Author + information
- Received August 3, 2011
- Revision received September 26, 2011
- Accepted October 11, 2011
- Published online February 14, 2012.
- J. Matthew Brennan, MD, MPH⁎,⁎ (, )
- David Dai, PhD⁎,
- Manesh R. Patel, MD⁎,
- Sunil V. Rao, MD⁎,
- Ehrin J. Armstrong, MD, MSc†,§,
- John C. Messenger, MD∥,
- Jeptha P. Curtis, MD¶,
- Kendrick A. Shunk, MD, PhD§,‡,
- Kevin J. Anstrom, PhD⁎,
- Eric L. Eisenstein, DBA⁎,
- William S. Weintraub, MD#,
- Eric D. Peterson, MD, MPH⁎,
- Pamela S. Douglas, MD⁎ and
- William B. Hillegass, MD, MPH⁎⁎
- ↵⁎Reprint requests and correspondence:
Dr. J. Matthew Brennan, Duke Clinical Research Institute, Duke University Medical Center, 2400 Pratt Street, Suite 800, Durham, North Carolina 27705
Objectives This study sought to assess percutaneous coronary intervention (PCI) for unprotected left main coronary artery (ULMCA) stenosis in routine U.S. clinical practice.
Background Percutaneous coronary intervention for ULMCA stenosis is controversial; however, current use and outcomes of ULMCA PCI in routine U.S. clinical practice have not been described.
Methods We evaluated 5,627 patients undergoing ULMCA PCI at 693 centers within the National Cardiovascular Data Registry Catheterization Percutaneous Coronary Intervention Registry for temporal trends in PCI use (2004 to 2008), patient characteristics, and in-hospital mortality. Thirty-month mortality and composite major adverse events (death, myocardial infarction, and revascularization) with drug-eluting versus bare-metal stents were compared using inverse probability weighted (IPW) hazard ratios (HRs) in a nonrandomized Medicare-linked (age ≥65 years) patient cohort (n = 2,765).
Results ULMCA PCI was performed in 4.3% of patients with ULMCA stenosis. Unadjusted in-hospital mortality rates ranged from 2.9% for elective cases to 45.1% for emergent/salvage cases. By 30 months, 57.9% of the elderly ULMCA PCI population experienced death, myocardial infarction, or revascularization, and 42.7% died. Patients receiving drug-eluting stents (versus bare-metal stents) had a lower 30-month mortality (IPW HR: 0.84, 95% confidence interval [CI]: 0.73 to 0.96), but the composite of major adverse events were similar (IPW HR: 0.95, 95% CI: 0.84 to 1.06).
Conclusions In the United States, ULMCA PCI is performed in <5% of patients with ULMCA disease and is generally reserved for those at high procedural risk. Adverse events are common in elderly patients and are related to patient and procedural characteristics, including stent type.
In the United States, revascularization of unprotected left main coronary artery (ULMCA) stenosis has traditionally been achieved with cardiac surgery. However, American College of Cardiology/American Heart Association coronary revascularization guidelines have upgraded ULMCA percutaneous coronary intervention (PCI) in coronary artery bypass grafting–eligible patients from a class III (i.e., “should not be performed”) to a Class IIb recommendation (i.e., “may be considered”) (1) based on recent evidence (2).
Against this background, we examined trends in ULMCA PCI and long-term clinical outcomes in U.S. practice using a Medicare-linked cohort from the NCDR (National Cardiovascular Data Registry) CathPCI (Catheterization Percutaneous Coronary Intervention) registry.
Analyses included all patients without a prior history of coronary artery bypass graft surgery but with a significant ULMCA stenosis at any of the 693 centers with CathPCI registry participation between January 1, 2004, and December 31, 2008 (Fig. 1). A subgroup of elderly patients with linkage to Medicare inpatient claims was identified for longitudinal follow-up.
To enable 30-month follow-up, registry files were linked to Medicare 100% inpatient fee-for-service claims, as previously described (3,4). The Duke University School of Medicine Institutional Review Board granted a waiver of informed consent and authorization for this study.
Primary endpoints included death and a composite of major adverse events (i.e., death, myocardial infarction, and repeat revascularization) and were assessed using Medicare inpatient claims, as previously described (3).
Baseline patient and procedural characteristics were summarized as counts and percentages for categorical variables and mean with standard deviations for continuous variables. The association of the ULMCA PCI rate with the year of index hospitalization was evaluated using a chi-square test of trend. Statistical significance was defined as a p value ≤0.05. SAS statistical software (version 9.1, SAS Institute, Cary, North Carolina) was used for all calculations.
Propensity scores (including 107 covariates) (3) were used to balance drug-eluting stents (DES) versus bare-metal stents (BMS) treatment groups (5), and inverse probability weighted estimators incorporating the propensity scores were used to balance treatment groups (6).
For each of the primary endpoints and within each of the pre-specified subgroups, cumulative incidence rates at 1, 12, and 30 months were calculated using inverse probability weighted estimators (7), and adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated based on sandwich-estimated standard errors (8).
High-urgency revascularization procedures included primary PCI, rescue or facilitated PCI for ST-segment elevation myocardial infarction, or those involving pre-procedural cardiogenic shock, or emergent or salvage status. Logistic EuroSCORE (European System for Cardiac Operative Risk Evaluation) values were calculated using the variable mapping algorithm in Online Table 1⇓ and classified as high (≥33.5%) or low (<33.5%) based on prior literature (9). Intravascular ultrasound use was specific to the left main lesion, but its timing (i.e., pre-PCI, post-PCI, or both) was unspecified.
A post hoc “cause of death” analysis was performed among patients surviving to hospital discharge. Additionally, 2 sensitivity analyses were performed to evaluate the potential influence of unmeasured confounders and treatment selection bias (Online Table 2).
From January 2004 to December 2008, 131,004 patients having ≥50% ULMCA stenosis were included in the CathPCI registry (Fig. 1), with 4.3% receiving ULMCA PCI. Compared with patients not treated with PCI, stented patients were older with a higher prevalence of both chronic comorbidities and acute pre-procedural clinical instability (Table 1).
Of the 5,627 patients who underwent PCI for ULMCA stenosis, 4,085 (72.6%) were ≥65 years of age, including 2,765 (69%) with successful linkage to Medicare inpatient claims files (Fig. 1, Online Table 1). Medicare-linked patients were largely representative of the overall Medicare-eligible cohort.
Temporal trends of ULMCA PCI in the NCDR CathPCI registry
Use of ULMCA PCI was rare at most centers (660 centers with <30 total procedures [“low” volume, <6 procedures per year]; 25 centers with 30 to 75 total procedures [“moderate” volume, 6 to 15 procedures per year]; and 8 centers with >75 ULMCA procedures over 5 years [“high” volume, >15 procedures per year]). The proportion of patients with ULMCA disease treated with ULMCA PCI at centers participating in the CathPCI registry increased slightly throughout this interval (3.8% to 4.9%, p < 0.0001) (Fig. 2A), and most ULMCA PCI cases occurred in lower-urgency settings (Fig. 2B). DES were used in a majority (81.8%) of lower-urgency procedures; however, DES and BMS use was similar in high-urgency procedures after 2006.
Overall ULMCA NCDR CathPCI cohort: in-hospital outcomes
Overall, 95% of ULMCA patients survived to hospital discharge. Stented ULMCA patients had a higher unadjusted in-hospital mortality rate compared with nonstented patients (13.1% vs. 4.6%, p < 0.0001). In-hospital mortality was associated with clinical urgency, with a greater in-hospital mortality observed among patients with high (45.1%) versus low (2.9%) clinical urgency and high (26.0%) versus low (3.2%) pre-procedural logistic EuroSCORE.
Medicare-linked ULMCA PCI cohort: longitudinal outcomes
Of the 2,765 patients with longitudinal follow-up, 57.9% experienced a major adverse event by 30 months (death: 42.7%; myocardial infarction: 8.2%; repeat revascularization: 17.5%).
By 30 months, a similar unadjusted incidence of death was observed across low-, moderate-, and high-volume centers for both lower-urgency (35.4% vs. 36.4% vs. 33.9%, p = 0.2) and higher-urgency ULMCA PCI procedures (65.1% vs. 73.2% vs. 59.9%, p = 0.1). Composite events for high-urgency patients were most common among centers with the highest ULMCA PCI volume (low volume: 72.7%; moderate volume: 76.9%; high volume: 80.0%; p = 0.7)—a trend that reached significance among lower-urgency patients (50.1%, 52.5%, and 60.3%, respectively; p = 0.0005).
The 30-month outcomes stratified by clinical features are presented in Figure 3. Patients with the highest unadjusted incidence of major adverse events by 30 months included those with high logistic EuroSCORE values (76.6% vs. 53.9% for low logistic EuroSCORE values, p < 0.001), high clinical urgency (74.2% vs. 52.8% for low urgency, p < 0.001) (Fig. 4), and bifurcation lesions (64.8% vs. 54.7% for ostial or mid-body, p < 0.001) (Fig. 5).
The relative hazard of death by 30 months was lower among DES-treated versus BMS-treated patients both before and after adjustment by propensity weighting (39.6% vs. 52.7%; unadjusted HR: 0.60, 95% CI: 0.52 to 0.68; adjusted HR: 0.84, 95% CI: 0.73 to 0.96). However, no risk-adjusted difference in composite events was observed (adjusted HR: 0.95, 95% CI: 0.84 to 1.06) (Fig. 6). The relative risk of death and composite major adverse events associated with the use of DES versus BMS was consistent across most of the subgroups examined (Figs. 3A and 3B).
This analysis represents the first national report of outcomes with ULMCA PCI in the United States and the largest reported experience of ULMCA PCI to date. It highlights 3 main findings. First, as of 2008, percutaneous treatment of ULMCA disease in the United States is uncommon and generally reserved for those with either a high burden of noncardiac disease or emergent clinical presentations. Second, elderly patients receiving ULMCA PCI are at high risk for major adverse events following the procedure, with 40% dying within the first 3 years of follow-up. Finally, the risk of adverse events appears related to certain procedural characteristics, including stent choice.
Temporal trends in ULMCA PCI
In the United States, the proportion of ULMCA patients treated with percutaneous revascularization remains low, but it is slowly increasing. Through 2008, most centers performed <1 ULMCA PCI per month, although an increase in the volume of lower-urgency procedures suggests increasing general acceptance of elective PCI for ULMCA at some centers. Consistent with international trends (10), DES remains the predominant stent platform for lower urgency procedures in the United States.
Outcomes following PCI for ULMCA stenosis
In the United States, ULMCA PCI is performed primarily in patients with a high burden of comorbidities and frequently in those with high-urgency clinical presentations. In this population, in-hospital mortality was substantially higher in those with ULMCA disease who were treated percutaneously; however, this early risk is driven by those with high-urgency, rather than lower-urgency, presentations.
Among ULMCA patients selected for PCI in the United States, 30-month mortality is high (42.7%), with a predominance of cardiovascular events. These findings are especially notable when compared to the 10% mortality reported at 30 months in the MAIN-COMPARE (Korean Revascularization for Unprotected Left Main Coronary Artery Stenosis: Comparison of Percutaneous Coronary Angioplasty Versus Surgical Revascularization) study (11), and the 4% mortality reported at 1 year in the SYNTAX (Synergy Between PCI With Taxus and Cardiac Surgery) trial (12). The increased patient age, burden of comorbidities, and clinical acuity of the NCDR ULMCA PCI cohort likely account for some portion of the difference in outcomes, and the outcomes reported here are more consistent with both early international reporting of ULMCA PCI outcomes in nonoperable patients (26.5% event-free survival at 9 months) (13) and those reported in octogenarians undergoing ULMCA PCI within the Canadian health system (56.7% event-free survival at 2 years) (14). These data highlight the need for continued surveillance of ULMCA PCI in the United States.
In addition to identifying several patient subgroups at increased risk of adverse outcomes (i.e., patients with bifurcation lesions), we have also identified important trends in procedural technique. For example, single-stent techniques, which have recently been associated with improved outcomes (15), were less common in this ULMCA PCI population (37.5%), but they were associated with improved long-term outcomes as compared with 2-stent techniques. Also, the use of intravascular ultrasound for procedural guidance was much less common (18%) than reported elsewhere (e.g., 78% in the MAIN-COMPARE registry) (16). In this cohort, intravascular ultrasound use was associated with a lower unadjusted incidence of both major adverse events (51% vs. 59%) and death (35% vs. 44%), adding to the growing body of nonrandomized literature supporting its use during ULMCA stenting (17).
Finally, our results are consistent with prior reports suggesting that DES are superior to BMS for the treatment of left main lesions (18). Here, DES were associated with improved survival when compared with BMS, but the mechanism by which stent type affects the risk of death in this population remains unclear.
These data represent the largest-ever experience with percutaneous revascularization of ULMCA disease, and the first-ever report of the U.S. national experience with these procedures; however, several limitations are notable. First, the reporting of diagnostic-only procedures is voluntary in most health systems and is believed to be less reliable at some centers. Second, the NCDR data does not allow a direct comparison of outcomes with percutaneous, surgical, and medical treatment strategies. Third, the longitudinal cohort is drawn from a Medicare FFS subgroup of the overall population in the United States; however, this cohort is highly representative of elderly patients undergoing percutaneous ULMCA revascularization in the United States. Fourth, outcomes are assessed using inpatient claims data and, therefore, are subject to inherent reporting errors and detection biases. For example, the exclusion of outpatient “observational” hospital visits potentially resulted in an underestimation of the incidence of repeat revascularization. Fifth, although 66% of the deaths with an attributable “cause” occurred in-hospital, the use of Medicare claims to ascertain the true “cause of death” is imprecise and should be viewed simply as hypothesis generating. Finally, the outcome comparison for patients treated with DES versus BMS is subject to unmeasured biases inherent to the device selection, which may favor DES (3). A comparison to clinical trial data is not possible in this setting, as no prior clinical trials have randomized ULMCA patients to DES versus BMS.
Implications for future research
These results emphasize the need for a well-powered randomized clinical trial examining the safety and effectiveness of percutaneous revascularization in patients with ULMCA disease, with attention to the generalizability of U.S. patients recruited for trial participation and best practices in procedural techniques for revascularization.
ULMCA PCI remains a relatively uncommon procedure at most U.S. centers and is primarily reserved for those at high risk for coronary artery bypass graft surgery. Poor outcomes following percutaneous revascularization in elderly ULMCA patients are common and are likely influenced by both patient and procedural characteristics.
The authors would like to thank Michael Booth, Rachel Dokholyan, and Erin LoFrese for their contributions to this paper. Mr. Booth, Ms. Dokholyan, and Ms. LoFrese did not receive compensation for their assistance, apart from their employment at the institution where the study was conducted.
For supplemental tables, please see the online version of this paper.
This project was sponsored by the Agency for Healthcare Research and Quality, United States Department of Health and Human Services, Rockville, Maryland, as part of the Cardiovascular Consortium and funded under Project ID 24-DKE-3 and Work Assignment Number HHSAA290-2005-0032—TO4-WA2 as part of the Developing Evidence to Inform Decisions about Effectiveness (DEcIDE) program. Dr. Patel is a consultant for Bayer Healthcare, Ortho-McNeil Jansen, and Genzyme. Dr. Rao has received research funding from Cordis Corporation. Dr. Messenger is the site principal investigator for the Resolute and EDUCATE studies; and has received funding from Medtronic, Inc. Dr. Curtis receives salary support from the ACC-NCDR; and has equity holding in Medtronic. Dr. Shunk has received research funding from Abbott Vascular and Siemens Medical Systems and consulting fees from SurgiVision and Revascular Therapeutics. Dr. Anstrom has received research funding and salary support from Alexion, AstraZeneca, Bristol-Myers Squibb, Eli Lilly, Innocoll Pharmaceuticals, Medtronic, Pfizer, and Proctor & Gamble; serves on the data safety monitoring boards of Pfizer and Vertex; and receives consulting fees from Abbott Vascular, Ikaria, Pacific Therapeutics, Bristol-Myers Squibb, and AstraZeneca. Mr. Eisenstein has received research funding from Medtronic and Eli Lilly. Dr. Peterson has received research funding from Bristol-Myers Squibb/Sanofi-Aventis, Merck, Lilly, Johnson & Johnson, Society of Thoracic Surgeons, and American College of Cardiology. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or by the United States Department of Health and Human Services. Additional support was obtained from the National Cardiovascular Data Registry, American College of Cardiology, Washington, DC.
- Abbreviations and Acronyms
- bare-metal stent(s)
- confidence interval
- drug-eluting stent(s)
- hazard ratio
- percutaneous coronary intervention
- unprotected left main coronary artery
- Received August 3, 2011.
- Revision received September 26, 2011.
- Accepted October 11, 2011.
- American College of Cardiology Foundation
- King S.B. 3rd.,
- Smith S.C. Jr..,
- Hirshfeld J.W. Jr..,
- et al.
- Douglas P.S.,
- Brennan J.M.,
- Anstrom K.J.,
- et al.
- Rosenbaum P.R.,
- Rubin D.B.
- Dewey T.M.,
- Brown D.,
- Ryan W.H.,
- Herbert M.A.,
- Prince S.L.,
- Mack M.J.
- Price M.J.,
- Kandzari D.E.,
- Teirstein P.S.
- Park D.W.,
- Seung K.B.,
- Kim Y.H.,
- et al.
- Morice M.C.,
- Serruys P.W.,
- Kappetein A.P.,
- et al.
- Ellis S.G.,
- Tamai H.,
- Nobuyoshi M.,
- et al.
- Rodés-Cabau J.,
- DeBlois J.,
- Bertrand O.F.,
- et al.
- Palmerini T.,
- Marzocchi A.,
- Tamburino C.,
- et al.
- Banning A.P.,
- Westaby S.,
- Morice M.C.,
- et al.
- MAIN-COMPARE Investigators,
- Park S.J.,
- Kim Y.H.,
- Park D.W.,
- et al.
- MAIN-COMPARE Investigators,
- Kim Y.H.,
- Park D.W.,
- Lee S.W.,
- et al.