**3. Results**

### **3.1 Effectiveness**

### *Long term survival*

The comparative long-term survival of OPCAB and CCABG is not well evaluated. The longest follow-up is in the OCTOPUS-study (6). Five years postoperatively 130/142 OPCAB-patients and 130/139 CCABG patients were alive (p=ns). Other studies with up to twelve months follow-up also failed to show any difference (7-13). Due to the relatively low risk of mortality associated with either operation, however, the statistical strength to detect any difference is not present in any of these studies.

 The largest randomized study, the ROOBY-trial (14) showed a trend towards higher mortality in the OPCAB group at one year follow-up (4.1% vs. 2.9%, p=0.15) and a significant difference in cardiac deaths only (2.7% vs. 1.3%, p=0.03). On the other hand, another study with a mean 3.8 years follow-up of 300 patients showed a trend in the opposite direction with 5 deaths in the OPCAB-group and 10 in the CCABG-group (15).

### *Graft patency*

Even with the use of contemporary cardiac stabilizers and intracoronary shunts, OPCAB remains more technically challenging than CCABG. Difficulties with positioning the heart may cause the surgeon to graft a less favourable part of the coronary artery. Performing the anastomosis is more difficult and may lead to stenosis at the anastomosis site. Furthermore, the coagulability of blood is increased after OPCAB compared to CCABG (16-20). Hence, a serious concern when introducing the OPCAB technique has been whether the number and quality of the grafts would be equivalent to what could be achieved using CPB.

In the vast majority of randomized, controlled trials, the OPCAB-patients tended to receive a lower number of grafts than the patients operated using CPB. In the largest studies and in a meta-analysis this difference was statistically significant (13-14, 21, 22) with a mean difference of 0.1-0.3 grafts. Several studies compared the number of grafts compared to a preoperative plan. In most of these studies, no difference was found (8-9,12, 14, 23), a few studies showed a difference in favour of CCABG, and one study found a difference in favour of OPCAB (24). In this study, however, the absolute number of grafts was 0.2 lower in the OPCAB-group.

Allmost all of the earlier studies showed a trend towards poorer graft patency in OPCAB patients. A single, smaller study found this difference to be statistically significant (25). Also, the proportion of patent grafts in the largest study, the ROOBY-trial, was 82.6% in the OPCAB group and 87.8% in the CCABG group (p<0.01) (14). This difference, however, did not result in a higher number of myocardial infarctions in the OPCAB group.

In studies performed by few, dedicated OPCAB surgeons, the difference in number of grafts was very small and not statistically significant (12, 24). The study by Khan (25), the SMART study (14), and the Best Bypass Surgery Study (26) differentiated the findings and found a higher proportion of occluded grafts at right and circumflex territories and fewer occlusions in the LAD territory. Lingaas et al only found differences in graft patency between OPCAB and CCABG to be significantly different when comparing vein grafts as opposed to internal mammary artery grafts (10).

### *Recurrent or persistent chest pain*

An important parameter is freedom from chest pain. In the Octopus trial (6, 24), 89.0% experienced freedom from chest pain in the OPCAB group compared to 89.3% in the CCABG-group (p=ns). At five years follow-up, these numbers were down to 82.3% and 87.7%, respectively (p=ns). At one year follow-up, ergometer testing was performed in 81% of the patients. It was found to be negative in 79.8% of CCABG patients and 83.1% of OPCAB patients (p=ns). In the SMART-study, chest pain was present at one-year follow-up in 0% of CCABG and 3% of OPCAB patients, respectively (p=ns) (12). In a separate publication, using a specific questionnaire on chest pain in the 400 patients involved in the BHACAS1 and BHACAS2-studies, no difference was found after a median follow-up of three years (27).

### *Reintervention*

166 Special Topics in Cardiac Surgery

recorded. This may have caused high-risk patients to be moved from one treatment group

From the late 1990ies to 2002 a significant technical development in stabilizing equipment led to a fast rise in the number of OPCAB procedures. From 2002 results from the first randomized studies failed to show a clear benefit and interest has cooled somewhat. A significant number of randomized studies have been conducted comparing very different end-points after OPCAB and CCABG. This chapter aims to review the results of these

Searching MEDLINE and Cochrane library using the terms » *OPCAB* «, » *off-pump* «, » *offpump* « OR »MIDCAB«, limited to English language june 23rd, 2010, provided 4788 abstracts that were read manually to find randomized, controlled trials. Two-hundred and twenty nine papers were retrieved and read before 90 papers, reporting results from 61

The comparative long-term survival of OPCAB and CCABG is not well evaluated. The longest follow-up is in the OCTOPUS-study (6). Five years postoperatively 130/142 OPCAB-patients and 130/139 CCABG patients were alive (p=ns). Other studies with up to twelve months follow-up also failed to show any difference (7-13). Due to the relatively low risk of mortality associated with either operation, however, the statistical strength to detect

 The largest randomized study, the ROOBY-trial (14) showed a trend towards higher mortality in the OPCAB group at one year follow-up (4.1% vs. 2.9%, p=0.15) and a significant difference in cardiac deaths only (2.7% vs. 1.3%, p=0.03). On the other hand, another study with a mean 3.8 years follow-up of 300 patients showed a trend in the opposite direction with 5 deaths in the OPCAB-group and 10 in the CCABG-group (15).

Even with the use of contemporary cardiac stabilizers and intracoronary shunts, OPCAB remains more technically challenging than CCABG. Difficulties with positioning the heart may cause the surgeon to graft a less favourable part of the coronary artery. Performing the anastomosis is more difficult and may lead to stenosis at the anastomosis site. Furthermore, the coagulability of blood is increased after OPCAB compared to CCABG (16-20). Hence, a serious concern when introducing the OPCAB technique has been whether the number and

In the vast majority of randomized, controlled trials, the OPCAB-patients tended to receive a lower number of grafts than the patients operated using CPB. In the largest studies and in a meta-analysis this difference was statistically significant (13-14, 21, 22) with a mean difference of 0.1-0.3 grafts. Several studies compared the number of grafts compared to a preoperative plan. In most of these studies, no difference was found (8-9,12, 14, 23), a few studies showed a difference in favour of CCABG, and one study found a difference in

quality of the grafts would be equivalent to what could be achieved using CPB.

studies to assess the comparative effectiveness and safety of the two techniques.

individual randomized, controlled trials, were identified.

any difference is not present in any of these studies.

to the other (4-5).

**2. Methods** 

**3. Results** 

*Graft patency* 

**3.1 Effectiveness**  *Long term survival* 

> Given the lower number of patent grafts in the OPCAB groups, a greater need for coronary re-intervention might be expected. Only few of the published trials have had long enough follow-up for this question to be evaluated. In the BHACAS-1 study, three percent of both OPCAB and CCABG patients had had a reintervention – either percutaneous or surgical – within a median three years follow-up (7). The longest follow-up, which was published by the Octopus trialists, reported 7.7% of OPCAB-patients and 5% of CCABG patients to have undergone reintervention after five years (6). In the ROOBY-trial, the proportion undergoing reintervention was 4.6% in the OPCAB group and 3.4% in the CCABG group at one year follow-up. Neither individual studies nor metanalyses found thiese differences to be statistically significant (21, 26).

### *Quality of life*

A number of studies compare self-reported, health related quality of life after OPCAB and CCABG. Medical Outcomes Study-Short Form 36 (MOS SF-36) is the most commonly used tool. In this questionnaire eight scales cover physical, mental, and social well-being (28). One study found a significantly higher score among CCABG-patients in one of the eight scales ("Role emotional") in contrast to another study who favoured OPCAB patients in the dimension "Social Relationships", using another questionnaire (29,30). In general, few significant inter-group differences have been found, given the multiple tests being performed.

Current Evidence of On-Pump Versus Off-Pump Coronary Artery By-Pass Surgery 169

risk of experiencing a clinical myocardial infarction during the following year (43). A number

In a study of myocardial biopsies, it was found that the concentration of reduced glutathion recovered more rapidly in CCABG than OPCAB patients (38). For the OPCAB-operations, a proximal snare was used for occlusion of the vessel while performing the anastomosis. This finding suggests that cardioplegia is better tolerated than occlusion. Still, a higher increase in CK-MB was found in the CCABG-group. Together, these findings suggest that the myocardium in the territory of the occluded vessel suffers more from occlusion but a less profound damage to the entire myocardium is caused by ischemia and reperfusion. Which of these two situations pose the largest threat to heart function is not clear. Gadolinium contrast enhanced magnetic resonance perfusion imaging, reflecting permanent damage to the myocardium, failed to detect a difference between treatment groups despite a higher release of Troponine- I in the CCABG-group (39). On this background, it was speculated whether some of the Troponine leak represented protein release from non-structurally bound cytosolic pools, rather than true myocardial necrosis. In another study, micro-dialysis was used to sample myocardial interstitial fluid during and after surgery (44). More abnormal values were found during CCABG than during OPCAB. It was not stated in the paper whether samples were

taken within or outside the area of the temporally occluded vessel during OPCAB.

significant in some of these studies and in meta-analyses (21). *Postoperative inotropic support and low cardiac output syndrome* 

studies reports a significant difference (12), while others do not (13, 46).

In the BHACAS-1 study, Heart rate and rhythm were continuously monitored for 72 postoperative hours. The incidence of postoperative atrial fibrillation was found to be 45% among CCABG patients as compared to 8% among patients operated with OPCAB (p<0.001) (45). A large number of later studies, including the BHACAS-2 study, have confirmed this tendency but with a much smaller difference between groups. The tendency is statistically

Need for inotropic drugs after the operation may reflect either transient or permanent heart failure. Several, larger studies do not report this end-point (13-14), but among the ones that do, there is a trend towards a higher incidence in patients operated using cardiopulmonary by-pass. A meta-analysis of 16 studies including 1655 patients found a need for inotropic support postoperatively in 23.6% of CCABG and 15.1% of OPCAB patients (p=0.04)(21). Other studies report the incidence of "low cardiac output syndrome", defined as need for intra-aortic balloon pump, need for inotropic drugs or pressor drugs. One of the larger

Theoretically, the use of cardiopulmonary by-pass may cause stroke by a number of different mechanisms. These include the manipulation of the ascending aorta for cannulation and clamping, gaseous or particulate emboli formed in the by-pass circuit, and

Also, post-operative atrial fibrillation may cause strokes in spite of adequate antithrombotic treatment. For these reasons, an important argument for favouring OPCAB has been the

In low risk patients, the risk of suffering a peri-operative stroke is between 1 and 1.5%. Hence, none of the individual, randomized, controlled trials have had the statistical strength

of confounding issues may, however, be relevant.

*Atrial fibrillation* 

**3.4 Neurological complications** 

accidental interruption of flow.

intention to reduce the rate of peri-operative strokes.

*Stroke* 

On the other hand, a significant increase in self-reported, health related quality of life is invariably found in both groups comparing preoperative and postoperative status (24, 27, 31).
