and Paul C. Taylor, M.D. ORIGINAL ARTICLES

Transcription

1 ORIGINAL ARTICLES Trends in Selection and Results of Coronary Artery Reoperations Floyd D. Loop, M.D., Bruce W. Lytle, M.D., Carl C. Gill, M.D., Leonard A. R. Golding, M.D., Delos M. Cosgrove, M.D., and Paul C. Taylor, M.D. ABSTRACT The first 1,000 patients undergoing isolated coronary artery reoperation were divided into four cohorts of 250 patients each 1969 to 1976; 1976 to 1979; 1979 to 1981; and 1981 into Graft failure as an indication for reoperation rose from 26% in Group 1 to 40% in Group 4, and the interval lengthened from 17 to 61 months, presumably a result of late closures 5 to 10 years postoperatively. Progressive atherosclerosis in previously ungrafted vessels has decreased from 62% in Group 1 to 23% in Group 4, a decline attributed to more complete revascularization initially. The frequency of threevessel disease, stenosis of the left main coronary artery, and left ventricular impairment continues to rise in candidates for reoperation. Yet, operative mortality has declined from 5% to 2%, and most other forms of perioperative morbidity have decreased significantly when the early years are compared with the later experience. The number of grafts per patient has increased from 1.4 to 2.3, and complete revascularization in reoperations has increased from 65% to 76%. After a mean of 29 months, graft patency was 81% overall in 154 patients restudied after reoperation. Patency was similar for grafts to arteries previously involved with graft failure and to arteries not previously grafted. Five-year actuarial survival for patients in the first three cohorts (mean, 57 months) was 89%. Survival was affected significantly by extent of disease and left ventricular performance before reoperation and by completeness of revascularization; survival was not affected by the reason for reoperation, that is, graft closure, progressive atherosclerosis, or both. Early in the evolution of coronary artery surgery, reoperations were performed largely From the Department of Thoracic and Cardiovascular Surgery, The Cleveland Clinic Foundation, Cleveland, OH. Presented at the Nineteenth Annual Meeting of The Society of Thoracic Surgeons, San Francisco, CA, Jan 17-19, Address reprint requests to Dr. Loop, The Cleveland Clinic Foundation, 9500 Euclid Ave, Cleveland, OH for incomplete revascularization and early graft closure, frequently within the first few postoperative years. For nearly 10 years, coronary artery reoperations were fraught with high mortality and morbidity attributed to inexperience, technical accidents, poor myocardial protection, prolonged oxygenator perfusion, hemorrhage, and increased blood requirements. In the late 1970s, results began to improve because of greater operative experience and better myocardial protection. In the 1980s, a second wave of patients requiring coronary artery reoperations is expected. In contrast to patients in the first wave in the 1970s, these patients often have had a successful first operation: symptoms were relieved or lessened and left ventricular function was preserved for 5 or more years. Angina recurred, however, because of progressive coronary atherosclerosis or because of late graft thrombosis or atherosclerosis. During 15 years of performing myocardial revascularization, we have compiled an experience of 1,000 coronary artery reoperations. Trend analysis of early and late results depicts changes in a series of reoperations in one institution and includes patient characteristics correlated with intervals between operations and intermediate results. Patients and Method This series includes 1,000 patients who had myocardial revascularization initially and then underwent one reoperation that consisted of bypass grafts only. These 1,000 consecutive patients have been divided into four cohorts of 250 each. Table 1 shows the time of the reoperation in years for each cohort. One group of patients (mainly in the first cohort) had had a Vineberg implant or a bypass graft with a Vineberg implant as the first operation. In addition, Table 1 contains the mean patient age at the first and 380

2 ~ ~~~~~ 381 Loop et al: Coronary Artery Reoperations Table 1. Coronary Artery Reoperations ( ) Variable Series ,000 Reoperation years First operation (no. of patients) Grafts only Implant only Graft(s) + implant Mean age, first operation (yr) Mean age, second operation (yr) Interval between operations (mo)a "Note the lengthening interval between the first and second operations. Table 2. Clinical Risk Factors of Patients in Reoperation Seriesa Factor (N = 250) (N = 250) (N = 250) (N = 250) Diabetes (on insulin) 15% 37 14% 35 12% 31 19% 48 Hypertension (on admission) 20% 34% 31 % 35% Cigarette smoking 30% 25% 24 % 20% Cholesterol mg/dl 37% 52% 56% 48% Triglycerides > 140 mg/dl 72 % 74 % 75 % 66% Family history of coronary artery disease 47% 44% 57% 61 % Interval myocardial infarction 18% 24% 21 % 26% "Second number for each category is the number of patients involved. second operations, and the interval between operations in months. The ratio of women to men was approximately the same as in our primary revascularization series. Overall, women made up 11.4% (114) of the patients having reoperation but accounted for 13.6% and 12.8% of patients in 1980 and 1981, respectively. Data for 1982 are incomplete. Clinical risk factors are depicted in Table 2. Diabetes requiring insulin treatment was documented twice as often as in patients having primary revascularization. As we have reported previously [ 11, our reoperation experience has been divided into catego- ries defined by angiographic indications: graft closure; progressive atherosclerosis in previously ungrafted vessels; or a combination of graft closure and progressive atherosclerosis. Table 3 shows the changes in the prevalence of these indications within each cohort. Although reoperations for graft closure are increasing, the interval between operations for this indication now exceeds 5 years. In contrast, reoperation for progressive coronary atherosclerosis is declining in frequency. Functional class for angina was categorized according to the New York Heart Association (NYHA) classification. Overall, severity of angina did not differ appreciably among cohorts.

3 ~ ~~~~ 382 The Annals of Thoracic Surgery Vol 36 No 4 October 1983 Table 3. Angiographic Indications and Mean Interval between Operations Indication and Interval Graft closure 65 (26%) 82 (33%) 88 (35%) 100 (40%) Interval (mo) Progressive atherosclerosis 154 (62%) 103 (41%) 68 (27%) 58 (23%) Interval (mo) Combined (graft closure and progressive 31 (12%) 65 (26%) 94 (38%) 92 (37%) atherosclerosis) Interval (mo) Number in parentheses is the prevalence of that indication within its respective cohort. The mean interval in months between operations is for patients operated on for that angiographic indication. Table 4. Extent of Coronary Atherosclerosis at Reoperation Involvement One vessel (9.2%) (7.2%) (4.0%) (2.8%) Two vessels (33.6%) (24.8%) (22.4%) (24.0%) Three vessels (57.2%) (68.0%) (73.6%) (73.2%) Left main coronary arteryb (10.8%) (14.4%) (19.2%) (18.0%) Greater than 50% estimated reduction in diameter of lumen. breported independently. The number of patients in Functional Class I (asymptomatic) ranged from 2 (0.8%) in 1 to 6 (2.4%) in 3. The number in Class I1 ranged from 56 (22.4%) in 4 to 72 (28.8%) in 2. The number in Class I11 ranged from 96 (38.4%) in 3 to 116 (46.4%) in 2, and in Class IV, from 58 (23.2%) in 2 to 85 (34.0%) in 4. The number in Angina Class I11 or IV ranged from 68.8% (172 patients) to 76.4% (191 patients) in all years surveyed. Table 4 reviews the extent of coronary atherosclerosis documented by coronary arteriography before reoperation. Major arterial obstruction greater than 50% (estimated) was considered severe coronary atherosclerosis. Single-vessel disease and double-vessel disease have declined in prevalence from the first to the fourth cohort, whereas triple-vessel disease and left main coronary artery disease have increased. Table 5 shows that fewer patients come to reoperation with normal left ventricular function today compared with patients in the first cohort operated on from 1967 t o In all years, approximately a third of the patients have lost previously normal left ventricular function between operations. Results Mortality and Morbidity Operative mortality and major morbidity, including blood usage, are shown in Table 6 for each of the four cohorts of 250 patients. Operative mortality has fallen from 4.8% in the first group of patients to 2% in the most recent cohort. Perioperative myocardial infarction (defined as the appearance of new Q waves) has decreased slightly, reoperation for bleeding has been halved, and reduced blood usage is evident in recent years. The rate of stroke is approximately the same throughout all years, and

4 383 Loop et al: Coronary Artery Reoperations Table 5. Left Ventricular lmpairment before Reoperation Normal (52.8%) (64.8%) (38.4%) (39.6%) Impaired (47.2%) (35.2%) (61.6%) (60.4%) Lost normal LV function in interval between operations (43.7%) (25.0%) (47.0%) (28.0%) LV = left ventricular. Table 6. Mortality and Morbidity" Variable Operative mortality (4.8%) (3.2%) (2.4%) Perioperative myocardial infarction (7.6%) (5.6%) (6.8%) Bleeding requiring reopening (8.8%) (11.6%) (7.6%) Mean units of blood required (2.0%) (5.2%) (4.4%) Neurological deficit Respiratory dysfunction Wound complication (2.0%) (3.2%) (1.2%) (1.2%) (7.6%) (2.8%) (2.4%) (1.6%) (1.6%) (2.4%) (0.4%) (0.4%) "All numbers are numbers of patients except values for mean blood units. respiratory and wound complications have decreased in the most recent cohort. In all years, the number of grafts per patient has increased from 1.4 to 2.3, and complete revascularization has improved from 65% to 76% (not significant). A comparison was made between the 470 patients who underwent reoperation under normothermic arrest with the 530 who received cold potassium cardioplegia. Operative mortality was 4.3% in the former group and 2.1% in the more recent cardioplegia series (p = 0.047). The number of grafts per patient in the group administered cardioplegia was 2.2 versus 1.5 with normothermic arrest, which reflects the more recent experience. The most striking difference between the two methods is seen in the percentages of complete revascularization: 64% with cardioplegia and 39% with normothermic arrest (p < 0.001). The incidence of perioperative myocardial infarction was 5.1% with cardioplegia compared with 7.7% for normothermic arrest (not significant). Operative Risk Factors A maximum likelihood estimation procedure similar to the one used in past reoperation reports [l] was used to estimate the probability of operative mortality as a function of multiple variables. The quantitative and numerical variables analyzed for risk determination were described in the earlier report. Four of the seven factors individually associated with operative mortality remained predictors in the logistic model: incomplete first operation, incomplete reoperation, left main coronary artery disease greater than 50%, and age.

5 384 The Annals of Thoracic Surgery Vol36 No 4 October 1983 Table 7. Vein Graft Atherosclerosis Method of Confirmation LAD RCA Circ. Total TOTALLY OCCLUDED VEIN GRAFTS Pathology report only Operative note only Surgeon and pathologist VISUALIZED VEIN GRAFTS Arteriogram only Arteriogram and operation Arteriogram and pathology report with or without operative note LAD = left anterior descending coronary artery; RCA = right coronary artery; Circ. = circumflex coronary artery. As described in the earlier report [l], the higher risk in patients given implants compared with those who had only bypass grafts before reoperation is attributed to several factors. These include the longer interval between operations, more severe angina before reoperation, higher incidence of multivessel disease including left main coronary artery stenosis, greater left ventricular impairment before reoperation, and a higher rate of incomplete revascularization at reoperation. In addition, reoperation occurred, in general, in the early 1970s. Graft Putency After a mean catheterization interval of 29 months, 154 patients underwent postreoperative coronary arteriography. According to the records, 90% were restudied because of suspected recurring angina or sign of ischemia. Vein graft patency was 77.9% for 199 grafts studied, and internal mammary artery graft patency was 95.5% in 44 grafts. Overall, 197 of 243 grafts (81.1%) were open. Graft patency also was determined for grafts anastomosed to vessels previously involved with graft failure. Of 62 vein grafts in that category, 77.4% were open and 96.3% of 27 internal mammary artery grafts were patent. Thus, of the grafts anastomosed to vessels previously involved with a closed graft, 74 of 89 (83.1%) were open. When graft patency was determined for new grafts to vessels previously not grafted, 78.1% of 137 vein grafts and 94.1% of 17 internal mammary artery grafts were open, for an over- all graft patency of 79.9% in 154 grafts to arteries not grafted at the first operation. Vein Graft Atherosclerosis Vein graft atherosclerosis was found in 140 patients and documented in 171 of 269 vein grafts (Table 7). These 140 patients constituted 16.8% of 831 primary vein graft recipients. In patients with vein graft atherosclerosis, the mean interval from the first operation to catheterization before reoperation was 87 months. Eighty of the 171 affected grafts (46.8%) were occluded at the graft origin (mean interval, 85 months) and not visible angiographically. In the remaining 91 grafts, atherosclerosis was visualized on coronary angiography. These grafts were viewed after a mean interval of 88 months between the first operation and the catheterization performed before reoperation. This analysis underestimates the incidence of vein graft atherosclerosis, because some totally occluded grafts were not inspected by the surgeon or the pathologist. Clinical Results and Longevity The clinical result was determined for 659 surviving patients in the first three cohorts (mean follow-up, 57 months). In this subset, 344 (52.2%) were asymptomatic, and 184 (27.9%) were rated in NYHA Functional Class 11, 77 (11.7%) in Class 111, and 54 (8.2%) in Class IV. Five-year survival was calculated for patients in the first three cohorts (750 patients). Fiveyear survival was 88.7% (deaths from all causes included). By extent of disease, survival was

6 385 Loop et al: Coronary Artery Reoperations PERCENT SURVIVAL PERCENT SURVIVAL ONEVESSEL n= 51 TWO VESSEL n THREE VESSEL n = 499 LEFTMAIN n= COMPLETE n= INCOMPLETE n= YEARS Fig 1. Five-year actuarial survival according to the extent of coronary atherosclerosis found before reoperation in 750 patients undergoing coronary artery reoperation (first three cohorts). Numbers at the end of the curves are the 5-year endpoints. The mean follow-up was 57 months. 50 I I 1 I YEARS Fig 3. Patients who received complete revascularization (all major arteries greater than 1 mm in diameter and narrowed 50% or more were grafted) had a 5-year survival of 92.3% after reoperation compared with 85.6% for those classified as having incomplete revascularization. PERCENT SURVIVAL PERCENT SURVIVAL 100 I- 8ol I! 84.w... NORMAL LV FUNCTION n.396 MILD LV DYSFUNCTION = 211 MODERATE LV DYSFUNCTION n = 108 SEVERE LV DYSFUNCTION : 68.m I 70 I --- GRAFTCLOSURE n = 237 COMBINED INDICATIONS n = 187 PROGRESSIVE ATHEROSCLEROSIS n = % YEARS Fig 2. Five-year survival correlated with left ventricular (LV) function before reoperation. Patients with normal left ventricular contraction or mild impairment achieved higher 5-year survival than those with moderate or generalized impairment YEARS Fig 4. Patients reoperated on for graft closure, progressive atherosclerosis, or both achieved nearly the same 5-year survival. highest among patients with single-vessel disease (94.8%) and lowest among those with left main coronary artery disease (83.7%) (reported independently) (Fig 1). In patients with normal left ventricular function before reoperation, 5- year survival was 90.6% (Fig 2). For those with mild left ventricular dysfunction (limited to one of five left ventricular segments), it was 89%; for those with moderate left ventricular dysfunction (two to three segments impaired), 84%; and for those with severe left ventricular dysfunction, 68.8%. Figure 3 shows the difference in 5- year survival between patients who had complete revascularization at reoperation and those who did not. Figure 4 shows 5-year survival according to the preoperative angiographic indica-

7 386 The Annals of Thoracic Surgery Vol36 No 4 October 1983 tions. The range was tight-between 90.8% survival for patients operated on for graft closure and 87.3% for patients operated on for progressive atherosclerosis only. Comment In geographically dispersed areas, selection trends from the 1970s indicate that initial referrals for operation for mild angina and singlevessel disease are declining and that patients with multivessel disease, especially triple-vessel disease, are increasingly being selected as primary surgical candidates [2]. The number of bypass grafts per patient has doubled in the past 15 years and, consequently, revascularization has become more complete. These trends should produce fewer reoperations during the first 10 postoperative years. More than half of our patients undergoing primary revascularization during 1967 through 1970 had single-vessel disease, and 17% of them underwent reoperation in the next 10 years. Single-vessel disease accounted for 18% of patients having operations in 1971, and the rate of reoperation dropped to 10%. The number of patients in the 1972 subset who had reoperation dropped further to 7%, and single-vessel disease made up 16% of that original subset. Assuming that approximately 7% of patients operated on today will undergo reoperation in the next 10 years, we may expect 12,000 to 14,000 coronary artery reoperations annually nationwide in the 1990s (Appendix). Many patients who required a reoperation had undergone their first coronary bypass at a young age, often in their late 40s or early 50s, which may indicate a predilection for more progressive atherosclerosis. The early reappearance of angina is frequently traced to technical factors, whereas late-onset angina is attributed to new native lesions [3]. We have found that the mean interval between operations has increased gradually from 43 to 77 months. In other words, patients today average 6.5 years between operations. The prevalence of graft failure as a reoperative indication rose from 26% in 1 (1967 to 1976) to 40% in 4 (1981 into 1982), and the interval between operations for this indication increased from 17 to 61 months. This finding implies that graft closure from technical reasons requiring early reoperation is de- clining and the interval is increasing because of closures that occur 5 or more years postoperatively. An increasing number of grafts per patient, which includes grafts of "moderate" lesions [4], has reduced the number of patients who require reoperation for progressive atherosclerosis in previously ungrafted vessels. Our experience indicates that reoperations are less hazardous today than in previous years. Higher risk in reoperation has been attributed to accidents during reentry, adhesions that produce technical difficulties, and longer time on the pump-oxygenator. Today, diffuse coronary atherosclerosis and deteriorated left ventricular function probably produce greater risk than the three technical factors just mentioned. The improvement in results, which includes lower operative mortality, fewer perioperative infarctions, fewer reoperations for bleeding, and improved blood conservation, as shown in Table 6, is not due to selection of easier cases. For example, comparison of the prevalence of triple-vessel disease (57%), stenosis of the left main coronary artery (ll%), and left ventricular impairment (47%) of 1 with the prevalence of these same variables in 4 shows that they increased to 73%, 18%, and 60%, respectively (p < 0.001, p = 0.02, and p < 0.001). Although vein graft atherosclerosis may occur in the first few postoperative years, it is most apt to appear 5 or more years after the first operation. Investigators at the Montreal Heart Institute [5] found that 46% of vein grafts open 6 to 18 months postoperatively showed angiographic evidence of atherosclerotic change 10 to 12 years po-stoperatively. This finding does not include grafts that were completely occluded. Therefore, more than half the vein grafts may become atherosclerotic within 10 years. Vein graft atherosclerosis looms as an important reason for reoperation between years 5 and 10 and beyond, and also poses a treacherous problem for the surgeon [6]. Atherosclerotic veins that have continued antegrade flow are predisposed to embolization, and 2 of the deaths in our series resulted from atherosclerotic emboli producing fatal infarction. Under these conditions, gentle handling and early ligation of the atherosclerotic vein may reduce this risk. In our experience, totally occluded atherosclerotic vein grafts do not emboiize distally.

8 387 Loop et al: Coronary Artery Reoperations Grondin and co-workers [7] categorized atherosclerotic changes visible on angiography into six patterns. Usually two or more exist in the same graft, and the most frequent type in our experience is diffuse irregularity. Focal lesions may occur in the form of filling defects or stenoses, and the surgeon may be tempted to interpose a new vein into the old graft. We caution against this maneuver because the atherosclerotic process is almost always evident throughout the vein and a new graft at or beyond the distal anastomosis of the old one is probably safer in the long run. We emphasize the marked disparity between late results of internal mammary artery grafts and vein grafts. The attrition rate of vein grafts after the first year is approximately 2% per year, but closure is reported to be as high as 55% by 10 to 12 years [5]. Although we have less experience with the internal mammary artery graft, there is no evidence that attrition accelerates after the fifth year, either in our own series or in that of Barner and colleagues [B]. In conclusion, the interval between operations is lengthening and graft closure is increasing in incidence as vein graft failures are documented between postoperative years 5 and 10. Since revascularization is initially more complete in patients with multivessel disease, progressive atherosclerosis in previously ungrafted arteries becomes a less frequent indication for reoperation. Myocardial protection and blood conservation highlight the technical changes resulting in lower reoperation mortality and morbidity. Angina was completely relieved in 65 to 75% of patients 5 years after the first operation; however, only 52% were angina free 5 years after reoperation. This probably can be attributed to more diffuse atherosclerosis and left ventricular asynergy. Graft patency determined in mainly symptomatic patients after reoperation was essentially the same as that found after the first operation. A 5-year survival of 89% was found for the first 750 patients having reoperation and is nearly the same regardless of the angiographic indication for reoperation. Appendix The number of reoperations expected in the 1990s is based on the following data and assumptions. 1. The number of operations per year in the past 8 years is as follows: , , , , , , , , There will be a linear increase in the number of primary operations from 1980 until 1990, and coronary artery operations will be 200,000 in The operative mortality was 3% from 1973 until Since 1978, there has been and will continue to be an operative mortality of 2%. 4. The cumulative survival after primary operation is estimated as follows: 5 years, 90%; 10 years, 80%; 15 years, 65%; and 20 years, 50%. There is a linear decline in survival between each of these time points. 5. The rate of reoperation is 7% at 10 years with a linear increase over time. Among the survivors at 10 years, there is a 10% chance of reoperation in the next 10 years. This also is assumed to increase linearly over time. Based on these estimates, the number of coronary artery reoperations annually in the United States is projected to be: , , , , , , , , , , ,700 References 1. Loop FD, Cosgrove DM, Kramer JR, et al: Late clinical and arteriographic results in 500 coronary artery reoperations. J Thorac Cardiovasc Surg 81:675, Alderman EL, Fisher L, Maynard C, et al: Determinants in coronary surgery in a consecutive patient series from geographically dispersed medical centers (CASS). Circulation 66:Suppl 1:6, Culliford AT, Girdwood RW, Isom OW, et al: Angina following myocardial revascularization: does time of recurrence predict etiology and influence results of operation? J Thorac Cardiovasc Surg 77:889, Cosgrove DM, Loop FD, Saunders CR, et al: Should coronary arteries with less than 50% stenosis be bypassed? J Thorac Cardiovasc Surg 82:520, Campeau L, Enjalbert M, Lesperance J, et al: Atherosclerosis and late closure of aortocoronary saphenous vein grafts: sequential angographic studies at 1 year, 5-7 years, and at years after surgery. Circulation 68:Suppl 2:1, Keon WJ, Heggtveit HA, Leduc J: Perioperative

9 388 The Annals of Thoracic Surgery Vol36 No 4 October 1983 myocardial infarction caused by atheroembolism. J Thorac Cardiovasc Surg 84:849, Grondin CM, Campeau L, Lesperance J, et al: Atherosclerotic changes in coronary vein grafts six years after operation. J Thorac Cardiovasc Surg 77:24, Barner HB, Swartz MT, Mudd JG, Tyras DH: Late patency of the internal mammary artery as a coronary bypass conduit. Ann Thorac Surg 34:408, 1982 Discussion DR. E. STANLEY CRAWFORD (Houston, TX): I congratulate Dr. Loop and his associates for their important and timely study of coronary artery reoperations and admire Dr. Loop's modest and clear presentation. This group assumed leadership in this field many years ago, and the work presented today is good evidence that this position is being well maintained under the guidance of Dr. Loop. My comment only will indicate how much I need to learn from them. It is still difficult or impossible for me to predict accurately the operative arterial findings from angiograms. Consequently, all patients with angina and multiple-vessel involvement are accepted for operation regardless of factors other than recent myocardial infarction, poor general surgical risk, and substantial reduction in left ventricular function. Of 3,125 patients having an operation between January, 1969, and December, 1981, all underwent a bypass procedure. Ninety-seven percent survived and experienced improvement. Redo reoperations were performed in 210 patients during this interval, 144 of whom were in the group of 3,125 patients and were reoperated on an average of little more than 5 years later for recurrent symptoms. The incidence of abnormal left ventricular function in these 144 patients increased from 24 to 44%, and the incidence of previous infarction increased from 43 to 58%. Although 1.9 grafts were inserted per patient at the first operation, 72% of the patients had incomplete revascularization because of diffuse obstruction or small vessels. Although all patients had obstructed grafts at reoperation, grafts were still patent in 52% of the patients, new disease had developed in 44%, and previously existing disease had increased in 20%. One or more grafts were inserted at the second operation, but 52% of the patients remained incompletely revascularized for the same reasons as at the first operation. In fact, of the 8 (5.6%) patients who died, 7 died in the operating room of ineffective grafting for reasons that partially existed at the time of the first operation. Other contributing factors were poor left ventricular function in 7 and occlusion of the left main coronary artery in 4. The overall 5-year survival after the second operation was 82%: 93% among patients with good ventricular function and 68% among those with poor function. I have learned several things from the work of Dr. Loop and his associates and from my own experience: (1) Identify at the first operation those patients who would be poor candidates for a second operation; (2) bypass all involved major branches at the first operation, including those with less than 50% obstruction; (3) replace or bypass the entire graft; (4) employ myocardial protection; (5) have myocardial assist devices available. DR. HASSAN NAJAFI (President): Thank you, Dr. Crawford, for your usual lively discussion and pertinent comments. It is unfortunate that a difficult five-hour operation for coronary artery disease results in two paragraphs of operative notes stating that such and such arteries were bypassed. In Chicago, we also try to "red flag" patients who probably should not be restudied or reoperated on. This comprises a very small group, but it is important to have a point of reference. DR. LOOP: My associates and I appreciate the kind remarks of Dr. Crawford and share his frustrations about the technical problems of diffuse atherosclerosis. His comments about progressive arterial disease coincide with our own experience. I had anticipated that someone might ask about clinical risk factors in this group of patients having reoperation, but time did not permit a detailed summary of our findings. I emphasize that hyperlipidemia was no more prominent in these patients than in those in our primary revascularization experience. Fewer smoked cigarettes before reoperation. One clinical finding bears watching and should be confirmed as other reoperation series are reported. We found twice as many diabetic patients on insulin treatment in the reoperation series (19%). Otherwise, the usual risk factors were no different from those we encountered before the first operation. Our experience shows that effective palliation can be achieved after reoperation in selected patients. As a surgeon, I can tell you that these reoperative procedures are much easier technically than they were even 5 years ago.