Hybrid Coronary Revascularization Versus Off-Pump Coronary Artery Bypass Grafting for the Treatment of Multivessel Coronary Artery Disease

Hybrid Coronary Revascularization Versus Off-Pump Coronary Artery Bypass Grafting for the Treatment of Multivessel Coronary Artery Disease Michael E. Halkos, MD, Thomas A. Vassiliades, MD, MBA, John S. Douglas, MD, Douglas C. Morris, MD, S. Tanveer Rab, MD, Henry A. Liberman, MD, Habib Samady, MD, Patrick D. Kilgo, MS, Robert A. Guyton, MD, and John D. Puskas, MD Division of Cardiothoracic Surgery and the Division of Cardiology, Emory University School of Medicine, Atlanta, Georgia ADULT CARDIAC Background. Hybrid coronary revascularization () combines a minimally invasive (3-cm anterolateral thoracotomy), sternal-sparing, off-pump left internal mammary artery left anterior descending (LIMA-LAD) coronary artery anastomosis with percutaneous coronary intervention (PCI) to non-lad coronary arteries. We compared outcomes of versus traditional off-pump coronary artery bypass grafting () for the treatment of multivessel coronary artery disease (CAD). Methods. Between October 8, 2003 and April 23, 2010, 147 patients with multivessel coronary disease were treated with at a US academic center. These were matched 4:1 to 588 contemporaneous patients treated with multivessel by sternotomy using an optimal matching algorithm with 8 preoperative variables: age, gender, ejection fraction, presence of diabetes, myocardial infarction (MI), number of diseased vessels, left main coronary artery disease, and Society of Thoracic Surgeons (STS) predicted risk of mortality (PROM) score. In-hospital major adverse events (MACCE) and the need for repeated revascularization during followup were compared between groups. All-cause mortality was determined using the Social Security Death Index (SSDI). Results. Matching produced groups with similar coronary anatomy and statistically similar preoperative risk factors. The incidence of MACCE was similar between groups (3/147 versus 12/588 ). During a median 3.2 years of follow up, the need for repeated revascularization was higher for than for (18/147 [12.2%] versus 22/588 [3.7%]; p < 0.001). The incidence of blood transfusion was higher for the group. Estimated 5-year survival was similar between groups (, 84.3% versus, 86.8%; p 0.61). Conclusions. Hybrid coronary revascularization is a minimally invasive treatment for multivessel CAD. Although repeated revascularization was greater with, both in-hospital and midterm outcomes were comparable with those of traditional. Further investigation into the comparative effectiveness of this alternative strategy is warranted. (Ann Thorac Surg 2011;92:1695 1702) 2011 by The Society of Thoracic Surgeons Hybrid coronary revascularization () is an alternative coronary revascularization strategy that combines a minimally invasive, sternal-sparing left internal mammary artery left anterior descending coronary artery bypass (LIMA-LAD) with percutaneous coronary intervention (PCI) to non-lad coronary lesions. Although historically this procedure was offered to patients felt to be at high risk for complications of traditional coronary artery bypass grafting (CABG) by median sternotomy, growing interest in this treatment for patients with multivessel coronary artery disease (CAD) has been Accepted for publication May 24, 2011. Presented at the Forty-seventh Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31 Feb 2, 2011. Address correspondence to Dr Halkos, Division of Cardiothoracic Surgery, Emory University School of Medicine, Emory University Hospital Midtown, 550 Peachtree St NE, 6th flr, Medical Office Tower, Atlanta, GA 30308; e-mail: mhalkos@emory.edu. fueled by patients and cardiologists desire for less invasive treatment options. CABG by median sternotomy provides results superior to multivessel PCI with drug-eluting stents (DES) for patients with severe multivessel CAD [1]. However these results were largely driven by the reduced need for repeated revascularization in the CABG cohort, with 12-month mortality and myocardial infarction (MI) rates similar between CABG and PCI groups. Multivessel stenting is now routinely performed in most centers in the United States given the lower restenosis rates with DES compared with bare metal stents [2]. The demonstrated survival benefit of CABG is largely Dr Halkos discloses that he has a financial relationship with Intuitive Surgical; Dr Vassiliades with Medtronic. 2011 by The Society of Thoracic Surgeons 0003-4975/$36.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2011.05.090

ADULT CARDIAC 1696 HALKOS ET AL Ann Thorac Surg MULTIVESSEL CORONARY ARTERY DISEASE TREATMENT 2011;92:1695 1702 attributable to the LIMA-LAD bypass, which provides excellent long-term durability [3]. Furthermore, PCI with DES to non-lad vessels may provide results similar to those of saphenous vein grafts (SVGs), although a direct comparison has not been performed. combines the durability of a LIMA-LAD graft by a minimally invasive approach with the less invasive PCI for non-lad lesions. Currently only 5 studies exist in the literature comparing with traditional CABG [4 8]. The purpose of this study was to compare clinical outcomes of patients undergoing with a matched cohort of patients undergoing traditional off-pump coronary artery bypass grafting () by sternotomy. Material and Methods This study used a matched cohort design to examine short-term and midterm outcomes between patients undergoing competing surgical approaches: and hybrid revascularization. Hybrid procedures were defined as those in which a minimally invasive LIMA-LAD bypass was performed in combination with PCI to at least 1 major non-lad coronary artery (right coronary artery or circumflex artery, or both). In all, 147 hybrid cases were identified querying the Emory University institutional Society of Thoracic Surgeons (STS) Adult Cardiac Database for all eligible cases during the entire period that has been performed, October 8, 2003 to April 23, 2010. The study was approved by the Institutional Review Board at Emory University. Each patient undergoing was matched with 4 patients undergoing using an optimal matching algorithm [9]. This algorithm sequentially matched each record of patients undergoing with potential controls by calculating the multivariate distance between the patients based on 8 variables that were available preoperatively. The algorithm chooses the set of matches that minimizes the sum of the multivariate distances across all possible sets of matches. The matching variables were chosen to include important risk factors for adverse outcomes, including STS predicted risk of mortality (PROM), patient age, gender, ejection fraction, presence of diabetes, myocardial infarction (MI) within 7 days of the operation, number of diseased vessels, and left main CAD with greater than 50% of vessel diameter compromised. STS PROM, which is a function of more than 30 preoperative variables, was weighted twice as high as any other matching factor because it is a conglomerate of each patient s entire risk profile. All preoperative variables were collected and compared between groups to assess comorbidity imbalances across the surgical groups. Indications and Contraindications for The relative indications for a hybrid approach include the presence of proximal or mid-lad stenosis amenable to minimally invasive LIMA-LAD bypass and the presence of non-lad lesions that are amenable to PCI. In general patients with atherosclerotic coronary disease felt to be high risk for PCI or those with a lower likelihood of procedural success or complete revascularization (chronic total occlusions, bifurcation lesions) were generally treated surgically with traditional. The decision to proceed with was made only after careful discussion between the surgeon, interventional cardiologist, referring cardiologist, and the patient. Patients were considered eligible for only if the surgeon thought that an equivalent LIMA-LAD bypass could be formed through a minimally invasive approach and if the interventional cardiologist could perform PCI to non-lad vessels with excellent technical success. Relative contraindications for included a nongraftable LAD, a hemodynamically unstable patient, previous sternotomy or left thoracotomy, severe pulmonary disease with inability to tolerate single-lung ventilation, body mass index greater than 40, and non-lad disease not felt to be successfully treated with PCI. This included long lesions requiring multiple stents, small-diameter vessels not amenable to PCI, bifurcation lesions, chronic total occlusions, lesions previously treated with PCI, or other more technically challenging lesions. Procedures Before 2009, the surgical component of the hybrid procedure was performed with the endoscopic atraumatic coronary artery bypass (EndoACAB) procedure, which has been described in detail [10]. Briefly this involves a thoracoscopic LIMA harvest through 3 small ports (1 camera, 2 operating ports). After LIMA harvest, the pericardium is opened longitudinally to expose the LAD. A spinal needle is passed through the anterior chest wall visualized endoscopically to identify the planned site for a tailored small anterolateral 3-cm to 4-cm thoracotomy incision to be made in the appropriate interspace. A soft tissue retractor is used to provide exposure over the LAD without rib spreading, and the anastomosis is performed manually using 8-0 polypropylene suture without cardiopulmonary bypass. Since 2009 the majority of hybrid procedures have been performed with robotic assistance. The da Vinci Surgical System (Intuitive Surgical Inc, Sunnyvale, CA) is used to harvest the LIMA, open the pericardium, and identify the optimal target site on the LAD. The remainder of the procedure is nearly identical to the EndoACAB procedure. For procedures, the order and timing of the surgical and percutaneous interventional procedures are determined by the patient s coronary anatomy and is a joint decision between the surgeon and the interventional cardiologist. In general the LIMA-LAD bypass is performed first to minimize complications associated with antiplatelet and anticoagulation medication. However in patients with critical coronary lesions, it is our policy to treat the culprit lesion first. Similarly, most patients undergo both procedures during the index hospitalization to ensure complete revascularization. The majority of patients in this study underwent a staged hybrid procedure, usually within 2 to 3 days of each other. Fewer than 10 patients in this study underwent a simultaneous hybrid procedure. The majority of patients

Ann Thorac Surg HALKOS ET AL 2011;92:1695 1702 MULTIVESSEL CORONARY ARTERY DISEASE TREATMENT received drug-eluting stents (DES) unless contraindications existed for long-term clopidogrel therapy. For patients who underwent the PCI portion of the procedure first, 600 mg of clopidogrel was given before PCI. These patients underwent the surgical portion of the procedure with ongoing clopidogrel therapy (75 mg/day). For patients who underwent the surgical portion of the procedure first (the majority) or those who underwent traditional, 150 mg of clopidogrel was administered 4 hours after the procedure if chest tube drainage was less than 100 ml/hour for 4 hours. These patients were then maintained on 75 mg clopidogrel/day for at least 6 weeks. An additional loading dose (300 mg) was given to patients undergoing at the time of the PCI procedure. All patients received preoperative and postoperative aspirin therapy. Outcomes Outcomes of interest included several postoperative inhospital end points: death, permanent stroke, MI, major adverse cardiac and cerebrovascular events (MACCE, the composite of death, stroke, and MI), renal failure, need for blood transfusion, length of stay from operation to discharge, atrial fibrillation, number of hours on a ventilator, and renal failure. Midterm outcomes included survival time, and time to reintervention, either by CABG or PCI. A medical records query was completed whereby all subsequent reinterventions (PCI or CABG) at Emory University hospitals were assessed and dates of the procedures were collected. The time between the patient s index surgery and any reintervention was calculated as a study end point. Repeated revascularization was defined as target vessel revascularization if the need for repeated revascularization involved a coronary artery initially treated with either bypass grafting or PCI. Repeated revascularization was defined as progression of native disease if repeated revascularization involved a native coronary artery that was not previously treated by either PCI or surgical revascularization. Patients who underwent with the LIMA-LAD bypass performed first had LIMA-LAD patency confirmed before proceeding with PCI. Thus angiographically detected defects of the LIMA-LAD anastomosis in patients who underwent may have 1697 undergone intervention regardless of clinical indications. PCI of the LIMA or LIMA anastomosis was performed after a joint decision between the cardiologist and surgeon if angiographic defects were detected during the subsequent PCI procedure. These were classified as repeated reinterventions even if there were no clinical indications for intervention. Midterm survival was determined by querying the Social Security Death Index (SSDI) to determine dates of death for all patients in the study. The SSDI database sensitivity (92.2%) is comparable to the National Death Index sensitivity [11]. Death dates were gathered from SSDI on March 14th, 2010; any patient alive on that date was considered a censored observation in survival analyses. Statistical Analysis Univariate analyses were performed to determine whether good matches were made between each case and the 4 controls using paired tests. Means were compared using a paired t test and frequencies were analyzed using McNemar s test for paired proportions. The means and proportions for each group are shown in Table 1. Similarly, postoperative outcomes were compared using these same paired tests. Also for descriptive purposes, standardized mean differences were calculated for each matching variable to demonstrate the relative distance between the hybrid means/ proportions and the means/proportions [12]. Time-to-event analyses (death, PCI, CABG, or reintervention) were conducted using Kaplan-Meier productlimit estimation. Survival curves were generated to compare the failure course of the groups for each end point. In the case of imbalance across groups with respect to any matching factor, Cox proportional hazards regression was used to adjust for the effect of the factor. Crude (unadjusted) and adjusted hazard ratios with 95% confidence intervals were compared to assess whether the imbalanced factor affected the survival estimates of the groups. The proportional hazards assumption was assessed by examining the correlation between Schoenfeld residuals and survival time among uncensored patients. The analysis was performed using SAS version 9.2 (SAS ADULT CARDIAC Table 1. Demographics of Matched Variables (N 147) Standardized Difference p Value Age (mean SD) 64.3 12.5 64.3 12.8 0.0% 0.99 Females (%) 168 (28.6) 56 (38.1) 20.2% 0.025 Diabetes mellitus (%) 209 (35.5) 58 (39.5) 8.3% 0.38 Ejection fraction (mean SD) 54.7 8.7 54.6 8.7 1.2% 0.91 STS PROM (mean SD) 0.018 0.021 0.020 0.023 9.1% 0.35 No. diseased vessels (mean SD) 2.5 0.6 2.4 0.6 16.7% 0.07 Left main coronary artery disease 50% stenosis (%) 111 (18.9) 27 (18.4) 1.3% 0.89 Myocardial infarction 7 days (%) 73 (12.4) 20 (13.6) 3.6% 0.7 hybrid coronary revascularization; off-pump coronary artery bypass grafting; STS PROM Society of Thoracic Surgeons predicted risk of mortality.

ADULT CARDIAC 1698 HALKOS ET AL Ann Thorac Surg MULTIVESSEL CORONARY ARTERY DISEASE TREATMENT 2011;92:1695 1702 Institute, Cary, NC). All comparisons were made at the 0.05 alpha level. Results A total of 147 patients underwent hybrid coronary revascularization combining a minimally invasive LIMA-LAD bypass with PCI of non-lad coronary vessels. These patients were matched 4:1 with contemporaneous patients based on 8 preoperative variables. Patients were well matched with no significant differences with the exception of female gender, with more female patients in the hybrid group (20.2% standardized difference) (Table 1). According to unmatched variables there were no significant differences between groups (Table 2). The use of the LIMA in both groups was greater than 99% (Table 3). The primary objective of the study was to examine the incidence of in-hospital MACCE and repeated revascularization events (Tables 4 and 5). The MACCE rate was the same between and (2.0%). There were 5 (0.9%) in-hospital deaths in the group compared with 1 (0.7%) in-hospital death in the group (p 0.72). The need for blood transfusion was significantly greater in the group compared with the group (56.0% versus 35.4%; p 0.001). Other in-hospital outcomes were similar between groups (Table 3). The need for repeated revascularization was greater in the group compared with the group (Table 5). Eighteen patients (12.2%) in the group underwent reintervention compared with 22 (3.7%) in the group (p 0.001). The median time of reintervention (operation or PCI) from the index procedure was 3.2 years. Of these patients, 13 (8.8%) in the group and 18 (3.1%) in the group underwent target vessel revascularization (p 0.002). Seven patients (4.8%) in the group and 5 (0.9%) patients in the Table 3. Perioperative Variables group underwent reintervention for progression of native disease in an artery not previously treated by operation or PCI (p 0.001). Details of reintervention are listed in Table 5. Seven (4.8%) patients underwent PCI of the LIMA or LIMA-LAD anastomosis in the group compared with 6 (1.0%) patients in the group. Only 1 of the 7 LIMA interventions was performed during the index hospitalization for an angiographically detected defect. There were 14 (2.4%) SVG interventions in the group and 5 (3.4%) interventions for in-stent restenosis in the group. The secondary end point for this study was midterm survival. Midterm survival was similar between groups, with estimated 5-year survival of 84.3% for the group and 86.8% for the group (Fig 1). Comment (N 147) p Value Arterial grafts (mean SD) 1.36 0.70 1.07 0.29 0.001 Vein grafts (mean SD) 1.67 1.06 0.08 0.46 0.001 Total grafts (mean SD) 3.02 1.07 1.16 0.55 0.001 LIMA or BIMA Used (%) 585 (99.5) 147 (100) 0.78 BIMA bilateral internal mammary artery; hybrid coronary revascularization; LIMA left internal mammary artery; off-pump coronary artery bypass grafting. To our knowledge, the current study represents 1 of the largest comparisons of to conventional CABG. In this study, patients undergoing had similar inhospital mortality and midterm survival rates as those undergoing. With the exception of a lower incidence of transfusion in the group, other in-hospital Table 2. Demographics of Unmatched Variables Demographics of Unmatched Variables (N 147) p Value White (%) 463 (79.0) 113 (76.9) 0.57 Renal insufficiency (%) 29 (4.9) 12 (8.2) 0.13 Renal failure/dialysis 10 (1.7) 6 (4.1) 0.08 dependent (%) Cerebrovascular disease (%) 98 (16.7) 29 (19.7) 0.38 Previous stroke (%) 58 (9.9) 18 (12.2) 0.40 NYHA class IV (%) 82 (20.3) 17 (15.5) 0.26 Current smoker (%) 296 (50.3) 62 (42.4) 0.08 Congestive heart failure (%) 100 (17.0) 20 (13.6) 0.32 Last creatinine value 1.19 1.10 1.44 1.65 0.09 (mean SD) Hypertension (%) 499 (84.9) 128 (87.1) 0.50 Chronic lung disease (%) 69 (13.4) 15 (10.2) 0.33 Peripheral vascular disease (%) 88 (15.0) 17 (11.6) 0.29 hybrid coronary revascularization; NYHA New York Heart Association; off-pump coronary artery bypass grafting. Table 4. Hospital Outcomes Hospital Outcomes (N 147) p Value Death (%) 5 (0.9) 1 (0.7) 0.84 Stroke (%) 4 (0.7) 1 (0.7) 1 MI (%) 3 (0.5) 1 (0.7) 0.8 MACCE (death stroke, 12 (2.0) 3 (2.0) 1 MI) (%) Blood transfusion (%) 329 (56.0) 52 (35.4) 0.001 Renal failure (%) 15 (2.6) 4 (2.7) 0.91 Hospital length of stay 6.1 4.7 6.6 6.7 0.48 (mean SD) Atrial fibrillation (%) 109 (18.5) 29 (20.1) 0.63 Hours on ventilator 22.7 89.5 17.0 30.8 0.28 (mean SD) Intensive care unit length of stay (mean SD) 52.7 87.8 57.4 145 0.70 hybrid coronary revascularization; MACCE major adverse cardiac and cerebrovascular events; MI myocardial infarction; off-pump coronary artery bypass grafting.

Ann Thorac Surg HALKOS ET AL 2011;92:1695 1702 MULTIVESSEL CORONARY ARTERY DISEASE TREATMENT Table 5. Details of Repeated Revascularization Events Repeated Revascularization Events (N 147) p Value All repeat revascularization 22 (3.7) 18 (12.2) 0.001 events (%) PCI (%) 21 (3.6) 16 (10.9) 0.001 CABG (%) 1 (0.2) 2 (1.4) 0.043 Target vessel 18 (3.1) 13 (8.8) 0.002 revascularization (%) Progression of native 5 (0.9) a 7 (4.8) 0.001 disease (%) Lesion in IMA or 6 (1.0%) 7 (4.8%) 0.001 IMA-LAD (%) Occlusion or stenosis of 14 (2.4) 0 (0.0) 0.06 SVG (%) In-stent restenosis (%) 0 (0.0) 5 (3.4) 0.001 a One patient in with both target vessel revascularization and progression of native disease in previously untreated vessel. CABG coronary artery bypass grafting; hybrid coronary revascularization; IMA internal mammary artery; LAD left anterior descending coronary artery; off-pump coronary artery bypass grafting; PCI percutaneous coronary intervention; SVG saphenous vein graft. 1699 outcomes were similar between groups. The incidence of repeated revascularization was higher in the group compared with the group. In smaller studies, both Kon and colleagues [6] and Reicher and associates [7] reported similar in-hospital outcomes between patients undergoing and those undergoing. In addition, Zhao and colleagues [5] reported their outcomes on 67 patients undergoing planned hybrid procedures, with clinical outcomes comparable to those undergoing traditional CABG. We have previously reported a reintervention rate for of 5.5% at 1 year [4]. In the current study with longer follow-up, the reintervention rate for increased to 12.2% (18 patients). In a smaller study by Reicher and colleagues [7] (39 patients), repeated revascularization rates were similar between and. has emerged as a viable alternative to both traditional CABG and multivessel PCI to provide complete revascularization for patients with multivessel CAD. As patients and referring cardiologists have sought less invasive treatment options for CAD, offers the advantages of both surgical and interventional treatments the durability and survival advantage of the LIMA-LAD bypass as well as the less invasive option of PCI. There are several well-described approaches for minimally invasive CABG, including robotic-assisted [13], endoscopic [10], minimally invasive direct (MID) CABG [6, 7], and totally endoscopic using robotic technology [14, 15]. Furthermore, the timing and order of procedures has also been reported, with excellent results for both simultaneous (same day [16] or 1 stop [5]) and staged procedures [6, 7, 10, 17 19]. The majority of our cases have been staged, with operations performed first when possible. It is our general policy to proceed with the operation first unless the culprit lesion involves a non- LAD coronary artery. In this case, PCI of the right coronary artery or circumflex artery, or both, is performed first, with the LIMA-LAD done 2 to 3 days after PCI. Performing the surgical session first allows for the operation to proceed without the risks of antiplatelet therapy needed for PCI. Furthermore, the logistic issue of coordinating 2 separate teams at the same time has hampered our ability to proceed with simultaneous or 1-stop. In centers with busy operating rooms, simultaneous PCI in a hybrid room occupies cardiac anesthesia and nursing personnel and may prevent or delay other surgical or interventional cases from proceeding as scheduled. Nonetheless we are currently exploring options to facilitate more same-day procedures at Emory. An important survival advantage of CABG is conferred by the routine use of the LIMA to bypass the LAD [3, 20]. Because of the higher associated restenosis rates with ADULT CARDIAC Fig 1. Kaplan-Meier 5-year survival estimates according to hybrid coronary revascularization () or off-pump coronary artery bypass grafting () (p 0.61).

ADULT CARDIAC 1700 HALKOS ET AL Ann Thorac Surg MULTIVESSEL CORONARY ARTERY DISEASE TREATMENT 2011;92:1695 1702 PCI of the LAD, surgeons and interventionalists generally agree that the LIMA is the most durable and effective treatment for proximal LAD disease. The reported incidence of SVG failure [21 23] and the lower restenosis rates with DES have made the optimal treatment of non-lad vessels (DES versus SVG) in the context of 3-vessel CAD more controversial [8, 24]. Thus combining positive features of both procedures with a hybrid approach is the impetus for collaboration between surgeons and cardiologists. In large observational studies by Hannan and colleagues [20] and others [25], CABG was shown to provide a survival advantage and lower rates of MI and repeated revascularization compared with DES for the treatment of multivessel CAD. In the SYNTAX (Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery) trial, treatment of multivessel CAD with DES was found to be inferior to CABG for the primary end point, 12-month MACCE [1]. However this outcome difference was largely driven by the increased rate of repeated revascularization in the DES group, with mortality and MI having outcomes similar to CABG at 12 months. Other studies have corroborated these findings in the DES era [26]. Furthermore, the severity of multivessel CAD has implications with regard to MACCE. In the SYNTAX trial, when patients were subdivided into tertiles of SYNTAX score, the lowest tertile group showed 1-year MACCE rates similar to those of CABG, although the study was not powered or designed for subgroup analysis. In the Arterial Revascularization Therapies Study II (ARTS II), which enrolled a broader cohort of patients with multivessel CAD, PCI with DES was associated with higher 5-year MACCE rates compared with a historical CABG cohort from ARTS I (Arterial Revascularization Therapies Study I), but patients in the lowest 2 SYNTAX score tertiles had similar MACCE rates compared with CABG [2, 27]. These studies suggest that the severity and complexity of CAD may impact MACCE and the need for repeated revascularization. This discussion is pertinent to the present study for several reasons. Although patients who underwent in this study had equivalent short- and long-term survival and no increased incidences of adverse in-hospital outcomes, the need for repeated revascularization was higher in patients undergoing procedures than in those undergoing multivessel. The need for target vessel revascularization occurred in 8.8% of patients undergoing compared with 3.0% of patients undergoing. These results suggest that careful patient selection for is important for optimal long-term outcomes. A patient with proximal LAD stenosis and focal lesions in the right coronary artery or circumflex artery (or both), eg, may be a more ideal candidate for than a patient with multisegment and diffuse disease, which would require more complex PCI and a larger stent burden. Other lesions that may prove challenging for PCI include heavily calcified lesions, long lesions, and bifurcated lesions. Future comparative studies with CAD severity classified by SYNTAX score may provide better guidelines and selection criteria for patients undergoing. In this study almost all hospital outcomes were similar between treatment groups, with the exception of the need for blood transfusion, which was less with. Others have reported lower complication rates, shorter hospital stays, and reduced stays in the intensive care unit [6 8]. Other advantages may include a shorter recovery time after discharge, quicker return to work, and better patient satisfaction [6]. Although midterm follow-up was available for all patients, the indications for intervention in the group and group were different. In the group, reintervention was driven by clinical indications only. However because the majority of patients underwent operation first and then PCI in the group, the majority ( 90%) of all LIMA grafts underwent angiographic assessment before PCI of non-lad lesions. Nonetheless only 1 of the 7 internal mammary artery interventions was performed during the index hospitalization for an angiographically detected defect. It is our current practice to angiographically assess all grafts performed with minimally invasive approaches (most often in the hybrid room) before discharge from the hospital unless contraindications exist. This stems from the observation from Zhao and associates [5] that up to 12% of all grafts have angiographic defects that are detected with completion angiography. A major limitation to this study is that patient information regarding repeated revascularization for both and was limited to patients who returned to Emory-affiliated hospitals after discharge for reintervention. Patients treated at other regional hospitals for repeated revascularization were not captured in this retrospective analysis. Although this limitation likely applied equally to both and groups, this variable cannot be accounted for; thus the repeated revascularization data must be interpreted with caution. In addition, although patients were matched according to 8 preoperative variables, including coronary anatomy, selection bias cannot be excluded despite the matching algorithm. Patients who underwent were carefully selected for the procedure according to surgeon and interventional cardiologist discretion. Patients excluded from were generally those with complex coronary anatomy. Conclusions In summary represents a safe and effective strategy with equivalent in-hospital outcomes and 5-year survival for carefully selected patients with multivessel CAD. Fueled by a growing desire for less invasive options, this strategy represents a valuable alternative for patients who would otherwise be treated exclusively by multivessel PCI and for selected patients referred for traditional CABG. The former represents an opportunity for patients to receive the long-term proven durability of a LIMA- LAD bypass, whereas the latter represents a minimally invasive option for patients who would otherwise be treated with a more invasive conventional operation. A

Ann Thorac Surg HALKOS ET AL 2011;92:1695 1702 MULTIVESSEL CORONARY ARTERY DISEASE TREATMENT randomized trial comparing with multivessel PCI or CABG, or both, would provide more definitive evidence regarding long-term outcomes and the need to repeat revascularization. Future comparative studies with CAD severity classified by SYNTAX score may provide better guidelines and selection criteria for patients undergoing. Until then careful patient selection and collaboration with cardiologists will be important determinants of outcomes. References 1. Serruys PW, Morice MC, Kappetein AP, et al. Percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease. N Engl J Med 2009; 360:961 72. 2. Serruys PW, Onuma Y, Garg S, et al. 5-year clinical outcomes of the ARTS II (Arterial Revascularization Therapies Study II) of the sirolimus-eluting stent in the treatment of patients with multivessel de novo coronary artery lesions. J Am Coll Cardiol. 2010;55:1093 101. 3. Loop FD, Lytle BW, Cosgrove DM, et al. Influence of the internal-mammary-artery graft on 10-year survival and other cardiac events. N Engl J Med 1986;314:1 6. 4. Vassiliades TA, Kilgo PD, Douglas JS, et al. Clinical outcomes after hybrid coronary revascularization versus offpump coronary artery bypass. Innovations 2009;4:299 306. 5. Zhao DX, Leacche M, Balaguer JM, et al. Routine intraoperative completion angiography after coronary artery bypass grafting and 1-stop hybrid revascularization results from a fully integrated hybrid catheterization laboratory/operating room. J Am Coll Cardiol 2009;53:232 41. 6. Kon ZN, Brown EN, Tran R, et al. Simultaneous hybrid coronary revascularization reduces postoperative morbidity compared with results from conventional off-pump coronary artery bypass. J Thorac Cardiovasc Surg 2008;135:367 75. 7. Reicher B, Poston RS, Mehra MR, et al. Simultaneous hybrid percutaneous coronary intervention and minimally invasive surgical bypass grafting: feasibility, safety, and clinical outcomes. Am Heart J 2008;155:661 7. 8. de Canniere D, Jansens JL, Goldschmidt-Clermont P, Barvais L, Decroly P, Stoupel E. Combination of minimally invasive coronary bypass and percutaneous transluminal coronary angioplasty in the treatment of double-vessel coronary disease: two-year follow-up of a new hybrid procedure compared with on-pump double bypass grafting. Am Heart J 2001;142:563 70. 9. Rosenbaum PR. Optimal Matching for Observational Studies. J Am Stat Assoc 1989;84:1024 32. 10. Vassiliades TA Jr, Douglas JS, Morris DC, et al. Integrated coronary revascularization with drug-eluting stents: immediate and seven-month outcome. J Thorac Cardiovasc Surg May 2006;131:956 62. 11. Schisterman EF, Whitcomb BW. Use of the Social Security Administration Death Master File for ascertainment of mortality status. Popul Health Metr 2004;2:2. 1701 12. Austin PC. The relative ability of different propensity score methods to balance measured covariates between treated and untreated subjects in observational studies. Med Decis Making 2009;29:661 77. 13. Kiaii B, McClure RS, Stewart P, et al. Simultaneous integrated coronary artery revascularization with long-term angiographic follow-up. J Thorac Cardiovasc Surg 2008;136: 702 8. 14. Katz MR, Van Praet F, de Canniere D, et al. Integrated coronary revascularization: percutaneous coronary intervention plus robotic totally endoscopic coronary artery bypass. Circulation 2006;114(1 Suppl):I473 6. 15. Gao C, Yang M, Wu Y, et al. Hybrid coronary revascularization by endoscopic robotic coronary artery bypass grafting on beating heart and stent placement. Ann Thorac Surg 2009;87:737 41. 16. Gilard M, Bezon E, Cornily JC, et al. Same-day combined percutaneous coronary intervention and coronary artery surgery. Cardiology 2007;108:363 7. 17. Holzhey DM, Jacobs S, Mochalski M, et al. Minimally invasive hybrid coronary artery revascularization. Ann Thorac Surg 2008;86:1856 60. 18. Vassiliades TA Jr, Reddy VS, Puskas JD, Guyton RA. Longterm results of the endoscopic atraumatic coronary artery bypass. Ann Thorac Surg 2007;83:979 84. 19. Bonatti J, Schachner T, Bonaros N, et al. Simultaneous hybrid coronary revascularization using totally endoscopic left internal mammary artery bypass grafting and placement of rapamycin eluting stents in the same interventional session. The COMBINATION pilot study. Cardiology 2008;110:92 5. 20. Hannan EL, Wu C, Walford G, et al. Drug-eluting stents versus coronary-artery bypass grafting in multivessel coronary disease. N Engl J Med 2008;358:331 41. 21. Alexander JH, Hafley G, Harrington RA, et al. Efficacy and safety of edifoligide, an E2F transcription factor decoy, for prevention of vein graft failure following coronary artery bypass graft surgery: PREVENT IV: a randomized controlled trial. JAMA 2005;294:2446 54. 22. Balacumaraswami L, Taggart DP. Intraoperative imaging techniques to assess coronary artery bypass graft patency. Ann Thorac Surg 2007;83:2251 7. 23. Barner HB. Operative treatment of coronary atherosclerosis. Ann Thorac Surg 2008;85:1473 82. 24. Byrne JG, Leacche M, Vaughan DE, Zhao DX. Hybrid cardiovascular procedures. JACC Cardiovasc Interv 2008;1:459 68. 25. Yan TD, Padang R, Poh C, et al. Drug-eluting stents versus coronary artery bypass grafting for the treatment of coronary artery disease: a meta-analysis of randomized and nonrandomized studies. J Thorac Cardiovasc Surg 2010:111. 26. Lee MS, Yang T, Dhoot J, Iqbal Z, Liao H. Meta-analysis of studies comparing coronary artery bypass grafting with drugeluting stenting in patients with diabetes mellitus and multivessel coronary artery disease. Am J Cardiol 2010;105:1540 4. 27. Sarno G, Garg S, Onuma Y, et al. Impact of completeness of revascularization on the five-year outcome in percutaneous coronary intervention and coronary artery bypass graft patients (from the ARTS-II study). Am J Cardiol 2010;106:1369 75. ADULT CARDIAC DISCUSSION DR JOSEPH F. SABIK (Cleveland, OH): I suppose a question I always have and 1 of the questions we try to deal with in these hybrid procedures, how do you manage the anticoagulants in terms of the antiplatelet agent? DR HALKOS: Well, our preference is to do the LIMA to LAD first, but our standard is really to treat the culprit lesion first. So in a patient who has a 99% RCA and an 80% LAD, we usually proceed with RCA stenting first, and we usually will proceed with a minimally invasive procedure with the patient on Plavix. We haven t had that much difficulty with bleeding with this minimally invasive approach, but it has been unpredictable at times. DR SABIK: How about in patients who have the IMA first, is there a certain sequence you follow?

ADULT CARDIAC 1702 HALKOS ET AL Ann Thorac Surg MULTIVESSEL CORONARY ARTERY DISEASE TREATMENT 2011;92:1695 1702 DR HALKOS: Well, patients who have the IMA first get loaded with Plavix in the recovery area after there are no significant signs of bleeding. We usually load with 150 mg 4 hours after surgery and then start them on 75 mg a day postoperatively. But that is pretty much our standard regimen for routine off-pump coronary bypass patients. One of the potential options in the future with regard to patients who get stented first is the new I.V. antiplatelet agent, cangrelor, which can be used as a bridge to surgery. It is a very short-acting antiplatelet agent, and we have used that in several cases as well. We hold the drug 2 hours before surgery and then are able to resume Plavix after we have completed the surgical portion of the procedure. anastomosis, and the question is whether those should be left alone and reimaged or whether they should be treated. Honestly, in our center it depends on which cardiologist is doing the procedure. Obviously the other part of your question is whether the anastomosis is being performed as accurately as it would be through a sternotomy, and we will have to follow that more carefully over the long term. I will tell you that since I started doing these, and Dr. Vassiliades taught me how to do these, we have been doing routine completion angiograms at least until we can document our results and confirm that we are providing technically excellent results. DR SABIK: What do you think the role of this will be in the future? Obviously 1 of the benefits of this type of approach really ensures that all the patients are having an ITA to their LAD, and as we know from our own registry, maybe only 95% of patients are getting an ITA. Do you see this as beneficial? DR HALKOS: I think it is interesting how these patients come to us. These aren t patients who would otherwise get coronary bypass surgery. These are usually patients who would otherwise get multivessel PCI. I can t stress the importance of collaborating with the cardiologists with regard to identifying patients who would benefit from a minimally invasive LIMA-LAD who have suitable non-lad anatomy for PCI, which is usually focal lesions in the RCA or circumflex artery. I think as we saw next door, I don t think this is an optimal approach for patients with severe complex CAD. The patients who would otherwise be treated with multivessel PCI are ideal candidates for a hybrid approach because of the survival advantage of the LIMA-LAD bypass. You can argue that a focal RCA lesion that I showed in the presentation can probably be treated as well with a drugeluting stent as it can with a vein graft. I don t have data to support that statement, obviously, but I think that is the argument. We are providing the survival advantage of a LIMA to LAD bypass with the minimally invasive nature of PCI. So it is sort of a best of both worlds strategy, but I think it has been to be done in carefully selected patients. DR GEORGE PALMER (Orlando, FL): Just a question about the use. You said since 2009 you all have been using robotic assistance. Could you define that a little more carefully? And also can you explain why the higher incidence of what I would call burned IMAs or nonusable IMAs in your assisted group? DR HALKOS: Great questions and anticipated questions. Robotic assistance at Emory mostly has involved a LIMA harvest. We open the pericardium and we use the scope during the robotic harvest to identify where on the LAD we plan our anastomosis. Then the remainder of the procedure is done through a 3- to 4-cm rib-sparing thoracotomy with a hand-sewn, off-pump anastomosis. The question regarding the 4.8% problem in the LIMA to LAD anastomosis or in the IMA itself in the hybrid group compared with the 1% problem in the LIMA in the off-pump group is a good question. We know that 1 of those patients was treated in the hospital for an angiographic defect detected prior to their PCI, and one of the talks during the Tech-Con yesterday addressed this. This wasn t a clinically indicated event, but there was some narrowing at the distal, at the toe, or just above the DR BRIAN CMOLIK (Cleveland, OH): Getting back to the LIMA to the LAD technical problems, can you tell me a little bit about whether you use any functional studies to look at whether or not to intervene on those LIMAs before you the let the cardiologist go and instrument those vessels? DR HALKOS: It s a great question. It s a little touchy, because obviously we are working closely with some of our cardiologists. The more seasoned and experienced cardiologists are more inclined to leave them alone, I guess because they have known from prior studies and experience that most of these problems may resolve themselves. They are not very aggressive, and obviously they will call me or call John to discuss these things before they proceed with an intervention, but we haven t worked out those details. DR HANNI SHENNIB (Phoenix, AZ): I think hybrid is a great procedure; however, I am sort of a little bit confused about your strategy for hybrid in those particular patients where you treat the culprit lesion first and then you do the LIMA to the LAD. There have always been the proponents of why don t you just do the culprit lesion and then stage your procedure at a later phase if you need to, ie, defer the secondary revascularization to another admission or interval when the patient becomes symptomatic or the lesion becomes positive on stress? Otherwise, why not just do an procedure? So my question to you is: have you looked at any other diagnostic tool such as FFRs to determine the need for more urgent synchronous revascularization versus staged revascularization or a hybrid procedure? DR HALKOS: I think a lot of that is semantics, but I understand your question. Almost all of our procedures have been staged, and if they proceed with an intervention on the right, we are not usually doing the LIMA to LAD at the same time or the same day. It is usually staged by several days, if not weeks if they are stable, and we do routinely use IVUS and FFR to provide physiologic measurements of either lesion if there is any concern. DR SHENNIB: Right, but my question is, why not just do during the same admission just a stent and then wait until the patient becomes symptomatic to deal with the rest? DR HALKOS: In an angiographically and physiologically significant lesion in the proximal LAD, we make a joint decision during our assessment to proceed with intervention and completely revascularize the patient.