Portland, 21 September 1960
In June 1970 a retired truck dispatcher in Spokane, Washington, fell off a ladder while painting his house and was fatally injured. There was nothing particularly unusual about the death of Philip Amundson: thousands of Americans die in domestic accidents every year. But his survival to the age of sixty-two, and the fact that he was well enough even to climb a ladder, constituted a medical miracle. Ten years earlier he had been a hopeless invalid, unable to walk more than a few steps without getting out of breath. Decades earlier a childhood infection had irreparably damaged one of the valves in his heart, and the organ was now starting to fail. Having exhausted all other options, Amundson’s cardiologist told him that his only hope of survival was an experimental treatment developed by a young surgeon, Albert Starr. Animal tests had delivered promising results, but his single human patient had died within hours.
On 21 September 1960, Starr removed Amundson’s diseased mitral valve and replaced it with a prosthetic valve, a silicone rubber ball imprisoned in a plastic cage. It worked like a dream: within hours Amundson was sitting up in bed and talking to his nurses. Permanently replacing a human mitral valve was an achievement of unprecedented daring, and doing so with a man-made device made it all the more remarkable. It was a development which promised a cure for hundreds of thousands of patients previously regarded as untreatable, and the hospital authorities knew that the world’s media would take an interest. But the operation was kept secret for three weeks, long enough for them to be confident that Amundson would make a full recovery. By mid-October the danger had passed, and the story became front-page news. ‘Heart Valve Replaced by Rubber Ball’, read a typical headline, accompanied by a photograph of an impassive Amundson holding a model of the device that had saved him.1
A handful of surgeons had implanted artificial valves before, with results that ranged from death to mild improvement, but nobody had achieved anything like this success; there was a unique sense of excitement about this operation that suggested it was something special. It had been preceded by two years of promising animal experiments, and Starr was optimistic that heart surgeons finally had a reliable and durable artificial valve which could save the lives of thousands. The device, designed by a retired engineer called Lowell Edwards, exceeded his wildest expectations. Within three years more than 6,000 sick patients had been given one,2 and many lived for decades. When he died ten years after his operation, Philip Amundson’s valve was still working perfectly. But this was not exceptional: in 2014 it was reported that a sixty-seven-year-old woman in Pennsylvania with a Starr–Edwards valve was alive and well forty-eight years after it had been implanted.3
Creating an artificial ‘spare part’ for the heart was a triumph of design and engineering – but even the most sophisticated prosthetic valves are no match for those we are born with. The human heart has four, and their function is simple: to keep the blood flowing in the right direction. Each consists of three tough fibrous leaflets (two in the case of the mitral valve) attached to a ring of tissue, the annulus. The leaflets meet in the centre of the valve aperture like the petals of a tightly closed flower, forming a seal which prevents blood from flowing backwards. When the valve opens these ‘petals’ rapidly unfurl, allowing blood forward into the next chamber or vessel, before snapping shut to ensure that none flows back the way it came.
The heart is often referred to as a pump, but it’s more useful to think of it as a pair of them working in parallel. Pump no.1 (the right side of the heart) sends deoxygenated blood towards the lungs, while pump no.2 (the left side) propels the freshly oxygenated blood to the rest of the body. Each consists of two chambers: a reservoir, known as the atrium, and the pumping chamber, the ventricle.
Every heartbeat has two phases, known as diastole and systole, during which the organ fills and then empties itself of blood. At the beginning of diastole the heart muscle is relaxed, and blood which has completed its journey around the body flows into the right atrium, like water filling a reservoir. From there it trickles through the tricuspid valve into the empty right ventricle. At the end of diastole the muscles of the atrium contract, raising the pressure in the chamber and forcing more blood through the tricuspid valve into the ventricle. A fraction of a second later systole begins and the muscle of the ventricle contracts. This has the same effect as squeezing a plastic ketchup bottle: the contents are compressed, producing an increase in pressure. Increased blood pressure makes the tricuspid valve slam shut, and opens the pulmonary valve. Like ketchup ejected from a squeezed bottle, blood is driven from the right ventricle into the pulmonary artery towards the lungs.
While deoxygenated blood is travelling through the right side of the heart, an essentially identical sequence is unfolding on the left. Freshly oxygenated blood returns from the lungs via the pulmonary veins and arrives in the left atrium. During diastole it flows through the mitral valve into the left ventricle; when systole begins and the ventricle contracts, the valve snaps shut and blood is propelled at high pressure through the aortic valve and into the aorta, to be distributed to the whole body. When the body is at rest this entire process of diastole and systole, known as the cardiac cycle, takes around a second.
If you listen to a healthy human heart you’ll hear the familiar rhythmic ba-boom, ba-boom, ba-boom, usually rendered by medics as ‘lub-dub’. The quieter first sound, the ‘lub’, is caused by the mitral and tricuspid valves closing at the beginning of systole; the louder ‘dub’ is the sound of the aortic and pulmonary valves slamming shut at the end of systole, as the pressure in the two ventricles falls. Sometimes a doctor will be able to hear an additional or anomalous sound through the stethoscope – called a ‘murmur’ – and this is often evidence of a problem. Among the possible causes of a heart murmur are a valve which has become narrowed and fails to open properly (known as stenosis), or one which no longer provides a tight seal, allowing blood to flow backwards (known as regurgitation).
Repairing or replacing the valves of the human heart was among the most intractable problems of twentieth-century surgery and one which, once countenanced, took more than sixty years to solve. The first suggestion that diseased valves might be amenable to surgical treatment was made by Herbert Milton, the principal medical officer at Kasr el-Aini Hospital in Cairo. In 1897, the year after Ludwig Rehn’s first successful repair of a cardiac wound, Milton wrote to The Lancet to describe his new method of opening the chest. This was itself of great significance, since the procedure he recommended – splitting the breastbone in order to gain access to the organs of the thorax – is now the incision most commonly used in open-heart surgery. In 1897 there were few operations requiring such a drastic opening of the chest, however, so Milton suggested circumstances in which it might in future be useful, including the removal of foreign bodies from the lungs. ‘If once a safe route is established a great field for surgical interference lies open,’ he wrote. ‘Heart surgery is still quite in its infancy, but it requires no great stretch of fancy to imagine the possibility of plastic operations in some, at all events, of its valvular lesions.’4
Why did Milton first think of the valves when considering what operations might be performed on the heart? Partly because they presented the most urgent problem. Specialists in the late nineteenth century knew a lot about valve disorders, their symptoms and the sounds they made when heard through a stethoscope. And there was no shortage of cases: doctors were contending with a veritable epidemic of valve disease, caused by an infection which – in the developed world, at least – has almost disappeared today: rheumatic fever.
The history of rheumatic fever is an interesting example of a disease metamorphosing before our eyes. Pathogens such as viruses or bacteria reproduce at such a rate that mutations can emerge quite quickly, often changing the nature of the illness they cause. When doctors wrote about rheumatic fever in the eighteenth century, the symptoms they described typically included fever and joint pain, or ‘rheumatism’. But around 1800 the pathogen appears to have evolved, and started to affect the tissues of the heart; still later, the brain became the disease’s main focus, giving rise to the strange involuntary jerking movements known as Sydenham’s chorea or St Vitus’s Dance.5 We now know that the pathogen involved is a bacterium, Streptococcus pyogenes. This usually causes nothing worse than a sore throat, but in some patients the antibodies produced by the immune system to combat the infection trigger tissue inflammation, causing the symptoms of rheumatic fever. In the developed world the disease has retreated steadily since 1900, but in poorer nations it continues to be a major problem, estimated to cause over 250,000 deaths a year.6
The first to make the connection between heart disease and rheumatic fever was David Pitcairn, a Scottish physician working at St Bartholomew’s Hospital in London. In 1788 he noticed that patients with rheumatism were more likely to show symptoms of heart disease, and conjectured that the two disorders had a common cause, which he called ‘rheumatism of the heart’.7 The theory came to wider attention in 1812, when William Charles Wells published a detailed study definitively proving the association. Wells had an unusually wide range of interests: he solved the mystery of where dew comes from,8 and proposed a theory of natural selection almost fifty years before Darwin.9 One case cited in his landmark paper was that of a young woman who in 1807 died from heart disease a few months after falling ill with rheumatic fever. Inside her heart were found a number of ‘excrescences resembling small warts’,10 several of which were on the mitral and aortic valves. These growths, known as vegetations, are a characteristic feature of rheumatic heart disease. Wells’s findings were confirmed by many other physicians: in 1909 a doctor at St Bartholomew’s reported that 99 of 100 rheumatic fever patients in his care were also suffering from diseased heart valves.11
Rheumatic fever was a common ailment, creating an army of patients with incurable cardiac problems. Particularly upsetting was the fact that a large proportion of them were children – in stark contrast to the situation today, when acquired heart disease predominantly affects the elderly. In 1898 a London physician, Daniel Samways, tentatively offered a suggestion. Writing in The Lancet, he predicted that one day mitral stenosis – a narrowing of the mitral valve caused by rheumatic vegetations – might be relieved by ‘slightly notching’ the valve orifice with a scalpel.12 Samways felt that if it were somehow possible to gain access to the inside of the heart, making a nick in one of the valve leaflets would increase the size of the aperture and thus increase the flow of blood. His suggestion was barely noticed; but four years later, when the eminent surgeon Sir Thomas Lauder Brunton made a similar suggestion, all hell broke loose.
Mitral stenosis is a distressing condition, causing debilitating breathlessness, fatigue and chest pain. Lauder Brunton’s frustration at being unable to ease his patients’ suffering was compounded by the realisation that during autopsies it was quite easy to cut open the constricted valve with a scalpel, and it occurred to him that it might be possible to do this on a living patient.13 He was cautious in his proposal, noting that the operation should only be attempted after it had been tested on animals. Nevertheless, it provoked outrage: an editorial in The Lancet noted sniffily that Lauder Brunton had ‘proceeded no further than the table of the dead-house in making his investigation’, and chided him for proposing a potentially dangerous operation without having first conducted his own experiments to establish conclusively whether it was feasible.14 There was some support, however: Daniel Samways wrote to back up his colleague, pointing out that the same idea had occurred to him.15
Opinion remained divided for some time. The great Rudolph Matas, pioneer of aneurysm surgery, believed the possibility ‘not such a chimera as many have supposed’.16 Ludwig Rehn was discouraging, writing in 1913 that in surgical terms the heart valves were ‘noli me tangere’ – out of bounds.17 Yet he was already out of date, for the previous year two French surgeons had attempted to operate on human patients. Théodore Tuffier tried to ease a case of aortic stenosis using an ingenious and non-invasive method: rather than cut into the heart he pressed his little finger into the wall of the aorta, forcing a pocket of the vessel downwards into the aortic valve in the hope that this would expand its narrowed opening. The patient was improved, though only temporarily.18 Tuffier’s compatriot Eugène Doyen tried something more drastic, inserting an instrument through the wall of the heart in an attempt to dilate a young girl’s constricted pulmonary valve; she died shortly afterwards.19
One problem surgeons faced was that they could not see what they were doing: opening the heart would remain impossible until the invention of cardiopulmonary bypass several decades later. To get around this difficulty, in the early 1920s two researchers at Washington University in St Louis, Evarts Graham and Duff Allen, invented an instrument for looking at the inside of the heart: the ‘cardioscope’, a metal tube about the size and shape of a fountain pen, fitted with a lens and a light bulb. The surgeon would make an incision between two ribs to expose the heart, then insert the cardioscope through a small hole made in the cardiac wall. By leaning over the patient’s chest and peering through this miniature telescope he could then survey the heart’s interior landscape – rather hazily, since blood is opaque and the lens thus had to be in direct contact with whatever the operator wanted to see. It was also fitted with a narrow blade which could be used to cut diseased valve leaflets. Results were mixed: the cardioscope was used to slit the mitral valves of twenty-two dogs, proving that the concept was sound,20 but when Allen attempted to treat three patients with mitral stenosis in August 1923 they all died. Disheartened, he abandoned his research.
Some months earlier a surgeon in Boston, Elliott Cutler, had performed the first mitral valve operation on a human patient, a procedure he only attempted after several years of scrupulous preparation in the laboratory.21 Cutler decided that simply cutting a slit in the diseased valve was not enough, since the new incision might heal, and instead wanted to remove a small section of the leaflet tissue. To accomplish this he asked a young research fellow, Claude Beck, to devise a new instrument which he called the cardiovalvulotome. It had a pair of cutting jaws, designed so that they would hold on to the excised tissue and prevent it from getting into the bloodstream.22 In the spring of 1923 Cutler’s colleague Samuel Levine told Cutler that he had found the ideal candidate for operation. She was a twelve-year-old girl with crippling mitral stenosis, the consequence of a bout of rheumatic fever two years earlier. She had been bedridden for six months and her heart was starting to fail: so little blood could pass through the narrowed valve that the pressure on one side had become dangerously high, and as a result the organ could no longer pump enough to the rest of the body.
At 8.45 a.m. on 20 May, Cutler operated. The cardiovalvulotome was not yet ready for use, so instead he used a long thin knife with a slightly curved blade. Suspecting that the heart would cope better with the trauma if it became habituated to gentle handling, Cutler put his hand underneath it and turned it over to look at the other side – perhaps the first surgeon ever to do so. He dripped adrenaline directly on to the heart to make its contractions more vigorous, and then plunged the knife into the left ventricle. He pushed it in until he felt an obstruction which he assumed to be the mitral valve, twisted the blade and made two little nicks in what he hoped were the valve leaflets. Having completed this terrifying procedure completely unsighted, he withdrew the knife and placed sutures at the point where it had entered the heart muscle. The operation had taken a little over an hour.23
At first the girl showed worrying signs of complications, but after four days she had made a remarkable recovery. Her doctors were so pleased with her progress that they allowed her out of bed to put in a surprise appearance at the hospital’s tenth birthday celebrations, which were in full swing downstairs. Though pleased with her condition, Cutler was guarded about the operation’s success. He was not sure what he had done, or whether it would help.24 Whatever benefit she gained from Cutler’s daring work was short-lived, as his patient survived for only two years before dying from pneumonia.25 The operation did achieve something, though: it proved that a patient could withstand the insult of having an instrument inserted into the cardiac chambers. It was enough to encourage Cutler to persevere, and he went on to operate on seven further patients using the cardiovalvulotome. None survived for longer than a week, and in 1929 he abandoned the operation as too dangerous.26
It was the first of many such disappointments: frequently over the next few decades a carefully thought-through idea would result only in failure. Sometimes surgeons wandered down many blind alleys before finding the correct path forward, or allowed themselves to be blinded by their own ingenuity, attempting a procedure long before anaesthetic technique and operating-theatre technology had attained the necessary sophistication. Such setbacks were an inevitable corollary to progress, but they came at substantial human cost. The many patients who died after agreeing to undergo an experimental procedure knew that medicine had nothing better to offer them – but this was scant consolation for their families, who had been given brief hope that their loved one might, after all, get better.
Two years after Cutler’s first operation, the London surgeon Henry Souttar tried a different method. Before taking up medicine Souttar had studied mathematics and engineering – experience which clearly influenced his thinking. ‘The problem is to a large extent mechanical,’27 he wrote, and he approached the heart valves as pragmatically as an engineer would a malfunctioning pump. In March 1925 he was introduced to Lily Hine, a sickly nineteen-year-old patient who had grown up in the slums of Bethnal Green, in the sort of insanitary conditions where rheumatic fever thrives.28 The illness had left her with a mitral valve which was grossly distorted: it would neither open nor close fully, a combination of stenosis and regurgitation.
Souttar felt that he could at least improve the stenosis, and on 6 May 1925 he conducted an audacious operation to separate the fused valve leaflets. Having opened Lily’s chest, he made a small incision in the left atrial appendage,fn1 a flap of heart muscle that sticks out from the heart’s left atrium. He quickly slipped a finger through the hole and used it to explore the inside of the atrium. He was immediately struck by the wealth of information this simple fingertip examination yielded: by feeling the mitral valve leaflets he could construct a mental picture of their condition, while a rush of blood travelling the wrong way told him that the valve was not closing properly. He had intended to cut the fused leaflets with a knife, but now worried that this would exacerbate the regurgitation. So he improvised, pushing his finger through the valve opening to enlarge it. As he removed his digit from the girl’s beating heart there was a terrible moment when a suture slipped, causing a sudden geyser of blood across the operating table. With great presence of mind Souttar checked the flow by gripping the end of the atrial appendage between finger and thumb while an assistant bound it tightly with silk.29
Lily Hine was sent to the country to recuperate, and after six weeks claimed to feel much better. The psychological impact of surgery is such, however, that patients often overestimate how much good it has done: Souttar noted that objective tests showed little obvious improvement to her overall condition. And though he was confident that the procedure was sound, he never had an opportunity to repeat it and prove its worth conclusively. Most cardiologists of the era believed (wrongly) that rheumatic fever was primarily a disease of the heart muscle, in which case operating on the valves was futile; London specialists refused to send Souttar another patient.30 This closed another chapter in experimental heart surgery: without any tangible success to show for their work, the few surgeons who had been willing to countenance the operation decided that further attempts would be futile.
Many years later Souttar would admit that he had been ‘perhaps too adventurous’,31 and indeed history shows him to have been twenty years ahead of his time; it was not until the 1940s that any real progress was made in attempts to repair the heart valves. In 1946 a young surgeon from South Carolina, Horace Smithy, developed a valvulotome, a cutting instrument similar to Elliott Cutler’s two decades earlier. It resembled a large metal syringe, with a plunger at one end and a pair of cutting jaws at the other. The jaws were sheathed in the cylinder of the device as it was introduced into the heart; when the surgeon depressed the plunger the jaws would ‘bite’ out a section of valve leaflet and retain it.
For Smithy the battle against valve disease was deeply personal: he was himself suffering from aortic stenosis, a discovery he made as a medical student when he first used a stethoscope to listen to his own heart.32 His experiments using the valvulotome on dogs were successful,33 and when he presented his research at a meeting of the American College of Surgeons in September 1947 it made a big impression: the success of the Blue Baby operation three years earlier had made journalists hungry for further advances in cardiac surgery. One of those who read the newspaper reports was Betty Lee Woolridge, a twenty-one-year-old woman from Ohio. She wrote to Smithy, explaining that she had mitral stenosis and had been given a year to live by her doctors. She begged him to attempt the operation on her: ‘Why use dogs when you can use human beings? If anything happened to me you might learn something which could help somebody else.’34
Smithy was at first reluctant, but Ms Woolridge was not to be denied. On the last day of January 1948 he operated, cutting out a small portion of her stenotic mitral valve. Ten days later she was well enough to fly back to Ohio, posing for newspaper photographers as she bade farewell to Smithy on the aircraft steps. Smithy was emboldened to operate on six more patients; four survived, but the overall results were disappointing, with only two showing ‘slight’ improvement.35
Smithy’s own health was in steep decline, adding a sense of acute urgency to his work. In May he approached Alfred Blalock to ask whether he would consider operating on him; nobody had so far attempted such a procedure on the aortic valve. Blalock was prepared to help, but a trial operation in Baltimore, at which Smithy assisted, was a disaster. One of those present, Denton Cooley, recalled Smithy’s reaction as their patient died: ‘I looked over at him and saw his face fall. He thought that this was his only chance at having a successful operation for himself.’36 His last hope had indeed evaporated: Blalock could not be persuaded to repeat the experience, and a few months later Smithy succumbed, aged thirty-four, to the condition he had sought to treat. On 29 October his death was reported under the headline ‘Surgeon Dies, Too Weak For Own Cure’.37
If only Smithy had survived a little longer! Meaningful progress was painfully close: within a few months three surgeons, working independently and miles apart, made the same important breakthrough in the treatment of mitral stenosis. The first was Charles Bailey of Hahnemann University Hospital, Philadelphia, who had funded his way through medical school by selling ladies’ underwear, a crucial sideline, as it turns out: contemplating his wares one morning he was struck by the structural similarity between a girdle and the mitral valve. Suspenders offered firm but flexible support to the stockings, just as the chordae tendineae, tough fibrous strings of tissue, anchor the valve leaflets to muscles on the inner wall of the heart. This insight informed much of his later thinking.38 A young artist friend helpfully supplied an illustration of a panty-girdle and a mitral valve side by side; his name was Walt Disney.39
Bailey, a great admirer of Smithy, had watched his colleague’s deterioration with great sadness. At a surgical meeting he listened to his heart and the ‘terrible rumble’ of the aortic stenosis that would kill him.40 Bailey needed no reminder of the devastating consequences of valve disease: at the age of twelve he had seen his father die of mitral stenosis, coughing up blood into a basin while his mother did her best to comfort him.41 Like Smithy, Bailey revisited Cutler’s operation, trying to relieve the stenotic valve by cutting a section out of its leaflets with an instrument which acted like a hole punch.
Things did not go well. Four of his early patients died, and one of the hospitals at which he worked banned him from making any further attempts. His other employer, Episcopal Hospital, tried to do the same, but after a heated argument with the head of cardiology Bailey was allowed to continue.42 On 10 June 1948 he finally had the success he needed. This time he used a new instrument, a knife attached to his forefinger like an extended claw: having inserted it through the heart wall he hooked the blade through the stenotic valve and used it to slit open the commissure, the point where the two leaflets had fused. He then removed the knife and used his finger to dilate the valve further, ensuring it could open fully once more.43
The operation saved not only the patient’s life – her symptoms were greatly relieved – but Bailey’s career. Ten days later she was well enough to travel a thousand miles to Chicago, where Bailey presented her to a room full of impressed surgeons. Evarts Graham, co-inventor of the cardioscope, praised Bailey’s ingenuity and made a point that remains as valid today as it was in 1948: ‘Unfortunately, as with any new surgical procedure, the pioneers get only the patients who are the worst possible risks.’44 Like Souttar and Cutler before him, Bailey had only been permitted to operate on those who were at death’s door; critically ill patients whose survival prospects were poor even if the surgery went as planned. After years in the animal laboratory Bailey was sure that his procedure was sound; it was only by saving one of these hopeless cases that he could persuade his physician colleagues to send him individuals more likely to survive major surgery. Many later surgeons risked livelihood and reputation as Bailey did, watching patient after patient die before finally proving that a new operation could save lives.
As Bailey travelled to Chicago on 9 June, another surgeon was already repeating his achievement. In Boston, Dwight Harken successfully opened a stenotic mitral valve using a valvulotome.45 Thinking that he was the first to do so, he hurriedly wrote an article documenting the case, unaware that Bailey had beaten him to it.46 Three months later, in London, Russell Brock took the bronze medal in this surgical race, performing the first in a series of successful mitral valve operations.47
Temperamentally, Brock, Bailey and Harken were quite a trio. Brock was a shy man, but hid this behind a brusque, even abrasive, manner. To his egalitarian American colleagues he must have appeared a throwback to an earlier age, for he belonged to an English tradition which cast senior medics as demigods; his juniors at the Brompton Hospital in London were expected to await his arrival on the front doorstep each morning, ready to conduct the great man on his daily rounds. Harken and Bailey were less intimidating figures, but they simply couldn’t stand each other. If they met at a medical conference the tension was palpable, and discussions frequently turned into blazing rows. Bailey had an entertaining explanation for their mutual antipathy: ‘My mother was redheaded; my daughter is redheaded. I never was, but Harken was … we just tore at each other with the classical vigor of redheaded people.’48 Many years later they would become good friends.
Though Harken and Brock used a blade to perform their first mitral valve operations, they both subsequently concluded that the Souttar ‘finger-fracture’ technique of using a forefinger to break open the fused valve leaflets was a vastly superior approach, and adopted it in all their operations.49 Removing tissue from the leaflets using a punch or valvulotome often rendered the valve incapable of closing fully; rather than curing the stenosis this converted it into regurgitation, which was almost as bad for the patient as the original condition. The designer of the cardiovalvulotome, Claude Beck, who himself went on to become a celebrated heart surgeon, later felt that his invention ‘probably delayed the development of the operation for mitral stenosis by some 20 years’.50
‘In many cases the finger alone splits the valve with an accuracy and speed that no instrument could rival,’ Brock observed in 1952, reporting excellent results in a series of 100 operations.51 Declaring that a ‘new field of surgery’ had been opened up, Brock suggested that the procedure, which he called mitral valvotomy, was now routine and should be performed wherever there was suitable expertise. Demand was overwhelming: rheumatic fever remained a major concern in the 1950s, with 250,000 Britons believed to be suffering from mitral stenosis.52 A further improvement arrived in 1954, when the French surgeon Charles Dubost invented a dilator, an instrument even more efficient than the finger at opening a stenosed valve. At its tip were two blunt blades which could be passed into the left atrium and then expanded like an umbrella, pushing apart the fused leaflets.53 Many other surgeons took up the new operation, and over the next decade thousands of patients with mitral stenosis found their quality of life radically improved as a result.
More challenging still was the problem of mitral regurgitation: there was no easy way to repair valve leaflets so damaged by disease that they would not fully close. Inserting a foreign body in the aperture was one method several surgeons employed to prevent blood flowing the wrong way: Harken suspended small spheres or spindles of Perspex in front of the valve,54 while Bailey tried an even more difficult technique, sewing the loose valve leaflets together with strips of pericardium.55 None of these operations was particularly successful: the condition tended to recur, often within a few months. A completely different approach was needed.
As any driver who has endured an unexpectedly expensive afternoon in a garage will aver, replacing a faulty part is sometimes better than repairing it. The first to realise that this applied to hearts as well as engines was Gordon Murray, the Toronto surgeon whose work on heparin proved so vital to the development of open-heart surgery. Murray was working as a junior registrar at the London Hospital in 1925 when Souttar performed his one and only mitral stenosis operation,56 but what really kindled his interest in the subject was a visit to the laboratory of Cutler and Beck and the opportunity to discuss their work with them.57
When Murray began his own research in 1936 he took a radical approach, attempting to create a new mitral valve with spare parts harvested from the patient’s own body. In a series of experiments on dogs he cut out the entire valve apparatus, both the leaflets and the chordae tendineae – the stocking and the suspender belt. He also removed several inches of the external jugular vein, a large vessel from the neck; this is one of several major veins which return blood from the head towards the heart, so it is possible to sacrifice one of them without causing serious problems. Having excised this vein graft, Murray then turned it inside out. This was a crucial detail, since the inner surface of veins and arteries, the endothelium, has one particularly desirable property: because it is in constant contact with the bloodstream it must necessarily deter clotting.58 Murray found a novel way of inserting the section of inverted vein through the cardiac wall and suturing it in place, all without opening the heart. The sling of vein tissue was fixed in position with slight tension, so that when the ventricle contracted in systole the increased pressure would push it flat against the mitral opening, like a newspaper blown against a railing – preventing any blood from flowing backwards into the atrium. Though a number of the twelve dogs operated on died during the operation or shortly afterwards, two survived, apparently quite healthy.59
Murray’s research was treated with what he described as ‘supercilious amusement’ by his colleagues, but he was absolutely in earnest, applying the technique to several patients; one was transformed from a bedridden invalid to an enthusiastic golfer,60 though the fate of the others was not recorded. But there was anger when Murray revealed that he had translated an essentially unproven therapy to humans after such poor results in the animal laboratory; moreover, heart valves were still considered the realm of the physician, not the surgeon. He was summarily banned from repeating the operation.61
By the late 1940s, when Murray resumed his research, heart surgery was an established and rapidly advancing discipline. Others were trying techniques similar to those he had pioneered: John Gibbon and his collaborator John Templeton used grafts of venous and pericardial tissue to reconstruct the tricuspid valves of dogs.62 Murray improved this procedure by wrapping the vein graft around a tendon taken from the forearm to strengthen the new leaflets, and succeeded in keeping one dog alive for an impressive seven years after surgery. His human patients did less well: eight of the ten survived, their condition described only as ‘fairly satisfactory’.63
But a still more exciting era was on the horizon. While all attempts thus far had refashioned existing body parts into valve substitutes, researchers had begun working on a mechanical artificial valve – a man-made replacement for human tissue. Charles Hufnagel started his research at Peter Bent Brigham Hospital in Boston, where Elliott Cutler was still professor of surgery. In 1947 he discovered that it was possible to insert a tube made from methyl methacrylate (a tough polymer now used as bone cement in hip replacements) into a dog’s aorta so that it acted as an inner lining to the vessel.64 No clotting ensued; this was a highly significant finding, as it proved that an artificial valve made from synthetic materials would be tolerated by the body. A couple of years later a surgeon from Oklahoma City, J. Moore Campbell, unveiled an experimental device which he had inserted into the aortas of dogs.65 A hard plastic sphere the size of a pea was confined in a plastic tube in such a way that when the heart was relaxed the ball blocked the entrance to the valve; when it contracted, the increased pressure in the aorta pushed it aside and blood could flow forwards through the tube.
By coincidence, Hufnagel had hit upon a very similar design for his own valve – but, unlike Campbell, he took the next leap and used it to treat a patient. The first person ever to be given an artificial heart valve was a thirty-one-year-old woman with severe aortic regurgitation. A bout of rheumatic fever at the age of five had seriously damaged her aortic valve, and since her early twenties she had been in decline, afflicted with chest pain and breathlessness. In September 1952 Hufnagel implanted his valve into the woman’s aorta.66 It did not replace her diseased one, because the device was too large to be implanted in its proper anatomical position at the top of the heart. It was placed instead in the descending aorta, some distance from the organ; this did not entirely correct the problem, but stopped around 75 per cent of the anomalous blood flow.67 The patient made a full recovery and was even well enough to work full time, something she had not done for almost a decade.
From the outset it was clear that Hufnagel’s valve was a major advance. If suitable patients were chosen they did well, and mortality was low.68 But it was an imperfect facsimile of human anatomy, and had another serious drawback: it was noisy, emitting a loud click with every heartbeat. Early recipients of the Hufnagel valve could be heard from the other side of a room; one had to give up playing poker because every time he received a good hand the valve started to click violently.69 Later improvements made it quieter, but not before some patients had become so distressed that they killed themselves, driven to distraction by the incessant ticking inside their chests.70
Artificial valves were now a fashionable area of research, stimulated by Hufnagel’s success and the arrival of cardiopulmonary bypass, which made it possible to cut open the heart; a bewildering variety of designs was produced and tested, with varying degrees of success. Before the advent of the medical device industry in the 1960s there was little money for research, which meant that most of these prototypes were made in garages or at kitchen tables. The first prosthetic mitral valve ever implanted in a human was a prime example of such a Heath Robinson contraption. It was invented by a surgeon from Sheffield, Judson Chesterman, who was an amateur motor mechanic and based his design on the valves of a car engine. The hospital plumber made him a prototype out of copper, but Chesterman rightly had doubts about placing a metal object inside the heart. The next model was constructed from Perspex by a lab technician, Clifford Lambourne, who polished the valve to the necessary high finish with a silk handkerchief during cinema outings with his wife. On 22 July 1955, Chesterman implanted this device into the heart of a thirty-four-year-old man. The patient died fourteen hours later; although he had become the first surgeon anywhere to place an artificial valve inside the human heart, Chesterman did not pursue his research.71
In September 1960, 200 experts from around the world, including many of the most eminent names in heart surgery, gathered at the Edgewater Beach Hotel in Chicago for a conference on the future of the artificial valve. Countless different approaches were discussed, with little consensus reached as to the ideal materials or design. There were many tales of failure; the only success story came from Nina Braunwald, a thirty-two-year-old surgeon at the National Heart Institute in Maryland. One of the first female cardiac surgeons, Braunwald had grown up in Brooklyn and completed her training under Charles Hufnagel. Working with her colleague Theodore Cooper, she set out to manufacture a prosthesis which would closely mimic the appearance and function of the original. She and Cooper took plaster casts of normal mitral valves, and used these to fabricate replicas from polyurethane; artificial chordae tendineae of woven Teflon were then attached so that the leaflets could be anchored to the inside surface of the heart.72 After successful testing in dogs, Braunwald implanted the valve in five patients with mitral regurgitation. Four died shortly after the operation, but a fifth lived on for three months – the first time a mitral implant had lasted more than a few hours.73 Among the first to congratulate her on the achievement was Albert Starr, who in his comments after her presentation revealed that he was working on a valve of his own, which he felt had ‘much promise’.
In an address drawing proceedings to a close, the conference chairman Dr Alvin Merendino acknowledged the frustration felt by many researchers, but urged them to stay positive: ‘Unfortunately, no one unveiled the valve. Yet, in all fairness, it must be said that the Conference ends on a note of real encouragement.’74 His optimism was entirely justified: two weeks later Albert Starr operated on Philip Amundson, and there was little doubt that the valve had arrived.
Starr’s interest in the problem of valve design had begun serendipitously towards the end of 1958, when a retired engineer called Lowell Edwards made an appointment to speak to him. The elderly man with a Parkinsonian tremor who turned up at Starr’s office did not at first impress, dressed in golf kit and trainers and offering some very strange ideas. Edwards explained that he had become interested in the human circulation, and felt that with help from a medical expert he could build an artificial heart. While agreeing that this was an exciting idea, Starr pointed out that ten years of research by surgeons had failed even to produce an artificial valve. So the two men agreed to work together on a less ambitious project: developing a mitral valve prosthesis.75 As Edwards left his office, Starr wondered whether this shabbily dressed old man who claimed to be a wealthy inventor was a crank; any lingering doubts were banished when he saw Edwards getting into a smart Cadillac parked outside.76
Though unconventional, Edwards was anything but a crank. Engineering was in his blood: decades earlier his father, an amateur mechanic, had built a generator – and a steam engine to power it – to provide the first electric light in his town. Inspired by his example, Lowell trained as an electrical engineer before moving into hydraulics. As a young man he developed a machine that used water jets to strip the bark off logs – tremendously useful in Oregon, where the timber industry dominated the local economy. But his most significant invention was a fuel pump for planes flying at high altitude. During the Second World War this booster pump was installed in almost all US military aircraft, and Edwards became a rich man. By the time he met Starr he was living in comfortable retirement, funded by the proceeds of more than sixty patents.77
Albert Starr was the younger man by almost thirty years, but there was no sense that he was the junior partner in this new enterprise. Another protégé of Alfred Blalock, at the age of thirty-two he was already an experienced surgeon and a military veteran, having served as a medic in the Korean War and performed more than a thousand operations in battlefield conditions. Starr and Edwards threw themselves into their research, meeting at least once a week to discuss possible materials and designs. They assumed, like most of Starr’s colleagues, that a successful prosthesis needed to resemble natural anatomy, and their first prototypes were closely modelled on the human mitral valve, with two leaflets made from flexible plastic.78 When Starr implanted this device in dogs the results were uniformly poor. The animals died within a couple of days: autopsy revealed that clots were forming on the stitches securing the valve to the heart, and growing until the valve was completely occluded. After several dogs had died Starr had a sudden insight: what if mimicking natural anatomy was a red herring? Maybe haemodynamics – the way the blood flows through the valve – was more important than what the device looked like.
Freed from trying to imitate the real valve, Starr and Edwards looked for an alternative. One obvious precedent was Hufnagel’s device, which made no effort to replicate human anatomy; another inspiration was a valve recently investigated by Henry Ellis at the Mayo Clinic in Minnesota.79 It had no leaflets, but instead used a plastic ball trapped inside a cage of three curved metal struts. The circular base of this valve, which was lined with a ring of cloth, was sewn into the mitral annulus between the left atrium and ventricle so that the cage protruded into the ventricle. When the heart was relaxed and pressure in the atrium was higher than in the ventricle, the ball was pushed to the other end of the cage, allowing blood to flow forwards into the ventricle. As the ventricle contracted, the increase in pressure would push the ball back into the aperture, blocking it and preventing any reverse flow of blood. This was a venerable design, long used by engineers as a valve in a range of applications – it can be traced all the way back to a patent filed in 1858 by one J. B. Williams for an ‘improved bottle stopper’.80
Working in a shed attached to the summer cabin he owned in northern Oregon,81 Edwards soon had a prototype and the immediate results were strikingly better: dogs implanted with the device now survived for weeks rather than days.82 But first attempts are never perfect, and his experiments gave Starr valuable information on where they were going wrong. Edwards had plenty of time to devote to the project, and was able to produce a new prototype valve for testing every few weeks, allowing Starr to assess a wide variety of designs.83 He implanted these into over forty dogs, with gradual improvement; one, a Labrador called Blackie, survived for thirteen months.84
Early in the summer of 1960 the hospital’s chief of cardiology, Herbert Griswold, visited Starr’s animal laboratory and found it full of healthy, happy dogs with mechanical valves clicking away inside them. Deeply impressed, Griswold urged Starr to transfer his research to humans, pointing out that he had dozens of patients with mitral disease who might benefit.85 Starr was at first reluctant: he had kept dogs alive with the valve for months, but this was no proof that it would perform adequately for years in a human. If a valve were to last for twenty years it would have to open and close more than 800 million times.86 Manufacturing a device so durable was at the boundaries of what engineering could achieve. Luckily, Starr’s assistants had devised a machine which opened and closed the valve an astonishing 6,000 times a minute, which meant that three weeks of testing simulated forty-three years inside a patient.87 When they showed Starr that valves put through this ordeal showed negligible wear, with the hard plastic ball virtually unchanged in diameter, he was finally persuaded that the device was ready for human trials.
His first patient was a young woman who had undergone two previous operations for mitral stenosis and was now so unwell that she was forced to spend all her time in an oxygen tent. Starr was pleasantly surprised to find the operation technically much easier than working on the smaller canine heart. When his patient had woken up from the anaesthetic, he and the hospital’s chief of medicine, Hod Lewis, went to see her. Starr watched with amusement as his boss bent over the patient with his stethoscope, his moustache twitching as he listened to the unaccustomed click of the device inside her heart.88 Everything seemed to indicate a successful outcome, until a few hours later she turned over in bed and suddenly died. An X-ray showed that the operation had left bubbles of air inside the heart which had escaped and lodged in the brain, killing her instantly.89
When Philip Amundson arrived for surgery, in poor health after two previous unsuccessful operations, Starr vowed that this would not happen again. On 21 September, Starr operated. Once he had opened the chest, he attached Amundson to the heart-lung machine so that the heart could be stopped. He made an incision in the left atrium, exposing the mitral valve. He then cut out the diseased leaflets, leaving just enough tissue to allow him to stitch the new valve in place. Twenty sutures were placed around the diameter of the annulus – tedious work, since they needed to be evenly spaced and placed at just the right depth. Next, the suture threads were attached to the fabric ring at the base of the prosthesis, which could then be gently lowered into place. The heart was allowed to fill with blood, and when Starr was satisfied that no air remained inside he closed his incision and turned off the bypass machine.90 For the first time in a decade, Philip Amundson had a fully functioning mitral valve.
Amundson made an excellent recovery, the best possible fillip for Starr’s confidence in the new device. Of his next six patients only one died; the other five were greatly improved, an excellent outcome given that Starr had been allowed to operate only on the sickest and most hopeless cases.91 When he presented his results at a surgical meeting early in 1961, Starr admitted that he had at first found the unnatural ball-and-cage valve ‘repugnant’; but its success could not be denied, and within months it was being used in hospitals all over America.92
It was not long before the Starr–Edwards device had competition. New designs appeared throughout the 1960s, and it became a mark of status for a surgeon to have his or her name attached to a valve: Braunwald, Cooley, DeBakey and Lillehei all received this honour. Some differed from the ball-and-cage pattern, using a free-floating or tilting metal disc to control the blood flow. But none proved as successful as the Starr–Edwards, which continued to dominate the market for years. Though it had triumphantly answered the surgeons’ prayer for a simple and reliable artificial valve, it was not without its faults. In particular, it was bulky: it did not work properly in some patients who had an unusually small aorta.
The search for something better led to one of the worst scandals in surgical history. In 1979 a new device, the Björk–Shiley convexo-concave valve, was released on to the market. Within months of the first implantation, reports started to emerge of patients dropping dead without warning. Owing to a production fault a part was prone to coming loose, falling into the bloodstream and causing catastrophic regurgitation, but – disgracefully – almost 86,000 devices were implanted in patients before the faulty models were finally withdrawn from sale. By 2005 more than 600 of them had failed, making the Björk–Shiley valve the most dangerous medical device ever to be used clinically.93 It later emerged that Pfizer, the company manufacturing the valve, had long known about the problem and hid the evidence from regulators, a transgression which cost it hundreds of millions of dollars in compensation and fines.94
This debacle might have destroyed all confidence in the safety of artificial valves, but luckily the Björk–Shiley device already had a successful competitor. In the early 1970s a young entrepreneur called Manny Villafana set up a biotechnology company to manufacture a valve with a novel design. Instead of using a ball or disc to regulate the blood it had two flaps like butterfly wings, attached with hinges to the centre of the valve opening. Villafana chose this ‘bileaflet’ design not for any scientific reasons, but as a marketing ploy: he reasoned that this would make it radically different from anything already available.95 Quite by accident, it turned out to be the best commercial decision he ever made. The device proved to be a vast improvement on ball-and-cage models, small enough to fit in every patient and offering a close approximation to the natural valve. First implanted in 1977, and soon imitated by other companies, the bileaflet mechanical valve proved so reliable that it is still widely used today.
Mechanical valves are known to be safe and effective: they last for decades, and patients can expect to live a normal life. Despite all this, fewer are being implanted every year. There is an alternative, one that was being developed in parallel with the Starr–Edwards valve and its successors – and half a century on, it is becoming the option of choice for many patients.
In the 1950s, as surgeons started to appreciate the myriad challenges of designing an artificial valve, Gordon Murray sought another solution. Ten years earlier, Robert Gross had pioneered the use of arterial grafts to treat coarctation, using pieces of blood vessel taken from cadavers to replace diseased sections of the aorta. Murray reasoned that this technique could be modified so that the graft included a functioning valve. In 1955 he operated on a twenty-two-year-old man, inserting an aortic valve taken from the body of a thirty-three-year-old who had died ten days previously.96 Like Hufnagel, Murray chose to implant it in the descending aorta, deciding that placing it in its anatomically correct position was too technically demanding. His patient’s recovery was rapid, and after eighteen months he was able to do hard manual labour. Eight further operations were equally successful, with the grafts still functioning well up to six years later.97
Though this was progress, a valve placed in an unnatural position was far from a cure. Murray’s grafts were implanted some 10 centimetres from the outlet of the left ventricle: although they reduced anomalous blood flow by around 50 per cent, a significant volume could still travel backwards from the major vessels of the upper half of the body and back into the heart. This residual regurgitation reduced the amount of blood pumped with each stroke, putting the organ under considerable strain. In the summer of 1962 a surgeon in London finally succeeded in placing an aortic graft in its natural position.
Donald Ross was born and raised in South Africa – where one of his classmates had been a certain Christiaan Barnard – before moving to the UK. He was interested in the idea of transplanting valves, but realised that without some method of preservation, suitable grafts would be difficult to come by; surgeons could not rely on a suitable donor having died in the previous week. He learned of the work of two researchers in Oxford, Carlos Duran and Alfred Gunning, who had found that if the valves were immersed in ethylene dioxide and freeze-dried they could be stored for long periods at room temperature.98
The first implementation of this useful technique happened more or less by accident. During an operation on 24 July 1962, Ross was attempting to repair the badly diseased aortic valve of a middle-aged man when, in his words, ‘the whole thing finally disintegrated and went down the sucker.’99 This was a calamity, since mechanical valves were not yet available in England. In desperation, Ross sent a colleague to get one of his experimental freeze-dried grafts and sewed it into the patient. This was merely a temporary measure: Ross intended to replace the homograft with a mechanical valve as soon as he could get hold of one. But that proved unnecessary, as the patient made a good recovery and lived for another three years. In his subsequent report Ross made a more radical proposal, suggesting that an even better replacement for a diseased aortic valve would be the patient’s own pulmonary valve, which could itself be replaced by a homograft. Though this might seem a complication too far, there was solid logic behind it: the two valves are almost identical, and although the pulmonary valve operates at lower pressure than the aortic, research had shown that when it was transplanted into the aorta it quickly thickened to compensate.100 Another five years passed before he put this plan into action, but the operation – known as the Ross procedure – quickly proved its worth. It was particularly effective with children, because the new aortic valve grew with the patient; many surgeons still use the procedure today.
Unlike many new techniques the use of freeze-dried homografts became an immediate success, and dozens more patients were soon operated on. There was much excitement about the development, which seemed to promise the most satisfactory solution to valve disease. But after a few years, patients started to return to hospital showing signs of valve failure, and when samples were examined under the microscope they showed worrying signs of deterioration. Homografts were not the answer after all.101
One of those watching with interest was a young doctor in Paris, Alain Carpentier. Destined to become one of the world’s great cardiac surgeons, Carpentier had developed a taste for innovation during his training, when he had come into the orbit of Robert Judet, a pioneer of the artificial hip. After some agonising he decided to specialise in cardiac surgery, excited by the rapid pace of developments in the field. In his first valve operations Carpentier used the Starr–Edwards prosthesis, but regular complications drove him to look for alternatives. He performed the first homograft implantation in Paris, but French law – which required a forty-eight-hour delay between death and recovery of a donor valve – made it virtually impossible to ensure that the grafts were safe to use.102
Most parts of the body cannot easily be transplanted, since alien tissue is soon detected by the immune system and attacked, a phenomenon known as rejection. The heart valves, though, are peculiar in that they are composed largely of collagen, a tough fibrous protein generally ignored by the immune system. This simplified the implantation of homograft valves, since grafts were unlikely to provoke rejection. It also raised an intriguing possibility: why not use grafts from another species? Animal collagen was no more likely to prompt an immune response than human tissue, and that way valves could be grown to order and made available wherever and whenever they were needed. These would be known as xenografts – from the Greek ‘xenos’, meaning ‘foreigner’.
For two years Carpentier and a colleague, Jean-Paul Binet, experimented with xenografts from various species. But which to use? What was needed was a valve of a suitable size which was anatomically similar to a human’s. The best match came from a gorilla they autopsied at Prince Rainier’s private zoo in Monaco. But the idea of breeding gorillas for their heart valves was patently ridiculous, and they found eventually that three animals were required to offer a variety of sizes: lambs provided a small valve, pigs a medium one, and calves the largest.103 In September 1965 they inserted a pig valve into the heart of a forty-seven-year-old patient in Paris, the first of eighty such procedures. Although the early results were excellent, degradation of the valve remained a problem, with many patients falling ill after two or three years. Carpentier had been using a mercury-based solution to preserve the grafts, and decided to look for a substance that would protect them for longer.
This was a chemical problem rather than a medical one, and Carpentier began his search by investigating the aldehydes, a group of compounds used in embalming and to tan leather. After working his way systematically through more than fifty of them, Carpentier alighted upon glutaraldehyde, a molecule already used in microscopy to ‘fix’ biological specimens in their pristine state for examination. When applied to valve tissue, Carpentier found that it not only made it invisible to the immune system but also strengthened it.104 After experiments showed that inflammation tended to appear at the junction between human and pig tissue, Carpentier began to mount the grafts in a polymer cloth frame; the addition of this man-made material created a composite graft for which Carpentier coined the term ‘bioprosthesis’ – an artificial device constructed from biological material.105
One further improvement lay in store, courtesy of a Swedish surgeon, Åke Senning, who had been working on a different approach. Dissatisfied with the rapid deterioration and unpredictable availability of homograft valves, he decided to make his own from scratch. After removing the diseased valve he used strips of fascia lata, tough tissue from the patient’s thigh, to construct a replacement.106 These worked well, and using the patient’s own tissue made rejection impossible. But each valve had to be painstakingly constructed leaflet by leaflet, an operation requiring terrific deftness. It was Donald Ross who realised that much time and effort could be saved if the fascia lata valves were constructed outside the body and mounted in a frame, which could then simply be sewn in place. Using this technique he was able to construct replacements for the aortic, mitral and tricuspid valves – sometimes using two or three in the same heart. The initial results were outstandingly good,107 but within three years the grafts started to fail, and Ross’s collaborator, the Romanian-born Marian Ionescu, started to look for a longer-lasting substance.
Ionescu realised that it was not necessary to use the patient’s own tissue; the success of porcine xenografts had proved that even valves from a different species could be well tolerated by the body. Building on Carpentier’s work, he used glutaraldehyde to preserve bovine pericardial tissue, the tough membrane surrounding the heart in cows. Once treated it was then cut into strips and fashioned into a valve with three leaflets, fastened into a cloth-covered wire frame.108 This new technique immediately opened a world of possibility: valves could be constructed well in advance, in a range of sizes, and stored more or less indefinitely at room temperature. A surgeon in need of a replacement valve could simply pick one of a suitable size off the storeroom shelf. Bovine pericardial bioprostheses would prove even more durable than pig valves, lasting up to twenty years without deterioration.
By the 1980s surgeons had two excellent alternatives for patients suffering from serious valvular disease: the mechanical valve or the bioprosthesis. But the two are not interchangeable. Although mechanical valves can last a lifetime, they introduce foreign materials into the body, increasing the risk of sudden blood clots. To minimise this danger, patients must take anticoagulant drugs for life. Tissue valves, on the other hand, are unlikely to cause clotting problems but have a maximum lifespan of around twenty years. In practice this means that younger patients are generally given a mechanical valve, while the elderly are more likely to receive a bioprosthesis.
Why, then, is curing valve disease not now simply a matter of selecting a suitable device and implanting it in the patient? That was indeed standard practice for many years, but today’s clinicians do everything in their power to avoid using devices they spent decades perfecting. Strange as that may seem, it is because surgical philosophy has evolved in parallel with technology. The heart valves we are born with have been perfected by millions of years of evolution, structures unmatched by even the most sophisticated prosthesis. Why replace the ideal valve with something inherently inferior?
This was the question Alain Carpentier asked himself one evening in 1967 as he left work at the Hôpital Broussais in Paris. Walking through its venerable arch to the street outside, he noticed that it resembled a mitral valve: the iron gates were the leaflets, anchored to the firm ‘annulus’ of the stone arch. It occurred to him that if the structure were partially destroyed, a decent architect would have little difficulty in rebuilding it, using some sort of physical support to restore the original geometry of the gates: ‘Obviously a surgeon would do the same for the mitral valve!’109
Carpentier’s insight would transform the treatment of valve disease.In many cases of mitral regurgitation, he realised, it should be possible to reconstruct rather than replace the diseased valve. His aim was to support the distorted annulus by implanting a firm ring around it, bringing together the leaflets so that they would meet in the centre of the valve and once more provide a tight seal. After testing several different designs, Carpentier concluded that kidney-shaped rings worked best.110 The new operation was known as annuloplasty, and in the years that followed its introduction Carpentier made an intensive study of the many ways in which valves could be distorted by disease. The result was a new range of techniques which could be used to repair the mitral valve: cutting out superfluous tissue, tightening floppy leaflets with sutures or even moving the chordae tendineae. This new arsenal in the battle against valve disease is sometimes known as ‘the French correction’, a pun coined by Carpentier himself.111 After thirty years of experience it was apparent that it offered something that no prosthesis ever could: a cure. A tissue valve is merely a temporary measure, while a patient with a mechanical device faces a lifetime on prescription drugs. In Carpentier’s words, ‘It is only cured if you make the effort to reconstruct the valve, to restore the mobility, to reshape the orifice as it should be, as designed by God Himself.’112
Surgeons have often been accused of playing God; many decades spent battling the ravages of valve disease finally taught them that imitating His works is sometimes preferable to replacing them.