Comparison of Percutaneous Versus Surgical Revascularization of Severe Unprotected Left Main Coronary Stenosis in Matched Patients

Comparison of Percutaneous Versus Surgical Revascularization of Severe Unprotected Left Main Coronary Stenosis in Matched Patients Sorin J. Brener, MDa,*, John M. Galla, MDa, Roosevelt Bryant III, MDb, Joseph F. Sabik III, MDb, and Stephen G. Ellis, MDa Coronary artery bypass grafting (CABG) has been the recommended treatment for patients with significant left main coronary artery (LMCA) stenosis. Advances in stent technology have invigorated investigations into the suitability of a percutaneous approach for these patients. Favorable short-term results from nonrandomized comparisons were previously reported. Patients (n ⴝ 97) who underwent percutaneous coronary intervention for severe (>70%) LMCA stenosis were matched in a 1:2 ratio with a cohort that underwent surgical revascularization (n ⴝ 190). The groups were similar for age, gender, European System for Cardiac Operative Risk Evaluation, left ventricular ejection fraction, history of myocardial infarction, and presence of renal disease. Kaplan-Meier estimates of 3-year mortality were similar for the PCI and CABG groups at 80% (95% confidence interval [CI] 68 to 88) versus 85% (95% CI 79 to 89, p ⴝ 0.14), respectively. Propensity score–adjusted 3-year mortality did not differ between groups (p ⴝ 0.22). Multivariable modeling identified only higher European System for Cardiac Operative Risk Evaluation (hazard rate 1.33, 95% CI 1.16 to 1.54, p <0.001) and the presence of diabetes mellitus (hazard rate 1.96, 95% CI 1.24 to 3.09, p ⴝ 0.004) as independent risks of mortality at 3 years. In conclusion, patients who underwent percutaneous revascularization of severe LMCA stenosis appeared to have 3-year survival equivalent to those who underwent CABG. Diabetes mellitus and advanced co-morbidity were the principal determinants of survival. These findings support the need for randomized trials with adequate follow-up to compare the 2 approaches. © 2008 Elsevier Inc. All rights reserved. (Am J Cardiol 2008;101:169 –172)

Several recent reports detailed immediate and short-term results of several single-center and registry cohorts with severe left main coronary artery (LMCA) stenosis that underwent percutaneous coronary intervention (PCI) with bare-metal or drug-eluting stents (DESs) or coronary artery bypass grafting (CABG).1– 4 We present long-term outcomes of 97 patients who underwent LMCA PCI matched to 190 patients who underwent CABG at our institution. Methods All patients who underwent LMCA stent placement by a single interventionalist (SJB) from 1997 to 2006 were analyzed, excluding those treated for acute myocardial infarction or acute coronary syndrome. Most were referred for this procedure because of multiple co-morbidities or patient or physician preference. A comparison cohort was selected from the institutional cardiovascular surgery database. The PCI and CABG registries are prospective repositories for data and have received waiver for clinical research from the institutional review board. Patients who underwent PCI were matched in a 1:2 ratio to patients who underwent Departments of aCardiovascular Medicine and bCardiothoracic Surgery, Cleveland Clinic, Cleveland, Ohio. Manuscript received June 13, 2007; revised manuscript received and accepted August 7, 2007. *Corresponding author: Tel: 216-444-0732; Fax: 216-444-8050. E-mail address: [email protected] (S.J. Brener). 0002-9149/08/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2007.08.054

isolated CABG (no concomitant valvular or aortic surgery) in the same period. Matching characteristics included, in decreasing order of importance, age, left ventricular ejection fraction, history of myocardial infarction, presence of diabetes mellitus, and procedure date. The primary outcome was all-cause 3-year mortality, assessed using the validated Social Security Death Index Registry5 and follow-up obtained in each registry. Because the 2 registries do not systematically collect information for subsequent myocardial infarction and revascularization or concomitant medical therapy, these end points were not reported. Until December 31, 2005, patients who underwent PCI were invited for routine surveillance angiography 3 to 6 months after the procedure. Regardless of treatment choice, the goal was to achieve as complete revascularization as possible. Vascular access was primarily transfemoral. When LMCA stenosis affected the bifurcation and the left circumflex artery was ⬎2 mm in diameter, simultaneous kissing stent deployment back to the origin of the LMCA was performed. If the simultaneous kissing stent technique was not used, final kissing balloon inflation was performed in the 2 branches. Stent choice and use of adjunctive pharmacologic or interventional therapies were at the discretion of the operator. All patients received aspirin and a loading dose ⱖ300 mg of clopidogrel. Aspirin and clopidogrel therapy were recommended indefinitely. CABG was carried out using standard techniques and preferentially used the internal thoracic artery for revascuwww.AJConline.org

170

The American Journal of Cardiology (www.AJConline.org)

larization of the left anterior descending artery. Cardiopulmonary bypass was used in all patients undergoing CABG. The European System for Cardiac Operative Risk Evaluation (EuroSCORE) was previously validated to predict short-term mortality and used to identify patients at high risk (EuroSCORE 6 to 12).6 Relevant factors that contributed to the EuroSCORE calculation included age, gender, ejection fraction, chronic renal insufficiency, peripheral vascular disease, and chronic pulmonary disease. Continuous variables were presented as mean ⫾ SD and were compared using Student’s t test or Mann-Whitney U test. Chi-square test or Fisher’s exact test was used to analyze differences in categorical variables. For each patient, a propensity score for undergoing PCI (0) or CABG (1) was derived using multivariable logistic regression analysis.7,8 The model included age, diabetes mellitus, previous myocardial infarction, heart failure class, EuroSCORE (which already included baseline characteristics and coronary heart disease risk factors), and number of diseased vessels. A univariate survival comparison was conducted using the Kaplan-Meier method and log-rank test. To characterize predictors of mortality, a multivariable Cox proportional hazards model including age, diabetes mellitus, EuroSCORE, revascularization strategy, and propensity score was created.9 All tests were 2 tailed and used ␣ ⫽ 0.05 to characterize statistical significance. Statistical analyses were performed using STATA SE, version 9.2 (StataCorp, College Station, Texas). Results Baseline characteristics of each patient group are listed in Table 1. Ninety-seven patients underwent PCI and were matched to 190 clinically similar patients who underwent revascularization using CABG. In the PCI group, distal LMCA involvement was present in 71 patients (72%) and 46 (65%) had a simultaneous kissing stent. Balloon angioplasty alone was performed in 5 patients, 37 patients received bare-metal stents, and the remaining 55 patients received a DES (30 sirolimus-eluting stents, 25 paclitaxel-eluting stents). Twelve patients were treated with rotational atherectomy debulking (Rotablator; Boston Scientific, Natick, Massachusetts) before stent placement because of severe calcification. All patients who underwent CABG except for 1 received an internal mammary artery graft. There were no statistically significant differences between groups with respect to the matched characteristics of age, history of myocardial infarction, or left ventricular ejection fraction. Diabetes mellitus was more prevalent in the PCI group. Patients in the CABG group had slightly more extensive coronary artery disease (2.84 ⫾ 0.36 vs 2.75 ⫾ 0.52 diseased vessels; p ⫽ 0.09) than patients treated using PCI. In the PCI cohort, 2.7 ⫾ 1.5 lesions were treated per patient, whereas patients in the CABG group received 3.3 ⫾ 1.0 grafts (distal anastomoses; p ⫽ 0.001). Average propensity score in the PCI group was 0.47 ⫾ 0.25 compared with 0.74 ⫾ 0.18 in the CABG group (p ⬍0.001). The c statistic for the propensity score was 0.80, indicating good discriminatory ability. Three patients (3%) in the PCI cohort died during their index hospitalization. There were no intraprocedural deaths.

Table 1 Baseline characteristics of the study population Variable

PCI (n ⫽ 97)

CABG (n ⫽ 190)

p Value

Age (yrs) Men Diabetes mellitus Hypertension Chronic renal insufficiency Previous myocardial infarction Current smoker Ejection fraction (%) Peripheral vascular disease EuroSCORE EuroSCORE ⱖ6 Follow-up (mo)

68 ⫾ 11 72 (72%) 42 (42%) 82 (82%) 7 (7%) 56 (56%) 11 (11%) 45 ⫾ 14 25 (25%) 4.6 ⫾ 2.2 37 (37%) 34 ⫾ 28

68 ⫾ 10 141 (74%) 25 (25%)* 141 (74%) 15 (8%) 112 (59%) 101 (57%) 44 ⫾ 11 70 (37%) 4.5 ⫾ 2.4 61 (32%) 122 ⫾ 106

0.44 0.64 0.004 0.12 0.81 0.54 ⬍0.001 0.58 ⬍0.001 0.62 0.30 ⬍0.001

Values expressed as mean ⫾ SD or number (percent). * Diabetes status was unknown in 91 patients who underwent CABG.

Eight patients (4%) in the matched surgical cohort died in the hospital (1 intraoperative death; p ⫽ 0.60). By 3 years, estimated survival rates in the PCI and CABG groups were 80% (95% confidence interval [CI] 68 to 88) and 85% (95% CI 79 to 89), respectively (p ⫽ 0.14; Figure 1). Propensity score–adjusted survival at 3 years confirmed these findings (p ⫽ 0.22; Figure 2). The only independent predictors of mortality at 3 years were higher EuroSCORE (hazard rate [HR] 1.33, 95% CI 1.16 to 1.54, p ⬍0.001) and diabetes mellitus (HR 1.96, 95% CI 1.24 to 3.09, p ⫽ 0.004). Age (p ⫽ 0.38), propensity score (p ⫽ 0.29), and PCI versus CABG (p ⫽ 0.27) were not significantly associated with mortality. In high-risk patients (EuroSCORE ⱖ6), only diabetes mellitus was independently predictive of mortality (HR 2.30, 95% CI 1.09 to 4.86, p ⫽ 0.03), whereas PCI versus CABG was not (p ⫽ 0.62). Of 64 patients who returned for surveillance angiography after LMCA PCI, only 2 needed repeated PCI of the LMCA and 1 proceeded to CABG. Discussion The principal finding of this analysis was that 3-year survival of selected patients who underwent PCI for severe LMCA stenosis was not inferior to that of similar patients who underwent CABG, even after adjusting for the propensity to undergo either type of revascularization. These results challenged the current guidelines for LMCA PCI in patients with stable coronary artery disease, which assign it a class IIa (level B) recommendation for patients not eligible for CABG and a class III (level C) for those eligible for CABG.10 One third of the cohort had very high risk of CABG measured using the EuroSCORE. Despite more complete revascularization in the CABG group (judged by the number of anastomoses compared with the number of treated lesions), only severity of co-morbid illnesses before revascularization and, in particular, diabetes mellitus, affected survival. It was notable that despite a careful attempt to match patients, the rate of diabetes was significantly higher in the PCI group. This observation was in contrast to recommendations of the Bypass or Angioplasty Revascularization Investigation (BARI) study,11 but agreed with the

Coronary Artery Disease/PCI or CABG for LMCA Stenosis

171

0.75

1.00

Survival in the PCI and CABG groups

0.00

0.25

0.50

P=0.14

0

10

20 Follow-up (months)

PCI

PCI (N) 97 CABG (N) 190

71 174

30

40

CABG

51 162

35 149

5

Figure 1. Unadjusted survival of patients who underwent PCI (blue line) and CABG (red line) at 3 years (p ⫽ 0.14).

-ln[-ln(Survival Probability)] 1 2 3 4

CABG

PCI

0

P=0.22

0

2

4

6

ln(analysis time)

Figure 2. Propensity score–adjusted survival of patients who underwent PCI (blue line) and CABG (red line) at 3 years (p ⫽ 0.22).

lack of effect BARI had on clinical practice after its publication.12 This study corroborated previous studies that showed similar rates of major adverse cardiac events at 6 to 15 months for patients who underwent PCI or CABG for severe LMCA stenosis.1–3,13 Only Chieffo et al1 used propensity analysis to correct for the inevitable selection bias of the revascularization techniques, but limited their follow-up to 1 year. The need for repeated revascularization after CABG or PCI has been the driving force behind differences in major adverse cardiac events in many comparisons of the 2 approaches to atherosclerotic heart disease.14 Compared with balloon angioplasty, CABG decreased the need for repeated revascularization by 90%,15 but this difference narrowed considerably with stenting. In the Arterial Revascularization Therapies Study (ARTS), rates of repeated revascularization were 3.5% for patients who underwent CABG and

16.8% for patients who underwent PCI.16 In the ARTS 2 study, rates of freedom from death, infarction, or revascularization were 93.6% for patients who underwent DES PCI, 91% for patients who underwent CABG at 1 year, and 81% and 84% at 3 years, respectively (P. Serruys, personal communication, Annual Meeting of the American College of Cardiology, New Orleans, Louisiana, March 2007), Even in the era of DESs, the need for repeated revascularization may be a concern in challenging lesions, such as the LMCA. In a small series, Price et al17 documented the need for repeated procedures in 38% of 50 patients treated with LMCA DESs and followed up with serial angiography at 3 and 9 months. The high recurrence rate was mostly caused by restenosis in the circumflex artery and may be explained by the less aggressive approach to bifurcation stenting compared with the strategy used in our cohort. The ongoing SYNergy between percutaneous coronary intervention with TAXus and cardiac surgery study (SYNTAX; 710 patients)

172

The American Journal of Cardiology (www.AJConline.org)

and PREmier of Randomized COMparison of Bypass Surgery Versus AngioplasTy Using Sirolimus-Eluting Stent in Patients With Left Main Coronary Artery Disease (PRECOMBAT; 600 patients) will soon provide additional information from patients randomly assigned to DES placement or CABG for severe LMCA disease. Inherent selection bias may limit the application of this analysis to a more general population, particularly because of the lack of ability to assess how often patients were refused for the alternative revascularization strategy. Although patients were matched with respect to important clinical characteristics, systematic differences between groups may have influenced results. A propensity analysis was used in an attempt to control for these deficits. It is possible that a larger cohort would have unmasked more significant differences in mortality between the 2 strategies. Interventions included in this analysis were performed over a long period during which techniques, therapies, and adjunctive medications were in sufficient flux to introduce heterogeneity into the management of this group and may have affected their outcomes. Surgical results at this institution were substantially better than the national average, and this may not be reflective of outcomes at less skilled institutions.18 Finally, we reported only mortality data and other clinically important measures, such as (recurrent) myocardial infarction. Stent thrombosis, stroke, or revascularization rates may sufficiently sway the argument in favor of either therapy. 1. Chieffo A, Morici N, Maisano F, Bonizzoni E, Cosgrave J, Montorfano M, Airoldi F, Carlino M, Michev I, Melzi G, et al. Percutaneous treatment with drug-eluting stent implantation versus bypass surgery for unprotected left main stenosis: a single-center experience. Circulation 2006;113:2542–2547. 2. Lee MS, Kapoor N, Jamal F, Czer L, Aragon J, Forrester J, Kar S, Dohad S, Kass R, Eigler N, et al. Comparison of coronary artery bypass surgery with percutaneous coronary intervention with drugeluting stents for unprotected left main coronary artery disease. J Am Coll Cardiol 2006;47:864 – 870. 3. Palmerini T, Marzocchi A, Marrozzini C, Ortolani P, Saia F, Savini C, Bacchi-Reggiani L, Gianstefani S, Virzi S, Manara F, et al. Comparison between coronary angioplasty and coronary artery bypass surgery for the treatment of unprotected left main coronary artery stenosis (the Bologna Registry). Am J Cardiol 2006;98:54 –59. 4. Tan WA, Tamai H, Park S-J, Plokker HWT, Nobuyoshi M, Suzuki T, Colombo A, Macaya C, Holmes DR Jr., Cohen DJ, Whitlow PL, Ellis SG. Long-term clinical outcomes after unprotected left main trunk percutaneous revascularization in 279 patients. Circulation 2001; 104:1609 –1614.

5. Curb JD, Ford CE, Pressel S, Palmer M, Babcock C, Hawkins CM. Ascertainment of vital status through the National Death Index and the Social Security Administration. Am J Epidemiol 1985;121:754 –766. 6. Nashef SA, Roques F, Michel P, Gauducheau E, Lemeshow S, Salamon R. European System for Cardiac Operative Risk Evaluation (EuroSCORE). Eur J Cardiothorac Surg 1999;16:9 –13. 7. Blackstone EH. Comparing apples and oranges. J Thorac Cardiovasc Surg 2002;123:8 –15. 8. Rubin DB. Estimating causal effects from large data sets using propensity scores. Ann Intern Med 1997;127:757–763. 9. Cox D. Regression models and life tables (with discussion). J R Stat Soc B 1972;34:187–220. 10. Smith SC Jr., Feldman TE, Hirshfeld JW Jr., Jacobs AK, Kern MJ, King SB III, Morrison DA, O’Neill WW, Schaff HV, Whitlow PL, et al. ACC/AHA/SCAI 2005 guideline update for percutaneous coronary intervention—summary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/SCAI Writing Committee to Update the 2001 Guidelines for Percutaneous Coronary Intervention). J Am Coll Cardiol 2006;47:216 –235. 11. The BARI Investigators. Seven-year outcome in the Bypass Angioplasty Revascularization Investigation (BARI) by treatment and diabetic status. J Am Coll Cardiol 2000;35:1122–9. 12. McGuire DK, Anstrom KJ, Peterson ED. Influence of the Bypass Angioplasty Revascularization Investigation National Heart, Lung, and Blood Institute Diabetic clinical alert on practice patterns: results from the National Cardiovascular Network Database. Circulation 2003;107:1864 –1870. 13. Chieffo A, Park SJ, Valgimigli M, Kim YH, Daemen J, Sheiban I, Truffa A, Montorfano M, Airoldi F, Sangiorgi G, et al. Favorable long-term outcome after drug-eluting stent implantation in nonbifurcation lesions that involve unprotected left main coronary artery: a multicenter registry. Circulation 2007;116:158 –162. 14. Mercado N, Wijns W, Serruys PW, Sigwart U, Flather MD, Stables RH, O’Neill WW, Rodriguez A, Lemos PA, Hueb WA, et al. One-year outcomes of coronary artery bypass graft surgery versus percutaneous coronary intervention with multiple stenting for multisystem disease: a meta-analysis of individual patient data from randomized clinical trials. J Thorac Cardiovasc Surg 2005;130:512–519. 15. Pocock SJ, Henderson RA, Rickards AF, Hampton JR, King SBR, Hamm CW, Puel J, Hueb W, Goy JJ, Rodriguez A. Meta-analysis of randomised trials comparing coronary angioplasty with bypass surgery [see comments]. Lancet 1995;346:1184 –9. 16. Serruys PW, Unger F, van Hout BA, van den Brand MJ, van Herwerden LA, van Es GA, Bonnier JJ, Simon R, Cremer J, Colombo A, et al. The ARTS study (Arterial Revascularization Therapies Study). Semin Interv Cardiol 1999;4:209 –219. 17. Price MJ, Cristea E, Sawhney N, Kao JA, Moses JW, Leon MB, Costa RA, Lansky AJ, Teirstein PS. Serial angiographic follow-up of sirolimus-eluting stents for unprotected left main coronary artery revascularization. J Am Coll Cardiol 2006;47:871– 877. 18. Society for Thoracic Surgery. Annual STS Report. Available at: www. STS.org 2002. Accessed June 10, 2007.