Early students of the circulation were preoccupied with structure rather than function. The field attained the status of an art in the collaboration of the anatomist, Andreas Vesalius (1540–1564) and the artist Jan Van Calcar, a student of Titian. Such studies did however stimulate studies of function. Leonardo De Vinci (1452–1519) wrote that the contraction of the heart lasted about half as long as the resting period, and he built glass models of the heart valves to illustrate their function (Fig. 1) 1614. In De Vinci's life-time Jean Fernel wrote the first book devoted solely to bodily function, and originated the disciplinary name, ‘physiology’. Harvey estimated the stroke volume of the ventricles by measuring their capacity and showed that the heart must pump out in a few minutes ‘a volume of blood greater than that contained in the whole body’.

Harvey described the circulation of blood in 1628, but it was more than 250 years later before meaningful progress occurred in the surgical treatment of cardiovascular disease. The concept of circulatory support began in 1812 when Le Gallois wrote ‘If one could substitute for the heart a kind of injection of arterial blood, either naturally or artificially made, one would succeed easily in maintaining alive indefinitely any part of the body whatsoever’. In the 1850s Brown Sequard performed morbid but enlightening experiments to this end: ‘after decapitation of a dog, cannulae were inserted into common carotid and vertebral arteries on both sides and oxygenated blood injected with a syringe. During 15 minutes that the perfusion lasted there were movement of the eyes and jaws’.

In practical terms, as in most fields of surgery, cardiac surgery began with the treatment of trauma. Successful suture of experimental cardiac wounds was achieved by Bloch in the 1880s. The first clinical success is attributed to Rehn in 1897. In 1902 Hill, in the United States performed successful cardiac suture on the kitchen table by the light of oil lamps. At the time Billroth made the much quoted comment that ‘any surgeon who would attempt an operation on the heart should lose the respect of his colleagues’. This concept was upheld by the profession for many years until it gradually lost credibility in the decades after the end of the Second World War.

In 1899 Provost and Batelli devised an effective method for defibrillating the heart of an animal by opening the chest and applying directly to it 240 V of alternating current. It was 1947 before this was first applied successfully to a patient. There are few reports of cardiovascular operations before 1900. In 1888 Matas opened an aneurysmal sac and performed intravascular suture of the arterial openings. Weil and Delorme advocated removal of scar tissue surrounding the heart for treatment of constrictive pericarditis in the 1890s, though this relatively simple procedure was not undertaken clinically until 1920.

The practical development of cardiovascular surgery has, therefore occurred almost entirely within the twentieth century. At a time when the investigation of cardiac disease relied simply on physical examination and ausculation, rheumatic valve disease was the principle interest. Brunton (1902) suggested that mitral stenosis should be amenable to surgical therapy; shortly afterwards Munro urged ligation of the patent ductus arteriosus. In 1912 Tuffier described a successful transaortic aortic valvotomy and in 1923 Cutler and Levine reported the first transventricular aortic valvotomy. Both operations were extremely adventurous for the time and received more criticism than acclaim (Fig. 2) 1615. In 1925 Sir Henry Souttar described a method for relieving rheumatic mitral stenosis by introducing a finger through the left atrial appendage (Fig. 3) 1616. The patient, a 19-year-old girl survived with good symptomatic relief. At the time Souttar wrote ‘The problem is to a large extent mechanical and as such should already be within the scope of surgery, were it not for the extraordinary nature of the conditions under which the problem must be attacked. In view of the extreme danger to the brain from even the shortest check to its blood supply, any manipulations which are carried out must therefore be executed in the full flow of the bloodstream and they must not perceptively interfere with the contractions of the heart.' Although Souttar and others clearly foresaw the possibility of surgical solutions for cardiac disease, heart surgery became a moral issue, in much the same way as did cardiac transplantation some 50 years later. Souttar readdressed the need for artificial circulatory support, but cardiac surgery was to remain limited by these constraints for a further 30 years. Perhaps this was just as well for there were no antibiotics, no blood transfusion service, and anaesthesia for open chest surgery had not been developed.

In 1945 Strieder closed a patent ductus arteriosus, but the patient died of infection. Fifteen months later Gross performed the first successful ligation of a patent ductus and subsequently described the procedure for division and suture of the dividend ends.

The demonstration that some types of heart disease could be treated surgically provided an added incentive for accurate diagnosis. An endoscopic approach to internal cardiac structures was first attempted by Jacobaeus and Liljestrand in 1914 (Fig. 4) 1617. In 1929 a technique for right heart catheterization was described by Forssman, who practised transvenous insertion of a ureteral catheter in cadavers and then performed the procedure on himself when his colleagues refused to assist him. Little attention was given to this method until 1941 when Cournand, Richards, and others began to use it more extensively. Subsequently, they and Forssman received a Nobel Prize for their work.

The Second World War provided an additional stimulus to progress in cardiac surgery. In 1944 Harken was in charge of the first American Surgical Chest Centre located with the 160th General Hospital of the United States Army in Cirencester. He became fascinated by the problem of cardiac wounds and foreign bodies in the heart. In all Harken carried out 134 operations for the removal of cardiovascular foreign bodies (Fig. 5) 1618, 78 of which were within or related to the great vessels and 56 within or closely applied to the heart. Not one of his patients died. He subsequently returned to Boston to develop surgery of the mitral and aortic valves. According to Harken the credit for suggesting surgery for congenital pulmonary stenosis should go to O'Shaunessy. A description of the technique appeared in the text of a Hunterian lecture he prepared before his death in 1940. The unpublished text contains a drawing of the stenotic valve and another depicting its incision with a valvulotome inserted through the right ventricle.

Surgery for coronary artery disease originated in 1943 when Beck attempted to create a new cardiac blood supply by removal of part of the epicardium and onlay of pectoral muscle against the surface of the heart. Nineteen years later Vineberg implanted the internal mammary artery into the wall of the left venticle as a treatment for angina. The first experimental procedure for coarctation of the aorta was described by Park and Blalock in 1944, using subclavian artery to bypass obstruction in dogs. Claggett of the Mayo Clinic used this method clinically (Fig. 6) 1619 and shortly afterwards Crafoord described resection of coarctation with end-to-end anastomosis. This procedure required proximal and distal cross-clamping of the aorta, a procedure that was only conceived after Crafoord had torn a patent ductus during ligation and was forced to clamp the aorta for 28 min while repair was undertaken. When no neurological damage resulted Crafoord reasoned that one should be able to occlude the aorta of a patient with coarctation for sufficient time to allow resection and reanastomosis.

After Blalock's experiments with the subclavian artery, Taussig suggested that this might be used to divert blood to the lungs of patients with congenital pulmonary stenosis, a concept which would revolutionize the treatment of cyanotic congenital heart disease (Fig. 7) 1620. At the time there were grave doubts as to whether a severely cyanosed patient with heart disease could withstand anaesthesia, let alone the temporary occlusion of a pulmonary artery. Nevertheless, the first operation in 1944 was a success and in 1946 Potts described a procedure in which a side-to-side anastomosis between the aorta and pulmonary artery could be performed to the same effect, with the use of partial occluding clamps. Even more ingenious was the creation by Blalock and Hanlon in patients with transposition of a large interatrial defect by closed means.

The enormous clinical benefit derived from relatively simple mechanical surgical procedures greatly increased the interest in the scope of cardiovascular surgery for both congenital and acquired defects. The prospect of performing surgical procedures on intracardiac structures fascinated many investigators who sought more and more ingenious methods for achieving their goals. Notable amongst these were Bailey, Harken, Brock, Smithy (who himself died from aortic stenosis), and Tubbs, who in addition to performing the first successful closure of a patent ductus with endocarditis, developed mitral commisurotomy. Mitral valve surgery produced many problems including torn atria, atrial thrombus and systemic embolism, dysrhythmias, and serious mitral regurgitation. Bailey's first successful mitral commisurotomy was performed in June of 1948, after four previous failures. He had already been refused operating time in three of the five hospitals in Philadelphia in which he had privileges. Further failures would close these hospitals to him. He therefore scheduled his fourth and fifth operations on the same day in two hospitals. The fourth patient died, but Bailey had a taxi waiting and completed a successful operation on the fifth patient before the hospital could cancel. A week later this patient was flown to Chicago and was presented triumphantly at a cardiological meeting there.

The simple concept of relief of obstruction in cardiovascular surgery continued in 1948 when, within a few weeks, Holmes Sellors then Brock slit the stenotic valve in patients with valvular pulmonary stenosis using a transventricular approach (Fig. 8) 1621. In 1950 Brock reported the use of blind right ventricular infundibular resection for the treatment of tetralogy of Fallot using a punch type instrument. Subsequently, a variety of blind procedures were used in an attempt to close atrial and ventricular septal defects (Fig. 9) 1622. These included the blind suture method of Gordon Murray in 1948, the circumclusion operation of Sondergaard in 1950, atrioseptopexy by Bailey in 1952, and the atrial well technique of Robert Gross in 1953 (Fig. 10) 1623. From 1952 a number of methods for catheterization of the left side of the heart were described, including the transbronchial, transthoracic, and transeptal approaches, and it became increasingly apparent that satisfactory reproducible and safe repair of more complex abnormalities required direct vision of intracardiac structures. The aortic valve in particular was never satisfactorily dealt with by closed techniques. Blind approaches either from the left ventricle with an expanding dilator or from the aorta with an operating tunnel stitched to an aortotomy incision were occasionally successful but there were many spectacular failures. Surgery of the aortic valve therefore had to await the advent of open heart surgery and a reliable valve prosthesis.

Initially, there were three separate approaches to the goal of intracardiac surgery on a heart which was not required to support the circulation. The principles of hypothermia and temporary circulatory arrest were established by Bigelow, whose experimental findings were reported in 1950. Moderate systemic hypothermia (32°C) achieved by topical cooling in association with venous inflow occlusion by temporary ligation of the superior and inferior vena cava gave the surgeon 8 to 10 min to close an atrial or ventricular septal defect before irreversible cerebral damage ensued. The first clinical success using this method was reported by Lewis in 1952, using cold rubber blankets, and was later repeated by Swan in 1953, using an iced water bath technique which was eventually adopted by others.

Swan's hypothermic bath technique was cumbersome, messy, and required a prolonged rewarming process. Brock therefore developed a technique for venovenous cooling with a cooling coil and then returned blood to the heart. The procedure was quicker and more reliable, but the patient had to be heparinized. It was never widely used.

The first successful open resection of infundibular stenosis and closure of ventricular septal defect was performed by Scott using hypothermia. Drew and Anderson subsequently used profound hypothermia with longer periods of arrest for intracardiac surgery in adults.

The second method, that of donor cross-circulation was originally considered by Blum and Megibow in 1950. Kerr (1952) and Southworth and Pierce (1952) used cross-circulation for partial or complete cardiac bypass in experimental animals, but experienced many technical difficulties by failing to use a pump to control the interchange of blood between donor and recipient. Andreasen and Watson (1952) presented a detailed physiological analysis of their experiments with cross-circulation and emphasized the value of pump control and the reduced flow concept based on their earlier azygos flow studies. Lillehei and Varco (1955) successfully applied donor cross-circulation to the clinical correction of ventricular septal defects, initially reporting results in eight patients with no mortality or significant morbidity. Lillehei's spectacular success obtained by linking the femoral vessels of an adult donor to the vessels of a child whose heart would be taken out of circuit for an intracardiac repair is one of the most exciting landmarks in the history of surgery (Fig. 11) 1624. Usually, the adult was a parent of the child. For the first operation the donor was an unrelated volunteer taxi driver. With further application of this technique it became apparent that the risks to the donor were unacceptable. An otherwise effective method of circulatory support was therefore abandoned in favour of cardiopulmonary bypass.

In 1935 Gibbon developed an apparatus which supported the circulation of a cat for 35 min while its pulmonary artery was occluded with a clamp. Over the next decade, the inability to oxygenate the blood efficiently hampered further developments until 1950, when it was recognized that turbulence in the oxygenator circuit could produce much higher rates of oxygenation. This discovery, known as the boundary phenomenon, increased the contact between red cells and oxygen. At the same time Tom Watson, Chairman of the board of IBM, provided financial and engineering support to build a heart–lung machine on a more sophisticated scale. Gibbon then began extracorporeal circulatory experiments on large dogs and performed successful intracardiac operations. Over a 3-year period mortality in animals fell from 80 per cent to 12 per cent, and by 1953 he was prepared to take on clinical cases.

The occasion presented itself when he repaired atrial septal defect with a large left to right shunt in an 18-year-girl. He used a pump oxygenator system: total bypass time was 45 min and for 26 min the circulation was taken over completely by the machine. The extracorporeal circuit consisted of a roller pump and an oxygenator made from stainless steel vertical screens over which the blood flowed producing turbulence with good oxygen exchange. The screens were enclosed in a plastic shell with oxygen bubbled directly into the closed environment. Coronary suction was used to recover the coronary sinus blood and a left ventricular vent was inserted to eliminate the risks of air embolus, since the heart continued to beat. The girl made an uneventful recovery. This achievement was the culmination of 22 years of investigation and generated an explosium of technological progress in cardiac surgery. Interestingly, at the time, it was not recognized as an important breakthrough, largely because the techniques of hypothermia and cross-circulation were more readily available to clinical surgery. Gibbons' following three patients died and he himself became disillusioned. It was a further 8 years before Sharp performed the first open pulmonary embolectomy, thus fulfilling Gibbons' original notion. Gibbons' stationary vertical screen oxygenator was adopted by Kirklin and later by Holmes Sellors. In Stockholm, Bjork devised an apparatus consisting of stainless steel discs rotating in a bath of blood. Melrose, improved the efficiency of Bjork's machine by rotating an inclined drum through which blood flowed slowly by gravity across series of shallow trough-like discs. This machine was used for supportive perfusion in sick patients with aortic stenosis: aortic valvotomy was still performed as a closed procedure but it was hoped that extracorporeal circulation would increase the likelihood of survival. In December 1953 a woman with severe aortic and mitral valve disease survived a double closed valvotomy performed using the Melrose machine. Six further patients were operated on, but only the first survived and clinical work was halted. In April 1957 the Hammersmith Hospital team operated on a 30-year-old woman with a secundum atrial septal defect and pulmonary hypertension. The postoperative course was stormy but the patient survived and is still alive and well. However, because atrial septal defects could be closed with hypothermia alone there was continued scepticism regarding the heart–lung machine. In an attempt to dispel this, the paediatric cardiologist, Bonham Carter referred 50 children with ventricular septal defects to Hammersmith Hospital, with no questions asked until all 50 operations had been completed. The operations were a resounding success, with an overall mortality rate of 20 per cent. This fell to only 4 per cent in those with uncomplicated defect without pulmonary hypertension or anatomical complications. These early operations used potassium citrate to arrest the heart. The Melrose technique was to inject 2 ml of a 25 per cent solution of potassium citrate in 50 ml of autologous heparinized blood into the aortic root (with the ascending aorta clamped). This produced ideal operating conditions on a flaccid bloodless heart, with rapid return to normal beating on release of the aortic cross-clamp. The technique, a forerunner of cold potassium cardioplegia, was used to treat over 100 patients without any apparent ill-effects. However, reports from North America suggested that the potassium solution caused myocardial damage and it was reluctantly abandoned in favour of electrical fibrillation with hypothermia, or coronary artery cannulation and perfusion.

Cardiopulmonary bypass using a pump oxygenator system developed rapidly and increased greatly the scope of cardiac surgery. Nevertheless, side-effects, particularly severe pulmonary dysfunction, bleeding tendencies, and renal failure, occurred frequently and were attributed to damage to blood components caused by the pump oxygenator system. The so-called ‘post-perfusion syndrome’ was attributed to the adverse effects from contact of blood with foreign surfaces. Drew addressed this problem by keeping the patient's lungs as the oxygenator and using a two-pump system to replace both right and left sides of the heart. To this he added a cooling coil and explored the use of profound hypothermia with body temperatures as low as 8 to 10°C. The first successful application of this technique was in January 1959 for closure of atrial and ventricular septal defects in a 4-year-old child. Apart from Ross, who has recently used the two-pump system for coronary artery surgery, others did not adopt this technique. Deep hypothermia is nevertheless used routinely with conventional cardiopulmonary bypass, and total circulatory arrest for surgery of congenital heart defects in infants and young children, and for extensive surgery on the thoracic aorta in adults. Barrett-Boyes pioneered the use of surface cooling in infants: he employed ice bags followed by core-cooling to 10 to 20°C with the pump oxygenator.

Modifications and refinement of the heart–lung machine continued during the 1960s and 1970s. The next major step was the development of the membrane oxygenator, a device which reduced trauma to blood cells by separating blood from gas. With the elimination of time constraints attention switched to the treatment of more complex congenital and acquired heart defects. Closed mitral valvotomy was well established for mitral stenosis, but there were no satisfactory procedures for aortic valve disease or mitral regurgitation. Various reparative procedures were attempted (annuloplasty by Wooller, pericardial inserts by Ross and Cleland) but there was clearly a demand for a valve prosthesis. Hufnagel designed and used a caged ball valve prosthesis for insertion in the descending thoracic aorta of patients with severe aortic regurgitation (Fig. 12) 1625. Further developments along similar lines by Starr, Harken, and McGovern led to the development of a satisfactory caged ball prosthesis for insertion in the subcoronary aortic position (Fig. 13) 1626. Initial attempts to use the prosthesis in the mitral area failed because of protrusion of the cage into the outflow tract of the left ventricle. Despite this, in 1962 Starr and Edwards reported a series of 16 patients who had undergone mitral valve replacement with a ball valve prosthesis: 10 survived. In 1964 Melrose reported both experimental and clinical use of a tilting disc valve made of polypropylene. The valves were manufactured and used clinically by Wooller and Logan. This prosthesis was haemodynamically sound and initially highly successful. With time more and more patients returned with extensive thrombus formation around the prosthesis, and its use was abandoned. Soon afterwards Bjork designed a similar tilting disc valve which was manufactured by the Shiley Company and used clinically in competition with the Starr valve from Edwards Laboratories. Some surgeons, including Brock, had serious reservations about implanting foreign material within the heart and great vessels. Brock was convinced that natural tissues from cadavers should be used and began by using an aortic tube homograft for repair of coarctation. Ross carried out the first aortic valve replacement using an aortic homograft as an emergency in London in 1962, the valve homograft having been prepared by Gunning in Oxford. Both he and Barrett-Boyes worked enthusiastically with homograft replacement, refining the methods of collection, preparation, and preservation. Later Ross varied the technique by transferring the native pulmonary valve to the aortic area and replacing the patient's pulmonary valve with an aortic valve homograft. Senning replaced the aortic valve with autogenous facia lata, initially with excellent early results. The procedure was later abandoned when valve failure followed shrinkage and calcification of the native material. Nevertheless, many surgeons found homograft valve replacement tedious and difficult to carry out and in order to overcome these drawbacks attempts were made to mount pericardium, dura mater, or homograft valves on a metal frame to which a sewing ring was attached. A little later a pig valve heterograft, again prepared by Gunning in Oxford, was implanted in a patient by Binet in France, and soon after Allison and Gunning in Oxford also used these pig valves. When Carpentier showed that glutaraldehyde preparation removed most of the antigenic properties of pig or calf valves, heterograft valves were stent mounted and have subsequently proven very successful.

Revascularization of the ischaemic myocardium remained a considerable challenge. As early as 1937 O'Shaunnessy had suggested the technique of omentopexy, whereby omentum was transferred in continuity from the abdomen to the surface of the heart to encourage collateral circulation. Pneumopexy, partial coronary sinus ligation, and pericardial abrasion were all tried, with dubious outcome. Vineberg's implantation of the internal mammary artery into myocardium provided a more promising line of attack. He showed that an open-ended artery buried in the myocardial sinusoids could establish vascular connections and supply additional blood to an ischaemic left ventricle. Effler and Favaloro adopted this technique enthusiastically and were well placed when Mason Soanes developed the technique of coronary arteriography. This revolutionized surgery for coronary artery disease by demonstrating the site of coronary occlusion and providing the concept of coronary bypass. Favaloro used autogenous saphenous vein, reporting the technique in 1968. Although coronary artery bypass surgery would eventually account for more than 80 per cent of cardiac operations the procedure was initially regarded with scepticism and relatively few operations were carried out for the next 5 years. Although it was clear that successful coronary artery bypass relieved angina, there were doubts about the ability to prolong life and its superiority over increasingly effective medical treatment of angina.

From 1973 to 1978 several British centres took part in a European multicentre randomized trial to compare long-term results of coronary bypass with those of medical management. The results together with those of American trials, and careful analysis of subsets with different degrees of coronary disease, established the superiority of surgery, particularly in patients with two and three vessel coronary disease, left main stem obstruction, and high-grade proximal left anterior descending obstruction. Impaired left ventricular function from previous myocardial infarction reinforces the decision for surgery and the superiority of surgical management.

In December 1967 Barnard astonished the world by performing the first clinical human cardiac transplant, based on the pioneering research work of Shumway at Stanford. The patient died from acute rejection 18 days later, but a second patient lived for 18 months and produced a wave of enthusiasm throughout the surgical world. In Britain Ross performed three transplants but the longest survivor lived for only 43 days and publicity surrounding excision of the beating donor heart brought the programme to an end. This led to a clear definition of brain death designed to convince the public and medical profession that an individual would be accepted as a donor only when independent life would never be possible. Fortunately, Shumway continued with his transplant programme when most others had given up in disillusionment. His remarkable achievements led English and Yacoub separately to resume the British programme in 1979.

Barnard CN. A human cardiac transplant: an interim report of a successful operation performed at Groote-Schuur Hospital, Capetown. S Afr Med J 1967; 41: 1271–4.

Blalock A. Cardiovascular surgery, past and present. J Thoracic Cardiovasc Surg 1966; 51: 153–64.

Blalock RC, Cambell M. Discussion on surgery of the heart and great vessels. Proc R Soc Med 1951;44: 995–1005.

Cleland WP. The evolution of cardiac surgery in the United Kingdom. Thorax 1983; 38: 887–96.

Drew CE, Anderson IM. Profound hypothermia in cardiac surgery. Lancet 1952; i: 748–50.

Gibbon JH. Application of a mechanical heart and lung apparatus to cardiac surgery. Minnesota Med 1954; 37: 171–80.

Gross RE. Surgical correction for coarctation of the aorta. Surgery 1945; 18: 673–78.

Harken DE. Foreign bodies in and in relation to the thoracic blood vessels and heart. I. Techniques for approaching and removing foreign bodies from chambers of the heart. Surg Gynecol Obstet 1946; 83: 117–25.

Kirklin JW. Open heart surgery at the Mayo Clinic: the 25th anniversary. Mayo Clin Proc 1980; 55: 339–41.

Kirklin JW, Pacifico AD. Surgery for acquired valvular heart disease. N Engl J Med 1973; 288: 133–38.

Lillehei CW, Cohen M, Warden HE, Zeigler NR, Varco RL. The results of direct vision closure of ventricular septal defects in eight patients by means of controlled cross circulation. Surg Gynecol Obstet 1955; 10: 447–66.

Souttar HS. The surgical treatment of mitral stenosis. Br Med J 1925; 2: 603–6.

Starr A, Edwards ML. Mitral replacement: clinical experience with a ball valve prosthesis. Ann Surg 1961; 154–726.

Vineberg A. Treatment of coronary artery insufficiency by implantation of the internal mammary artery into the left ventricular myocardium. J Thoracic Surg 1952; 23: 42–54.