Coronary artery disease accounts for many patient visits to internists. Many of these patients now have bypass grafts. After a variable symptom-free period following bypass surgery, often these patients see their physician because of recrudescent angina. Fortunately, as one set of studies identified the developments that cause these symptoms, other studies provided clues about measures that prevent or minimize graft failure and thus increase the benefits of coronary artery bypass grafting. Therefore, to care for these patients effectively, general practitioners, internists, cardiologists, and cardiac surgeons must be knowledgeable about the findings of these studies. This article reviews bypass graft disease and highlights important aspects of its prevention and management.

Methods

Top Methods Author & Article Info References

A search of MEDLINE from 1968 to 1994 identified articles on saphenous vein and arterial grafts and on repeat coronary operation. Often particularly informative articles helped to focus the search by referencing relevant articles from which a particular theme could be developed. Some institutions, such as the Cleveland Clinic and the Montreal Heart Institute, did many studies of bypass grafts and disease. The former in particular was disproportionately the source of much useful information. Such articles were also a rich source of relevant referenced material.

Saphenous Vein Grafts

The saphenous vein graft is used most commonly because it is relatively plentiful, readily accessed, and easily harvested. Although it also provides adequate flow to the recipient artery, its tendency to occlude is an important drawback.

Saphenous Vein Graft Attrition

Determining an accurate failure rate of saphenous vein grafts was difficult. Among other factors, the observed saphenous vein graft failure rate depends on the reason for restudying the graft. Attrition rates were higher when patients were restudied for symptoms compared with patients restudied solely for a research protocol [5]. In addition, angiography underestimates the severity of lesions in native coronary arteries and especially in vein grafts [6, 7]. Other difficulties I encountered when reviewing the literature included the failure of some investigators to specify whether the occlusion rate was per distal anastomosis, per graft, or per patient [8]. This information is important because saphenous vein grafts may be individual, sequential, or branched Y grafts and are usually placed in multiple arteries in the same patient [8]. Furthermore, in longitudinal studies, not only are consecutive patients not often studied [8], but the same patients are rarely studied successively. Finally, available data on rates of graft occlusion, although not completely irrelevant, may not reflect current experience and the effect of the lessons for prevention revealed by earlier studies. It is difficult to believe, for instance, that within the first month after surgery the graft occlusion rate from thrombosis is still as high as 10% (which was first reported more than a decade ago), despite current knowledge about readily and widely implemented measures that prevent endothelial injury.

The reported attrition rate has varied for the reasons previously noted. The attrition rate is highest, 8% to 12%, soon after coronary artery bypass grafting (that is, within the first 4 weeks) [2, 3, 8-10]. By the end of the first year, 12% to 20% of the saphenous vein grafts have closed [2, 3, 8-10]. The occlusion rate subsequently slows to an annual rate of 2% for the next 4 or 5 years [3]. After 5 years, the occlusion rate doubles to 4% per year, so that by 10 years after coronary artery bypass grafting about 50% of the grafts have closed [3]. However, although most grafts are diseased 10 years after bypass grafting, 70% to 80% of grafts that appear normal or minimally diseased at 5 years remain so 10 years after surgery [2, 11].

Mechanisms of Graft Closure and Histopathologic Findings of Graft Lesions

Three processes cause saphenous vein graft failure. Although these processes are time-dependent and one or the other predominates at particular points, their occurrences overlap. Thrombosis accounts for graft failure within the first month but continues to occur as long as 1 year after coronary artery bypass grafting. Fibrointimal hyperplasia occurs predominantly after 1 month to 5 years. Vein graft atherosclerosis may begin as early as the first year but is fully developed only after about 5 years [3, 12-14].

Thrombosis

Endothelial injury and technical errors predispose veins to thrombosis [3, 12] (Table 1). Certain aspects of saphenous vein harvesting may cause endothelial injury. Apart from the direct physical injury to the vein graft that can occur during harvesting, the pressure used to distend the graft to detect leaks may, if uncontrolled, be as high as 300 mm Hg [12, 13]. Such pressures damage the endothelium. Other important causes of endothelial injury during vein harvesting include ischemia of the vein wall caused by transient loss of luminal blood and vasa vasorum; the nonphysiologic pH of the distending fluid when a fluid other than blood is used to support and dilate the vein before implantation; and exposure to the high and unaccustomed pressures in the arterial circulation [13]. No less important are the effects of various risk factors, including smoking and hypercholesterolemia, that may induce or perpetuate endothelial injury.

Endothelial injury to vein grafts after harvesting include denudation of the intima that can extend into the media. Damage to the endothelium causes endothelial dysfunction, which decreases production of prostacyclin [15] and nitric oxide [16]. Both of these substances inhibit growth and platelet activation, adhesion, and aggregation. For this reason, these type III injuries, which by definition not only denude the endothelium but also disrupt the internal elastic lamina and the media [17], cause platelet activation, adhesion, and aggregation. Subsequent activation of the clotting cascade eventually causes thrombosis.

Technical errors that increase the likelihood of thrombosis include errors in performing the distal anastomosis; excessive or insufficient graft length, which causes kinks and linear tension in the graft; and mismatched sizes of graft and recipient artery, which can compromise flow, promote turbulence, or both.

Low graft flow rate is one of the most important factors Table 1 that increase the likelihood of early graft occlusion [8]. Circumstances that cause a low graft flow rate are small luminal size (< 1.5 mm) of the grafted artery and decreased distal runoff due to severe disease in the recipient artery. The specific artery grafted is also important: The occlusion rate of a saphenous vein graft to the right coronary and circumflex arteries is higher than that for left anterior descending grafts [10]. Other factors that predict early graft failure include endarterectomy of the grafted artery, local atheroma at the arteriotomy site, extension of the arteriotomy into a branch vessel, postoperative smoking, and hyperlipidemia. For sequential grafts, side-to-side anastomoses have higher patency than do end-to-side anastomoses [18], presumably because the former have greater flow. In general, side-to-side and end-to-side anastomoses involving the diagonal and left anterior descending arteries have higher patency rates than do the same types of anastomoses involving other vessels [18]. Nonsequential end-to-side anastomoses to single vessels have higher patency rates than do sequential end-to-side anastomoses to the same position, but rates are similar when the arteries involved in sequential grafts are diagonal and left anterior descending [18].

Fibrointimal Hyperplasia

Fibrointimal hyperplasia is also caused by endothelial injury and invariably occurs to some degree [13, 14, 19]. Again, endothelial dysfunction resulting from injury impairs prostacyclin and nitric oxide production [15, 16] and thus decreases the growth and platelet inhibition attributed to these substances. Subsequent platelet activation may release platelet-derived growth factor and basic fibroblast growth factor, which stimulate the smooth-muscle cells of the media to migrate to the intima and to proliferate and synthesize fibrous tissue, including collagen and proteoglycans [20]. The latter process causes the intimal thickening or hyperplasia that generally results in a 25% decrease in the luminal vessel diameter [8]. Evidence shows that basic fibroblast growth factor released from injured endothelial cells [21] is important in fibrointimal hyperplasia. Recent data from a canine model showed that the density of receptors for this growth factor is increased when the vein graft is distended at 200 mm Hg [22].

Vein Graft Atherosclerosis

Vein graft atherosclerosis usually occurs with fibrointimal hyperplasia and conforms to the same pathogenetic paradigm as arterial atherosclerosis [23, 24], with some variations. Thus, cellular elements, including platelets, macrophages, and smooth muscle cells, are involved. Growth factors released by these elements govern and modulate their interaction and behavior. Lipids, invariably present in lesions of atherosclerosis, are taken up from the blood, processed by macrophages, and deposited in the nascent lesions [23]. The risk factors for saphenous vein graft atherosclerosis are the same as those for native coronary disease and include hyperlipidemia, smoking, and diabetes [25, 26]. However, in some studies, hypertension was not a risk factor for vein graft disease [25], although some investigators implicate it as a risk factor for fibrointimal hyperplasia [27, 28]. Hemodynamic forces such as high turbulence, low flow velocity, and low shear rate also predispose vessels to atheroma formation [29]. For example, the high turbulence at the proximal (aortic) and distal (host) artery anastomoses is thought to be the main reason why these sites are particularly prone to occlusion.

Vein graft atherosclerosis, however, may be different in other important respects from native coronary atherosclerosis [30] (Table 2). Vein graft atherosclerosis tends to be diffuse and concentric, whereas native coronary lesions are more focal, proximal, and eccentric. The fibrous cap of vein graft disease tends to be either completely absent or thin and poorly developed [30, 31]. Foam cells, which are more numerous, and lipid debris are thus exposed to the circulation in vein graft disease. The lesions have a heavy intimal inflammatory infiltrate that may include multinucleate giant cells laden with lipids [31]. Arterial atherosclerosis, on the other hand, tends to be contained within the subendothelium; the fibrous cap is well developed and overlies a center containing cellular elements and lipid debris. Saphenous vein graft lesions are friable and fragile; when disturbed during coronary intervention or reoperation, atherosclerotic debris may occlude the vessel distally. This causes perioperative myocardial infarction, which increases the mortality risk of reoperation for patients with patent but diseased saphenous vein grafts [32, 33]. In contrast, native coronary atherosclerosis is not friable or fragile and thus is less prone to embolism [30]. Because it is rich in lipids, the saphenous vein graft lesion may rupture and become blocked by a thrombus late in its course; this may cause acute coronary syndrome if the thrombus is occlusive or nearly so [34, 35].

Consequences of Saphenous Vein Graft Disease

When sufficiently severe, saphenous vein graft disease blocks flow through the graft either partially or completely and thereby prevents the saphenous vein graft from providing an alternative route of blood supply to the territory served by a stenosed coronary artery. This causes myocardial ischemia that manifests typically as angina pectoris, myocardial infarction, new or worsening heart failure, arrhythmias, or sudden death [36]. The patency rate of saphenous vein grafts in symptomatic patients is lower than the rate in asymptomatic patients [5], an indication that graft failure causes symptoms.

Randomized and nonrandomized studies [37-40] showed that with time and graft attrition, the initial benefits of coronary artery bypass grafting diminish and finally disappear. The survival curves tend to converge for patients treated medically and surgically [37-40]. For example, a recent report from the Veterans Affairs Coronary Surgery Study Group showed that 18 years after coronary artery bypass grafting, the survival benefit of bypass surgery was lost, even for the angiographic high-risk group, and the difference in the mortality rates disappeared for those treated medically and those who had coronary artery bypass grafting [37]. Although this was partly the result of crossover to surgery by patients initially randomized to receive medical treatment, graft attrition was probably the predominant influence.

Determinants of Prognosis of Saphenous Vein Graft Disease

Determinants of the prognosis of saphenous vein graft disease include those related to the graft (that is, graft-specific) and those unrelated to the graft. The most important graft-related determinant of prognosis is probably the age of the saphenous vein graft [30, 41]. A recent study [30] showed that regardless of the severity of saphenous vein graft disease, grafts older than 5 years were associated with higher mortality rates than were grafts that were less than 5 years old. The poor prognosis of late vein graft disease is probably related to the presence of atherosclerosis. The histopathologic features of these lesions make them more likely than native coronary artery disease to cause complications and death [30].

The location of the graft (that is, the host vessel to which the graft is anastomosed) is also a strong determinant of prognosis. Disease of left anterior descending grafts was associated with substantially higher mortality than was disease of the saphenous vein graft associated with either the circumflex or the right coronary artery [30]. This may be explained by the relatively large area consistently supplied by the left anterior descending graft, which is always connected to the main trunk of this vessel. In contrast, other single grafts are usually anastomosed to branches (posterior descending artery and obtuse marginal branches) rather than to the main trunk of the other two major coronary arteries and thus do not usually supply as large an area as the left anterior descending graft.

When the two graft-specific determinants of prognosis occurred together, their effect on prognosis was cumulative and particularly powerful. Thus, if grafts to the left anterior descending artery older than 5 years had more than 50% stenosis, patient survival 2 and 5 years after catheterization was 70% and 50%, respectively [30]. This outcome is worse than for stenosis of the native left anterior descending artery greater than 50%, in which 2-and 5-year survival rates were 97% and 80%, respectively [30]. On the other hand, for grafts less than 5 years old, patients did relatively well regardless of location of disease. Survival at 5 years was 92%, and at 10 years it was 76%. Even for patients with 75% to 99% stenosis of the saphenous vein graft to the left anterior descending artery, survival at 5 years was 94% if the graft age at catheterization was less than 5 years [30].

The other predictors of higher mortality rates in patients with saphenous vein graft disease [30] are unrelated to the presence of the graft and also predict higher mortality rates in patients with native coronary disease [1]. These include advanced age, moderate to severe left ventricular dysfunction, and the presence of left main disease or three-vessel disease. Left ventricular dysfunction and three-vessel or left main disease adversely affect survival rates regardless of graft age [30].

The Internal Mammary Artery

The left and right internal mammary arteries may be used as coronary artery bypass grafts. The left internal mammary artery is used more commonly because its proximity to the left anterior descending coronary artery and diagonal vessels makes it the most convenient arterial graft for the most important coronary artery. Indeed, it was the first coronary bypass graft, used by Vasilii Kolesov in Leningrad in February 1964 [42], 3 years before Favaloro performed the first saphenous vein bypass graft of the coronary circulation [43]. However, in the early days of coronary artery bypass grafting, concern about its lower flow rate compared with the saphenous vein graft [44, 45] and its more technically demanding application made the internal mammary artery somewhat unpopular with surgeons and raised doubts about its future as a graft. Today the internal mammary artery is preferred for coronary artery bypass grafting [46-49] because many studies have documented its advantages over the saphenous vein graft (Table 3). Its patency rate is consistently higher than that of the saphenous vein graft [50-59]: at 10 years, it is 80% higher than that of the saphenous vein graft. The higher patency rate leads to a higher survival rate [54, 60, 61].

The most common angiographic abnormalities in patients with internal mammary artery grafts are progression of lesions or development of new lesions in native vessels not previously grafted. Occlusion of the internal mammary artery is rare; when it does occur, it is not due to the usual causes. Intimal hyperplasia does not usually occur and atherosclerosis is rare [62, 63]. Mechanical or technical factors such as angulation of the anastomotic site, stretching, and other inadvertent injury during surgery usually cause occlusion. A characteristic lesion seen in the internal mammary artery, the distal thread sign or string sign, occurs in 2% of patients [47]. The distal 3 to 5 cm of the graft shows uniformly narrowed lumen to less than 1 mm. Mechanical and physiologic factors thought to contribute to occlusion are stretching of a short graft, ligation of the internal mammary vein, cautery-induced heat injury, hematoma of the pedicle, competitive flow from a grafted coronary artery with subcritical lesions, and severe left ventricular hypertrophy [47].

Attributes of the Internal Mammary Artery That Make It Suitable as a Graft

The internal mammary artery is strikingly resistant to atherosclerosis [62, 63]. In an autopsy study of 215 internal mammary artery segments, Kay and associates [62] found that only 4.2% had more than a 25% decrease in lumen diameter. In no case was lesion severity greater than 50%. The ratio of the thickness of the intima to the thickness of the media is a measure of the propensity to atherosclerosis. Sims [63] compared the values of this ratio obtained for the internal mammary artery with those obtained for the left anterior descending artery. For the left anterior descending artery, which is prone to atherosclerosis, this ratio ranged from 1.0 to 10.0; for the internal mammary artery in patients as old as 50 years, the ratio was less than 0.1 and was never more than 0.6, even by age 70 years. In another study, only 2 of 58 patients had substantial atherosclerosis of the internal mammary artery: One patient had 25% to 50% stenosis and the other had stenosis greater than 50% [64]. The low prevalence of internal mammary artery atherosclerosis was also shown in Finland, which has one of the highest rates of atherosclerosis. Although 45 of 160 (25%) consecutive patients examined at autopsy had coronary atherosclerosis, only 5 of 160 (3.1%) had notable stenosis in the internal mammary artery [65].

Why is the internal mammary artery relatively immune to atherosclerosis? First, available evidence implicates its physiologic character. The internal elastic lamina of the internal mammary artery is intact, and the relative absence of fenestrations is thought to prevent smooth-muscle cells of the media from migrating to the intima in the process of atherogenesis [66, 67]. Second, the endothelial function of the internal mammary artery is enhanced compared with that of veins [16, 17, 68, 69]. This is manifested in greater production of prostacyclin [16, 69] and nitric oxide [17, 69]. Both of these endothelial-derived factors powerfully inhibit growth and platelet aggregation. Furthermore, a recent study [21] showed that, compared with the saphenous vein graft, the internal mammary artery has a lower density of receptors for basic fibroblast growth factor, which is mitogenic for smooth-muscle cells and promotes their migration from the media to the intima during atherogenesis. Furthermore, in response to endothelial injury caused by distention at high pressure, the density of receptors for basic fibroblast growth factor increases much less dramatically in the internal mammary artery than in the saphenous vein graft. All these effects inhibit the process of atherogenesis [68, 69]. To the extent that any substance produced by the graft endothelium is flushed down to the recipient vessel, it seems likely, although unproven, that this protection from atherosclerosis may extend to the recipient artery.

Other properties that make the internal mammary artery suitable as a bypass graft include its small size, which better matches the diameter of coronary arteries than does the saphenous vein; the absence of valves and varicosities; and decreased turbulence of flow through the internal mammary artery compared with veins, presumably as a result of the greater similarity in geometry of the internal mammary and coronary arteries [70]. In addition, the elastin and collagen support of the arterial wall is more suited to arterial pressures and enables the internal mammary artery to autoregulate its flow by changing its size with time to match flow demands [71]. Finally, the internal mammary artery is also suitable because the organ it supplies, the sternum, appears to tolerate partial loss of its blood supply [72] without unacceptable morbidity and mortality.

The internal mammary artery has a relatively thin wall with a lumen-to-outer media diameter less than 350 µ m and vasa vasorum that do not penetrate the media [73-75]. Thus, it is suitable as a free graft because it does not depend on its vasa vasorum for nourishment as its media is thin enough to make diffusion from the lumen adequate for nutritional support. Furthermore, its endothelial function is not impaired when used as a free graft: Prostacyclin production was not substantially different in the free compared with the attached graft [76].

Clinical Benefits of the Internal Mammary Artery Graft

The internal mammary artery is the preferred graft for coronary artery bypass because its patency rate after the procedure is clearly superior to that of the saphenous vein graft Figure 1 [50-57]. The statistics reported depend somewhat on the reason for obtaining the study angiogram; patency rates observed when angiography was done for symptoms tend to be slightly lower than those observed when angiography was required by a research protocol. At 10 years, patency of the internal mammary artery (mainly the left internal mammary artery) is about 90% [58], compared with approximately 50% for the saphenous vein graft [3]. Data available for patency of the internal mammary artery at 15 to 21 years show that this 90% patency is maintained [59]. Although the left internal mammary artery has been studied more extensively, available data show that the patency rate of the right internal mammary artery is similar to that of the left internal mammary artery [53, 77, 78].

View larger version (20K): [in this window] [in a new window]   Figure 1. Patency of internal mammary artery and saphenous vein grafts at 1-year intervals. The number of patients restudied at each interval is noted. (Reprinted with permission [N Engl J Med. 1986; 314:1-6].).

The higher patency rate of the internal mammary artery provides clinical benefits. Foremost among these is a higher survival rate for patients with at least one internal mammary artery graft compared with patients with only vein grafts [46, 54, 56, 57]. The relative survival benefit appears higher in patients at high risk, such as those with left ventricular dysfunction Figure 2 [54]. Although this improvement in survival rate was seen even in single-vessel disease, patients with multivessel disease benefited more than did patients with single-vessel disease [54]. Thus 10-year actuarial survival rates for internal mammary artery recipients compared with patients who received only saphenous vein grafts were 93% compared with 88% (P = 0.05) for single-vessel disease; 90% compared with 80% (P < 0.001) for two-vessel disease; and 83% compared with 71% (P < 0.001) for three-vessel disease [54]. For single-vessel disease involving the left anterior descending coronary artery, the benefit of the internal mammary artery became apparent after 5 years and increased with time: In patients with the internal mammary artery graft compared with those with vein graft, survival rates were, respectively, 95% compared with 91% at 10 years, 88% compared with 77% at 15 years, and 80% compared with 65% at 18 years (P < 0.001) [61]. In keeping with their higher risk status, patients with moderate to severe left ventricular dysfunction derived greater survival benefit from grafting with the internal mammary artery rather than only saphenous veins than did patients with normal or nearly normal left ventricular function Figure 2 [54]. At 10 years, the survival benefits of the former were increased by 27% from 60% to 77% (P < 0.001), whereas those of the latter increased only 12% from 79% to 88% (P < 0.001) [54]. The survival benefit derived from the internal mammary artery has been observed in men and women, in patients older than 65 years and in younger patients [57], and has persisted for as long as 20 years [60].

View larger version (19K): [in this window] [in a new window]   Figure 2. Ten-year survival in relation to left ventricular performance in the two groups of patients, one with at least one internal mammary artery graft and the other with exclusively saphenous vein grafts. The patients receiving internal mammary arteries had significantly higher survival rates in both categories of left ventricular performance, compared with patients who received only vein grafts (patients with normal left ventricular function or mild impairment, P < 0.001; patients with moderate or severe impairment, P < 0.001). IMA equals internal mammary artery. (Reprinted with permission [N Engl J Med. 1986; 314:1-6].).

The survival benefit of the internal mammary artery stems from its other clinical benefits. It improves left ventricular performance [79] and decreases the rate of recurrence of symptoms, myocardial infarction, congestive heart failure, and of reoperation [56, 57, 60].

Indications for Using the Internal Mammary Artery

Whenever possible, the internal mammary artery should be used to revascularize the most important parts of the left ventricle, particularly the anterior wall if it is viable [48]. Therefore, the left internal mammary artery is used more frequently than the right internal mammary artery, usually as an in situ rather than a free graft. Bypass of left anterior descending lesions, especially proximal ones, almost invariably with the left internal mammary artery, is like reinforcing the backbone of the vascular supply of the heart and thereby making the heart better able to withstand vascular insults from other sites. If the anterior wall is already infarcted, it is even more important to ensure that the diseased coronary artery that supplies a larger territory is bypassed with an internal mammary artery or other arterial graft. When reoperation is needed, its risks should be justified by an effort to provide maximum benefit by making every attempt to use an internal mammary artery to revascularize the largest possible area of myocardium.

Because the attached in situ internal mammary artery does not need an aortic anastomosis, the internal mammary artery should be used in patients with calcific atheroma of the ascending aorta, the so-called no-touch aorta [80] where there is concern about embolism of calcific material [48]. Internal mammary artery grafts are also recommended for young patients [48]. Older patients who are at higher risk also derive greater relative benefit from any proven life-prolonging measure. The internal mammary artery is useful when saphenous vein grafts are unavailable or unsuitable or are needed for peripheral vascular surgery [48].

Sometimes the internal mammary artery may be used as a free graft. First, the free graft provides additional length, making a distal coronary artery accessible. This is necessary because the in situ mammary pedicle is too short to reach the distal branches of the right coronary and circumflex arteries and would otherwise have to be stretched, risking stenosis of the internal mammary artery-coronary artery anastomosis. Second, it seems better, when necessary, to use the right internal mammary artery as a free graft because the surgeon can thus avoid crossing the midline, which would be almost unavoidable if the in situ right internal mammary artery were used to provide blood flow to the left anterior descending, diagonal, or upper circumflex branches. An internal mammary artery graft that crosses the midline is vulnerable to injury at reoperation [81, 82]. Using a free right internal mammary artery to revascularize the anterior descending or circumflex arteries decreases the risk for injury at reentry. The free internal mammary artery graft has a higher patency rate than does the saphenous vein graft [81]. For best results, the free internal mammary artery should not be stripped of its surrounding pedicle of tissue [83].

When To Be Cautious in Using the Internal Mammary Artery

The internal mammary artery is not the preferred graft in all clinical circumstances Table 3 [81]. During an emergency, such as surgical rescue after a failed angioplasty, if the patient is unstable or in shock, the surgeon should place a high premium on expeditiously restoring blood flow to the ischemic myocardium. The saphenous vein graft is valuable then. Patients whose left ventricular performance is severely impaired and who have a low cardiac output may be unsuitable for internal mammary artery grafting because they may experience low blood pressure, which would decrease flow in the internal mammary artery [81].

Some surgeons have thought that patients older than 75 years should not receive an internal mammary artery [81], presumably because their longevity is already limited. However, available data indicate that the operative mortality rate is not higher in patients older than 75 years than in those aged 65 to 75 years, even when bilateral internal mammary artery grafts are used [84]. Furthermore, elderly patients who have successful coronary artery bypass grafting live longer than do their age-, sex-, and race-matched counterparts in the general population [85]. Importantly, internal mammary artery grafting improves survival rates of patients 70 years and older. In a study of 723 such patients, Gardner and coworkers [86] reported 4-year survival rates of 86% compared with 77% (P < 0.01) with internal mammary artery and without internal mammary artery, respectively. Age greater than 75 years therefore is not a valid reason to avoid the internal mammary artery for coronary bypass. Besides, such older patients frequently lack suitable saphenous veins or have a heavily calcified aorta, leaving the surgeon no option but to use an internal mammary artery.

Patients who had chest wall irradiation or mastectomy, especially if bilateral, are not candidates for grafting with the internal mammary artery because such patients tend to have smaller and possibly damaged internal mammary arteries that are unsuitable for grafting [81]. Because the flow demand of a severely hypertrophied left ventricle with exceptionally large coronary arteries may be better satisfied, at least in the short term, by saphenous vein grafts, Loop and colleagues [81] advised surgeons to avoid the internal mammary artery in such cases. They recommend that the more expeditious saphenous vein graft be used if the operation planned is extensive, such as coronary artery surgery combined with carotid endarterectomy, ventricular aneurysmectomy, or valve repair or replacement. The internal mammary artery should be avoided if extensive brachiocephalic and subclavian disease with murmurs are present because such lesions compromise flow to the internal mammary artery [81]. Because harvesting the internal mammary artery requires opening the pleura, lung function can deteriorate in persons with lung disease. Therefore, the internal mammary artery should not be used if lung function is already impaired. Finally, the internal mammary artery is not recommended if the surgeon lacks proper experience.

Bilateral Internal Mammary Artery Grafting and Its Potential Advantages

If one internal mammary artery confers clear advantages over vein grafts, do two internal mammary artery grafts increase the benefit that much more? Available data from observational studies suggest that they may. The study by Cameron and colleagues [56, 60] indicates that, compared with patients who received one internal mammary artery graft, patients who received two internal mammary artery grafts were less likely to have angina and myocardial infarction or to need reoperation. These benefits translated into a higher survival rate with two than with one internal mammary artery: 100% compared with 91% at 5 years; 89% compared with 82% at 10 years; and 86% compared with 72% at 14 years in patients who received two internal mammary artery grafts (n = 38) and in all patients who received internal mammary artery grafts (n = 532), respectively [56]. Cumulative survival rates at 20 years were 38% for patients who received vein grafts, 50% for those who received only one internal mammary artery graft, and 63% for the group with bilateral internal mammary artery grafts (P = 0.004) [60]. Other studies [77, 84, 87, 88] echoed these findings. Fiore and associates [77] observed 15-year survival rates of 74% and 59% (P = 0.05) among hospital survivors who received double or single internal mammary artery grafts, respectively. At 15 years, there was also a significant improvement in total event-free survival (32% compared with 18% [P < 0.01]), which included freedom from recurrent myocardial infarction (75% compared with 59% [P < 0.03]) and from recurrent angina (36% compared with 27% [P < 0.03]). Galbut and coworkers [84] showed that patients older than 65 years benefited from bilateral internal mammary artery grafting; the 8-year survival rate was 67.9% compared with 60.7% (P < 0.03) with bilateral and single internal mammary artery grafting, respectively. Green and associates [87] also reported a significantly decreased incidence of angina (10% compared with 32% [P < 0.001]) and decreased incidence of postoperative myocardial infarction (1.4% compared with 5.7% [P < 0.03]) at 5 years with bilateral compared with single internal mammary artery grafting, respectively. However, findings from one randomized study [70] suggested that a second internal mammary artery graft did not provide a significant survival advantage. Because follow-up in that study was for only 4 years, a period that may not have been long enough to detect a difference in survival, final judgment must be reserved. An important consideration may be the risk status of the patients studied. Perhaps bilateral mammary artery grafting would show a survival advantage in patients at higher risk, such as those with moderately to severely impaired left ventricles who urgently need more reliable myocardial blood supplies because they are less likely to withstand further ischemic insults and deterioration of left ventricular function. Of course, these patients would have a higher operative mortality rate. Whether the expected gain in survival would justify such a risk should be addressed in a randomized trial before bilateral internal mammary artery grafting can be recommended for such patients.

Complications of Internal Mammary Artery Grafts

Use of one internal mammary artery should not increase morbidity rates [48], although it was a risk factor for sternal wound infection in one study [89]. Loop and colleagues [48] have, however, noted several drawbacks of internal mammary artery grafting. These include increased operating time, which improves with experience; increased bleeding, which is more problematic with multiple arterial grafts and preventable by careful inspection of the pedicle during and after mobilization; and brachial plexus injury related, perhaps, to excessive opening of the sternotomy retractor and therefore readily preventable [48].

Use of bilateral internal mammary artery grafts increases operative morbidity Table 3 [90-93]. The most severe complication is sternal wound infection thought to be related to the loss of blood flow to the sternum after such operations. However, other risk factors must be present for this risk to be realized [90]. Kouchoukos and colleagues [91] reported that sternal wound infection occurred 3.5 times as frequently with bilateral internal mammary artery grafting as with grafting using only one internal mammary artery. In various reports, risk factors for this complication included obesity [91-93], prolonged mechanical ventilation [91], advanced age [91], increased operating time [89, 92], and diabetes mellitus [89, 90-92]. Diabetes increases the risk up to five times [84, 90, 92, 94]. For example, investigators at Emory University [94] reported that the incidence of wound complications was 9.3% in patients with diabetes compared with 2.5% in those without diabetes (P < 0.001) during bilateral internal mammary artery grafting. Other complications of bilateral internal mammary artery grafting are increased risk for bleeding, increased requirement for blood products, and increased frequency of superficial chest wound infection, sterile dehiscence, and delayed healing [91].

One other drawback of internal mammary artery grafts is the increased susceptibility of the internal mammary artery to injury at reoperation [79, 82, 95]. In one study [96], the presence of an internal mammary artery graft was a risk factor for death at reoperation. However, this finding probably reflected early experience because this risk is now considered manageable [79]. The first time the internal mammary artery is used for grafting, the surgeon can apply various measures to protect the internal mammary artery and reduce the risk for injury during future repeated sternotomy. Such measures include covering the internal mammary artery anteriorly with a pericardial or pleural flap or with polytetrafluoroethylene [79] or polyhydroxybutyrate [97]. An alternative protective maneuver involves opening the adjacent pleural space to provide a partial intrathoracic course for the graft [79].

Finally, use of the internal mammary artery also complicates myocardial protection at reoperation. Retrograde cardioplegia is often necessary to reach and protect areas of myocardium perfused by a patent internal mammary artery [95]. Such areas would otherwise remain warm and prone to ischemic injury. Deep hypothermia with circulatory arrest is another option [95].

Alternative Arterial Grafts

Attributes

Not every artery can serve as an arterial graft [98]. Certain attributes make an artery suitable as a coronary bypass graft. First, the territory the artery perfuses under normal conditions must be such that loss of the artery does not cause unacceptable morbidity. For example, even after both internal mammary arteries are used for coronary artery bypass grafting, the resulting loss of blood flow to the sternum, which they normally supply [72], is only partial [99, 100] and causes-morbidity infrequently [88, 92, 94]. Second, to be used as a graft, an artery must be infrequently susceptible to atherosclerosis [66, 101], a property that depends on an intact internal elastic lamina [73] and on enhanced endothelial production of prostacyclin and endothelium-derived nitric oxide [102, 103].

An artery to be used as an attached in situ graft must be located close to the coronary arteries it will supply and have adequate length to enable anastomosis to the recipient coronary artery without undue tension.

Finally, to serve as a free graft, an artery must derive sufficient nourishment from the lumen alone without relying on the vasa vasorum, which do not function in a free graft. In such arteries, lumen-to-outer media distance is less than 350 µ m and the vasa vasorum are confined to the adventitia, an indication that their role in arterial nourishment is not vital. Otherwise, their nutritional role is indispensable and their loss when the artery is used as a free graft could result in arterial wall ischemia and necrosis [74, 104], which cause graft closure.

Indications

The surgeon may use alternative arterial grafts when no other conduits are available [105], such as after one or more reoperations, with vein stripping, or when internal mammary arteries are unsuitable, as after mastectomy or chest irradiation. In patients with impaired lung function, use of alternative arterial grafts avoids entry into the pleural cavity with risk of further deterioration of pulmonary function [106, 107]. In young patients, the following circumstances may call for use of alternative arterial grafts: hyperlipidemia [105]; need for revascularization of two-vessel disease such that it is wise to preserve an internal mammary artery for possible future use; and when complete revascularization is desired with only arteries [107], thus presumably ensuring longer graft patency.

The Right Gastroepiploic, Inferior Epigastric, and Radial Arteries as Alternative Arterial Grafts

The right gastroepiploic artery and the right and left inferior epigastric arteries are used as alternative arterial grafts, the latter exclusively as a free graft and the former both as an attached in situ graft and as a free graft [108]. Both have the requisite biological properties and therefore are relatively immune to atherosclerosis [73, 102, 103]. Their short-term patency is good, but long-term patency is still unknown [109, 110].

Some of the drawbacks to using the right gastroepiploic artery include the need to enter the peritoneal cavity to harvest it [105], increased harvest time, and a propensity to vasospasm [110, 111]. Furthermore, its vasa vasorum have been shown in one study [98], but not in another study [73], to penetrate the media, a property that would undermine its suitability for use as a free graft.

The inferior epigastric artery is superficial. Harvesting is therefore relatively easy [106, 107]. However, because of its location, surgeons can only use it as a free graft, a role for which it is suited because its vasa vasorum are confined to the adventitia and do not penetrate the media [73, 106]. Its relatively short length restricts its use to restoring blood flow to coronary arteries closest to the aorta. These include the left anterior descending, the diagonal, intermediate branch, and the first obtuse marginal artery [112]. Surgeons may use it when vein grafts and internal mammary arteries are not available [113]. It may also be used when preoperative pulmonary function is already impaired, with the likelihood of postoperative pulmonary complications if the surgeon enters the pleura [106].

Surgeons abandoned the radial artery 20 years ago because of a high closure rate. In retrospect, this appears to be partly related to mishandling during harvesting. Currently, some surgeons are giving the radial artery a second chance after some grafts initially thought to be closed at early postoperative study were discovered open at subsequent study more than 15 years later [114]. Several attributes make the radial artery attractive for bypass grafting [114]. It matches the size of most coronary arteries because it is somewhat larger in diameter than the internal mammary artery. It is also longer than other arterial grafts and can reach all potential targets on the heart's surface. Technically, its thick resilient wall facilitates anastomosis to both the aorta and the recipient coronary artery. Surprisingly, a recent study found that despite its thick wall, its vasa vasorum do not penetrate its media [73], suggesting that they have little or no role in arterial nourishment. This property would be indispensable because surgeons use it exclusively as a free graft. Finally, the radial artery is readily accessed and harvested without the risks incurred by entering a major body cavity. Limited data on short-term patency are encouraging. In one study, patency was 93.5% after a mean of 9.2 months [114]. In another [115], patency of the radial artery graft was 94.1% at 9.4 months, which compared favorably with 100% patency for the left internal mammary artery in the same patients [115]. However, only a favorable long-term patency rate will revive the radial artery as a viable alternative to the internal mammary artery. Should that occur, the radial artery will probably become one of the preferred grafts.

Preventing Saphenous Vein Graft Disease

Surgical Measures

Prevention of coronary artery bypass graft disease must start at surgery because events at surgery, especially the way in which veins are harvested and the type of graft used, alter the risk of coronary artery bypass graft disease. The incidence of atherosclerosis in arterial grafts is much lower than that in vein grafts. Therefore, the most important preventive measure at surgery is use of suitable arterial grafts and, when feasible, multiple arterial grafts. Because use of these arterial grafts is technically more demanding, only persons with the requisite skill and experience should mobilize and implant them, taking care to avoid undue tension on the implanted graft.

Because it is frequently not possible to use more than one arterial graft, for the foreseeable future most patients who have coronary artery bypass grafting will receive primarily vein grafts. To decrease endothelial injury and thereby the risk for graft closure, surgeons should adopt optimal techniques for harvesting vein grafts [13]. Surgeons appear to have their preferred methods to accomplish this, but some general recommendations include careful dissection of the saphenous vein to avoid injury to its wall, ligation of tributaries 2 to 3 mm away from the wall of the vessel, irrigation and gradual distention of the vein at relatively low pressures of no more than 100 mm Hg, perfusion of the vein with heparinized autologous blood [13], and prompt atraumatic implantation of the vein graft because the extent of endothelial injury increases with the delay from harvesting to implantation [17]. Finally, evidence suggests that performing the proximal anastomosis first and subsequently distending the vein through the proximal anastomosis before implantation of the distal anastomosis causes less endothelial injury than would be the case if the anastomoses were performed in reverse order [116]. In addition to these measures, the surgeon should try to match graft size to recipient artery size, avoid kinks and tension during implantation, and minimize the anastomosis of grafts sequentially in the circumflex and right coronary system because these are particularly prone to closure [18].

Medical Measures

Despite the surgeon's best efforts, the endothelium of vein grafts will receive some injury. Thus, whenever vein grafts are used, the surgeon and cardiologist or internist must prescribe drugs to prevent the platelet activation and aggregation that are consequences of endothelial injury. Antiplatelet drugs, including a combination of aspirin and dipyridamole [117-119], aspirin alone [120, 121], and ticlopidine alone [122], administered soon after surgery drastically reduce the occlusion rate of saphenous vein grafts. For instance, Chesebro and associates [118] showed that a regimen of persantine started before surgery and aspirin started 7 hours after surgery reduced the occlusion rate from 10% to 2% at 1 month and from 23% to 11% at 1 year. Another study [121] showed that dipyridamole probably does not reduce the rate of occlusion; in fact, aspirin alone reduced the occlusion rate by as much as a combination of aspirin and persantine. In addition, this study showed that to be effective, antiplatelet therapy had to be started on or before the second day after surgery; and the best results were obtained when treatment began less than 48 hours after surgery [121].

If it is not begun before surgery, antiplatelet therapy appears not to increase bleeding complications, including excessive amount of chest tube drainage, the requirement of blood transfusion, and the need for reoperation to stop bleeding [120]. Ticlopidine started on the second day after surgery decreased the graft occlusion rate from 13.4% to 7.1% (P < 0.05) 10 days after surgery; from 24% to 15% (P < 0.02) after 6 months; and from 26.1% to 15.9% (P < 0.01) after 1 year [122]. There is some indication that the occlusion rate of internal mammary arteries may be affected by antiplatelet therapy. In the randomized study done by Goldman and coworkers [120], the occlusion rate of the internal mammary artery grafts was 0% (0 of 131) in the patients treated with aspirin before surgery compared with 2.4% (3 of 125) (P = 0.081) in the patients who received placebo before surgery. This benefit was not significant and, even if real, seems modest, given the price exacted by increased bleeding when aspirin was started before surgery. Another study [123] showed that addition of dipyridamole to aspirin did not improve the outcome.

Considering that thrombosis is the most important cause of vein graft occlusion within the first year after coronary artery bypass grafting, would anticoagulation reduce the rate of graft occlusion? Two studies [124, 125] examined the effect of warfarin on the rate of graft occlusion. In one, no effect was observed [124]; in the other, the authors found a trend toward some benefit, but the difference observed was not significant [125]. Another study evaluated oral anticoagulants for internal mammary artery grafts and showed a high rate of adverse complications [123]. Therefore, prophylactic anticoagulation with warfarin cannot be recommended.

Hypolipidemic agents may also reduce the occurrence of coronary artery bypass graft disease, especially because increased levels of low-density lipoprotein and apoprotein B and decreased levels of high-density lipoprotein predict an increased rate of disease [26]. The cholesterol-lowering atherosclerosis studies I and II showed that a combination of colestipol and niacin decreased the occurrence of new disease and the progression of established disease in bypass grafts [126, 127]. New lesions in bypass grafts developed in only 16% of treated patients compared with 38% of patients receiving placebo (P = 0.006) [127]. Thus, the study showed that lowering lipid levels conferred some benefit even in patients whose lipid levels were not considered to be high.

Because oxidation of low-density lipoproteins features prominently in the process of atherogenesis [23, 24, 128-130], prevention of such oxidation might be beneficial. So far no one has studied the effect of antioxidants on the occurrence of atherosclerosis in grafts. Recent reports indicating that daily intake of vitamin E was associated with a decreased incidence of atherosclerosis [131, 132] are interesting, but until definitive studies are done, no one can say to what extent such results can be extrapolated to patients with bypass grafts.

Percutaneous Transluminal Angioplasty and Other Percutaneous Interventions

Conventional Balloon Angioplasty

The outcome of conventional balloon angioplasty of saphenous vein grafts is generally worse than that achieved with native vessels and depends partly on the location of the lesion and the age of the graft but mainly on lesion characteristics. Pooled data from 12 studies [133] showed that the initial success rate is slightly higher for lesions located either distally or in the body of the graft than for proximal lesions. The restenosis rate is as high as 58% in the proximal location, 52% in the body of the graft, and 28% in the distal location [133]. In general, the most satisfactory result is obtained in short, focal lesions at a distal site in grafts less than 5 years old. The success rate is more than 90%, with a complication rate of less than 2% and restenosis rate of about 30%. At the other extreme is a chronic total occlusion of an old vein graft, for which the initial success rate is about 50%, the complication rate is more than 10%, and the restenosis rate is more than 60%. In between are long lesions, lesions in grafts older than 4 to 6 years, diffuse vein graft disease, intragraft thrombus, and lesion at a proximal location or in the body of the graft. Although the expected success rate is still more than 90% in these cases, the complication rate is more than 5% and the restenosis rate is at least 45%.

The unsatisfactory result of conventional angioplasty of saphenous vein grafts provided the impetus for performing alternative coronary interventions short of reoperation. These are directional and extraction atherectomy and stenting.

Directional Atherectomy

A recent randomized trial comparing directional vein graft atherectomy with angioplasty showed that the initial success rate (89% compared with 79%) and initial gain in luminal diameter (1.45 compared with 1.12 mm; P < 0.001) were higher with atherectomy than with angioplasty. However, this was achieved at the expense of a higher frequency of complications, especially non-Q-wave myocardial infarction, which occurred at a higher rate with directional atherectomy (16% compared with 9.6%; P = 0.09) because of a higher frequency of distal embolism (13.4% compared with 5.1%; P = 0.012), presumably caused by passing the bulky atherectomy device through the graft and by actively manipulating and debulking the fragile and friable lesion [134]. Largely as a result of this, the composite of adverse end points, including death, myocardial infarction, emergency bypass surgery, and acute closure, was more common with directional atherectomy than with angioplasty (20.9% compared with 12.2%; P = 0.059). Despite the initial favorable result with atherectomy, the restenosis rate at 6 months was similar for both interventions—45.6% for atherectomy compared with 50.5% for angioplasty. Thus, overall directional atherectomy has no clear advantages over conventional balloon angioplasty.

Transluminal Extraction Atherectomy

Transluminal extraction atherectomy for vein grafts appealed to interventional cardiologists because it can aspirate clot and atheromatic debris [135] and thus theoretically would cause fewer embolic events. Limited experience shows that this expectation has not been fulfilled. In one study, emboli formed at distal sites at a rate of about 12% [135], which matches the rate of 7% to 13% noted with directional atherectomy of vein grafts [134, 136-138]. No-reflow and abrupt closure also occurred at rates of 9% and 5%, respectively. Despite the frequent requirement for adjunctive balloon angioplasty to achieve optimal primary angiographic results, the restenosis rate of 69%, which included a 29% rate of total occlusion, is rather high [135]. Thus, the outcome of extraction atherectomy for vein grafts is not superior to that of conventional angioplasty.

Stenting

Stenting with the Palmaz-Schatz coronary and biliary stents is the one percutaneous intervention that achieved the best results in old diseased vein grafts [139]. In the largest series reported, the initial success rate was about 98%. The overall rate of angiographic restenosis (that is, diameter stenosis more than equals 50%) at 3 to 6 months was 17%. This tended to be higher in patients with diabetes (29%) than in those without diabetes (13%; P = 0.07) and in those with lesions 10 mm long (27%) than in those with shorter lesions (15%; P = 0.28). In comparison, de Feyter [135] reported an overall restenosis rate of 42% after conventional angioplasty of vein grafts. In other studies, this rate was as high as 68% [140]. The low residual stenosis and therefore the larger lumen diameter achieved after stenting probably accounts for the low rate of restenosis with these stents [141, 142]. The one drawback of stenting is the elaborate anticoagulation regimen, which not only prolongs hospital stay and increases costs but also increases the rate of vascular complications requiring surgical intervention. However, stenting holds promise as an effective treatment for vein graft atherosclerosis. The Percutaneous Transluminal Coronary Angioplasty Versus Coronary Stenting of De Novo Saphenous Vein Grafts trial, a randomized, multicenter trial now in progress comparing conventional angioplasty with Palmaz-Schatz stenting, will further clarify the latter's role in treating vein graft disease.

Reoperation

Risks and Complications

Repeat operation challenges the technical skills of the surgeon more than primary coronary artery bypass grafting and is associated with a higher risk for operative death and perioperative myocardial infarction [143]. Therefore, the physician should recommend reoperation only after determining that its potential benefits outweigh the risks.

To recommend reoperation wisely, cardiologists and general internists must appreciate its complexities and subtleties (Table 4). It is technically more difficult for several reasons. The heart is encased in scar tissue and adheres to the sternum. The surgeon must therefore dissect it free, but delicately. The dissection is tedious and time-consuming. It exposes the patient to a substantial risk for injury to the heart and, because vessel identification is difficult, to coronary arteries and patent grafts, including the internal mammary artery, especially if the latter crosses the midline [79]. The prolonged period from induction of anesthesia to cardiopulmonary bypass exposes the heart to ischemia, perioperative infarction, and left ventricular dysfunction. This has some implications. First, the physician must not count on reoperation to rescue failed angioplasty in a patient with bypass grafts because it may not be expeditious enough. Second, unstable patients must be stabilized, if necessary with intra-aortic balloon pump, before reoperation.

Effecting cardioplegia is another aspect of reoperation fraught with risks. A patent internal mammary artery may rewarm the heart. The pattern of coronary supply is often unconventional; if grafts are patent, the routes of perfusion are often multiple, making cardioplegia difficult to achieve and associated with the risk for inadequate myocardial protection [143]. Antegrade delivery of cardioplegia through old patent vein grafts may dislodge friable atherosclerotic debris and thrombi, which may form distal emboli and cause ischemia, infarction, and death [143]. Indeed, Gonzalez-Santos and colleagues [33] reported a perioperative mortality rate of 13% during reoperation if the saphenous vein grafts are patent but diseased, as opposed to 3% when saphenous vein grafts are completely occluded and therefore not an embolic risk. Akins and colleagues [96] observed a higher rate of perioperative myocardial infarction for patients with diseased grafts than for patients who did not have diseased grafts: 8.3% compared with 4.5% (P = 0.05), respectively. To reduce these risks, many surgeons deliver cardioplegia retrograde, use deep hypothermia, or both [79, 143, 144]. Retrograde cardioplegia is particularly important if a patent internal mammary artery supplies extensive amount of myocardium [95]. Finally, during reoperation, revascularization may be incomplete if the patient lacks suitable bypass conduits.

Most studies show that the operative mortality and morbidity rates for repeat coronary artery bypass grafting are higher than those for the primary procedure. In their study of the largest group of patients reported, Lytle and colleagues [143] observed an operative mortality rate of 3.4% for reoperation compared with 1% for a first operation. Factors noted in that series to increase the operative mortality risk were presence of left main disease, New York Heart Association functional class III or IV, advanced age, and incomplete revascularization, whereas advanced age, hypertension, and abnormal left ventricular function predicted late death on subsequent follow-up [143]. Akins and colleagues [96] observed that the following factors predicted hospital death: nonelective operation, perioperative myocardial infarction, previous myocardial infarction, and presence of internal mammary artery graft from previous coronary artery bypass grafting. Other investigators [145-147] identified female sex, poor left ventricular function, unstable angina, recent myocardial infarction, ejection fraction of less than 30%, urgent or emergent revascularization, and interval to reoperation longer than 10 years as factors that predict perioperative death.

The rate of perioperative myocardial infarction in the large Cleveland Clinic series was 8% for repeat operation and less than 1% for first operation [143], and this may explain their observation that more deaths were related to myocardial dysfunction. The subsequent rate of symptom return was higher, but the symptoms were milder than those occurring after primary coronary artery bypass grafting [143]. Other complications included damage to the innominate vein, to the right internal mammary artery, and to coronary arteries. Atherosclerosis debris also obstructed some of the bypass grafts [143].

Predictors of Survival Benefit of Reoperation

Who should have reoperation? Without randomized trials, no one can answer this question confidently [148]. Nonetheless, some observational studies identified principles and considerations that help determine who would benefit from reoperation. Because it carries considerably higher morbidity and mortality rates than does primary bypass surgery, reoperation should be offered only to stable patients who are at considerable risk for death without the procedure. Generally, the higher the risk status of the patient, the greater is the likelihood that he or she will benefit from reoperation.

What factors determine risk status? The first important determinant is the age of vein grafts [30, 41]. For grafts older than 5 years, atherosclerosis is prominent. Given the greater propensity of lesions to thrombosis, graft closure, and emboli, it is a more virulent and dangerous process in vein grafts than in native vessels [30]. Thus, risk status is high with grafts older than 5 years [30, 41]. In the observational study of Lytle and colleagues [41], reoperation improved survival in patients with late (after 5 years) saphenous vein graft stenosis even if they had only class I and II angina. It did not improve survival in patients with early (< 5 years) vein graft disease [41]. The importance of late vein graft stenosis is further high-lighted by the finding that the presence of an internal mammary artery improves survival in late vein graft disease. The immunity of the internal mammary artery to atherosclerosis greatly magnifies its importance in altering the patient's risk status after 5 years, when atherosclerosis looms large as a cause of morbidity and death. As a corollary, if graft age is less than 5 years, the internal mammary artery's importance diminishes and risk status tends to be low regardless of the graft type [41].

Another important determinant of risk status is the total size of viable myocardium susceptible to ischemia and infarction. Before the first coronary bypass operation, the number of diseased vessels is a good indication of the size of the jeopardized myocardium and informs the decision to do primary coronary artery bypass grafting [1]. However, after bypass grafting, the normal pattern of coronary blood supply to the heart is disrupted [143]. Any particular segment of the left ventricular myocardium may be perfused wholly or in part by a graft, a nonrevascularized native vessel, or both, either of which could be diseased. Thus, the clinician must consider the patient's complete vascular condition and estimate the size of the myocardium at risk given the geography of the grafts, native coronary vasculature, and the location of lesions. The larger the myocardium at risk, the higher the patient's risk status is and the more readily justified reoperation becomes. However, although the number of diseased vessels is no longer as useful a measure of the amount of jeopardized myocardium as before the first operation, patients who had more severe coronary artery disease before the first operation generally are more likely to have more areas at risk as a result of graft disease. In keeping with this, Lytle and associates [30] showed that patients with left main and three-vessel disease are at higher risk and thus are more likely to benefit from reoperation [41].

Although the pattern of coronary flow is generally disrupted by bypass grafting, the left anterior descending coronary artery is unique in two respects. First, it is consistently the most important of the three major coronary arteries [1]. Second, when it is bypassed, the graft is usually connected to the spine or main trunk of this vessel rather than to branches of the vessel, as is usually true for the circumflex marginal arteries and the posterior descending branch of the dominant artery. Consequently, the graft to the left anterior descending coronary artery most predictably supplies an area of the myocardium larger than that supplied by any other single graft. Thus, the anterior wall with its blood supply from the left anterior descending coronary artery may be considered the backbone of the left ventricle. Therefore, risk status strongly depends on whether the anterior wall is viable.

If the anterior wall has already been infarcted, the patient's risk status is probably high and the additional criteria needed to justify repeat operation are less stringent. If the anterior wall is viable, then the surgeon must determine how secure its blood supply is. The degree to which its blood supply is secure depends on the type and age of graft to the left anterior descending coronary artery, and on whether any part of this vascular system is stenosed. Thus, its blood supply is most secure if the graft to the left anterior descending coronary artery is an internal mammary artery and both graft and recipient native artery are free of lesions; in this case, justifying reoperation is difficult because survival with medical therapy matches that observed with reoperation [41]. Conversely, the blood supply is not secure if a lesion exists in the graft to this vascular system and especially if the graft is a vein older than 5 years [30]. Reoperation is more readily justified in this latter case because it confers a survival benefit [41]. Lytle and colleagues [30, 41] showed that patients with disease of old vein grafts to the left anterior descending artery are a particularly high-risk group and therefore benefit the most from reoperation [41]. In their observational study, the repeated procedure improved survival in patients with stenosis of old (> 5 years) saphenous vein grafts to the left anterior descending artery, particularly if lesion severity was more than 50% (Figure 3). For patients with late (after more than 5 years) stenotic saphenous vein grafts to the left anterior descending artery, survival rates at 2 and 4 years were, respectively, 84% and 74% for the patients who had repeated operation and 76% and 53% for those treated medically (P = 0.004) Figure 3 [41].

View larger version (19K): [in this window] [in a new window]   Figure 3. Survival of patients with late stenoses in saphenous vein grafts to the left anterior descending coronary artery. Patients having reoperation had improved survival rates. (Reprinted with permission from Mosby-Year Book, Inc., and from Lytle WB, Loop FD, Taylor PC, Goormastic M, Stewart RW, Novoa R, McCarthy P, Cosgrove DM. The effect of coronary reoperation on the survival of patients with stenosis in saphenous vein bypass grafts to coronary arteries. J Thorac Cardiovasc Surg. 1993; 105:609.).

The principle that the left anterior descending artery territory is usually the most important area likely to be supplied from a single source must be qualified. First, the size of the anterior wall supplied by a graft to the left anterior descending artery depends on the location of the lesion originally bypassed. If the lesion developed proximally before any branches, the territory is inordinately large and important [1]. If the bypassed lesion occurred after several branches, the territory is smaller and the importance of the left anterior descending graft is commensurately diminished. Second, the pattern of supply from a graft or native vessel or both sometimes increases the importance of another vessel. For instance, after bypass of a strategically placed lesion, such as in the proximal segment of the circumflex or right coronary artery, a single graft could control the blood supply of the entire territory previously fed by a large major coronary artery. This situation might also arise when a graft to an other-wise modest-sized area provides collateral supply to a larger one, thereby increasing its importance.

The other major determinant of risk is the status of left ventricular function: The more severe the left ventricular dysfunction, the higher is the risk status [30, 41]. With severe left ventricular dysfunction, therefore, reoperation is easier to justify, especially if the anterior wall is already infarcted or is still viable but jeopardized and can be revascularized.

Guidelines for Management

General Medical Procedures

In many respects, medical treatment of coronary artery bypass graft disease mirrors that of native coronary artery disease, because both diseases have common manifestations—myocardial ischemia and infarction, heart failure, ventricular arrhythmias, and sudden death. However, some aspects of treatment for coronary artery bypass graft disease relate to the unique presence of a graft and the fact that the exact time of placement of a vein graft, and therefore of the potential for the three kinds of bypass graft disease, is known.

After coronary artery bypass grafting, in the absence of contraindications and regardless of the absence of symptoms, physicians should treat patients with an antiplatelet drug such as aspirin, which decreases the rate of graft closure. All of these patients will also benefit from hypolipidemic agents for secondary prevention [126, 127]. Because lipid levels in such patients are high enough to promote atherosclerosis, lowering the lipid levels would probably be beneficial to the patient even if the lipid levels are not considered high. If angina develops, the patient should be given ß-blockers, nitrates, and calcium channel blockers. If left ventricular dysfunction occurs, an angiotensin-converting enzyme inhibitor is added.

Patients Eligible for Cardiac Catheterization

Cardiac catheterization clarifies the patient's risk status and enables the clinician to determine if the patient will benefit from coronary interventional therapy or repeated operation. Therefore, when symptoms return after coronary artery bypass grafting, cardiac catheterization should be done given any of the following conditions: 1) Angina class is higher than II; 2) time since bypass grafting is 5 or more years; 3) left ventricular dysfunction is present; or 4) left main or three-vessel disease is present. In the absence of these factors, catheterization is appropriate if the patient is older than 59 years; has ST-segment depression, inverted T waves, or both on the electrocardiogram; or has peripheral vascular disease [1]. Otherwise, the physician should refer the patient for catheterization if the result of a stress test, preferably with thallium, is strongly positive [1].

Patients Eligible for Reoperation and for Percutaneous Angioplasty or Other Nonoperative Intervention

The factors that qualify patients for reoperation include the status of left ventricular function, the total size of the myocardium jeopardized by graft and native vessel disease, the condition of the anterior wall (infarcted or viable), graft age (that is, time elapsed since bypass surgery), and the state of perfusion of a viable anterior wall, including type and age of its graft and severity of lesion of the graft to the left anterior descending artery.

In general, reoperation is more readily justified Figure 4 if the patient's risk status is high as a result of severe left ventricular dysfunction; infarcted anterior wall; a viable anterior wall with its perfusion in jeopardy, especially if jeopardized by disease in an old vein graft; extensive viable area of the affected left ventricle; or vein grafts older than 5 years. Conversely, reoperation is more difficult to justify given normal left ventricular function or mild left ventricular dysfunction; a viable anterior wall with secure perfusion, especially if secured by an internal mammary artery; vein graft age of less than 5 years; or a relatively small segment of the left ventricle in jeopardy.

View larger version (38K): [in this window] [in a new window]