‘Few things are more distressing to a physician than to stand beside a suffering patient who is anxiously looking to him for that relief of pain which he feels utterly unable to afford. Perhaps there is no class of cases in which such occurrences as this take place frequently as in some kinds of cardiac disease in which angina pectoris forms at once the most prominent and the most painful and distressing symptom.’ (T. Lauder Brunton. On the use of nitrate of amyl in angina pectoris. Lancet, 1867; ii: 97–8.)

In the early 1900s medical treatment of coronary artery disease was well established but of limited value. Efforts were then made to improve the situation by surgery. Three distinct and differing lines of approach were conceived. First, efforts were made to relieve the anginal syndrome by blocking the nervous innervation of the heart. Attempts were made to interrupt the cervical and dorsal sympathetic nervous system by injections of procaine or alcohol. Dorsal laminectomy and section of the posterior nerve roots was used in 1927 by Singer. Subsequently Raney advised surgical division of the upper thoracic preganglionic fibres only, in order to avoid Horner's syndrome and hypersensitivity to adrenaline. These procedures were often successful, though Sir James MacKenzie felt that the anginal syndrome should be preserved in order to restrain the patient from undertaking unduly severe exertion resulting in ventricular fibrillation.

Efforts were then made to reduce cardiac workload by reducing metabolic rate. This had a physiological basis through reducing oxygen requirement in the areas of limited myocardial coronary reserve. Total thyroidectomy was undertaken to produce myxoedema which in turn moderated the cardiovascular response to adrenaline. The scale of this undertaking was subsequently reduced to simple ligation of the superior and inferior thyroid arteries; this preserved the recurrent laryngeal nerves and parathyroid glands which were often damaged during total thyroidectomy. Thyroidectomy was said to provide symptomatic relief in 80 per cent of patients, though they often became physiologically and psychologically vegetative.

Increased understanding of the pathophysiology of angina subsequently led to direct efforts to increase myocardial blood supply. O'Shaughnessy sutured omentum to the exposed surface of the heart, a concept based on the work of Morrison and Wilkey who had demonstrated the omentum's property of vascularization. Beck attempted to accomplish the same type of surface vascularization by mobilizing the pectoralis minor muscle and suturing the free medial end to the surface of the heart. Carter used the medial surface of the left lung. After many of these early attempts it was possible to demonstrate vascular communications between the grafted tissue and surface vessels of the epicardium. Autopsy frequently showed infarction of the deeper myocardial layers with no changes in superficial areas revascularized by such grafts.

Observations of vascularized adhesions developing between the heart and pericardium following pericarditis stimulated the use of pericardial irritants such as powdered talc or bonemeal. Powdered asbestos was eventually shown to be the most effective irritant, causing an aggressive granulomatous reaction and much capillary vascularity between the pericardium and epicardium.

Pratt in 1898 had retrogradely perfused the excised heart of a cat by way of the coronary sinus. In 1937 Gross ligated the coronary sinus in experimental animals in an attempt to cause congestion of the myocardium and enhance intrinsic vascularity. The aim was to stimulate an overdevelopment of the normal communications between the deeper myocardial sinusoids and the cardiac chambers. Incidentally, these experiments showed the superficial coronary sinus system of venous drainage of the heart to be expendable. The technique was then applied to patients without great success, but the concept of retrograde perfusion of the myocardium evolved. Roberts in 1943 suggested that an ischaemic myocardium might be revascularized by anastomosis of a large artery with the coronary sinus. Beck then performed a considerable amount of animal experimentation, finally developing a reliable surgical technique to arterialize the coronary sinus by connecting a vein graft from the aorta. This resulted in retrograde flow of arterial blood through the capillary bed and into the distal coronary arterial tree. (Retrograde instillation of cardioplegia solution through the coronary sinus is now a method of choice for myocardial protection in patients with severe diffuse coronary artery disease.) Soon afterwards Murray described deliberate resection of the atheromatous portion of a left anterior descending coronary artery and its replacement by a vascular graft. He employed distal perfusion of the coronary segment to maintain myocardial viability during the anastomosis. Vineberg then implanted the bleeding end of the left internal mammary artery within the muscle of the left ventricular wall. His belief was that anastomoses would form with other coronary arterial branches. While this method was essentially restricted to the territory of the anterior portion of the left ventricle and intraventricular septum, intervascular communication was demonstrated in over 60 per cent of experimental animals. The Vineberg procedure was the preferred method until the advent of direct myocardial revascularization by aortocoronary bypass grafts.

Saphenous vein aortocoronary bypass grafting, introduced in the late 1960s, offered substantial benefits over previous therapeutic modalities, especially for intractable angina (Fig. 1) 1759. Coronary artery bypass grafting relieved symptoms in most patients with angina unresponsive to maximal medical management. It is now known that 85 per cent of patients with angina will have their symptoms abolished or unequivocally reduced after coronary artery bypass grafting. Unfortunately, early and widespread acceptance of bypass surgery for severe angina was delayed by early clinical trials which compared surgical and medical treatment for less pressing indications. These were designed to allow surgery in any individual who developed problematic angina; a situation which affected 25 to 40 per cent of patients in the medical arms of most major trials leading to cross-over. Another confounding factor was the rapid improvement in both medical and surgical management of coronary disease. Revascularization techniques and myocardial preservation have improved dramatically since the late 1960s, with developments such as cold potassium cardioplegia, oxygenated blood cardioplegia, retrograde cardioplegia, multiple-dose cardioplegia, use of the internal mammary artery as a conduit, and intra-aortic balloon pumping for cardiac support. Medical treatment has improved with the introduction of &bgr;-blockers and calcium channel blockers, the documentation of the efficacy of nitrates, and the recognition of the importance of thrombotic occlusion and thrombolysis in ischaemic heart disease.

The best known co-operative studies were conducted in the 1970s and early 1980s to evaluate the role of coronary bypass in relatively stable patients with angina. The Veterans Administration (VA) Cooperative Study, the Coronary Artery Surgery Study (or CASS), and the European Coronary Surgery Study were all multicentre studies which examined the efficacy of coronary artery bypass grafting compared to medical management in different clinical situations.

The VA co-operative study screened more than 5000 patients from January 1970 to December 1974; 686 were entered into the final phase of the study. Ninety had severe left main coronary disease and 596 had multiple vessel disease. Patients were randomized to medical or surgical management. The majority of all surgical patients received saphenous vein grafts: the operative mortality rate was 5.6 per cent. When patients were stratified according to number of vessels grafted, operative mortality at 30 days was 0 per cent for one vessel, 6.1 per cent for two vessels, and 7.3 per cent for three vessels. Graft angiography, performed in 84 per cent of patients, documented 69 per cent of grafts to be patent at 10 to 15 months.

One of the most important findings of the VA co-operative study was that chronic stable angina in association with high grade left main coronary artery disease is a definite indication for coronary artery bypass grafting and that survival is clearly improved in this group of patients. In the 596 patients with coronary artery lesions in vessels other than the left main coronary artery, no clear short-term (at 36 months) or long-term survival advantage (at 7 or 11 years) was documented. After 36 months, 87 per cent of the medical group and 88 per cent of the surgical group were alive; at 7 years, 70 per cent of the medical group and 77 per cent of the surgical group were alive (see Fig. 2 1760). Nevertheless, a statistically significant difference in survival was found in patients with coronary lesions in locations other than the left main coronary artery who were in two high-risk subgroups: patients with three-vessel coronary artery disease and impaired left ventricular function, and patients with prior myocardial infarction, hypertension, or resting ST depression on electrocardiogram. Only 52 per cent of the medically treated patients with three-vessel disease and impaired ventricular function were alive at 7 years, compared to 76 per cent of those treated surgically; survival was 37 per cent and 50 per cent, respectively at 11 years. Survival in patients with two of the three clinical risk factors (hypertension, previous myocardial infarction, and resting ST depression) was better in those patients treated surgically than in those treated medically at 7 years. Patients who fulfilled criteria for both high risk subgroups derived the greatest benefit from surgery: 76 per cent 7-year survival compared to 36 per cent 7-year survival in those treated medically, and 11-year survivals of 54 per cent and 24 per cent, respectively. Figure 3 1761 shows these data as survival curves.

The Coronary Artery Surgery Study (‘CASS’) was a multicentre North American investigation which enrolled patients from August 1975 to May 1979 to assess the effect of coronary surgery on mortality and selected non-fatal end-points. The registry included 24 959 patients who underwent coronary arteriography during the enrolment period. Thirty-three per cent of patients were enrolled at study centres that did not participate in the final randomized study or were studied in a pilot trial; 16 626 patients were screened for entry into the randomized trial. Subsequently, subjects were excluded if they had normal or minimal disease by angiography, non-operable disease by angiography, Class III or Class IV angina, left main coronary artery disease, prior coronary bypass grafting, congestive heart failure, or were older than 65. After all exclusions, 780 patients were randomized to medical or surgical therapy. Mortality in both groups was low. Annual mortality rates in the surgical arm were 0.7 per cent, 1.0 per cent, and 1.55 per cent for patients with single-, double-, and triple-vessel disease. Mortality in the medical group was remarkably low at 2.4 per cent, 1.2 per cent, and 2.1 per cent, for single-, double-, and triple-vessel disease, respectively. No statistically significant difference in survival between the two groups was noted at 5 years (Figs 4, 5) 1762,1763,1764. However, it is important to note that each year 5 per cent of the medical group underwent coronary artery bypass grafting for amelioration of progressive symptoms and that these patients remained in the medical group for data analysis.

Among 160 patients with ejection fractions of less than 0.5 and three-vessel disease, survival at 5 years was better in those treated surgically, although this was not statistically significant. A subsequent analysis of the same patients at 7 years revealed that patients with triple vessel disease and ejection fractions higher than 0.34 but lower than 0.50 had an improved 7-year survival following surgery compared with medical treatment (84 per cent vs. 70 per cent).

An analysis of non-randomized patients from the CASS trial with either Class I or Class II angina who refused randomization and those patients with Class III or Class IV angina who were excluded from the study by design revealed that non-operatively treated patients with Class III or IV angina, three-vessel disease, and normal left ventricular function had only a 74 per cent 5-year survival, compared to 92 per cent among those treated surgically. A similar comparison of patients with Class III and Class IV angina, three-vessel disease, and abnormal left ventricular function revealed a 5-year survival rate of 82 per cent following surgery and only 52 per cent for those treated medically. In comparison, patients with Class I or II angina and normal left ventricular function had a 5-year survival in excess of 92 per cent following either medical or surgical treatment. These data were interpreted as confirming the importance of surgical treatment in those patients with severe angina.

Another purpose of the CASS study was to examine quality of life. Patients treated surgically had significantly less chest pain, fewer activity limitations, and required less therapy with nitrates and &bgr;-blockers. At 5 years, surgically treated patients had significantly longer treadmill times, less exercise-induced angina, and less ST segment elevation than medically treated patients. However, these improvements in physiological findings were not reflected by an increase in employment or recreational status. Thus, patients who are moderately symptomatic after infarction, or who have chronic stable angina, should be managed medically initially. If symptoms worsen or the patient becomes dissatisfied with his or her lifestyle, coronary artery bypass grafting should be recommended. Interestingly, studies of quality of life and work status performed in Britain demonstrated considerably better rehabilitation statistics.

The third major randomized trial was the European coronary surgery study. This multicentre prospective randomized trial studied 768 men under the age of 65 with mild to moderate angina, 50 per cent or more stenosis of at least two major coronary arteries, and normal left ventricular function. In this study, coronary artery bypass grafting improved survival overall in patients with three vessel disease, and in patients with proximal left anterior descending artery lesions in association with either two- or three-vessel disease (Fig. 6) 1765. Interestingly, patients with left main coronary artery disease treated surgically did not have a statistically significant increase in survival compared to those treated medically. However, the number of patients with left main disease was small (59). Symptoms (including anginal attacks, use of &bgr;-adrenergic blockers and nitrates, and exercise performance) were significantly ameliorated in the surgically treated patients at 5 years. Operative mortality was 3.6 per cent. In patients with three-vessel disease the 5-year survival was 94 per cent following surgery and 82 per cent in those treated medically (Fig. 7) 1766. If an important stenosis of the left anterior descending artery was present, a 5-year survival of 92.7 per cent was achieved following surgery, whereas only 82 per cent of medically treated patients were alive at this time (Fig. 8) 1767. Both findings were statistically significant. In patients with evidence of left ventricular dysfunction and ischaemic ST segment depression of greater than 1.5 mm, surgery improved survival to 91.7 per cent at 5 years compared to 79 per cent survival without surgery. The European study therefore documents significantly improved survival in patients with stenosis of the proximal third of the left anterior descending artery associated with disease in one or two other vessels and in patients with an abnormal electrocardiogram at rest showing ST segment depressions of more than 1.5 mm.

Patients with cardiogenic shock secondary to myocardial infarction are particularly difficult to manage, and these attracted the interest of early medical and surgical teams. Intra-aortic balloon pumping stabilized the condition of many of these patients; aggressive medical management with early angiography and early bypass surgery improved the survival of those who could not be weaned from the pump. By 1973, the Massachusetts General Hospital had seen an improvement from 20 per cent survival at 1 year with intra-aortic balloon pumping alone to 37 per cent survival at 1 year with intra-aortic balloon pump support plus early angiography and coronary artery bypass grafting.

These major and sometimes contradictory landmark studies have produced inevitable controversies surrounding indications for coronary artery bypass grafting. Each study had problems with compensating for important improvements in treatment and cross-over from medical to surgical groups. Differences in the types of patients enrolled make comparisons between the different studies difficult. Nevertheless, broad clinical indications for coronary artery surgery can be agreed upon and have been summarized by Cohen (Table 1) 506.

Subsequent single-centre studies of patients in high risk groups confirmed the substantial benefits of surgery over medical treatment. For patients with triple vessel disease and impaired left ventricular function the difference in survival for medical versus surgical treatment is 53 per cent versus 89 per cent. When survival data are analysed ( Table 2 507 and Table 3 508 it is clear that the subgroup of patients shown in Table 4 509 have a better prognosis with surgical treatment. Angioplasty has recently replaced or delayed surgery, becoming the primary intervention in a proportion of patients.

Stable angina is precipitated by exercise or emotion, subsides with rest or calm, and responds to vasodilator therapy. It may seriously restrict ability to work and quality of life. The decision to recommend coronary artery surgery depends on the likelihood of prolonged symptomatic relief considered unattainable by medical means, the likelihood of prolonged survival, and the balance of operative risk. Age plays little part in selection: surgery is undertaken routinely on patients over the age of 80. Urgent operations for myocardial ischaemia carry a higher mortality than does elective surgery. In the elderly the mortality rate associated with emergency surgery may reach 20 per cent, due to a combination of long-standing diffuse coronary disease and impaired left ventricular function.

Coexisting disease and patient-controlled factors should be introduced in general selection criteria. Smoking adds to the risk of any elective surgical procedure, and surgery should not be offered to patients who refuse to stop smoking. Obesity adversely affects outcome, but weight loss is difficult to achieve in a patient with restricted activity. For those with co-existing medical problems or cancer a life expectancy of more than 3 years would seem a reasonable discriminating factor: patients with lymphoma or colonic cancer with favourable prognosis are suitable for revascularization, whereas those with advanced breast or gastric tumours are unlikely to accrue lasting benefit. Chronic renal dialysis or renal transplantation do not discriminate against coronary surgery, and coronary artery bypass grafting is often performed on dialysis patients in preparation for transplantation. Coronary surgery may also usefully precede or be undertaken in conjunction with surgery for carotid artery disease or abdominal aortic aneurysm. Patients with recently identified pulmonary gastrointestinal or renal malignancies may be candidates for revascularization as a first step to radical surgery or pharmacological control of their disease. The prognosis for the non-cardiac illness must be a guide to selection: patients with haematological malignancies or lymphoma represent a higher surgical risk due to haemorrhagic tendencies or decreased resistance to infections.

Failure of medical treatment to provide effective symptomatic relief and to improve the quality of life remains the primary indication for surgery in chronic stable angina. The importance of coronary anatomy and left ventricular function in selection of patients for revascularization is considered in Table 5 510. Those selected for elective revascularization usually have severe or poorly controlled symptoms and a positive exercise test, or have a positive submaximal exercise test within 3 weeks of myocardial infarction. Such patients frequently have a significant left main coronary artery stenosis or triple vessel obstructive disease with impaired left ventricular function. Patients with two-vessel or single-vessel disease in which the proximal left anterior descending artery is stenosed, and those with obstructed dominant circumflex or right coronary arteries with a large territory at risk should also be considered for surgery or angioplasty. In practice, many patients with single or two vessel disease are first submitted for angioplasty, which offers symptomatic relief for those with anatomically simpler obstructive lesions at reduced cost. Nevertheless, progression of coronary disease together with early failure or late recurrence of angina results in patients with single- or two-vessel disease returning for surgery. The use of elective angioplasty in patients with more straightforward disease has inevitably increased the degree of complexity of surgery, and consequently the operative mortality rate.

Although early studies suggested impairment of left ventricular function as an incremental risk factor in coronary bypass operations, experience has now shown that revascularization can be safely performed in patients with ejection fractions of 15 per cent or less. Factors which significantly complicate elective revascularization include low cardiac output with a cardiac index less than 1.87 and severe pulmonary hypertension with pressures above 55 mmHg. Diffuse distal coronary disease, particularly when calcified, technically complicates the procedure, as does extensive aortic calcification. The significance of congestive heart failure remains debatable, since it is often difficult to differentiate irreparably injured from stunned myocardium capable of improvement after revascularization. In general, coronary bypass is less successful in patients who do not have demonstrable myocardial ischaemia. Patients with intermittent acute ischaemia causing pulmonary oedema may benefit greatly, whereas those with more chronic cardiac failure due to myocardial scarring may not. Dipyridamole thallium scanning may be helpful in separating the different types of chronic myocardial injury. A scheme for decision making in chronic stable angina is shown in Table 6 511.

‘Unstable angina’ is defined as the new appearance of angina or the change in a previously stable pattern of angina culminating in angina at rest or minimal activity. Significant left main stem disease is seen in 12 to 15 per cent of these patients, and two-thirds have left main or significant triple vessel disease with large areas of myocardium at risk. Many patients are already on triple medical therapy. The remaining third, mostly with new onset angina, have single- or double-vessel disease that may be treated with angioplasty. Whatever the distribution of disease and degree of impairment of left ventricular function it is advisable to stabilize the patient's condition by intravenous nitrate therapy together with systemic heparinization before proceeding to cardiac catheterization and either angioplasty or surgery (Table 7) 512. We consider that all such patients should undergo cardiac catheterization before hospital discharge, even if they become stable and do not require urgent surgery. In the Oxford practice, two-thirds of all revascularization procedures are performed on an urgent basis for unstable angina. &bgr;-Blockade or calcium channel blockers and nitrates should be administered until the time of operation. Administration of heparin allows aspirin to be discontinued, a consideration which has important implications for platelet function and perioperative bleeding. Since many patients suffer recurrent ischaemia on discontinuation of heparin, it should be continued until admission to the anaesthetic room. The vogue for intra-aortic balloon pumping in patients with unstable angina has passed, and it is now unusual to use the balloon pump for this indication. However, the balloon pump may be life-saving for patients with haemodynamic instability where surgery may be delayed for logistic reasons.

Acute transmural myocardial infarction usually occurs after complete coronary thrombotic occlusion. The process evolves by recruitment of ischaemic myocardial cell populations, its final extent depending on the degree of collateral circulation. Most deaths are due to left ventricular failure or dysrhythmia from an ischaemic focus. Early restoration of coronary flow limits the extent of myocardial necrosis and reduces the number of deaths from left ventricular failure. In the mid-1970s urgent coronary artery bypass grafting was undertaken if the patient could be investigated within 6 h of the onset of pain. Intra-aortic balloon pump support was used preoperatively, particularly in those with cardiogenic shock with a left ventricular end-diastolic pressure below 25 mmHg, cardiac index below 2.0, and ejection fraction of less than 30 per cent. Protagonists of emergency revascularization in acute infarction have achieved a 2 per cent hospital mortality rate (Table 8) 513 compared with 11.5 per cent for conventional therapy (when surgery was undertaken within 6 h of symptomatic onset). The mortality rates at 3 years were 6 per cent for surgically treated patients and 20.5 per cent for medically managed patients. When patients with cardiogenic shock were excluded from analysis, surgical mortality was 1.2 per cent, and that in the medical group 9.3 per cent. Acute anterior myocardial infarction carries a higher mortality risk than inferior infarction, but this was reduced from 16.5 per cent to 7 per cent by surgery. Patients revascularized within 5 h showed significantly better ejection fractions than those revascularized at 10 h: late revascularization failed to improve global ejection fraction significantly. Despite its initial success, emergency revascularization following acute myocardial infarction has practically disappeared with the evolution of pharmacological thrombolysis, together with angioplasty techniques. A combination of these methods can provide early non-surgical revascularization with similar effects on limitation of infarct size and outcome.

Patients with recurrent anginal pain after acute myocardial infarction have threatened extension of the area of injury or new infarction in separate areas with compromised blood flow. In general, surgery is avoided within the first 48 h of infarction, during which time the myocardium is stunned. However, continued instability may dictate urgent intervention. Operative risk depends on the size of the original infarction, degree of left ventricular failure, and temporal proximity to infarction. When continued ischaemia or haemodynamic instability necessitate surgery soon after an infarction, intra-aortic balloon pump counterpulsation is of great value.

Acute myocardial ischaemia during or after percutaneous transluminal coronary angioplasty may necessitate urgent surgical revascularization. Both clinical and ECG findings of ischaemia referrable to the dilated vessel are criteria for intensive medical therapy, including vasodilators and calcium channel blockers. Heparin or thrombolytic agents are frequently used prior to further attempts at dilatation. Resolution of pain and elevated ST/T wave segments may follow medical treatment or further percutaneous transluminal angioplasty, in which case the patient can be returned to the coronary care unit, maintained on continuous intravenous heparin. Continuing pain and ST wave changes despite medical therapy and repeat catheter manipulation requires urgent coronary bypass surgery ( Fig. 9(a 1768, b)). If an operating theatre is unavailable for logistic reasons the intra-aortic balloon pump can be used. In some units emergency portable cardiopulmonary bypass facilities are available and are initiated by percutaneous placement of catheters in the femoral artery and vein.

Reversed segments of autogenous saphenous vein have been the mainstay of coronary artery surgery since its conception (Fig. 10) 1769. Patients who have undergone bilateral stripping of varicose veins are at a serious disadvantage since the short saphenous vein is difficult to harvest without resort to the prone position, which is clearly dangerous in patients with unstable symptoms. Veins from the arm are thin-walled and of different character; their use is complicated by high early occlusion rates. Sequential vein grafting by leading one or two lengths of vein across multiple coronary branches may compensate for restricted availability of good quality saphenous vein (Fig. 11) 1770. Before vein harvest the legs should be carefully inspected for evidence of previous surgery, deep vein thrombosis, or varicosities: the long saphenous vein is often of satisfactory quality even in the presence of superficial varicosities. Excess saphenous vein from donors of the same blood group, either preserved in antibiotic solutions or freeze-dried, has been used, but the technique was abandoned after proven acute rejection and thrombosis of vein grafts in one patient. Factors which influence the long-term patency of saphenous vein grafts are shown in Table 10 515. Perhaps the single most important factor is the extent of disease in the native coronary vessel and the ‘run off’ of blood from the site of anastomosis. After documentation of greatly improved patency rates following the use of the internal mammary artery rather than the saphenous vein most surgeons prefer to use a single or both mammary arteries for the great majority of patients ( Table 11 516, Fig. 12 1771). Although sequential grafts can be performed with mammary arteries, a single graft to the left anterior descending coronary is the most frequently employed technique ( Fig. 13 1772 (a, b)). The left internal mammary artery can be used to graft intermediate or circumflex branches and, if necessary, the right internal mammary artery can be brought through the transverse sinus to the circumflex territory. The right internal mammary artery can be used to graft the right coronary proximal to the acute margin of the heart (Fig. 14) 1773. If necessary, either internal mammary artery can be detached from the subclavian artery and used as a free graft. The impressive patency of arterial grafts has encouraged the use of free radial artery grafts and mobilization of the gastroepiploic artery, particularly in repeat coronary surgery. Synthetic conduits such as Gore-Tex, and biological grafts including bovine umbilical and carotid arteries, have been employed when autogenous conduits are not available. In general these have proved unsatisfactory with high early rates of occlusion. The majority of coronary operations are undertaken with a combination of a single internal mammary artery and saphenous vein. The influence of an internal mammary graft on survival has been shown by the Cleveland Clinic group (Table 12) 517. The effect of aspirin and dipyridamole treatment on vein graft patency is shown in Table 13 518.

Favaloro initially described coronary bypass grafting, using saphenous vein harvested from the lower extremity in 1969. The technique involved partial occlusion of the ascending aorta and performance of an end-to-side vein to aortic anastomosis. Distally, an end-to-side or an end-to-end anastomosis was performed using interrupted sutures. All operations were performed with the patient heparinized, cannulated, and on cardiopulmonary bypass through a median sternotomy. Femoral cannulation was typically used for arterial inflow. Favaloro emphasized the importance of accurate cineangiography for successful coronary surgery: this was developed at the Cleveland Clinic by Sones. Many variations of the technique have now been developed. Coronary revascularization is most frequently performed using hypothermic cardioplegic arrest, with distal anastomoses preceding proximal anastomoses. Alternatively, proximal anastomoses can be performed first during cooling. Other surgeons still prefer hypothermic fibrillatory arrest, and recently, limited enthusiasm has developed for warm cardioplegic revascularization.

This technique is widely used. Although many variations exist, the following is a brief description of a representative technique. The anatomy and nomenclature of the coronary arteries is shown in Fig. 15 1774.

The heart is exposed through a median sternotomy and the pericardium is opened to expose the heart and great vessels. The heart is inspected and the quality and calibre of the coronary arteries are assessed. The left internal mammary artery is harvested from the anterior chest wall, dividing all intercostal branches. For maximal length, this artery is harvested from the level of the subclavian vein to its distal bifurcation into the musculophrenic and superior epigastric arteries (Fig. 12) 1771. This vessel is used routinely in all except the most unstable patients, for whom concerns about poor early flow rates due to spasm may mitigate against its use. Papaverine is used to dilate the vessel. The right internal mammary artery can be similarly harvested for use as a pedicled or free graft. Concomitantly, saphenous vein is harvested from the lower extremities. After the vein has been meticulously excised it is carefully distended with heparinized blood, and any remaining open branches are ligated. Proximal and distal ends of the vein are carefully identified: the vein must be placed in reversed fashion so that venous valve do not obstruct flow.

Heparin is administered systemically and the activated clotting time (ACT) is monitored throughout the remainder of the operation to ensure adequate anticoagulation is achieved (ACT >400 s, twice control levels). Additional heparin is administered as indicated. Most surgeons cannulate the ascending aorta at the base of the innominate artery for arterial inflow. A single two-stage venous cannula is placed in the right atrium. Proximal openings drain the right atrium and superior vena cava; distal openings drain the inferior vena cava. Alternatively, the venae cavae can be separately cannulated via the right atrium. If antegrade cardioplegia is used (as described in Section 34.2 246), a smaller calibre cannula is placed in the ascending aorta at a planned proximal vein graft anastomotic site. Some surgeons prevent left ventricular distension by using a ‘vent’ to drain blood away through cannulae placed directly into the left ventricle through the left ventricular apex, or through the right superior pulmonary vein or pulmonary artery.

After all cannulae have been satisfactorily placed, systemic cooling is begun by cooling the pump perfusate. This provides added protection by minimizing rewarming via collateral flow, and lengthens the time available for repairs should any cardiopulmonary bypass circuit accident occur. In Oxford we limit cooling to 32°C and ensure thorough rewarming before discontinuing bypass.

As soon as systemic cooling is started the aorta is cross-clamped proximal to the aortic cannula. Cold cardioplegic solution is administered—via the ascending aortic root cannula for antegrade cardioplegia or via a coronary sinus cannula for retrograde cardioplegia. Either technique brings about the desired hypothermic diastolic cardioplegic arrest. We also employ cold saline solution for topical cooling, although temporary left phrenic nerve palsy may result.

Once satisfactory cardioplegic arrest occurs, the desired distal anastomoses are performed. A coronary arteriotomy is made distal to lesions identified by cineangiography (Fig. 16) 1775. Distal anastomoses are typically performed with a running, non-absorbable suture of polypropylene, although some surgeons prefer interrupted sutures, especially at the heel of the anastomosis. Additional cardioplegia can be administered via newly placed grafts, protecting the myocardium beyond the coronary occlusion. When the internal mammary artery is used as a pedicled graft, this anastomosis is performed last. The native coronary is often severely and diffusely diseased, and no obvious position is found to perform the anastomosis ( Fig. 17 1776 (a)). Under these circumstances an endarterectomy can be performed to remove the atheromatous core. The right coronary artery is usually opened and endarterectomized from its bifurcation into posterior descending and left ventricular branches ( Fig. 17 1777 (b)). Endarectomy in the left anterior descending coronary is avoided wherever possible since the septal branches shear off and perioperative myocardial infarction may result.

Systemic rewarming is begun during completion of the final distal anastomosis; the aortic cross-clamp is then removed promptly. A partially occluding clamp is applied to the ascending aorta (Fig. 18) 1778 and proximal anastomoses are sewn using a running non-absorbable suture (Fig. 19) 1779. On removal of the partially occluding clamp after the completion of the proximal anastomoses, air bubbles are meticulously evacuated from vein grafts to avoid coronary artery air embolism. As systemic rewarming progresses, sinus rhythm returns spontaneously or is reinstituted using direct current cardioversion.

The patient is then weaned from the cardiopulmonary bypass circuit, allowing the heart to fill slowly but progressively, and to resume cardiac ejection. Once the surgeon is satisfied with cardiac performance without cardiopulmonary bypass support, all cannulae are removed and protamine is administered to reverse systemic heparinization. Temporary pacing wires are attached to the ventricle and/or the atrium: temporary abnormalities in conduction are frequent in the early hours after cardiopulmonary bypass. The pericardial and chest cavities are drained with large thoracostomy tubes connected to 20 cm of negative pressure. The sternum is closed with interrupted steel wires (Fig. 20) 1780.

Before the institution of cardiopulmonary bypass a partially occluding clamp is placed on the ascending aorta, and a maximum of two or three proximal venous anastomoses are performed (Fig. 21) 1781. During completion of the final proximal anastomosis, cardiopulmonary bypass is initiated with systemic cooling. As with the previous technique, venting of the left ventricle may be employed via the right superior pulmonary vein, left ventricular apex, pulmonary artery, or the ascending aorta. The aorta is cross-clamped and cold diastolic hypothermic cardioplegic arrest is achieved by infusing cold cardioplegic solution proximal to the aortic cross-clamp. The coronary artery with the most severe lesion(s) is grafted first. This technique allows subsequent doses of cardioplegia administered via the ascending aortic cardioplegia cannula to reach myocardium beyond coronary artery occlusions. If the internal mammary artery is to be used as a pedicled graft, this anastomosis is performed last (Fig. 22) 1782. Systemic rewarming is begun early and the aortic cross-clamp removed promptly after the last distal anastomosis. Care is taken to avoid embolization of air down the saphenous vein grafts.

This technique is used by a minority of surgeons. Cannulation of the ascending aorta is performed after the patient has been systemically heparinized. The right atrium is cannulated for venous return and cardiopulmonary bypass is initiated, with systemic cooling to 32°C. Ventricular fibrillation is induced and the left ventricle is decompressed via the left ventricular apex, aorta, or right superior pulmonary vein. Because the heart is not arrested, mean aortic perfusion must be maintained at 80 to 100 mmHg to maintain a perfusion gradient across the myocardium. The most severely narrowed coronary arteries are bypassed first, permitting distal perfusion after completion of the anastomosis. Performance of distal anastomoses requires aortic cross-clamping with myocardial ischaemia for 10 to 15 min. Stay sutures around the coronary artery maintain a bloodless field. At the end of the distal anastomosis the aortic cross-clamp is removed and the myocardium is reperfused while the proximal aortic anastomosis is performed with a partially occluding side-clamp. This process is repeated until all vein grafts are completed. Internal mammary artery grafts are performed last.

Systemic rewarming is begun during the last distal anastomosis. Reinstitution of normal sinus rhythm may require cardioversion. The operation is then completed in standard fashion.

It should be noted that some surgeons do not use systemic hypothermia. Fibrillation is reversed during the proximal aortic anastomoses to reduce myocardial oxygen demand.

Improvements in anaesthetic technique, myocardial protection, and cardiopulmonary bypass technology have led to shorter, safer, and more effective operations. This has enabled radical changes in postoperative care. Conventional postoperative intensive care with positive pressure ventilation can be avoided for more than 90 per cent of adult cardiac patients, irrespective of age and priority. We conduct cardiopulmonary bypass at mild to moderate hypothermia (32–35°C) using the short acting anaesthetic agent propofol, discontinuation of which allows early recovery of consciousness. Thorough rewarming prevents shivering in the recovery period, which may increase oxygen demand by 400 per cent. Extubation can be achieved within 2 h in more than 80 per cent of patients, irrespective of age, priority, or severity of disease. Venous return and cardiac output improve with discontinuation of intermittent positive pressure ventilation. Inotropic support is seldom required, even for patients with poor left ventricular function. Glyceryl trinitrate infusion, and occasionally sodium nitroprusside, are used to control hypertension and afterload. Perioperative myocardial infarction or myocardial stunning may require inotropic support with adrenaline or dobutamine. Balloon pumping is required in the 2 to 5 per cent of patients with myocardial infarction or ejection fraction less than 15 per cent (Fig. 23) 1783. Catastrophic intraoperative myocardial infarction or failure to wean patients with end-stage left ventricular failure from cardiopulmonary bypass may be managed by left ventricular assist devices or resort to cardiac transplantation. In a series of 1000 adult patients undergoing coronary or valve surgery only 10 required reintubation after early extubation. Most patients leave hospital after 5 to 7 days. Antihypertensive medication is continued and 150 mg aspirin prescribed daily to preserve graft function. Antianginal agents are suspended.

Even the earliest reports of coronary artery bypass grafting showed very low mortality rates for such an extensive operation. Favalaro reported an operative mortality rate of less than 5 per cent in 1967, and other centres reported similar results: such initial low mortality is remarkable when one considers the many advances made in surgical techniques since that time. Improved myocardial preservation reduced the mortality rate in elective operations to less than 1 per cent by the early 1980s. However, the increasing use of emergency surgery after percutaneous transluminal coronary angioplasty and the application of coronary surgery to more ill and elderly patients has increased operative mortality rates once again. Most recent reports for all patients undergoing bypass grafting show an operative mortality rate of approximately 4 to 5 per cent. This is probably due to the fact that more straightforward cases are now undergoing coronary angioplasty instead of bypass surgery. Surgery is now restricted to a higher risk population.

Operative mortality rates for patients with normal left ventricular function undergoing elective coronary bypass remain low: the CASS trial reported an operative mortality rate of 3.1 per cent in 9369 elective operations. Patients with poor left ventricular function have higher mortality. Older patients, especially those more than 75 years old, are also at increased operative risk. An overall mortality of 2.3 per cent was reported by the Cleveland Clinic, and this rose progressively with advancing age, reaching 4.7 per cent in those older than 75. Although several reports have documented a decreased efficacy of coronary bypass surgery in women, women do not have an increased operative mortality.

Elective and emergency reoperations are associated with an increased risk: overall operative mortality in the CASS study rose to 5.3 per cent when reoperations were included. Similarly, an increased operative mortality of 3.4 per cent was reported for reoperations by the Cleveland Clinic Group in 1987. Scarring of the heart and great vessels makes dissection difficult and accounts for much of this operative mortality. Embolization of atherosclerotic material from old grafts leads to markedly increased perioperative myocardial infarction rates in patients undergoing coronary reoperations.

The onset of thrombolytic treatment in the 1980s produced a group of patients who had residual critical stenoses present in important coronary arteries on cineangiography. Coronary artery bypass grafting is safe and effective in these circumstances and mortality is not increased. Urgent coronary revascularization can be performed early after acute myocardial infarction, and it is especially beneficial if completed within 4 to 8 h after infarction. However, few centres have been able to provide the amount of manpower necessary for this formidable logistical task. Recent studies have shown improved results with revascularization performed at increasing lengths of time after myocardial infarction using an amino acid-enriched myocardial cardioplegia agent.

The safety of urgent coronary artery bypass grafting after failed percutaneous transluminal angioplasty is controversial. Substantially increased mortality in patients who are rushed to the operating room after failed angioplasty has been reported (as high as 12 per cent). Logistically, few surgical centres can afford to have an operating room standing by in case of disaster.

Major complications of coronary artery bypass grafting include stroke and perioperative myocardial infarction. Mediastinitis is a serious complication of open heart procedures and occurs in 1 to 2 per cent of patients: mortality can be as high as 20 per cent. Diabetics have an increased risk of wound infections and mediastinitis, but the incidence remains 5 per cent or less. The overall cerebrovascular accident rate is less than 1 per cent. Major strokes are more common in patients with heavily calcified and atherosclerotic ascending aortas and advanced carotid arterial disease. Minor neurological impairment occurs in a larger number of patients, but is clinically unimportant and resolves quickly. On the whole, leg wound problems are the most troublesome sequelae of coronary surgery and are attributable to poor technique (Fig. 24(a–c)) 1784,1785,1786.

Long-term results of coronary bypass grafting are dependent upon patent grafts. Reversed saphenous vein graft patency rate is approximately 80 to 85 per cent at 1 year, and 45 to 50 per cent at 10 years. The internal mammary artery grafted to the left anterior descending artery has a much improved early and late patency rate: 95 to 97 per cent at 1 year and up to 90 per cent at 10 years. Many surgeons now feel that elective revascularization of the left anterior descending artery should be performed with an in-situ left internal mammary artery graft. When angina recurs within the first year of surgery, early graft occlusion is the likely cause and is due to inadequate run off, technical failure or, occasionally, post-pericardotomy syndrome (Fig. 25) 1787. Angina recurring between 1 and 5 years after treatment is usually due to progression of native disease associated with progression of vein graft ‘atheroma’ and occlusion. Reoperation or angioplasty of a graft or native vessel can then be performed if medical treatment proves ineffective. Operative mortality is slightly greater than for first time operations.

CASS principal investigators and their associates. Coronary artery surgery study (CASS). A randomized trial of coronary artery bypass surgery. Circulation, 1983; 68: 939–50.

European coronary surgery study group. Long term results of prospective randomised study of coronary artery bypass surgery in stable angina pectoris. Lancet, 1982; ii: 1173–80.

Foster ED, et al. Comparison of operative mortality and morbidity for initial and repeat coronary artery bypass grafting: The coronary artery surgery study (CASS). Registry experience. Ann Thoracic Surg, 1984; 38: 563–570.

Kaiser GC, et al. Survival following coronary artery bypass grafting in patients with severe angina pectoris (CASS): an observational study. J Thoracic Cardiovasc Surg, 1985; 89: 513–24.

Loop FD, et al. Influence of the internal mammary artery graft on 10 year survival and other cardiac events. N Engl J Med, 1986; 314: 1–6.

Lytle BW, et al. Coronary bypass surgery early after thrombolytic therapy for acute myocardial infarction. Ann Thoracic Surg, 1986; 41: 176–83.

Parsonjnet V, et al. Emergency operation after failed angioplasty. J Thoracic Cardiovasc Surg, 1988; 96: 198–203.