Introduction

Off-pump coronary artery bypass grafting (OPCAB) is unlikely to ever become a standard operation that an average surgeon can routinely perform. OPCAB is a technically demanding procedure that is associated with negative results in some cases, including incomplete revascularization and conversion to emergency on-pump surgery, which can result in serious complications. Ideally, coronary artery bypass grafting (CABG) should provide complete revascularization in one operation and should provide a better long-term outcome than percutaneous coronary intervention (PCI). If these two ideals are not realized, the procedure can no longer be described as “CABG.”

Previous large randomized studies have indicated a significantly worse outcome of OPCAB than the on-pump technique [1–4], which has led to less frequent use of this technique worldwide. However, any experienced OPCAB surgeon will find it difficult to agree completely with these observations. As might be expected, experienced OPCAB surgeons are able to maintain a consistent level of quality in on-pump CABG. If a rate of conversion to on pump of more than 10 % is observed [1–4], the technique can no longer be described as “OPCAB,” and the surgeon should train to master adequate technique for OPCAB. Experienced off-pump surgeons have no doubt that OPCAB has several advantages for high-risk patients and can provide an equal or superior outcome to on-pump CABG in both the short- and long terms. In the present review, we discuss the current frontline strategy for high-quality OPCAB.

Major previous studies

A number of previous prospective randomized trials and meta-analyses have failed to show a benefit of OPCAB in terms of either the short- or long-term clinical outcome. The first adequately powered randomized trial to compare OPCAB with on-pump CABG was the Randomized On/Off Bypass (ROOBY) trial [1, 4], which strongly questioned the value of OPCAB. The trial enrolled 2,203 selected patients from 18 Veterans Affairs medical centers and 53 attending surgeons. The participating surgeons were required to document that they had performed at least 20 off-pump CABG procedures, including some in which complete revascularization was achieved for all vascular territories of the heart. The median extent of pre-study experience with off-pump surgical procedures was only 50 cases. Furthermore, a number of inexperienced cardiothoracic trainees (postgraduate years 6–10) were designated before randomization as the primary surgeons. The proportion of residents among the primary surgeons was 55.4 % in the off-pump and 64.0 % in the on-pump group. The STS risk scores and patient risk factors were equally distributed between the two groups. The primary 30-day results and primary long-term (30 days–1 year) end points were evaluated. The secondary end points included the completeness of revascularization, graft patency at 1 year, and neuropsychological testing results. There were no differences in the 30-day mortality or short-term major adverse cardiovascular events between the two groups. The OPCAB patients received significantly fewer grafts per patient, with 2.9 ± 0.9 in the off-pump group and 3.0 ± 1.0 in the on-pump group. However, the difference between the two groups was smaller than expected. Disappointingly, the conversion rate to on-pump surgery reached 12.5 %. No significant differences were observed in the mortality or major adverse cardiovascular events at 30 days. However, after 1 year, the OPCAB patients had a higher incidence of cardiac-related death and major adverse events and a lower rate of graft patency. All of these differences may be explained by the poor technique of the surgeons who were not sufficiently familiar with the off-pump technology. In addition, the study was based on the data from more than 10 years ago, which were collected in the early days of OPCAB, which has continued to develop since that time. It certainly takes longer to master the off-pump technique than on-pump CABG. The study simply demonstrates that technically deficient coronary revascularization leads to poor outcomes and also confirms the technical difficulty of the OPCAB technique.

A recent important study which remedied the shortcomings of the ROOBY trial is the CABG Off- or On-Pump Revascularization Study (CORONARY) [5, 6], which enrolled 4,752 patients from 79 centers in 19 countries. The participant surgeons were more experienced, having had at least 2 years of surgical practice, and having completed more than 100 procedures involving the specific technique. No procedure was admitted in which a trainee had acted as the primary surgeon. Moreover, compared to the ROOBY trial, the enrolled patients were more likely to be elderly and female, and to have left main or triple vessel disease, and to have undergone urgent surgery. They thus had more severe pathologies and were therefore more representative of a real-world background. It is well known that the advantageous effects of the off-pump technique are greatest in such high-risk patients. At 30 days, the primary composite outcome (death, stroke, myocardial infarction, new renal failure requiring dialysis at 30 days) showed no significant inter-group differences (9.8 % in the off-pump group vs. 10.3 % in the on-pump group). The use of the off-pump technique resulted in reduced rates of transfusion, reoperation for bleeding, respiratory complications, and acute kidney injury, but led to an increased risk of early revascularization. At 1 year, there were no significant differences between the two groups with respect to the primary composite outcome, the rate of repeat revascularization, quality of life, or neurocognitive function. Although the CORONARY study had fewer grafts per patient and a lower rate of complete revascularization, it was generally accepted that the study design and the skill level of the participating surgeons were superior to those of most surgeons. The study successfully demonstrated that the level of surgical expertise may significantly affect the early and long-term outcomes.

Arterial grafting

Multiple arterial grafting

Since the development of the OPCAB technique, the trend in revascularization strategies has been toward in situ all-arterial grafting because of the benefits of the aorta no-touch technique and better long-term clinical outcomes. A number of previous studies have shown the excellent effects of using arterial conduits, which maintain a high rate of long-term patency and provide improved clinical outcomes [7–9]. There are now three reliable in situ arteries [the two internal thoracic arteries (ITA) and the right gastroepiploic artery (GEA)] and one free graft (the radial artery) that can be used. The use of the ITA is generally known to be associated with low rates of mortality and reintervention. A number of important reports have demonstrated that bilateral ITA use offers the best long-term survival and the lowest rates of reintervention [10–12]. Since Buxton et al. [13] and Lytle et al. [12] first demonstrated the long-term efficacy of bilateral ITA grafting, it has been gaining acceptance among surgeons. Grafting of the bilateral ITAs to the left coronary system and additional grafting of the GEA to the distal right coronary artery (RCA) have been reported to provide good long-term outcomes [14–16]. Using three in situ arterial grafts, such as the bilateral ITAs and the GEA, can allow for complete avoidance of aortic manipulation. The now ongoing arterial revascularization trial (ART), the first randomized trial to compare the clinical results between bilateral and single internal mammary coronary artery bypass grafting, is expected to give evidence of the clinical advantages of bilateral ITA grafting with a 10-year follow-up [17].

How to use both the ITAs

The level of clinical evidence for the long-term effects of using bilateral ITAs has increased during the past decade. If the patient requires reconstruction of both the left anterior descending artery (LAD) and the circumflex artery (CX) area, the left and right ITAs should be routinely used in combination. It is, however, unclear how the two ITAs should be grafter and to which area in order to achieve the best outcomes. OPCAB surgeons should discuss the best grafting model for both ITAs. Two common combinations for ITA placement are as follows: first, grafting of the right ITA (RITA) to the LAD across the midline and grafting of the LITA to the CX; second, grafting of the RITA to the CX through the transverse sinus and grafting of the LITA to the LAD. With regard to which graft is better, the general understanding is that the early clinical results, graft patency rate, and technical difficulty are similar for these two combinations. We prefer the RITA to LAD and LITA to CX combination, because we often experience cases that require CX–CX sequential grafting (Figs. 1, 2). The LAD reconstruction is the most important in CABG, so we think that it should be anastomosed individually by one ITA. It has been argued that the RITA is too short to graft to the distal LAD, but this applies mainly for cases with a pedicled RITA. When the RITA is skeletonized with an ultrasonic scalpel, it is long enough to reach the distal LAD (Figs. 3, 4). Some studies have shown good results for composite Y- or T-grafts of the free RITA to the in situ LITA, which may increase the number of anastomoses.

Fig. 1
figure 1

In situ three arteries (right internal thoracic artery to LAD, left internal thoracic artery to circumflex artery, gastroepiploic artery to distal right coronary artery)

Full size image
Fig. 2
figure 2

Sequential anastomoses with left internal thoracic artery

Full size image
Fig. 3
figure 3

In situ skeletonized three arteries

Full size image
Fig. 4
figure 4

Skeletonized right internal thoracic artery is enough long to reach the distal LAD

Full size image

Skeletonization with an ultrasonic scalpel

For high-quality OPCAB, the skeletonization technique is now essential to optimize the condition of the arterial graft. Skeletonization has many advantages, such as the avoidance of early spasms, easy identification of bleeding, improved vessel quality, achievement of functional lengthening and a larger caliber with maximum flow, ease in performing sequential anastomosis, and preservation of the sternal blood flow. Skeletonization for harvesting the ITA has gradually gained acceptance among surgeons. Higami et al. [18] first described ultrasonic ITA skeletonization and revealed its technical feasibility and advantages. They showed that the skeletonized ITA averaged 4 cm longer than the pedicled conduit and had a free flow rate greater than 100 mL/min, which is at least 20 % higher than that of the pedicled ITA. A number of previous studies revealed that skeletonization of the ITA lowers the risk of deep sternal wound infection, even in diabetic patients with bilateral ITA use [19–21]. Shi et al. [22] examined the endothelial function of the GEA skeletonized with an ultrasonic scalpel in an ex vivo experiment using acetylcholine and bradykinin. They concluded that skeletonization with an ultrasonic scalpel is as safe as non-skeletonized dissection in preserving the contractile properties and endothelium-dependent and endothelium-independent relaxations of the GEA. Fukata et al. [23] also reported that thermal degeneration was limited to the surrounding connective tissue, and that the media or intima of the vessel wall was not affected by the ultrasonic scalpel.

In our experience, using an ultrasonic scalpel improves the technical ease of the procedure, shortens the harvesting time, and increases the effective length and free flow of the ITAs. Surgeons should routinely master the ultrasonic skeletonization technique for the effective use of the ITAs in OPCAB surgery.

The right gastroepiploic artery

The history of using the right GEA in coronary revascularization was started in the 1960s, when Bailey reported the Vineberg implantation of the GEA into the posterior area of the heart. The use of the GEA for direct coronary artery anastomosis was first reported by Pym [24] and Suma [25] in 1987, but still has not gained acceptance among cardiovascular surgeons worldwide. The reasons for its less frequent use include concerns about insufficient flow capacity and vasospasm, and the risk of competitive flow causing graft failure [26].

The GEA is the main branch of the gastroduodenal artery, which arises from the right hepatic artery. In contrast to the IMA, the GEA has fewer elastic lamellae in its media and is classified as a muscular artery [27]. The GEA provides at least 20 cm of useable length, which can reach all areas of the heart. Skeletonization using an ultrasonic scalpel makes the GEA longer and wider, so that it can be anastomosed at a more proximal position than the pedicled GEA. The suitable targets of the in situ GEA are the distal right coronary artery or the distal circumflex artery. It is feasible to perform sequential grafting using the skeletonized GEA due to its sufficient length and diameter.

Several studies have indicated a good patency rate for GEA grafting in the early postoperative period. First, in 1996, Voutilainen et al. [28] reported an angiographic late patency rate for GEA graft (82.1 % at 5 years) that was close to that of the internal thoracic arteries and superior to that of the saphenous vein graft (SVG). Hirose et al. [29] reported a series of 1,000 isolated CABG cases using the GEA that showed a patency rate of 98.7 % at 1 year, 91.2 % at 3 years, and 84.4 % at 5 years. Suma et al. [30] reported on 20 years of experience with GEA grafting that showed a cumulative patency rate for GEA grafts of 97.1 % at 1 month, 92.3 % at 1 year, 85.5 % at 5 years, and 66.5 % at 10 years. According to these limited data, the long-term patency rate of the GEA ranged from 82 to 86 % at 5 years, which is superior to that of the SVG, but inferior to that of ITA grafts. However, in the previous studies, the GEA graft was used in a pedicled fashion.

The patency rate of the skeletonized GEA is reported to be better than that of the non-skeletonized GEA. Kim et al. [31] evaluated the early and 1-year postoperative results of skeletonized GEA grafting to the RCA and found an excellent early patency rate of 98.3 % and a 1-year patency rate of 92.0 %. There is even less data available on the long-term patency of the skeletonized GEA, for which there are no accurate figures. However, in 2013, we presented a single-center report indicating that the cumulative patency rate of the skeletonized GEA was 97.8 % at 30 days, 96.7 % at 1 year, 96.0 % at 3 years, 94.7 % at 5 years, and 90.2 % at 8 years after surgery, which indicated that it was superior to the pedicled GEA or saphenous vein grafts [32] (Fig. 5).

Fig. 5
figure 5

The cumulative patency rate of the skeletonized GEA was 97.8 % at 30 days, 96.7 % at one year, 96.0 % at three years, 94.7 % at five years, and 90.2 % at eight years after surgery

Full size image

The skeletonized GEA is a reliable conduit and assumes a large role in the total arterial OPCAB strategy. Since 2001, we have consistently used the GEA in a skeletonized fashion using ultrasonic scissors. Our technique for harvesting the skeletonized GEA using an ultrasonic scalpel was previously described and has proved to be simple and safe [27, 33]. Removal of the surrounding tissue makes the GEA longer and wider, so that it can be anastomosed at a more proximal position than the pedicled GEA. The intraoperative maximal dilation and maximal length may make a major contribution to better long-term GEA patency.

Flow competition may occur when the GEA is anastomosed to a coronary artery with low-grade stenosis. As a general rule, we use the GEA only when the coronary artery stenosis is severe. Shimizu et al. [26] examined the difference in flow characteristics between the GEA and SVG using a Doppler-tipped guidewire and found that, when the coronary artery stenosis was moderate, the flow volume of the GEA was less than that with the SVG. This difference was explained as a consequence of flow competition between the GEA and the native coronary artery. We think that late graft occlusion may occur in association with flow competition. In our 2013 study, we also discovered through a multivariate analysis that low-grade (<75 %) target vessel stenosis was a strong risk factor for late GEA occlusion. We therefore recommend again that the GEA should, in principle, be used only for cases of more severe stenosis.

Surgeons have not necessarily achieved a sufficient understanding of the potential of the GEA, which has been undervalued. We think that the GEA will play a critical role in OPCAB procedures with in situ all-arterial grafting, and that surgeons should discuss and acquire a detailed grasp of the potential of the GEA.

The radial artery

The radial artery is a convenient graft that has many advantages, including the ease of harvesting, sufficient length for grafting to almost any coronary artery territory, and a suitable caliber size to match the coronary artery. It is frequently used as the second graft of choice after the ITAs. The first use of the radial artery in CABG was reported by Carpentier et al. [34], but it was soon abandoned because of disappointing early angiographic outcomes due to the unpredictable development of spasms. After improvement in the harvesting technique, management through pharmacologic dilatation, and the use of calcium channel blockers, the radial artery was reviewed on the basis of encouraging reports in the 1990s. The radial artery is a muscular artery and has more spastic characteristics than other arterial conduits. To prevent spasms, intravenous nitrates or calcium channel blockers should be used intraoperatively as well as postoperatively until the patient can take oral medication. To preserve the endothelial function, the surgeon should take a care to avoid excessive manipulation of the radial artery and should not mechanically dilate it during harvesting.

Efforts to achieve an aorta no-touch strategy in the OPCAB procedure have led to increased use of the RA as a composite graft. Although the patency rate of radial artery composite grafts is acceptable, direct aortocoronary bypass is superior to composite grafting. Careful consideration is therefore needed given the risk of undesirable blood-flow distribution due to an unbalanced vascular area. The early (~12 months) patency rate of the radial artery is favorable and can be expected to be more than 90 %. Recently, the long-term (~5 years) patency rate has been reported to be between 70 and 98 %. The patency of the radial artery is more significantly influenced by target vessel stenosis than are other conduits. It is said that if the proximal stenosis is 90 % or more, the radial artery patency is similar to that of ITA-RA composite grafts and direct aortocoronary bypass. Therefore, when the proximal native stenosis is mild, the radial artery should be used for direct aortocoronary anastomosis. Under the current conditions in aorta no-touch OPCAB, the radial artery is a supplemental conduit used in cases where the bilateral ITAs are not available, or where complete reconstruction of the left side coronary area cannot be achieved with the bilateral ITAs alone.

Grafting pattern

Composite grafting

The use of the bilateral ITAs enables OPCAB operations to be performed without manipulation of the ascending aorta using an in situ or Y-configuration. Many surgeons currently use composite T- or Y-grafting with the free right ITA or use the radial artery attached end to side to the left ITA. A common arrangement is the composite T-graft, whereby free ITA grafts are attached proximally end to side to the in situ ITA. Although several studies [35–37] reported that the clinical and angiographic results of composite grafting were equivalent to those of individual grafting, there are some contrary opinions that composite grafts may be susceptible to the detrimental effects of flow competition with the native coronary artery when used for a mildly stenosed target vessel [35, 38, 39].

When the myocardial area revascularized with a composite ITA graft is large, a major concern is that the single attached LITA may be unable to supply enough blood. Studies using transit-time Doppler techniques have indicated that construction of composite arterial grafts results in a significant increase in flow through the left ITA. The amount of flow supplying each region depends on the severity of coronary stenosis and the coronary vascular area. Composite arterial grafting causes splitting of the internal thoracic artery flow to various myocardial regions based on the differences in native blood flow. Nakajima et al. [38] presented an angiographic study of 362 patients with composite T-grafts, which found competitive flow in 14.6 % of the composite grafts and occlusion in 3.6 % of the patients. Manabe et al. [39] showed that the angiographic outcomes of composite grafts were closely related to the severity of stenosis in the target coronary artery. They found that composite grafting was an independent predictor of graft occlusion or the string sign in cases with target vessels exhibiting mild stenosis. The failure of an ITA graft to the LAD as a result of competitive flow deprives the patient of the main benefit expected from the operation, such that an equivalent result could have been achieved using a simple in situ ITA. Lev-Ran et al. [35] reported that the early results of bilateral ITA grafting with T-grafts were comparable with those of in situ grafts, but increased angina and decreased midterm survival led them to recommend in situ grafting whenever technically possible. However, composite grafting plays a crucial role in these procedures, because it eliminates the need for proximal anastomosis to the ascending aorta and conserves extra lengths of the arterial graft for additional grafting. We therefore think that composite T- or Y-grafting with two ITAs should be employed for selected patients with severe stenosis in the LAD and the marginal branches of the circumflex artery.

In situ grafting

Arterial grafts are known to narrow diffusely or to occlude when used under low-flow conditions. Even in vessels with a stenosis degree as low as 50 %, however, the patency rate of an in situ ITA to LAD graft was over 90 %. This excellent patency of in situ ITA to LAD grafts was long-lasting and remained high up to 15 years or more after surgery. Although the results of composite grafts are acceptable, almost all reports showed a better patency rate and long-term outcomes for in situ than composite grafts. Reduced patency rates for free RITA grafts have been demonstrated when these grafts are connected proximally to the aorta.

The main concerns associated with in situ right ITA grafting to the LAD are insufficient length and the proximity of the crossover right ITA to the sternum, which could compromise a subsequent repeat sternotomy. Refinements in ITA skeletonized harvesting techniques have increased the graft length and improved the distal free flow, and may reduce postoperative sternal wound complications. If the length of the crossover right ITA is not sufficient to comfortably reach the desired anastomotic site on the LAD, we consider using the T-graft arrangement. However, in our OPCAB cases treated during the last 2 years, the T-graft technique was implemented in only 0.5 % of cases. Thus, a skilled skeletonization technique for ITA harvesting can resolve the issues associated with an insufficient length of the RITA graft in almost all cases. We prefer in situ RITA grafting anterior to the aorta because of the technical ease and the patency rate equivalent to in situ LITA grafting to the LAD (Fig. 5).

A major concern in the anterior retrosternal RITA crossover route is the potential risk of damage to the artery during repeat sternotomy, and we have therefore taken comprehensive measures to prevent injury to the crossover RITA. During this approach, the RITA is moved into a tunnel in the right pericardium and directed leftward across the midline of the ascending aorta toward the LAD. Mediastinal fat is used to cover the RITA. A space is thus maintained between the crossover RITA and the posterior table of the sternum for future resternotomy. This maneuver allows free space in the aorta and provides a safe distance between the crossover ITA and the sternum.

The use of sequential grafting is essential to achieve complete revascularization with only in situ arterial grafts. The two ITAs in combination with the right gastroepiploic artery provide three sources of blood supply. When OPCAB is planned using the aorta no-touch technique and bilateral ITA grafts, the right coronary artery can be bypassed with an in situ right GEA or an extension of the composite ITA graft. The LAD should be reconstructed individually with one (in most cases the right) ITA. We often use the in situ left ITA as a sequential graft for CX reconstruction. The left ITA may be used for up to two sequential anastomoses, but three anastomoses are difficult. The skeletonized right GEA is suitable for sequential grafting and has sufficient length and diameter for up to three or even four sequential anastomoses.

Of our consecutive elective OPCAB cases treated during the last 2 years, over 90 % were performed using an in situ all-arterial grafting technique with an aorta no-touch policy. Thus, in almost all cases, complete revascularization could be achieved with three in situ arterial conduits (both ITAs and the right GEA). We use the SVG in specific cases, such as in patients with previous gastrectomy, an RCA target with mild stenosis, or with a severely calcified right GEA.

The aorta no-touch technique with all-arterial grafting

Efforts to reduce the risk of stroke during CABG have led to the widespread use of an aorta no-touch strategy with the off-pump technique. The incidence of stroke in patients undergoing routine CABG is reported to be 1–3 %, most commonly as a result of ascending aortic embolic phenomena. It has been hypothesized that the best strategy to reduce stroke risk appears to be to completely avoid handling the aorta. The main indication for a no-touch aorta OPCAB operation is patients who have a severely atherosclerotic or porcelain ascending aorta. Kim et al. [40] reported a lower incidence of perioperative stroke with the no-touch aorta OPCAB technique, but similar stroke rates between OPCAB with aortic manipulation and conventional on-pump CABG. Kapetanakis et al. [41] reported that the stroke rate of on-pump CABG was 1.5 times (2.2 vs. 1.6 %) that of off-pump CABG with partial aortic clamping and three times (2.2 vs. 0.8 %) that of no-touch aorta OPCAB. Based on their OPCAB results, Misfeld et al. [42] found that 1.4 % (81/5779) of patients who underwent aortic manipulation had strokes, compared with 0.5 % (29/5619) who were treated using the aortic no-touch approach. Lev-Ran et al. [43] reported one neurological event among 429 consecutive patients (0.2 %) in the no-touch OPCAB group, which compared favorably with a stroke rate of 2.2 % observed in the side-clamp OPCAB group. As another advantage, Kim et al. [44] demonstrated that OPCAB with complete avoidance of aortic manipulation significantly reduced the incidence of postoperative atrial fibrillation compared with conventional CABG (11.4 vs. 21.1 %). Even though the no-touch technique may be the best clinical practice, it may not be applicable to every patient and is not routinely applied in many centers. Despite all of the above evidence in favor of the no-touch technique, the current clinical setting still includes saphenous vein grafting with proximal anastomosis in patients with multi-vessel disease.

There is no doubt that the aorta no-touch technique has strong advantages in preventing the embolic complications associated with a diseased ascending aorta. Surgeons should endeavor to apply the aorta no-touch technique to reduce additional preventable embolic events.

OPCAB in high-risk cases

Poor left ventricular heart function

Unstable hemodynamics and a low left ventricular ejection fraction are frequently cited as contraindications to OPCAB. The major reasons given are the difficulty of optimal coronary artery exposure in cases with a low ejection fraction associated with enlarged ventricles and hemodynamic instability or severe rhythm disturbances during the displacement of the heart. Many surgeons prefer using CPB because hemodynamic instability, hypotension induced by ventricular arrhythmias, and cardiac arrest are frequent problems encountered in this specific group of patients. CABG using CPB has become safer due to recent developments in myocardial protection techniques, and in the majority of patients, the damaging effects on the myocardium are minimal and reversible, but patients with left ventricular (LV) dysfunction have very poor reserve and even slight damage to the myocardium may have significant consequences. Hemodynamic deterioration is the greatest concern during displacement of the heart in OPCAB surgery. Displacement of the beating heart may be well tolerated in patients with good LV function, but hemodynamic compromise occurs more often in patients with severe LV dysfunction, making it difficult to achieve complete revascularization using the OPCAB technique. However, several pioneering surgeons have taken on the challenge of adopting this technique for these high-risk patients. Ascinone et al. [45] reported a surgical mortality of 7 % in a series of 74 OPCAB patients with an ejection fraction (EF) <30 %. Arom et al. [46] found a surgical mortality rate of 2 % in their series of 45 OPCAB patients with a low EF (<30 %). Shennib et al. [47] compared the clinical outcomes of OPCAB and conventional CABG in patients with poor LV function (EF < 35 %) and reported that the surgical mortality was lower in the former group (3.2 vs. 10.9 %). However, there were fewer distal anastomoses per patient (2.8 vs. 3.9) in the OPCAB group. Recently, Keeling et al. [48] reported a large and propensity-matched OPCAB trial in patients with a low EF, in which they showed that OPCAB was associated with a significantly lower adjusted risk of death, stroke, and major adverse cardiac events, and that the prolonged intubation and postoperative transfusion rates were also significantly lower in the OPCAB group.

In actuality, a skilled OPCAB surgeon will not be discouraged by cases with an enlarged heart or LV dysfunction. The change in the ventricular filling volume during the displacement of the heart may produce only a marginal effect on the enlarged heart. A concentric hypertrophic heart is more irritably affected by a change in the filling volume. An enlarged ventricular chamber can buffer the change in the filling volume and minimize the impact on hemodynamic stability. The regional motion of the ventricular wall of an enlarged heart is smaller than that in a normal heart, which makes it easy to anastomose. When the familiarity with the OPCAB technique increased, data began to emerge citing the potential benefits of the avoidance of CPB for high-risk subgroups. Although it remains unproven, we feel that the greatest benefits from the avoidance of extracorporeal circulation will be observed not in low-risk individuals, but in patients with significant comorbidities.

Acute coronary syndrome

Patients with evolving acute coronary syndrome, defined as a continuum from unstable angina through non-ST-segment elevation MI to ST-segment elevation MI, are at a high risk of requiring surgery. The current indications [49, 50] for emergency CABG surgery in ASC patients are limited to those presenting with evolving myocardial ischemia refractory to optimal medical therapy, the presence of left main stenosis (≧50 %) and three-vessel disease, ongoing ischemia despite successful or failed PCI, complicated PCI, cardiogenic shock accompanied by complex coronary anatomy, or life-threatening ventricular arrhythmias thought to be caused by myocardial ischemia. An aggressive surgical approach, when appropriate, has been shown to be the superior approach for patients presenting with moderate- to high-risk ACS and is recommended by US and European guidelines for treating these patients. It is generally thought that the outcome in these patients may depend on factors such as the timing of the operation, the left ventricular function, the presence of collateral flow, and the presence of hemodynamic instability.

The adoption of the OPCAB technique in ACS patients undergoing emergency surgical revascularization remains controversial. Great efforts were made to identify high-risk subgroups that may benefit more from off-pump strategies. Some have speculated that patients with active ischemia may experience a reduced extent of myocardial damage during off-pump surgery, perhaps through avoidance of ischemic arrest and reperfusion injury, earlier revascularization of the culprit lesion, attenuation of the no-flow phenomenon, and reduction in myocardial edema, but clinical results have so far proved inconsistent, although it is logical that CPB and cardioplegic arrest may further damage the already jeopardized myocardium. These patients with ongoing myocardial ischemia are classically thought to benefit from cardioplegic protection, which reduces the metabolic demand and resuscitates the ischemic myocardium. However, avoiding the inflammatory reactions provoked by CPB and the global ischemia of cardioplegic arrest may be of equal or greater benefit to the already injured myocardium. Several retrospective studies have indicated that these high-risk patients may benefit most from avoiding CPB.

The surgical mortality when using conventional arrested heart CABG techniques in these patients was high (1.6–32 %) and strongly depended on the preoperative hemodynamic condition. Every et al. [51] reported that, in their 1,299 acute myocardial infarction (AMI) patients who underwent CABG, there was no difference in the hospital mortality rate between patients who underwent the operation during the first 24 h after admission and those who underwent the operation later in their hospital stay. In a multicenter analysis of 2,099 patients who underwent conventional CABG within 24 h after AMI, the hospital mortality was 14 %. Tomasco et al. [52] indicated a similar mortality rate of 13.4 % for patients operated on within 24 h after AMI. A retrospective analysis describing the outcomes of 12,988 unselected patients with ACS demonstrated that early CABG was an independent predictor of lower mortality [53]. Serge et al. found a markedly lower mortality rate of 1.6 % for this subset of patients.

On the other hand, there are very few examples in the literature that present the clinical results of off-pump CABG in ACS cases. Locker et al. demonstrated that the timing in itself is not a significant predictor of early mortality in off-pump CABG, but is a significant predictor in patients undergoing an operation with CPB. Rastan et al. demonstrated that beating-heart techniques implemented during active ischemia reduced in-hospital adverse outcome measures, such as the transfusion requirements, need for extensive inotropic support, prolonged ventilation time, intensive care unit stay, and in-hospital stroke rate [54]. Fattouch et al. reported [55] that in the on-pump group, the incidence of early death in patients presenting with and without shock was 27 and 3.7 %, but in the off-pump group, the corresponding rates were 7.5 and 0 %. They concluded that preoperative cardiogenic shock increases the early mortality after CABG in patients with AMI, specifically in those undergoing on-pump surgery. Biancari et al. [56] showed that off-pump CABG may reduce ischemic injury to the myocardium and suggested some early benefit of this technique for patients presenting with active ischemia and unstable angina. Moscarelli et al. [57] concluded from a meta-analysis of OPCAB outcomes in patients with ACS that OPCAB may have a beneficial effect on the 30-day mortality in hemodynamically stable patients undergoing emergency revascularization.

In the setting of OPCAB procedures, preoperative intra-aortic balloon pumping (IABP) therapy improves the cardiac performance and facilitates access to the target vessels. IABP support is beneficial to the displaced heart in terms of maintaining the hemodynamic stability during OPCAB [58]. Delayed use of IABP was associated with increased mortality, whereas the prevalence of perioperative myocardial infarction and in-hospital mortality was significantly lower among patients in whom IABP support was initiated preoperatively. IABP therapy should therefore be performed preoperatively in selected high-risk patients with definite indications. The most common indications for preoperative IABP placement before OPCAB were severe left main coronary artery disease, unstable angina, left ventricular dysfunction, recent acute myocardial infarction, and congestive heart failure (REFs).

In ACS cases, one of the important factors is the sequence of grafting, which is individualized based upon the patient’s coronary artery disease pattern. In patients with severe left main coronary artery stenosis, the LAD is grafted first. Collateralized vessels should be grafted before collateralizing vessels. Generally, culprit coronary arteries should be reconstructed after other easily accessible targets are grafted. Experienced OPCAB surgeons should promptly decide on the sequence of grafting in ACS cases according to the accessibility of the target coronary artery.

At present, the value of the OPCAB technique in ACS cases is unclear, and there is currently insufficient evidence to determine the effect of the OPCAB technique on ACS patients. However, we think that OPCAB will play a special role in ACS treatment in the future.

Diabetic patients

Diabetes mellitus is a strong risk factor that worsens the clinical outcome of coronary artery bypass graft (CABG) surgery in the long term, as well as in the short term [59–62]. The reasons are that diabetic patients have more extensive, diffuse, multiple and rapidly progressive coronary vessel disease, and more extra-cardiac comorbidities. It is well known that diabetic patients experience a survival benefit with CABG as opposed to percutaneous coronary intervention (PCI) [63]. Off-pump CABG (OPCAB) has now gained acceptance as an advantageous technique that provides particular benefits in diabetic patients compared to on-pump conventional CABG [64–66]. The population of diabetic patients undergoing CABG has been steadily increasing worldwide, especially in Japan, and now accounts for up to 40 % of all CABG patients. To bring the clinical outcome for diabetes patients up to the same level as for non-diabetic patients, surgeons need to consider the best strategy for CABG.

Cardiopulmonary bypass has been reported to be associated with more complications and to induce greater oxidative stress in diabetic than in non-diabetic patients (21). Some authors have found reduced surgical morbidity in diabetic patients undergoing OPCAB compared with conventional CABG [64–66]. For example, Magee et al. [64] reported the influence of diabetes on the mortality and morbidity in comparison with 346 off-pump and 2,545 on-pump cases. They found that OPCAB in diabetic patients was associated with a significant reduction in postoperative morbidity, although the survival advantage was controversial. In a comparison of 540 OPCAB cases and 475 on-pump cases using propensity-adjusted regression analysis, Emmert et al. [65] also reported that OPCAB offers lower mortality and superior postoperative outcomes in diabetic patients. Renner et al., who compared 355 diabetic patients undergoing OPCAB with 502 undergoing on-pump CABG, found that OPCAB was associated with significantly lower rates of 30-day mortality and postoperative complications and significantly decreased 6-month and 1-year mortality rates [66]. Thus, OPCAB has several advantages for diabetic patients compared to conventional on-pump CABG.

The current trend in off-pump revascularization strategies is toward in situ all-arterial grafting. Hwang et al. [67] found that diabetes did not affect the long-term survival or clinical events in patients with multi-vessel coronary disease who underwent total arterial off-pump revascularization. We think that the advantageous effects of total arterial OPCAB can minimize the adverse contributions of diabetes mellitus, resulting in outcomes equivalent to those in non-diabetic patients. Off-pump total arterial revascularization with the in situ skeletonized bilateral ITA and the GEA using the aorta no-touch technique is an optimal strategy that can improve the clinical outcome for diabetic patients.

Summary

Almost all of the previous prospective randomized trials and meta-analyses have failed to show a benefit of OPCAB in either the short- or long-term clinical outcome. General observations from relevant reports agree that, compared to on-pump conventional CABG, OPCAB results in fewer distal anastomoses, a higher rate of incomplete revascularization, inferior graft patency, equal rates of early mortality and morbidity, and worse clinical outcome, without any noted benefit in the long-term. OPCAB is a more technically demanding procedure, requiring time for surgeons to master, and is obviously stressful for both the surgeon and anesthesiologist. However, all of the negative points associated with OPCAB can be explained by poor surgical technique. OPCAB will therefore presumably, and unfortunately, never become a standard operation that the average surgeon can routinely perform. It certainly takes longer to master the off-pump technique than on-pump CABG. However, skilled off-pump surgeons have noted that OPCAB has several advantages for high-risk patients and can provide an equal or superior outcome to on-pump CABG in both the short- and long terms. It will therefore not disappear as a surgical alternative. Pioneering OPCAB surgeons, including ourselves, will continue with strenuous efforts to optimize our OPCAB skills for treating patients with ischemic heart disease.