Give Online: Help shape patient care for generations to come.
Cleveland Clinic Logo



Request an Appointment



Contact us with Questions

Expand Content

10 year survival

Influence of the Internal-Mammary Artery Graft on 10-Year Survival and Other Cardiac Events

In patients who undergo coronary-artery bypass procedures, the operative technique and the pathoanatomy of the grafted coronary artery are the immediate determinants of graft patency. After the fifth postoperative year, however, the major determinant of long-term graft patency appears to be the type of bypass conduit. The saphenous-vein graft is prone to progressive intimal proliferation. Angiographic studies have shown that a 2 percent per year vein-graft attrition rate from the 1st to the 7th postoperative year increases to approximately 5 percent per year from the 7th to the 12th year. At the end of the first 10postoperative years, serial coronary arteriograms show that only 38 to 45 percent of aortocoronary vein grafts are open and have a normal appearance. Between the 7th and 12th postoperative years, lesions that resemble vein-graft atherosclerosis were seen in nearly half of arteriographic studies, and this incidence did not include atherosclerosis in closed grafts.

Intrinsic vein-graft disease has been correlated with abnormal levels of serum cholesterol and triglycerides, diabetes, elevated low-density lipoprotein. The gradual rise in the risk of sudden death beginning 3 ½ years after saphenous-vein bypass grafting is probably due more to vein-graft closure than to progression of native coronary artery disease, and it portends an accelerating decline in survival rates and an increase in rates of re-operation and other cardiac events by 7 to 10 years after operation.

In contrast, atherosclerosis has been observed rarely in the internal-mammary-artery conduit, and the patency of those grafts in the later years after surgery greatly exceeds that in vein grafts. Several studies have shown that after 7 to 10 years, the patency of internal-mammary-artery grafts is 85 to 95 percent.

The anterior descending coronary artery to which the left internal mammary artery is most frequently grafted has been shown to be a strategically important artery. High-grade proximal stenosis in this vessel is a risk factor for premature death. Acute obstruction of the proximal anterior descending artery is more likely to cause a fatal myocardial infarction than are acute obstructions of any of the other coronary arteries (except the left main artery). These facts suggest that use of the internal mammary artery as the graft to the anterior descending artery may improve long-term survival. To examine this possibility, we performed an observational study.


From among the first 1000 patients undergoing isolated coronaryarte4y bypass grafting each year in The Cleveland Clinic Foundation from January 1,1979, we chose for our study those who had a narrowing of 50 percent or more of the anterior descending artery, with or without other coronary arterial lesions, and who received a vein graft or an internal-mammary-artery graft to the anterior descending artery, with or without other vein grafts,. Because of the large number of cardiac operations and the complexity of large-scale follow-up we routinely studied only the first 1000 consecutive patients who underwent isolated coronary-artery bypass annually. This number provides large and uniform cohorts each year for an analysis.

We excluded patients with Stenosis of 70 percent or more in the left main coronary artery; those who underwent emergency coronary -artery bypass grafting; those who had previous cardiac operations; and those who received bilateral internal-mammary-artery grafts, sequential internal-mammary-artery grafts, internal-mammary-artery grafts to vessels other than the anterior descending artery, or free internal-mammary-artery grafts. Patients who died in the hospital after operation were also excluded; this included 12(0.52 percent) of the 2318 patients who received internal-mammary-artery grafts and 64 (1.73 percent) of the 3689 patients who received vein grafts. Thus, the patients studied included 2306 who had an internal-mammary-artery graft to the anterior descending artery, with or without vein grafts to other coronary arteries, and 3625 who had only vein grafts.

Severe coronary atherosclerosis is defined as luminal narrowing of about 50 percent or more in one or more of the major coronary arteries. Left ventricular functional status is categorized as normal or mild dysfunction (one of five segments of the left ventricular wall with impaired motion), moderate dysfunction (two segments with impaired function). Normal status or mild impairment corresponds with an ejection fraction of 0.50 or more, and moderate or severe impairment with an ejection fraction of less than 0.50. Complete revascularization is defined as grafting to all major coronary arteries (anterior descending, large marginal circumflex, and right coronary) that have sever atherosclerosis.

Follow-up information was obtained between October 1983 and July 1984. Clinical information about survival and subsequent cardiac events was obtained by letters and telephone calls from trained personnel. Although the personnel were not blinded, they followed standard procedures and were unaware of the purpose of the study and, in most instances, of the details of the patient’s operative procedure. If death or cardiac events had occurred, the patient’s physicians were contracted by telephone to document the events and details of any hospitalizations. Of the 5931 patients, 17 could not be contracted for follow-up study. The mean follow-up period was 8.7 years-8.5 for patients with an internal-mammary-artery graft and 8.8 years for patients without such a graft.

Clinical, angiographic, and operative differences between the two groups were tested for statistical significance by t-test and chi-square tests as appropriate. Death (from any cause), post-hospitalization myocardial infarction, hospitalization for cardiac causes, cardiac re-operations, and total cardiac events were described actuarially by the Kaplan-Meier method. The generalized Wilcoxon test was used to examine curve differences. A group of variables that are known to affect long-term survival was forced into each of two Cox models, on of which also included "use of internal mammary artery" as a variable. The relative fit of the two models to the data was examined by a test of maximum likelihood ratios. Also, the relative risk of each cardiac event among patients without an internal-mammary-artery graft, as compared with that among patients with such a graft, was examined by use of the Cox model that included the internal mammary artery as a variable.

The variables entered into both Cox models were age (<50, 50 to 54,55,to 59, and > 60 years); sex; preoperative angina (none, class 1 or 2, or class 3 or 4, classified according to the Canadian grading system except for class 2, which was defined as angina resulting from walking one block); extent of severe coronary atherosclerosis (one-, two-, or three-vessel disease); left ventricular function (normal or mild, or moderate or severe impairment); and year of operation. The distribution of these variables in the two study groups is shown in Table 1.

Among 2306 patients who received an internal-mammary-artery graft and 3625 who received only vein grafts, 855 patients (mean interval, 36 months) and 1445 patients (mean interval, 39 months), respectively, underwent a postoperative cardiac catheterization at some time during the follow-up period. When more than one post-operative study was done, only the results of the most recent one were considered. In both groups, recurring or suspected angina was in the indication for repeat catheterization in 55 percent of patients. In the 45 percent who were asymptomatic, the repeat catheterization was performed routinely, usually between one and two years after the bypass. The data on patency are presented separately for those in whom the information was obtained one year postoperatively and for those who were evaluated annually thereafter. In the comparison of patency rates, the Mantel-Haenszel chi-square procedure was used to test for differences in patency, adjusted for the number of postoperative years.



Ten year actuarial survival (excluding hospital deaths) for patients who received an internal-mammary-artery graft was 86.6 percent, as compared with 75.9 percent for those who has saphenous-vein grafts (univariate analysis, P<0.0001). Among patients with single-vessel (anterior descending artery) coronary artery graft had a 93.4 percent survival rate, as compared with 88.0 percent for those with a saphenous-vein graft (P=0.05) Among patients with two-vessel disease, the 10-year survival for those who received an internal-mammary-artery graft was 90.0 percent, as compared with 79.5 percent for patients receiving only saphenous-vein grafts (P<0.0001). In the group with three-vessel disease, 82.6 percent of those with an internal-mammary-artery graft survived 10 years, as compared with 71.0 percent of the patients with saphenous-vein grafts (P<0.0001).

Among patients with normal to mild impairment of left ventricular function, actuarial 10-year survival was 87.6 percent for those with an internal-mammary-artery graft and 78.5 percent for those with only a vein graft (P<0.0001); in those with moderate to severe left ventricular dysfunction, survival was 76.5 percent and 60.4 percent, respectively (P=0.0002). Among the 2078 men with an internal-mammary-artery graft, survival was 87.0 percent, as compared with 76.4 percent among the 3233 with a vein graft only (P<0.0001). Women with internal-mammary-artery graft (n=217) had a 10-year actuarial survival of 82.7 percent, as compared with 71.3 percent among women with saphenous-vein grafts only (n=386) (P=0.14).

The Cox model that included the internal-mammary-artery graft as a variable explained the data regard to survival better than the model without that graft as a variable (P<0.0001). Absence of the graft increased the risk of death by 1.61 (95 percent confidence interval, 1.41 to 1.85) in the years after hospital discharge.

Cardiac Events

Patients who received internal-mammary artery grafts had higher actuarial 10-year freedom from late myocardial infarctions, hospitalization for cardiac causes, cardiac re-operations, and total cardiac events than did patients with vein grafts only. The Cox model that contained the internal-mammary-artery graft as a variable explained the data regarding each of these events better than did the other model (P<0.0001 in each case). Thus, under the Cox model that included the mammary-artery graft, patients with vein grafts only had 1.41 (95 percent confidence interval, 1.2 to 1.65) times the risk of later myocardial infarction, as compared with those who received an internal-mammary-artery graft; 1.25 (1.13 to 1.38) times the risk of cardiac hospitalization; 2.00 (1.57 to 2.56) times the risk of cardiac re-operations; and 1.27 (1.16 to 1.39) times the risk of all late cardiac events (P<0.0001 for all comparisons).

The univariate analysis revealed no differences in terms of recurrent or persistent angina or hospitalization for arrhythmia between those with the mammary-artery graft and those with the vein graft only. However, when the internal-mammary-artery graft was used as a variable in the Cox model, patients with vein grafts had 1.2 (1.06 to 1.39) times the risk of postoperative angina as did those with an internal –mammary-artery graft P=0.004). Hospitalization for arrhythmia was not different in the two groups, according to univariate analysis (Table 2) or analysis with the Cox model (P=0.9).

Graft Patency

The mean patency rate of the internal-mammary-artery grafts (n=855, all to the anterior descending artery) was 96.0 percent, whereas that of the vein grafts (to the anterior descending artery; n=1445) was 81.1 percent (Mantel-Haenszel, P<0.0001). Among all vein grafts, patency was 76.7 percent. Among saphenous-vein grafts placed in vessels other than the left anterior descending artery, patency was 74.2 percent.


A properly designed randomized trial is the best method for testing the hypothesis that the use of the internal mammary artery as a bypass graft to the anterior descending artery-even in patients with multi-vessel disease-increases the survival rate above that associated with revascularization with vein grafts only. However, no such trial is planned and there are several reasons why such an investigation may not be feasible. First, such a trial would require considerable expense and the results would not be known for at least 10 years.

Moreover, present knowledge about late patency rates would bias the offering of the internal mammary artery and saphenous vein as comparable conduits in a trial. Consequently, an observational study with appropriate statistical control may be the most practical approach for the continued study of what appears to be a major, controllable variable that affects long-term outcome after coronary-artery surgery-i.e., graft patency.

The limitations of an observational study include the possibility that confounding variables may not be recognized or measured. To a much lesser extent, randomized clinical trials may also be subject to subtle patient differences that cannot always be appreciated. Our nonrandomized patient groups showed some differences affected the results was minimized by the method of analysis that used the Cox models. Nevertheless, our conclusions must be viewed in the context of a nonrandomized study with its inherent limitations.

From 1971 through 1979, the personal preferences of the surgeon determined the choice of graft to the anterior descending artery. These decisions about conduits were not related to patent characteristics, except that the mammary-artery graft was generally not used in emergency situations or in patients with severe left main coronary artery disease or extensive brachiocephalic atherosclerosis. The surgeons who participated in the study used both internal mammary-artery and saphenous-vein grafts in their patients. During whose years, we recognized that early patency was higher in internal-mammary-artery grafts, but we did not know whether those grafts improved survival or reduced postoperative cardiac events.

Thus, on the basis of individual patencies of the internal-mammary-artery and saphenous-vein grafts, the skills of the surgeons were judged to be comparable. After 1979, sequential arteriograms disclosed that late patency for the mammary grafts was significantly higher than that for the vein grafts. This information, coupled with our finding of vein-graft atherosclerosis as an increasingly common reason for reoperations, persuaded us that the internal mammary artery might be the preferable graft to the anterior descending artery, and clearly led to preferential use of this conduit. To eliminate this bias from the study described here, we did not include patients operated on after 1979.

This comparative study provides support for the inference that the use of an internal-mammary-artery graft, rather than a saphenous-vein graft, to the anterior descending coronary artery improves long-term survival rate. Since this was a study of the late results of grafting procedures, patients who died in the hospital during the perioperative period were excluded. The very small number of patients lost to follow-up and the length of the follow-up were crucial in our analysis and lessened the possibility that unidentified events biased the results.

Improved survival with the mammary-artery graft was evident at 5 years, and the improvement was even greater at 10 years. In both groups, the mortality rate was more marked in the latter part of the follow-up period, especially in patients with saphenous-vein grafts. Although the patency rate of the mammary-artery grafts remained high, a decrement in patency did occur in the group receiving those grafts, particularly in patients with three-vessel disease. This suggests that vein-graft closure to other coronary arteries or progression of the native vessel disease had had an effect.

During the years of the study, we were meticulous in the procurement of veins from the leg , avoiding clamping or full-thickness handling of the veins, as well over-distension. It is possible that surgeons are even more careful about handling saphenous veins today, but there is no evidence that late closure rates are affected by this extra care. In fact, reported rates of vein-graft atherosclerosis are similar for different methods of preparation.

This analysis and the series described by Okies and others found that fewer late cardiac events occurred when an internal mammary artery than when a saphenous-vein graft was used. Thus, the incidences of later myocardial infarction, reoperation, and death were reduced by use of the mammary-artery graft. The small effect of the use of the internal mammary artery on the return of angina supports the ideas that angina may be generated by any ischemic area of myocardium that is supplied by a closing vein graft, and that preservation of flow to the area of distribution of the anterior descending artery is crucial for extending longevity and preventing major cardiac events.

The lack of effect of the use of the internal-mammary-artery graft on hospitalization for arrhythmia may be related to the probability that this complication is associated primarily with myocardial scarring that is already present a the time of coronary-artery bypass grafting. Why does the internal-mammary-artery graft have a favorable influence on survival and reduce the incidence of serious late morbidity?

Consistently high early and late graft patency is one explanation, but patency may not be synonymous with improved perfusion. Several studies support our contention that an internal-mammary-artery graft to the anterior descending coronary artery supplies adequate perfusion for a protracted period. In patients with one-vessel disease who received either a single internal-mammary-artery or a single vein graft, graded exercise testing has shown comparably improved performance. Also, patients with multi-vessel disease who received one or more internal-mammary-artery grafts alone or in combination with vein grafts had improved exercise tolerance. Regional myocardial perfusion at rest and after isoproterenol infusion has been assessed in internal-mammary-artery and saphenous-vein grafts by measurement of a washout of radioactive xenon in the distribution of the graft.

Comparison flow meter studies of internal-mammary-artery and saphenous-vein grafts that were performed during surgery have demonstrated that flow through the saphenous-vein graft is superior. Sequential arteriograms of internal-mammary-artery grafts, however, frequently-show that those grafts, when anastomosed to coronary arteries, enlarge over time. This information suggests that the internal mammary artery is capable of adapting its size to provide blood flow sufficient to meet the oxygen demands of the myocardium. Most important, in large clinical series, angina relief with internal-mammary-artery grafts has been found to be equal to that provided by saphenous-vein grafts for five to nine years.

The results of coronary-artery surgery must be viewed over the long term. There is a striking difference in the late development of atherosclerosis in bypass conduits between patients with internal-mammary-artery grafts and those with saphenous-vein grafts. It is remarkable that the internal mammary artery, a vessel comparable in size to a coronary artery and subject to the same risk factors and biochemical environment, should have such a low incidence of atherosclerosis up to 10 years after the bypass operation. Yet, the incidence of atherosclerosis in the in situ internal mammary artery is so low that preoperative arteriography is not warranted (an exception might be cases in which candidates for reoperation have had bilateral saphenous-vein ligation and stripping). Perhaps the vasoactive properties of the artery are protective against atherosclerosis (the inert saphenous vein has no such properties).

This important difference in late clinical results between patients who have had an internal-mammary-artery graft to the anterior descending coronary artery and those who have received vein grafts only. The favorable late results of grafting the internal mammary artery to the anterior descending coronary artery are not achieved at an increased perioperative risk, in our experience, no additional wound or pulmonary complications have occurred during hospitalization for this type of grafting. Some surgeons find that anastomosis of the internal mammary artery is more difficult and increases the operating time. Moreover, mobilizing the internal mammary artery from the chest wall and preparing it for anastomosis requires patience and more delicate operative procedures. As surgeons gain personal experience with the conduit, however, their increase in confidence will allow wider application of this type of grafting, and the operating time for the procedure will approach that required for vein grafts.

Better long-term results and less graft atherosclerosis should become strong incentives for surgeons to use this bypass conduit, despite the additional effort required. Although this study provides no information about the use of the internal mammary artery in reoperations, in patients with left main coronary artery disease, or in those who undergo emergency revascularization, the favorable influence of this type of grafting on long-term morbidity and mortality suggests that its application in such patients should be studied. Extension of the use of the internal mammary artery through bilateral grafting has resulted in excellent long-term survival for selected patient subsets. And the use of sequential mammary-artery grafts is currently being evaluated. The therapeutic implications of our study are that most patients undergoing bypass surgery who have a severe stenosis of the anterior descending artery should receive an artery, along with other grafts as indicated.

We are indebted to Dr. Eugene Braunwalt for encouragement; to Drs. John W. Kirklin and Eugene Blackstone for advice and editorial assistance; and to Eric K. Christiansen (Administrator, Department of Thoracic and Cardiovascular Surgery), Emily Wagstaff (Supervisor, Cardiovascular Information Registry), and Carol Slivka (Programmer Analyst, Cardiovascular Information Registry) For their contributions.

For Tables - please see - Reproduced from Loop F, Lytle B, Cosgrove D, Stewart R, Goormastic M, Williams G, Golding L, Gill C, Taylor P, Sheldon W, Proudfit W. Influence of the Internal Mammary Artery Graft on 10-Yearr Survival and Other Cardiac Events, NEJM, Vol 314: 1-6, 1986.

Reproduced from Loop F, Lytle B, Cosgrove D, Stewart R, Goormastic M, Williams G, Golding L, Gill C, Taylor P, Sheldon W, Proudfit W. Influence of the Internal Mammary Artery Graft on 10-Year Survival and Other Cardiac Events, NEJM, Vol 314: 1-6, 1986, with permission from ©The Massachusetts Medical Society. All rights reserved.

This information is provided for education purposes only and is not intended to replace the medical advice of your doctor or health care provider. While we strive to keep our website current, medical practices sometimes change quickly. Please consult your health care provider for advice about a specific medical condition or contact the Cleveland Clinic if you would like an appointment.

Talk to a Nurse: Mon. - Fri., 8:30 a.m. - 4 p.m. (ET)

Call a Heart & Vascular Nurse locally 216.445.9288 or toll-free 866.289.6911.

Schedule an Appointment

Toll-free 800.659.7822