the PCI and CABG groups (table 3). The adjusted risk of TVR was still significantly higher in the PCI group than the CABG group. In the propensity score-matched cohorts, the PCI group also had similar cumulative event rates of death and the composite of death, Q-wave MI, or stroke compared with the CABG group, but the long-term cumulative rate of TVR was consistently higher in the PCI group than the CABG group (table 4). Formal testing for interactions showed that the long-term rates of death and the composite of death, Q-wave MI, or stroke between PCI and CABG were consistent in multiple subgroups such as diabetes (HR 1.41, 95% CI 0.81 to 2.44, p=0.226) and acute coronary syndrome (HR 1.25, 95% CI 0.82 to 1.91, p=0.291).

Figure 1

Kaplan–Meier curves for outcome in patients who underwent percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG). (A) Overall mortality; (B) composite of death, Q-wave myocardial infarction (MI), or stroke; (C) target vessel revascularisation (TVR).

Table 3

HR for clinical outcomes in patients undergoing PCI compared with CABG and in patients undergoing complex stenting compared with simple stenting

Table 4

HR for clinical outcomes in patients performing PCI compared with CABG and in patients receiving complex stenting compared with simple stenting among propensity matched patients

Clinical outcomes between the simple stenting group and the complex stenting group

Observed event-free survival and crude RR according to the stenting techniques in the PCI group are presented in figure 2 and table 3. There were no significant differences in long-term outcomes of death and composite of death, Q-wave MI, or stroke between the two stenting groups, whereas the rate of TVR was significantly higher in the complex stenting group. After thorough adjustment of baseline covariates with multivariate Cox proportional hazards regression model and IPTW, the long-term risks of death and composite of death, Q-wave MI, or stroke were also similar between the two stenting groups (table 3). The adjusted risk of TVR was still significantly higher in the complex stenting group than in the simple stenting group. In the propensity score-matched cohorts, the overall findings were consistent (table 4).

Figure 2

Kaplan–Meier curves for outcome in patients who received complex or simple stenting. (A) Overall mortality; (B) composite of death, Q-wave myocardial infarction (MI), or stroke; (C) target vessel revascularisation (TVR).

Clinical outcomes between the simple stenting group and the CABG group

The baseline characteristics of the simple stenting group and the CABG group are presented in supplementary table 2 (available online only). After propensity score matching, there was no longer any significant difference for any covariate in the matched cohorts. There were no significant differences in the long-term outcomes of death and the composite of death, Q-wave MI, or stroke between the simple stenting group and the CABG group, whereas the rate of TVR was significantly higher in the simple stenting group. These findings were consistent after adjusting with multivariate Cox proportional hazards regression model and IPTW, and in the propensity score-matched cohorts (see supplementary tables 3 and 4, available online only).

Discussion

In this large, multicentre cohort analysis evaluating clinical outcomes during a median follow-up of 4.2 years according to the treatment modality for unprotected left main bifurcation disease, we found that there was no significant difference in the risk of death and composite of death, Q-wave MI, or stroke between the PCI and CABG groups. In contrast, the cumulative rate of TVR was significantly higher in the PCI group than in the CABG group. These findings suggest that PCI using DES provides comparable long-term clinical outcomes as CABG for unprotected left main bifurcation disease, specifically in terms of safety outcomes such as death and composite of death, Q-wave MI, or stroke, but not in terms of efficacy outcomes such as TVR. In the PCI-treated cohort, the simple stenting group had a significantly lower rate of TVR than the complex stenting group, while providing similar long-term risks of death and composite of death, Q-wave MI, or stroke.

Compared with left main ostial and shaft disease, left main bifurcation disease remains challenging for PCI even after the liberal use of DES. Procedural risk, technical complexity, higher revascularisation rate and concern for the potential risk of stent thrombosis are factors to consider when deciding the treatment modality for left main bifurcation disease.13–15 In a large observational study involving 1111 patients with unprotected left main disease treated with DES, patients with left main bifurcation disease had a significantly increased cumulative rate of composite of death, MI and TVR compared with those with left main ostial and shaft disease.16 In addition, the rate of target-lesion revascularisation (TLR) during 3 years was higher in patients with left main bifurcation stenosis compared with left main ostial and shaft stenosis from the J-Cypher registry.13 However, the paucity of concrete evidence of long-term (at least 5 years) safety and efficacy of PCI from large-scale registries or randomised trials for left main bifurcation disease may be the main factor deterring the robust introduction of PCI for left main bifurcation disease into clinical practice.2 In this regard, this study provides invaluable long-term clinical outcome data for unprotected left main bifurcation disease, along with the recently reported PRECOMBAT trial in which 383 patients with unprotected left main bifurcation disease were enrolled.9 While the current study has the intrinsic weakness of a non-randomised registry design compared with the prospective, randomised PRECOMBAT clinical trial, it offers some strengths: (1) We recruited only patients with true left main bifurcation disease from 18 cardiac centres throughout the country; we did not include patients with left main ostial or shaft disease and those with distal left main stenosis and a stenosis less than 50% involving ostia of LAD and/or LCx arteries. (2) We successfully achieved a longer-term follow-up in nearly all patients up to 5.2 years (median 4.2 years). (3) Because this registry was planned before the PRECOMBAT trial, there were no overlapping patients and more centres participated in this registry. (4) Our results remained valid even after rigorous adjustment using IPTW and multivariable Cox proportional hazards regression analysis and in propensity matched cohorts, which underscores the comparability of populations and outcome differences being mainly driven by treatment option.

For the treatment of unprotected left main bifurcation disease with PCI in the DES era, the optimal strategy of stenting still remains a major issue, specifically the superiority of simple stenting techniques to complex stenting techniques. The J-Cypher registry analysing 380 patients with unprotected left main bifurcation disease reported that patients who underwent complex stenting had significantly higher rates of cardiac death and TLR than those who underwent simple stenting during 2.6 years of follow-up.13 However, another study involving 777 Italian patients with left main bifurcation disease found no significant difference in the cumulative event rates of death and MI, but significantly higher rate of TLR in the complex stenting group during 1.5 years follow-up.16 By observing the clinical outcomes of 556 DES-treated patients with unprotected left main bifurcation disease during 4.2 years of follow-up, we also found no significant differences in the long-term clinical outcomes except TVR in the complex stenting group.

This study has several limitations and strengths. First, because this is a non-randomised observational study, unmeasured confounders and procedural bias might have affected the results. However, we performed rigorous adjustments of our data to reduce the impact of treatment selection bias and potential confounding using IPTW, propensity score matching and multivariable Cox proportional hazards regression analysis. Participating centres showed similar selection patterns of treatment modality and similar clinical outcomes according to the selection of treatment modality. Furthermore, this study drew a conclusion from a relatively large sample size of patients with unprotected left main bifurcation disease and a longer follow-up duration. Because of a concern for late catch-up phenomena in DES-treated lesions,17 our study is valuable in providing extended clinical observational outcomes in this highest risk subset of patients undergoing PCI.2 Second, because we did not integrate the angiographic characteristics of left main bifurcation lesions into our database, the relation between lesion anatomy of the left main bifurcation and treatment selection between PCI versus CABG or between simple stenting versus complex stenting techniques could not be analysed. However, we tried to constitute a homogenous cohort of left main bifurcation disease by including only patients with greater than 50% stenosis involving both the distal left main and at least one of ostia of LAD or LCx arteries. Furthermore, we collected data on the EuroSCORE and SYNTAX score in all patients. In the present study, by adjusting all the variables of the EuroSCORE and SYNTAX score with the use of IPTW and multivariable Cox proportional hazards regression analysis, we tried to integrate clinical and anatomical factors having a significant impact on the selection of treatment modalities and long-term clinical outcomes into our cohort of unprotected left main bifurcation disease.

Conclusion

As the first report on a comparison of long-term clinical outcomes associated with PCI using DES and CABG in patients with unprotected left main bifurcation disease, we found that PCI yielded favourable safety outcomes of death and the composite of death, Q-wave MI or stroke compared with CABG during 4.2 years follow-up; however, it failed to improve the efficacy outcome of TVR. Complex stenting seems to be associated with similar long-term safety outcomes compared with simple stenting but with a significantly increased rate of TVR in this setting.

Reference

View Abstract

Footnotes

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethics approval was provided by the institutional review board at each hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.