A Comparison of Bilateral with Single Internal Mammary Artery Grafts on Postoperative Mediastinal Drainage and Transfusion Requirement

Berroeta, Clarisse MD*; Benbara, Abdel MD*; Provenchère, Sophie MD*; Ajzenberg, Nadine MD, PhD†; Benessiano, Joelle PhD‡; Depoix, Jean-Pol MD*; Desmonts, Jean-Marie MD*; Iung, Bernard MD§; Philip, Ivan MD*

doi: 10.1213/01.ane.0000242514.49621.0c
Cardiovascular Anesthesia: Editorial

The superiority of the left internal mammary artery (LIMA) graft over autogenous saphenous vein as a bypass conduit in coronary artery bypass surgery has been well established. Early and late patency rates of bilateral internal mammary artery (BIMA) grafts exceed those of vein grafts, and patients who receive BIMA have improved long-term survival rates and more freedom from reoperations and other cardiac events. But because of other concerns, particularly the question of increased risk of postoperative bleeding, controversy still surrounds the perioperative period. In the present study we sought to determine whether BIMA grafting was an independent risk factor of postoperative bleeding and of blood product use in patients undergoing primary elective coronary artery revascularization. For this purpose, 33 consecutive patients scheduled for BIMA grafting were matched with 66 patients operated on by single LIMA grafting. Patients in the LIMA group had significantly less postoperative mediastinal drainage than those in the BIMA group (median: 722 vs 920 mL, P = 0.0001). Fifty-six patients received blood products (56% vs 51% in LIMA and BIMA groups, respectively; P = 0.67). In multivariate analysis, BIMA and operative duration were independent predictors of increased postoperative drainage. Nevertheless, in logistic regression, BIMA was not significantly associated with blood product use, unlike precardiopulmonary bypass hematocrit and duration of surgery (OR and 95% CI: 0.89 [0.80–0.96] P = 0.01; 1.009 [1.001–1.019] P = 0.04, for an increase of 1% in hematocrit and 1 min in duration of surgery, respectively). In conclusion, these data support the idea that BIMA graft slightly increases postoperative drainage but not transfusion requirement.

IMPLICATIONS: Bilateral internal mammary artery (BIMA) graft increases postoperative drainage. However, the clinical relevance of this risk remains low, since bleeding is moderate and does not lead to a higher transfusion requirement. Thus, the practice of BIMA should not be discouraged for fear of increased blood loss.

From the *Département Anesthésie-Réanimation; †Laboratoire d'Hématologie; ‡Centre d'Investigation Clinique; and §Service de Cardiologie, Hôpital Bichat-Claude Bernard Assistance Publique, Hôpitaux de Paris, France.

Accepted for publication August 3, 2006.

Address correspondence to: Dr C. Berroeta, Département d'Anesthésie, Hôpital Bichat-Claude Bernard 46, rue Henri Huchard, 75877 Paris Cedex 18, France. Address e-mail to clarisse.berroeta@bch.ap-hop-paris.fr. Reprints will not be available from the author.

Article Outline

The superiority of the left internal mammary artery (LIMA) graft over autogenous saphenous vein as a bypass conduit in coronary artery bypass graft (CABG) surgery has been well established. Early and late patency rates of bilateral internal mammary artery (BIMA) grafts exceed those of vein grafts and patients who receive BIMA have improved long-term survival rates and more freedom from reoperations and other cardiac events (1–3). The better size match with the coronary artery, the single anastomosis, and the biochemical and physical qualities of this conduit for the most important coronary artery in the human circulation, the left anterior descending artery (LAD), are clear advantages over saphenous vein graft. Consequently, the use of bilateral internal thoracic artery grafts have been also proposed (4–8).

Several authors have examined the early and long-term outcomes in groups of patients undergoing single versus BIMA grafting (9). Operative mortality did not differ between strategies (7) and one study even reported a protective effect (2). In a large retrospective study, Lytle et al. (6) concluded that patients who received two internal mammary arteries (IMA) had decreased risk of death, reoperation, and angioplasty compared with patients who received single IMA. As freedom from cardiac events is a main target of any revascularization procedure, BIMA grafting should be proposed in patients younger than 75 yr, especially if life expectancy is more than 10 yr (10).

However, while the concept of BIMA grafting is becoming well established to improve late survival, controversy still surrounds the perioperative period, particularly concerning the increased risk of sternal wound infections and postoperative bleeding (7,11).

An increased risk of bleeding has been reported after BIMA grafting, leading some teams to consider this type of surgery as high risk for bleeding (12,13). However, the consequences on reexploration for bleeding remain controversial (12,14–16). Furthermore, no data are available concerning blood product use.

The objective of the present study was to determine whether BIMA grafting was an independent risk factor of postoperative bleeding and of blood product use in patients undergoing coronary artery revascularization. For this purpose, consecutive patients scheduled for BIMA grafting were matched with patients operated on by LIMA grafting during the same period.

Back to Top | Article Outline


Patient Population

During six consecutive months, patients undergoing CABG surgery with internal thoracic artery under cardiopulmonary bypass (CPB) were eligible for the study. The protocol was approved by our local ethics committee, which waived the need of written informed patient consent. The choice of the type of operation was made by the surgeon in accordance with the cardiologist. Patients undergoing CABG with BIMA grafts were included in the BIMA group. Each of these patients was matched with two patients in the LIMA group, during the same period, based on the number of grafts, the surgeon, age within 5 yr, preoperative hematocrit (within 3%), and body mass index (within 2 kg/m2). Case matching was performed from our local database by two individuals (NA, BI) who were blinded to the postoperative outcome.

Preoperative factors for noninclusion were: emergent surgery, IV administration of heparin, antiplatelet drug other than aspirin (clopidogrel), thrombolysis, history of heart surgery, chronic renal failure (baseline serum creatinine >200 μmol/L) or hepatic disease, thrombocytopenia (platelet count <150 G/L), coagulopathy disease, anemia (hematocrit <35%), and weight <50 kg. Intraoperative exclusion factors were administration of aprotinin and requirement of intraaortic balloon pump.

Back to Top | Article Outline

Anesthetic and Surgical Management

Premedication was standardized for all patients; β-blocker drugs were given until the day of surgery. Aspirin and low-molecular-weight heparin were continued until the day before surgery. Standardized anesthesia was used in all patients: induction by hypnomidate (0.3–0.4 mg · kg−1, fentanyl (3–5 μg · kg−1), pancuronium bromide (0.1 mg · kg−1) and maintenance of anesthesia by fentanyl (total dose 13–15 μg · kg−1) and isoflurane until CPB and then propofol.

Monitoring techniques were also standardized: electrocardiogram (ECG), arterial and pulmonary artery catheters. During CPB normothermia (bladder temperature >36°C) was maintained with a perfusion temperature of 37°C. CPB was nonpulsatile; membrane oxygenators were always used. Flow rates of 2.4 L · min−1 · m−2 were used. Myocardial protection was achieved by intermittent anterograde cold blood cardioplegia and a warm reperfusion just before the removal of the aortic cross-clamp. Heparin (300 IU · kg−1) was administered by the surgeon in the right atria before cannulation. An activated clotting time (ACT) >400 s was required for the onset of CPB, as well as during CPB (if necessary an additional dose of 5000 IU of heparin was injected). After the end of CPB, protamine sulfate was administered at the ratio 1:1. Before sternal closure and chest tube insertion, each thoracic cavity was emptied.

Tranexamic acid (Exacyl®; Sanofi Winthrop, Gentilly, France) was administered to all patients (20 mg · kg−1 at the incision, and then 2 mg · kg−1 · h−1 until the end of surgery). IV prophylactic antibiotic treatment over the first 24 h consisted of cefamandol (60 mg · kg−1 · d−1), or vancomycin (30 mg · kg−1 · d−1) plus aminoglycoside (gentamicin, 3 mg · kg−1 · d−1) in case of allergy to cephalosporins.

Postoperative care was delivered in an intensive care unit (ICU) by anesthesiologists, and then in the ward by cardiologists. Aspirin was resumed on day one after surgery in all patients. As routinely performed in our institution, the cardiac troponin I concentration was measured before cardiac surgery and at the 20th postoperative hour (Dade-Behring, RXL HM). Mediastinal drainage was measured at hourly intervals in the ICU and mediastinal drains were removed after the 36th postoperative hour, when blood loss was <100 mL over 4 h. For each patient the quantity of mediastinal tube drainage at 24 h and before removal (total drainage) was recorded. All perioperative variables were prospectively recorded and retrospectively analyzed.

Back to Top | Article Outline

Surgical Procedures

All operations were undertaken through median sternotomy. IMAs were dissected with electrocautery, and small-size branches were clipped or cauterized. The artery was harvested with both accompanying veins as well as a generous surrounding pedicle of tissue. Left and right IMA were prepared with diluted solution of papaverine applied topically. The left IMA was always directed to the LAD artery. The right IMA was always severed from the subclavian artery.

Back to Top | Article Outline

Transfusion Policy

Transfusion therapy was directed toward the goal of maintaining the hematocrit above 20% during CPB and 25% after CPB. Postoperatively, packed red blood cells (PRBCs) were transfused for a hematocrit <25%. Excessive bleeding was defined as a drainage >200 mL · h−1 for the initial 3 h and >100 mL · h−1 thereafter. Treatment of excessive bleeding included protamine, if the activated partial thromboplastin time or ACT were elevated, platelet transfusion, for a count of <100 g/L or–suspected platelet dysfunction, and fresh-frozen plasma (FFP) transfusion, for an International Normalized Ratio >1.5 (or prothrombin time >16.1 s) when continuing bleeding could not be resolved by platelet transfusion.

Back to Top | Article Outline

Primary End Point

Total mediastinal drainage was recorded in all patients. The total number of blood units and of each type of blood product (PRBCs, FFP, and platelets) transfused was also recorded. Patients were coded as having been transfused or not.

Back to Top | Article Outline

Statistical Analysis

The sample sizes of both groups were calculated as follows: According to a previous study in our institution (17), we took a standard deviation of 500 mL for postoperative bleeding. For an expected difference of 300 mL, and assuming a two-tailed a value of 0.05 and a β-risk of 0.20, we determined that at least 99 patients should be analyzed in our study (33 and 66 patients in BIMA and LIMA groups respectively).

Results are expressed as mean ± sd, median [25th–75th percentiles], or percentage as appropriate. The comparison of hematocrit changes between groups was made by a two-factor analysis of variance for repeated measures. The link between postoperative blood loss and perioperative variables was studied by the Mann–Whitney test or with the Spearman's rank order test. The comparison of patients with or without transfusion was assessed by Student's t-test or the χ2 test.

The variables identified by univariate analysis with P < 0.20 were included in two multivariate models:

a generalized linear regression model was used for the quantitative prediction of bleeding. Because of the skewed distribution of bleeding, modeling was performed on its logarithmic transformation.

a multiple logistic regression procedure for blood product use. Variables were selected using a backward procedure with P = 0.05.

All tests were two-tailed and P < 0.05 was considered as significant. Analysis was performed on SAS statistical software (8.2; Cary, NC).

Back to Top | Article Outline



Ninety-nine patients were included in the study. During this period of time, 226 patients underwent CABG. Five patients had emergency surgery and were not included (1 in the BIMA group). Seven were excluded because aprotinin was administered (preoperative administration of clopidogrel, n = 4, among which two were in the BIMA group; redo surgery, n = 3), and two were excluded for postoperative intraaortic balloon pump requirement. Both groups of patients (BIMA versus LIMA) were not different with regards to perioperative variables, except for CPB duration (Table 1). The median number of grafts was 3 (range: 2–4). Case matching was always realized except for surgeons (two cases) and for the number of grafts (one case). Reoperation for bleeding was necessary in one patient in each group. The postoperative cardiac troponin I concentration did not differ between groups. Three patients died; one in the BIMA group and two in the LIMA group (after removal of mediastinal drains, on days 5, 6, and 10 respectively).

Back to Top | Article Outline

Postoperative Mediastinal Drainage

Median mediastinal drainage was 840 mL [720–1190]. Patients of the BIMA group had significantly more postoperative mediastinal drainage than those in the LIMA group (Table 1, P = 0.0001; odds ratio and 95% CI for a bleeding >1000 mL:2.15 [0.92–5.01]). As shown in Table 2, BIMA and operative duration were the strongest predictors of increased postoperative drainage. In multivariate analysis, these two factors remained independent predictors of increased postoperative drainage (P = 0.01 and 0.04, respectively).

Back to Top | Article Outline

Requirement for Homologous Transfusion

Hematocrit changes during the perioperative period were similar in both groups. Fifty-six patients (56%) were transfused: 56 patients received PRBCs, 14 received platelets, and eight received FFP. No significant difference was found between groups (Table 1, P = 0.67). In univariate analysis, variables significantly associated with homologous transfusion requirement were body weight, female sex, low left ventricular ejection fraction, low pre-CPB hematocrit, and duration of surgery (Table 3). In logistic regression, pre-CPB hematocrit and duration of surgery were significantly associated with blood product use (Table 4). The use of BIMA was not significantly associated with transfusion requirement.

Back to Top | Article Outline


Our results demonstrate that the use of BIMA graft is an independent risk factor for increased chest drainage volume, but does not increase transfusion requirement.

The present study is a single-center study in which consecutive patients were included during a short period of time and managed by the same surgical and anesthetic teams. This ensured homogeneous perioperative care. To limit the potential bias due to the absence of randomization, the authors performed a case–control study by carefully matching perioperative data of each patient and the use of multivariate analysis to adjust for the other risk factors, bleeding, and transfusion. Even at institutions experienced with extensive arterial grafting, patients receiving BIMA grafts have been selected subgroups, and their selection has been based on patient-related and surgeon-related factors. In the literature, no randomized trials have compared LIMA and BIMA strategies.

The clinical benefit of the left IMA to the LAD bypass graft is now a well-proven principle of CABG surgery (9,15,18,19). However, controversy still persists concerning the clinical advantages and/or perioperative complications of BIMA grafting. Several studies suggest that BIMA graft is associated with improved long-term graft patency, superior survival, and decreased risks of angina recurrence, myocardial infarction, and coronary reoperation (15,20). Nevertheless, higher perioperative morbidity has been related to an increased rate of postoperative bleeding and wound infection (7,12–14,21,22). The increased risk of postoperative bleeding found in the present study is in agreement with the literature (12,13). This may be attributed to an enlarged endothoracic wound and a prolonged operative duration.

The clinical consequences of increased postoperative bleeding can be assessed by the reoperation rate and blood product use. Previous studies suggest that reoperation for bleeding occurred two to three times more frequently after BIMA (21,22). More recent data has not confirmed this point (15,16), and our study was not designed to explore it. Few data were available concerning transfusion requirement. In the present study, BIMA grafting was not an independent risk factor for blood product use, as opposed to data reported by Cosgrove et al. (7). In our study, the two predictors of transfusion were low preoperative hematocrit and operative duration. In the literature, preoperative red blood cell mass is one of the strongest preoperative factors for transfusion requirement (23,24).

Finally, several hypotheses may explain the apparent discrepancy between the enhanced postoperative bleeding that we report and the absence of more frequent transfusion requirement. First, it must be pointed out that, although statistically significant, the observed difference in blood loss remains limited between techniques, at around 200 mL. Conversely, Taggart et al. reported that the use of BIMA grafts increases mean postoperative blood loss by approximately 400 mL (25). Moreover, we found that changes in hematocrit were similar in both groups which suggest that blood loss (unmeasured) from the saphenous vein dissection may be under-estimated, and that total chest drainage volume does not correspond to actual hemoglobin loss. Consequently, postoperative bleeding should be designated more appropriately as postoperative “drainage.” To have actually measured total hemoglobin loss, it would have been necessary to measure hematocrit in the mediastinal and pleural effusion. Nevertheless, it cannot be measured easily.

Our transfusion rate, around 50%, may appear important. However, recent papers report similar results (24,26), and thus our data can be generalized to other institutions. To avoid a potential bias we excluded patients receiving clopigrel. Indeed, bleeding and requirement of transfusion appeared higher in these patients and they routinely received full-dose aprotinin. In a recent study, 79% of patients receiving clopidogrel <5 days before surgery were exposed to blood products in the absence of an antifibrinolytic drug (27). At the time of our study, few patients (4%) received clopidogrel preoperatively. At the present time this percentage remains small, at around 15%, which suggests that our results still can be generalized.

The clinical relevance of our results is that the choice of BIMA grafts should not be discouraged for fear of increased blood loss. Consequently, BIMA should influence neither preoperative antiplatelet therapy (preoperative interruption of aspirin), nor the choice of antifibrinolytic drug (aprotinin rather than tranexamic acid).

We conclude that these data confirm that BIMA graft increases the amount of postoperative drainage. However, the clinical relevance of our findings remains low since the increase in bleeding is moderate, and does not lead to increased transfusion requirement.

Back to Top | Article Outline


1. Loop FD, Lytle BW, Cosgrove DM, et al. Influence of the internal-mammary-artery graft on 10-year survival and other cardiac events. N Engl J Med 1986;314:1–6.
2. Leavitt BJ, O'Connor GT, Olmstead EM, et al. Use of the internal mammary artery graft and in-hospital mortality and other adverse outcomes associated with coronary artery bypass surgery. Circulation 2001;103:507–12.
3. Edwards FH, Clark RE, Schwartz M. Impact of internal mammary artery conduits on operative mortality in coronary revascularization. Ann Thorac Surg 1994;57:27–32.
4. Rizzoli G, Schiavon L, Bellini P. Does the use of bilateral internal mammary artery (IMA) grafts provide incremental benefit relative to the use of a single IMA graft? A meta-analysis approach. Eur J Cardiothorac Surg 2002;22:781–6.
5. Pick AW, Orszulak TA, Anderson BJ, Schaff HV. Single versus bilateral internal mammary artery grafts: 10-year outcome analysis. Ann Thorac Surg 1997;64:599–605.
6. Lytle BW, Blackstone EH, Loop FD, et al. Two internal thoracic artery grafts are better than one. J Thorac Cardiovasc Surg 1999;117:855–72.
7. Cosgrove DM, Lytle BW, Loop FD, et al. Does bilateral internal mammary artery grafting increase surgical risk? J Thorac Cardiovasc Surg 1988;95:850–6.
8. Myers WO, Berg R, Ray JF, et al. All-artery multigraft coronary artery bypass grafting with only internal thoracic arteries possible and safe: a randomized trial. Surgery 2000;128:650–9.
9. Taggart DP, D'Amico R, Altman DG. Effect of arterial revascularisation on survival: a systematic review of studies comparing bilateral and single internal mammary arteries. Lancet 2001;358: 870–5.
10. Calafiore AM, Di Giammarco G, Teodori G, et al. Late results of first myocardial revascularization in multiple vessel disease: single versus bilateral internal mammary artery with or without saphenous vein grafts. Eur J Cardiothorac Surg 2004;26:542–8.
11. Grossi EA, Esposito R, Harris LJ, et al. Sternal wound infections and use of internal mammary artery grafts. J Thorac Cardiovasc Surg 1991;102:342–6; discussion 6–7.
12. Uva MS, Braunberger E, Fisher M, et al. Does bilateral internal thoracic artery grafting increase surgical risk in diabetic patients? Ann Thorac Surg 1998;66:2051–5.
13. Gansera B, Gunzinger R, Angelis I, et al. End of the millennium— end of the single thoracic artery graft? Two thoracic arteries— standard for the next millenium? Early clinical results and analysis of risk factors in 1,487 patients with bilateral internal thoracic artery grafts. Thorac Cardiovasc Surg 2001;49:10–15.
14. Walkes JC, Earle N, Reardon MJ, et al. Outcomes in single versus bilateral internal thoracic artery grafting in coronary artery bypass surgery. Curr Opin Cardiol 2002;17:598–601.
15. Endo M, Nishida H, Tomizawa Y, Kasanuki H. Benefit of bilateral over single internal mammary artery grafts for multiple coronary artery bypass grafting. Circulation 2001;104:2164–70.
16. Cassese M, Speziali G, Martinelli GL, Diena M. Reduced complication rate in bilateral mammary artery-to-coronary artery bypass grafting. Ann Thorac Surg 1998;65:1841–2.
17. Provenchere S, Plantefeve G, Hufnagel G, et al. Renal dysfunction after cardiac surgery with normothermic cardiopulmonary bypass: incidence, risk factors, and effect on clinical outcome. Anesth Analg 2003;96:1258–64.
18. Endo M, Tomizawa Y, Nishida H. Bilateral versus unilateral internal mammary revascularization in patients with diabetes. Circulation 2003;108:1343–9.
19. Cohn L. Use of the internal mammary artery graft and in-hospital mortality and other adverse outcomes associated with coronary artery bypass surgery. Circulation 2001;103:483–4.
20. Stevens LM, Carrier M, Perrault LP, et al. Single versus bilateral internal thoracic artery grafts with concomitant saphenous vein grafts for multivessel coronary artery bypass grafting: effects on mortality and event-free survival. J Thorac Cardiovasc Surg 2004;127:1408–15.
21. Fiore AC, Naunheim KS, Dean P, et al. Results of internal thoracic artery grafting over 15 years: single versus double grafts. Ann Thorac Surg 1990;49:202–8; discussion 8–9.
22. Barner HB, Standeven JW, Reese J. Twelve-year experience with internal mammary artery for coronary artery bypass. J Thorac Cardiovasc Surg 1985;90:668–75.
23. Moskowitz DM, Klein JJ, Shander A, et al. Predictors of transfusion requirements for cardiac surgical procedures at a blood conservation center. Ann Thorac Surg 2004;77:626–34.
24. Dial S, Delabays E, Albert M, et al. Hemodilution and surgical hemostasis contribute significantly to transfusion requirements in patients undergoing coronary artery bypass. J Thorac Cardiovasc Surg 2005;130:654.
25. Taggart DP, Djapardy V, Naik M, Davies A. A randomized trial of aprotinin (Trasylol) on blood loss, blood product requirement, and myocardial injury in total arterial grafting. J Thorac Cardiovasc Surg 2003;126:1087–94.
26. Bybee KA, Powell BD, Valeti U, et al. Preoperative aspirin therapy is associated with improved postoperative outcomes in patients undergoing coronary artery bypass grafting. Circulation 2005;112:I286–92.
27. van der Linden J, Lindvall G, Sartipy U. Aprotinin decreases postoperative bleeding and number of transfusions in patients on clopidogrel undergoing coronary artery bypass graft surgery: a double-blind, placebo-controlled, randomized clinical trial. Circulation 2005;112:I276–80.
© 2006 International Anesthesia Research Society