N with increased early mortality after coronary artery

Transcription

1 Factors Influencing Long-Term (10-Year to 15=Year) Survival After a Successful Coronary Artery Bypass Operation W. Dudley Johnson, MD, Jerold B. Brenowitz, MD, and Kenneth L. Kayser, MS Milwaukee Heart Surgery Associates, S.C., and St. Mary's Hospital, Milwaukee, Wisconsin The long-term survival of 6,181 consecutive patients who underwent a coronary bypass operation from 1972 to 1986 was evaluated in relation to certain risk factors: age at the time of operation, sex, preoperative left ventricular function, presence of diffuse coronary artery disease, and previous coronary bypass operation. Advanced age markedly reduced the 10-year and 15-year survival. When compared with the general population, however, the survival of patients in the older age groups (over 60 years) exceeded that of the general population at 10 and 15 years. In patients under 55 years of age, the 10-year and 15-year survival is less than that of the general population. Male and female patients of comparable ages had virtually identical 10-year and 15-year survival rates. Severe left ventricular dysfunction, previous coronary bypass operation, and diffuse coronary artery disease requiring coronary artery endarterectomy all adversely affected the long-term survival rates. Within all age groups studied there was a significant benefit in survival with either one or more mammary artery bypass grafts. The data presented provide an important historical control group that can be used in evaluating the long-term results after successful coronary bypass operations. (Ann Tliornc Siirg 1989;48:19-25) umerous patient-related risk factors are associated N with increased early mortality after coronary artery bypass operations. In our own series of over 7,000 patients followed since 1972, we have consistently identified five variables that significantly increase operative mortality [l, 21: advanced age, left ventricular dysfunction, previous coronary artery bypass operation, female sex, and "diffuse" coronary artery disease. Associated illnesses such as diabetes mellitus, renal insufficiency, peripheral or extracranial occlusive vascular disease, and chronic obstructive pulmonary disease, as well as the number of diseased coronary arteries and the presence of left main coronary artery stenosis, are all related to increased early mortality in numerous other studies [> 131. Our data agree with much of this, but we have not been collecting information on many of these variables long enough to define their effect on long-term results. Consequently, this communication will only focus on the 10-year to 15-year results of coronary bypass operation as related to the patients' age at the time of operation, preoperative left ventricular function, sex, the presence of diffuse coronary artery disease, and whether or not previous coronary bypass procedures had been done. In addition, the effects of single or multiple mammary artery bypass grafts will be evaluated. Presented at the Twenty-fourth Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Sep 2628, Address reprint requests to Dr Brenowitz, Milwaukee Heart Surgery Associates, S.C., 2315 N Lake Drive, Suite 1007, Milwaukee, WI Patients and Methods The records of 7,211 consecutive patients who underwent coronary bypass operations from August 1, 1972, through June 30, 1986, were reviewed. For the purpose of this study 450 patients who died within 90 days of operation were considered operative deaths and were excluded. One hundred fifty-three patients with combined valve replacement and coronary artery bypass grafting have been previously reported [l] and were also excluded from this analysis. Finally, 17 patients with unknown birth dates and 410 foreign patients were excluded, leaving a study group of 6,181 patients, 5,189 men and 992 women. Follow-up was obtained through a mail and telephone survey, in addition to a review of the United States National Death Index," and is available for virtually all patients. The patients were divided into seven subgroups based on their age at the time of operation: group A, 593 patients, less than 45 years; group B, 748 patients, 45 to 50 years; group C, 1,117 patients, 50 to 55 years; group D, 1,362 patients, 55 to 60 years; group E, 1,154 patients, 60 to 65 years; group F, 778 patients, 65 to 70 years; and group G, 429 patients, more than 70 years. Our definition of diffuse coronary artery disease, as well as the indication for coronary artery endarterectomy and the techniques employed, have been reported [2, 141. Data were analyzed by the,$ test and the methods of Cutler and Ederer [15]. Life tables for the United States population were obtained from the United States Public Health Service [ 161. Determinations of left ventricular ejection fraction were ' National Death Index, 3700 East-West Highway, Hyattsville, MD by The Society of Thoracic Surgeons /89/$3.50

2 ~~.~ 1% JOHNSON ET AL Ann Thorac Surg 100 gol Yr ".._ r I I I I % Fig 2. Effect of patient's age on long-term surviual for all 6,282 patients in the study. For standard errors, see Table 2. not available for many patients operated on early in the series. Therefore severe left ventricular dysfunction is defined as two or more akinetic wall segments or a left ventricular ejection fraction less than If there is global hypokinesis or the left ventricular ejection fraction is less than 0.20 we assign a double risk factor. Results Age and Sex Age at the time of operation has a profound effect on late survival (Fig 1; Table 1). Ten-year actuarial survival decreased from 78.3% (standard error [SE], 2.2%) in patients under 45 years of age to 60.4% (SE, 2.6%) in patients aged 65 to 70 years. Similarly, the 15-year survival decreased from 64.6% (SE, 4.9%) in patients under 45 years of age to 35.0% (SE, 6.6%) in patients aged 65 to 70 years. Because of these dramatic differences, it is imperative to look at the proportion of patients in various age groups whenever a specific risk factor is evaluated. In evaluating the effect of sex on long-term survival, it was noted that females had a much higher representation in older age groups (Fig 2). However, when we compared sexes in comparable age groupings, survival was virtually identical (Fig 3). Comparison With the United States Population When the survival of our patients was compared with that of age- and sex-matched groups in the general popula- Table 1. Standard Errors for Figure 1" ~ Age Range Follow-up (yr) (Yr) Under M Over a All numbers are given as percentages. NA = not available NA L o 15 c - 10 f!i n r70 Age In Years Fig 2. Distribiition of age groups and sex for all 6,182 patients in the study. tion, some very interesting facts were observed. The thrust of this analysis was to compare survival during the first 5 years after operation with the second 5 years in the same patients. Therefore, we considered only patients operated on more than 5 years ago. In addition, patients with severe left ventricular dysfunction, repeat operation, multiple coronary artery endarterectomies, or female sex were excluded, leaving a cohort of 2,605 men with no additional risk factors; a group that should theoretically have the best long-term results. In the age groups under 55 years, the 5-year survival approximates that of the general population. There is a marked acceleration in annual mortality, however, in the second and third 5-year periods after operation. In the older age groups (over 60 years) the survival of the surgical patients exceeds that of the general population throughout the period observed (Fig 4). Left Ventricular Function and Repeat Bypass Operation Except for age, left ventricular dysfunction at the time of coronary bypass operation is the dominant factor affecting long-term survival. All age groups are equally distributed among the classifications of ventricular function ( p = 0.7). Patients undergoing repeat bypass operation and patients requiring multiple coronary artery endarterectomies were excluded from this analysis. The results, summarized in T b 70 5: 60 4: ' & fa n Men Age at Time of Surgery '10 Year Results Only Women 0 10 Year Survival 15 Year Survival * = 2 1 Stnd. Error Fig 3. Sex, age, and long-term survival for all 6,182 patients in the study.

3 ~ Ann Thorac Surg JOHNSON ET AL 21 Age at Surgery Range ( Mean yr Fig 4. Actitarid siirviual for rrrde patients with no risk factors by age versus Llnited States Public Health Service riilzite niales; 2,605 patients: iiinle, no risk factors (severe ueiitriciilnr dys fii I ict io 11, repeat opera t ion, tzuo or inore coronary endarterectorrries), operated on before yr Years after surgery LLLLLJ 0 5. Study Group 1 \, I\ U.S.P.H.S. White Males \. 3 = f 1 Standard Error Figure 5, are dramatic. Ten-year survival is reduced from 79.1% (SE, 0.8%) in patients with normal or near-normal left ventricular function to 49.0% (SE, 2.1%) in patients with severe left ventricular dysfunction, to 35.1% (SE, 5.5%) in patients with global dysfunction. Figure 6 summarizes the comparison between patients having their primary coronary bypass operation and patients having a repeat operation. Patients with severe left ventricular dysfunction or multiple coronary artery endarterectomies were excluded. The remaining group was homogeneous in age distribution (p = 0.75). Clearly there is a marked disparity in survival after 5 years, with a substantial increase in the annual mortality of the group having repeat operation. Difftlse Coronary Artery Disease Figure 7 summarizes the comparison between those patients with conventional grafts only, those requiring one coronary endarterectomy, and those requiring multiple coronary artery endarterectomies. Patients with severe left ventricular dysfunction and those having repeat coronary bypass operations were excluded. As there was a statistically significant age disproportion in the older age groups (over 65 years) (p < 0.001), with more elderly patients requiring single or multiple coronary artery en- darterectomies, only patients under 65 years were considered. This group was homogeneous in age distribution (p = 0.7). Furthermore, as we have only compiled endarterectomy data since 1978, there is no survival data beyond 10 years for these patients. Up until 10 years, however, there is only a slightly decreased long-term survival in patients with one or more coronary artery endarterectomies compared with those patients with conventional grafts only. In addition, there was no statistically significant difference when patients requiring a single right coronary artery endarterectomy were compared with those requiring a single left coronary artery endarterectomy [2]. Mlzmmary Artery Bypass Grafts When analyzing the effects of mammary artery bypass grafts on long-term survival, the age distribution of patients was not homogeneous (p < 0.001). As expected, a higher percentage of young patients received one or two mammary artery bypass grafts. The patients were therefore divided into three groups: under 50 years, 50 to 60 years, and 60 to 70 years. Patients with severe left ventricular dysfunction and those undergoing repeat bypass operation or multiple coronary artery endarterectomies were excluded. Within each age group there was a signif _ Normal or Mild LV Dysfunction N= =-=-ere LV Dysfunction N=740 bai Hypokinesia,.. EF (0.20 N-114 '1 3 =?: 1 Standard Error r r I I 0 1 ; Fig 5. Efects of left uentricirlar (LV) firnctroiz on survival for 4,372 patients; multiple endarterectoiiiy and repeat bypass operation excluded Primary CAB Surgery N= Repeat CAB Surgery N=626 3 = f 1 Standard Error I -" I 1-0 ; i 10 Fi<g 6. Coinparison of priinary uersus repeat coronary artety bypass (CAB) operation for 4,143 patients 7uith no other risk factors.

4 22 JOHNSON ET AL Ann Thorac Surg n 40- -Multiple Coronary Artery Endarterectomies N=458 bypass operation, female sex, and diffuse coronary disease $ = t 1 Standard Error requiring coronary artery endarterectomy. Insulin-dependent diabetes mellitus is an additional well-documented I I I icant benefit in survival with either one or more mammary artery bypass grafts, although there did not appear to be any increased benefit with multiple mammary artery grafts. Not surprisingly, this benefit became more obvious the longer the patients were followed (Fig 8). A complete analysis of the long-term results of mammary artery grafts is currently underway and is beyond the scope of this paper, but there does not appear to be any increased benefit when the mammary artery is used to bypass the left anterior descending coronary artery (598 patients: 10-year survival, 83.5%; SE, 1.7%; 15-year survival, 62.2%; SE, 5.7%) rather than a diagonal or circumflex marginal branch (846 patients: 10-year survival, 82.1%; SE, 1.5%; 15-year survival, 63.6%; SE, 4.3%) a- 2.$ sa Single Mammary Bypass N=387 i Multiple Mammary Grafts N=173 Vein Grafts Only N=232 t 1 Standard Error A.-I T= I Comment A number of studies have been published that describe the long-term (over 10 years) results after a coronary bypass operation [3, 5, 8, 9, 121. Certainly there is an ongoing need to precisely evaluate survival in the various subgroups of patients that undergo coronary bypass operations; it is important information that can be used to compare the durability of different modes of treatment. In addition, it supplies valuable predictive knowledge to the patients, their families, and referring physicians. For the purpose of this communication, all patients who died within 90 days of coronary bypass operation were considered operative deaths. Although surgical mortality is usually limited to 30 days, modern therapies can often support patients with multiple complications for far longer than that before they die; therefore only patients who survived 90 days were defined as having had a "successful" operation. As all of the risk factors we looked at markedly increased operative mortality, by only looking at the long-term results of 90-day survivors, we have virtually eliminated the effects of operative fatalities on long-term survival curves. Equally important, it must be emphasized that the cumulative effects of operative and long-term deaths are additive for an individual or group of patients. The purpose of this paper, however, is only to look at the effects of preoperative risk factors on the long-term survival after a successful operation a B i._ Single Mammary Bypass N= Multiple Mammary Grafts N= Vein Grafts Only N= $ = r 1 Standard Error."I I lb sa 60 K - n Multiple Mammary Grafts N= Vein Grafts Only N=545 $ = t 1 Standard Error I I I I 0o 1 i C Fig 8. Survival by number of mammary artery bypass grafts for 3,288 patients with no risk factors: (A) age less than 50 years; fb) age 50 to 60 years; fc) age 60 to 70 years. 15

5 Ann Thorac Surg 1989;48:19-25 JOHNSON ET AL 23 A patient s age at the time of coronary bypass operation is one of the most significant predictors of early mortality [2, 6, 10, 13, 141. Not unexpectedly, this is also true for late mortality in the series presented here (see Figs 1, 4) as well as others [5, 91. What was surprising was the markedly increased mortality in the second and third 5-year periods after operation that were observed in the younger age groups (under 55 years). In the older age groups (over 60 years) the survival after coronary bypass operation exceeded that of the general population throughout the period of observation. This is almost identical to the findings of Cosgrove and associates [9] in their 10-year results after primary myocardial revascularization. Of additional importance is the fact that, in the youngest groups, most of the patients had mammary artery bypass grafts (under 45 years: patients, 74%; 45 to 50 years: patients, 77.4%). Also, all of the operations were primary in patients with good left ventricular function and without diffuse coronary artery disease. The addition of any risk factor or multiple risk factors, or the failure to use mammary artery grafts, would make the disparity between the surgical group and the general population even greater. As our operative mortality for this type of patient is only 0.5% [2], the attrition curves for the patient group, early and late mortality combined, would be virtually identical to those seen in Fig 4. Female sex is another well-accepted predictor of increased early mortality after coronary bypass operations [l, 6, 101. In our series, as well as others [12, 131, there is a much higher proportion of women patients in the older age groups (see Fig 2), almost certainly because of the protective effects of estrogen in premenopausal women. Richardson and Cyrus [13] found a significantly lower (p < 0.001) 5-year survival in women. Killen and coworkers [12] found that the absolute survival of female patients in their series was as good as that observed in men; the female patients, however, did not achieve the expected survival pattern of a matched general population that was observed in their male patients. Whereas the absolute survival of women in our series was not as good as that of men, the 10-year and 15-year survival in equivalent age groups was virtually identical (see Fig 3). As the survival of women in the general population is better than men of the same age, there is still an increased risk of late death in female patients, although it is not as great as we had expected. Left ventricular dysfunction profoundly affects operative [l, 2, 6, 7, 101 and late mortality [3, 8, 9, 11, 131. It is commonly the most significant predictor of both early and late death [5]. Our own results (see Fig 5) are so clear-cut that little discussion is required. The 15-year survival decreases from 59% in patients with normal or nearnormal left ventricular function to 28% in patients with severe left ventricular dysfunction. The desirability of doing an operation-when indicated-before serious injury to the left ventricle occurs is obvious. Severe left ventricular dysfunction is not a contraindication to coronary bypass operation, as the long-term survival of many of these patients is abysmal when treated medically; it does, however, substantially increase the operative risk and reduce long-term survival. The operative risk of a patient having a repeat bypass procedure is at least two to three times the risk of a first operation in a similar patient [ In addition, Rutherford and colleagues [ll] found that a prior cardiac operation is a significant predictor of late death ( p < 0.01) when compared with primary operations. Even eliminating the increased operative risk by only considering 90-day survivors, we found a significant and persistently increased late mortality in our patients undergoing repeat operations (see Fig 6). There is no clear explanation for this, but it is probably due in part to the more diffuse nature of the coronary artery disease seen in patients having repeat operations, and the decreased capability to completely revascularize many of these patients [20]. We have recently described our entire experience with coronary artery endarterectomy in the treatment of diffuse coronary artery disease [2, 141. Operative mortality is increased in the presence of diffuse coronary artery disease, whether or not endarterectomy is performed [2, 101; it is even higher in patients requiring multiple endarterectomies. The long-term survival of these patients, as summarized in Figure 7, was therefore better than we had expected, and is further testament to the durability of the techniques we use. The clinical status of the patients and graft patency data have been published [2, 141. Finally, the data presented in Figure 8 confirm the significant and persistent value of mammary artery bypass grafts in improving long-term survival in all age groups up to 70 years old, for periods up to 13 years after operation. The work done at the Cleveland Clinic has proved, virtually beyond doubt, the durability of mammary artery bypass grafts and their beneficial effect [9,21]. Although we have not yet completed a comprehensive study of our patients with mammary artery grafts, two interesting trends are apparent. First, there does not seem to be any additional benefit to multiple mammary artery grafts compared with single mammary artery grafts. Also, there seems to be no difference in the long-term results when a single mammary artery graft was done to the left anterior descending coronary artery (598 patients; 10-year survival, 83.5%; SE, 1.7%; 15-year survival, 62.2%; SE, 5.7%) rather than a diagonal or circumflex marginal branch (846 patients; 10-year survival, 82.1%; SE, 1.5%; 15-year survival, 63.6%; SE, 4.3%). We cannot explain this as yet, but only point it out, as the beneficial effects described by Loop and associates [21] related only to patients with mammary artery grafts to the left anterior descending coronary artery. Possible reasons for these curious observations have to do with our philosophy regarding mammary artery bypass grafts. We routinely use the mammary artery to bypass the most critically obstructed left-sided coronary artery; it is not used in diffusely diseased coronary arteries where endarterectomy is required, nor routinely in vessels less than 1.5 mm in size. Thus a mammary artery is almost always placed into a relatively normal distal coronary artery of large diameter that has a critical proximal obstruction, and that appears to be the most important vessel for a particular patient, whether that is the left anterior descending or a large diagonal or circumflex branch. Having bypassed the most seemingly important

6 24 JOHNSON ET AL Ann Thordc Surg vessel with a mammary artery, combined with the expected long-term patency of over 90% and the excellent long-term survival of these patients, it will probably take substantially larger numbers of patients followed for longer periods of time to demonstrate the expected additional benefit that would be gained by a second mammary artery bypass graft to a less important coronary artery. The data presented here on the long-term survival of patients having a coronary bypass operation and various risk factors that affect that survival provide information on the durability of an important mode of therapy for patients with severe coronary artery disease. There are a number of invasive procedures being developed for patients in whom medical treatment is not indicated. Our work, and that of several other groups, establishes a historical control group that can be used to evaluate the relative merits of other invasive or noninvasive treatments, both short-term and long-term. Perhaps more important, it is obvious from the data presented that any meaningful discussion of long-term results must take into account the preoperative condition of the patient under consideration, particularly age and ventricular function. References Johnson WD, Kayser KL, Pedraza I'M, Brenowitz JB. Combined valve replacement and coronary bypass surgery. Chest 1986;90: Brenowitz JB, Kayser KL, Johnson WD. Results of coronary artery endarterectomy and reconstruction. J Thorac Cardiovasc Surg 1988;95:1-10. Lawrie GM, Morris GC Jr, Calhoon JH, et al. Clinical results of coronary bypass in 500 patients at least 10 years after operation. Circulation 1982;66(Suppl 1):l-5. Lawrie GM, Morris GC Jr, Glaeser DH. Influence of diabetes mellitus on the results of coronary bypass surgery. JAMA 1987;256: Lawrie GM, Morris GC Jr, Baron A, et al. Determinants of survival 10 to 14 years after coronary bypass: analysis of preoperative variables in 1,448 patients. Ann Thorac Surg 1987;44: Kennedy JW, Kaiser GC, Fisher LD, et al. Multivariate discriminant analysis of the clinical and angiographic predictors of operative mortality from the Collaborative Study in Coronary Artery Surgery (CASS). J Thorac Cardiovasc Surg 1980;80: McCormick JR, Schick EC Jr, McCabe CH, et al. Determinants of operative mortality and long-term survival in patients with unstable angina. J Thorac Cardiovasc Surg 1985; 89: Lytle BW, Kramer JR, Golding LR, et al. Young adults with coronary atherosclerosis: 10 year results of surgical myocardial revascularization. J Am Coll Cardiol 1984;4: Cosgrove DM, Loop FD, Lytle BW, et al. Determinants of 10 year survival after primary myocardial revascularization. Ann Surg 1985;202: Wright JG, I'ifarre R, Sullivan HJ, et al. Multivariate discriminant analysis of risk factors for operative mortality following isolated coronary artery bypass graft. Chest 1987;91: Rutherford ID, Whitlock RML, McDonald BW, et al. Multivariate analysis of the long-term results of coronary artery bypass grafting performed during 1976 and Am J Cardiol 1986; Killen DA, Reed WA, Arnold M, et al. Coronary artery bypass in women: long-term survival. Ann Thorac Surg 1982; 34: Richardson JV, Cyrus RJ. Reduced efficacy of coronary artery bypass grafting in women. Ann Thorac Surg 1986;42(Suppl): S Brenowitz JB, Kayser KL, Johnson WD. Triple vessel coronary artery endarterectomy and reconstruction. J Am Coll Cardiol 1988;11: Cutler SJ, Ederer F. Maximum utilization of the life table method in analyzing survival. J Chronic Dis 1958;8: Feinleib M, Hurley PL, Kapantais G. National Center of Health Statistics: vital statistics of the United States. Vol 11, Part A, Mortality, Section 6:lO. Washington, DC: US Government Printing Office, Johnson WD, King JF. Coronary bypass surgery in diabetic patients. In: Scott RC, ed. Clinical cardiology and diabetes. New York, Futura Publishing Co, 1981: Johnson WD, Brenowitz JB, Saedi SF. Coronary artery bypass surgery reoperations. In: Adams AS, Roberts AJ, eds. Cardiac colloquy. Boston: Adams Publishing Groups, 1987: Brenowitz JB, Johnson WD, Kayser KL, et al. Coronary artery bypass grafting for the third time or more. Results of 150 consecutive cases. Circulation 1988;78(Suppl 1): Lytle BW, Loop FD, Cosgrove DM, et al. Fifteen hundred coronary reoperations. Results and determinants of early and late survival. J Thorac Cardiovasc Surg 1987;93: Loop FD, Lytle BW, Cosgrove DM, et al. Influence of the internal-mammary-artery graft on 10-year survival and other cardiac events. N Engl J Med 1986;314:1-6. DISCUSSION DR H. NEWLAND OLDHAM, Jr (Durham, NC): I enjoyed reading the manuscript that was provided by Dr Brenowitz. This report of the long-term results in 6,000 operative patients by the Milwaukee group clearly describes the factors influencing 10-year and 15-year survival after a coronary bypass operation. The Duke Coronary Artery Data Bank contains approximately 8,000 surgical patients, and also has a roughly equivalent group of medical patients. An adjusted Cox model comparison of survival out to 12 years shows a significant increase in long-term survival after surgical treatment in the overall population after adjusting for baseline inequalities. Three major factors are associated with the significant long-term survival improvement with operation. These are the extent of coronary artery disease, poor medical prognosis due to factors such as age or ventricular dysfunction, and a more recent operative date. This improvement in surgical over medical survival is consistent comparing 1970, 1977, and 1984 even though in each successive time period higher-risk patients were accepted for operation. The improvement is quite dramatic in patients with reduced ventricular function. Thus, the larger the medical risk the greater the benefit from operation; or, said another way, the sicker the patient the greater the benefit. This increase in surgical survival has progressed so that by 1984, almost all subgroups have shown an improvement. Expressed as a hazard ratio, by 1984 the 95% confidence limits demonstrate benefit in all groups except single-vessel disease. It is important to stress that the subgroups of patients with reduced long-term surgical survival are not bad surgical patients.

7 Ann Thorac Surg JOHNSON ET AL 25 These are the same patients who, if treated medically, will have an even worse prognosis. Thus, selection of only low-risk patients deprives patients who would achieve the greatest survival benefit from operation. In this context I wonder if Dr Brenowitz would expand his thoughts on the use of internal mammary artery bypass grafts. I was interested to note that single or double mammary artery grafts equally improved survival and that use of an internal mammary artery graft to virtually any coronary artery seemed to provide the same long-term survival benefit. DR BRENOWITZ: The answer is I don t know. We re going to look at that more carefully. These are raw data; it s very surprising. I think that it is clear that in every age group that we studied, from the youngest to the oldest patients, the presence of a mammary artery graft increased survival throughout the entire period of observation, up to 15 years, and the difference between patients receiving mammary artery grafts and patients receiving only vein grafts was more dramatic the longer the follow-up. We have not looked at it carefully enough to know the effect of left ventricular function on those groups, how many of the patients had endarterectomies, or how many patients had redo operations, and I think some of these other factors might play a role. I personally think that if the patient has a large lateral marginal branch or a very large diagonal branch, that is probably a much better artery to put the mammary artery into than a diffusely diseased and small left anterior descending coronary artery. Dr Johnson has been doing that for quite some time. So we tend to put the mammary artery, when we use a single left or right mammary artery, into the best artery available and not necessarily just into the anterior descending. Why two or three mammary bypass grafts don t seem to do better than one I cannot explain. I would like to say two things in closing. First, we hear a lot about angioplasty and other methods of treatment, and people are talking about what they call long-term results. I think Dr Akins was more precise when he talked about intermediate-term results. Certainly 10-year and now 15-year results are available for bypass grafting. Furthermore, it is clear from our data that as time goes on the results seem to deteriorate, although not as much if you use mammary artery bypass grafts. I think that this is important information that can be used to compare the durability of various treatments. And finally, even in the best groups of patients it is obvious that after 5 or 6 years the annual mortality rate increases significantly. Clearly other things are required to try to improve the survival in these patients, such as risk factor modification, diet, exercise, and cessation of smoking.