Taking the Medicine: A Short History of Medicine’s Beautiful Idea, and our Difficulty Swallowing It

by Burch, Druin

  The other useful class of drugs for heart failure were the opiates, like morphine and heroin. They made people comfortable. If someone was having angina or a heart attack, opiates took away the pain. If the patient’s heart was beating weakly, and the lungs filling with fluid as a result, the drugs eased the awful feelings of suffocation and drowning. To this day, no one has done the tests to reveal whether opiates help people live or die in either situation.

  Only in 1960 did it become routine for doctors to press on people’s chests when their hearts stopped beating. Up until then, the accepted response was to crack open someone’s ribs, stick a hand inside, grab their heart and squeeze. Death rates from heart disease have now decreased by two thirds, but improvements in resuscitation techniques account for a tiny fraction of that change. Much more of the benefit comes from the fact that you are less likely now to be a smoker.1 A large proportion, however, comes from drugs. Clot busters in the hours after a heart attack are the most dramatic. Others, like the whole cocktail that wards off future heart attacks, are dull in comparison. You take them every day for years and they make you feel no better. They extend your life – and your ability to enjoy it in good health – all the same.

  What these drugs have in common is that all of them make only moderate differences. None of them is like streptomycin for tuberculous meningitis, or penicillin for someone dying of overwhelming infection. None of them has effects that are big enough to understand on the basis of a doctor’s personal experience. None of them, in fact, has effects big enough to be understood without the use of properly designed trials of thousands of people. Despite that, each one of them has saved vast numbers of lives. Taken together they have changed the nature of human heart disease.

  The first and most important of all these drugs is aspirin.

  Fever and pain are both aspects of the body’s response to physical insult or derangement. The fever is aimed at making life more difficult for invading organisms, whose metabolisms are potentially more vulnerable to temperature changes than our own. Pain and malaise are also things that the body creates intentionally, part of its method for protecting itself. If you feel unwell then you do less, leaving more energy available for healing. Or perhaps you avoid using an injured part of your body, allowing it to rebuild more reliably. Celsus, a first-century AD Roman writer, described fever and pain as being two of the components of inflammation, the body’s overall reaction to illness or trauma. Rubor, calor, dolor and tumour were his four sonorous cardinal signs – redness, heat, pain and swelling.

  To a large extent, Celsus was accurate. The body does have a generalised response to injury. We still use the term ‘inflammation’, although doctors are also fond of ‘inflammatory cascade’ (probably because it is longer). Along with the four responses that Celsus mentioned, there are others. The immune system is mobilised through various hormones. And the blood’s tendency to clot is increased, a helpful response given that bodily insults are often caused by trauma that makes you bleed.

  Aspirin, in a different way from thalidomide but with some of the same effects, damps down overall inflammation, interfering with the chain of biochemical changes that underlie it. Thalidomide’s mechanism is not fully understood, partly because it seems to act on a variety of different biochemical reactions. Aspirin seems more precise, chiefly inhibiting an enzyme called cyclo-oxygenase which is used by the body to control the production of locally acting pro-inflammatory hormones. That is why the same drug that reduces a fever also helps relieve pain and ease swelling. And in line with its overall ability to interfere with the processes of inflammation, aspirin makes your blood less likely to clot.

  In the 1960s, with the developed world increasingly frightened of heart disease, blood clots were gaining a lot of attention. Most heart disease seemed to be caused by blockages of the heart’s arteries, fatty plaques that steadily developed over time. That left a puzzle as to how something that grew slowly could cause very sudden problems. The emerging conclusion was that the fatty lumps within the blood vessels really did enlarge gradually, and as they did so their surfaces became increasingly inflamed. Eventually, triggered by the inflammation, clots formed, shutting off the small spaces down which blood was flowing. If that happened in a coronary, a vessel supplying the heart with blood, you had a heart attack. A chunk of your heart muscle died. If it was a big enough portion, so did you.

  The story of Lawrence Craven is a curious one, bridging the world of anecdote and quackery with the emerging one of reliable testing. Born in Iowa in 1883, Craven studied science and medicine in Minnesota, then spent the First World War as a captain in the US military. The war over, he moved to California and devoted the rest of his life to working as a family doctor. That included a lot of minor surgery. Removing large numbers of tonsils and adenoids was once thought to be an essential part of good medical care. The procedures were believed, on the basis of medical experience and medical judgement, to be extremely useful. Actual research has tended to say otherwise and rates of both operations have now plummeted. For Lawrence Craven, however, they were a big part of his working life.

  Craven’s first aspirin-related publication was a letter to the Annals of Western Medicine and Surgery. It resulted partly from his experience of performing these minor operations. He felt confident that aspirin made the blood less liable to clotting, and while this was not a new finding he did not think that it was widely enough appreciated. Craven worried about two things. First, he noted that rates of bleeding after removing someone’s tonsils or adenoids were on the increase. He put this down to wider use of aspirin for pain relief. Second, he wondered whether aspirin could prevent the blood clots that caused heart attacks.

  Before Craven, other doctors tried something called dicoumarol for the same purpose. Dicoumarol was discovered as the result of natural observations on cattle in 1921. Across North America, cows occasionally dropped down dead. Sometimes they were found to have small external wounds that simply never clotted. Other times they looked entirely healthy, but when vets opened them up they found massive internal bleeding. Frank Schofield, a Canadian vet, discovered that what the cows had in common was the consumption of a particular plant, sweet clover.

  ‘Sweet clover disease’ was a disorder of blood-clotting, resulting from cows eating badly cured hay that had become mouldy. Karl Link and other chemists at the University of Wisconsin managed to separate, crystallise and determine the structure of the causative molecule in the hay. Dicoumaral, as it was called, belonged to a class of naturally occurring substances called coumarins, used for their smell and their flavour in perfumes and in drinks. (Coumarins have a fragrance similar to freshly mown hay; chilled Bison grass vodka is a good example.) In 1948 Link suggested that an altered form of dicoumarol might be useful as a rat poison. In view of its being developed with the help of the Wisconsin Alumni Research Foundation, and being a coumarin, Link called the new substance warfarin.

  Although clinicians had shown interest in the effects of dicoumarol on man, Link found that they were reluctant to try out the new warfarin. The fact it had been ‘originally promoted to exterminate’ put them off. Then, in 1951, a member of the American military attempted to kill himself using the poison. Changing his mind after five days of taking only moderate doses, he reported to a naval hospital and was looked after until the drug had worn harmlessly out of his system.

  Warfarin turned out to be not only more powerful but also more clinically predictable than dicoumarol, and rapidly became the favoured medical means for reducing the blood’s inclination to clot. The drug was powerful, effective and clearly dangerous. The side effects were dramatic enough to be obvious. On balance, doctors were not sure if it was useful. Aspirin, Craven suggested, was safer.

  As a theory to raise, it was inspired. As something to believe in without proof, it was appalling. Aspirin was clearly a less potent drug that dicoumarol or warfarin. The side effects were rarer. Craven’s assumption was that it must therefore be safe. The idea that, li
ke dicoumarol, it offered a balance of risks, did not occur to him. To prove that aspirin reduced the blood’s ability to clot, Craven swallowed aspirin until his nose began to bleed. Milder side effects, in his mind, meant ones that could be written off. Milder benefits just meant benefits.

  Led on by this mental arithmetic of unthinking optimism, Craven set out to test his theory. His approach was medieval. He decided aspirin was likely to work and started handing out prescriptions. His letter to the Annals of Western Medicine and Surgery reported his experience of giving it daily to 400 people for two years. None of them, he said, had suffered a heart attack. Initially he thought it should be taken by all men between thirty and ninety years of age, later he dropped that to those who were forty-five to sixty-five, overweight and under-exercised.

  In 1953 Craven wrote to the Mississippi Valley Medical Journal. By then he reported having given daily aspirin to around 1,500 people. As before, he found that not one of them had suffered a heart attack while taking the drug. His justifications for prescribing it are interesting:

  The value of Aspirin (acetylsalicylic acid) in the general prophylaxis of coronary occlusion is suggested by observations accumulated during the past seven years. Concededly, the effectiveness of any type of prophylactic treatment is difficult to prove, and this applies especially to a procedure aiming merely at nonspecific prevention. Observations on healthy subjects can never be made under strictly scientific conditions, and resulting figures are only within limits suitable for statistical evaluation. Such findings may therefore merely have the value of preliminary impressions, and will be substantiated or refuted by subsequent clinical research. But as long as the field of general prophylaxis of coronary thrombosis is still outside the limits of present-day research procedures, preliminary observations may still be of practical importance provided: 1. the measure is safe in all subjects and throughout the entire extended period of medication; 2. the observations are not in opposition to the trend and results of clinical and experimental research; and 3. it is well understood that the findings were not arrived at under strictly scientific conditions.

  The first sentence is fine. Leaving aside for a moment Craven’s choice of recommending aspirin to everyone he knew, his observation that it reduced heart attacks was reasonable. Starting off with a suggestion of benefit, and knowing that it is only a suggestion, was a good way to begin. The second and third sentences show the mental difficulties facing an obviously intelligent and educated doctor in the years after the Second World War. Craven was aware that his methods were not scientific. That is, they relied on impressions rather than tests, and he knew that impressions were not always enough to outweigh the fantasies of the investigator. Yet not only was he unaware of methods for accurately comparing two groups of people treated with aspirin, he decided that unreliable methods were good enough to be going on with. He pointed out the flaws in his beliefs but he trusted them anyway. He had a hunch that his judgement was sound.

  Craven’s speculation that research techniques might improve was mistimed. The methods of the randomised, double-blind, placebo-controlled trial were already in print. The MRC streptomycin trial was published in 1948, when Craven began his ‘experiments’.2 Five years later he was still unaware of it. But perhaps the strangest conclusion that Craven reached is that aspirin ‘is safe in all subjects and through the entire extended period of medication’. He had tested it on himself until his nose fountained with blood. He was proposing it as an alternative to the dicoumarol which was known to kill by causing haemorrhage. He had written about aspirin itself causing dangerous post-operative bleeding. Deciding that it was now entirely and at all times safe was exceedingly odd. We might not be totally certain of aspirin’s benefits, Craven was saying, but it can’t do any harm. Let’s give it to everyone.

  Eventually, Craven claimed to have used aspirin on 8,000 of his friends and patients. He reported that ‘not a single case of detectable coronary or cerebral thrombosis occurred among patients who faithfully adhered to this regime during a period of eight years’. It is an astonishing statement. Very few treatments are so effective as to completely eliminate all cases of a disease, and we know that aspirin is not one of them.

  What was the explanation for Craven’s perfect results? Most likely there is a clue in the sentence’s careful qualification. Craven says that there were no heart attacks (coronary thromboses) or strokes (cerebral thromboses) among those who ‘faithfully adhered to this regime during a period of eight years’. The implication is that there were both strokes and heart attacks, but that on close questioning Craven was able to satisfy himself that the patients who suffered had not been 100 per cent reliable when it came to taking their pills. This was probably true. No one is perfect at swallowing daily tablets. The advantage of not using a placebo control is that you can always come up with an explanation that handily accounts for whatever it is you want to explain away. Nine out of his 8,000 patients died, and Craven was thorough enough to make sure autopsies were performed. Some showed ruptures of the aorta, the main blood vessel emerging from the heart. Rather than being worried that these internal bleeds were due to aspirin, Craven felt vindicated that his patients had not died of heart attacks.

  Craven is an excellent representation of the sort of thinking that held medicine back. He was an astute, industrious, intelligent and well-intentioned man, but those qualities could not compensate for methodological errors in the way he came to his conclusions. His suspicion of aspirin’s benefits was correct, but it was a lucky guess – particularly lucky for the 8,000 people to whom he prescribed it on that basis.

  The medical profession ignored Craven for all the wrong reasons. Not because of the flaws in his approach but because he was an obscure family doctor without the academic warrant to come up with interesting ideas. Cardiologists were appalled that Craven could not explain how aspirin stopped blood from clotting. It says something about medical thinking that this was deemed to be a valid criticism. It was entirely clear that aspirin interfered with the blood’s ability to clot. The cardiologists who held the power and prestige, however, did not feel that this was important. There was no satisfying theory to explain how aspirin prevented blood from clotting; therefore the fact that it obviously did so was not interesting. What mattered was not demonstrable fact but missing theory.

  Lawrence Craven, despite his aspirin, died from a heart attack. It was 1957 and he was seventy-four. No one responded to his thoroughly appropriate calls for more interest in aspirin and more studies of its effects. In fact, no one looked at the therapeutic potential of aspirin for heart disease for years. ‘The reasons for the delay’, said the New York Times in 1991, ‘are not clear but they partly reflect the tendency of scientists to insist on understanding the biological mechanism for a treatment before studying it.’ That was a polite way of saying that doctors got more pleasure from complicated theories than from simple tests. The New York Times also commented on why Craven’s ideas were not taken up by pharmaceutical companies. Their interest was more reliably pragmatic, and much better suited to favouring proven effects over absent theories. The way the patent system was arranged, though, discouraged them from looking at aspirin. Patents get given out for drugs rather than for uses. Governments, in other words, have set up a system that rewards companies for coming up with new compounds but not for finding fresh uses for existing ones. Discovering a way of making aspirin save millions of lives was simply not something that offered strong financial incentives. Patent laws are there to reward those willing to stake their money on research. When it comes to encouraging the exploration of existing drugs, they fail.

  From 1967, reports began appearing that aspirin interfered with platelets, tiny fragments within blood that cause it to clot. That helped get doctors a little more interested, but it was not decisive. The actual way in which aspirin exerted its influence was still unclear. ‘Credit for influencing physicians to prescribe aspirin for heart problems’, says Gabriel Khan, a Canadian cardiologist,
in his 2005 Encyclopedia of Heart Diseases, ‘must be given to John Vane.’ Dr Khan’s justification is that it was John Vane who finally described the molecular mechanism by which aspirin operated.

  Lovely as it was, in terms of showing how it was that aspirin affected platelets and, through them, blood clots and then cardiologists, Vane’s work was not what prodded physicians into motion. John O’Brien did more. A haematologist (blood specialist) working in Portsmouth, England, he was unaware of Craven but also published a paper (this time in 1963) showing that aspirin prevents blood clots. O’Brien’s innovation was to show, by using measurements of platelet stickiness, that fairly routine doses of aspirin had clear effects. O’Brien was deliberately trying to find drugs that inhibited blood clots. Before reading about aspirin, he tried a range of other drugs. Some of them stopped clots, but only at doses that were likely to kill people. Once he felt confident about aspirin, O’Brien followed his first paper up with a Lancet article in 1968. There he suggested a trial of the drug for preventing heart attacks. By then others were thinking along the same lines.

  Sometime in 1968, O’Brien met up with Peter Elwood, a doctor whose interest in epidemiology had led to him to take up a post with Archie Cochrane in Wales. O’Brien had already tried to persuade the MRC to pay for a trial of aspirin in heart disease, but his calculations of the number of people required made the prospect too expensive. The effects of aspirin were likely to be moderate, reasoned O’Brien, and so large numbers of patients needed to be enrolled into any trial that was going to be able to distinguish the effects of aspirin from the shifting fortunes of chance. Elwood came up with an improvement. What was needed, he realised, was not so much a large number of people as a large number of actual heart attacks. A trial recruiting only those at high risk of heart attacks would need fewer people. And since one heart attack often predicted another, you had a straightforward way of identifying a large group of people who were definitely at risk.