Should off-pump CABG be abandoned?

YES

I believe on-pump coronary artery bypass graft (CABG) surgery should be our primary operation, with off-pump CABG reserved only for certain situations.

In talking about these two surgical options, there are basically three main issues of interest: procedural outcomes, the quality of revascularization, and long-term effectiveness.

Procedural outcomes: In the 1990s, there was great excitement for the potential for off-pump CABG. We thought survival would be better, stroke less common, neurocognitive outcomes improved, etc., but after several randomized studies and a number of well-controlled observational studies, we found that, in terms of the important procedural outcomes – death, MI, stroke, and acute renal failure – there were actually no differences between off- and on-pump CABG. We saw this in ROOBY, with low-risk patients and less experienced surgeons; in the Canadian CORONARY trial, with experienced off-pump surgeons; and in the European GOPCABE trial, where the surgeons were very experienced and the patients were high risk.

There are some benefits with off-pump CABG in terms of what I call the "reversible" complications of surgery: fewer transfusions, less postoperative atrial fibrillation, fewer respiratory infections; but I believe these things have to be balanced against some of the negative effects of off-pump surgery in terms of its quality of revascularization and long-term effectiveness.

Probably one of the biggest disappointments we’ve had with off-pump CABG is that, in the trials, we didn’t see any major difference in neurocognitive dysfunction. What we learned is that cardiopulmonary bypass really is not that bad in terms of neurocognitive dysfunction after surgery.

One issue that tends to be forgotten is the risk associated with conversion. In a hemodynamically unstable patient, converting from off-pump to on-pump is associated with higher risks of death, bleeding, renal failure, stroke, respiratory failure, and GI complications. In CORONARY, with experienced surgeons, the conversion rate was almost 7.9%, and in GOPCABE, it was 9.7%. Even with experienced surgeons, conversions do occur, and they have negative consequences.

Quality of revascularization: We know that long-term survival is related to completeness of revascularization. The more ischemic myocardium that is left at risk after CABG, the more likely the patients is to have an MI or die, and this has been very clearly demonstrated in the surgical literature as well as the percutaneous coronary intervention literature. It has been shown multiple times that fewer grafts are done in off-pump patients, compared with on-pump patients. Indeed, in a recent Cochrane meta-analysis of more than 7,000 patients in 57 trials, there were significantly more grafts done in on-pump than off-pump cases.

Also in terms of graft patency, again, multiple studies have shown lower graft patency in off-pump patients, either directly or by means of showing a higher reintervention rate after off-pump as compared with on-pump surgery. In both CORONARY and GOPCABE, the repeat revascularization rate at 30 days was higher for off-pump CABG.

Long-term effectiveness: The long-term results from ROOBY are very sobering: significantly higher 1-year cardiac mortality in the off-pump arm and higher 1-year composite adverse events. Again, in the 2012 Cochrane meta-analysis, with more than 10,000 patients in 75 trials, mortality was significantly higher with off-pump patients, compared with on-pump, with a hazard ratio of 1.24 (95% CI, 1.01-1.53). When the authors of the Cochrane meta-analysis removed what they called the "studies with bias," this signal was even stronger. Also in observational studies, such as one published by Racz et al. in 2004 (J. Am. Coll. Cardiol. 2004; 43: 557-64), the best survival was found to be in patients who had on-pump CABG and the worst in those who had undergone off-pump CABG.

In summary, procedural outcomes with off-pump and on-pump CABG are similar, albeit with lower reversible complications in the off-pump patients, but a greater conversion risk, along with its associated complications. The quality of the revascularization is worse in off-pump patients, the completeness of revascularization is less and re-intervention rates are higher, and the bottom line, of course, it that there is higher long-term mortality after off-pump CABG.

"Should off-pump CABG be abandoned?" As our default procedure, yes. The vast majority of our patients are best served with on-pump revascularization.

Dr. Joseph Sabik is the chairman of the department of thoracic and cardiovascular surgery and the Sheik Hamdam Bin Rashid Al Maktoum Distinguished Chair at the Cleveland Clinic. He disclosed that he performs both off-pump and on-pump CABG.

References

Anesthesiology 2005;102:188-203

N. Engl. J. Med. 2009;361:1827-37

N. Engl. J. Med. 2012;366:1489-97

N. Engl. J. Med. 2013;368:1179-88

Cochrane Database Syst. Rev. 2012; Mar 14. 3: CD007224 [doi:10.1002/14651858.CD007224.pub2])

J. Am. Coll. Cardiol. 2004;43:557-64

NO

Should off pump bypass be abandoned? Absolutely not, but let’s do it well.

The rationale for why off-pump CABG should be the preferred strategy is simple: cardiopulmonary bypass (CPB) entails extracorporeal circulation, aortic cannulation and clamping, global MI, hypothermia, and hemodilution, among other potentially deleterious phenomena. There are morbidities that can be attributed to these entities, and off-pump bypass avoids those effects by mechanically stabilizing each coronary artery target individually, while the rest of the heart beats and supports normal physiologic circulation. There is an important caveat to this, however, and that is if – and perhaps only if – a complete revascularization with precise anastomoses can be accomplished off pump, then the patient will in fact benefit.

At Emory University in Atlanta, we did a prospective, randomized trial on my own patients, which we called SMART (Surgical Management of Arterial Revascularization) trial. What we found was that in 200 unselected, consecutive patients undergoing either off- or on-pump CABG, we had lower myocardial enzyme release, fewer transfusions, more rapid extubation, and a shorter length of stay in hospital with off-pump CABG.

Completeness of revascularization is a very important issue. In SMART, the number of grafts per patient was exactly the same: 3.39 per patient with off-pump CABG and 3.4 with on-pump CABG. We coined the phrase "Index of Completeness of Revascularization," which we defined as the number of grafts we planned to do in examining the arteriogram prior to randomization and surgery divided by the number of grafts we actually did. We found no difference here, meaning we were able to do the operation we planned to do (1.00 vs. 1.01; P = not significant). Moreover, for the lateral wall, which is technically more difficult to reach in a beating heart, the number was similar in the off- and on-pump groups (0.97 vs. 0.98; P = not significant). We also used a similar percentage of arterial grafts in both groups.

CPB was an independent predictor of the need for transfusion by multivariate analysis with an odds ratio of 2.42 (P = .0073) and was associated with a longer length of stay by 1 day (5.1 days for off-pump and 6.1 for on-pump; P = .005 Wilcoxon).

Creatine phosphokinase of muscle band and troponin I release was about half as much in the off-pump group as in the on-pump group (P less than .001 Wilcoxon), and the rates of death, stroke, MI, angina, and reintervention were similar at both 30 days and 1 year, as was graft patency and quality of life. Off-pump CABG costs at 1 year were $1,955 less than on-pump, but this difference was not statistically significant (P = .08).

At 8 years, survival in SMART was still similar between groups (P = .33), as was graft patency in the small number of patients who had CT angiograms. PET scan results similarly showed no significant difference in ischemia between these two groups (P = .62). One patient in each group has had a percutaneous reintervention, and none have had a repeat coronary bypass in 10 years.

To see if these results could be replicated nationally, we turned to the STS database and looked at North American centers that performed more than 100 on-pump CABGs and more than 100 off-pump surgeries. This gave us 42,477 patients (16,245 off pump and 26,232 on pump) at 63 North American centers. We included the 2.2% of off-pump cases that were converted to on-pump cases in the off-pump group.

After risk adjustment for 32 variables, for the outcomes of death, stroke, MI, and major adverse cardiac events, off-pump bypass outperformed on-pump bypass in this huge cohort of patients from around the country. Looking at less-significant outcomes – renal failure, dialysis, sternal infection, reoperation, atrial fibrillation, prolonged ventilation, and length of stay greater than 14 days – all of them favored off-pump bypass.

We then looked at the Emory dataset (14,766 consecutive patients, 48% of whom had off-pump CABG and 52% on-pump) to see which patients benefitted more. For patients in the two lower quartiles of predicted risk, there was no difference in operative mortality. In the higher two risk quartiles, there was a mortality benefit with off-pump CABG, with a risk reduction for operative mortality of about 55% in the highest risk patients (P less than 0.001).

Logistic regression confirmed that there was an interaction between surgery type and predicted risk, and we now know that low-risk patients do not have the survival benefit of avoiding CPB. They do fine with on-pump CABG, but higher risk patients have a benefit from avoiding CPB and the higher the predicted risk, the greater the benefit to the patient.

We went back to the STS database and we looked at whether this applies only at some centers or all centers, some surgeons or all surgeons. We looked at almost a million cases, 210,469 of which were at sites that had a large off-pump CABG volume. With the usual adjustments, off-pump CABG was associated with significant reduction of risk of death, stroke, renal failure, any morbidity or mortality or prolonged length of stay, compared with on-pump bypass. This benefit was even more pronounced after adjustment for surgeon effect. Once again, the greater reduction was enjoyed by those patients who had the highest preoperative risk. In all predicted risk quartiles, off pump bypass reduced risk of death and stroke and that magnitude increased with increasing predicted risk of mortality. This was seen in large-volume centers and low-volume centers.

Similar results were seen in a recent multicenter, randomized, prospective trial by Lemma et al. that assigned 411 patients to either off- or on-pump coronary bypass (J. Thorac. Cardiovasc. Surg. 2012;143:625-31). There was reduced early mortality and morbidity among higher-risk patients. Interestingly, in this study, they used an experience-based randomization scheme, in which they had surgeons within each center who like to do off-pump CABG and those who like to do on-pump CABG, and each surgeon had hundreds of cases under his belt. For the primary endpoint, a composite of death, MI, stroke or TIA, renal failure, acute respiratory distress syndrome, or reoperation for bleeding, the rates were 5.8% of off-pump and 13.3% for on-pump patients (odds ratio, 2.5; P = .01).

I think the conclusions are clear, but not everyone has reached these same conclusions. In ROOBY, the results were different. Although the study was well conducted, it enrolled low-risk patients, in whom avoidance of CPB was unlikely to improve the expected excellent outcomes. And the operations were performed by residents, with supervising attendings who themselves only had to do 20 total career off-pump cases to be eligible. I think this lack of experience is well demonstrated by the 12.5% conversion rate from off-pump cases to on-pump in that trial. ROOBY enrolled the wrong patients and used the wrong surgeons.

In the CORONARY trial, conducted by Dr. Lamy in Canada but enrolling patients from 19 countries outside of Canada, there was no difference in the primary endpoint of death, stroke, MI, and renal failure at 30 days, but there was a decrease in transfusion, reoperation for bleeding, acute kidney injury, or respiratory complications. There wasn\'t a difference in stroke, interestingly, but the surgeons in this trial appropriately converted a hundred patients from on pump to off pump to avoid manipulating a calcified aorta. This was a good, well-conducted trial.

However, the primary outcome in CORONARY differed when assessed according to EuroSCORE. When the EuroSCORE was low, on pump outperformed off pump. When the EuroSCORE was high, off pump outperformed on pump.

These two trials offer important perspective: the ROOBY trial, enrolling low-risk patients, was actually in favor of on-pump CABG. The CORONARY trial, enrolling higher-risk patients, had a slight benefit in favor of off-pump CABG, and this was particularly evident in the Canadian cohort of 830 randomized patients, in which the primary outcome was, in fact, statistically significantly better in the off-pump group at 9.2% vs. 13.7%.

At the end of the day, I think it matters in whom you do off-pump CABG and how well you do it. It may not be for every patient or for every surgeon; off-pump CABG requires a focused and sustained effort to master a new set of physical and psychological skills to accomplish precise and complete revascularization. When we can do this, I think, we offer better outcomes for our patients.

Dr. John Puskas is the chairman of the department of cardiac surgery at Mount Sinai Beth Israel in New York. Dr. Puskas disclosed that he also does both on- and off-pump CABG. He received royalties from coronary surgical instruments marketed by Scanlan, as well.

References

JAMA 2004;291:1841-9

Ann. Thorac. Surg. 2009;88:1142-7

N. Engl. J. Med. 2009;361:1827-37

N. Engl. J. Med. 2010;362:851

N. Engl. J. Med. 2012;366:1489-97

N. Engl. J. Med. 2013;368:1179-88

Ann. Thorac. Surg. 2011;91:1836-42

J. Thorac. Cardiovasc. Surg. 2012;143:625-31

YES

I believe on-pump coronary artery bypass graft (CABG) surgery should be our primary operation, with off-pump CABG reserved only for certain situations.

In talking about these two surgical options, there are basically three main issues of interest: procedural outcomes, the quality of revascularization, and long-term effectiveness.

Procedural outcomes: In the 1990s, there was great excitement for the potential for off-pump CABG. We thought survival would be better, stroke less common, neurocognitive outcomes improved, etc., but after several randomized studies and a number of well-controlled observational studies, we found that, in terms of the important procedural outcomes – death, MI, stroke, and acute renal failure – there were actually no differences between off- and on-pump CABG. We saw this in ROOBY, with low-risk patients and less experienced surgeons; in the Canadian CORONARY trial, with experienced off-pump surgeons; and in the European GOPCABE trial, where the surgeons were very experienced and the patients were high risk.

There are some benefits with off-pump CABG in terms of what I call the "reversible" complications of surgery: fewer transfusions, less postoperative atrial fibrillation, fewer respiratory infections; but I believe these things have to be balanced against some of the negative effects of off-pump surgery in terms of its quality of revascularization and long-term effectiveness.

Probably one of the biggest disappointments we’ve had with off-pump CABG is that, in the trials, we didn’t see any major difference in neurocognitive dysfunction. What we learned is that cardiopulmonary bypass really is not that bad in terms of neurocognitive dysfunction after surgery.

One issue that tends to be forgotten is the risk associated with conversion. In a hemodynamically unstable patient, converting from off-pump to on-pump is associated with higher risks of death, bleeding, renal failure, stroke, respiratory failure, and GI complications. In CORONARY, with experienced surgeons, the conversion rate was almost 7.9%, and in GOPCABE, it was 9.7%. Even with experienced surgeons, conversions do occur, and they have negative consequences.

Quality of revascularization: We know that long-term survival is related to completeness of revascularization. The more ischemic myocardium that is left at risk after CABG, the more likely the patients is to have an MI or die, and this has been very clearly demonstrated in the surgical literature as well as the percutaneous coronary intervention literature. It has been shown multiple times that fewer grafts are done in off-pump patients, compared with on-pump patients. Indeed, in a recent Cochrane meta-analysis of more than 7,000 patients in 57 trials, there were significantly more grafts done in on-pump than off-pump cases.

Also in terms of graft patency, again, multiple studies have shown lower graft patency in off-pump patients, either directly or by means of showing a higher reintervention rate after off-pump as compared with on-pump surgery. In both CORONARY and GOPCABE, the repeat revascularization rate at 30 days was higher for off-pump CABG.

Long-term effectiveness: The long-term results from ROOBY are very sobering: significantly higher 1-year cardiac mortality in the off-pump arm and higher 1-year composite adverse events. Again, in the 2012 Cochrane meta-analysis, with more than 10,000 patients in 75 trials, mortality was significantly higher with off-pump patients, compared with on-pump, with a hazard ratio of 1.24 (95% CI, 1.01-1.53). When the authors of the Cochrane meta-analysis removed what they called the "studies with bias," this signal was even stronger. Also in observational studies, such as one published by Racz et al. in 2004 (J. Am. Coll. Cardiol. 2004; 43: 557-64), the best survival was found to be in patients who had on-pump CABG and the worst in those who had undergone off-pump CABG.

In summary, procedural outcomes with off-pump and on-pump CABG are similar, albeit with lower reversible complications in the off-pump patients, but a greater conversion risk, along with its associated complications. The quality of the revascularization is worse in off-pump patients, the completeness of revascularization is less and re-intervention rates are higher, and the bottom line, of course, it that there is higher long-term mortality after off-pump CABG.

"Should off-pump CABG be abandoned?" As our default procedure, yes. The vast majority of our patients are best served with on-pump revascularization.

Dr. Joseph Sabik is the chairman of the department of thoracic and cardiovascular surgery and the Sheik Hamdam Bin Rashid Al Maktoum Distinguished Chair at the Cleveland Clinic. He disclosed that he performs both off-pump and on-pump CABG.

References

Anesthesiology 2005;102:188-203

N. Engl. J. Med. 2009;361:1827-37

N. Engl. J. Med. 2012;366:1489-97

N. Engl. J. Med. 2013;368:1179-88

Cochrane Database Syst. Rev. 2012; Mar 14. 3: CD007224 [doi:10.1002/14651858.CD007224.pub2])

J. Am. Coll. Cardiol. 2004;43:557-64

NO

Should off pump bypass be abandoned? Absolutely not, but let’s do it well.

The rationale for why off-pump CABG should be the preferred strategy is simple: cardiopulmonary bypass (CPB) entails extracorporeal circulation, aortic cannulation and clamping, global MI, hypothermia, and hemodilution, among other potentially deleterious phenomena. There are morbidities that can be attributed to these entities, and off-pump bypass avoids those effects by mechanically stabilizing each coronary artery target individually, while the rest of the heart beats and supports normal physiologic circulation. There is an important caveat to this, however, and that is if – and perhaps only if – a complete revascularization with precise anastomoses can be accomplished off pump, then the patient will in fact benefit.

At Emory University in Atlanta, we did a prospective, randomized trial on my own patients, which we called SMART (Surgical Management of Arterial Revascularization) trial. What we found was that in 200 unselected, consecutive patients undergoing either off- or on-pump CABG, we had lower myocardial enzyme release, fewer transfusions, more rapid extubation, and a shorter length of stay in hospital with off-pump CABG.

Completeness of revascularization is a very important issue. In SMART, the number of grafts per patient was exactly the same: 3.39 per patient with off-pump CABG and 3.4 with on-pump CABG. We coined the phrase "Index of Completeness of Revascularization," which we defined as the number of grafts we planned to do in examining the arteriogram prior to randomization and surgery divided by the number of grafts we actually did. We found no difference here, meaning we were able to do the operation we planned to do (1.00 vs. 1.01; P = not significant). Moreover, for the lateral wall, which is technically more difficult to reach in a beating heart, the number was similar in the off- and on-pump groups (0.97 vs. 0.98; P = not significant). We also used a similar percentage of arterial grafts in both groups.

CPB was an independent predictor of the need for transfusion by multivariate analysis with an odds ratio of 2.42 (P = .0073) and was associated with a longer length of stay by 1 day (5.1 days for off-pump and 6.1 for on-pump; P = .005 Wilcoxon).

Creatine phosphokinase of muscle band and troponin I release was about half as much in the off-pump group as in the on-pump group (P less than .001 Wilcoxon), and the rates of death, stroke, MI, angina, and reintervention were similar at both 30 days and 1 year, as was graft patency and quality of life. Off-pump CABG costs at 1 year were $1,955 less than on-pump, but this difference was not statistically significant (P = .08).

At 8 years, survival in SMART was still similar between groups (P = .33), as was graft patency in the small number of patients who had CT angiograms. PET scan results similarly showed no significant difference in ischemia between these two groups (P = .62). One patient in each group has had a percutaneous reintervention, and none have had a repeat coronary bypass in 10 years.

To see if these results could be replicated nationally, we turned to the STS database and looked at North American centers that performed more than 100 on-pump CABGs and more than 100 off-pump surgeries. This gave us 42,477 patients (16,245 off pump and 26,232 on pump) at 63 North American centers. We included the 2.2% of off-pump cases that were converted to on-pump cases in the off-pump group.

After risk adjustment for 32 variables, for the outcomes of death, stroke, MI, and major adverse cardiac events, off-pump bypass outperformed on-pump bypass in this huge cohort of patients from around the country. Looking at less-significant outcomes – renal failure, dialysis, sternal infection, reoperation, atrial fibrillation, prolonged ventilation, and length of stay greater than 14 days – all of them favored off-pump bypass.

We then looked at the Emory dataset (14,766 consecutive patients, 48% of whom had off-pump CABG and 52% on-pump) to see which patients benefitted more. For patients in the two lower quartiles of predicted risk, there was no difference in operative mortality. In the higher two risk quartiles, there was a mortality benefit with off-pump CABG, with a risk reduction for operative mortality of about 55% in the highest risk patients (P less than 0.001).

Logistic regression confirmed that there was an interaction between surgery type and predicted risk, and we now know that low-risk patients do not have the survival benefit of avoiding CPB. They do fine with on-pump CABG, but higher risk patients have a benefit from avoiding CPB and the higher the predicted risk, the greater the benefit to the patient.

We went back to the STS database and we looked at whether this applies only at some centers or all centers, some surgeons or all surgeons. We looked at almost a million cases, 210,469 of which were at sites that had a large off-pump CABG volume. With the usual adjustments, off-pump CABG was associated with significant reduction of risk of death, stroke, renal failure, any morbidity or mortality or prolonged length of stay, compared with on-pump bypass. This benefit was even more pronounced after adjustment for surgeon effect. Once again, the greater reduction was enjoyed by those patients who had the highest preoperative risk. In all predicted risk quartiles, off pump bypass reduced risk of death and stroke and that magnitude increased with increasing predicted risk of mortality. This was seen in large-volume centers and low-volume centers.

Similar results were seen in a recent multicenter, randomized, prospective trial by Lemma et al. that assigned 411 patients to either off- or on-pump coronary bypass (J. Thorac. Cardiovasc. Surg. 2012;143:625-31). There was reduced early mortality and morbidity among higher-risk patients. Interestingly, in this study, they used an experience-based randomization scheme, in which they had surgeons within each center who like to do off-pump CABG and those who like to do on-pump CABG, and each surgeon had hundreds of cases under his belt. For the primary endpoint, a composite of death, MI, stroke or TIA, renal failure, acute respiratory distress syndrome, or reoperation for bleeding, the rates were 5.8% of off-pump and 13.3% for on-pump patients (odds ratio, 2.5; P = .01).

I think the conclusions are clear, but not everyone has reached these same conclusions. In ROOBY, the results were different. Although the study was well conducted, it enrolled low-risk patients, in whom avoidance of CPB was unlikely to improve the expected excellent outcomes. And the operations were performed by residents, with supervising attendings who themselves only had to do 20 total career off-pump cases to be eligible. I think this lack of experience is well demonstrated by the 12.5% conversion rate from off-pump cases to on-pump in that trial. ROOBY enrolled the wrong patients and used the wrong surgeons.

In the CORONARY trial, conducted by Dr. Lamy in Canada but enrolling patients from 19 countries outside of Canada, there was no difference in the primary endpoint of death, stroke, MI, and renal failure at 30 days, but there was a decrease in transfusion, reoperation for bleeding, acute kidney injury, or respiratory complications. There wasn\'t a difference in stroke, interestingly, but the surgeons in this trial appropriately converted a hundred patients from on pump to off pump to avoid manipulating a calcified aorta. This was a good, well-conducted trial.

However, the primary outcome in CORONARY differed when assessed according to EuroSCORE. When the EuroSCORE was low, on pump outperformed off pump. When the EuroSCORE was high, off pump outperformed on pump.

These two trials offer important perspective: the ROOBY trial, enrolling low-risk patients, was actually in favor of on-pump CABG. The CORONARY trial, enrolling higher-risk patients, had a slight benefit in favor of off-pump CABG, and this was particularly evident in the Canadian cohort of 830 randomized patients, in which the primary outcome was, in fact, statistically significantly better in the off-pump group at 9.2% vs. 13.7%.

At the end of the day, I think it matters in whom you do off-pump CABG and how well you do it. It may not be for every patient or for every surgeon; off-pump CABG requires a focused and sustained effort to master a new set of physical and psychological skills to accomplish precise and complete revascularization. When we can do this, I think, we offer better outcomes for our patients.

Dr. John Puskas is the chairman of the department of cardiac surgery at Mount Sinai Beth Israel in New York. Dr. Puskas disclosed that he also does both on- and off-pump CABG. He received royalties from coronary surgical instruments marketed by Scanlan, as well.

References

JAMA 2004;291:1841-9

Ann. Thorac. Surg. 2009;88:1142-7

N. Engl. J. Med. 2009;361:1827-37

N. Engl. J. Med. 2010;362:851

N. Engl. J. Med. 2012;366:1489-97

N. Engl. J. Med. 2013;368:1179-88

Ann. Thorac. Surg. 2011;91:1836-42

J. Thorac. Cardiovasc. Surg. 2012;143:625-31

YES

I believe on-pump coronary artery bypass graft (CABG) surgery should be our primary operation, with off-pump CABG reserved only for certain situations.

In talking about these two surgical options, there are basically three main issues of interest: procedural outcomes, the quality of revascularization, and long-term effectiveness.

Procedural outcomes: In the 1990s, there was great excitement for the potential for off-pump CABG. We thought survival would be better, stroke less common, neurocognitive outcomes improved, etc., but after several randomized studies and a number of well-controlled observational studies, we found that, in terms of the important procedural outcomes – death, MI, stroke, and acute renal failure – there were actually no differences between off- and on-pump CABG. We saw this in ROOBY, with low-risk patients and less experienced surgeons; in the Canadian CORONARY trial, with experienced off-pump surgeons; and in the European GOPCABE trial, where the surgeons were very experienced and the patients were high risk.

There are some benefits with off-pump CABG in terms of what I call the "reversible" complications of surgery: fewer transfusions, less postoperative atrial fibrillation, fewer respiratory infections; but I believe these things have to be balanced against some of the negative effects of off-pump surgery in terms of its quality of revascularization and long-term effectiveness.

Probably one of the biggest disappointments we’ve had with off-pump CABG is that, in the trials, we didn’t see any major difference in neurocognitive dysfunction. What we learned is that cardiopulmonary bypass really is not that bad in terms of neurocognitive dysfunction after surgery.

One issue that tends to be forgotten is the risk associated with conversion. In a hemodynamically unstable patient, converting from off-pump to on-pump is associated with higher risks of death, bleeding, renal failure, stroke, respiratory failure, and GI complications. In CORONARY, with experienced surgeons, the conversion rate was almost 7.9%, and in GOPCABE, it was 9.7%. Even with experienced surgeons, conversions do occur, and they have negative consequences.

Quality of revascularization: We know that long-term survival is related to completeness of revascularization. The more ischemic myocardium that is left at risk after CABG, the more likely the patients is to have an MI or die, and this has been very clearly demonstrated in the surgical literature as well as the percutaneous coronary intervention literature. It has been shown multiple times that fewer grafts are done in off-pump patients, compared with on-pump patients. Indeed, in a recent Cochrane meta-analysis of more than 7,000 patients in 57 trials, there were significantly more grafts done in on-pump than off-pump cases.

Also in terms of graft patency, again, multiple studies have shown lower graft patency in off-pump patients, either directly or by means of showing a higher reintervention rate after off-pump as compared with on-pump surgery. In both CORONARY and GOPCABE, the repeat revascularization rate at 30 days was higher for off-pump CABG.

Long-term effectiveness: The long-term results from ROOBY are very sobering: significantly higher 1-year cardiac mortality in the off-pump arm and higher 1-year composite adverse events. Again, in the 2012 Cochrane meta-analysis, with more than 10,000 patients in 75 trials, mortality was significantly higher with off-pump patients, compared with on-pump, with a hazard ratio of 1.24 (95% CI, 1.01-1.53). When the authors of the Cochrane meta-analysis removed what they called the "studies with bias," this signal was even stronger. Also in observational studies, such as one published by Racz et al. in 2004 (J. Am. Coll. Cardiol. 2004; 43: 557-64), the best survival was found to be in patients who had on-pump CABG and the worst in those who had undergone off-pump CABG.

In summary, procedural outcomes with off-pump and on-pump CABG are similar, albeit with lower reversible complications in the off-pump patients, but a greater conversion risk, along with its associated complications. The quality of the revascularization is worse in off-pump patients, the completeness of revascularization is less and re-intervention rates are higher, and the bottom line, of course, it that there is higher long-term mortality after off-pump CABG.

"Should off-pump CABG be abandoned?" As our default procedure, yes. The vast majority of our patients are best served with on-pump revascularization.

Dr. Joseph Sabik is the chairman of the department of thoracic and cardiovascular surgery and the Sheik Hamdam Bin Rashid Al Maktoum Distinguished Chair at the Cleveland Clinic. He disclosed that he performs both off-pump and on-pump CABG.

References

Anesthesiology 2005;102:188-203

N. Engl. J. Med. 2009;361:1827-37

N. Engl. J. Med. 2012;366:1489-97

N. Engl. J. Med. 2013;368:1179-88

Cochrane Database Syst. Rev. 2012; Mar 14. 3: CD007224 [doi:10.1002/14651858.CD007224.pub2])

J. Am. Coll. Cardiol. 2004;43:557-64

NO

Should off pump bypass be abandoned? Absolutely not, but let’s do it well.

The rationale for why off-pump CABG should be the preferred strategy is simple: cardiopulmonary bypass (CPB) entails extracorporeal circulation, aortic cannulation and clamping, global MI, hypothermia, and hemodilution, among other potentially deleterious phenomena. There are morbidities that can be attributed to these entities, and off-pump bypass avoids those effects by mechanically stabilizing each coronary artery target individually, while the rest of the heart beats and supports normal physiologic circulation. There is an important caveat to this, however, and that is if – and perhaps only if – a complete revascularization with precise anastomoses can be accomplished off pump, then the patient will in fact benefit.

At Emory University in Atlanta, we did a prospective, randomized trial on my own patients, which we called SMART (Surgical Management of Arterial Revascularization) trial. What we found was that in 200 unselected, consecutive patients undergoing either off- or on-pump CABG, we had lower myocardial enzyme release, fewer transfusions, more rapid extubation, and a shorter length of stay in hospital with off-pump CABG.

Completeness of revascularization is a very important issue. In SMART, the number of grafts per patient was exactly the same: 3.39 per patient with off-pump CABG and 3.4 with on-pump CABG. We coined the phrase "Index of Completeness of Revascularization," which we defined as the number of grafts we planned to do in examining the arteriogram prior to randomization and surgery divided by the number of grafts we actually did. We found no difference here, meaning we were able to do the operation we planned to do (1.00 vs. 1.01; P = not significant). Moreover, for the lateral wall, which is technically more difficult to reach in a beating heart, the number was similar in the off- and on-pump groups (0.97 vs. 0.98; P = not significant). We also used a similar percentage of arterial grafts in both groups.

CPB was an independent predictor of the need for transfusion by multivariate analysis with an odds ratio of 2.42 (P = .0073) and was associated with a longer length of stay by 1 day (5.1 days for off-pump and 6.1 for on-pump; P = .005 Wilcoxon).

Creatine phosphokinase of muscle band and troponin I release was about half as much in the off-pump group as in the on-pump group (P less than .001 Wilcoxon), and the rates of death, stroke, MI, angina, and reintervention were similar at both 30 days and 1 year, as was graft patency and quality of life. Off-pump CABG costs at 1 year were $1,955 less than on-pump, but this difference was not statistically significant (P = .08).

At 8 years, survival in SMART was still similar between groups (P = .33), as was graft patency in the small number of patients who had CT angiograms. PET scan results similarly showed no significant difference in ischemia between these two groups (P = .62). One patient in each group has had a percutaneous reintervention, and none have had a repeat coronary bypass in 10 years.

To see if these results could be replicated nationally, we turned to the STS database and looked at North American centers that performed more than 100 on-pump CABGs and more than 100 off-pump surgeries. This gave us 42,477 patients (16,245 off pump and 26,232 on pump) at 63 North American centers. We included the 2.2% of off-pump cases that were converted to on-pump cases in the off-pump group.

After risk adjustment for 32 variables, for the outcomes of death, stroke, MI, and major adverse cardiac events, off-pump bypass outperformed on-pump bypass in this huge cohort of patients from around the country. Looking at less-significant outcomes – renal failure, dialysis, sternal infection, reoperation, atrial fibrillation, prolonged ventilation, and length of stay greater than 14 days – all of them favored off-pump bypass.

We then looked at the Emory dataset (14,766 consecutive patients, 48% of whom had off-pump CABG and 52% on-pump) to see which patients benefitted more. For patients in the two lower quartiles of predicted risk, there was no difference in operative mortality. In the higher two risk quartiles, there was a mortality benefit with off-pump CABG, with a risk reduction for operative mortality of about 55% in the highest risk patients (P less than 0.001).

Logistic regression confirmed that there was an interaction between surgery type and predicted risk, and we now know that low-risk patients do not have the survival benefit of avoiding CPB. They do fine with on-pump CABG, but higher risk patients have a benefit from avoiding CPB and the higher the predicted risk, the greater the benefit to the patient.

We went back to the STS database and we looked at whether this applies only at some centers or all centers, some surgeons or all surgeons. We looked at almost a million cases, 210,469 of which were at sites that had a large off-pump CABG volume. With the usual adjustments, off-pump CABG was associated with significant reduction of risk of death, stroke, renal failure, any morbidity or mortality or prolonged length of stay, compared with on-pump bypass. This benefit was even more pronounced after adjustment for surgeon effect. Once again, the greater reduction was enjoyed by those patients who had the highest preoperative risk. In all predicted risk quartiles, off pump bypass reduced risk of death and stroke and that magnitude increased with increasing predicted risk of mortality. This was seen in large-volume centers and low-volume centers.

Similar results were seen in a recent multicenter, randomized, prospective trial by Lemma et al. that assigned 411 patients to either off- or on-pump coronary bypass (J. Thorac. Cardiovasc. Surg. 2012;143:625-31). There was reduced early mortality and morbidity among higher-risk patients. Interestingly, in this study, they used an experience-based randomization scheme, in which they had surgeons within each center who like to do off-pump CABG and those who like to do on-pump CABG, and each surgeon had hundreds of cases under his belt. For the primary endpoint, a composite of death, MI, stroke or TIA, renal failure, acute respiratory distress syndrome, or reoperation for bleeding, the rates were 5.8% of off-pump and 13.3% for on-pump patients (odds ratio, 2.5; P = .01).

I think the conclusions are clear, but not everyone has reached these same conclusions. In ROOBY, the results were different. Although the study was well conducted, it enrolled low-risk patients, in whom avoidance of CPB was unlikely to improve the expected excellent outcomes. And the operations were performed by residents, with supervising attendings who themselves only had to do 20 total career off-pump cases to be eligible. I think this lack of experience is well demonstrated by the 12.5% conversion rate from off-pump cases to on-pump in that trial. ROOBY enrolled the wrong patients and used the wrong surgeons.

In the CORONARY trial, conducted by Dr. Lamy in Canada but enrolling patients from 19 countries outside of Canada, there was no difference in the primary endpoint of death, stroke, MI, and renal failure at 30 days, but there was a decrease in transfusion, reoperation for bleeding, acute kidney injury, or respiratory complications. There wasn\'t a difference in stroke, interestingly, but the surgeons in this trial appropriately converted a hundred patients from on pump to off pump to avoid manipulating a calcified aorta. This was a good, well-conducted trial.

However, the primary outcome in CORONARY differed when assessed according to EuroSCORE. When the EuroSCORE was low, on pump outperformed off pump. When the EuroSCORE was high, off pump outperformed on pump.

These two trials offer important perspective: the ROOBY trial, enrolling low-risk patients, was actually in favor of on-pump CABG. The CORONARY trial, enrolling higher-risk patients, had a slight benefit in favor of off-pump CABG, and this was particularly evident in the Canadian cohort of 830 randomized patients, in which the primary outcome was, in fact, statistically significantly better in the off-pump group at 9.2% vs. 13.7%.

At the end of the day, I think it matters in whom you do off-pump CABG and how well you do it. It may not be for every patient or for every surgeon; off-pump CABG requires a focused and sustained effort to master a new set of physical and psychological skills to accomplish precise and complete revascularization. When we can do this, I think, we offer better outcomes for our patients.

Dr. John Puskas is the chairman of the department of cardiac surgery at Mount Sinai Beth Israel in New York. Dr. Puskas disclosed that he also does both on- and off-pump CABG. He received royalties from coronary surgical instruments marketed by Scanlan, as well.

References

JAMA 2004;291:1841-9

Ann. Thorac. Surg. 2009;88:1142-7

N. Engl. J. Med. 2009;361:1827-37

N. Engl. J. Med. 2010;362:851

N. Engl. J. Med. 2012;366:1489-97

N. Engl. J. Med. 2013;368:1179-88

Ann. Thorac. Surg. 2011;91:1836-42

J. Thorac. Cardiovasc. Surg. 2012;143:625-31