All-artery multigraft coronary artery bypass grafting with only internal thoracic arteries possible and safe: A randomized trial

All-artery multigraft coronary artery bypass grafting with only internal thoracic arteries possible and safe: a randomized trial William O. Myers, MD, Richard Berg, MS, Jefferson F. Ray, MD, John W. E. Douglas-Jones, MD, Hope S. Maki, MD, Richard H. Ulmer, MD, Bernard R. Chaitman, MD, and Richard A. Reinhart, MD, Marshfield, Wis, St Louis, Mo, and Greenville, NC

Background. The internal thoracic artery (ITA) bypass to the left anterior descending coronary artery is of proven benefit in multigraft coronary artery bypass. Total ITA grafts, if reoperation is averted by avoiding saphenous vein grafts (SVGs), are attractive. The safety of the total ITA graft operation (allITA) is a concern. Methods. A randomized trial of multiple-ITA bypass graftings with the use of bilateral sequential ITA without SVGs was performed. Control patients received 1 ITA plus SVG. Inclusion criteria were those used in the Coronary Artery Surgery Study, extended to age 76 years, and any angina class, except emergent. One hundred sixty-two patients were randomized (81 patients per group) from January 1, 1990, to December 31, 1994. Results. Baseline traits were similar as were cross-clamp times, pump times, and number of arteries bypassed (average, 4.3 arteries). Patients who received multiple ITA grafts had no myocardial infarctions, per reference laboratory. One patient died, and 2 patients returned for bleeding. The ITA-SVG group had similar results. The all-ITA group experienced successful completion in 93% of cases. Complications did not differ from control patients. Conclusions. Early and 5-year outcomes were not different between the all-ITA group and the ITA with SVGs group. We believe experienced surgeons can safely extend the ITA to multibypass coronary artery bypass without use of SVG to achieve an all-ITA operation. (Surgery 2000;128:650-9.) From the Departments of Cardiovascular and Thoracic Surgery and Cardiology, Marshfield Clinic, and the Department of Clinical Research, Marshfield Medical Research and Education Foundation, Marshfield, Wis; the Division of Cardiology, St Louis University Medical Center, St Louis, Mo; and the Brody School of Medicine, East Carolina University, Greenville, NC

THE INTERNAL THORACIC ARTERY (ITA) used in coronary bypass grafting is a superior conduit.1,2 It is resistant to atherosclerosis and, when used as a bypass to the left anterior descending (LAD) coronary artery, offers improved late survival3,4 and reduction in reoperation rates.4,5 However, Presented at the 57th Annual Meeting of the Central Surgical Association, Chicago, Ill, March 2-4, 2000. Supported in part by Marshfield Clinic Physician Research Funds, Marshfield Medical Research Foundation Disease Specific Restricted Funds, and the Gwen D. Sebold Award. Reprint requests: William O. Myers, MD, Department of Cardiovascular and Thoracic Surgery, Marshfield Clinic, 1000 North Oak Ave, Marshfield, WI 54449. Copyright © 2000 by Mosby, Inc. 0039-6060/2000/$12.00 + 0 11/6/108113 doi:10.1067/msy.2000.108113

650 SURGERY

patients with left side ITA (LITA) to the LAD coronary artery (LITA-LAD) do come back for repeat operations because of the failure of vein grafts to other branches of the coronary system and present a technical challenge. It becomes attractive to perform multiple bypass coronary artery bypass grafting (CABG) with only the 2 ITA grafts (all-ITA). Pioneers in this concept are Sauvage et al6 and Green et al.7 Our study was undertaken to bring the power of randomization to bear on the question of comparative safety and early results, with a comparison of an all-ITA bypass multigraft operation and the “standard” operation with the LITALAD plus saphenous vein grafts (SVGs). Data from the National Heart, Lung, and Blood Institute Coronary Artery Surgery Study (CASS) showed that the LITA used to supply the LAD

Surgery Volume 128, Number 4

along with vein grafts to other targets reduced the risk of death in the first 5 years after surgery by 35%.4 We did not know whether the use of the ITA grafts for other bypasses increased or decreased the protection offered by the single ITA bypass. However, such an “extended ITA operation” offered the possibility that other ITA bypasses also might last longer than vein grafts. We proposed taking the ITA grafts as semiskeletonized arteries to have a long enough graft to reach all distal vessels. Would this treatment reduce the durability of the LITA-LAD? How about sequential grafts? Would the smaller superior epigastric and musculophrenic terminal ITA branches used to reach the most distal coronary arteries reliably supply adequate early blood flow to avoid perioperative myocardial infarctions, postoperative rhythm disturbance, or slow recovery? PATIENTS AND METHODS Patients were recruited between January 1, 1990, and December 31, 1994, a 5-year intake period for the study. For the randomizing surgeons, patients were seen by a Marshfield Medical Research Foundation research assistant to present the study, verify eligibility, and obtain informed consent. The Department of Biostatistics maintained the randomization scheme, and the assignment was obtained by telephone by the surgeon in the operating room before the start of the operation. Rarely, did a patient who signed up for randomization change his or her mind—nor the referring doctor—and specify which operation was wanted. The decision was honored, but if randomization had already been done, the patient was kept in the group assigned for analysis. These few patients were carried as “crossovers” with analysis by treatment assigned, not treatment received. The surgical assistant recorded details of the operation in our standard database forms. A supplementary form was used to answer the question of whether the assigned operation had been performed and whether totally or partially. Exclusions. Patients who qualified for this study were aged 76 years or younger, were referred for nonemergency primary isolated coronary operations that required 2 or more bypass grafts, had 70% or less left side main stenosis and 35% or greater ejection fraction, and were without lower extremity peripheral vascular disease. Excluded were patients in whom the saphenous veins had been used for another operation and those patients with lower extremity amputation. Additional exclusions were patients with insulin-dependent diabetes and those

Myers et al 651

patients with body weight more than 30% over the Metropolitan Life Insurance tables. Patients who required concurrent cardiac or noncardiac operations were also excluded. Patients with renal failure and CABG in preparation for another operation were excluded. If ventricular aneurysm or concomitant carotid surgical procedures were planned or if another disease made survival of more than 5 years doubtful, those patients were excluded. Patients who had communication or comprehension problems, patients who refused blood transfusions, and patients who lived at a distance that would make follow-up difficult were excluded. Baseline data were obtained and recorded on a standard dataform by the nurse clinician or the catheterization laboratory technologist. These data were entered into an Oracle database (Oracle Corp, Redwood City, Calif) maintained by the Cardiology Department. Operative details were also recorded in the cardiology database and in the cardiac surgical database maintained in our cardiac surgery office. From discharge, death, and postdischarge follow-up, possible postoperative complications were monitored.8 Baseline data are given in Table I. Laboratory data, including postoperative cardiac enzymes, were obtained by chart abstraction by the research coordinator. These were maintained in an SPSS database (SPSS Inc, Chicago, Ill) at the Research Foundation. Electrocardiographic data. The protocol required that each patient have a 12-lead electrocardiogram before the procedure and before hospital discharge. All electrocardiograms were interpreted at the St Louis University Central Electrocardiogram and Myocardial Infarction Classification Laboratory. Each electrocardiogram was coded independently by trained central laboratory staff blind to the patient’s clinical history and treatment assignment. Serial comparison of sequential tracings was performed with a modified Novacode system to identify patients with new electrocardiogram changes in the Minnesota code.9-11 Postprocedure myocardial infarction was defined by the presence of a new 2-grade (Minnesota code) worsening in the Q wave on the postprocedure versus the preprocedure electrocardiogram. Electrocardiogram data are presented in Table II. The patients were followed at 30 days and then every 6 months, until the last of the randomized patients had reached 5 years after the procedure. Research assistants in the Clinical Research Support Unit handled follow-up calls or survey letters. Data were reviewed at 2 and 5 years, and statistical comparisons made.

652 Myers et al

Surgery October 2000

Table I. Baseline data All-ITA group Variable Mean age (y) Men Women Mean body mass index (kg/m2) Mean diastolic blood pressure (mm Hg) Mean systolic blood pressure (mm Hg) Mean pulse rate (beats/min) Tobacco use Never Former Current History of Diabetes Hypercholesterolemia Hypertension Family history of coronary artery disease before age 60 y Employment status Retired Fully employed Part-time Disabled Cardiac Noncardiac Unemployed, other Unknown Use of cardiac medications Beta blockers Calcium channel blockers Diuretics Digitalis Antiarrhythmics Nitrates Antiplatelets/anticoagulants Hypolipidemics Unstable angina Progressive angina Subacute myocardial infarction (≥ 4 d) Congestive heart failure Positive thallium Positive exercise Mean no. of arterial segments with ≥ 70% stenosis 50%-69% stenosis Mean left ventricular Score Ejection fraction Mitral regurgitation None Slight Moderate Severe Mitral valve prolapse (%) *t

n 81† 63 18 79‡ 73‡ 73‡ 71‡ 77‡ 24 31 22 81‡ 11 55 46 65‡ 31 81‡ 38 34 3 0 0 1 5 81‡ 30 41 13 3 2 51 53 17 14 22 13 4 13 26

62.6 61.5 64.2 29.0 78.3 136.2 71.2 31.2% 40.3% 28.6% 13.6% 67.9% 56.8% 47.7% 46.9% 42.0% 3.7% 0.0% 0.0% 1.2% 6.2%

n 81‡ 61 20 79‡ 78‡ 78‡ 76‡ 81‡ 24 31 26 81‡ 12 44 38 71‡ 37 81‡ 33 35 6

Average/%* 62.8 61.9 63.6 29.2 77.4 137.1 70.9

P value .86 .78 .85 .89 .63 .83 .88

29.6% 38.5% 32.1%

.89

14.8% 54.3% 46.9%

.82 .08 .21

52.1%

.61

40.7% 43.2% 7.4%

37.0% 50.6% 16.0% 3.7% 2.5% 63.0% 65.4% 21.0% 17.3% 27.2% 16.0% 4.9% 16.0% 32.1%

0 1 1 5 81‡ 32 31 10 3 3 37 49 10 9 19 12 2 14 34

39.5% 38.3% 12.3% 3.7% 3.7% 45.7% 60.5% 12.3% 11.1% 23.5% 14.8% 2.5% 17.3% 42.0%

.75 .11 .50 1.00 1.00 .03 .52 .14 .26 .59 .83 .68 .83 .19

73‡ 73‡

3.6 1.4

77‡ 77‡

3.4 1.4

.57 .96

68‡ 75‡

7.0 60.8%

77‡ 81‡

7.4 62.6%

.41 .48

51 11 5 0 68‡

76.1% 16.4% 7.5% 0.0% 4.4%

60 14 2 1 77‡

77.9% 18.2% 2.6% 1.3% 6.5%

test for averages; Fisher exact test for percents. of patients available in the category; less than 81 represents missing data.

†Number

Average/%*

Standard group

0.0% 1.2% 1.2% 6.2%

.45

.72

Myers et al 653

Surgery Volume 128, Number 4 Table II. Electrocardiogram results All-ITA group Variable Baseline Q-wave codes† Definite myocardial infarction Possible myocardial infarction ST-depression codes Any code T-wave codes Any code Postoperative Q-wave codes† 2-Step worsening 1-Step worsening ST depression codes New or worse codes T-wave codes New or worse codes

n 80‡ 10 9 73‡ 14 73‡ 30 80‡ 0 5 67‡ 9 67‡ 23

Standard group

% 12.5 11.3 19.2 41.1 0.0 6.3 13.4 34.3

n 81‡ 11 5 73‡ 22 73‡ 37 80‡ 3 2 66‡ 9 66‡ 30

%

P value*

13.6 6.2

1.00 .54§

30.1

.12

50.7

.24

3.8 2.5

.25 1.00¶

13.6

.97

45.5

.19

*Fisher’s

exact test. †St Louis laboratory. ‡Number of patients available in category; less than 80 represents an electrocardiogram, ST, or T that could not be interpreted because of conduction disturbance. §Probability value for definite and possible myocardial infarction combined. A 2-step change is considered a new definite myocardial infarction; a 1-step change is considered a new possible myocardial infarction. ¶Probability value for 1- and 2-step changes combined.

Blood products and other resources. Data were obtained from the blood bank, and financial data were obtained from Marshfield Clinic and St Joseph’s Hospital in Marshfield. Surgical technique. For the standard operation, saphenous veins were obtained by open surgical technique. Gentle handling was stressed. The vein was tested and distended with buffered heparinized saline solution with a Bonchek pressure-limiting balloon (Cobe Cardiovascular, Arvada, Colo). Standard median sternotomy was used. The left sternal one half was elevated with a retractor (Rultract; Cleveland, Ohio). The ITA pedicle was taken down with a strip of fascia and muscle for the standard operation, and generally the veins remained attached proximally. The pedicle had some bulk, which we thought might be an advantage for the control patients that the study patients would not have. Nevertheless, we felt it was important that, because they were to have a “standard operation,” this method be followed as part of their operations. The harvest of the ITA grafts was the main difference between the 2 operative procedures. Although there was some variation among surgeons, in the case of the all-ITA, the internal thoracic fascia was opened on the medial side of the vessels along the edge of the sternum, with no incision lateral to the vessels. The fascia and transversus thoracic muscles were retracted to expose the vessels, forming a shelf from which the artery and accompanying veins

could be gently dissected together. The branches were ligated with Hemoclips (PillingWeck Surgical, Research Triangle Park, NC) as distally as possible from the ITA. The terminal branches beyond the bifurcation were taken as far as necessary to do the proposed operation. The superior end of the pedicle was dissected up to the subclavian vein and beyond, up near the subclavian artery itself, although we usually did not see this artery. We almost always saw the phrenic nerve. The internal thoracic vein was clipped and cut at its origin. The semiskeletonized artery was then brought down alongside the phrenic nerve, entering the mediastinum high, with the pericardium reflected laterally as described by Sauvage.12 This placement of the LITA brings it quite deep into the mediastinum and away from injury during future reoperation. The artery comes just anterior to the left superior pulmonary vein. It can easily be brought to the marginal branches of the circumflex coronary artery and onto the inferior surface of the heart where the 2 terminal branches can be used to supply a posterior descending and a posterolateral coronary artery. The right side ITA (RITA) is dissected in the same manner. The course of the artery is somewhat different, depending on the planned use of this vessel in the operation (Table III). Fifteen minutes are often required for the preparation of the route of passage for each ITA into the pericardium and ensuring hemostasis. The dissection on the left side is carried

654 Myers et al

Surgery October 2000

Table III. LITA and RITA graft patterns used in the all-ITA group Graft patterns

No. of patients

ITA graft patterns LITA with 2-4 distals (average, 2.75) BITA with 2 single distals BITA with 1 single RITA distal and LITA sequence (2 or 3 distals), both subclavian based BITA with both ITA grafts as sequences (4-6 distals) both subclavian based (RITA crossing great vessels) BITA, RITA Y graft (3-5 distals; average, 4.25) RITA grafts Distal RCA system (PDA, PL, AM, RCA) Anterior descending system (LAD, diagonals, septal) Crossing anterior to aorta Back of aorta and PA (via transverse sinus) Anterior to SVC Retrocaval As Y graft based on LITA LITA grafts Anterior system (LAD, diagonals, septal) Circumflex and right (obtuse marginals, PL, PDA) Anterior system and origin of RITA Y

8 7 13 46 5 17 22 27 11 16 5 22 53 5

BITA, Bilateral ITA; RCA, right side coronary artery; PDA, posterior descending artery; PL, posterior lateral coronary artery; AM, acute marginal; PA, pulmonary artery.

down to the transverse aorta and the pulmonary artery so that the LITA comes straight down the mediastinum. On the right side, the same considerations hold. The pericardium is reflected laterally. The ITA lies posterior to the subclavian vein, entering the mediastinum deep to the vein. It can enter back of the superior vena cava and back of the aorta and go through the transverse sinus to either the anterior descending system or to the circumflex. Most of the time the RITA was taken to the anterior descending with the terminal branches to the diagonal and mid or distal LAD. If the RITA is to supply the right side coronary system, it is brought down the trough lateral to the vena cava and onto the inferior surface of the heart. It reaches nicely to the inferior wall arteries. All other non-ITA harvest surgical techniques were the same in both groups. Some technical modifications were used throughout with the aim of maintaining blood supply to the sternum. The pleura was left intact. The strap muscles were not severed. The mediastinum was drained with 2 flat drains (Axiom 7; Axiom Medical, Compton, Calif) in both groups of patients. Operating loupes (×4.5) and 8-0 polypropylene were used for the ITA anastomoses and ×3.5 or ×4.5 and 7-0 polypropylene for vein grafts. RESULTS Patient characteristics. Randomization gave us 2 equal groups of patients (Table I). Patient age ranged from 36 to 76 years (mean, 62.7 years), with

women approximately 3 years older than men. Women were 22% of the trial in both groups. The same proportions of patients who had never smoked (approximately 30%), patients who had formerly smoked (approximately 40%), and patients who currently smoked (30%) were present in study patients and control patients. Diabetes (non-insulin dependent), hypercholesterolemia, and hypertension did not differ significantly between groups. The use of cardiac medications was not significantly different, except for nitrate use of 63% of the patients in the all-ITA group versus 46% in the control group (P = .03). Fairly severe coronary artery disease was present in both groups, averaging more than 3 stenoses of 70% or greater. Q-wave evidence for myocardial infarction before the operative procedure was present in 23.8% of the ITA group and 19.8% of the control group (Table II). Preoperative ST segment depression was slightly less frequent in the ITA group (19.2%) than the control group (30.1%). The difference was not statistically significant. T-wave abnormalities were frequent, occurring in 41% to 51% of the 2 study groups.13 Distribution of ITA grafts. Eighty patients in both groups had LITA grafts used in their bypass grafting. One patient in each group did not have this graft; and, according to the operative reports, the LITA was thought to be inadequate for use. It should be noted that the minimum requirement for the 2 arms of the study was at least 2 coronary

Myers et al 655

Surgery Volume 128, Number 4 Table IV. Operative data All-ITA group Variable

n

Assigned or intended operation totally accomplished 75/81 Assigned or intended operation partially accomplished 6 Crossovers 3 Bypasses 4.33 ± 1.29 ITA bypasses 80 Average ITA bypass per patient with ITA 4.1 ± 1.4 Patients with > 1 ITA bypass 79 Pure multiple ITA 72 ITA bypasses 0 1 1 1 2 11 3 15 4 17 5 22 6 13 7 1 Operation time 77 Anesthesia time 76 Pump time (min) 81 Cross-clamp time (min) 81 Packed cells (units) 81 Fresh frozen plasma 81 Platelets 81 Albumin 81 Hespan 81 CK-MB at 0 h (ng/mL) 80 CK-MB at 16 h (ng/mL) 80 AST level (units/mL) 81 Low myocardial temperature (°C) 76 High myocardial temperature (°C) 76

Standard group

%/Average*

n

%/Average*

92.6% 7.4%

98.8% 1.3%

97.5% 88.9%

79/80 1 1 4.27 ± 1.26 80 1.6 ± 0.65 43 2

1.2% 1.2% 13.6% 18.5% 21.0% 27.2% 16.1% 1.2% 5 h 38 min 6 h 58 min 135.9 101.1 2.7 1.5 1.17 4.4 0.11 18.2 17.6 38.2 10.8 19.2

1 37 37 6 0 0 0 0 78 80 81 81 81 81 81 81 81 80 80 81 79 79

1.2% 45.7% 45.7% 7.4% 0.0% 0.0% 0.0% 0.0% 4 h 44 min 5 h 57 min 133.5 99.9 2.6 1.6 0.85 4.0 0.12 21.2 19.7 36.4 10.1 18.3

98.8%

P value

.758 98.8% 53.1% 2.5%

<.001 <.001 .64 .79 .77 .88 .51 .27 .88 .10 .45 .49 .15 .38

CK-MB, Creatin kinase muscle band; AST, aspartate aminotransferase. *t test for averages; Fisher exact test for percentages.

artery segments bypassed with the ITA grafts for the study group and at least 1 vein graft and 1 ITA for the control group. Multiple coronary bypass was defined as 2 or more coronary segments bypassed. Table III shows the graft patterns used for the ITA in the all-ITA patient group. Eight patients had only the LITA used for their “multiple” bypass operation. This was usually to the diagonal or other high left side artery and the LAD. For 49 patients, the RITA was used to the anterior system. It crossed anterior to the aorta in 22 patients; but because we were concerned about the possible injury to the RITA at any future reoperation, we increasingly took it back of the aorta and pulmonary artery through the transverse sinus. We found the retrocaval position useful, with the RITA in its natural plane back of the venous system.

Operative data comparison. The assigned operation was totally accomplished in 92.6% of patients in the all-ITA group (75 cases; Table IV). In 6 patients, the assigned operation was partially accomplished (ie, a supplemental vein graft was needed somewhere in the system). Operation time and anesthesia time were significantly longer on average for the all-ITA group. However, pump time and cross-clamp time were similar. Blood bank usage was similar in the 2 groups. The average number of bypass graftings (both venous and arterial) in the 2 groups was equal at 4.3. The average number of ITA bypass grafts was 4.1 for the patients in the all-ITA group and 1.6 for the patients in the standard operation group. In the allITA group 97.5% of the patients had more than 1 ITA bypass graft, and in the standard operation group, 53.1% had more than 1 ITA bypass (account-

656 Myers et al

Surgery October 2000

Table V. Costs Variable Charges ($) Hospital anesthesia Hospital surgery Total hospital Sum of clinic and hospital charges (admission–30 d after discharge) *Mean

All-ITA group*

Standard group*

P value†

768 ± 184 10,086 ± 1,547 29,060 ± 6,006 46,671 ± 13,987

790 ± 398 9,460 ± 2,081 27,965 ± 6,230 43,455 ± 12,646

.558 <.001 .120 .056

± SD. rank sum test.

†Wilcoxon

Table VI. Postoperative complications Variable Conditions leaving the OR Stable OR complications Length of stay (mean ± SD) Ventilator > 48 h Exploration for bleeding Enzyme myocardial infarction‡ Pneumothorax Atrial fibrillation Thoracentesis Stroke Hospital deaths (same admission) Postdischarge events (30 d) Leg wound treatment Sternal wound Hospital (elsewhere) Hospital (Marshfield) Recurrent angina Death (30 d)

All-ITA group (n;%)

Standard group (n;%)

P value*

81 (100) 3 (3.7)† 7.9 ± 2.13 1 (1.23) 2 (2.47) 2 (2.47) 4 (4.94) 7 (8.64) 0 (0.00) 1 (1.23) 0 (0.00)

81 (100) 0 (0.0) 7.68 ± 2.24 0 (0.00) 4 (4.94) 2 (2.47) 1 (1.23) 8 (9.88) 2 (2.47) 1 (1.23)§ 1 (1.23)§

— .25 .65 — — 1.00 .366 1.00 — — —

2 5 1 3 0 1

(2.50) (6.17) (1.23) (3.70) (0.00) (1.23)

1 0 2 3 1 1

(1.23) (0.00) (2.47) (3.70) (1.23) (1.23)

— .06 — — — —

OR, Operating room. *Fisher exact test for percent; t-test for length of stay. †Operating room complications included: coagulopathy (distal anastomosis repair required), other (back on pump); cardiogenic shock stabilized with intraaortic balloon pump; injury to ITA that required vein graft. ‡CK-MB 6% or 40 ng/mL at 8 or 16 hours. §Same patient. Hospitalized for any reason.

ed for by the fact that, by the time we started the study, the sequential LITA to diagonal and LAD had become fairly standard). Surgical costs were significantly higher in the all-ITA group (Table V). Operative death and morbidity comparisons. All patients in the 2 groups were stable on leaving the operating room. An operating room complication was recorded in 3 of the all-ITA group, including 1 injury to an ITA that required vein graft substitution, with none in the standard operation group. There was 1 stroke and 1 early postdischarge death in each group (Table VI). Thirty-day postdischarge events. There were 5 sternal wound complications in the all-ITA group, all of which were superficial and responded to outpatient treatment (Table VI).

Two-year and 5-year death and morbidity comparisons. Patients have currently been followed for between 54 and 114 months (median, 90 months), with 161 of 162 patients followed at least 60 months (or dead). At 5 years, only 8 deaths had occurred (all-ITA group, 5 deaths; control group, 3 deaths), with no significant difference in overall survival rates (P = .39). Relatively few cardiac events have been observed at 5 years, with 1 myocardial infarction in each group and a total of 5 patients with percutaneous transluminal coronary angioplasty (PTCA) or repeat bypass (all-ITA group, 4 patients; standard group, 1 patient). There is no significant difference in cardiac event-free survival rates (P = .82; Table VII). Figs 1 and 2 are plots of overall and event-free survival, respectively.

Myers et al 657

Surgery Volume 128, Number 4

Fig 1. Overall survival represents all-cause deaths, including surgical and 30-day deaths (Kaplan-Meier estimates). The probability value was ascertained by the log rank test. Minimums are per group. Two-year and 5-year survival with 95% confidence intervals is given in Table VII.

Fig 2. Event-free survival represents freedom from cardiac death, PTCA, or repeat CABG and hospitalization for myocardial infarction, congestive heart failure, or angina (Kaplan-Meier estimates). The probability value was ascertained by the log rank test. Minimums are per group. Twoyear and 5-year event-free survival with 95% confidence intervals is given in Table VII.

Table VII. Twenty-four and 60-month outcomes Variable At 24 mo. Patients with follow-up Deaths* (n) Survival† (%) Hospitalized since operation‡ (n; %) Myocardial infarction since operation (n; %) PTCA or repeat bypass (n; %) Event-free§ survival† (%) Employed (n; %) At 60 mo. Patients with follow-up Deaths* (n) Survival†(%) Hospitalized since operation‡ (n; %) Myocardial infarction since operation (n; %) PTCA or repeat bypass (n; %) Event-free§ survival† (%) Employed (n; %)

P value

ITA group

Standard group

81 3 96.3 (89.7-98.7) 4 (4.9) 0 (0.0) 1 (1.2) 93.8 (86.1-97.3) 30 (37.0)

81 2 97.5 (91.4-99.3) 6 (7.4) 0 (0.0) 0 (0.0) 92.4 (84.4-96.5) 27 (33.3)

1 1 .74

80 5 93.8 (88.4-97.3) 8 (10.0) 1 (1.2) 4 (5.0) 88.6 (81.4-93.9) 22 (27.5)

81 3 96.3 (92.0-98.7) 11 (13.6) 1 (1.2) 1 (1.2) 86.1 (78.1-92.1) 21 (25.9)

.39 .63 1 .21 .82 .86

1 .75

*Includes surgical and 30 day-deaths. †Product-limit survival estimates and 95% confidence limits. ‡Hospitalized for angina, congestive heart failure, myocardial infarction, percutaneous transluminal coronary angioplasty, or repeat bypass grafting. §Combined event includes cardiac death and hospitalization for angina, congestive heart failure, myocardial infarction, percutaneous transluminal coronary angioplasty, or repeat bypass grafting. Log rank test for equal hazard rates with the use of all available follow-up data.

658 Myers et al

DISCUSSION As it became clear that atherosclerosis in vein grafts was a limiting factor to graft durability, more interest developed in the use of the ITA grafts, either sequential or bilateral grafts to multiple target vessels. Jones14 and Tector et al,15 among others, contributed papers supporting more than 1 bypass with the ITA graft. Other authors weighed in with the same concept, and it appeared that a new operation was soon to come on the scene and that an opportunity existed to study a new surgical operation before it became widespread. We each did more and more cases with that concept in mind, extending the use of the ITA grafts as we became comfortable with the technical aspects and safety. Beginning in January of 1989 and continuing for 6 months, 1 of us (W.O.M.) adopted a policy of an all-ITA bypass, with SVGs only as necessary on the basis of the report by Sauvage.6,12 A randomized trial with a registry was approved by our Institutional Review Board. In the meantime, the all-ITA operation with bilateral ITA grafts did not become widely accepted, and other artery bypass grafts were introduced to supplement the use of the LITA-LAD bypass graft.16-22 Buxton et al23 reported improved survival with bilateral ITA grafts compared with single ITA graft at 10 years. A recent report from the Cleveland Clinic database covering 10 or more years by Lytle et al5 used extensive statistical adjustments to retrospectively compare bilateral ITA with single ITA operations. The authors found greater survival, freedom from reoperation, and freedom from PTCA, all of which favored the patients with the bilateral ITA grafts. Many of these patients actually had an extended ITA operation similar to this report with multiple bypasses with no or few SVGs. A previous study by Cosgrove et al24 looked at comparative surgical risk, with findings similar to ours. Hirotani et al25 reported the safety of bilateral ITA grafts in patients with diabetes. The present study results suggest that an all-ITA bypass operation is possible with the use of just the ITA grafts, usually without the need for supplementation with the radial artery, the gastroepiploic artery, or SVGs. Early results of our randomized trial show equal safety and equal early outcome. We think there is a place for this operation that seems to be coming into its own, and we feel that our data support its continued use and broader application. We thank Nancy Rizner and Carla Finck for data management, Jane Carl for follow-up interviews, and Alice Stargardt for manuscript preparation.

Surgery October 2000 REFERENCES 1. Grondin CM, Campeau L, Lesperance J, Enjalbert M, Bourassa MG. Comparison of late changes in internal mammary artery and saphenous vein grafts in two consecutive series of patients 10 years after operation. Circulation 1984;70:I208-12. 2. Mills NL, Everson CT. Vein graft failure. Curr Opin Cardiol 1995;10:562-8. 3. Cameron A, Davis KB, Green G, Schaff HV. Coronary bypass surgery with internal-thoracic-artery grafts–effects on survival over a 15-year period. N Engl J Med 1996;334:216-9. 4. Cameron A, Davis KB, Green GE, Myers WO, Pettinger M. Clinical implications of internal mammary artery bypass grafts: the Coronary Artery Surgery Study experience. Circulation 1988;77:815-9. 5. Lytle BW, Blackstone EH, Loop FD, Houghtaling PL, Arnold JH, Akhrass R, et al. Two internal thoracic artery grafts are better than one. J Thorac Cardiovasc Surg 1999;117:855-72. 6. Sauvage LR, Wu H-D, Kowalsky TE, Davis CC, Smith JC, Rittenhouse EA, et al. Healing basis and surgical techniques for complete revascularization of the left ventricle using only the internal mammary arteries. Ann Thorac Surg 1986;42:449-65. 7. Green GE, Sosa JA, Cameron A. Prospective study of feasibility of routine use of multiple internal mammary artery anastomoses. J Cardiovasc Surg 1989;30:643-7. 8. Myers WO, Ray JF III. The use of single and multiple internal mammary artery bypass in reoperative coronary bypass surgery: the Marshfield experience. In: Myers WO, guest editor. Cardiac surgery: state of the art reviews. Philadelphia: Hanley & Belfus Inc; 1992. p. 375-95. 9. Prineas RJ, Crow RS, Blackburn H. The Minnesota code manual of electrocardiographic findings: standards and procedures for measurement and classification. Boston (MA): PSG Inc; 1982. p. 1-229. 10. Rautaharju PM, Calhoun HP, Chaitman BR. Novacode serial ECG classification system for clinical trials in epidemiologic studies. J Electrocardiol 1992;24:179-87. 11. Chaitman BR, Zhou SH, Tamesis B, Rosen A, Terry AB, Zumbehl KM, et al. Methodology of serial ECG classification using an adaptation of the Novacode for Q wave myocardial infarction in the Bypass Angioplasty Revascularization Investigation (BARI). J Electrocardiol 1996;29:265-77. 12. Sauvage LR. Extensive myocardial revascularization using only internal thoracic arteries for grafting the anterior descending, circumflex, and right systems. In: Myers WO, guest editor. Cardiac surgery: state of the art reviews. Philadelphia: Hanley & Belfus Inc; 1992. p. 397-419. 13. Chaitman BR, Alderman EL, Sheffield LT, Tong T, Fisher L, Mock MB, et al. Use of survival analysis to determine the clinical significance of new Q waves after coronary bypass surgery. Circulation 1983;67:302-9. 14. Jones EL. Extended use of the internal mammary-coronary artery bypass. J Card Surg 1986;1:13-21. 15. Tector AJ, Amundsen S, Schmahl TM, Kress DC, Peter M. Total revascularization with T grafts. Ann Thorac Surg 1994;57:33-9. 16. Pym J, Brown PM, Charrette EJP, Parker JO, West RO. Gastroepiploic-coronary anastomosis. J Thorac Cardiovasc Surg 1987;94:256-9. 17. Mills NL, Everson CT. Right gastroepiploic artery: a third arterial conduit for coronary artery bypass. Ann Thorac Surg 1989;47:706-11. 18. Carpentier A, Guermonprez JL, Deloche A, Frechette C,

Surgery Volume 128, Number 4

19. 20.

21.

22.

23.

24.

25.

DuBost C. The aorta-to-coronary radial artery bypass graft. Ann Thorac Surg 1973;16:111-21. Foster ED, Kranc MAT. Alternative conduits for aortocoronary bypass grafting. Circulation 1989;79:I34-9. Borger MA, Cohen G, Buth KJ, Rao V, Bozinovski J, Liaghati-Nasseri N, et al. Multiple arterial grafts: radial versus right internal thoracic arteries. Circulation 1998;98:II714. Acar C, Jebara VA, Portoghese M, Beyssen B, Pagny JY, Grare P, et al. Revival of the radial artery for coronary artery bypass grafting. Ann Thorac Surg 1992;54:652-60. Pym J, Luffman B, Parry M. Total arterial revascularization of the heart: intentional or inevitable. AACN Clin Issues 1997;8:9-19. Buxton BF, Komeda M, Fuller JA, Gordon I. Bilateral internal thoracic artery grafting may improve outcome of coronary artery surgery: risk-adjusted survival. Circulation 1998;98:II1-6. Cosgrove DM, Lytle BW, Loop FD, Taylor PC, Stewart RW, Gill CC, et al. Does bilateral internal mammary artery grafting increase surgical risk? J Thorac Cardiovasc Surg 1988;95:850-6. Hirotani T, Kameda T, Kumamoto T, Shirota S, Yamano M. Effects of coronary artery bypass grafting using internal mammary arteries for diabetic patients. J Am Coll Cardiol 1999;34:532-8.

DISCUSSION Dr Brian L. Cmolik (Cleveland, Ohio). What we do know from collecting data in the Society of Thoracic Surgeons database over a period of many hundreds of thousands of patients over many years is that the one predictor of long-term survival in patients after the coronary bypass is a patent ITA to the LAD. What remains unclear, but what we suspect, is that additional arterial grafts may confer additional survival and symptom-free benefits to our patients. Dr Myers is one of the original surgeons to move to all arterial grafts. You can see that this technically demanding operation can be performed safely and effectively with the use of both internal thoracic arteries and not a reliance on any additional arterial conduits (eg, radial artery, gastroepiploic artery). Many cardiac surgeons feel there is definitely an advantage to using pediculized arterial conduits. Dr Myers has demonstrated that pediculized conduits can be harvested within the chest, avoiding incisions in either the leg or the arm, and used effectively to graft the anterior, the posterior, and lateral surfaces of the heart. First, we know that all of these arteries are not created equally, and you are asking these ITAs to supply multiple run-off beds in these patients. Did you do anything to measure flow?

Myers et al 659

Second, was there any pharmacologic treatment used for these arteries, either calcium channel blockers or nitrates, to ensure or preserve graft patency? Third, more and more of us are moving away from the use of the cardiopulmonary bypass machine to achieve revascularization. Currently 80% of our group’s cases are accomplished without the use of the cardiopulmonary bypass machine. Do you think the extended use of the ITA will be feasible in an off-pump situation? Dr William H. Baker (Maywood, Ill). Did you perform any studies to try and demonstrate how many of your anastomoses were still patent over time? I know the patients did well. But they might have done well with only 1 or 2 of the anastomosis being patent instead of 3. Dr Myers. Regarding the question of do we measure flows: we did not measure flows. We have in the past. We started out years ago trying to measure flows with the electromagnetic flow probe. We could never really quite make it work. So we tried other methods such as dripping blood from the artery into the medicine cup to see how much; we did that for 1 or 2 years. But by the time we started this study, we were really confident of the practicality of using the ITA, and we believed we could certainly revascularize the entire heart. So no patients in this study had any flow measurement in either their vein grafts or their arterial bypasses. We did use dilute papaverine to dilate the ITAs. We used a 2-mm DLP (Medtronic, Minneapolis, Minn) white plastic cannula in 1 of the terminal branches and a Bonchek (Cobe Laboratories) pressure-limiting balloon and distended the arteries. That helped us to find leaking from the branches on the side that we had not clipped. We found that useful. We did not use any systemic pharmacologic agents other than nitroglycerin. Concerning off-pump CABG, we currently are doing some off-pump cases, but not with multiple internal thoracic grafts. Off-pump patients are getting 1 graft, so of course they are quite highly selected. I do not know whether multiple grafting with ITA grafts can be done. We have not tried it. Dr Baker, that is a really good question. One half of these could be occluded, and we would not know it, right? I think that certainly happened in the saphenous vein graft days. We do not know the answer to that. We have not had patients come back for routine restudies. I do not think there is any way to do that nowadays. We did that in the early days of the coronary artery surgery study. We had 10% of the patients come back, and we restudied them. But I think that time is gone.