Bilateral vs Single Internal Mammary Artery Bypass Grafting – Do we have the answer?

The paper by Iribarne et al (1) describes a retrospective study comparing the long term outcome of propensity matched groups of patients undergoing either bilateral (BIMA) and single (SIMA) internal mammary coronary artery bypass grafting, with emphasis on the need for repeat revascularization. The study originated from the Northern New England Cardiovascular Disease Study Group, which has been a leader in examining outcomes after cardiovascular surgery and quality improvement.

Coronary artery bypass grafting (CABG) has continued to be a principle method to treat advanced coronary disease. With improved percutaneous coronary intervention (PCI) and medical treatment, CABG has been forced to improve as well. The introduction of the single internal mammary artery (SIMA) graft was first demonstrated by Loop et al (2) to significantly prolong survival and improve outcome of CABG. The use of 2 (or bilateral) internal mammary artery grafts was described by Suzuki et al in 1973 (3). It would make sense that if the SIMA is good, two internal mammary grafts would be better. Indeed, Lytle (4) reported from the Cleveland Clinic that patients who received CABG with BIMA grafts did do better than if they receive a SIMA graft. Since this publication, many studies have demonstrated what seems to be common sense, that patients who receive bilateral internal mammary artery grafts (5, 6) or multiple arterial grafts of other types, should do much better in the long run. However, patients generally receive BIMA CABG if they are young, in otherwise good health and are expected to live a long time. Nearly all studies comparing BIMA with SIMA have been unmatched or propensity matched studies, with one exception (7). While propensity matched, retrospective studies do provide important information when prospective, randomized trials are not available, they remain limited. The study at hand is no exception.

The study by Iribarne et al examined patients from a cohort of 47,984 individuals undergoing CABG in Northern New England between 1992 and 2014. From these, 1,482 BIMA patients and the same number of SIMA patients were compared after a mean of 12 years. The authors reported a greater freedom from repeat revascularization among BIMA patients compared to SIMA (HR: 0.78, p=0.009). Among the matched cohort, 19% of SIMA patients underwent repeat revascularization whereas 15 % among BIMA patients (p=0.004) required re-intervention. This may explain the better survival at 12 years in the BIMA group (HR 0.70, p<0.001). While the differential rate of revascularization is significant, this represents only a 4 % overall difference between groups. As expected, the vast majority (94%) of repeat revascularization procedures were PCI in both groups.

Unfortunately, selection bias may have played a role in the different outcomes between BIMA and SIMA patients. The unadjusted in hospital mortality rate was more than double in the SIMA group vs the BIMA group, as was the rate of stroke. The post operative length of stay was also significantly higher in the SIMA group. This suggests that greater consideration for BIMA was given to lower risk patients. After propensity matching there was a 50% greater mortality in the SIMA group compared to the BIMA group of patients. Furthermore, and perhaps more importantly, after propensity matching there were a median of 3 bypass grafts performed in the SIMA group vs 4 grafts in the BIMA group. Thus, it is difficult to separate out the role of number of mammary grafts performed from the health and compliance of the patient or the number of total grafts performed. None the less, in this study, patients having undergone SIMA CABG did have a considerably higher rate of major adverse cardiovascular events compared to patients in the BIMA cohort, suggesting a mechanism whereby BIMA CABG may be superior.

Propensity matching does not take into account factors that are not analyzed. Whereas the matching in this study was generally good, it did not take into consideration factors than may have a significant impact on patient outcome after CABG. These factors include frailty, quality of bypass targets, quality of vein and arterial conduits, economic status, compliance with medications or secondary prevention after CABG. These may have been markedly different between the 2 cohorts. These issues can only be addressed in a prospective, randomized trial. The only such trial to date has been the Arterial Revascularization Trial (ART), recently presented at the 2016 American Heart Association Scientific Sessions and published in the New England Journal of Medicine (7). In this study, 3102 CABG patients were randomly assigned to undergo SIMA or BIMA coronary grafting in 28 cardiac surgical centers in seven countries. The primary outcome was death from any cause at 10 years. Of the patients, 1554 were assigned to undergo SIMA and 1548 to undergo BIMA grafting. At 5 years follow-up, the rate of death was 8.7% in the BIMA group and 8.4% in the SIMA group (p=0.77), and the rate of the composite of death from any cause, myocardial infarction, or stroke was 12.2% and 12.7%, respectively (p=0.69). What was remarkable in the ART trial, was the rate of sternal wound complication was 3.5% in the BIMA group versus 1.9% in the SIMA group (p=0.005), and the rate of sternal reconstruction was 1.9% versus 0.6% (p=0.002). There was no difference in the rate of revascularization, being just over 6% in each group. Like in the present study by Iribarne et al, there was no difference in outcome at 5 years. It is very possible that at 10 years follow-up, the outcome curves will diverge in the ART trial. Thus, there is no inconsistency between the ART trial and the study by Iribarne. The 10 year results of the ART trial will be presented in the near future and are anxiously awaited.

In summary, the study by Iribarne et al published in this issue of Circulation is a well performed, propensity matched, retrospective analysis of BIMA vs SIMA an average of 12 years following surgery. However, it also demonstrated the limitations of such a study. The findings are similar to other such studies comparing multiple arterial grafting with SIMA CABG. The 10 year results of the ART trial will give us more definitive information on whether BIMA is truly better than SIMA bypass grafting. As with other controversial topics such as off pump vs on pump CABG, the role of valve repair vs replacement for ischemic mitral valve insufficiency, and the utility of cell therapy for heart disease, the opinions remain divided. The real improvement observed with BIMA or other methods of multi-arterial grafting will only be decided after multiple prospective, randomized trials are completed. Even then, opinions may remain divided.