Author + information
- Received June 4, 2019
- Revision received June 20, 2019
- Accepted June 21, 2019
- Published online September 2, 2019.
- Zaza Samadashvili, MDa,
- Thoralf M. Sundt III, MDb,
- Andrew Wechsler, MDc,
- Joanna Chikwe, MDd,
- David H. Adams, MDe,
- Craig R. Smith, MDf,
- Desmond Jordan, MDf,
- Leonard Girardi, MDg,
- Stephen J. Lahey, MDh,
- Jeffrey P. Gold, MDi,
- Mohammed H. Ashraf, MDj and
- Edward L. Hannan, PhDa,∗ (, )@UAlbanySPH
- aSchool of Public Health, University at Albany, State University of New York, Rensselaer, New York
- bCardiac Surgical Division, Massachusetts General Hospital, Boston, Massachusetts
- cDepartment of Cardiothoracic Surgery, Drexel University, Philadelphia, Pennsylvania
- dDepartment of Cardiothoracic Surgery, Stony Brook School of Medicine, Stony Brook, New York
- eDepartment of Surgery, Mount Sinai Hospital, New York, New York
- fDepartment of Surgery, Columbia-Presbyterian Medical Center, New York, New York
- gDepartment of Cardiothoracic Surgery, Weill Cornell Medical Center, New York, New York
- hDivision of Cardiothoracic Surgery, University of Connecticut, Storrs, Connecticut
- iChancellor, University of Nebraska Medical Center, Omaha, Nebraska
- jDepartment of Cardiothoracic Surgery, Kaleida Health, Buffalo, New York
- ↵∗Address for correspondence:
Dr. Edward L. Hannan, School of Public Health, State University of New York, University at Albany, One University Place, Rensselaer, New York 12144-3456.
Background Despite recent guideline statements, there is still wide practice variation in the use of multiple arterial grafts (MAGs) versus single arterial grafts (SAGs) for patients with multivessel disease undergoing coronary artery bypass graft surgery. This may be related to differences in findings between observational and randomized controlled studies.
Objectives This study sought to compare intermediate-term MAG and SAG outcomes with enhanced matching to reduce selection bias.
Methods New York’s cardiac registry identified 63,402 multivessel disease patients undergoing coronary artery bypass graft surgery between January 1, 2005, and December 31, 2014, to compare outcomes (median follow-up 6.5 years) for patients receiving SAGs and MAGs. SAG and MAG patients were propensity matched using 38 baseline characteristics to reduce selection bias. The primary endpoint was mortality, and secondary endpoints included repeat revascularization and a composite endpoint of mortality, acute myocardial infarction, and stroke.
Results Before matching, 20% of procedures employed MAG. At 1 year, there was no mortality difference between matched MAG and SAG patients (2.4% vs. 2.2%, adjusted hazard ratio [AHR]: 1.11; 95% confidence interval [CI]: 0.93 to 1.32). At 7 years, MAG patients had lower mortality (12.7% vs. 14.3%, AHR: 0.86; 95% CI: 0.79 to 0.93), a lower composite outcome (20.2% vs. 22.8%, AHR: 0.88; 95% CI: 0.83 to 0.93), and a lower repeat revascularization rate (11.7% vs. 14.6%, AHR: 0.80; 95% CI: 0.74 to 0.87). At 7 years, the subgroups for which MAG did not have a lower mortality rate included patients with off-pump surgery, 2-vessel disease with right coronary artery disease, recent acute myocardial infarction, renal dysfunction, and patient ≥70 years of age.
Conclusions Mortality and the composite outcome were similar between MAG and SAG patients at 1 year, but lower for MAG after 7 years. Patients of higher volume MAG surgeons experienced lower MAG mortality.
- CABG surgery
- mortality differences
- multiple arterial revascularization
- multiple vessel disease
- observational study
Outcomes for isolated coronary artery bypass graft (CABG) surgery have improved markedly in the world over the course of the last 3 decades, with rates in New York State decreasing from 3.52% for in-hospital mortality in 1989 to 1.67% for in-hospital/30-day mortality in 2016 (1). Part of the reason for this improvement in outcomes is likely to have been the increased use of left internal thoracic artery (LITA) grafts to the left anterior descending coronary artery (LAD). The use of LITA-LAD grafts for patients with LAD disease now exceeds 95% in the United States (2), and U.S. and European guidelines both have Class I recommendations for the use of LITA-LAD (3,4).
The superior graft patency and patient survival with arterial grafts in comparison to saphenous vein grafts in the LAD has led to speculation that arterial grafts may also be superior to vein grafts in other coronary arteries, and there have been numerous studies that have compared outcomes for multivessel disease patients receiving a single LITA-LAD and saphenous vein grafts to patients with multiple arterial grafts (MAGs). These mostly observational studies have concluded that the use of MAG, usually bilateral internal thoracic artery (BITA) grafts, are associated with a significant mortality benefit (5–12), and BITA has become a Class IIa recommendation in patients with a long life expectancy in recent guidelines (3,13).
However, randomized trials that have compared BITA and single internal thoracic artery (SITA) have not identified a survival advantage for BITA (14–18), and a recent meta-analysis of observational studies found that unmeasured confounders may explain the survival advantage of BITA in observational studies (19). The purpose of this study is to use New York State Registry data to compare longer-term (follow-up 6.5 years) outcomes for patients with multivessel coronary artery disease who undergo CABG surgery with a single arterial graft (SAG) with patients who undergo the procedure with MAGs. The study also assesses the extent of selection bias by examining early outcomes. In addition, the study compares outcomes for important patient subgroups and explores variations in surgeon choice of procedure and surgeon outcomes for MAG as a function of MAG volume.
The databases used to conduct the study were New York’s Cardiac Surgery Reporting System (CSRS) and New York’s Vital Statistics file. CSRS was created in 1989 to evaluate and improve the quality of CABG surgery and cardiac valve procedures in New York through the risk adjustment of outcomes and dissemination of reports to hospitals, surgeons, and the public. It contains demographics; numerous pre-procedural patient risk factors; procedural information (off-pump vs. on-pump surgery, total number of grafts used, number of arterial grafts used, and so on), perioperative complications (acute myocardial infarction [AMI], stroke, and so on); hospital and surgeon identifiers; admission, discharge, and procedure dates; and discharge disposition for all CABG and valve surgery performed in nonfederal hospitals in the state. Data from the CSRS are checked annually for completeness and for accuracy by matching the records to the Statewide Planning and Research Cooperative System (SPARCS), New York’s administrative acute care database. The accuracy of risk factors in the system is checked by using New York’s utilization review agent to audit samples of cases from selected hospitals each year.
Vital statistics data were matched to CSRS using unique patient identifiers to obtain deaths that occurred after discharge following the index procedure. SPARCS data were used to obtain AMI and stroke hospitalizations following the index hospitalization. The study was limited to New York State residents to minimize the chance of post-discharge outcomes occurring outside of New York.
Patients and hospitals
All 81,147 patients with multivessel coronary artery disease undergoing isolated CABG surgery between January 1, 2005, and December 31, 2014, in New York State nonfederal hospitals were identified. Patients were then excluded if they did not receive any arterial graft or received only a single conduit during the index surgery (n = 5,750); had an AMI within a day before the index surgery, had unstable hemodynamics, had malignant ventricular arrhythmia, were classified as emergent priority, or experienced an emergency transfer from percutaneous coronary intervention (PCI) or catheterization (n = 4,818); were out-of-state residents (n = 3,147); or had an invalid social security number (n = 5,407). Out-of-state residents were excluded because their deaths or subsequent hospitalizations could not be ascertained using New York Vital Statistics and SPARCS data. The final study sample consisted of 63,402 patients, 50,773 of whom (80%) received SAG and 12,629 of whom (20%) received MAG. The median follow-up period was 6.3 years with an interquartile range (IQR): 3.8 to 8.9 years for the SAG patients and 7.1 with IQR: 4.4 to 9.6 years for the MAG patients (6.5 years with IQR: 4.1 to 9.1 years for all patients combined).
Hospitals in the study include all 42 hospitals in which CABG procedures were performed by 246 surgeons during the period of the study. The median annual surgeon volume for all SAG procedures was 26.8 with an IQR of 9.5 to 46.0, and the median annual surgeon volume for all MAG procedures was 4.9, with an IQR of 2.0 to 12.5. The median annual hospital volume for all MAG procedures was 21.9, with an IQR of 10.8 to 45.1.
Mortality, AMI, stroke, major adverse cardiac events (MACE) (any of the 3 previous outcomes), and repeat revascularization were compared for the 2 procedures at 1 and 7 years following the index procedure. In-hospital sternal wound infection rates were also compared. Variation among surgeons in the use of the MAG versus SAG and the surgeon volume–mortality relationship for MAG procedures were examined.
Because patients were not randomized to MAG versus SAG and because many patient risk factors had large differences in prevalence between the 2 groups of patients, propensity score matching was used to minimize selection bias by identifying a set of MAG/SAG pairs matched on numerous patient risk factors, specific vessels diseased, surgeon volume, completeness of revascularization, number of conduits, total surgeon CABG volume, and type of surgery (off-pump, on-pump). See Table 1 for a complete list.
The propensity score was derived by developing a nonparsimonious logistic regression model that predicted the probability that a given patient would receive MAG based on all risk factors available in the registry. The propensity score was then used to match patients on a 1-to-1 basis so as to minimize the overall distance in propensity scores between the groups (20–23). Patients were matched exactly for a few important factors (year of surgery, off-pump surgery, 3 or more conduits used during the index surgery, complete revascularization [number of total conduits greater or equal to number of diseased vessels], left main trunk disease, 3-vessel disease, 2-vessel disease with right coronary artery involvement, previous AMI within 24 h, previous AMI within 1 to 7 days, left ventricular ejection fraction <50%, diabetes, renal failure [serum creatinine >1.5 mg/dl or dialysis], and age ≥70 years) and were matched with regard to other characteristics as long as their estimated log-odds from the logistic regression model were no more than 0.2 SD apart. Standardized differences in the prevalence of propensity model variables were then calculated (24).
The propensity-matched pairs were then used to analyze differences in mortality, AMI, stroke, MACE, and repeat revascularization rates between the 2 groups. To eliminate further differences within the matched pairs, a Cox proportional hazards model was used to calculate the hazard ratio (HR) for the 2 procedures after adjusting for all risk factors in the propensity score.
A hierarchical logistic regression model with MAG/SAG as the dependent variable was used to examine the extent of differences among surgeons in the choice of procedure. The hierarchical model was used because the independent variables being examined were at 2 different levels: patient and surgeon. The first level of the model contained all significant patient-level predictors of MAG, including demographics, coronary disease severity, and comorbidities. The individual surgeons were the only variables used at the second level of the model. The relationship between MAG mortality and total annual MAG provider (surgeon and hospital) volume was also tested. All tests were 2-sided and conducted at the 0.05 level, and all analyses were conducted in SAS version 9.2 software (SAS Institute, Cary, North Carolina).
Table 1 presents the frequencies for various baseline characteristics describing operator CABG volume, surgery type, patient demographics, coronary artery disease severity, and comorbidities. Many characteristics demonstrated significant differences, with MAG patients being younger; being more likely to be male; having lower prevalences of diabetes, cerebrovascular disease, COPD, and renal dialysis; having higher body surface area; and having higher left ventricular ejection fractions. As indicated, 12,629 of 63,402 procedures (19.9%) were MAG procedures.
The median number of grafts/patient was 3 (IQR: 3 to 4). The median percentage of grafts that were arterial grafts was 50% (IQR: 33% to 67%). Off-pump surgery was associated with a higher rate of incomplete revascularization (26% vs. 21%).
The propensity-matching process resulted in a total of 10,828 matched pairs of patients. Table 2 demonstrates that the propensity-matched pairs are very similar with respect to risk factor prevalences, with no patient characteristics having a standardized difference larger than 10%. Online Table 1 presents the propensity model that predicts the choice of MAG, and Online Figure 1 shows propensity-matched and unmatched cases as a function of the estimated probability of MAG being chosen. As indicated, a large percentage of unmatched cases occur for SAG patients whose propensity to receive MAG is ≤0.1.
MAG patients had a higher rate of in-hospital sternal wound infections (1.01 vs. 0.62; p = 0.002). Table 3, the Central Illustration, and Figures 1 and 2 present comparative outcomes for propensity-matched MAG and SAG patients at 1 and 7 years. As demonstrated in Table 3 and the Central Illustration, 1-year mortality was not different for MAG compared with SAG (2.4% vs. 2.2%; adjusted hazard ratio [AHR]: 1.11; 95% confidence interval [CI]: 0.93 to 1.32). AMI, stroke, and MACE rates were not different at 1 year, but MAG patients had lower repeat revascularization rates.
Online Table 2 presents significant risk factors for mortality of all patients at any time during the course of the study. For propensity-matched patients at 7 years, mortality was significantly lower in MAG patients (12.7% vs. 14.3%; AHR: 0.86; 95% CI: 0.79 to 0.93), see Table 3 and the Central Illustration. Similar results occurred when limiting surgeon CABG volume to at least 50 procedures/year (AHR: 0.85; 95% CI: 0.78 to 0.93). MACE, AMI, and repeat revascularization rates were also lower (20.2% vs. 22.8%, AHR: 0.88; 95% CI: 0.83 to 0.93) for MACE, (5.2% vs. 6.2%, AHR: 0.85; 95% CI: 0.75 to 0.95) for AMI, and 11.7% vs. 14.6%, AHR: 0.80; 95% CI: 0.74 to 0.87) for repeat revascularization (Table 3, Figures 1 and 2). Of the 2,303 propensity-matched patients who underwent subsequent PCI, a total of 4,146 lesions were treated; 73% were in native coronary arteries that were grafted in the index CABG procedure, 20% in the grafts themselves, and 7% in native, previously untreated coronary arteries. Of the 49 patients who underwent repeat CABG surgery, 12 of the procedures were in previously untreated arteries and 37 were in previously treated arteries.
Table 4 presents the 7-year mortality for selected subgroups after further adjustment of differences in matched pairs using Cox regression analyses. Subgroups of patients for whom MAG patients did not have lower mortality were patients who were ≥70 years of age, had off-pump surgery, 2-vessel disease with right coronary artery disease, an AMI within 20 days before surgery, and renal dysfunction.
After excluding cases that did not undergo complete revascularization during the index hospitalization, 7-year mortality rates for MAG and SAG patients remained significantly different for the 2 arms (11.5% vs. 13.0%, AHR: 0.87; 95% CI: 0.79 to 0.96).
Surgeon and hospital volume
The surgeon and hospital volumes for MAG procedures associated with each patient were calculated as the average of the total annual MAG procedures performed by the patient’s surgeon in the study period. The MAG surgeon volumes for all patients ranged from 0 to 67 for the study period, with an IQR from 1 to 10. The intraclass correlation coefficient was 0.49, 95% CI: 0.44 to 0.54, and the median odds ratio was 5.54, 95% CI: 4.60 to 6.56, indicating that after accounting for significant patient risk factors, there was substantial variation in surgeon choice of procedure. The highest surgeon volume tertile (≥10) of MAG procedures had significantly lower 7-year mortality than the other tertile when restricted to MAG patients only (AHR for highest volume tertile versus other tertiles: 0.86; 95% CI: 0.75 to 0.99; p = 0.046). The hospital average annual MAG volumes ranged from 0 to 149 for the study period, with an interquartile range from 10 to 60. The hospitals with average annual MAG volume ≥50 did not have significantly lower 7-year mortality than other hospitals when restricted to MAG patients only (AHR for hospital volume ≥50 versus hospital volume <50: 0.93; 95% CI: 0.82 to 1.06; p = 0.25).
There is considerable debate about the use of multiple arterial grafts for patients with multi-vessel coronary artery disease. Most observational studies have found survival benefits for BITA grafts in comparison to patients receiving a SITA graft, and current guidelines have a Class IIa recommendation for the use of BITA in patients with a long life expectancy (3,13). However, recent randomized controlled trials (RCTs) do not confirm these survival benefits (14–18). The controversy is further complicated by the fact that most studies compare SITA and BITA, whereas radial artery (RA) grafts have been shown to be good alternatives for second artery grafts. Consequently, a comparison of MAG versus SAG outcomes may be more relevant. Our study found that at 1 year, there were no differences between the MAG and SAG patients in mortality, MACE, or stroke. This suggests that at least some of the inevitable selection bias that is introduced by using an observational study has been controlled for by using propensity score matching. These results are different from the findings of Gaudino et al. (19), who reported that in a meta-analysis of 12 recent propensity score–matched observational studies, the mortality reduction for BITA compared with SITA was significant at 1 year, and the mortality reduction at 1 year (incidence rate ratio: 0.70; 95% CI: 0.60 to 0.82) was very similar to what it was at the end of the follow-up period (incidence rate ratio: 0.77; 95% CI: 0.70 to 0.85). We hypothesize that our propensity matching may have been more successful than earlier observational studies because we used several additional variables in the matching process, including completeness of revascularization, number of conduits used, and specific vessels diseased. Importantly, we also matched on surgeon volume groups to minimize any bias caused by higher-volume surgeons with lower risk-adjusted mortality rates choosing 1 of the procedures (MAG or SAG) more frequently. Also, our study was >10 times larger than earlier observational studies.
At 7 years (median follow-up 6.5 years), our study found a difference in favor of MAG in mortality (12.7% vs. 14.3%, AHR: 0.86; 95% CI: 0.79 to 0.93, and in MACE (20.2% vs. 22.8%, AHR: 0.88; 95% CI: 0.83 to 0.93). The largest and most recent RCT to compare our results to is the ART (Arterial Revascularization Trial), which compared 10-year outcomes for BITA and SITA (18). The ART study concluded that in an intention-to-treat analysis at 10 years, there were no differences in mortality between BITA (n = 1,548) and SITA (n = 1,554), 20.3% versus 21.2%, HR: 0.96; 95% CI: 0.82 to 1.12. Also, there was no significant difference for MACE (death, AMI, or stroke), 24.9% versus 27.3%, HR: 0.90; 95% CI: 0.79 to 1.03. Although our 7-year follow-up findings disagreed with the MACE findings of the ART study, it is notable that the 2 studies had almost identical HRs (0.90 and 0.88), but our study reached statistical significance with a sample size >3 times as high. In addition to the unmeasured confounders in our observational study, there are other reasons why our findings may be different from the ART findings. First, in the SITA group in the ART study, 21.8% of the patients also received a RA graft, which has been demonstrated to be superior to saphenous vein grafts (8,24,25). As noted by the ART authors, this may be a key confounder because it is likely to benefit the SITA group by adding an arterial graft to the second most important coronary artery (18). In fact, when the ART analyses were redone using actual receipt of 1 versus 2 or more arterial grafts, there was a survival benefit identified for the latter group (18.6% vs. 23.1%, HR: 0.81; 95% CI: 0.68 to 0.95). In our study, which could not distinguish between RA and internal thoracic artery grafts, these patients were included in the MAG group. Thus, the significantly lower mortality for MAG found in our study corresponds to the findings of the ART study with actual treatment rather than intention to treat. As noted in the ART study, the ongoing ROMA (Randomized Comparison of the Clinical Outcome of Single Versus Multiple Arterial Grafts) trial will compare SAGs with MAGs of any type, so patients with a LITA and a radial graft will be assigned to the MAG group (26).
Second, 13.9% of the ART patients in the (intention-to-treat) BITA group underwent SITA although assigned to the BITA group, and these patients (who presumably experienced worse outcomes) were assigned to the SAG group in our study because we had no information about intention to treat (18). Third, compliance with guideline-directed medical therapy is likely to narrow the outcomes difference between the 2 groups, and the ART study experienced a very high compliance rate (18). Although no data of this type were available in our study, it is unlikely that it was as high. Fourth, approximately 40% of procedures in the ART study were performed off pump, compared with 18% in our study. However, MAG and SAG patients undergoing off-pump surgery had almost identical risk-adjusted mortality in our study (AHR: 0.98; 95% CI: 0.82 to 1.17), so if the percentage of off-pump patients in our study had been as high as in the ART study, it is very possible the MAG patients would not have experienced a lower mortality.
With respect to subgroups of interest in our study, we found that the mortality advantage of MAGs was limited to patients who are younger than 70 years of age. We hypothesize that the multiple comorbidity burden that is more common in the elderly as well as the impact of a prolonged anesthesia pre-bypass time are contributing factors in this finding. This is an important finding in that it should serve to focus subsequent RCT substudies on the impact of age on optimal choice of procedure. The group of patients with 2-vessel disease including right coronary artery involvement (i.e., the absence of both LAD and left circumflex coronary artery involvement) was investigated because this is a group of patients that may have received a LITA and a RA in the MAG group (i.e., would not have received both a LITA and right internal thoracic artery). The results show that there is no mortality benefit of MAG for this group of patients. Instead, the mortality benefit is restricted to patients with both the LAD and left circumflex coronary arteries diseased. This is another finding that may be worth confirming in subsequent RCTs. As noted above, the mortality advantage for MAG patients was limited to on-pump patients, and this highlights another subgroup that should be examined in subsequent RCTs.
Another finding of the study was that surgeon volume of MAG procedures was significantly related to mortality for MAG procedures. Also, despite generally low surgeon volumes of MAG procedures, there was considerable intersurgeon variation in the use of MAGs.
First and foremost, it is an observational study and the lack of randomization means that patients in the SAG group were undoubtedly sicker than their MAG counterparts. We attempted to minimize this selection bias by propensity matching patients and we found that after matching, patients in the 2 groups had very similar profiles with respect to the numerous risk factors available in our registry. It is also a very good sign that unlike earlier observational studies on this topic (19), the 1-year mortality, MACE, AMI, and stroke rates were nearly identical for the 2 groups, implying that no differences occurred at a time too early to be attributable to saphenous vein graft failure. Nevertheless, bias related to variables not contained in our registry could not be removed. Some of the unmeasurable confounders include general health and life expectancy, access to care, long-term follow-up, and graftability of the target vessels (19).
Also, we could not distinguish between internal thoracic artery grafts and RA grafts because this information was not available in the registry during the study period. This is of interest because some earlier studies have demonstrated superior outcomes for internal thoracic artery grafts compared with RA grafts, and it would have been interesting to compare outcomes among MAG patients as a function of the type of arterial grafts they received (21). We did not have access to the harvesting technique, and this information may have impacted relative sternal wound infection rates and graft patency. Our definition of complete revascularization may also be inaccurate in certain patients because we cannot match grafting to the number of critical lesions. It is possible some vessels with critical lesions could not be grafted, leading surgeons to graft less severely diseased vessels in another territory. In addition, we did not have information about the location of sequential grafts, which were probably more frequent in the SAG group and may have possibly affected relative outcomes. We were unable to identify crossovers from MAG to SAG. Many of these patients, who were attributed to SAG, experienced intraoperative difficulties, and probably had worse longer-term outcomes. Also, secondary medical prevention may have been better in the MAG group, and the imperfect definition of complete revascularization may have favored the MAG group. The combination of these caveats could explain the modest survival and MACE advantage that we found for the MAG group.
Our findings show a 7-year mortality and MACE advantage for MAG patients, and although these findings are at variance with the findings from a recent RCT, the reason for this may be strongly related to differences in patient mix and in assignment of patients to study groups (MAG vs. SAG rather than BITA vs. SITA). Also, our study has identified a few hypotheses to examine in future RCTs, including outcome differences between MAG and SAG patients as a function of off-pump versus on-pump surgery, age, specific vessels diseased, and type of arterial grafts used.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: In selected patients with multivessel disease undergoing coronary bypass surgery, use of MAGs is associated with better intermediate-term outcomes than SAGs.
TRANSLATIONAL OUTLOOK: Randomized studies should compare outcomes for older patients and those with other specific characteristics undergoing complete or incomplete revascularization or off-pump surgery.
The authors thank Kimberly S. Cozzens, Rosemary Lombardo, and the cardiac surgery departments of the participating hospitals for their tireless efforts to ensure the timeliness, completeness, and accuracy of the registry data.
Dr. Sundt has served on a clinical events committee for Medpace. Dr. Wechsler has received consulting fees from Bioventrix. Dr. Chikwe has received speakers honoraria from Edwards Lifesciences. Dr. Adams has served as co-principal investigator on trials for Medtronic and Abbott; and his institution receives royalties for intellectual property related to the development of valve repair products from Edwards Lifesciences and Medtronic. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Listen to this manuscript's audio summary by Editor-in-Chief Dr. Valentin Fuster on JACC.org.
- Abbreviations and Acronyms
- adjusted hazard ratio
- acute myocardial infarction
- bilateral internal thoracic artery
- coronary artery bypass graft
- confidence interval
- Cardiac Surgery Reporting System
- hazard ratio
- interquartile range
- left anterior descending coronary artery
- left internal thoracic artery
- major adverse cardiac events
- multiple arterial graft
- percutaneous coronary intervention
- radial artery
- randomized controlled trial
- single arterial graft
- single internal thoracic artery
- Statewide Planning and Research Cooperative System
- Received June 4, 2019.
- Revision received June 20, 2019.
- Accepted June 21, 2019.
- 2019 American College of Cardiology Foundation
- ↵New York State Department of Health. Adult Cardiac Surgery in New York State: 2014-2016. January 2019. Available at: https://www.health.ny.gov/statistics/diseases/cardiovascular/heart_disease/docs/2014-2016_adult_cardiac_surgery.pdf. Accessed June 1, 2019.
- Hillis L.D.,
- Smith P.K.,
- Anderson J.L.,
- et al.
- Yi G.,
- Shine B.,
- Rehman S.M.,
- Altman D.G.,
- Taggart D.P.
- Buttar S.N.,
- Yan T.D.,
- Taggart D.P.,
- Tian D.H.
- Schwann T.A.,
- Habib R.H.,
- Wallace A.,
- et al.
- Royse A.G.,
- Brennan A.P.,
- Ou-Young J.,
- et al.
- Taggart D.P.,
- Altman D.G.,
- Gray A.M.,
- et al.
- Taggart D.P.,
- Benedetto U.,
- Gerry S.,
- et al.
- Gaudino M.,
- DiFranco A.,
- Rahouma M.,
- et al.
- Rosenbaum P.R.
- Gaudino M.,
- Benedetto U.,
- Fremes S.,
- et al.
- Gaudino M.,
- Alexander J.H.,
- Bakaeen P.G.,
- et al.