Risk and Results of Bypass Grafting Using Bilateral Internal Mammary and Right Gastroepiploic Arteries

Transcription

1 Risk and Results of Bypass Grafting Using Bilateral Internal Mammary and Right Gastroepiploic Arteries Olivier Jegaden, MD, Armand Eker, MD, Pietro Montagna, MD, Jean Ossette, MD, Guy De Gevigney, MD, G6rard Finet, MD, Alain Saint Pierre, MD, Didier Revel, Roland Itti, MD, and Philippe H. Mikaeloff, MD Department of Cardiovascular Surgery, Hopital Cardiologique Louis Pradel, Lyon, France From January 1990 to June 1994, 240 patients (mean age, years) underwent myocardial revascularization with the exclusive use of in situ bilateral internal mammary and right gastroepiploic arteries. Left ventricular function was normal in 34% of patients, moderately impaired in 58.5%, and severely impaired in 7.5%. The mean number of distal anastomoses was 3.5 ± 0.7 and the rate of complete myocardial revascularization was 80%. Early mortality was 0.4%, and complications occured in 20 patients: myocardial infarction, 1.6%; intraaortic balloon pump, 0.8%; reoperation for bleeding, 0.8%; and mediastinitis, 0.4%. Early (15th postoperative day) angiographic control of grafts was performed in 51 patients; the rate of functional and patent anastomoses was 100% for internal mammary arteries and 96% for gastroepiploic arteries. Early functional results (3 ± 1 postoperative months) were studied in 141 patients during exercise test with medical treatment: 99% were symptom-free and 14% had ischemic modification of electrocardiograms. A 2-year postoperative functional assessment without medical treatment was performed during exercise test in 66 patients: 98% were symptom-free and 26% had ischemic modification of electrocardiograms; during the same procedure, thallium myocardial scintigraphy was obtained in 50 patients and 18 patients had moderate ischemic defect on exercise. Ischemic modifications of electrocardiograms and defects seen on thallium scintigraphy were correlated significantly with incomplete revascularization (p < 0.05). Ischemic defects seen on thallium scintigraphy demonstrated limited gastroepiploic artery and right internal mammary artery flow at the maximum level of exercise in at least 8% and 4% of patients, respectively. The 2-year and 4-year actuarial survival was 96.5% + 4.2%; the rate of late cardiac events was 0.6% per patientyear. This technique can be achieved with minimal operative risk and offers good functional results and midterm survival rate. (Ann Thorac Surg 1995;59:955-60) he internal mammary artery (IMA) has become the T conduit of choice and the reference material in the field of coronary artery bypass surgery [1-3]. Myocardial revascularization using both IMAs now is performed increasingly with a reasonable operative risk [4, 5l. The advantages of bilateral IMA grafts over single IMA grafts are still a controversial subject, doubtlessly because in the series reported [6, 7] patients underwent saphenous vein bypass associated with IMA bypass. The technique of sequential IMA bypass allows the surgeon to increase the number of arterial anastomoses [8, 9]. The right gastroepiploic artery (GEA) may be considered as a third reliable arterial pedicle for myocardial revascularization, as reported by Suma [10], Mills [11], Lytle [12], and their co-workers, with good midterm results [13]. In this study, we report our experience with myocardial revascularization based on the exclusive use of both IMAs and GEA, and we try to answer the two following Accepted for publication Dec 29, Address reprint requests to Dr Jegaden, Hopital Cardiologique Louis Pradel, BP Lyon-Monchat, Lyon Cedex 03, France. questions: Can this technique be a reasonably routine technique; and do the observed results justify its use? Material and Methods Patients From January 1990 to June 1994, 240 patients underwent myocardial revascularization with exclusive use of both IMAs and GEA. They represent 30% of all patients undergoing isolated coronary artery bypass grafting during the same period and performed by the same surgeon (O.J.), from 10% in 1990 to 45% in 1994 (Fig 1). The series consisted of 220 male and 20 female patients with a mean age of years (range, 26 to 80 years). All patients presented with disabling angina (Canadian Cardiovascular Society class III or IV), and 15 of them exhibited signs of heart failure (New York Heart Association class II or III); 49% of patients had a previous myocardial infarction. All patients had three-vessel disease, and 28 patients had a left main coronary artery stenosis. The mean left ventricular ejection fraction was (range, 0.23 to 0.86); left ventricular function was normal in 82 patients, moderately altered in 140 patients, and severely altered 1995 by The Society of Thoracic Surgeons /95/$ (95)00017-F

2 956 JEGADEN ET AL Ann Thorac Surg BOTH IMAs AND GEA GRAFTING 1995;59: % % -~- 3 Art.Grafts X "*- 2 Art. Grafts 45% 33% Table 2. Sites of Anastomoses Site RIMA LIMA GEA Total LAD Diagonal Marginal RCA PDA PLA Total % 10% I I I, I ~ 22% Fig 1. Yearly rates of coronary artery bypass surgery using three arterial (Art.) grafts (both internal mammary arteries [IMAsl and right gastroepiploic artery [GEA], 240 patients), two arterial grafts (both IMAs, 161 patients; one IMA and GEA, 95 patients), and one arterial graft (one IMA associated with sequential vein graft, 308 patients). (~r 1994 includes cases between January and June only.) (left ventricular ejection fraction <0.35) in 18 patients. At the beginning of our experience, this technique was reserved for young patients with stable angina and good left ventricular function; progressively, indications were extended (see Table 1) to aged patients (16%), unstable angina (12%), severe left ventricular dysfunction (7.5%), left main coronary artery stenoses (12%), reoperations (4%), and patients with a metabolic risk (11%). Table 1. Preoperative Patient Characteristics % Characteristic No. Percent Age (y) < to > CCS class IIl IV NYHA class 0 to I II to III 15 6 Previous myocardial infarction Left main stenosis Diabetes 20 8 Chronic dialysis 6 3 Previous CABG 9 4 LVEF < to > CABG = Coronary Artery Bypass Grafting; CCS = Canadian Cardiovascular Society; LVEF = left ventricular ejection fraction; NYHA = New York Heart Association. GEA = right gastroepiploic artery; LAD = left anterior descending coronary artery; LIMA = left internal mammary artery; PDA = posterior descending coronary artery; PLA - posterolateral coronary artery; RCA - right coronary artery; R1MA right internal mammary artery. Operative Technique All procedures were performed by the same surgeon (O.J.) and according to the same protocol. All patients received a Swan-Ganz catheter (Baxter Healthcare Corp, Edwards Division, Santa Ana, CA) inserted before induction of standard anesthesia. A high-dose aprotinin protocol was used. Median sternotomy was done without electrocoagulation. Standard cardiopulmonary bypass techniques were applied using a COB-CML membrane oxygenator (COB BCT Inc, Denver, CO) with moderate hemodilution. For myocardial protection, slushed ice was used as topical cooling, and an oxygenated cardioplegic solution [14] was administered after aortic crossclamping using the combined antegrade/retrograde perfusion. Exclusive use was made of in situ bilateral IMA and GEA grafts. Internal mammary artery grafts were dissected without electrocoagulation, using metallic clips for the hemostasis of branches. The GEA was dissected according to Suma and associates' description [10]. Periadventitial infiltration of grafts with papaverine was performed systematically. The distal part of the arterial grafts was sectioned in the heparinized patient just before cardiopulmonary bypass. Distal anastomoses were performed under continuous aortic cross-clamping and done with 8-0 polypropylene continuous suture; magnification of 3.5 was used. Complete myocardial revascularization was defined as bypass of all clinically significant coronary artery lesions. The first GEA anastomosis was performed to bypass the right coronary artery or its branches. The GEA pedicle was brought through a hole in the diaphragm, passing the stomach and the liver anteriorly. Depending on the site of the anastomosis, the length of the GEA pedicle was as short as possible to obtain a large GEA for the anastomosis without tension on the GEA pedicle. The orientation of the GEA pedicle and anastomosis varied according to the site of the bypass: 90-degree angle for distal right coronary artery bypass, 0 degrees for the bypass of the posterior descending coronary artery and the posterolateral coronary artery. The sites of GEA grafting are shown in Table 2. Left IMA anastomoses then were performed for bypass of the diagonal arteries, marginal branches of the circum-

3 Ann Thorac Surg JEGADEN ET AL ;59: BOTH IMAs AND GEA GRAFTING flex artery, or both. Right IMA anastomosis finally was performed, generally for bypassing the left anterior descending coronary artery. In 7 patients with markedly dilated left ventricle, the right IMA was too short to allow left anterior descending artery revascularization and therefore the IMAs were cross-used: the left IMA for left anterior descending artery bypass and the right IMA for bypass of the diagonal and marginal arteries. In 101 patients IMAs were used as sequential grafts (6 right and 95 left IMAS): side-to-side anastomoses were performed first, then end-to-side anastomoses according to Jones and associates' description [15]. The sites of IMA grafting are shown in Table 2. The mean number of distal anastomoses per patient was (range, 3 to 6), and increased from 3 in 1990 to 3.8 in 1994 (p < 0.01) (Fig 2). The rate of complete revascularization was 80%, increasing from 66% in 1990 to 90% in 1994 (p < 0.05) (see Fig 2). Mean aortic clamping time was minutes and mean cardiopulmonary bypass time was minutes. After cardiopulmonary bypass, blood flow in the arterial grafts should be optimal and mean systemic arterial pressure should exceed 70 mm Hg, with inotropic support, if required, to avoid the vicious circle of low blood flow in arterial grafts leading to low cardiac output leading to perioperative myocardial infarction. Intraaortic balloon pumps were used where necessary. Follow-up Early mortality was postoperative mortality occurring within the first month. Postoperative angiographic control of bypass was proposed to patients operated on from January 1990 to December 1991; it was carried out on the 15th day after operation, using selective catheterization of IMA and GEA grafts. Early postoperative functional results were evaluated on the basis of exercise stress testing with medical treatment ( months after operation); it was done on a bicycle, consisting of 2-minute stages with increments % ~ 7O 6O 66 j 90% ot i i i! Anast. Number: 3_+0 3.1_ _+0,5 3.6_ Fig 2. Bypass grafting using bilateral internal mammary and right gastroepiploic arteries: yearly rates of complete myocardial revascularization and yearly mean number of distal anastomoses (Anast.) per patient. There was a signifi'cant increase from 1990 to 1994 (p < 0.05). (* 1994 includes cases between January and June only.) Table 3. Complications After Bypass Grafting Using Bilateral Internal Mammary and Right Gastroepiploic Arteries Complication No. Percent Myocardial infarction Intraaortic balloon pump Reoperation Bleeding Sternal disjunction Mediastinitis Intestinal occlusion Respiratory failure Phrenic paralysis Stroke of 20 W. Moderate ischemic modification on electrocardiograms was defined as a less than 2-mm difference during exercise with a quick recovery at rest; a sign of severity was defined as a more than 2-ram difference during exercise or as low recovery at rest. Follow-up information was collected yearly or more frequently as necessary through direct patient contact, from responses to mailed questionnaires, or by telephone interview with surviving patients or family members. A prospective study of functional results 2 years after operation was conducted, based on exercise stress testing and thallium myocardial scintigraphy without medical treatment. In July 1994, no patient was lost to follow-up and mean postoperative follow-up was months (range, 1 to 54 months). Statistical Analysis Values of continuous variables are expressed as the mean ± the standard deviation. Comparison of means was conducted using Student's t test. Significance of differences in discrete variables was evaluated using X 2 analysis. Patient survival was calculated according to the actuarial method, including the early postoperative mortality. Values of p less than 0.05 were considered statistically significant. Results Mortality and Morbidity Early (first month) postoperative mortality was 0.4%; 1 patient with porphyria died on the sixth postoperative day of no cardiac cause during an episode of acute abdominal porphyria. Early postoperative morbidity was 8.3% (Table 3). There were 4 cases of perioperative myocardial infarction (2 lateral, 1 apical, 1 inferior). In 2 patients, intraaortic balloon pumping was initiated in the operating room, after cardiopulmonary bypass, to prevent low cardiac output and was maintained for 48 hours: these 2 patients presented with unstable angina and severe left ventricular dysfunction (left ventricular ejec-

4 958 JEGADEN ET AL Ann Thorac Surg BOTH IMAs AND GEA GRAFTING 1995;59: Table 4. Functional Results (Ischemic Modifi'cation on Electrocardiograms)" Complete Incomplete p Test Total Revascularization Revascularization Value Early 20/ /21 <0.05 test 2-Year test 17/66 7/ <0.01 a There was a significant correlation between incomplete myocardial revascularization and ischemic modifications on electrocardiograms during early (141 patients) and 2-year (66 patients) exercise stress testing. tion fraction, 0.25 and 0.32, respectively). Inotropic support for maintaining mean systemic arterial pressure greater than 70 mm Hg after cardiopulmonary bypass was used in 108 patients (45 %). Reoperation was required in 6 patients (see Table 3). ntraaortic balloon pump requirement was more frequent in patients with unstable angina (7%; p < 0.05) and severe left ventricular dysfunction (11%; p < 0.05); the rate of postoperative myocardial infarction was higher in patients with unstable angina (3.5%; not significant) and a previous operation (11%; p < 0.05). Graft Patency Of the 52 patients operated on from January 1990 to December 1991, 51 underwent early angiographic examination (15th day after operation) of the arterial grafts and anastomoses. Angiographic control of the GEA graft alone was performed in 18 patients and of the three grafts in 33 patients. The 117 examined grafts were patent, without string effect and with excellent graft adaptation: 51 GEA, 33 right IMA, 26 left IMA, and 7 sequential left IMA grafts. The 161 inspected anastomoses were patent; two GEA anastomoses were nonfunctional because of flow competition. The rate of functional and patent anastomoses was 100% for IMAs and 96% for GEAs, without a significant difference. Early Functional Results Early exercise stress testing with medical treatment was performed in 141 nonrandomized patients representative of the larger samples. The test was maximal or submaximal for 68% of patients; mean maximum exercise level was 123 ± 28 W (range, 60 to 180 W) with a mean double product of 28,466 z 7,466. Ninety-nine percent of patients were asymptomatic; moderate ischemic modifications on electrocardiograms, without signs of severity, were observable in 14% of patients and significantly correlated (p < 0.05) with incomplete myocardial revascularization (Table 4). Survival No patient was lost to follow-up. Mean postoperative follow-up was months (range, 1 to 54 months). There were three late, noncardiac deaths: two of stroke and one of suicide. Actuarial survival at 2 years and 4 years was 96.5% _+ 4.2% (Fig 3). In this series, cumulative follow-up was patient-years and rate of mortality was 0.89% per patient-year. In the 236 surviving patients, we observed no myocardial infarct, and no coronary reoperation or angioplasty was required; only 2 patients with incomplete myocardial revascularization presented, at 6 and 8 months after operation, respectively, with nondisabling recurrent angina controlled by medical treatment. In this series the rate of late cardiac events was 0.6% per patient-year. Functional Results at 2 Years Of the 73 patients with a more than 2 years' follow-up, 66 agreed to be included in a prospective evaluation of functional results at 2 years. During exercise stress testing without medical treatment, 65 patients (98%) were asymptomatic and 17 patients exhibited moderate ischemic modifications on electrocardiograms, with signs of severity in 1 case. Exercise test was maximal in 35 patients and submaximal in 31 patients; mean maximum exercise level was 136 _+ 32 W (range, 60 to 200 W) and mean double product was 31,615 ± 6,322. Ischernic modifications on electrocardiograms were significantly correlated (p < 0.01) with incomplete myocardial revascularization (see Table 4). During the same procedure, thallium myocardial scintigraphy was performed in 50 patients: the results were normal in 18 patients, 14 patients exhibited gap images due to preoperative or perioperative myocardial infarction (anterior, 4; lateral, 2; inferior, 8), and 18 patients had moderate ischemic defects on exercise, reversible after redistribution (anterior, 5; lateral, 2; inferior, 11), associated with a gap image of myocardial infarction in 4 cases (anterior, 1; lateral, 1; inferior, 2). Only 1 patient with a postoperative recurrent angina ischemic thallium defect was symptomatic. Of the 18 patients with ischemic defects on thallium scintigraphy, 11 patients had undergone early angiographic control of the GEA anastomoses and 5 patients of the IMA anastomoses with a good early patency of the anastomoses examined. Ischemic defects seen on thallium scintigraphy with exercise were correlated significantly (p < 0.01) with ischemic modifications on electrocardiograms and incomplete myocardial revascularization (Table 5). However, the ischernic defect on % loo ot o I I I I % I I I I I 1'2 1'8 2'4 3'o 3'6 4'2 4'8 Months n = Fig 3. Actuarial survival after bypass grafting using bilateral internal mammary and right gastroepiploic arteries (early deaths included). Numbers below figure indicate patients at risk during follow-up.

5 Ann Thorac Surg JEGADEN ET AL ;59: BOTH IMAs AND GEA GRAFTING Table 5. Thallium Myocardial Scintigraphy at 2 Years a Ischemic Thallium Defect Yes No Variable (n = 18) (n = 32) p Value Previous infarction NS Incomplete revascularization 11 6 (0.01 Early grafts control GEA NS IMA 5 7 NS Recurrent angina 1 0 NS ECG ischemic modification 12 4 (0.01 a There was a significant correlation between ischemic defect on thallium scintigraphy, incomplete myocardial revascularization, and ischemic modification on electrocardiograms. ECG = electrocardiographic; GEA right gastroepiploic artery; IMA = internal mammary artery,; NS not significant. thallium scintigraphy could only be explained by insufficient blood output on maximum exercise of the GEA graft in 4 patients and of the right IMA graft in 2 patients; these patients had complete myocardial revascularization, no preoperative or postoperative myocardial infarction on the same myocardial wall a good early patency of the graft concerned, and the ischemic defect seen on thallium scintigraphy was limited and reversible after redistribution. Comment For the past 20 years, our team's reference technique in the field of coronary surgery has been the association of left IMA graft, for the LAD, with sequential venous graft for the other coronary arteries. Renewed interest in bilateral IMA bypass grafting [4-7] and the use of the GEA as a third arterial graft [10-13] appeared consistent with the concept of myocardial revascularization with exclusive use of arterial grafts. The aim of this technique was to improve the long-term functional results, owing to the known excellent long-term patency of IMA grafts [1-3, 6] and midterm patency of the GEA grafts [13]. This aim could only be envisaged with the prospect of acceptable operative risks and controllable perioperative mortality and morbidity factors [16, 17], although arterial bypass grafting is technically more exacting [10, 15]. Indications were very selective in our early experience. Progressively, indications were extended to high-risk patients, and this technique was used whenever possible. Criteria for nonindication are anatomic: insufficient blood flow of the arterial grafts, less than 1.5-ram diameter of arteries, excessive heart dilatation without possibility of using in situ arterial grafts, developing gastroduodenopancreatic lesions, or history of supramesocolic abdominal operation. In addition, caution should be used in patients with cardiogenic shock or severe left ventricular dysfunction (left ventricular ejection fraction <0.35), when there is a high risk of low postoperative cardiac output and low flow in the arterial grafts. In such indi- cations, the technique combined with preventive intraaortic balloon pump use is warranted in young patients when angina is in the foreground, as in a few of our cases; usually, in these indications, we prefer to use one single IMA graft in association with a sequential venous bypass graft to avoid increasing the operative risk [18]. Considering the selection criteria of patients and intraoperative and postoperative precautions, the operative risk observed in this series proved acceptable [4, 13, 19, 20]. Postoperative mortality is low and perioperative morbidity is under control. For comparison, during the same period two arterial pedicles (both IMAs or one IMA and GEA) were used in 32% of patients and the early mortality was 0%, and one IMA associated with a sequential venous graft was used in 38% of patients and the early mortality was 1.6% (see Fig 1). One case of intestinal occlusion was the only complication observed due to the use of the GEA, which confirms the good tolerance and absence of ischemia in a stomach deprived of its right gastroepiploic blood supply [16]. In our early experience with GEA bypass grafting we had observed a few cases (5%) of GEA anastomotic stenosis or occlusion on the early GEA angiographic control, which was performed systematically in the first 100 patients who received GEA bypass grafting from January 1990 to December 1991 [21]; therefore, when the diameter of GEA was less than 2 mm or when the coronary wall was thickened, a piece of the coronary wall was resected so as to open the GEA anastomosis. The early rate of functional and patent GEA anastomoses in our series (96%) is not different from that reported in other series [11-13, 16, 22] and that obtained with IMA (100%). We observed no early difference in patency between right and left IMA grafts [6, 19]. Excellent early patency of anastomoses observed in this series can be expected not to alter with time, and to be a token of good long-term results [20, 23]. Functional results on exercise show that the aims of operative indications were reached consistently. The significant correlation between asymptomatic ischemic modifications on electrocardiograms and incomplete myocardial revascularization led us to increase the number of anastomoses and complete myocardial revascularizations with the use of left IMA as a sequential graft. The preliminary results of the 2-year prospective study of patient functional evaluation confirms the reliability of this myocardial revascularization technique [24, 25], although results are available only on patients who entered the study early. The future rates of asymptomatic ischemic defects shown by thallium scintigraphy and silent residual myocardial ischemia are expected to drop in parallel with the increased rate of complete myocardial revascularization performed in this series. No criteria in favor of grafts occlusion were observed. However, our study shows than in at least 8% of cases for GEA and 4% for right IMA, arterial graft blood flow is insufficient at maximum exercise level and causes silent myocardial ischemia detectable by thallium scintigraphy. This finding may be interpreted as a caveat: one should be cautious when using GEAs or distal right IMAs of small

6 960 JEGADEN ET AL Ann Thorac Surg BOTH IMAs AND GEA GRAFTING 1995;59: diameter [21, 26, 27]. The rate of lateral ischemic defects seen on thallium scintigraphy was low and the use of the left IMA as a sequential graft was not involved. Actuarial survival in this series proves statisfactory, comparable with that of the series reported for bilateral IMA grafts [6, 20]; in addition, survival is associated with a low rate of late cardiac events. Patient follow-up is still too short to anticipate long-term results; however, longterm survival will not be compromised by early postoperative mortality and the quality of long-term functional results theoretically will depend more on further development of coronary heart disease than on alteration of arterial bypass. The results we report for this series show that the initial goals of the technique have been reached. Myocardial revascularization based on exclusive use of both IMAs and the GEA can be achieved with minimal operative risk and offers good functional results and midterm survival rate, which justify its use. Bilateral IMA and GEA bypass grafting has become a routine technique that we apply whenever possible. However, results remain to be confirmed by long-term evaluation. References 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: Acinapura AJ, Rose DM, Jacobowitz IJ, et al. Internal mammary artery bypass grafting: influence on recurrent angina and survival in 2,100 patients. Ann Thorac Surg 1989;48: Edwards FH, Clark RE, Schwartz M. Impact of internal mammary artery conduits on operative mortality in coronary revascularization. Ann Thorac Surg 1994;57: Kouchoukos NT, Wareing TH, Murphy SF, Pelate C, Marshall WC. Risks of bilateral internal mammary artery bypass grafting. Ann Thorac Surg 1990;49: Accola KD, Jones EL, Craver JM, Weintraub WS, Guyton RA. Bilateral mammary artery grafting: avoidance of complications with extended use. Ann Thorac Surg 1993;56: 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: Galbut DL, Traad EA, Dorman MJ, et al. Coronary bypass grafting in the elderly. Single versus bilateral internal mammary artery grafts. J Thorac Cardiovasc Surg 1993;106: Van Sterkenburg SM, Ernst SM, Brutel de la Riviere A, et al. Triple sequential grafts using the internal mammary artery. An angiographic and short-term follow-up study. J Thorac Cardiovasc Surg 1992;104: Palatianos GM, Bolooki H, Horowitz MD, et al. Sequential internal mammary artery grafts for coronary artery bypass. Ann Thorac Surg 1993;56: Suma H, Fukumoto H, Takeuchi A. Coronary artery bypass grafting by utilizing in situ right gastroepiploic artery: basic study and clinical application. Ann Thorac Surg 1987;44: Mills NL, Everson CT. Right gastroepiploic artery: a third arterial conduit for coronary artery bypass. Ann Thorac Surg 1989;47: Lytle BW, Cosgrove DM, Ratliff NB, Loop FD. Coronary artery bypass grafting with the right gastroepiploic artery. J Thorac Cardiiovasc Surg 1989;97: Suma H, Wanibuchi Y, Terada Y, Fukuda S, Takayama T, Furuta S. The right gastroepiploic artery graft. Clinical and angiographic midterm results in 200 patients. J Thorac Cardiovasc Surg 1993;105: Jegaden O, Girard C, Finer G, et al. Myocardial protection with Hamburg oxygenated erystalloid cardioplegic solution for multiple coronary bypass and multivalvular replacement. J Cardiovasc Surg 1991;32: Jones EL, Lattouf O, Lutz JF, King SB. Important anatomical and physiological considerations in performance of complex mammary-coronary artery operations. Ann Thorac Surg 1987;43: Suma H, Wanibuchi Y, Furuta S, Takeuchi A. Does use of gastroepiploic artery graft increase surgical risk? J Thorac Cardiovasc Surg 1991;101: He GW, Ryan WH, Acuff TE, et al. Risk factors for operative mortality and sternal wound infection in bilateral internal mammary artery grafting. J Thorac Cardiovasc Surg 1994; 107: Jegaden O, De Gevigney G, Montagna P, et al. Late survival up to 20 years after isolated coronary bypass surgery using internal mammary artery in patients with severe left ventricular dysfunction. J Cardiovasc Surg 1994;35: Ramstr6m J, Lund O, Cadavid E, Oxelbark S, Thuren JB, Henze AC. Right internal mammary artery for myocardial revascularization: early results and indications. Ann Thorac Surg 1993;55: Galbut DL, Traad EA, Dorman MJ, et al. Seventeen-year experience with bilateral internal mammary artery grafts. Ann Thorac Surg 1990;49: Jegaden O, Montagna P, Rossi R, Revel D, Mikaeloff P. Utilisation de l'art6re gastro-6piplo'ique droite dans la ehirurgie de pontage coronaire. Exp6rience personnelle de 100 cas. Lyon Chir 1992;88: Perrault LP, Carrier IV[, Hebert Y, et al. Clinical experience with the right gastroepiploic artery in coronary artery bypass grafting. Ann Thorac Surg 1993;56: Lytle BW, Loop FD, Cosgrove DM, Ratliff NB, Easley K, Taylor PC. Long-term (5 to 12 years) serial studies of internal mammary artery and saphenous vein coronary bypass grafts. J Thorac Cardiovasc Surg 1985;89: Kusukawa J, Hirota Y, Kawamura K, et al. Efficacy of coronary artery bypass surgery with gastroepiploic artery. Assessment with thallium 201 myocardial scintigraphy. Circulation 1989;80(Suppl 1):I Isomura T, Hisatomi K, Hirano A, Hayashida N, Matsuzoe S, Ohishi K. Clinical evaluation with exercise performance in twenty patients who underwent coronary artery bypass grafting with both the gastroepiploic and internal thoracic arteries. J Thorac Cardiovasc Surg 1993;105: Nakao T, Kawaue Y. Effect of coronary revascularization with the right gastroepiploic artery. Comparative examination of angiographic findings in the early postoperative period. J Thorac Cardiovasc Surg 1993;106: Mills NL, Hockmuth DR, Everson CT, Robart CC. Right gastroepiploic artery used for coronary artery bypass grafting. Evaluation of flow characteristics and size. J Thorac Cardiovasc Surg 1993;106: