Article Text


Original article
Bilateral internal mammary artery bypass grafting: long-term clinical benefits in a series of 1000 patients
  1. Batric Popovic1,
  2. Damien Voillot1,
  3. Pablo Maureira2,
  4. Fabrice Vanhuyse2,
  5. Nelly Agrinier3,
  6. Etienne Aliot1,
  7. Thierry Folliguet2,
  8. Jean Pierre Villemot2
  1. 1Département de Cardiologie, CHU Nancy, Nancy, France
  2. 2Service de Chirurgie des Maladies Cardiovasculaires et Transplantations, CHU Nancy, Nancy, France
  3. 3Epidémiologie et Evaluation Cliniques, CHU Nancy, Nancy, France
  1. Correspondence to Dr Batric Popovic, Département Cardiologie, CHU Nancy, rue de Morvan, Vandoeuvre les Nancy 54500, France; batricpopovic{at}


Objective Bilateral internal mammary arteries (BIMA) remain widely underused in coronary artery bypass grafting (CABG). In this study, we aim to investigate the early and long-term outcomes of BIMA grafts in isolated CABGs.

Design Single-centre retrospective observational study.

Setting University Hospital, Nancy.

Patients 1000 consecutive patients undergoing elective, isolated, primary, multiple CABGs using BIMA grafts and supplemental venous grafts for multi-vessel coronary disease.

Main outcome measures In-hospital mortality and major morbidity, and long-term all-cause mortality.

Results Mean age of the overall population was 60±15 years. A left ventricular ejection fraction (LVEF) ≤45% was found in 28% of the patients and 27.1% of the patients were diabetics. Comorbidities were represented by chronic renal failure, chronic obstructive pulmonary disease and peripheral artery disease in 11, 11.7 and 27.3% of the cases, respectively. The in-hospital mortality rate was 2.8%. Early postoperative morbidity included myocardial infarction (2.2%), stroke (0.9%), mesenteric ischaemia (0.7%) and mediastinitis (2.2%).

The Kaplan–Meier 8-year survival rates for patients less than 65 and between 65 and 74 years of age were 88% and 66%, respectively (p<0.01). Multiple regression analysis showed that patients’ age 65 years or greater at baseline (OR 2.3; 95% CI 1.3 to 4, p<0.001), acute coronary syndrome (OR 1.9; 95% CI 1.1 to 3.4, p=0.02), chronic renal failure (OR 2.7; 95% CI 1.4 to 5.2, p<0.001), peripheral artery disease (OR 3.1; 95% CI 1.8 to 5.5, p<0.001) and LVEF ≤45% (OR 2.6; 95% CI 1.4 to 4.5, p<0.001) were independent predictors of long-term cardiovascular mortality.

Conclusions Our longitudinal analysis presents encouraging data concerning operative risk of BIMA grafting and provides excellent long-term survival in appropriately selected patients.

Statistics from

Coronary artery bypass grafting (CABG) is an established treatment of myocardial ischaemia symptoms, which significantly improves survival among certain subgroups of patients.1 Despite advances in percutaneous coronary intervention (PCI), surgical treatment remains the standard of care for patients with severe multi-vessel or left main coronary artery disease, and provides a lower rate of subsequent revascularisation.1 The use of a single internal mammary artery (SIMA) rather than vein graft to the left anterior descending coronary artery has become the standard operation largely due to excellent long-term patency into the first and the second decade.2 ,3 The choice of an IMA graft is indeed supported by the absence of neointimal hyperplasia,4 relative resistance to atherosclerosis and preserved endothelial function of the IMA grafts.5

The superior outcome associated with SIMA grafting has quickly encouraged the use of bilateral IMA (BIMA).6 ,7 However, BIMA grafting is technically more challenging, which still prevents its widespread use.

The aim of our study is to present the short-term and long-term clinical outcomes of BIMA grafting in a large cohort of patients with multi-vessel disease who underwent isolated CABG.


Patient selection

From April 1996 to December 2009, 4210 patients underwent isolated CABG at our university centre. From this group, we identified 1000 patients who underwent primary, isolated surgical revascularisation using BIMA grafting. We excluded patients who experienced isolated CABG using exclusively saphenous vein grafts (SVGs) (20 patients), single IMA graft (205 patients) or IMA and additional SVG (2985 patients). Patients who underwent concomitant repair of valve, cardiac rupture, ventricular aneurysm or ascending aorta were also excluded.

Coronary artery disease was defined as a reduction of the vessel diameter by ≥70% in one view on coronary angiography. The presence of stenosis ≥70% in the left anterior descending, circumflex or right coronary system was used as the criterion for single-vessel, double-vessel or triple-vessel diseases. Before myocardial revascularisation, the heart team, including cardiologists and surgeons, systematically identified the target vessels according to myocardial viability. All myocardial revascularisations were guided by myocardial viability attested by cardiac echocariography, and were completed by dobutamine echocardiography, nuclear imaging or nuclear magnetic resonance according to the physician's judgement. Revascularisation was considered as complete if every significant target vessel was grafted.

Surgical technique

BIMA grafting was performed using standard on-pump or off-pump bypass techniques at the discretion of the operating surgeon. Myocardial preservation during cardiopulmonary bypass involved normothermic, intermittent, anterograde and retrograde blood cardioplegia.

The left internal mammary artery (LIMA was harvested as a pedicle and grafted exclusively to the left coronary system. The right IMA was mostly harvested as a pedicle and grafted to the left coronary system or the right coronary artery and, in a minority of cases, used as a free graft. A free right IMA was used when the length of the conduit was too short, or if the distal anastomotic site was unreachable with a pedicled right IMA. The composite graft included an end-to-side anastomosis of the free right IMA on an in situ LIMA.

The right gastroepiploic artery or radial artery as a third arterial conduit was not used in our study.

Definition of terms and data collection

The in-hospital course was studied in terms of procedural characteristics, vital status, renal status, perioperative and postoperative red blood-cell transfusions, infectious complications, myocardial infarction, cerebrovascular events, mesenteric ischaemia, emergency repeat coronary and/or peripheral revascularisation procedures and rhythmic disorders.

Patients did not undergo systematic control angiography during the follow-up period. Perioperative myocardial infarction (within 7 days after intervention) was defined as a creatine kinase-MB five times or more the upper limit of normal, with new Q waves in two contiguous leads on the postoperative electrocardiogram or the development of new bundle branch block. In the case of elevated creatine kinase levels at baseline, myocardial infarction was defined as an increase of more than 50% over the baseline values after intervention. After this postoperative period, myocardial infarction was defined as the presence of new pathological Q waves or the development of a new left bundle branch block with increased cardiac marker levels (creatine kinase-MB≥3 times the upper limit of normal). The cause of death and the occurrence of myocardial infarction were analysed by the review committee.

Cardiovascular deaths included death resulting from an acute myocardial infarction, heart failure, stroke, pulmonary embolism and peripheral artery disease (aortic aneurysm rupture or mesenteric ischaemia). Cardiovascular events included cardiovascular deaths, myocardial infarction, stroke and subsequent target vessel revascularisation.

All survivors were interviewed by telephone to obtain information regarding subsequent coronary or peripheral ischaemic events and cardiac medical therapy. If subsequent hospitalisation, death or other events had occurred, the patient's physician or appropriate hospital record department were interviewed to document the events.

Statistical analysis

Continuous variables are presented as the mean±SD and compared using the Student's t test or the Wilcoxon test. Categorical data are presented as numbers and frequencies and compared using the χ2 or Fisher's exact tests.

For univariate analyses, we used crude logistic regression models to identify factors associated with cardiovascular mortality. Factors associated with cardiovascular mortality were entered into a stepwise logistic regression model with a significance level to enter set at 0.2 and a significance level to stay set at 0.05.

All statistical analyses were performed using SAS V.9.2 software (Cary, North Carolina, USA).


Patient characteristics

The patients’ baseline and operative characteristics are presented in tables 1 and 2.

Table 1

Clinical and demographical characteristics of the study groups

Table 2

Procedural characteristics of bilateral internal mammary arteries group

Compared with the overall population who experienced isolated CABG using LIMA+SVG during this period (2285 patients), patients who underwent BIMA graft were of younger age (60±15 vs 65.4±14 years p<0.01) and male patients (87.4% vs 77% p<0.01), with low incidence of comorbid conditions including diabetes mellitus (27.1% vs 37.2% p<0.01), peripheral artery disease (27.3% vs 40% p<0.01) and chronic obstructive pulmonary disease (COPD; 11.7% vs 14.3% p=0.02). We did not observe any difference between the two groups concerning the extension of coronary disease, preoperative left ventricular ejection fraction (LVEF) or initial clinical symptoms. The mean number of grafts per patient was 2.47±0.5, and the total revascularisation was obtained in 80% of the cases. No additional radial graft was used.

The LIMA was harvested almost exclusively as a pedicle and grafted on the left coronary system. The LIMA was grafted on the right coronary artery and circumflex as a free graft in one and five patients, respectively. A sequential anastomosis technique for the left anterior descending artery and diagonal branch was used in 32 patients.

The right IMA graft was anastomosed to the left coronary system and the right coronary artery in 79.5% and 20.5% of the patients, respectively. The right IMA was harvested as a pedicle in majority of the cases and was grafted on the circumflex artery as a free graft or on an in situ LIMA in 40 patients.

Early results

The in-hospital overall mortality rate and cardiovascular death rate were 2.8% and 1.6%, respectively (table 3). Early postoperative morbidity included myocardial infarction (2.2%), ischaemic stroke (0.9%), mesenteric ischaemia (0.7%) and acute limb ischaemia (0.6%).

Table 3

In-hospital postoperative course of bilateral internal mammary arteries group

The cumulative rate of cardiovascular events (cardiovascular deaths, myocardial infarction, coronary revascularisation, stroke, mesenteric or limb ischaemia) during the in-hospital course represented 7.9%.

Supraventricular arrhythmias necessitating pharmacological cardioversion occurred in 23.1% of the patients during the 48 h following the procedure. Among them, patients with pre-existing permanent or paroxysmal atrial fibrillation represented 4.6% of the cases and the new onset after CABG within the first 48%. Sternitis requiring redo sternal surgery and local skin infections occurred in 5.9% of the cases. This complication included superficial sternal wound infections without sepsis and positive microbiology in 37 patients (3.7% of the cases), and deep wound infections leading to sepsis in 22 patients (2.2% of the cases). Among patients who experienced mediastinitis, two patients died during the in-hospital course.

Long-term results

The long-term follow-up (mean follow-up: 6.2±5 years) revealed 103 deaths, including 42 cardiovascular deaths. During this follow-up, death or cardiovascular events occurred in 196 patients. We recorded myocardial infarction and stroke in 8 and 17 patients, respectively, and PCI in 95 patients. Among them, BIMA and SVG failure were noted in 29 and 21 patients, respectively, whereas 45 patients underwent subsequent PCI of coronary arteries that were not initially grafted.

The Kaplan–Meier 8-year survival rates for patients less than 65 years and between 65 and 74 years of age were 88% and 66%, respectively (p<0.01) (figure 1). The Kaplan–Meier 8-year freedom from cardiac adverse events was 68% in the whole population.

Figure 1

Kaplan–Meier curves for the probability of death from any cause.

Multiple regression analysis showed that patient age 65 or above at baseline (OR 2.3; 95% CI 1.3 to 4, p<0.001), acute coronary syndrome (OR 1.9; 95% CI 1.1 to 3.4, p=0.02), chronic renal failure (OR 2.7; 95% CI 1.4 to 5.2, p<0.001), peripheral artery disease (OR 3.1; 95% CI 1.8 to 5.5, p<0.001) and LVEF ≤45% (OR 2.6; 95% CI 1.4 to 4.5, p<0.001) were independent predictors of long-term cardiovascular mortality (table 4).

Table 4

Factors associated with long-term cardiovascular mortality of bilateral internal mammary arteries group


We present a large, single-centre observational study of surgical coronary revascularisation using BIMA grafts, which represented a sizeable proportion of all procedures performed during this period (24% of the cases). Obviously, the choice of revascularisation using BIMA grafts was directly influenced by well-known comorbid factors, such as diabetes mellitus, COPD, obesity and the age of elderly patients.8 ,9 Our cohort included relatively young patients (mean age: 60±15 years) with a lower incidence of diabetes mellitus and peripheral artery disease as compared with patients who underwent LIMA–SVG grafting. In our study, the early clinical outcome after CABG was consistent with other contemporary reports of a 1%–2% mortality rate and a 1%–2% morbidity rate for each of the following morbidities: myocardial infarction, stroke and renal failure regardless of surgical revascularisation strategy.1 ,10 ,11

Sternal complications are the main limitation in the use of BIMA grafts.

One concern with the widespread use of BIMA is the risk of impaired sternal healing, especially in diabetic, elderly, and obese patients.12

The previously published higher incidence of sternal infections in BIMA graft compared with SIMA graft was confirmed by a recent Arterial Revascularisation Trial analysis.10

In our study, patients experiencing wound complications were mostly diabetics (45%), which in accordance with previous studies recommended not to use BIMA in these groups of patients.8 Matsa et al8 specified that the increased risk of infection in diabetic patients was especially prominent in obese female patients. However, diabetic patients have the most to gain from BIMA grafts, and harvesting the IMA with a skeletonisation technique, preserving collaterals and sternal blood supply, reduces this adverse outcome.13

We also confirmed the better favourable long-term outcomes of BIMA, including survival and freedom from major adverse cardiac events similar to those in previously published studies compared with SIMA and SVGs.8 ,14 At 8 years, the Kaplan–Meier survival rates for patients below 65 years and between 65 and 74 years of age were 88% and 66%, respectively, and freedom from cardiac adverse events for the overall population was 68%. BIMA grafting seems to produce better long-term outcomes than SIMA grafting during the second postoperative decade, when 75% of the vein conduits are occluded or severely diseased.6 This vein graft failure leads to worse survival, late myocardial infarction, a high risk of recurrent angina and the need for further intervention.15 The use of BIMA grafting in elderly patients remains controversial.14 ,16 The evolution in surgical techniques and patient selection for CABG due to demographical changes has led to its application in older and sicker patients with more complex disease. In a large series, Pevni et al9 showed higher mortality in patients more than75 years of age, whereas according to Galbut et al,14 the late survival rate seemed to be better with BIMA grafting than SIMA grafting Thus, we consider that life expectancy is the key point of the revascularisation strategy more than a cut-off and we should not exclude some specifically high-risk subgroups who have the most to gain from this surgical revascularisation strategy.

The concept of global vascular risk is also recognised as an important predictor of major events in patients with coronary artery disease.17 ,18 Our data confirm that the presence of peripheral arterial disease remains an independent predictive factor for all-cause mortality in patients regardless of the surgical revascularisation strategy.18

The significantly greater risks of CABG among patients with peripheral arterial disease were not unexpected. These patients tended to be older with comorbidities, such as chronic renal dysfunction and pulmonary insufficiency, and thus represented a high-risk population. The extent of coronary artery disease in these patients was also more diffuse, which, in part, contributed to incomplete revascularisation and long-term graft occlusion.

Non-ST-elevation acute coronary syndrome represented the initial clinical presentation in 41% of our cases. The rationale for early, non-emergent CABG after non-ST elevation acute coronary syndrome is primarily to reduce cardiac adverse events such as infarction, death and recurrent myocardial ischaemia and to preserve ventricular function.19 On the other hand, early CABG strategy is associated with an increase in inflammation and platelet dysfunction associated with plaque rupture.20 As there is no randomised study comparing early and delayed CABG strategies, the appropriate timing is generally individualised based on the symptoms, haemodynamic status and coronary anatomy.19

If thrombotic risks and haemodynamic instability remain risk factors in the postoperative course, the long-term prognosis of patients with acute coronary syndrome undergoing CABG is also influenced by the high proportion of patients with more comorbidities, diffuse coronary disease and left ventricular dysfunction.

A significant proportion of patients included in our study presented with left ventricular dysfunction, a well-known factor associated with both an increased risk of perioperative mortality and a worse late outcome regardless of surgical strategy.21 ,22 Traditionally, the progression of heart failure has been attributed to left ventricular remodelling and has been thought to be unrelated to the causes of left ventricular dysfunction.22

The available data suggest that comorbidities and the presence and extent of coronary artery disease may accelerate the progression of heart failure, explaining the higher mortality among ischaemic compared with non-ischaemic heart failure patients.23

The evidence for the long-term impact of CABG in coronary patients with heart failure remains uncertain and is often limited to observational cohorts.24–26 In recent studies, CABG was associated with improved survival, especially in those with lower LVEF.27 The benefit concerned patients with ischaemic but viable myocardium who subsequently underwent revascularisation.27 ,28

However, the lack of interaction between myocardial viability status and the benefit from CABG in STICH trial indicates that assessment of myocardial viability alone should not be the deciding factor in selecting the best therapy for these patients. Other structural predictive parameters such as left ventricular volumes and ischaemic mitral regurgitation should be taken into account for patient selection.

Therefore, further stratified analyses should be encouraged to identify the exact benefit from BIMA grafting in this situation.


Our study was a non-randomised, retrospective and observational study with inherent bias.

The lack of a control group is one of the limitations of this study. A propensity score-matched study using a control group with patients who underwent LIMA+SVG revascularisation during the same period would have made our analysis more powerful. However, the interest of our study was to obtain the most complete and conclusive cardiovascular data during a long-term follow-up, which is unusual in such large series.

Other variables may also influence the results of surgical revascularisation as the quality of coronary vessels requiring bypass grafting and the location of both grafts.

Another limitation of the study is the low use of off-pump grafting reducing the postoperative rate of stroke.29 However, postoperative rate of stroke remains low and this low rate is probably directly influenced by the patient selection bias: a younger population with a lower proportion of peripheral artery disease. Although the comparison of patients undergoing single IMA versus BIMA grafting has been largely studied, the hybrid approaches using BIMA and elective PCI using drug-eluting stents should represent the next step forward towards an optimisation of myocardial revascularisation.30


Our longitudinal analysis confirms that bilateral IMA grafting presents encouraging data concerning periprocedural course and provides excellent long-term survival. This surgical revascularisation strategy should be widely used in appropriately selected patients.

View Abstract


  • Contributors BP designed data collection tools, monitored data collection for the study. He is guarantor. DV monitored data collection and revised the draft paper. NA wrote the statistical analysis plan, cleaned and analysed the data and drafted and revised the paper. PM, FV, EA, TF, JPV initiated the collaborative project, analysed the data and drafted and revised the paper.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethics committee of the authors’ institution.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles