- Email: [email protected]
Comparison of Intermediate-Term Outcomes of Coronary Artery Bypass Grafting Versus Drug-Eluting Stents for Patients ≥75 Years of Age
Comparison of Intermediate-Term Outcomes of Coronary Artery Bypass Grafting Versus Drug-Eluting Stents for Patients ‡75 Years of Age Edward L. Hannan, PhDa,*, Ye Zhong, MD, MSa, Peter B. Berger, MDb, Gary Walford, MDc, Jeptha P. Curtis, MDd, Chuntao Wu, MDe, Ferdinand J. Venditti, MDf, Robert S.D. Higgins, MDg, Craig R. Smith, MDh, Stephen J. Lahey, MDi, and Spencer B. King III, MDj Several randomized controlled trials and observational studies have compared outcomes of percutaneous coronary interventions (PCIs) with drug-eluting stents (DESs) and coronary artery bypass grafting (CABG), but they have not thoroughly investigated the relative difference in outcomes for patients aged ‡75 years. In this study, a total of 3,864 patients receiving DES and CABG (1,932 CABG-DES pairs) with multivessel coronary disease were propensity matched using multiple patient risk factors and were compared with respect to 3 outcomes (mortality, stroke/myocardial infarction [MI]/mortality, and repeat revascularization) at 2.5 years with a mean follow-up of 18 months. The mortality rates (DES/CABG hazard ratio 1.06, 95% confidence interval 0.87 to 1.30) and the stroke/MI/mortality rates (DES/CABG hazard ratio 1.15, 95% confidence interval 0.97 to 1.38) for the 2 procedures were not significantly different. Repeat revascularization rates were significantly higher for patients who received DESs. In conclusion, older patients experienced similar mortality and stroke/MI/mortality rates for CABG and PCI with DES, although repeat revascularization rates were higher for patients undergoing PCI with DES. Ó 2014 Elsevier Inc. All rights reserved. (Am J Cardiol 2014;113:803e808)
The purposes of this study are to compare patient outcomes of percutaneous coronary interventions (PCIs) with drug-eluting stents (DESs) and coronary artery bypass grafting (CABG) for patients aged 75 years with coronary artery disease and to examine how relative outcomes vary for 4 important risk factors for revascularization (ejection fraction [EF], chronic obstructive pulmonary disease [COPD], diabetes, and proximal left anterior descending [LAD] disease [with stenosis 70%]). Methods End points in the study included 2.5-year mortality, stroke/myocardial infarction (MI)/mortality, and repeat revascularization. The mean follow-up time was 1.5 years; all procedures performed from January 1, 2008 to December 31, 2010 were followed through December 31, 2010. a University at Albany, State University of New York, Albany, New York; Geisinger Medical Center, Danville, Pennsylvania; cJohns Hopkins University, Baltimore, Maryland; dYale School of Medicine, New Haven, Connecticut; ePenn State Hershey College of Medicine, Hershey, Pennsylvania; fAlbany Medical Center, Albany, New York; gThe Ohio State University, Columbus, Ohio; hColumbia-Presbyterian Medical Center, New York, New York; iUniversity of Connecticut Health Center, Storrs, Connecticut; and jSt. Joseph’s Health System, Atlanta, Georgia. Manuscript received August 5, 2013; revised manuscript received and accepted November 7, 2013. This work was supported by grant RC1HL099122 from the National Institutes of Health, Bethesda, Maryland. See page 808 for disclosure information. *Corresponding author: Tel: (518) 402-0333; fax: (518) 402-0414. E-mail address: [email protected] (E.L. Hannan). b
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2013.11.035
The primary databases used for the study were New York State’s clinical registries for PCI and for CABG, the Percutaneous Coronary Interventions Reporting System (PCIRS) and the Cardiac Surgery Reporting System (CSRS), respectively. These registries contain detailed information on patient demographics, risk factors, hemodynamic state, left ventricular function, coronary vessels diseased and attempted, complications, procedure choices, provider identifiers, discharge status, and in-hospital adverse outcomes. All common data elements in the 2 databases have identical definitions. PCIRS also contains information on the type(s) of device used for each attempted lesion, including the type and brand of stent used. Completeness of data reporting is monitored by matching PCIRS to New York’s acute care hospital discharge database, the Statewide Planning and Research Cooperative System (SPARCS) and to the Department of Health’s Ambulatory Surgery Database. PCIRS and CSRS records were matched with SPARCS records using unique hospital identifiers along with patient identifiers and admission, surgery, and discharge dates. Also, in-hospital outcomes are matched to SPARCS to ensure accuracy, and the Department of Health’s utilization review agent audits samples of records from hospitals to ensure the accuracy of risk factor reporting. SPARCS data were also used to identify emergency admissions with MI or stroke as the principal diagnoses in the follow-up period. Patient identifiers in the PCIRS and CSRS were used to link patients in the index revascularization procedure to future admissions in PCIRS and CSRS to identify subsequent revascularization. All staged PCIs were excluded from the repeat revascularization measure. In addition, patient identifiers were used to link the www.ajconline.org
804
The American Journal of Cardiology (www.ajconline.org)
Table 1 Differences before propensity matching in patient characteristics for patients aged 75 years receiving coronary artery bypass grafting (CABG) and drug-eluting stents (DES) in New York State: January 1, 2008 to December 31, 2010
Table 2 Differences after propensity matching in patient characteristics for patients aged 75 years receiving coronary artery bypass grafting (CABG) and drug-eluting stents (DES) in New York State: January 1, 2008 to December 31, 2010
Variable
p Value
Variable
<0.0001
Age groups (yrs) 75e79 80e84 85 Women Hispanic ethnicity Race White Black Other BMI (kg/m2) <16.5 16.5e18.4 18.5e24.9 25.0e30.0 20.1e34.9 35.0e40.0 >40.0 EF (%) <20 2e29 30e39 40e49 50 Previous MI (days) 1e7 8e14 15e20 21 None Carotid/cerebrovascular disease Peripheral vascular disease Hemodynamically unstable Heart failure None At current admission Before current admission Malignant ventricular arrhythmia COPD Diabetes mellitus Renal function Serum creatinine level/no dialysis (mg/dl) <1.2 1.2e1.5 1.6e2.0 2.1e2.5 2.6e3.0 >3.0 Dialysis Number of narrowed coronary arteries 2, With proximal LAD 2, Without proximal LAD 3, With proximal LAD 3, Without proximal LAD
CABG (n ¼ 2,628)
DES (n ¼ 4,299)
56.7 33.9 9.4 38.4 6.2
50.8 33.3 15.9 45.6 9.0
91.0 5.9 3.0
82.7 10.4 6.9
0.3 1.1 30.4 42.2 18.9 5.2 1.8
0.2 1.4 30.7 40.1 18.8 6.1 2.7
1.4 6.0 13.5 19.4 59.8
0.7 3.9 6.8 12.6 75.9
19.2 6.4 1.2 19.3 54.0 26.2 15.8 0.3
19.1 1.8 0.4 10.3 68.5 12.8 11.8 0.1
79.5 16.6 4.0 1.4 27.5 33.7
86.4 10.5 3.1 0.4 8.9 31.8
54.9 28.2 10.7 2.9 1.0 0.6 1.7
57.9 28.4 8.4 2.4 0.6 0.6 1.7
Age groups (yrs) 75e79 80e84 85 Women Hispanic ethnicity Race White Black Other BMI (kg/m2) <16.5 16.5e18.4 18.5e24.9 25.0e30.0 20.1e34.9 35.0e40.0 >40.0 EF (%) <20 2e29 30e39 40e49 50 Previous MI (days) 1e7 8e14 15e20 21 None Carotid/cerebrovascular disease Peripheral vascular disease Hemodynamically unstable Heart failure None At current admission Before current admission Malignant ventricular arrhythmia COPD Diabetes mellitus Renal function Serum creatinine level/no dialysis (mg/dl) <1.2 1.2e1.5 1.6e2.0 2.1e2.5 2.6e3.0 >3.0 Dialysis Number of narrowed coronary arteries 2, With proximal LAD 2, Without proximal LAD 3, With proximal LAD 3, Without proximal LAD Data are presented as percentage. BMI ¼ body mass index.
<0.0001 <0.0001
0.09
<0.0001
<0.0001
<0.0001 <0.0001 0.04 <0.0001
<0.0001 <0.0001 0.11 0.006
<0.0001 17.1 13.4 38.2 31.4
21.1 49.0 11.1 18.7
CABG DES Standardized (n ¼ 1,932) (n ¼ 1,932) Difference
Data are presented as percentage. BMI ¼ body mass index.
55.3 33.6 11.1 41.3 7.0
52.8 35.1 12.1 40.5 7.2
5.0 3.1 3.2 1.5 0.6
88.8 7.4 3.8
88.5 7.3 4.2
1.1 0.4 2.4
0.4 1.2 29.8 40.7 19.8 6.0 2.2
0.2 1.1 30.9 41.3 18.8 5.4 2.4
2.9 0.5 2.4 1.1 2.6 2.5 1.4
0.9 5.3 12.0 18.3 63.5
1.2 5.1 10.7 17.5 65.5
2.5 1.2 3.9 2.0 4.2
20.8 4.1 0.8 16.9 57.4 20.7 14.7 0.2
20.6 2.8 0.7 14.9 61.0 18.2 13.8 0.2
0.4 7.4 1.2 5.5 7.4 6.4 2.7 0.0
82.0 14.8 3.3 0.8 18.2 33.9
83.6 13.5 2.9 0.6 15.3 32.6
4.4 3.7 2.1 3.1 7.8 2.6
55.9 28.8 9.7 2.3 0.8 0.7 1.7
56.6 28.3 9.5 2.7 0.7 0.5 1.7
1.4 1.2 0.9 2.6 1.2 2.8 0.0
21.4 18.1 26.7 33.8
23.7 18.7 23.9 33.8
5.5 1.6 6.7 0.1
Coronary Artery Disease/CABG and PCI Outcomes
805
Table 3 2.5-Year outcomes for propensity-matched patients aged 75 years receiving coronary artery bypass grafting (CABG) and drug-eluting stents (DES) in New York State: January 1, 2008 to December 31, 2010 Long-Term Outcomes
Mortality Stroke/MI/Mortality Repeat revascularization
Observed Rate
Kaplan-Meier Survival Estimates
CABG
DES
CABG
DES
11.3 14.3 3.0
10.7 14.7 15.7
15.8 19.8 4.5
16.9 21.9 24.1
Adjusted HR (95% CI) DES/CABG
p Value
1.06 (0.87e1.30) 1.15 (0.97e1.38) 7.48 (5.61e9.98)
0.58 0.12 <0.0001
Data are presented as percentage.
Figure 1. Kaplan-Meier survival curves for propensity-matched CABG and DES patients during 2.5-year follow-up.
index procedure to New York State’s vital statistics data to identify deaths that occurred after discharge. A total of 90,686 patients were confirmed to have undergone PCI with DES and no bare-metal stents, and a total of 26,323 patients were confirmed to have undergone isolated CABG surgery from January 1, 2008 to December 31, 2010. Of these patients, we sequentially excluded those who were aged <75 years (90,940 patients), patients who had previous revascularization procedure (10,349), had preprocedural cardiogenic shock (29), left main disease (2,441), an MI in the 24 hours before the index procedure (1,065), were from
out-of-state (394), had single-vessel disease (4,846), or had multiple revascularization procedures in the same admission (18). All other patients undergoing revascularization from January 1, 2008 to December 31, 2010 in New York State (6,927, of which 4,299 underwent DES and 2,628 underwent CABG surgery) were used to identify one-to-one propensitymatched pairs of patients with similar risk profiles. These propensity-matched patients were then followed through the end of 2010 to compare the mortality, stroke/MI/mortality, and repeat revascularization for patients receiving DES and CABG. A total of 59 different hospitals were in the study.
806
The American Journal of Cardiology (www.ajconline.org)
Table 4 2.5-Year outcomes for propensity-matched subsets of patients aged 75 years receiving coronary artery bypass grafting (CABG) and drugeluting stents (DES) in New York State: January 1, 2008 to December 31, 2010 Risk Factors
Mortality EF 40% <40% COPD No Yes Diabetes No Yes Proximal LAD disease No Yes Any of the above risk factors None Any Stroke/MI/Mortality EF 40% <40% COPD No Yes Diabetes No Yes Proximal LAD disease No Yes Any of the above risk factors None Any Repeat revascularization EF 40% <40% COPD No Yes Diabetes No Yes Proximal LAD disease No Yes Any of the above risk factors None Any
Matched Pairs
Adjusted HR (95% CI) DES/CABG
p Value
1,251 80
1.01 (0.76e1.34) 1.05 (0.49e2.27)*
0.94 0.89
1,384 98
1.14 (0.89e1.46) 1.16 (0.52e2.58)*
0.30 0.41
909 261
0.83 (0.59e1.16) 0.76 (0.39e1.50)
0.28 0.44
617 534
0.88 (0.60e1.30) 1.17 (0.77e1.80)
0.52 0.47
166 1,202
0.58 (0.22e1.53)* 1.23 (0.97e1.56)
0.52 0.09
1,251 80
1.04 (0.81e1.32) 1.67 (0.72e3.89)*
0.78 0.24
1,384 98
1.25 (1.00e1.56) 0.78 (0.39e1.55)*
0.046 0.48
909 261
1.00 (0.75e1.33) 1.01 (0.59e1.73)
0.99 0.97
617 534
0.96 (0.69e1.34) 1.28 (0.88e1.85)
0.82 0.19
166 1,202
0.65 (0.30e1.45)* 1.27 (1.03e1.56)
0.29 0.02
1,251 80
7.42 (5.23e10.53) 7.95 (1.66e38.15)*
<0.0001 0.01
1,384 98
7.17 (5.16e9.96) 19.12 (2.90e125.88)*
<0.0001 0.0008
909 261
6.61 (4.39e9.95) 14.89 (5.93e37.40)
<0.0001 <0.0001
617 534
6.25 (3.96e9.87) 5.41 (3.14e9.30)
<0.0001 <0.0001
5.00 (2.23e11.24)* 7.78 (5.33e11.35)
<0.0001 <0.0001
166 1,202
* HR was obtained using PROC PHREG to produce a stepwise regression analysis, because of small sample size.
Patients receiving DES and CABG in the study were compared for numerous patient characteristics, including demographics, co-morbidities, ventricular function, preprocedural MI, hemodynamic state, and vessels diseased. Chi-square tests were used to determine significant differences in the use of the 2 types of revascularization for each patient characteristic. Because patients were not randomized to DES and CABG, and because many of the prevalences of the characteristics mentioned previously differed between the 2 groups, propensity score matching was used to identify sets of DES-CABG surgery pairs matched on those characteristics so that the selection bias associated with our observational study could be minimized. The propensity score was derived by developing a logistic regression model that predicted the probability that a given patient would receive CABG surgery on the basis of the patient characteristics mentioned previously. This value was used to match patients without replacement on a one-to-one basis so as to minimize the overall distance in propensity scores between the groups.1e4 Differences between the 2 matched samples in the prevalence of propensity model variables were tested using standardized differences in the observed prevalence of the variables in the matched groups.5 The propensity-matched pairs were then used to analyze differences in outcomes between DES and CABG surgery. Remaining differences in prevalences of propensity model variables were further reduced using multivariable (Cox proportional hazards) models for adjustment along with robust standard errors to control for clustering of patients in matched pairs, taking into account that the samples were matched. Adverse outcome rates of DES and CABG surgery were also compared for 4 patient characteristics (EF, COPD, diabetes, and proximal LAD disease [with stenosis 70%]) that were preselected on the basis of their high risk and use in other studies. This was done for each characteristic by comparing differences in each of the outcomes for matched pairs containing the same value of that characteristic (e.g., all matched pairs in which both patients were diabetics). Adjusted hazard ratios (HRs) were obtained through the same method mentioned previously. All tests were 2-sided and conducted at the 0.05 level, and all analyses were conducted in SAS 9.1 (SAS Institute, Cary, North Carolina). Results A total of 6,927 patients aged 75 years (4,299 patients who received DES and 2,628 CABG) were subjected to propensity matching and 3,864 patients (1,932 pairs and 74% of all patients receiving CABG) were propensity matched. The variables used in the propensity model are all the variables in Table 1, and the C statistic for the models was 0.81. Table 1 presents differences in patient characteristics for patients receiving DES and CABG in New York from 2008 to 2010. As indicated, older patients, women, Hispanics, nonwhites, patients with higher EF and no previous MI, patients who were less likely to have had several different co-morbidities (carotid or cerebrovascular disease, peripheral vascular disease, hemodynamic instability, heart failure, malignant ventricular arrhythmia, COPD, and renal dialysis),
Coronary Artery Disease/CABG and PCI Outcomes
and patients with fewer diseased coronary vessels and absence of proximal LAD disease had higher prevalences of PCI with DES than CABG. Table 2 demonstrates that the DES-CABG pairs that were propensity matched are quite similar, with all of the standardized differences below 10%. Table 3 indicates that there were no significant differences in mortality or stroke/MI/mortality between CABG and DES for propensity-matched patients aged 75 years. However, patients aged 75 years receiving DES had significantly higher risk-adjusted rates of repeat revascularization than propensity-matched patients undergoing CABG. Figure 1 demonstrates that crossovers occur at about 1 year for mortality and at about 8 months for stroke/MI/mortality. Subgroup analyses for 4 groups of high-risk patients (low EF, COPD, diabetes, and proximal LAD disease [70% stenosis]) are presented for each of the outcome measures in Table 4. For mortality, there were no significant differences for any of the subsets. CABG was associated with lower stroke/MI/mortality rates for patients without COPD and for patients with 1 of the 4 risk factors. For repeat revascularization, CABG surgery was associated with significantly lower rates for all subsets of patients. Discussion Numerous comparative effectiveness studies have been conducted to compare outcomes of CABG and PCI with stents for patients with multivessel coronary artery disease.6e20 However, there is a need to continually update studies of this nature because of the evolution of the 2 procedures, particularly PCI, which has evolved from balloon angioplasty to bare-metal stenting to multiple generations of DES. In 2009, Hlatky et al9 pooled data from 10 randomized controlled trials with a median follow-up of 5.9 years. The investigators found that there was no significant difference in mortality between CABG and PCI (CABG/PCI HR 0.91, 95% confidence interval [CI] 0.82 to 1.02, p ¼ 0.12).9 Also, CABG was associated with significantly lower mortality for patients aged 65 years (but not in patients aged <65 years).9 None of the 10 randomized controlled trials involved the use of DES. In the SYNergy between PCI with TAXus and CABG surgery (SYNTAX) trial for patients with 3-vessel disease or left main disease, Mohr et al12 found that at 5 years, CABG was associated with a significantly lower major adverse cardiac and cerebrovascular event rate (26.9% vs 37.3%, p <0.0001). MI and repeat revascularization rates were also significantly lower for patients receiving CABG (3.8% vs 9.7%, p <0.0001% and 13.7% vs 25.9%, p <0.0001, respectively).12 In the Future REvascularization Evaluation in patients with Diabetes mellitus: Optimal management of Multivessel disease (FREEDOM) trial, which assessed outcomes of CABG versus PCI in diabetics with multivessel disease in the DES era with a median follow-up of 3.8 years, Farkouh et al16 found that for death from any cause, nonfatal MI, or nonfatal stroke, the rates were 18.7% versus 26.6%, respectively, p ¼ 0.005. Results were not examined for different age ranges. In the ASCERT (American College of Cardiology Foundation and The Society of Thoracic Surgeons Collaboration on the Comparative Effectiveness of Revascularization Strategies) study by Weintraub et al,14 the authors used matched data from The Society of Thoracic
807
Surgeons and Centers for Medicare and Medicaid Services from 2004 to 2008 to show that at 4-year follow-up, there was lower mortality for older (>65 years) patients with multivessel disease undergoing CABG than PCI (16.4% vs 20.8%, CABG/PCI risk ratio 0.79, 95% CI 0.76 to 0.82). An observational study by Hannan et al13 using 2003 to 2004 data from New York’s cardiac registries for patients with multivessel disease found that CABG was associated with lower mortality (adjusted CABG/PCI HR 0.80, 95% CI 0.65 to 0.97) and mortality/MI (adjusted HR 0.75, 95% CI 0.63 to 0.89) than DES. CABG was also associated with lower MI/mortality rates for patients aged 80 years (adjusted HR 0.74, 95% CI 0.56 to 0.96).13 A New York study by Wu et al15 that followed 2003 to 2005 patients who underwent CABG and had DES through 2008 found that CABG patients were associated with higher 5-year survival rates than DES patients (CABG/DES HR for propensity-matched patients 0.71, 95% CI 0.67 to 0.77). Significantly lower risks of death were also found for CABG for all age groups, although the level of significance was lower for older patients.15 A limitation of the studies mentioned previously is that they either did not look at the relative outcomes for patients aged 75 years, or when they did study the relative outcomes of CABG and PCI by age, they used subgroup analyses that were underpowered by virtue of the analysis method used. The purpose of our study was to compare relative outcomes of CABG and DES in patients aged 75 years as the primary focus of the analyses. This enabled us to then examine key risk factors for those patients to see if the relative outcomes of the 2 procedures differed as a function of whether the patients had those risk factors. We found that the older patients had similar mortality rates and stroke/MI/mortality rates, for DES and CABG, but significantly lower repeat revascularization rates when they underwent CABG. Also, CABG was associated with lower stroke/MI/mortality rates for patients without COPD and for patients with 1 of the 4 individual risk factors that were examined. However, the power to draw conclusions about the 4 individual risk factors we examined in our study is limited due to small sample sizes. As with all observational studies, a caveat of our study is the selection bias introduced by not randomizing patients to the 2 treatments. We have tried to minimize this bias by propensity matching patients using the very large set of patient risk factors available in our CABG and PCI registries, both of which contain essentially the same risk factors with the same definitions. After propensity matching, we found that patients receiving CABG and DES were very similar with regard to the variables used in the matching process, as evidenced by acceptable low values of percent standardized differences. Any remaining differences, although small, were further adjusted for using Cox proportional hazards models. Nevertheless, a remaining threat to selection bias is unmeasured risk factors (e.g., malignancy, dementia, and frailty) that are not contained in the registries, and this omission could have introduced a bias. Also, previous valve surgery and significant valve disease are not in the PCI registry and therefore could not be used for matching patients. Also, compared with randomized controlled trials of CABG and PCI, the absence of frequent mandated follow-up in an observational study that could increase the use of dual
808
The American Journal of Cardiology (www.ajconline.org)
antiplatelet therapy should benefit CABG more in a relative sense and this could bias the outcomes in favor of CABG. In addition, although certain subgroups of CABG patients experienced lower stroke/MI/mortality rates, these strokes and MIs were identified using SPARCS data, and these data are not as precisely defined (International Classification of Diseases, Ninth Revision, codes are used) or audited as the clinical registry data used for other data elements in our study. Also, we had access to only New York State data for deaths after discharge, MI and stroke admissions after discharge, and repeat revascularization, so we could not capture these events outside New York State. To minimize the probability that these events would occur out-of-state, we limited the study to residents of New York State. Nevertheless, patients could have moved to another state or country and either died, been admitted for an MI, or undergone repeat revascularization there, and these events would have been missed by our study. However, there is no reason why there should be a bias in favor of either type of treatment with respect to missed patients, and an earlier study demonstrated that there was not a bias in this regard.21 Acknowledgment: The authors thank New York State’s Cardiac Advisory Committee (CAC) for their encouragement and support of this study and Kimberly S. Cozzens, MA and Cynthia Johnson and the cardiac catheterization laboratories of the participating hospitals for their tireless efforts to ensure the timeliness, completeness, and accuracy of the registry data.
10.
11.
12.
13.
14.
15.
Disclosures The authors have no conflicts of interest to disclose. 1. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41e55. 2. Rosenbaum PR. Optimal matching for observational studies. JASA 1989;84:41e55. 3. Ming K, Rosenbaum PR. A note on optimal matching with variable controls using the assignment algorithm. J Comput Graph Stat 2001;10: 455e463. 4. Gu XS, Rosenbaum PR. Comparison of multivariate matching methods, structures, differences, and algorithms. J Comput Graph Stat 1993;2:405e420. 5. Klein JP, Moeschberger ML. Survival Analysis: Techniques for Censored and Truncated Data. New York: Springer-Verlag; 1997. 6. Booth J, Clayton T, Pepper J, Nugara F, Flather M, Sigwart U, Stables RH; the SoS Investigators. Randomized, controlled trial of coronary artery bypass surgery versus percutaneous coronary intervention in patients with multivessel coronary artery disease. Circulation 2008;118:381e388. 7. Serruys PW, Ong AT, van Herwerden LA, Sousa JE, Jatene A, Bonnier JJRM, Schonberger JPMA, Buller M, Bonser R, Disco C, Backx B, Hugenholtz PG, Firth BG, Unger F. Five-year outcomes after coronary stenting versus bypass surgery for the treatment of multivessel disease: the final analysis of the Arterial Revascularization Therapies Study (ARTS) randomized trial. J Am Coll Cardiol 2005;46:575e581. 8. Malenka DJ, Leavitt BJ, Hearne MJ, Robb JF, Baribeau YR, Ryan TJ, Helm RE, Kellett MA, Dauerman HL, Dacey LJ, Silver MT, VerLee PN, Weldner PW, Hettlemen BD, Olmstead EM, Piper WD, O’Connor GT. Comparing long-term survival of patients with multivessel coronary disease after CABG or PCI: analysis of BARI-like patients in northern New England. Circulation 2005;112(suppl I):I-371eI-376. 9. Hlatky MA, Boothroyd DB, Bravata DM, Boersma E, Booth J, Brooks MM, Carrie D, Clayton TC, Danchin N, Flather M, Hamm CW, Hueb WA, Kahler J, Kelsey SF, King SB, Kosinski AS, Lopes N, McDonald KM, Rodriguez A, Serruys P, Sigwart U, Stables RH, Owens DK, Pocock SJ. Coronary artery bypass surgery compared with
16.
17.
18.
19.
20.
21.
percutaneous coronary interventions for multivessel disease: a collaborative analysis of individual patient data from ten randomized trials. Lancet 2009;373:1190e1197. Morrison DA, Sethi G, Sacks J, Henderson W, Grover F, Ramanathan KB, Weiman D, Saucedo J, Antakli T, Paramesh V, Pett S, Vernon S, Birjiniuk V, Welt F, Krucoff M, Wolfe W, Lucke JC, Mediratta S, Booth D, Murphy E, Ward H, Miller L, Kiesz S, Barbiere C, Lewis D; Investigators of the Department of Veterans Affairs Cooperative Study #385, the Angina With Extremely Serious Operative Mortality Evaluation (AWESOME). Percutaneous coronary intervention versus coronary artery bypass graft surgery for patients with medically refractory myocardial ischemia and risk factors for adverse outcomes with bypass: a multicenter, randomized trial. J Am Coll Cardiol 2001;38: 143e149. King SB III, Kosinski AS, Guyton RA, Lembo NJ, Weintraub WS; for the Emory Angioplasty Versus Surgery Trial. Eight-year mortality in the Emory Angioplasty versus Surgery Trial (EAST). J Am Coll Cardiol 2000;35:1116e1121. Mohr FW, Morice MC, Kappetein AP, Feldman TE, Ståhle E, Colombo A, Mack MJ, Holmes DR Jr, Morel MA, Van Dyck N, Houle VM, Dawkins KD, Serruys PW. Coronary artery bypass graft surgery versus percutaneous coronary intervention in patients with three-vessel disease and left main coronary disease: 5-year follow-up of the randomized, clinical SYNTAX trial. Lancet 2013;381:629e638. Hannan EL, Wu C, Walford G, Culliford AT, Gold JP, Smith CR, Higgins RS, Carlson RE, Jones RH. Drug-eluting stents vs. coronary artery bypass grafting in multivessel coronary disease. N Engl J Med 2008;358:331e341. Weintraub WS, Grau-Sepulveda MV, Weiss JM, O’Brien SM, Peterson ED, Kolm P, Zhang Z, Klein LW, Shaw RE, McKay C, Ritzenthaler LL, Popma JJ, Messenger JC, Shahian DM, Grover FL, Mayer JE, Shewan CM, Garratt KN, Moussa ID, Dangas GD, Edwards FH. Comparative effectiveness of revascularization strategies. N Engl J Med 2012;366:1467e1476. Wu C, Camacho FT, Zhao S, Wechsler AS, Culliford AT, Lahey SJ, King SB 3rd, Walford G, Gold JP, Smith CR, Jordan D, Higgins RS, Hannan EL. Long-term mortality of coronary artery bypass graft surgery and stenting with drug-eluting stents. Ann Thorac Surg 2013;95: 1297e1305. Farkouh ME, Domanski M, Sleeper LA, Siami FS, Dangas G, Mack M, Yang M, Cohen DJ, Rosenberg Y, Solomon SD, Desai AS, Gersh BJ, Magnuson EA, Lansky A, Boineau R, Weinberger J, Ramanathan K, Sousa JE, Rankin J, Bhargava B, Buse J, Hueb W, Smith CR, Muratov V, Bansilal S, King S 3rd, Bertrand M, Fuster V; FREEDOM Trial Investigators. Strategies for multivessel revascularization in patients with diabetes. N Engl J Med 2012;367:2375e2384. Farooq V, van Klaveren D, Steyerberg EW, Meliga E, Vergouwe Y, Chieffo A, Kappetein AP, Colombo A, Holmes DR Jr, Mack M, Feldman T, Morice MC, Ståhle E, Onuma Y, Morel MA, GarciaGarcia HM, van Es GA, Dawkins KD, Mohr FW, Serruys PW. Anatomical and clinical characteristics to guide decision making between coronary artery bypass surgery and percutaneous coronary intervention for individual patients: development and validation of the SYNTAX score II. Lancet 2013;381:639e650. Heub W, Lopes N, Gersh BJ, Soares PR, Ribeiro EE, Pereira AC, Favarato D, Rocha AS, Hueb AC, Ramires JA. Ten-year follow-up survival of the Medicine, Angioplasty, or Surgery Study (MASS II): a randomized controlled clinical trial of 3 therapeutic strategies for multivessel coronary artery disease. Circulation 2010;122:949e957. Rodriguez AE, Baldi J, Fernandez Pereira C, Navia J, Rodriguez Alemparte M, Delacasa A, Vigo F, Vogel D, O’Neill W, Palacios IF; ERACI II Investigators. Five-year follow-up of the Argentine randomized trial of coronary angioplasty with stenting versus coronary bypass surgery inpatients with multiple vessel disease (ERACI II). J Am Coll Cardiol 2005;46:582e588. Serruys PW, Morice MC, Kappetein AP, Colombo A, Holmes DR, Mack MJ, Stahle E, Feldman TE, van den Brand M, Bass EJ, Van Dyck N, Leadley K, Dawkins KD, Mohr FW; SYNTAX Investigators. Percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease. N Engl J Med 2009;360:961e972. Hannan EL, Racz MJ, McCallister BD, Ryan TJ, Arani DT, Isom OW, Jones RH. A comparison of three-year survival after coronary artery bypass graft surgery and percutaneous transluminal coronary angioplasty. J Am Coll Cardiol 1999;33:63e72.
The purposes of this study are to compare patient outcomes of percutaneous coronary interventions (PCIs) with drug-eluting stents (DESs) and coronary artery bypass grafting (CABG) for patients aged 75 years with coronary artery disease and to examine how relative outcomes vary for 4 important risk factors for revascularization (ejection fraction [EF], chronic obstructive pulmonary disease [COPD], diabetes, and proximal left anterior descending [LAD] disease [with stenosis 70%]). Methods End points in the study included 2.5-year mortality, stroke/myocardial infarction (MI)/mortality, and repeat revascularization. The mean follow-up time was 1.5 years; all procedures performed from January 1, 2008 to December 31, 2010 were followed through December 31, 2010. a University at Albany, State University of New York, Albany, New York; Geisinger Medical Center, Danville, Pennsylvania; cJohns Hopkins University, Baltimore, Maryland; dYale School of Medicine, New Haven, Connecticut; ePenn State Hershey College of Medicine, Hershey, Pennsylvania; fAlbany Medical Center, Albany, New York; gThe Ohio State University, Columbus, Ohio; hColumbia-Presbyterian Medical Center, New York, New York; iUniversity of Connecticut Health Center, Storrs, Connecticut; and jSt. Joseph’s Health System, Atlanta, Georgia. Manuscript received August 5, 2013; revised manuscript received and accepted November 7, 2013. This work was supported by grant RC1HL099122 from the National Institutes of Health, Bethesda, Maryland. See page 808 for disclosure information. *Corresponding author: Tel: (518) 402-0333; fax: (518) 402-0414. E-mail address: [email protected] (E.L. Hannan). b
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2013.11.035
The primary databases used for the study were New York State’s clinical registries for PCI and for CABG, the Percutaneous Coronary Interventions Reporting System (PCIRS) and the Cardiac Surgery Reporting System (CSRS), respectively. These registries contain detailed information on patient demographics, risk factors, hemodynamic state, left ventricular function, coronary vessels diseased and attempted, complications, procedure choices, provider identifiers, discharge status, and in-hospital adverse outcomes. All common data elements in the 2 databases have identical definitions. PCIRS also contains information on the type(s) of device used for each attempted lesion, including the type and brand of stent used. Completeness of data reporting is monitored by matching PCIRS to New York’s acute care hospital discharge database, the Statewide Planning and Research Cooperative System (SPARCS) and to the Department of Health’s Ambulatory Surgery Database. PCIRS and CSRS records were matched with SPARCS records using unique hospital identifiers along with patient identifiers and admission, surgery, and discharge dates. Also, in-hospital outcomes are matched to SPARCS to ensure accuracy, and the Department of Health’s utilization review agent audits samples of records from hospitals to ensure the accuracy of risk factor reporting. SPARCS data were also used to identify emergency admissions with MI or stroke as the principal diagnoses in the follow-up period. Patient identifiers in the PCIRS and CSRS were used to link patients in the index revascularization procedure to future admissions in PCIRS and CSRS to identify subsequent revascularization. All staged PCIs were excluded from the repeat revascularization measure. In addition, patient identifiers were used to link the www.ajconline.org
804
The American Journal of Cardiology (www.ajconline.org)
Table 1 Differences before propensity matching in patient characteristics for patients aged 75 years receiving coronary artery bypass grafting (CABG) and drug-eluting stents (DES) in New York State: January 1, 2008 to December 31, 2010
Table 2 Differences after propensity matching in patient characteristics for patients aged 75 years receiving coronary artery bypass grafting (CABG) and drug-eluting stents (DES) in New York State: January 1, 2008 to December 31, 2010
Variable
p Value
Variable
<0.0001
Age groups (yrs) 75e79 80e84 85 Women Hispanic ethnicity Race White Black Other BMI (kg/m2) <16.5 16.5e18.4 18.5e24.9 25.0e30.0 20.1e34.9 35.0e40.0 >40.0 EF (%) <20 2e29 30e39 40e49 50 Previous MI (days) 1e7 8e14 15e20 21 None Carotid/cerebrovascular disease Peripheral vascular disease Hemodynamically unstable Heart failure None At current admission Before current admission Malignant ventricular arrhythmia COPD Diabetes mellitus Renal function Serum creatinine level/no dialysis (mg/dl) <1.2 1.2e1.5 1.6e2.0 2.1e2.5 2.6e3.0 >3.0 Dialysis Number of narrowed coronary arteries 2, With proximal LAD 2, Without proximal LAD 3, With proximal LAD 3, Without proximal LAD
CABG (n ¼ 2,628)
DES (n ¼ 4,299)
56.7 33.9 9.4 38.4 6.2
50.8 33.3 15.9 45.6 9.0
91.0 5.9 3.0
82.7 10.4 6.9
0.3 1.1 30.4 42.2 18.9 5.2 1.8
0.2 1.4 30.7 40.1 18.8 6.1 2.7
1.4 6.0 13.5 19.4 59.8
0.7 3.9 6.8 12.6 75.9
19.2 6.4 1.2 19.3 54.0 26.2 15.8 0.3
19.1 1.8 0.4 10.3 68.5 12.8 11.8 0.1
79.5 16.6 4.0 1.4 27.5 33.7
86.4 10.5 3.1 0.4 8.9 31.8
54.9 28.2 10.7 2.9 1.0 0.6 1.7
57.9 28.4 8.4 2.4 0.6 0.6 1.7
Age groups (yrs) 75e79 80e84 85 Women Hispanic ethnicity Race White Black Other BMI (kg/m2) <16.5 16.5e18.4 18.5e24.9 25.0e30.0 20.1e34.9 35.0e40.0 >40.0 EF (%) <20 2e29 30e39 40e49 50 Previous MI (days) 1e7 8e14 15e20 21 None Carotid/cerebrovascular disease Peripheral vascular disease Hemodynamically unstable Heart failure None At current admission Before current admission Malignant ventricular arrhythmia COPD Diabetes mellitus Renal function Serum creatinine level/no dialysis (mg/dl) <1.2 1.2e1.5 1.6e2.0 2.1e2.5 2.6e3.0 >3.0 Dialysis Number of narrowed coronary arteries 2, With proximal LAD 2, Without proximal LAD 3, With proximal LAD 3, Without proximal LAD Data are presented as percentage. BMI ¼ body mass index.
<0.0001 <0.0001
0.09
<0.0001
<0.0001
<0.0001 <0.0001 0.04 <0.0001
<0.0001 <0.0001 0.11 0.006
<0.0001 17.1 13.4 38.2 31.4
21.1 49.0 11.1 18.7
CABG DES Standardized (n ¼ 1,932) (n ¼ 1,932) Difference
Data are presented as percentage. BMI ¼ body mass index.
55.3 33.6 11.1 41.3 7.0
52.8 35.1 12.1 40.5 7.2
5.0 3.1 3.2 1.5 0.6
88.8 7.4 3.8
88.5 7.3 4.2
1.1 0.4 2.4
0.4 1.2 29.8 40.7 19.8 6.0 2.2
0.2 1.1 30.9 41.3 18.8 5.4 2.4
2.9 0.5 2.4 1.1 2.6 2.5 1.4
0.9 5.3 12.0 18.3 63.5
1.2 5.1 10.7 17.5 65.5
2.5 1.2 3.9 2.0 4.2
20.8 4.1 0.8 16.9 57.4 20.7 14.7 0.2
20.6 2.8 0.7 14.9 61.0 18.2 13.8 0.2
0.4 7.4 1.2 5.5 7.4 6.4 2.7 0.0
82.0 14.8 3.3 0.8 18.2 33.9
83.6 13.5 2.9 0.6 15.3 32.6
4.4 3.7 2.1 3.1 7.8 2.6
55.9 28.8 9.7 2.3 0.8 0.7 1.7
56.6 28.3 9.5 2.7 0.7 0.5 1.7
1.4 1.2 0.9 2.6 1.2 2.8 0.0
21.4 18.1 26.7 33.8
23.7 18.7 23.9 33.8
5.5 1.6 6.7 0.1
Coronary Artery Disease/CABG and PCI Outcomes
805
Table 3 2.5-Year outcomes for propensity-matched patients aged 75 years receiving coronary artery bypass grafting (CABG) and drug-eluting stents (DES) in New York State: January 1, 2008 to December 31, 2010 Long-Term Outcomes
Mortality Stroke/MI/Mortality Repeat revascularization
Observed Rate
Kaplan-Meier Survival Estimates
CABG
DES
CABG
DES
11.3 14.3 3.0
10.7 14.7 15.7
15.8 19.8 4.5
16.9 21.9 24.1
Adjusted HR (95% CI) DES/CABG
p Value
1.06 (0.87e1.30) 1.15 (0.97e1.38) 7.48 (5.61e9.98)
0.58 0.12 <0.0001
Data are presented as percentage.
Figure 1. Kaplan-Meier survival curves for propensity-matched CABG and DES patients during 2.5-year follow-up.
index procedure to New York State’s vital statistics data to identify deaths that occurred after discharge. A total of 90,686 patients were confirmed to have undergone PCI with DES and no bare-metal stents, and a total of 26,323 patients were confirmed to have undergone isolated CABG surgery from January 1, 2008 to December 31, 2010. Of these patients, we sequentially excluded those who were aged <75 years (90,940 patients), patients who had previous revascularization procedure (10,349), had preprocedural cardiogenic shock (29), left main disease (2,441), an MI in the 24 hours before the index procedure (1,065), were from
out-of-state (394), had single-vessel disease (4,846), or had multiple revascularization procedures in the same admission (18). All other patients undergoing revascularization from January 1, 2008 to December 31, 2010 in New York State (6,927, of which 4,299 underwent DES and 2,628 underwent CABG surgery) were used to identify one-to-one propensitymatched pairs of patients with similar risk profiles. These propensity-matched patients were then followed through the end of 2010 to compare the mortality, stroke/MI/mortality, and repeat revascularization for patients receiving DES and CABG. A total of 59 different hospitals were in the study.
806
The American Journal of Cardiology (www.ajconline.org)
Table 4 2.5-Year outcomes for propensity-matched subsets of patients aged 75 years receiving coronary artery bypass grafting (CABG) and drugeluting stents (DES) in New York State: January 1, 2008 to December 31, 2010 Risk Factors
Mortality EF 40% <40% COPD No Yes Diabetes No Yes Proximal LAD disease No Yes Any of the above risk factors None Any Stroke/MI/Mortality EF 40% <40% COPD No Yes Diabetes No Yes Proximal LAD disease No Yes Any of the above risk factors None Any Repeat revascularization EF 40% <40% COPD No Yes Diabetes No Yes Proximal LAD disease No Yes Any of the above risk factors None Any
Matched Pairs
Adjusted HR (95% CI) DES/CABG
p Value
1,251 80
1.01 (0.76e1.34) 1.05 (0.49e2.27)*
0.94 0.89
1,384 98
1.14 (0.89e1.46) 1.16 (0.52e2.58)*
0.30 0.41
909 261
0.83 (0.59e1.16) 0.76 (0.39e1.50)
0.28 0.44
617 534
0.88 (0.60e1.30) 1.17 (0.77e1.80)
0.52 0.47
166 1,202
0.58 (0.22e1.53)* 1.23 (0.97e1.56)
0.52 0.09
1,251 80
1.04 (0.81e1.32) 1.67 (0.72e3.89)*
0.78 0.24
1,384 98
1.25 (1.00e1.56) 0.78 (0.39e1.55)*
0.046 0.48
909 261
1.00 (0.75e1.33) 1.01 (0.59e1.73)
0.99 0.97
617 534
0.96 (0.69e1.34) 1.28 (0.88e1.85)
0.82 0.19
166 1,202
0.65 (0.30e1.45)* 1.27 (1.03e1.56)
0.29 0.02
1,251 80
7.42 (5.23e10.53) 7.95 (1.66e38.15)*
<0.0001 0.01
1,384 98
7.17 (5.16e9.96) 19.12 (2.90e125.88)*
<0.0001 0.0008
909 261
6.61 (4.39e9.95) 14.89 (5.93e37.40)
<0.0001 <0.0001
617 534
6.25 (3.96e9.87) 5.41 (3.14e9.30)
<0.0001 <0.0001
5.00 (2.23e11.24)* 7.78 (5.33e11.35)
<0.0001 <0.0001
166 1,202
* HR was obtained using PROC PHREG to produce a stepwise regression analysis, because of small sample size.
Patients receiving DES and CABG in the study were compared for numerous patient characteristics, including demographics, co-morbidities, ventricular function, preprocedural MI, hemodynamic state, and vessels diseased. Chi-square tests were used to determine significant differences in the use of the 2 types of revascularization for each patient characteristic. Because patients were not randomized to DES and CABG, and because many of the prevalences of the characteristics mentioned previously differed between the 2 groups, propensity score matching was used to identify sets of DES-CABG surgery pairs matched on those characteristics so that the selection bias associated with our observational study could be minimized. The propensity score was derived by developing a logistic regression model that predicted the probability that a given patient would receive CABG surgery on the basis of the patient characteristics mentioned previously. This value was used to match patients without replacement on a one-to-one basis so as to minimize the overall distance in propensity scores between the groups.1e4 Differences between the 2 matched samples in the prevalence of propensity model variables were tested using standardized differences in the observed prevalence of the variables in the matched groups.5 The propensity-matched pairs were then used to analyze differences in outcomes between DES and CABG surgery. Remaining differences in prevalences of propensity model variables were further reduced using multivariable (Cox proportional hazards) models for adjustment along with robust standard errors to control for clustering of patients in matched pairs, taking into account that the samples were matched. Adverse outcome rates of DES and CABG surgery were also compared for 4 patient characteristics (EF, COPD, diabetes, and proximal LAD disease [with stenosis 70%]) that were preselected on the basis of their high risk and use in other studies. This was done for each characteristic by comparing differences in each of the outcomes for matched pairs containing the same value of that characteristic (e.g., all matched pairs in which both patients were diabetics). Adjusted hazard ratios (HRs) were obtained through the same method mentioned previously. All tests were 2-sided and conducted at the 0.05 level, and all analyses were conducted in SAS 9.1 (SAS Institute, Cary, North Carolina). Results A total of 6,927 patients aged 75 years (4,299 patients who received DES and 2,628 CABG) were subjected to propensity matching and 3,864 patients (1,932 pairs and 74% of all patients receiving CABG) were propensity matched. The variables used in the propensity model are all the variables in Table 1, and the C statistic for the models was 0.81. Table 1 presents differences in patient characteristics for patients receiving DES and CABG in New York from 2008 to 2010. As indicated, older patients, women, Hispanics, nonwhites, patients with higher EF and no previous MI, patients who were less likely to have had several different co-morbidities (carotid or cerebrovascular disease, peripheral vascular disease, hemodynamic instability, heart failure, malignant ventricular arrhythmia, COPD, and renal dialysis),
Coronary Artery Disease/CABG and PCI Outcomes
and patients with fewer diseased coronary vessels and absence of proximal LAD disease had higher prevalences of PCI with DES than CABG. Table 2 demonstrates that the DES-CABG pairs that were propensity matched are quite similar, with all of the standardized differences below 10%. Table 3 indicates that there were no significant differences in mortality or stroke/MI/mortality between CABG and DES for propensity-matched patients aged 75 years. However, patients aged 75 years receiving DES had significantly higher risk-adjusted rates of repeat revascularization than propensity-matched patients undergoing CABG. Figure 1 demonstrates that crossovers occur at about 1 year for mortality and at about 8 months for stroke/MI/mortality. Subgroup analyses for 4 groups of high-risk patients (low EF, COPD, diabetes, and proximal LAD disease [70% stenosis]) are presented for each of the outcome measures in Table 4. For mortality, there were no significant differences for any of the subsets. CABG was associated with lower stroke/MI/mortality rates for patients without COPD and for patients with 1 of the 4 risk factors. For repeat revascularization, CABG surgery was associated with significantly lower rates for all subsets of patients. Discussion Numerous comparative effectiveness studies have been conducted to compare outcomes of CABG and PCI with stents for patients with multivessel coronary artery disease.6e20 However, there is a need to continually update studies of this nature because of the evolution of the 2 procedures, particularly PCI, which has evolved from balloon angioplasty to bare-metal stenting to multiple generations of DES. In 2009, Hlatky et al9 pooled data from 10 randomized controlled trials with a median follow-up of 5.9 years. The investigators found that there was no significant difference in mortality between CABG and PCI (CABG/PCI HR 0.91, 95% confidence interval [CI] 0.82 to 1.02, p ¼ 0.12).9 Also, CABG was associated with significantly lower mortality for patients aged 65 years (but not in patients aged <65 years).9 None of the 10 randomized controlled trials involved the use of DES. In the SYNergy between PCI with TAXus and CABG surgery (SYNTAX) trial for patients with 3-vessel disease or left main disease, Mohr et al12 found that at 5 years, CABG was associated with a significantly lower major adverse cardiac and cerebrovascular event rate (26.9% vs 37.3%, p <0.0001). MI and repeat revascularization rates were also significantly lower for patients receiving CABG (3.8% vs 9.7%, p <0.0001% and 13.7% vs 25.9%, p <0.0001, respectively).12 In the Future REvascularization Evaluation in patients with Diabetes mellitus: Optimal management of Multivessel disease (FREEDOM) trial, which assessed outcomes of CABG versus PCI in diabetics with multivessel disease in the DES era with a median follow-up of 3.8 years, Farkouh et al16 found that for death from any cause, nonfatal MI, or nonfatal stroke, the rates were 18.7% versus 26.6%, respectively, p ¼ 0.005. Results were not examined for different age ranges. In the ASCERT (American College of Cardiology Foundation and The Society of Thoracic Surgeons Collaboration on the Comparative Effectiveness of Revascularization Strategies) study by Weintraub et al,14 the authors used matched data from The Society of Thoracic
807
Surgeons and Centers for Medicare and Medicaid Services from 2004 to 2008 to show that at 4-year follow-up, there was lower mortality for older (>65 years) patients with multivessel disease undergoing CABG than PCI (16.4% vs 20.8%, CABG/PCI risk ratio 0.79, 95% CI 0.76 to 0.82). An observational study by Hannan et al13 using 2003 to 2004 data from New York’s cardiac registries for patients with multivessel disease found that CABG was associated with lower mortality (adjusted CABG/PCI HR 0.80, 95% CI 0.65 to 0.97) and mortality/MI (adjusted HR 0.75, 95% CI 0.63 to 0.89) than DES. CABG was also associated with lower MI/mortality rates for patients aged 80 years (adjusted HR 0.74, 95% CI 0.56 to 0.96).13 A New York study by Wu et al15 that followed 2003 to 2005 patients who underwent CABG and had DES through 2008 found that CABG patients were associated with higher 5-year survival rates than DES patients (CABG/DES HR for propensity-matched patients 0.71, 95% CI 0.67 to 0.77). Significantly lower risks of death were also found for CABG for all age groups, although the level of significance was lower for older patients.15 A limitation of the studies mentioned previously is that they either did not look at the relative outcomes for patients aged 75 years, or when they did study the relative outcomes of CABG and PCI by age, they used subgroup analyses that were underpowered by virtue of the analysis method used. The purpose of our study was to compare relative outcomes of CABG and DES in patients aged 75 years as the primary focus of the analyses. This enabled us to then examine key risk factors for those patients to see if the relative outcomes of the 2 procedures differed as a function of whether the patients had those risk factors. We found that the older patients had similar mortality rates and stroke/MI/mortality rates, for DES and CABG, but significantly lower repeat revascularization rates when they underwent CABG. Also, CABG was associated with lower stroke/MI/mortality rates for patients without COPD and for patients with 1 of the 4 individual risk factors that were examined. However, the power to draw conclusions about the 4 individual risk factors we examined in our study is limited due to small sample sizes. As with all observational studies, a caveat of our study is the selection bias introduced by not randomizing patients to the 2 treatments. We have tried to minimize this bias by propensity matching patients using the very large set of patient risk factors available in our CABG and PCI registries, both of which contain essentially the same risk factors with the same definitions. After propensity matching, we found that patients receiving CABG and DES were very similar with regard to the variables used in the matching process, as evidenced by acceptable low values of percent standardized differences. Any remaining differences, although small, were further adjusted for using Cox proportional hazards models. Nevertheless, a remaining threat to selection bias is unmeasured risk factors (e.g., malignancy, dementia, and frailty) that are not contained in the registries, and this omission could have introduced a bias. Also, previous valve surgery and significant valve disease are not in the PCI registry and therefore could not be used for matching patients. Also, compared with randomized controlled trials of CABG and PCI, the absence of frequent mandated follow-up in an observational study that could increase the use of dual
808
The American Journal of Cardiology (www.ajconline.org)
antiplatelet therapy should benefit CABG more in a relative sense and this could bias the outcomes in favor of CABG. In addition, although certain subgroups of CABG patients experienced lower stroke/MI/mortality rates, these strokes and MIs were identified using SPARCS data, and these data are not as precisely defined (International Classification of Diseases, Ninth Revision, codes are used) or audited as the clinical registry data used for other data elements in our study. Also, we had access to only New York State data for deaths after discharge, MI and stroke admissions after discharge, and repeat revascularization, so we could not capture these events outside New York State. To minimize the probability that these events would occur out-of-state, we limited the study to residents of New York State. Nevertheless, patients could have moved to another state or country and either died, been admitted for an MI, or undergone repeat revascularization there, and these events would have been missed by our study. However, there is no reason why there should be a bias in favor of either type of treatment with respect to missed patients, and an earlier study demonstrated that there was not a bias in this regard.21 Acknowledgment: The authors thank New York State’s Cardiac Advisory Committee (CAC) for their encouragement and support of this study and Kimberly S. Cozzens, MA and Cynthia Johnson and the cardiac catheterization laboratories of the participating hospitals for their tireless efforts to ensure the timeliness, completeness, and accuracy of the registry data.
10.
11.
12.
13.
14.
15.
Disclosures The authors have no conflicts of interest to disclose. 1. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41e55. 2. Rosenbaum PR. Optimal matching for observational studies. JASA 1989;84:41e55. 3. Ming K, Rosenbaum PR. A note on optimal matching with variable controls using the assignment algorithm. J Comput Graph Stat 2001;10: 455e463. 4. Gu XS, Rosenbaum PR. Comparison of multivariate matching methods, structures, differences, and algorithms. J Comput Graph Stat 1993;2:405e420. 5. Klein JP, Moeschberger ML. Survival Analysis: Techniques for Censored and Truncated Data. New York: Springer-Verlag; 1997. 6. Booth J, Clayton T, Pepper J, Nugara F, Flather M, Sigwart U, Stables RH; the SoS Investigators. Randomized, controlled trial of coronary artery bypass surgery versus percutaneous coronary intervention in patients with multivessel coronary artery disease. Circulation 2008;118:381e388. 7. Serruys PW, Ong AT, van Herwerden LA, Sousa JE, Jatene A, Bonnier JJRM, Schonberger JPMA, Buller M, Bonser R, Disco C, Backx B, Hugenholtz PG, Firth BG, Unger F. Five-year outcomes after coronary stenting versus bypass surgery for the treatment of multivessel disease: the final analysis of the Arterial Revascularization Therapies Study (ARTS) randomized trial. J Am Coll Cardiol 2005;46:575e581. 8. Malenka DJ, Leavitt BJ, Hearne MJ, Robb JF, Baribeau YR, Ryan TJ, Helm RE, Kellett MA, Dauerman HL, Dacey LJ, Silver MT, VerLee PN, Weldner PW, Hettlemen BD, Olmstead EM, Piper WD, O’Connor GT. Comparing long-term survival of patients with multivessel coronary disease after CABG or PCI: analysis of BARI-like patients in northern New England. Circulation 2005;112(suppl I):I-371eI-376. 9. Hlatky MA, Boothroyd DB, Bravata DM, Boersma E, Booth J, Brooks MM, Carrie D, Clayton TC, Danchin N, Flather M, Hamm CW, Hueb WA, Kahler J, Kelsey SF, King SB, Kosinski AS, Lopes N, McDonald KM, Rodriguez A, Serruys P, Sigwart U, Stables RH, Owens DK, Pocock SJ. Coronary artery bypass surgery compared with
16.
17.
18.
19.
20.
21.
percutaneous coronary interventions for multivessel disease: a collaborative analysis of individual patient data from ten randomized trials. Lancet 2009;373:1190e1197. Morrison DA, Sethi G, Sacks J, Henderson W, Grover F, Ramanathan KB, Weiman D, Saucedo J, Antakli T, Paramesh V, Pett S, Vernon S, Birjiniuk V, Welt F, Krucoff M, Wolfe W, Lucke JC, Mediratta S, Booth D, Murphy E, Ward H, Miller L, Kiesz S, Barbiere C, Lewis D; Investigators of the Department of Veterans Affairs Cooperative Study #385, the Angina With Extremely Serious Operative Mortality Evaluation (AWESOME). Percutaneous coronary intervention versus coronary artery bypass graft surgery for patients with medically refractory myocardial ischemia and risk factors for adverse outcomes with bypass: a multicenter, randomized trial. J Am Coll Cardiol 2001;38: 143e149. King SB III, Kosinski AS, Guyton RA, Lembo NJ, Weintraub WS; for the Emory Angioplasty Versus Surgery Trial. Eight-year mortality in the Emory Angioplasty versus Surgery Trial (EAST). J Am Coll Cardiol 2000;35:1116e1121. Mohr FW, Morice MC, Kappetein AP, Feldman TE, Ståhle E, Colombo A, Mack MJ, Holmes DR Jr, Morel MA, Van Dyck N, Houle VM, Dawkins KD, Serruys PW. Coronary artery bypass graft surgery versus percutaneous coronary intervention in patients with three-vessel disease and left main coronary disease: 5-year follow-up of the randomized, clinical SYNTAX trial. Lancet 2013;381:629e638. Hannan EL, Wu C, Walford G, Culliford AT, Gold JP, Smith CR, Higgins RS, Carlson RE, Jones RH. Drug-eluting stents vs. coronary artery bypass grafting in multivessel coronary disease. N Engl J Med 2008;358:331e341. Weintraub WS, Grau-Sepulveda MV, Weiss JM, O’Brien SM, Peterson ED, Kolm P, Zhang Z, Klein LW, Shaw RE, McKay C, Ritzenthaler LL, Popma JJ, Messenger JC, Shahian DM, Grover FL, Mayer JE, Shewan CM, Garratt KN, Moussa ID, Dangas GD, Edwards FH. Comparative effectiveness of revascularization strategies. N Engl J Med 2012;366:1467e1476. Wu C, Camacho FT, Zhao S, Wechsler AS, Culliford AT, Lahey SJ, King SB 3rd, Walford G, Gold JP, Smith CR, Jordan D, Higgins RS, Hannan EL. Long-term mortality of coronary artery bypass graft surgery and stenting with drug-eluting stents. Ann Thorac Surg 2013;95: 1297e1305. Farkouh ME, Domanski M, Sleeper LA, Siami FS, Dangas G, Mack M, Yang M, Cohen DJ, Rosenberg Y, Solomon SD, Desai AS, Gersh BJ, Magnuson EA, Lansky A, Boineau R, Weinberger J, Ramanathan K, Sousa JE, Rankin J, Bhargava B, Buse J, Hueb W, Smith CR, Muratov V, Bansilal S, King S 3rd, Bertrand M, Fuster V; FREEDOM Trial Investigators. Strategies for multivessel revascularization in patients with diabetes. N Engl J Med 2012;367:2375e2384. Farooq V, van Klaveren D, Steyerberg EW, Meliga E, Vergouwe Y, Chieffo A, Kappetein AP, Colombo A, Holmes DR Jr, Mack M, Feldman T, Morice MC, Ståhle E, Onuma Y, Morel MA, GarciaGarcia HM, van Es GA, Dawkins KD, Mohr FW, Serruys PW. Anatomical and clinical characteristics to guide decision making between coronary artery bypass surgery and percutaneous coronary intervention for individual patients: development and validation of the SYNTAX score II. Lancet 2013;381:639e650. Heub W, Lopes N, Gersh BJ, Soares PR, Ribeiro EE, Pereira AC, Favarato D, Rocha AS, Hueb AC, Ramires JA. Ten-year follow-up survival of the Medicine, Angioplasty, or Surgery Study (MASS II): a randomized controlled clinical trial of 3 therapeutic strategies for multivessel coronary artery disease. Circulation 2010;122:949e957. Rodriguez AE, Baldi J, Fernandez Pereira C, Navia J, Rodriguez Alemparte M, Delacasa A, Vigo F, Vogel D, O’Neill W, Palacios IF; ERACI II Investigators. Five-year follow-up of the Argentine randomized trial of coronary angioplasty with stenting versus coronary bypass surgery inpatients with multiple vessel disease (ERACI II). J Am Coll Cardiol 2005;46:582e588. Serruys PW, Morice MC, Kappetein AP, Colombo A, Holmes DR, Mack MJ, Stahle E, Feldman TE, van den Brand M, Bass EJ, Van Dyck N, Leadley K, Dawkins KD, Mohr FW; SYNTAX Investigators. Percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease. N Engl J Med 2009;360:961e972. Hannan EL, Racz MJ, McCallister BD, Ryan TJ, Arani DT, Isom OW, Jones RH. A comparison of three-year survival after coronary artery bypass graft surgery and percutaneous transluminal coronary angioplasty. J Am Coll Cardiol 1999;33:63e72.