Preparation

After the induction of anesthesia, patients are positioned, prepped, and draped in an identical fashion to an on-pump procedure. Even though OPCAB allows for minimally invasive approaches, including small thoracotomy, endoscopic, and robotic-assisted coronary artery bypass, the most common approach remains via median sternotomy, which facilitates multivessel grafting. A median sternotomy may be routinely accomplished via a limited skin incision (10-12 cm) and allows the surgeon to visualize the operative field from an orientation that is familiar and similar to on-pump procedures. This facilitates target-vessel identification as well as harvesting of the internal mammary arteries. Additionally, should conversion to conventional bypass become necessary, a median sternotomy allows easy access for cannulation for CPB. The limited skin incision should ideally be centered on the lower portion of the sternum and extend to the xiphoid. The linea alba is further incised as the lower portion of the skin incision is undermined. This leaves the upper sternum cosmetically unscathed and also permits wider opening of the lower portion of the sternum, which provides more space for cardiac displacement during off-pump grafting of the lateral and inferior walls.

Conduit Harvesting

During left and/or right internal mammary artery harvest, we routinely skeletonize the vessel using the harmonic scalpel (Harmonic Synergy Blade, Ethicon, NJ, USA) in order to optimize the length of the vessel(s) while minimizing trauma to the chest wall. Unlike ONCAB, in OPCAB the heart is not decompressed, and the extra artery length is often necessary to avoid tension on the anastomosis during rightward displacement for lateral or inferolateral wall grafting. Dividing or removing the endothoracic fascia, skeletonizing the internal mammary artery during harvest, and dividing the left pericardium vertically toward the left phrenic nerve at the level of the pulmonary artery are all adjunct techniques that provide for extra length and less tension on the LIMA-LAD anastomosis. After dividing the mammary artery distally, we inject with a soft silastic-tipped needle each mammary artery with a solution comprised of 19 mL of the patient’s blood, 10 mg of milrinone (in a concentration of 1 mg/mL) and 1000 U of heparin (1000U/mL solution) plus 9 ml of buffered plasmalyte. The instillation of approximately 5 mL of this solution into the lumen of the mammary artery completely resolves any spasm, thereby creating an ideal conduit for the bypass. Skeletonized mammary arteries are then clipped distally and placed in a “jacuzzi” bath of this same solution, within a stoppered 5 cc syringe positioned in the apex of each thoracic cavity. Radial artery and saphenous vein conduits are simultaneously harvested endoscopically during internal mammary artery harvest. It is our practice to administer 2500 U of heparin before endoscopic conduit harvest to minimize thrombus formation within the conduit during the harvest. Concern over graft quality with endoscopic vein harvest has prompted increased vigilance in atraumatic harvest technique to ensure optimal conduits for bypass. We use the same solution described above to instill within and to bathe the radial arteries conduits after harvest until the time of grafting.

Pericardiotomy

After single or bilateral internal mammary artery harvest, the heparin dose is administered, and the arterial conduits are divided distally. Once all conduits are obtained and checked, the retractor is positioned in the sternal incision. A more caudal placement reduces traction on the brachial plexus, and generally facilitates mobilization of the heart for positioning. Current retractors used for beating-heart surgery come with attachable devices to aid in positioning the heart as well as stabilizing the target artery.

The pericardium is then opened in an inverted T incision and incised laterally along the diaphragm to facilitate cardiac displacement. It is essential to free the left lateral pericardium from the diaphragm to allow the pericardium to be retracted to displace the heart and effectively expose the lateral wall of the left ventricle. Nonetheless, the phrenic nerves must be identified and carefully preserved during pericardial incision and mobilization. Several pericardial traction sutures are placed to assist with exposure and lateral displacement of the heart; these stitches are positioned away from the opening margin of the pericardium fairly deep in the lateral wall of the pericardium in order to maximize the exposure and take advantage of the rolling ability of the heart within the semicircular circumference of the pericardium. To avoid compression on the right heart during lateral displacement, the right pericardium can be dissected along the diaphragm, or the right pleural space can be opened widely to allow the heart to fall into the right chest during lateral displacement; this trick is particularly useful in the setting of cardiomegaly. Additionally, 1 or 2 rolled towels placed under the right limb of the sternal retractor help to elevate the right side of the sternum to allow the heart to be displaced toward or into the right chest.

An important traction suture is the “deep stitch,” which is placed approximately two-thirds of the way between the inferior vena cava and left pulmonary vein at the point where the pericardium reflects over the posterior left atrium (Figure 1). Care should be taken with placement of this suture to avoid the underlying descending thoracic aorta, esophagus, left lung, and adjacent inferior pulmonary vein. While passing the deep stitch, the right-sided pericardial traction sutures should be relaxed to prevent compression of caval inflow. The deep stitch should be covered with a rubber catheter to prevent laceration of the epicardium during retraction. Furthermore, the manual elevation and compression of the heart required to take this stitch may be poorly tolerated in patients with marginal hemodynamics or significant left main coronary artery disease. In that case, grafting and reperfusion of the left anterior descending coronary artery should be accomplished before placing the deep pericardial traction suture.

Recently we have adopted the use of a long 2-inch wide ribbon gauze to aid in displacement of the heart from the pericardial well. This technique was first described in scientific detail by Sergeant.[30] We have chosen to combine the deep stitch and its rubber catheter with the gauze sling by simply entrapping the midpoint of the ribbon gauze between the strands of the deep stitch and fixing it to the posterior pericardium by tightening the catheter down onto it with a strong hemostat (Figure 1). The two limbs of the ribbon gauze are then used to facilitate mobilization and displacement of the heart to expose various coronary artery targets during grafting. The ribbon gauze is easily removed when the deep stitch and its catheter are removed after all grafts are completed and good graft flows have been documented.

Figure 1

Placement of the “deep stitch.” Surgeon’s view. The heart is elevated, and the suture is placed approximately two-thirds of the way between the inferior vena cava and left pulmonary vein at the point where the pericardium reflects over the posterior left atrium. The deep stitch should be covered with a soft rubber catheter to prevent laceration of the epicardium during retraction. The purpose of the deep stitch is to elevate the heart up and out of the pericardial well.

Aortic Evaluation

Before beginning to perform the distal anastomoses, a final plan of grafting has to be confirmed: all the targets should be visualized, and the ideal conduit for each target should be chosen in advance. It is extremely important at this stage to confirm the possibility of safely connecting the proximal anastomosis to the ascending aorta or opting for a no-aortic-touch technique, which requires I- and/or T-graft composite conduits to be built between the internal mammary arteries and free radial artery conduits before beginning distal coronary anastomoses. The grafting strategy should be confirmed and verbally shared with all members of the operating team prior to beginning distal anastomoses. Epiaortic ultrasonography is used in all our patients undergoing cardiac surgery to evaluate and grade the ascending aorta. It provides both the surgeon and the anesthesiologist with a simple, fast, noninvasive, and inexpensive tool for assessing the extent of atheromatous disease in the ascending aorta and should be performed in every cardiac surgical operation in preparation for aortic cannulation/clamping or selection of an alternative clampless technique. Through the operative incision, the ascending aorta, from the aortic root to the origin of the innominate artery, is scanned directly by the surgeon using an ultrasound probe connected to an echocardiography ultrasound scanner. Any of a variety of commercially available linear array ultrasound probes may be placed inside a sterile sleeve filled with sterile saline and used intraoperatively to interrogate the ascending aorta. The ultrasound imaging probe manufactured by Medistim Inc (Oslo, Norway) has a 128-element transducer that is designed for direct cardiac use and is sterilized and applied directly to the ascending aorta, operating at frequencies between 8 and 18MHz. The information obtained often dictates changes in operative strategy, depending on the grade of atherosclerosis. Similarly, it allows the surgeon to individualize placement of proximal clampless facilitating anastomotic devices to minimize the risk of atheroembolism. Multiple studies using epiaortic ultrasound for aortic screening have demonstrated improvement in clinical outcomes as a result of intraoperative modifications of the surgical plan. .[31],[32],[33],[34]

Exposure

Optimal target-vessel exposure and 3-dimensional stabilization are essential for OPCAB surgery. Cardiac positioners and stabilizers have greatly increased the ability to manipulate the heart with minimal hemodynamic compromise. Two systems routinely adopted are the Medtronic Octopus Tissue Stabilizer and Starfish or Urchin Heart Positioner (Medtronic, Inc, Minneapolis, MN) and the Getinge ACROBAT stabilizer and XPOSE positioner (Getinge, Wayne, NJ, USA). “Apical” suction devices allow distraction and manipulation of the heart by creating a vacuum-type seal to the epicardial surface. They are generally placed on the apex to expose the anterior wall (LAD territory) and inferior wall (posterior descending territory) of the heart and may also be placed on the acute margin to expose the right coronary artery (Figure 2). They are frequently placed slightly away from the apex, especially to the left of the apex, to expose the lateral wall and branches of the left circumflex coronary artery (Figure 3). Because these suction-based cardiac positioning devices pull rather than push to accomplish cardiac displacement, the heart is not compressed, functional geometry is maintained, and hemodynamic stability is preserved. Importantly, these suction-based cardiac positioning devices are best used in combination with deep pericardial tractions suture(s) and/or ribbon gauze. The traction stitch and ribbon gauze elevate the posterior pericardium, and verticalize and rotate the heart so that the suction-based cardiac positioning devices are not obliged to exert as much traction force on the epicardium, thereby allowing use of lower levels of suction and decreasing the likelihood of epicardial hematomas or tears, which can be problematic. The anterior wall vessels often require only the coronary stabilizer for adequate exposure. The stabilizer is positioned along the caudal aspect of the retractor toward the left, with the retractor arm placed out of the way to prevent interference during the anastomosis. The location of these devices on the sternal retractor also requires consideration. For the lateral and inferior wall vessels, the cardiac positioner is usually placed on the surgeon’s side at the most cephalad location of the retractor. The coronary stabilizers can then be placed on either side. A general rule is to place the stabilizer in the assistant’s way instead of the surgeon’s in order to prevent these devices from obstructing the surgeon’s view or interfering with hand positioning during suture placement.

Figure 2

Exposure of the inferior wall with aid of an apical suction device. View from the surgeon’s side of table. The cardiac suction device is placed on the apex allowing the heart to be easily displaced to expose the inferior wall vessels. The apical suction device may be attached to the retractor on the surgeon’s side, as shown, or on the assistant’s side.

Figure 3

Exposure of the lateral wall with aid of an apical suction device. View from head of table. The cardiac positioning device is placed slightly off of the apex. A coronary stabilizer may be used additionally to expose lateral wall vessels, such as the obtuse marginal vessels.

As mentioned, in addition to the positioners and stabilizers, manipulating the traction sutures and gauze slings can greatly enhance exposure. The purpose of the “deep stitch” is to elevate the heart up and out of the pericardial wall. When this suture (covered by a rubber catheter) is retracted toward the patient’s feet, it elevates the heart toward the ceiling and points the apex vertically with remarkably little change in hemodynamics. The two limbs of the ribbon gauze can then be used to fine-tune the position of the heart to optimize exposure of the desired coronary artery target. When the deep suture is retracted toward the patient’s left side, the heart rotates from left to right, exposing the lateral wall vessels. Variable tension on this stitch will enhance exposure to both the anterior and lateral wall. Again, the two limbs of the ribbon gauze are used to rotate the heart further. These maneuvers are performed before utilizing the suction-based positioning device. During exposure of the lateral wall, the left-sided pericardial sutures should be pulled taut, and the right-sided sutures completely relaxed to avoid compression on the right heart during cardiac displacement. It is critically important that pericardial sutures on both the right and left sides are never under tension simultaneously when displacing the heart to expose coronary targets, as this will generally lead to diminished venous return to the heart and subsequent hypotension.

Manipulation of the operating table can also facilitate exposure. Placing the patient in steep Trendelenburg position exposes the inferior wall. Turning the table sharply toward the right will aid with exposure of the lateral wall targets. Usually, little manipulation is required for grafting the anterior wall vessels. Occasionally, a warm, moist laparotomy pad can be placed adjacent to the “deep stitch” to assist with elevating and rotating the heart out of the pericardium, although this is rarely necessary when the ribbon gauze sling technique is used.

The coronary stabilizer devices consist of pods of suction cups that immobilize the target area by creating a vacuum between the epicardial surface and the stabilizer arm. This allows for construction of the anastomoses to take place in a relatively motionless field, approximately recreating the same anastomotic situation as in an arrested heart. In preparation for each distal anastomosis, a soft silastic vessel loop mounted on a blunt needle (Retractotape, Quest Medical, Inc, Allen, TX) is placed widely around the proximal vessel for transient occlusion. For inferior wall vessels, the tails of this vessel loop can be displaced posteriorly and caudally by tying a more posterior pericardial suture loosely around the vessel loop (Figure 4). The pericardial retraction suture serves as a “pulley” that not only enhances coronary exposure and the surgeon’s view, but also keeps this retraction stitch from interfering with the fine sutures during construction of the anastomosis (Figure 4). Similarly, this maneuver can be done for some lateral wall targets.

Figure 4

Use of retractor tape to create a “pulley” for optimizing inferior wall target vessel exposure. Close-up surgeon’s view. A soft silastic retractor tape is mounted on a blunt needle and placed widely around the target vessel proximally for transient occlusion. For inferior wall vessels, this suture can be displaced posteriorly and caudally by tying a more posterior pericardial suture loosely around the retractor tape. A medium clip secures the suture and retractor tape in place, which can then be retracted to transiently occlude the artery.

Although a well-trained first assistant is necessary for providing an effortless anastomosis, the second assistant also plays a major role in exposure. The field is kept free of blood with a humidified CO2 blower (DLP, Medtronic) and Cell Saver (Haemonetics, Braintree, MA), which are managed by the scrub nurse or second assistant during the opening and anastomosis of each distal coronary artery target. Occasionally an epicardial fat retractor can be used to expose the coronary target in patients with a large amount of epicardial fat. The second assistant usually stands to the right of the surgeon, though better exposure by this assistant standing at the head of the bed, to the surgeon’s left, may be achieved during anastomosis of the inferior wall or lateral wall targets. In chronically occluded vessels that have collateral and/or retrograde flow, bleeding into the field can be controlled with another vessel loop distally or an intracoronary shunt.

A final preparatory measure is to place temporary atrial and ventricular pacing cables before positioning the heart if it seems likely that intraoperative pacing will be useful. As the heart is rotated toward the right, visualization of the right atrium is more difficult, making placement of temporary clip electrodes on the right atrium challenging. It may be necessary to pace the left atrial appendage in rare circumstances; in this case, it is important to remember how friable that structure can be.

Sequence of Grafting

Careful assessment of the coronary angiogram is imperative. During on-pump cases, the location and number of vessels requiring bypass usually suffice during evaluation of the catheterization films. However, when planning for OPCAB, particular attention should be paid to the direction of collateral flow between coronary vessels, the presence of intramyocardial vessels, the size of the distal targets, the degree of stenosis, the complexity of coronary disease, and the number of lateral wall vessels requiring grafting. Careful attention must be paid to the sequence of grafting because regional myocardial perfusion is temporarily interrupted in the beating heart (Table 1). As a general rule, the collateralized vessel(s) is grafted first, and the collateralizing vessel grafted last. For example, in patients with an occluded right coronary artery with a posterior descending artery supplied by collaterals from the left anterior descending artery, grafting the left anterior descending first would not only leave the anterior wall ischemic, but also disrupt flow to the septum, inferior wall, and right ventricle during the LAD anastomosis. Thus, a more prudent approach would involve grafting the posterior descending artery first, then providing inflow to this graft (either via an in situ IMA or clampless proximal anastomosis) to ensure adequate flow while the proximal left anterior descending is temporarily occluded for construction of the LAD anastomosis. An alternative would be to graft the LAD first, with preemptive use of an intracoronary shunt. Another scenario that may pose problems is a large, moderately stenotic right coronary artery. Not uncommonly, temporary occlusion of this artery will result in profound bradycardia and hypotension. In these circumstances, the surgeon must be prepared to use an intracoronary shunt or promptly provide temporary epicardial pacing.

If LAD and diagonal vessels are generally considered less challenging to expose, stabilize and graft, an increasing level of difficulty applies for the right coronary artery, posterior descending artery, proximal and lateral obtuse marginal vessels and the ramus intermedius. The ramus intermedius artery is challenging to expose and stabilize because compression on the right ventricular outflow tract is poorly tolerated. The introduction of suction devices to position the heart has decreased the hemodynamic alterations associated with exposing this difficult part of the heart. The apex of the heart is retracted toward the patient’s right hip, and the table is rotated rightward toward the surgeon. This area of the heart tends to be tethered by the nearby pericardial reflection, reducing its mobility and making exposure more difficult. Beware aggressive incision of the pericardium here; the left phrenic nerve passes nearby. Additionally, a large left atrial appendage can hinder visualization when grafting near the atrioventricular groove. Rapid recovery of regional myocardial function prior to subsequent occlusion of other target arteries is essential for successful multivessel OPCAB. Additional options include a “proximals first” approach, which allows adequate regional perfusion after completion of each distal anastomosis. Of course, this is implicit in an all-arterial anaortic approach in which composite conduits are constructed before beginning distal anastomoses and each coronary target may be reperfused as soon as it is grafted. Although concern for myocardial protection during OPCAB stems from the brief periods of coronary occlusion necessary to visualize distal target vessels, adequate perfusion can be achieved by maintaining adequate systemic perfusion pressure, selective use of coronary artery shunts, careful use of traction sutures and stabilizers, and proper sequencing of graft anastomoses.[35]

Distal Anastomosis

After cardiac positioning and coronary stabilization, the silastic vessel loop can be placed, and the epicardium overlying the coronary artery is dissected at the chosen anastomotic site. If there are concerns about hemodynamic stability during regional ischemia, the proximal vessel can be test-occluded for 2 to 5 minutes. During this time the graft can be prepared. This gives the surgeon some assurance before committing to the anastomosis by creating an arteriotomy. After a brief period of reperfusion of 2 to 3 minutes, the vessel can be reoccluded and the artery prepared for anastomosis. This strategy of “preconditioning” is not used routinely, but can be useful in cases of important regional ischemia. Careful and judicious placement of intracoronary shunts is important because at least one study demonstrated significant endothelial injury with the use of intracoronary shunts. [35] Nonetheless, intracoronary shunts are used quite routinely, especially in large coronary arteries with large areas of subtended myocardium at risk. Moreover, intracoronary shunts are an important adjunct to the teaching of OPCAB to residents and fellows, since they protect the back wall of the coronary artery from inadvertent suturing and permit the less experienced surgeon more time to complete a precise anastomosis. The anastomosis is otherwise performed in a manner identical to on-pump grafting. It is essential to continue communication with the anesthesia team so that adequate steps can be promptly taken if hemodynamic conditions deteriorate.

For example, if pulmonary artery pressures begin to rise and mean arterial pressures begin to fall during a lateral wall anastomosis, several steps must be taken preemptively to avoid cardiovascular collapse. The first sign of impending deterioration may be a flattening of the continuous pulmonary arterial pressure tracing on the intraoperative monitor. Gently relaxing the cardiac positioner or coronary stabilizer can often improve hemodynamics. Optimizing table positioning, inotropes, vasopressors, fluid boluses, or pacing may also help. However, if it appears that hemodynamic conditions are deteriorating despite these interventions, then the safest next step is to place an intracoronary shunt, release both the coronary stabilizer and cardiac positioner, return the heart to the pericardial space, and allow hemodynamics to recover. At this point, a decision must be made to either convert “electively” to an on-pump procedure, possibly including on-pump with the heart beating, or complete the procedure off-pump with the benefit of the intracoronary shunt. Another option that is frequently used in patients at high risk for complications of CPB is the use of an intra-aortic balloon pump (IABP). An IABP can provide valuable mechanical support during cardiac displacement and positioning to enable safe and controlled completion of a distal anastomosis or entire CABG case that would otherwise require cardiopulmonary bypass. When needed, the IABP is best utilized as part of “Plan A” and is placed under sterile conditions with intraoperative TEE guidance at the beginning of the operation.

Proximal Anastomosis

Unlike on-pump coronary artery bypass, OPCAB provides the opportunity to minimize or completely avoid manipulation of the aorta. Avoiding partial clamping during proximal anastomoses can be achieved by performing proximal anastomoses to in situ arterial grafts, or using clampless proximal automated anastomotic connectors or facilitating devices.[36] This may be particularly relevant in patients with advanced aortic atheromatous disease detected by epiaortic ultrasound. Over the past decade, our practice has evolved to a routine all-arterial, no-aortic-touch OPCAB technique, completely avoiding any manipulation of the ascending aorta in the large majority of patients. While this approach has obvious advantages, it is a technically challenging procedure. A second choice to avoid aortic clamping is to use a commercially available device for clampless proximal anastomoses. These include the Heartstring III (Getinge Cardiovascular, Wayne, NJ) or PAS-Port Proximal Anastomosis System (Cardica, Redwood City, CA). The Heartstring device creates a hemostatic seal with the inner surface of the ascending aorta that allows the creation of a handsewn anastomosis with a relatively bloodless field. After completion of the anastomosis, the device is removed by unwinding the sealing cup from the aorta before tying down the suture; there is no foreign material other than suture material left in the anastomosis (Figure 5). However, this device still requires a handsewn anastomosis to be performed between the graft and the aorta, and can be associated with some blood loss. [37] The PAS-Port Proximal Anastomosis System was specifically designed to create a consistent anastomosis between a saphenous vein graft and the aorta during either on-or off-pump coronary bypass surgery. It is a fully integrated, automated system that cuts the aortotomy and attaches the vein graft to the aorta in seconds, producing consistent, reproducible anastomoses independent of surgical technique and skill. Compared with earlier devices, the PAS-Port system allows the endothelium of the vein graft to be untouched during the loading and deployment process; [38] however, there is a small amount of metallic foreign material left within the graft lumen.

Figure 5

Use of a Heartstring device to perform a proximal anastomosis without aortic crossclamping. Close-up surgeon’s view. After use of a standard aortic puncture device, the Heartstring device (A) creates a hemostatic seal with the inner surface of the ascending aorta that allows the creation of a handsewn anastomosis (B) within a relatively bloodless field. After completion of the anastomosis, the device is removed by unwinding the sealing cup from the aorta before tying down thesuture; there is no foreign material other than suture material left in the anastomosis.

A number of large retrospective studies have shown OPCAB to be associated with a decreased incidence of perioperative stroke compared with conventional on-pump CABG.[39] Conversely, randomized studies comparing the 2 revascularization strategies have often failed to show a significant advantage with OPCAB regarding postoperative stroke. This apparent contradiction may be explained by the fact that stroke is a relatively rare occurrence after CABG surgery, and even the largest randomized studies are underpowered to find a difference in the incidence of stroke. [40] A single-center institution reviewed more than 12,000 patients who underwent primary isolated CABG, and compared the incidence of stroke in patients with a complete no-touch aortic manipulation technique (inflow from the mammary arteries) vs patients who underwent proximal anastomosis with a proximal facilitating device, vs patients where a proximal clamp on the aorta was adopted. No-touch aortic technique and proximal facilitating device were associated with a statistically significant reduction of perioperative stroke when compared with proximal aortic side clamp techniques. [41] In OPCAB, we generally do not leave temporary epicardial atrial and ventricular wires, unless pacing wires were required during surgery or bradycardia or low left ventricular ejection fraction make a requirement for postoperative pacing likely.

All-Arterial No-Aortic-Touch OPCAB: State of the Art Surgical Coronary Revascularization

No-aortic-touch (as known as ‘anaortic’) coronary artery bypass grafting is a method of off-pump surgical coronary artery revascularization which avoids aortic manipulation by utilizing composite grafts with blood supply from one or both internal mammary arteries (IMAs) or the gastro-epiploic artery. Frequently, all-arterial grafts are employed. Typically, bilateral IMAs serve as the primary blood source. This technique provides specific benefits to those patients who cannot tolerate cardiopulmonary bypass and aortic cross-clamping e.g., individuals with a porcelain ascending aorta and/or mobile atheroma in the ascending aorta or aortic arch, by reducing the risk of disseminated emboli or aortic injury/dissection from aortic cannulation and cross-clamping. (Figure 6) [42] The rationale for all-arterial, no-aortic touch OPCAB is based on three important concepts: 1. All-arterial grafting prolongs life; bilateral IMAs and radial arteries are usually both better than veins for long-term graft patency. 2. No-aortic-touch OPCAB, by avoiding any manipulation of the ascending aorta, is associated with lower risk of stroke and death, and especially benefits high-risk patients. 3. Therefore, precisely performed, all-arterial, no-aortic-touch (anaortic) OPCAB may deliver the best possible long-term outcomes with the least possible perioperative morbidity and mortality. The evidence in support of the use of bilateral IMAs over single IMA dates to the 1990s, with landmark data provided by Lytle and colleagues. [43] Their findings demonstrated a significant survival benefit of bilateral IMAs, with the magnitude of such benefit increasing through 20 postoperative years. The benefits of bilateral ITAs usage have been subsequently reiterated by multiple studies, both at a multicenter [44],[45]and at a meta-analytic level[46]. Puskas et al. later demonstrated that such long-term benefit persists even amongst diabetic patients, provided conduits are skeletonized – which preserves retrosternal microperfusion, thus minimizing the risk for deep sternal wound infection.[47],[48],[49] Additionally, the routine use of vancomycin paste is a powerful adjunct to achieve both intraoperative hemostasis of the sternal bone and effective prevention of deep sternal wound infection. In our experience, optimal consistency of the paste can be achieved with 1 gram of vancomycin mixed with 3-4 drops of plasmalyte. The use of the radial artery, whose long-term patency has been shown to be similar to IMA and superior to saphenous vein conduits [50],[51] is a key to advancing total arterial revascularization. Importantly, a post hoc analysis of the ART trial showed that when radial artery was used to supplement single or bilateral IMAs, this multi-arterial configuration was associated with lower risk for mid-term major adverse cardiac events. [52] Additionally, a meta-analysis of propensity-matched studies (including more than 10,000 patients) by Gaudino et al. showed a statistically significant mortality benefit with three versus two arterial grafts. [53] As a final demonstration of the benefits of multiple arterial grafting, an elegant analysis by Taggart et al. has shown that the cumulative risk of major cardiovascular events increases with increasing number of venous grafts and decreases with increasing number of arterial grafts; the risk of death increased proportionally with the use of any vein grafts. [54] Coronary artery bypass surgery conducted on an arrested heart relies on clamping the ascending aorta. This procedure involves substantial manipulation of the aorta and poses notable risks of atheroembolism, with external pressure and forces leading to the possible detachment of atherosclerotic plaque from the inner lining of the aorta. Additionally, cardiopulmonary bypass itself can dislodge atherosclerotic plaques or atheroma, by the well-described ‘sandblasting effect’.[55],[56],[57] As a consequence, surgical approaches involving aortic manipulation or cardiopulmonary bypass entail higher perioperative neurologic injury rates, especially in patients with severe aortic atherosclerosis. Anaortic off-pump coronary artery bypass techniques, which avoid aortic manipulation entirely, are recognized as safe and efficient in mitigating neurologic complications. This is supported by Class 1B recommendations in the 2018 ECS/EACTS guidelines, highlighting the efficacy of no-aortic-touch strategies specifically in patients with ascending aortic atherosclerosis. [58] Indeed, the implementation of the no-aortic-touch technique has shown to significantly decrease the early stroke rate. [59] Additionally, a network meta-analysis by Zhao et al. [39] has demonstrated a net benefit of anaortic OPCAB not only in terms of postoperative stroke, but also mortality, postoperative myocardial infarction, renal failure, bleeding, and atrial fibrillation (Figure 7).

Figure 6

No-aortic-touch (as known as ‘anaortic’) coronary artery bypass grafting. Both mammary arteries are utilized as the primary blood source to achieve total arterial revascularization together with a composite graft of radial artery. The right internal thoracic artery (RITA) is extended in end-to-end fashion with a short segment of radial artery (RA) to reach the inferior wall and bypass the posterior descending artery (PDA). The left internal thoracic artery (LITA) is used to bypass the left anterior descending artery (LAD) and a ’Y’ graft anastomosis of radial artery (RA) branching from the left internal thoracic artery (LITA) is used to reach the last branch of the Circumflex artery (Cx marginal).

Figure 7

League Tables for CABG With and Without Manipulation of the Aorta. Outcomes shown for (A) stroke, (B) mortality, (C) myocardial infarction, (D) renal failure, (E) bleeding complications, and (F) atrial fibrillation following CABG with and without manipulation of the aorta (OR and 95% CI). OR < 1 means the treatment in top left is better. anOPCABG = anaortic off-pump coronary artery bypass grafting; OPCABG-PC = off-pump coronary artery bypass grafting with partial clamp; OPCABG-HS = off-pump coronary artery bypass grafting with the Heartstring system; OR = odds ratio. Reproduced with permission from [39]

Beating Heart Coronary Artery Bypass

Performing beating-heart coronary artery bypass with CPB support is especially useful in certain clinical scenarios, such as acute coronary syndromes with cardiogenic shock or in patients with severe left ventricular dysfunction. [60], [61] Some such patients already have tenuous hemodynamics and will not tolerate cardiac positioning and displacement during routine OPCAB maneuvers. In this approach, the patient is fully heparinized, the ascending aorta and right atrium are cannulated and the anastomoses are completed on CPB with the heart beating but decompressed. This provides for hemodynamic support and also eliminates the global ischemic insult and aortic cross-clamp associated with cardioplegic arrest. Grafting can then be performed in a similar sequence as OPCAB. This approach has been supported in several recent studies, [62],[63] specifically in the setting of acute coronary syndrome, as it provided improved outcomes with lower postoperative morbidity and mortality in patients undergoing emergency myocardial revascularization. We find this approach useful in the unusual situation when we need to complete one or two more distal anastomoses and hemodynamic stability is insufficient, despite all routine OPCAB maneuvers. It is important to maintain normal coronary perfusion pressures during normothermic beating-heart on-pump CABG to avoid inadvertent myocardial ischemia.

Nonsternotomy OPCAB Approach

The competitive status of percutaneous transluminal coronary angioplasty and stenting has stimulated an interest in minimally invasive direct coronary artery bypass grafting (MIDCAB). MIDCAB avoids cardiopulmonary bypass, permitting surgeons to pursue a less invasive approach that spares the sternum, expedites the postoperative course, and decreases the overall length of stay. Left thoracotomy was first adopted in the 1990s as an alternative route for off-pump LITA-to-LAD grafting. Initially, specialized retractors and instruments simplified the LITA harvesting and allowed surgery through a small anterolateral thoracotomy. Internal thoracic artery harvesting was further improved by the introduction of the da Vinci Surgical System (Intuitive Surgical), which facilitates harvesting of single or double ITAs and also allows completion of anastomosis to the anterior and lateral wall of the heart. Importantly, the left thoracotomy approach is also used in patients undergoing reoperative CABG as an alternate route of access to the lateral wall, thereby avoiding a repeat sternotomy and potential injury to patent anterior grafts. In this approach, proximal inflow may be from the in situ ITA or from proximal anastomosis of conduits to the descending thoracic aorta.

Clinical outcomes after OPCAB and ONCAB have been compared for more than two decades, with enrollment across many centers including hundreds of thousands of patients. Despite the abundance of literature, there is still no consensus regarding the optimal bypass strategy, especially in low-risk patients. In higher risk patients, it appears in recent studies that OPCAB may reduce both morbidity and mortality. This is not surprising; it is rational to expect the benefit of OPCAB to be greatest in those higher-risk patients for whom aortic cannulation, cross-clamping and CPB are more likely to lead to adverse outcomes. Studies may be divided into prospective randomized trials and observational retrospective analyses. Prospective trials avoid some of the selection bias and confounding inherent to retrospective and observational analyses, but are themselves subject to unmeasurable enrollment bias, since typically both the referring cardiologist and enrolling surgeon must share equipoise prior to enrollment. Moreover, due to resource constraints, prospective randomized studies are smaller and thus statistically underpowered to detect incremental differences in morbidity or mortality rates following CABG among patient groups in whom those events are infrequent. This remains true despite the recent completion and publication of 3 large, multicentered randomized trials—ROOBY, [65] GOPCAB, [66] and CORONARY [67],[68] —as the patient sample sizes required to demonstrate a significant difference in mortality would be more than 50,000 patients, and similar sample sizes would be required to detect differences in stroke and myocardial infarction. [68] Retrospective and observational analyses provide the large cohort size and long duration of follow-up to sufficiently power these studies to detect small but important differences in outcomes. However, retrospective studies are themselves inherently limited by biases, the most important of which is selection bias, which may persist despite the use of propensity matching and other advanced statistical methodologies designed to control for confounding biases. Taken together, both types of studies can provide valuable information to guide clinical practice.

Early Outcomes

The aforementioned three large, multicentered, randomized trials have all reported no difference between OPCAB and ONCAB with regard to 30-day mortality, nonfatal stroke, nonfatal myocardial infarction, and new renal failure within 30 days after randomization (Table 2). These studies varied in terms of patient populations and surgeon expertise, and thus, the similar near-term mortality and morbidities across these studies support the belief that both techniques are similarly effective in the near-term, at least for low-risk and mixed-risk populations of patients.

More specifically, the CORONARY trial (the CABG Off or On Pump Revascularization Study)[67],[69] and the GOPCABE[66] trial (German Off-Pump CABG Trial in Elderly Patients) reported results following randomization of 4752 and 2539 patients, respectively. In comparison with the previous individual largest randomized trial (ROOBY Trial with 2203 patients),[65] the CORONARY and GOPCABE trials were designed to enroll sicker, older, and an appropriate proportion of female patients: these changes resulted in a cohort more similar to the composition of patients actually operated on, thereby increasing external validity of the results. In all three trials, approximately 60% to 70% of patients had triple-vessel disease. The predicted 30-day mortality risk was 1.9% in the ROOBY trial, 3.8% in the GOPCABE trial, and 80% of patients in the CORONARY had a EuroSCORE of 0.5 (i.e., a predicted mortality of approximately 2%). Both the 30-day and the 1-year outcomes for the 3 trials are individually summarized in Tables 2 and 3, respectively. For these trials, there was no significant difference in the 30-day composite primary endpoint between OPCAB and ONCAB, or in the individual incidence of death, MI, stroke, or need for new renal dialysis. Both the CORONARY and GOPCABE trials reported an increased risk of repeat revascularization within 30 days for OPCAB, with respective figures of 0.7% vs 0.2%, and 1.3% vs 0.4%.

Though controversial, compelling evidence supports the theory that OPCAB is associated with decreased operative mortality and morbidity (Figure 8), particularly for high-risk patients, such as those with high risk scores based on Society of Thoracic Surgeon (STS) risk scoring calculations (Figure 9), older patients, patients with renal failure, patients with reduced left ventricular ejection fraction, and patients who have undergone redo CABGs.[11],[13],[14],[65],[66],[67],[69],[68] Less controversial is that OPCAB offers a variety of anaortic techniques that have been shown to be complimentary in reducing the risk of stroke especially in patients with atherosclerosis of the ascending aorta.[39]

Figure 8

Risk-adjusted odds ratios for major adverse events in all patients undergoing on-pump coronary artery bypass (ONCAB, patterned bar) vs off-pump procedures (OPCAB, filled bar). MACE indicates major adverse cardiac events; MI, myocardial infarction.

Figure 9

Comparison of observed mortality rates following off-pump coronary artery bypass grafting (OPCAB) and coronary artery bypass grafting (CABG) on cardiopulmonary bypass (CPB), based on patient preoperative Society of Thoracic Surgery (STS) predicted mortality risk score.

Long-Term Outcomes

Two potential benefits of ONCAB relative to OPCAB are increased completeness of revascularization and higher quality anastomoses, both of which would likely impact long-term survival following CABG. Randomized trials have consistently reported increased number of grafts with ONCAB vs OPCAB, even when studies require very experienced surgeons,[65],[66],[67] and decreased completeness of revascularization has been shown to be associated with decreased long-term survival.[71] However, three randomized trials that followed patients to 5 years and beyond reported no survival difference and almost identical survival curves.[65],[69],[72],[73] Furthermore, long-term follow-up of patients randomized to either treatment in modern studies has consistently shown no difference in individual outcomes, including stroke, repeat revascularization, myocardial infarction, need for reintervention, renal failure requiring dialysis, or composite endpoints of the above, whereas the ROOBY trial—the first of the modern multicentered randomized trials—found superior graft patency at 1-year with ONCAB, [65] this study allowed inexperienced surgical trainees to participate as the primary surgeon, a design element that has been widely questioned and corrected in subsequent trials. A meta-analysis of 10,709 patients across 3 randomized trials and 10 selected retrospective studies who were followed for more than 5 years reported that OPCAB was associated with increased long-term mortality (hazard ratio, 1.05 [95% CI, 1.00-1.13]). [74] Of note, randomized controlled trials contributed less than 3% of the weight to this meta-analysis outcome, and this analysis did not include the results of CORONARY, the largest RCT of OPCAB. Therefore, it is unclear whether the discrepancy between randomized trials and this meta-analysis is due to increased power due to larger sample size with the meta-analysis, or selection bias among the predominantly observational cohort. Notably, it has been repeatedly reported that nonrandomized patients undergoing OPCAB generally have higher preoperative comorbidities than those undergoing ONCAB, as surgeons and centers with expertise in OPCAB tend to avoid CPB in higher-risk patients. Mid- and long-term survival rates have demonstrated similarity between OPCAB and ONCAB patient cohorts. In a study by Hannan et al., the 3-year survival rates were indistinguishable between OPCAB and ONCAB groups, with an unadjusted survival rate of 89.4% versus 90.1% (p = 0.20). [75] Similarly, a study by Puskas et al. including over 12,000 patients showed equivalent 10-year survival rates across both OPCAB and ONCAB groups. [76] The study by Angelini et al. on long-term follow-up, spanning 6 to 8 years, found no disparity in survival outcomes between OPCAB and ONCAB patients (hazard ratio, 1.24; 95% CI 0.72–2.15, p = 0.44). [77] Puskas et al. reported similar findings in their randomized trial, demonstrating comparable survival rates at a mean follow-up of 7.5 years. [73] Conversely, the ROOBY trial, a randomized controlled trial by Shroyer et al., indicated a higher 1-year composite outcome for OPCAB patients, although other individual endpoints did not significantly differ. [65] Sensitivity analysis revealed higher 1-year cardiac-related mortality in the OPCAB group (2.7% vs. 1.3%, p = 0.03). [65] A more recent randomized controlled trial of patients with triple-vessel coronary artery disease undergoing OPCAB versus ONCAB has shown similar incidence of major adverse cardiovascular events and postoperative complications when surgery is performed in the same setting by an expert surgeon in both methods. [78] In a propensity-matched analysis from the UK National Database, Chan et al. have demonstrated a similar in-hospital mortality in octogenarians undergoing OPCAB versus ONCAB, regardless of the revascularization technique used. In that large database study, OPCAB was however associated with a lower incidence in postoperative neurological events, in spite of a higher need for dialysis. [79] As detailed above, when OPCAB is performed with total arterial grafts and in a no-aortic-touch fashion, superior outcomes have been reported. [80], [81],[82],[83],[84]

Converted Patients

Conversion rates of OPCAB to ONCAB amongst randomized trials have ranged from 0% to over 20%, [68] and reported mortality for this converted group ranged from 6% to 15%. In the recent CORONARY study, 8% of the OPCAB cohort was converted, one-half of whom were converted due to hypotension or ischemia, and the other half were converted for small or intramuscular coronaries. [74] Conversion is a crucial outcome to assess because it is a reliable surrogate for surgeon experience/expertise and because conversion to ONCAB is associated with increased mortality and morbidity both in the near term and long term, particularly when patients are converted emergently. [74]