The Rise and Fall of Modern Medicine

by James Le Fanu

  In 1943 Albert Schatz, a young postgraduate, joined Waksman’s department and set out to work through the many species of actinomycetes, searching for the elusive chemical that might kill bacteria in humans.

  I generally began my work in the laboratory between 5 and 6 in the morning, and continued until midnight or even later. I also prepared my own meals and ate in the laboratory. I worked at this intensity for several reasons. First, I was fascinated by what I was doing; it intrigued me no end. Secondly, I fully realised how important it would be to find antibiotics effective in treating diseases caused by bacteria and even more so tuberculosis. Finally, with an income of only $40 per month I simply could not afford to buy an automobile or have much social life. I therefore literally worked day and night in the laboratory. Many times I even slept there because it was already so late I could get only one or two hours of sleep at the most.10

  The research technique Schatz used was simple enough. In the ‘streak test’ he ‘streaked’ a strain of the actinomycetes bacteria in a line across an agar plate and then at right-angles he ‘streaked’ one or other of the bacteria that cause disease in humans. After a few days he looked to see whether its growth had been inhibited. Within a few months he had found streptomycin, which was dramatically effective in inhibiting the growth of tuberculosis germs. It did not take long to show that streptomycin cured guinea pigs experimentally infected with tuberculosis and soon afterwards a 21-year-old woman, Patricia, became the first patient to be cured with the new drug. Described as ‘a pretty dark-haired farmer’s daughter’, she was ‘wracked by persistent coughing, emaciated from consumption and pale, bathed in a fever of sweat and dying from widespread tuberculosis in both lungs’. Between 20 November 1944 and 7 April 1945 she received five courses of streptomycin injections, each of which lasted ten to eighteen days. Her temperature fell and the infection in her left lung melted away, implying her tuberculosis had at least been partially cured. But it had not been easy. Tuberculosis did not respond to treatment like other bacterial infections, as was only too clear from the protracted course of her treatment. There was, however, a storybook ending. Patricia recovered fully and, after leaving the sanatorium, married and became the mother of three fine children.11,12,13

  The second anti-tuberculous drug, para-amino salicylic acid or PAS, is a chemical variant of salicylic acid closely related to aspirin. In contrast to the long gestation and hard labour that led to streptomycin, PAS was the product of a single imaginative stroke of genius by a Danish doctor, Jorgen Lehmann. In 1940, while working in neutral Sweden, Lehmann received a letter from an old friend, Frederick Bernheim, containing a short article – just two paragraphs, 300 words in total – he had published in the journal Science. It reported the results of an experiment in which he had observed that the addition of just 1mg of aspirin to a culture of tuberculous bacilli increased the amount of oxygen they consumed by almost 100 per cent.14 There was no obvious reason why this should happen, but Lehmann, after thinking about the matter for a while, wrote to a Swedish pharmaceutical company, suggesting that were they to make a chemical variant of aspirin (or salicylic acid) to produce para-amino salicylic acid, it would be an effective treatment for tuberculosis. His reasoning was as follows: aspirin increased the amount of oxygen consumed by tubercle bacilli, so a chemical variant of aspirin might act as a ‘competitive inhibitor’, blocking its consumption of oxygen so the bacteria could not survive. It was a brilliant, logical and highly original idea – which, as will be seen, would subsequently lead to the discovery of several other important drugs. But regrettably, at the time no one believed him. His medical colleagues had experienced so many false dawns in the treatment of tuberculosis that it was only natural for them to be sceptical about a suggestion that a variant of aspirin could be the answer, but Lehmann finally overcame these well-justified prejudices.15 In a short article published in The Lancet on 5 June 1946 – ‘PAS and the Treatment of Tuberculosis’ – Lehmann reported a couple of short case histories, describing a fall in temperature and an improvement in appetite in two patients with tuberculosis. There was, however, at the time no evidence that PAS, unlike streptomycin, could actually cure tuberculosis in the lungs.16

  So when the Tuberculosis Trials Committee in Britain convened later in the same year streptomycin was at the time the only drug deemed worthy of investigation. Bradford Hill argued persuasively that the manner of that investigation should take the form of a clinical trial in which the decision about who should be given the drug and who should act as a ‘control’ should be determined ‘at random’. And why was this so important?

  The principle of the clinical trial could not be more straightforward: a simple experiment where the efficacy of a remedy is tested by comparing the outcome in those given it with that in a similar group who are not. If there is a measurable improvement in those receiving the remedy it can be presumed that it works. If there is not, it does not. The essence of a trial, then, is a comparison of the outcome between two groups. This is a modest enough scientific experiment and scarcely novel, having been used by the Parisian physician Pierre Charles Louis in investigating the benefits of bleeding in pneumonia (in which he found ‘no apparent difference in intensity or duration of symptoms between those bled and not bled’)17 and famously by the Scottish ship’s surgeon James Lind when investigating the merits of several cures, including lemons, for scurvy, whose devastating effect on British naval expeditions had become a major impediment to further expansion of the Empire.18

  The method of comparison in the clinical trial requires that the treatment be alternated, giving it to one patient but not the next, and so on, allocating patients to the ‘treatment’ or ‘control’ group at random. Bradford Hill explained it thus:

  There are various ways in which the random allocation of patients to the treatment groups can be carried out . . . preferably it is done by the construction of an order of allocation, unknown in advance to the clinician. Thus using OT to stand for Orthodox Treatment and NT for New Treatment we may have as the order for patients consecutively brought into the trial OT NT NT OT OT OT NT and so forth.19

  He then goes on to emphasise

  the advantages of this random allocation of patients ensures three things: it ensures that neither our personal idiosyncrasies nor our lack of balanced judgement has entered into the construction of the different treatment groups; it removes the danger that believing we may be biased in our judgement we endeavour to allow for that bias and by thus ‘leaning over backwards’ introduce a lack of balance from the other direction; and, having used a random allocation, the sternest critic is unable to say that quite probably the groups were differentially biased through our predilections or our stupidity.

  This would all seem entirely reasonable and the lucidity of Bradford Hill’s prose only makes it seem even more so. From the beginning of 1947 three London hospitals started admitting patients into the first streptomycin trial, in which fifty-five patients were to be given streptomycin for four months, the results being compared with fifty-two ‘controls’ who were to be ‘treated’ with bed-rest and, if necessary, the collapse treatment of the lung that Bradford Hill himself had undergone almost thirty years earlier. The allocation to ‘treatment’ or ‘control’ was determined by a series of random numbers devised by Bradford Hill and placed in a set of sealed envelopes.

  By the end of six months, twenty-eight of those given the streptomycin were markedly improved and only four had died, compared to fourteen deaths among those unfortunate enough to have been randomly allocated to the ‘control’ group. This finding, that those given streptomycin ‘did better’ than those who were not, scarcely justifies the considerable time and energy devoted to the organisation of a randomised trial but, in a way that Bradford Hill could scarcely have anticipated, his insistence that streptomycin be objectively evaluated was to be completely (if tragically) vindicated. There was a fundamental limitation in the use of streptomycin to treat tuberculosis. Patients certainly improv
ed but the requirement that treatment should last several months guaranteed that some of the tubercle bacilli would become resistant to the streptomycin and, when this happened, their condition subsequently deteriorated again.20 The streptomycin trial had been so intelligently organised that it was a straightforward matter to assess the potential seriousness of this problem of resistance, and the clear statistical presentation of the results brought home the full gravity of the problem – as revealed by an ‘update’ published three years later, by when thirty-two of the fifty-five patients treated with streptomycin had died.21

  The authoritative verdict of Bradford Hill’s first trial was thus much more compelling than the resolution of the issue of whether or not streptomycin worked. Rather, it made it absolutely clear that streptomycin represented ‘a false dawn’ where an initial impressive improvement in a patient’s condition was followed by a subsequent relapse that was closely related to the development of resistance. From this perspective there could be no greater vindication of Bradford Hill’s espousal of the objective evaluation of new treatments over the subjective impression of doctors.

  His methodical approach immediately pointed to the next step, which was to repeat exactly the same trial but this time, in the hope of combating the problem of resistance, combining streptomycin with Lehmann’s PAS. This second trial started in December 1948 and exactly one year later, long before the study was complete, the organisers took the unprecedented step of issuing an interim communication that they had ‘demonstrated unequivocally that the combination of PAS with streptomycin considerably reduces the risk of the development of streptomycin-resistant strains of tubercle bacilli’.22 With the publication of the full results in November 1950 the benefits of the combination of the two drugs was glaringly obvious: thirty-three of the participants in the first trial had become resistant to streptomycin, compared to only five in the second trial. No longer, as happened at the first trial, did patients respond initially to treatment only to die several years later from a recrudescence caused by resistance to streptomycin. With streptomycin and PAS the survival rate soared to 80 per cent.23

  This was not the end of the story. Over the next ten years tuberculosis treatment became ever more refined and successful, with the introduction of other drugs, notably isoniazid in 1952 and rifampicin in 1970.24,25 It soon became clear that the combination of three drugs was even better than two, and that if given continuously for a period of up to two years then virtually every patient could be cured of the disease. This happy situation persisted until the late 1980s when the difficulties of treating AIDS patients with tuberculosis led to the emergence of tubercle bacilli that were ‘multiply resistant’ to all anti-tuberculous drugs, raising the spectre that once again tuberculosis would become, as it had been prior to 1950, essentially an ‘incurable’ disease.

  As a final reminder of that time, it is appropriate to recall the fate of George Orwell, who died in 1950 aged forty-seven, just a few months before the results of the combination of PAS and streptomycin were published. Orwell had first been diagnosed as having tuberculosis back in 1938. His condition improved with bed-rest and lengthy convalescence in Morocco, but he suffered a relapse in 1946, soon after moving to a remote farmhouse on the island of Jura in the Hebrides, where he had gone to complete the last and greatest of his books, Nineteen Eighty-Four. He managed, through the influence of David Astor, the proprietor of the Observer, to obtain a small quantity of streptomycin. Initially all went well and a month after starting treatment he wrote to his friend Julian Symons: ‘I have been having the streptomycin and it is evidently doing its stuff. I haven’t gained much weight but I am better in every way.’ He was, however, one of the unlucky ones who developed a strong allergic reaction to the drug, in the form of a terrible rash and blisters such that he was no longer able to continue with treatment. His tuberculosis returned and he died in University College Hospital on 21 January 1950 from a massive haemorrhage into both his lungs. The famous literary critic Cyril Connolly wrote subsequently:

  The tragedy of Orwell’s life is that when at last he achieved fame and success he was a dying man and knew it. He had fame and was too ill to leave his room, money and nothing to spend it on, love in which he could not participate; he tasted the bitterness of dying. But in his years of hardship he was sustained by a genial stoicism, by his excitement about what was going to happen next and by his affection for other people.26

  Orwell’s fate has profound symbolic significance. Like the experience of the Oxford policeman Albert Alexander, who was the first person to receive penicillin, Orwell’s brush with streptomycin is a reminder to future generations of the difference that anti-tuberculosis drugs would make to so many people’s lives. Orwell died on the cusp of the paradigm shift. Another couple of years and he would have been spared the bitterness of a premature death to live on for several more decades. Who knows what else he might have achieved?

  In the aftermath of the brilliant and lucid manner in which tuberculosis had been shown to be a treatable disease, the Randomised Controlled Trial (shortened to RCT) blossomed, just as Bradford Hill had hoped, to become the standard way of evaluating new drugs. As his protégé Richard Doll observed in 1982: ‘Few innovations have made such an impact on medicine as the controlled clinical trial designed by Sir Austin Bradford Hill . . . thirty-five years later the structure, conditions of conduct and analysis of the currently standard trials are, for the most part, the same. Its durability is a monument to Sir Austin’s scientific perception, common sense and concern for the welfare of the individual.’27 A minority were unconvinced. In a letter to the British Medical Journal in 1951, ‘a blast of the trumpet against the monstrous regiment of mathematics’, a physician from Sunderland, Dr Grant Waugh, comments on

  the outbreak in epidemic form of a disease of pseudo scientific meticulosis. The symptoms of the condition are characterised by: a) evidence of a certain degree of cerebral exaltation; b) an inherent contempt for those who cannot understand logarithms; and c) the replacement of humanistic and clinical values by mathematical formulae. The systemic effects of this disease are apparent; patients are degraded from human beings to pricks in a column, dots in a field, or tadpoles in a pool; with the eventual elimination of the responsibility of the doctor to get the individual back to health.28

  Behind the bombast Dr Waugh was making a serious point, for, as will be seen, clinical trials were not infallible and when improperly conducted could give rise to false conclusions that could not be rectified by any amount of objectivity conferred by ‘randomisation’. The RCT, however, was to prove utterly indispensable in the evaluation of the explosion of new drugs that occurred in the 1950s and 1960s. The thalidomide tragedy in 1960 forced governments around the world to insist that all new drugs be formally tested for their effectiveness and safety in randomised controlled trials as a requirement for the granting of a product licence. Thus Bradford Hill established the gold standard by which the merits of modern drug therapy must be measured.29

  Epidemiological Proof: The Case of Smoking and Lung Cancer

  Bradford Hill’s second indestructible achievement in his annus mirabilis of 1950 was to show that smoking causes lung cancer. Nowadays this seems so obvious as to be unremarkable, but back in 1950 it was not, for the simple reason that as a direct consequence of two world wars in thirty years virtually everyone smoked. Tobacco had proved as much of a solace in the trenches at Passchendaele as during the London Blitz and, when not calming the nerves, ‘a smoke’ was at least something to accompany the endless cups of tea that filled the long, empty hours so characteristic of total war, when citizens were unable to pursue their legitimate occupations. It is easy to appreciate how difficult it could be to show that smoking caused lung cancer if everyone smoked, because both those with and those without the disease would be smokers. Indeed, only statistical methods could resolve this question, because statistics can see ‘below the surface’ of things to identify relationships that would otherwise remain obs
cure.

  It has been noted that lung cancer replaced tuberculosis in a metaphorical sense as part of the paradigm shift from one pattern of disease to another, but lung cancer also replaced tuberculosis in a literal sense in 1950, as for the first time the number of deaths from the disease – 13,000 – exceeded those from tuberculosis.30 And while the toll of tuberculosis rapidly receded over the next few years under the onslaught of anti-tuberculosis drugs, that of lung cancer soared. There are further interesting comparisons. The tragedy of both diseases was that their victims died young or, in the case of lung cancer, relatively young, in their fifties and sixties. And, just as tuberculosis prior to 1950 was essentially an incurable disease, so was lung cancer. Indeed lung cancer was the more grievous of the two illnesses as, with the very infrequent exception of those in whom the disease was detected early enough to be surgically removed, most patients died within eighteen months.31 From this perspective it is almost impossible to overstate the importance of Bradford Hill’s implication of smoking, as this dreadful, untreatable, escalating disease suddenly became ‘preventable’ through the simple expedient of people not smoking. And it is almost impossible to overstate just how significant this was for the subsequent development of medicine, as over the next fifty years this example of the ‘preventability’ of lung cancer was enormously influential in promoting the notion that most cancers and other common causes of death might also be preventable by similar ‘lifestyle’ changes (as will be explored in detail in the final section of this book).