R standard procedure, accounting for 4% to 6% of all. Bilateral Internal Mammary Artery Grafts in Reoperative and Primary Coronary Bypass Surgery

Transcription

1 Bilateral Internal Mammary Artery Grafts in Reoperative and Primary Coronary Bypass Surgery David L. Galbut, MD, Ernest A. Traad, MD, Malcolm J. Dorman, MD, Paul L. DeWitt, MD, Parry B. Larsen, MD, Paul A. Kurlansky, MD, Jaclyn H. Button, MS, Joan M. Ally, ARNP, and Thomas 0. Gentsch, MD Cardiovascular and Thoracic Surgical Associates, Miami Heart Institute, Miami, Florida Bilateral internal mammary artery grafting is recognized as a preferred method of myocardial revascularization. However, its efficacy in coronary bypass reoperation has not been clearly established. From January 1982 through June 1989, 88 patients underwent coronary bypass reoperation with bilateral internal mammary artery grafts. Results were compared with those for a subset of 88 patients receiving primary revascularization with bilateral internal mammary artery grafts who were computer matched for sex, age, left ventricular function, anginal classification, and left main coronary artery disease. In each group, 62.5% (55 patients) had unstable angina, 43.2% (38 patients) had reduced ejection fraction, and 21.6% (19 patients) in the reoperation group and 20.5% (18 patients) in the reference group had left main coronary artery disease. Hospital mortality for the reoperation group was 6.8% (6 patients) and for the reference group, 3.4% (3 patients). No significant difference was found in the incidence of reoperation for bleeding, sternal infection, or stroke in the two groups. The incidence of respiratory insufficiency in the reoperation group was 13.6% (12 patients) and in the reference group, 3.4% (3 patients) (p < 0.015). Recurrent angina occurred in 13.7% (10 patients) of patients in the reoperation group and 13.3% (10 patients) in the reference group. The long-term survival at 5 years for the reoperation group was 85.3% & 5.6% (2 standard error of the mean) and for the reference group, 91.6% f 3.1%. No significant difference was found in the equality of survival distribution for the two groups. The results of this comparative study demonstrate that bilateral internal mammary artery grafting can be accomplished with an acceptable operative risk in patients undergoing reoperation. Further, survivors of reoperation experience long-term survival benefits and functional improvement similar to those receiving primary coronary bypass surgery. (Ann Thorac Surg 1991;52:20-8) eoperation for coronary artery disease has become a R standard procedure, accounting for 4% to 6% of all coronary bypass operations performed annually [ 11. The incidence of reoperation in patients followed up 10 and 15 For editorial comment, see page 3. years after primary myocardial revascularization at a major institution approaches 17% and 38%, respectively [2]. Vein graft atherosclerosis has been the major cause of recurrent symptoms and is the leading angiographic indication for coronary bypass reoperation [2-41. Despite increased surgical experience and major advances in the technology of myocardial revascularization, several studies [2-61 have documented increased operative risk and decreased postoperative benefit in patients undergoing reoperation compared with those receiving primary revascularization. The internal mammary artery (IMA) has demonstrated Presented at the Thirty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Dorado, Puerto Rico, Nov 8-10, Address reprint requests to Dr Galbut, 1150 NW 14th St, Suite 207, Mixnl, FL superior early and long-term patency compared with venous conduits [ Patients receiving a left IMA graft to the anterior wall compared with a saphenous vein graft during primary myocardial revascularization have experienced improved long-term survival and a reduction in the rate of reoperation and late cardiac events [7, 8, 10, 12, 131. These findings are further supported by our 17-year experience [14] with bilateral IMA grafts. Although the use of single IMA grafts in coronary bypass reoperation has been occasionally reported, the use of bilateral IMA grafts has been rarely described [24, 6, 13, 151. Also, the influence of the IMA compared with the saphenous vein on the long-term benefits of reoperative survivors has not been clearly established [3]. Recognizing the superiority of the IMA as a conduit in primary coronary artery revascularization, our group has employed bilateral IMA grafting as the surgical treatment of choice for candidates for reoperation. The purpose of this study is to analyze and compare retrospectively our experience with 88 consecutive patients undergoing reoperation using bilateral IMA grafts with a computermatched reference group. Further, this study focuses on the assessment of operative risk and the resulting longterm benefits by The Society of Thoracic Surgeons /91/$3.50

2 Ann Thorac Surg 1991;52:2&8 GALBUT ET AL 21 Material and Methods Patient Population Table 2. Comparison of Clinical Risk Factors by Patient Group" Between January 1982 and June 1989, 88 consecutive Reoperation Reference patients underwent reoperation for revascularization of Factor group group p Value the myocardium with bilateral IMA and supplemental Cigarette smoking 69 (78.4) 64 (72.7) NS vein grafts. This cohort was compared with a reference Family history of coronary 64 (72.7) 59 (67.0) NS group of 88 patients who underwent primary bilateral artery disease IMA operation without concomitant valve replacement or Previous myocardial infarction 63 (71,6) 55 (62.5) NS aneurysm operation during the same period. The two Hypertension 39 (44.3) 44 (50.0) NS groups were computer matched for sex, age, left ventric- Diabetes 22 (25.0) 22 (25.0) NS ular function, anginal classification, and the presence of left main coronary artery disease (greater than 50% steno- Hyperli!Jidemia 27 (30.7) 15 (17.0) <0.05 sis). Numbers in parentheses are percentages Table 1 presents a comparison of selected preoperative NS = not significant clinical variables for the two groups. None of the variables achieved significance, thus ensuring that the two groups were comparable with regard to the clinical characteristics tients (53.4%), triple-vessel disease. Though more triplemeasured. The preoperative symptoms were defined by vessel disease was present in the reoperation group the New York Heart Association classification system. (65.9% versus 53.4%), no significant difference was There was no significant difference in the functional achieved. Left main coronary artery disease was present classification of patients undergoing reoperation versus in 19 patients (21.6%) in the reoperation group and in 18 those in the reference group. In each group, 94.3% of the patients (20.5%) in the reference group; this difference patients were in New York Heart Association class 111 or was also not significant. class IV. Patients in unstable condition constituted the In the reoperation group, dominant right coronary group with class IV symptomatology. artery circulation was found in 84 patients (95.5%) and Coronary clinical risk factors in the two patient groups dominant left coronary artery circulation in 3 patients included cigarette smoking, family history of coronary (3.4%). Balanced circulation was demonstrated angioartery disease, previous myocardial infarction, hyperten- graphically in 1 patient (1.1%). In the reference group, the sion (diastolic pressure greater than 90 mm Hg), diabetes mellitus, and hyperlipidemia (serum cholesterol level greater than 5.17 mmol/l [200 mg/100 ml]). The results of the comparison of these clinical risk factors are shown in Table 2. No significant differences were noted between the two patient groups except for the risk factor hyperlipidemia (30.7% versus 17.0%; p < 0.05). Preoperative angiographic findings in the reoperation group revealed that 30 patients (34.1%) had double-vessel disease (greater than 70% stenosis) and 58 patients (65.9%), triple-vessel disease. In the reference group, 41 patients (46.6%) had double-vessel disease and 47 pa- Table l. Comparison of Preoperative Clinical Variables by Patient Group" Variable Sex Male Female Age (Y) Mean Range Unstable angina Ejection fraction <0.50 Left main disease a Numbers in parentheses are percentages. NS = not significant. Reoperation Reference p group group Value 77 (87.5) 77 (87.5) 11 (12.5) 11 (12.5) NS ? 9.5 NS (62.5) 55 (62.5) NS 38 (43.2) 38 (43.2) NS 19 (21.6) 18 (20.5) NS angiographic findings were similar with 81 patients (92.0%) having dominant right coronary artery circulation, 6 patients (6.8%) having dominant left coronary artery circulation, and 1 patient (1.1%) having balanced circulation. There were a total of 263 coronary artery bypass grafts performed in the reoperation group (mean, 2.99 per patient; range, two to five). In the reference group, the total was 310 grafts (mean, 3.52 per patient; range, two to six). The mean number of grafts performed in the reference group was significantly higher than in the reoperation group ( p < 0.01). The type of conduit and corresponding recipient arteries for the two groups are shown in Table 3. In the reoperation group, there were 20 sequential left IMA and three sequential right IMA grafts. The right IMA was placed through the transverse sinus and grafted to the circumflex or obtuse marginal artery in 28 patients (31.8%). Two left IMA and 12 right IMA conduits were used as free grafts. In the reference group, 30 sequential left IMA and two right IMA grafts were used. The right IMA was placed through the transverse sinus and grafted to the circumflex or obtuse marginal artery in 46 patients (52.3%). One left IMA and two right IMA conduits were used as free grafts. Indications for Reoperation The major clinical indication for reoperation in all patients was recurrent symptoms. The majority of patients (62.5%) experienced unstable angina, which was found to be refractory to medical management. In patients with stable

3 22 GALBUT ET AL Ann Thorac Surg 1991;52:20-8 Table 3. Coronary Arteries Grafted and Type of Conduits by Patient Group LAD Diag cx RCAIPDCA Total Conduit Reop Ref Reop Ref Reop Ref Reop Ref Reop Ref Left IMA Right IMA Vein Total Cx = circumflex coronary artery; Diag = diagonal branch; IMA = internal mammary artery; LAD = left anterior descending coronary artery; PDCA = posterior descending coronary artery; RCA = right coronary artery; Ref = reference group; Reop = reoperation group. angina (37.5%), the underlying coronary artery and graft anatomy revealed severe atherosclerotic disease, which correlated with markedly positive exercise testing. Data on the angiographic indications for reoperation were available for 84 patients. Primary graft failure was defined as closure or atherosclerosis causing a 70% or greater degree of stenosis. The major indications for reoperation were graft failure or the combination of graft failure and progression of native coronary artery disease (greater than 70% stenosis), affecting 38 patients (45.2%) and 39 patients (46.4%), respectively. Progression of the native coronary artery disease with normally functioning grafts was seen in 7 patients (8.3%). This finding reflects a low incidence of incomplete revascularization at the initial operation. The mean number of grafts per patient was 2.69 (range, one to six) and the interval between the first operation and reoperation ranged from 6 months to 18 years (mean interval, 8.9 years). Operative Technique Details of the operative technique, including IMA mobilization, orientation, and anastomotic construction in bilateral IMA grafting, have been previously described [ 141. Reoperation, however, presents the surgeon with additional technical challenges. A major consideration in reoperation is the safety of sternal reentry. Preoperative assessment with a lateral chest roentgenogram and computed axial tomography was used to define the relationship of the anterior mediastinum and posterior table to the sternum. Visualization of both IMAs was accomplished and the orientation of functioning venous grafts was determined during preoperative cardiac catheterization. Patients with increased risk of reentry injury or instability included (1) those with mediastinal structures or patent coronary grafts adhering to the sternum, (2) those with coronary circulation dependent on atherosclerotic grafts, and (3) elderly women with osteoporotic sternums. In these patients, exposure of the femoral vessels for possible peripheral cannulation was performed before repeat sternotomy. Twenty-two patients (25.0%) with severe myocardial ischemia, substantially impaired left ventricular function, or a combination of these and with acceptable peripheral circulation had preoperative intraaortic balloon support. The sternotomy was performed with either the Sarns or Stryker saw after head elevation, lung deflation, and blood pressure control to approximately 90 mm Hg with vasodilating agents in the majority of patients. Peripheral cannulation with temporary cardiopulmonary bypass and decompression of the heart during the sternotomy was used occasionally in high-risk patients. After an uneventful reentry in these patients, cardiopulmonary bypass was discontinued during the initial dissection and mobilization of the IMAs. After the sternotomy, the ascending aorta and right atrium were exposed before mobilizing the IMAs. Cardiopulmonary bypass was initiated before completion of lysis of left ventricular adhesions. Great care was exercised not to enter or manipulate atherosclerotic but patent vein grafts. Such vein grafts were not routinely removed or ligated, particularly when the proximal native circulation was occluded. The IMA conduits either spanned anastomotic lesions or were placed distally. New bypass grafts were occasionally constructed to the myocardial regions supplied by old but minimally atherosclerotic grafts. Proximal new venous or free IMA anastomoses were made in the hood of the old aortovenous anastomosis, a site that is usually patent, thinner, and more pliable than the surrounding aorta. Myocardial protection during aortic cross-clamping was accomplished with topical cooling and antegrade cold (4" to 5 C) blood cardioplegia repeated every 20 minutes to maintain the septa1 temperature between lo" and 15 C. Retrograde coronary sinus cardioplegia has been used during the past 2 years. In most patients, aortic root venting was also employed. The mean cardiopulmonary bypass time was * 65.9 minutes (range, 31 to 366 minutes) for the reoperation group and * 45.6 minutes (range, 55 to 260 minutes) for the reference group (p < 0.01). The mean duration of aortic cross-clamping was minutes (range, 21 to 200 minutes) for the reoperation group and minutes (range, 32 to 176 minutes) for the reference group (p < 0.01). Patient Follow-up Perioperative data were obtained from each patient's hospital record. Follow-up information was collected yearly or more frequently as necessary through direct patient contact, from the patient's personal physician, from responses to mailed comprehensive questionnaires, or by telephone interview with surviving patients or family members. The follow-up interview included information regarding current symptoms, medications being

4 Ann Thorac Surg 1991:52:2&8 GALBUT ET AL 23 taken, and activity level. Moreover, patients were asked to describe their functional capability so as to assess their status based on the criteria of the New York Heart Association. Autopsy reports and death certificates, when available, provided additional information. Data collected were entered into a computer-assisted reporting system. Two computer-compatible data collection instruments-a patient registration form and a follow-up form-were completed for each patient. These data collection instruments provided standardized reporting of each patient s clinical status preoperatively and postoperatively. Stat is t ical Analysis Data are presented in the text and tables as frequency distributions and simple percentages. Values of continuous variables are expressed as the mean 2 the standard deviation. Univariate analysis of all discrete variables comparing the reoperation group with the reference group was accomplished with the x test with the appropriate degrees of freedom. Yates correction for continuity was applied where applicable. Comparison of means for continuous variables was conducted using an unpaired Student s t test. Patient survival is expressed by actuarial analysis according to the method of Cutler and Ederer [16] and by linearized occurrence rates. The algorithm of Lee and Desu [17] was used to test the equality of survival distribution for the two patient groups. Data collected were subjected to both quantitative and qualitative analysis with the aid of the Patient Analysis and Tracking System software [18]. Differences resulting in a p value of less than 0.05 were considered significant. Results Mortality The hospital mortality for the reoperation group was 6.8% (6 patients) and for the reference group, 3.4% (3 patients). No significant difference was observed in the hospital mortality for the two groups. Moreover, the presence of left main coronary artery disease, impaired left ventricular function, or unstable angina did not influence hospital mortality in the two groups. Hospital Complications Table 4 presents a comparison of selected operative variables by patient group. The hospital complications reported in the two groups included reoperation for bleeding, sternal infection, respiratory failure with intubation for more than 48 hours or tracheostomy (or both), perioperative myocardial infarction, and stroke. Myocardial infarction was defined as electrocardiographic findings of a new onset of Q waves. Statistical comparison of the occurrence rates in the two groups revealed that patients having reoperation experienced a significantly higher incidence of respiratory failure (p < 0.015). Moreover, a greater number of patients in the reoperation group required mechanical support (intraaortic balloon pump) than patients in the reference group (p < 0.001). The intraaortic balloon pump was used in the reoperation group in 25.0% (22 patients) preoperatively, 9.1% (8 Table 4. Comparison of Operative Results by Patient Group Variable Reoperation Reference p group group Value Reoperation for bleeding 5 (5.7) 4 (4.5) NS Sternal infection 3 (3.4) 2 (2.3) NS Respiratory failure 12 (13.6) 3 (3.4) c0.015 Perioperative myocardial 7 (8.0) 2 (2.3) NS infarction Stroke 3 (3.4) l(1.1) NS Mechanical support (IABP) 31 (35.2) 13 (14.8) <0.001 Mean hospitalization (days) 16.9? ? 5.6 <0.038 a Numbers in parentheses are percentages. IABP = intraaortic balloon pump; NS = not significant. patients) intraoperatively, and 1.1% (1 patient) postoperatively. In the reference group, it was used in 10.2% (9 patients) preoperatively, 3.4% (3 patients) intraoperatively, and 1.1% (1 patient) postoperatively. The reported incidence of the complications reoperation for bleeding, sternal infection, myocardial infarction, and stroke generated no significant difference between the groups. A significantly greater number of patients in the reference group experienced no complications compared with the reoperation group (88.6% versus 70.5%; p < 0.003). The mean hospital stay for patients in the reference group was significantly shorter than for those in the reoperation group ( versus 16.9 * 18.5 days; p < 0.038). Late Cardiac Events The linearized occurrence rate and number of late cardiac events in the 82 hospital survivors in the reoperation group were as follows: nonfatal myocardial infarction, 0.78% per patient-year (two events) and reoperation, 0.78% per patient-year (two events). For the 85 hospital survivors in the reference group, nonfatal myocardial infarction occurred at a rate of 0.95% per patient-year (three events). Late nonfatal myocardial infarction was diagnosed and documented in both groups through a review of patient hospital records, physician records, and the patient s medical history. No significant difference was noted in the incidence of late nonfatal myocardial infarction between the groups. Long-Term Follow-up The average duration of follow-up for hospital survivors in the reoperation group (n = 82) was 37.4 months (range, 1 to 81.6 months) and in the reference group (n = 85), 44.4 months (range, 1 to 76.8 months). The cumulative follow-up for the reoperation group was 256 patient-years and for the reference group, 315 patient-years. The linearized late mortality rate for the reoperation group was 2.73% per patient-year and for the reference group, 2.54% per patient-year. At the completion of the current follow-up, 73 (89.0%) of the hospital survivors in the reoperation group and 75 (88.2%) in the reference group were alive. Two patients (2.4%) in each group were lost to follow-up. The current

5 24 GALBUT ET AL Ann Thorac Surg 1991;52:2&8 Table 5. Current Status of Hospital Survivors by Patient Groupa Status Reoperation Reference group (n = 82) group (n = 85) Alive 73 (89.0) 75 (88.2) Lost to follow-up 2 (2.4) 2 (2.4) Deaths (late mortality) 7 8 Cardiac related 6 (7.3) 3 (3.5) Noncardiac related 0 (0) 3 (3.5) Cause unknown l(1.2) 2 (2.4) a Numbers in parentheses are percentages. status of hospital survivors and the causes of late death are presented in Table 5. The survival data for the 82 patients in the reoperation group and the 85 patients in the reference group discharged from the hospital are shown in Figure 1. At 5 years, survival was 85.3% & 5.6% (+- the standard error of the mean) in the reoperation group and 91.6% & 3.1% in the reference group. The equality of survival distribution for the two groups was tested, and no significant difference was noted. The preoperative and postoperative New York Heart Association classification of patients and their migration pattern are shown in Figure 2 for the reoperation group and in Figure 3 for the reference group. In the reoperation group, all of the patients improved at least one class except 3. One of these patients remained in class 111 and 2, in class IV. Preoperatively, 93.2% of this patient group were in either class I11 or IV, and at the completion of the current follow-up, 93.2% were in either class I or 11. In the reference group, all of the current survivors improved at least one class except 2 patients, 1 who remained in class I11 and 1 who went from class I1 to class 111. This group demonstrated an enhanced level of functional improvement as evidenced by the fact that 93.3% of the patients were in either class I11 or IV preoperatively and 96.0% were in either class I or I1 at the conclusion of the study. This enhanced level of functional improvement in the reference group did not achieve significance. Graft Putency Seventeen patients in the reoperation group were restudied from 6 months to 53 months postoperatively (mean interval, 25.6 months) for recurring symptoms suggestive of angina. Of the 20 left IMA anastomoses, 17 (85.0%) were patent, and of the 17 right IMA anastomoses, 13 (76.5%) were patent. Of the 12 saphenous vein grafts, eight (66.7%) were patent. Comment The IMA is the preferred conduit in coronary bypass grafting. Although its use is technically more demanding, it has intrinsic qualities that make it relatively immune to the three biological modes of graft failure affecting venous conduits: early thrombosis, intermediate subintimal fibrosis, and late atherosclerosis [8, 15, 191. These characteristics contribute to a reduced rate of reoperation in patients receiving IMA grafts compared with those receiving only vein grafts [2, 12, 131. In recent years, use of the IMA has been praised and expanded with bilateral and sequential grafts mostly in primary coronary bypass procedures [6, 141. Coronary artery reoperation is being performed with greater frequency and has been associated with increased operative mortality and decreased long-term improvement [2-61. The use of bilateral IMA grafts in reoperation has rarely been reported [24, 6, 13, 151 and has not been carefully analyzed. Further, comparisons of primary and reoperative coronary bypass results have usually included groups of patients with differing clinical characteristics undergoing operation during the same interval. This study compares two homogeneous groups. Preoperative clinical variables that have been associated with increased operative mortality, including female sex, advanced age, unstable angina, reduced ejection fraction, and the presence of left main coronary artery disease [2CL23], were equally distributed in both groups in the present study. The incidence of other clinical risk factors, Fig 1. Actuarial survival by group for patients discharged from the hospital. :L I - PRIMARY -- REOPERATION L o b MONTHS AFTER OPERATION

6 Ann Thorac Surg 1991;52:20-8 GALBUT ET AL 25 U Fig 2. Preoperative and postoperative Neu, York Heart Association functional class for patients undergoing reoperative corona ry bypass surgery (current survivors). including cigarette smoking, family history of coronary artery disease, prior myocardial infarction, hypertension, and diabetes, were also found to be similar in both groups. However, patients in the reoperation group had a significantly higher incidence of hyperlipidemia (p < 0.050). There is substantial evidence that elevated plasma lipoprotein levels are a principal correlate of the development of atherosclerosis in vein grafts or native coronary circulation in patients studied 10 years after primary bypass grafting [l]. Further, it has been demonstrated that patients with hyperlipidemia are more likely to require reoperation. In the present study, the profile of selected clinical characteristics mirrors that reported in the literature in recent years [2,3]. Patients were older, with a mean age of 62.1 years in the reoperation group and 61.9 years in the reference group. Approximately one fifth of the patients were 70 years of age or older, 22.7% in the reoperation group and 20.5% in the reference group. Almost two thirds (62.5%) of the patients had unstable angina and required urgent or emergent operation. The literature reports that angiographic indications for reoperation have varied depending on the interval after primary coronary artery bypass grafting and the extent of revascularization. In the 1970s and early 1980s, incomplete revascularization during the first operation was more common and often contributed to the progression of native coronary artery disease [3, 51. In recent years, vein graft failure secondary to atherosclerosis, alone or in combination with progression of the native coronary artery disease, has been reported in the majority of patients [2-51. This reflects an improved standard of more complete revascularization during the initial procedure and a longer interval between operations. Graft atherosclerosis usually develops in the fourth to fifth year postoperatively and continues to progress thereafter. At 10 years, approximately 40% of vein conduits are occluded, and 50% of the rest have major stenoses [l]. In the present study, patients in the reoperation group were seen after a mean postoperative interval of almost 9 years. These patients received a mean of 2.69 grafts at the initial procedure. Graft atherosclerosis, alone or in concert with progression of native coronary artery disease, was responsible for the need of reoperation in 91.6% (77 patients) in the present study. Myocardial preservation may be more difficult to accomplish during reoperation compared with the primary bypass procedure. This is due to a more complex coronary circulation, which is supplied through atherosclerotic vein grafts and stenotic or occluded native arteries. Antegrade cardioplegia may also promote atheroembolism in diseased vein grafts. In 1988, the use of retrograde cardioplegia was reported to enhance myocardial protection [24]. This method, which provides a uniform distribution of cold cardioplegia, is particularly helpful in a reoperation and may wash out any graft atheroembolic debris in the distal coronary arteries. The approach to functioning primary vein grafts is also important. Manipulation of the graft should be minimized to reduce the risk of atheroembolism. Some authors [l-31 have recommended ligation and division of all diseased vein grafts. Others have advocated retaining diseased but nonoccluded grafts to dependent myocardial regions that become the recipients of IMA grafts during reoperation [6]. In the present series, aneurysmal or severely diseased vein grafts were usually, but not uniformly, ligated. Functioning grsfts older than 5 years without focal or diffuse atherosclerosis were not routinely replaced. Hospital mortality is not increased with bilateral IMA grafting. Our experience [14] with 1,087 patients under- (PRE-OP) L I (POST-OP) Fig 3. Preoperative and postoperative New York Heart Association functional class for patients undergoing primary coronary bypass surgery (current survivors).

7 26 GALBUT ET AL Ann Thorac Surg 1991;52:2C-8 going primary coronary artery revascularization revealed a 2.7% hospital mortality (29 patients). The computermatched reference group in this study had a comparable rate of 3.4% (3/88 patients). Although the mortality of 6.8% (6/88 patients) in the reoperation group was greater, the difference did not achieve significance. Recently published hospital mortality rates for repeat coronary bypass operation have ranged from 2.7% to 7.1% [24, 61. The Coronary Artery Surgery Study registry review [5], representing a multicenter experience, reported a reoperative hospital mortality rate of 5.3% (15/283 patients). It has been reported that reoperation is a significant predictor of hospital mortality in coronary bypass procedures [ In several studies [l, 3, 5, 61, a significant difference in hospital mortality between primary and reoperative coronary revascularization has been reported. Cameron and colleagues [4], however, reported no significant difference in the hospital mortality of patients undergoing reoperation versus those having primary revascularization during a 10-year interval (4.7% versus 2.4%, respectively). Our findings are similar but both groups were matched for established predictors of mortality. The use of bilateral IMA grafts in primary coronary artery bypass grafting has not been associated with an increased incidence of sternal infection, reoperation for mediastinal bleeding, perioperative myocardial infarction, or respiratory failure. The incidence of sternal wound infection has been found to be related more to diabetes and advanced age than to bilateral IMA grafts [25]. In this study, the reoperation group experienced greater hospital morbidity, with 29.5% (26/88 patients) having hospital complications compared with 11.4% (10/88 patients) in the reference group ( p < 0.003). The increased incidence of respiratory failure in patients having reoperation in the present study is very significant. This may be a function of longer operating and cardiopulmonary bypass periods, increased blood and colloid transfusions, and greater use of topical saline slush for myocardial protection. The greater incidence of perioperative myocardial infarction in the reoperation group is similar to other reports [3, 41. This reflects the increased difficulty of myocardial preservation and the inherent risk of atheroemboli from diseased but functioning vein grafts. The goals of coronary artery revascularization are to prolong survival, eliminate or ameliorate angina, and reduce the need of a subsequent procedure. Five-year survival after repeat coronary bypass procedures has ranged from 85% to 90% [3, 41. Patients with advanced age, hypertension, incomplete revascularization, and impaired left ventricular function had reduced late survival [3]. Cameron and associates [4] reported a 10-year survival of 82% in patients undergoing reoperation, a figure similar to the rate in patients undergoing primary revascularization. In contrast, Lytle and co-workers [3] showed a 75% survival rate at 10 years in patients having reoperation, which was inferior to the results achieved after primary bypass grafting. In this series, the 5-year survival rate approached 85.3% * 5.6% in the reoperation group and 91.6% 2 3.1% in the reference group. Statistical comparison of the equality of survival distribution resulted in no significant difference, thus demonstrating a similar pattern of long-term patient survival for the two groups. Several studies [24] have recorded a reduction in the relief of angina in coronary reoperation compared with primary revascularization. Loop [2] found approximately 50% of the patients who underwent reoperation had recurrent angina 10 years after reoperation. Cameron and colleagues [4] reported 66% of such patients had recurrent angina compared with 35% of patients having primary revascularization followed up to 5 years. In the present series, 86.3% (63173 patients) and 86.7% (65175 patients) of survivors in the reoperation and reference groups, respectively, were asymptomatic or free from angina. This similarity of postoperative functional status in both groups differs from what is found in the literature [24] and may suggest the influence of bilateral IMA grafts. The postoperative patency of the isolated IMA graft in the reoperation group is comparable with the results in primary revascularization and supports this inference [8]. The frequency of reoperation for clinically symptomatic patients is increasing. The results of this comparative study demonstrate that reoperation with bilateral IMA grafts can be accomplished with an acceptable operative risk. Moreover, patients having reoperation experience long-term survival and functional improvement comparable with those undergoing primary revascularization. Bilateral IMA grafts appear to be the conduit of choice in reoperation and may reduce the need for a third procedure. We thank Dr George Ebra, research consultant, for his technical assistance in the preparation of this report and Mrs Kathie Roy, RN, and Mr Brian Galbut for their efforts in data collection. References 1. Foster ED. Reoperation for coronary artery disease. Circulation 1985;72(Suppl 5): Loop FD. A 20-year experience in coronary artery reoperation. Eur Heart J 1989;1O(Suppl H):7W. 3. Lytle BW, Loop FD, Cosgrove DM. Fifteen hundred coronary reoperations. Results and determinants of early and late survival. J Thorac Cardiovasc Surg 1987;93: Cameron A, Kemp HG Jr, Green GE. Reoperation for coronary artery disease. Ten years of clinical foi'5jw-up. Circulation 1988;78(Suppl 1):15% Foster ED, Fisher LD, Kaiser GC, Myers WO, principal investigators of CASS and their associates. Comparison of operative mortality and morbidity for initial and repeat coronary artery bypass grafting: the Coronary Artery Surgery Study (CASS) registry experience. Ann Thorac Surg 1984;38: Jones EL, Lattouf OM, Weintraub WS. Catastrophic consequences of internal mammary artery hypoperfusion. J Thorac Cardiovasc Surg 1989;98: Campeau L, Enjalbert M, Lesperance J, et al. The relation of risk factors to the development of atherosclerosis is saphenous vein bypass grafts and the progression of disease in the

8 Ann Thorac Surg 1991;52:2M GALBUT ET AL 27 native circulation: a study 10 years after aortocoronary bypass surgery. N Engl J Med 1984;311: 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; 7o(s~ppi 1):20a Lytle BW, Loop FD, Cosgrove DM, et al. Long-term (5 to 12 years) serial studies of internal mammary artery and saphenous vein coronary bypass grafts. J Thorac Cardiovasc Surg 1985;89:24a 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: Ivert T, Nuttunen K, Landou C, Bjork VO. Angiographic studies of internal mammary artery grafts 11 years after coronary artery bypass grafting. J Thorac Cardiovasc Surg 1988;96: 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: Cameron A, Green GE, Kemp HG Jr. Role of internal mammary artery in reoperations for coronary artery disease. Adv Cardiol 1988;36: Galbut DL, Traad EA, Dorman MJ, et al. Seventeen-year experience with bilateral internal mammary artery grafts. Ann Thorac Surg 1990;49: Green GE. Reoperations and the internal mammary artery. Adv Cardiol 1988;36: Cutler SJ, Ederer F. Maximum utilization of the life table method in analyzing survival. J Chronic Dis 1958;8: Lee E, Desu M. Computer program for comparing K samples with right-censored data. Comput Programs Biomed 1972;2: Anonymous. Patient analysis and tracking system, users guide. Portland, OR Dendrite Systems, Subramanian VA, Hernandez Y, Tack-Goldman K, Grabowski EF, Weksler BB. Prostacyclin production by internal mammary artery as a factor in coronary artery bypass grafts. Surgery 1986;100: Naunheim KS, Fiore AC, Arango DC, et al. Coronary artery bypass grafting for unstable angina pectoris: risk analysis. Ann Thorac Surg 1989;47: Parsonnet V, Dean D, Bernstein AD. A method of uniform stratification of risk for evaluating the results of surgery in acquired adult heart disease. Circulation 1989;79(Suppl 1): Galbut DL, Chen R, Traad EA, et al. Predicting operative mortality in coronary bypass surgery: the influence of unstable angina [Abstract]. Presented at the 12th Interamerican Congress of Cardiology, Rio de Janeiro, Brazil, July 23-28, Grover FL, Hammermeister KE, Burchfiel C, et al. Initial report of the Veterans Administration preoperative risk assessment study for cardiac surgery. Ann Thorac Surg 1990; 50: Diehl JT, Eichhorn EJ, Konstam MA, et al. Efficacy of retrograde coronary sinus cardioplegia in patients undergoing myocardial revascularization: a prospective randomized trial. Ann Thorac Surg 1988;45: Cosgrove DM, Lytle BW, Loop FD, et al. Does bilateral internal mammary artery grafting increase survival risk? J Thorac Cardiovasc Surg 1988;95:85M. DISCUSSION DR DENIS H. TYRAS (New Hyde Park, NY): I have a little problem with some of the definitions and terms that you use. Would you please clarify your definition of unstable angina because it struck me that there was an extraordinarily high incidence in this series. I think we need to be precise in our definitions. When you compared your data with those in the preceding presentation from Nashville, I was impressed by the difference in the reoperation groups. In your reoperation group, about half of the patients had normal ventricular function. That is at variance with the data presented by the Nashville group, yet your operative mortality was significantly higher than theirs. I realize that your numbers are small, and I worry about the possibility of a type I1 error; you have twice the operative mortality in the reoperation group as in the reference group, but it has not achieved significance. This could be simply because of the small size of the sample. I also take issue with the common practice of quoting perioperative myocardial infarction rates by looking at Q waves. I think this underestimates the incidence of infarction. To be more precise, it is necessary to look at cardiac isoenzyme levels. I would appreciate your addressing these points. DR GALBUT I think there are three points, and I will do my best to respond to them. First, is the issue of unstable angina. The definition used in this study included (1) patients who had an accelerated pattern (the majority were given intravenous nitroglycerin); (2) patients who had a myocardial infarction and continued to have postinfarctional pain for a period of up to 30 days; and (3) patients with impending infarction who would fit the category of the intermediate coronary syndrome. These were the categories that were incorporated into our group with unstable angina, all carefully delineated not just by the cardiologist but also by myself and my partners. Another point concerned the difference in mortality between our presentation and the previous presentation. I did not have an opportunity to review the details of that paper or look at the number of patients with unstable angina to match the same preoperative variables. However, if you have two groups of 88 patients or 95 patients, 1 or 2 deaths may be significant, but statistical methods were employed and are described in our manuscript. I am not certain there is a major difference in the mortality figures. The third point had to do with perioperative myocardial infarction. I concur with you, and that is why definitions must be stringent when presentations are made. Not being certain about the accuracy of the data regarding isoenzymes, we relied on new-onset Q waves. DR JOSEPH M. CRAVER (Atlanta, GA): First, would you elaborate on the strategy of using the right IMA in dealing with thickened aortas and distal graft stenosis? Second, what is your thought in regard to using IMAs to

9 28 GALBUT ET AL Ann Thorac Surg 1991;52:204 rebypass a stenotic, but not critically stenosed, vein graft, such as a 70% graft lesion? DR GALBUT I will address the second question first. There is an important issue as to whether a patent but diseased vein graft should be removed or ligated or excised. We have not routinely removed all vein grafts that are functioning but stenotic. We also have not routinely ligated them, particularly when a vein graft is 70% or 80% occluded and services an area of the heart that is graft dependent. Where we have placed an IMA, we have usually left that vein graft alone to function. Another consideration is that if the IMAs are small, there may be a little limitation on the degree of spanning of lesions that can actually be accomplished. If done, it has to be done carefully. Specifically, if we have a graftdependent circulation with a vein graft that is diseased, we place an IMA to that myocardial bed. We do not remove the vein graft. With reference to the strategy of use of the IMAs, we believe very strongly that the isolated IMA conduit affords additional length, and because the pedicle is not thick, it is easier to place this vessel out of harm's way. In our series, the left IMA was used to accomplish 108 grafts, 20 of which were sequential, primarily to a diagonal and the left anterior descending coronary artery. The left IMA was grafted to the left anterior descending, diagonal (mostly sequentially), occasionally to the circumflex artery, and occasionally, when the posterior descending coronary artery was in a left-dominant circulation, directly to the posterior descending coming from the left coronary artery. The right Ih4A was grafted in 14 patients in front of the aorta and pulmonary artery to the left anterior descending coronary artery, to the diagonal a few times, to the circumflex many times, and in 40 patients to the right coronary artery. In many instances, the right IMA was placed through the transverse sinus. About three or four were free grafts where the proximal anastomosis was constructed to the hood of an old saphenous vein graft that was occluded. Three of the right IMAs were used as sequential grafts. There were 64 vein grafts. DR LYNN H. HARRISON (Marrero, LA): I was interested in your data regarding the incidence of mediastinal wound complications. Although you pointed out that there was no significant difference between the reference group and the reoperation group, it struck me that those numbers were a little high in both cases. What was the mortality associated with this complication in your series? Do you have any hesitancy about using bilateral IMAs in the subset of patients who are diabetic? DR GALBUT I do not remember the exact percentage of wound complications. I think it was about 3.5% in the reoperation group and about 2% in the reference group. We believe that perhaps the most important contraindication to use of bilateral IMA grafts is diabetes. Diabetes is associated with a higher incidence of sternal wound infection than we think when bilateral IMAs grafts are used. Another group in whom it would be preferable to not place bilateral IMA grafts is elderly women with osteoporotic sternums.