Keeping the coronary arteries in place during beating-heart bypass necessitates using one of three types of coronary stabilizers: compression-type stabilizers, which use two-pronged forks that grasp the epicardium on either side of the artery to be bypassed; suction-type stabilizers, which use forks whose grasping surfaces have suction ports that improve their grip without the need for greater downward force; and capture-type stabilizers, fenestrated platforms through which silicone elastic tapes are passed, looped under the coronary artery, and pulled tight to hold the epicardium against the underside of the platform (Figs. 25.15, 25.16 and 25.17).

In some patients, the left hemisternum can block the surgeon’s view of the proximal obtuse marginal and posterolateral branches of the circumflex arteries, making it difficult to graft coronary arteries onto the lateral and posterolateral areas of the heart. Attempts to correct this problem by forcing the heart to the right with stabilization devices may compromise both blood flow and stability. However, several devices and techniques have been developed to improve exposure while maintaining hemodynamic stability during OPCAB. Deep pericardial sutures (Fig. 25.18), first described by Lima [79], can be tightened to create a pericardial ridge that supports the base of the lateral left ventricle, allowing the heart to be rotated to the right so that the apex points upward through the sternotomy incision. This provides adequate exposure of the lateral and inferior left ventricle. Apical suction devices, which enable the surgeon to pull the apex of the heart away from the base and pivot the cardiac axis, can also be used for this purpose; animal studies have shown that these devices are as effective as deep pericardial sutures [80, 81] and may cause less hemodynamic interference [80]. When it is difficult to expose the lateral wall, right pleurotomy and right vertical pericardiotomy (Fig. 25.19) may be used in addition to deep pericardial sutures to allow the heart to herniate into the right pleural space while maintaining right ventricular end-diastolic volume (Fig. 25.20). Performing asymmetric right hemisternal elevation in addition to the preceding techniques permits even greater lateral wall exposure, bringing the proximal obtuse marginal and posterolateral branches into the center of the operative field.

Maintaining a bloodless surgical field is necessary for constructing the coronary anastomoses during OPCAB. Normally, this is done by using temporary vascular occluders (e.g., clamps, vessel loops, snares, intracoronary shunts) on the coronary arteries. Because these devices can injure the arteries or cause endothelial dysfunction, some investigators have tested the use of a thermosensitive polymer (LeGoo; Pluromed, Woburn, MA), which is liquid at room temperature but solidifies into a gel at body temperature, to prevent blood flow into the surgical field during OPCAB. An initial human trial showed that the polymer could completely block blood flow from the coronary arteries without further intervention in 91 % of attempts [82]. Subsequently, other investigators found that using the polymer more often resulted in satisfactory hemostasis than using vessel loops (88.0 % vs. 60.7 %) and shortened the total anastomotic time (12.8 vs. 15.1 min). Thirty-day adverse event rates did not differ significantly between the two groups [83]. However, the polymer is not necessarily a perfect alternative to occlusive devices; results of an ex vivo study of the polymer’s effect on the IMA indicate that it may cause endothelial injury and smooth muscle dysfunction [84].

As stated in the 2011 ACCF/AHA Guideline for Coronary Artery Bypass Graft Surgery [85], there is not yet complete agreement about the relative safety and efficacy of OPCAB and conventional CABG with CPB. The difficulty in comparing the two approaches stems in part from the low incidences of mortality and significant morbidity associated with both techniques [86–97]. However, in the past decade, several large clinical trials, meta-analyses, and registry studies have been performed that had sufficient sample sizes to address this problem. A 2005 meta-analysis of 37 randomized, controlled trials of OPCAB suggested that it is associated with significantly decreased atrial fibrillation, transfusion and inotrope requirements, respiratory infections, ventilation time, intensive care unit stay, hospital stay, and overall in-hospital and (1-year) post-discharge costs compared with standard CABG [39]. Likewise, a randomized controlled trial (CORONARY) performed in 4,752 patients at 19 centers associated OPCAB with less blood transfusion, lower frequencies of reoperation for bleeding, and lower 30-day rates of acute kidney injury and respiratory complications than CABG [98]. These findings may be explained by the results of several small randomized trials, which suggest that OPCAB is associated with lower postoperative CK-MB [97, 99, 100] and troponin levels [97, 101] and less activation of complement factors C5a and TCC [102] than is standard CABG. Additionally, these studies suggest that OPCAB patients may have less blood loss during surgery [100] and a higher postoperative hematocrit [97].

Other studies of the short-term outcomes of CABG and OPCAB have produced less definitive results. Several studies have found in-hospital or 30-day mortality to be lower with OPCAB than with CABG [103–106], but a similar number of studies found no difference between the two procedures [98, 107–110]. With regard to short-term stroke risk, a meta-analysis of 59 randomized trials (n = 8,961) associated OPCAB with a 30 % reduction in 30-day stroke risk [107]. This is consistent with the results of two large, propensity-matched registry studies (CREDO-Kyoto [111] [n = 6,323] and a 2008 study by Puskas et al. [104] [n = 12,812]) and a large retrospective study (n = 13,108) [105], all of which showed lower short-term stroke rates in OPCAB patients than in CABG patients. In contrast, the ROOBY trial [109], the CORONARY trial [98], and a very large (n = 63,047) Veterans Administration registry study [108] found no differences in short-term stroke rates.

With regard to long-term outcomes, the evidence is similarly mixed. One study found that 10-year survival was somewhat poorer in OPCAB patients than in propensity-matched CABG patients [112]. However, Puskas et al.’s registry study [104] and the SMART trial [104] found no differences in long-term survival. This is despite the fact that several studies have associated CABG with better short- and long-term graft patency than OPCAB [109, 110, 113], although this finding is not universal [114].

Among CABG and OPCAB patients surveyed an average of 10.8 months after surgery, OPCAB patients reported having better physical functioning and emotional well-being [115]. However, other studies found that choice of procedure made no difference in terms of quality of life at 18 months [116] or in functional class or treadmill performance at 1 year [110].

Two retrospective studies have identified gender differences in CABG and OPCAB outcomes. In both studies, OPCAB was associated with lower 30-day mortality than CABG, but only in women [103, 104]. One of these studies also found that OPCAB enhanced 1-year survival in women [103]; the other study found that women had poorer 10-year survival than men regardless of which procedure they underwent [104].

A few studies have compared the results of CABG or OPCAB with those of PCI. One such study found that that OPCAB and percutaneous angioplasty with sirolimus-coated stents were associated with similar early and late mortality rates, but that the OPCAB patients had a lower rate of reoperation and return of angina [117]. Likewise, a registry study that found no differences between CABG and OPCAB in terms of outcomes showed that both procedures produced lower 18-month revascularization rates than PCI [116]. In contrast, a single-center, retrospective study found that the 2-year rates of survival and freedom from major adverse cardiac events were both higher in patients who underwent bypass procedures (83 % of which were OPCAB) than in patients who underwent PCI [118].

Early OPCAB procedures were limited in terms of the adequacy of revascularization and the quality and reproducibility of the distal anastomosis, leading surgeons to perform slightly fewer grafts in OPCAB patients than in conventional CABG patients [92, 93]. However, subsequent refinement of devices and techniques for lateral and inferior wall exposure (see above) has enabled surgeons experienced with OPCAB to perform grafts to all aspects of the heart, and recent randomized trials show that the average number of grafts constructed is at most only slightly smaller in OPCAB procedures than in CABG operations (the largest difference being 3.0 for OPCAB vs. 3.2 for CABG) [98, 109, 114].

Several retrospective nonrandomized series have shown that OPCAB may be associated with decreased operative risk in some high-risk groups [86, 119–121], including elderly patients, patients with reduced ejection fraction, and patients with renal failure [87, 122–127]. In addition, OPCAB may be well suited for patients with left main disease [128]. Furthermore, several reports have documented that OPCAB is associated with decreased risk in repeat CABG patients [129–131]. This is because reoperative OPCAB often can be performed without dissecting out the ascending aorta or manipulating diseased grafts. It is less clear whether OPCAB is better than standard CABG for patients with left ventricular dysfunction [132, 133], although such patients may benefit from OPCAB with temporary support from a ventricular assist device [134].

In dealing with heavily diseased but patent vein grafts, many surgeons reject OPCAB because of the risk of graft embolization while the heart is being repositioned. However, in patients with patent LIMA-to-LAD grafts, a tailored left thoracotomy approach for reoperative grafting of the lateral wall allows the construction of a graft from the descending thoracic aorta to lateral wall branches and avoids the potential complications of LIMA-graft mobilization [135, 136]. Similarly, lower sternotomy and upper abdominal incisions have been described for constructing grafts to the inferior wall with the right gastroepiploic artery [137, 138]. Long-term patency data for these types of grafts are not yet available.

Off-pump bypass is ideal for patients with significant atheroma or calcification of the ascending aorta. In these cases, free grafts are attached either to the side of the in situ LIMA, as first described by Tector and colleagues [139], or to the innominate, subclavian, or axillary artery if it is disease-free [140]. Alternatively, free grafts can be attached to a disease-free area of the aorta with an automated anastomotic device [141], or with a clampless anastomotic facilitator that obviates the need for a partial-occluding clamp. Recent series suggest that using these clampless approaches to OPCAB results in low perioperative stroke and mortality rates [141, 142], perhaps lower than those associated with conventional CABG [143, 144]. Palpation and epiaortic echocardiography, when used together, can frequently identify an appropriate dime-sized area of relatively normal aorta. Using OPCAB techniques enables relatively easy revascularization of these high-risk patients and avoids certain hazards associated with cannulation and cross-clamping of the diseased aorta.

The incidence of new onset of subtle neurocognitive dysfunction (NCD) after CABG has been reported to be between 5 and 60 %, depending on the sensitivity of the tests used. Specific deficits include short-term memory loss, impairment in performing simple calculations, and personality and mood disturbances [145–149]. Many factors related to CPB have been implicated, including SIR, alterations in cerebral blood flow, and microembolization caused by the CPB circuit or by cannulation and cross-clamping [147, 149, 150].

A few studies have associated OPCAB with a lower incidence of NCD than conventional CABG with CPB [150–152]. However, other studies have shown no significant difference [109, 110, 153–158]. The difference may be partly attributable to the time frame examined; a meta-analysis of 13 studies by Sun et al. [159] showed that CABG patients had more NCD than OPCAB patients did during the first 3 months after surgery, but this difference was no longer apparent when the patients were tested 6 months and 1 year postoperatively. Although several groups have reported that OPCAB is associated with significantly less SIR than is CABG [87, 160, 161], the magnitude of SIR has not been shown to correlate with the severity of NCD in individual patients [162]. The failure of OPCAB to reduce postoperative NCD has been largely attributed to the liberation of microemboli during application and removal of the partial-occlusion aortic clamp. Although OPCAB eliminates the need for a perfusion catheter for CPB or a cross-clamp for cardiac arrest, using the side-biting partial occlusion clamp, which is routinely done at many centers during the construction of proximal anastomoses, may be equally traumatic [163]. However, at least one small, randomized, prospective study has associated OPCAB with reduced microembolization—but not a reduced incidence of NCD—compared with standard CABG [164]. In several series, completely eliminating aortic manipulation by performing OPCAB with the exclusive use of in situ and composite grafts produced significantly lower rates of stroke than conventional CABG with CPB or OPCAB performed with a side-biting clamp [164–167]. Additional series in which aortic manipulation was minimized by using an automated anastomotic device (Cardica Inc., Redwood City, CA) to perform the proximal anastomosis or in which the Heartstring proximal seal system (Maquet Cardiovascular, Wayne, NJ) was used to avoid application of the partial occlusion clamp have shown similar results [141–144].

First introduced by Benetti and colleagues in 1995 [170], the minimally invasive direct coronary artery bypass (MIDCAB) procedure involves performing a 7-cm anterior lateral thoracotomy, usually in the 5th intercostal space. By this route, single-vessel off-pump bypass can be performed, using the LIMA to bypass the LAD. Several early reports showed that this technique substantially shortened hospital length of stay relative to standard CABG, and it also reduced transfusion requirements [135, 171–173]. One study also associated MIDCAB with a reduced incidence of atrial fibrillation [174], but two other studies did not support this finding [175, 176]. Results of small randomized trials suggest that, in patients with stenosis of the proximal LAD, MIDCAB is associated with a longer in-hospital recovery period [177] and higher costs than is direct stenting [178], but that MIDCAB produces similar or better short- and long-term outcomes [179, 180]. Nonetheless, in the United States, only a few centers regularly perform MIDCAB today.

Early in the MIDCAB experience, many U.S. centers mobilized the LIMA by using direct visualization through the thoracotomy (Figs. 25.21 and 25.22). This method was often technically demanding, and the forceful retraction of the chest wall necessary to expose the LIMA could cause substantial early postoperative pain. Furthermore, LIMA grafts produced by this technique may be too short. Nevertheless, U.S. centers that perform MIDCAB have frequently reported excellent results [181–185].

More recently, there has been resurgent interest in videoscopic LIMA mobilization (Figs. 25.23, 25.24 and 25.25). With this technique, the left lung is collapsed while right-lung ventilation is maintained, the left pleural space is insufflated with CO₂, and the mammary artery is harvested from the back of the chest wall with 5-mm videoscopic instruments. This technique avoids the aggressive chest wall retraction required for direct mammary artery harvest and allows the entire length of the artery to be safely mobilized. Although this technique is associated with a significant learning curve, its success has been well documented [186–190].

For patients with multivessel CAD, combining a LIMA-to-LAD bypass (i.e., MIDCAB) with endovascular intervention in the circumflex and right coronary arteries may be a good alternative to standard CABG. A multicenter series [191] and a small, single-center series [192] have shown that videoscopic mobilization of the LIMA and opening of the pericardium, using a voice-operated robotic arm to manipulate the scope, enables the surgeon to place the chest wall incision more accurately. This permits construction of the LIMA-to-LAD anastomosis without a rib spreader; instead, only soft tissue retractors and a specially-designed, bed-mounted stabilizer are used to expose the anastomotic site through the natural intercostal space [188]. These patients are also taken to the catheterization lab, where stents are placed in the circumflex and right coronary arteries. This is typically done several days or even weeks before or after the surgical portion of the procedure [193, 194], although both stages can be performed on the same day [195]. The hybrid MIDCAB approach significantly reduces postoperative pain [196], and preliminary patency and outcome data seem promising [194], but comparative studies are needed to justify the expense associated with this approach before it can be widely adopted.

Several reports have documented totally videoscopic LIMA-to-LAD procedures in which surgical robots were used to mobilize the LIMA and perform the sutured anastomosis, both on-pump [197–199] and off-pump (Figs. 25.26 and 25.27) [197, 200, 201]. These robots use master-slave servomotors and microprocessor control to precisely emulate and miniaturize the surgeon’s movements, enabling the surgeon to manipulate videoscopic instruments with multiple degrees of freedom (Fig. 25.28). A high-resolution, 3-dimensional display provides stereopsis with a high degree of magnification (Fig. 25.29). Tremor filtering and motion scaling can be adjusted to the surgeon’s preferences, enabling tremendous precision, which compensates somewhat for the lack of tactile feedback (although tactile feedback systems are under development [202]).