The Testing of Drugs in Man
PROFESSOR J. H. BURN. M.D., F.R.S., was the author of the Twenty-Ninth Stephen Paget Memorial Lecture, which was delivered on Tuesday, 6th December, 1960, in the Physiology Lecture Theatre,UniversityCollege,Gower Street,London, W.C.1. The President of the Research Defence Society, the Right Honourable the Earl of Halsbury, was in the chair.
The President announced that Professor Burn sent his apologies because he had been taken ill and was unable to be present.
"I am sure you will wish me to send a message of good wishes to him and say how sorry we are that he is not here", said the Earl of Halsbury. Professor Burn had, however, made arrangements for his paper to be read for him by Dr. J. M. Walker, ofWorcesterCollege,Oxford.
Dr. Walker prefaced his reading of the paper by saying that as a former pupil and later colleague of Professor Burn he was very happy to do this, although he regretted that the audience would miss Professor Burn's enthusiasm.
The Testing of Drugs in Man
By PROFESSOR J. HAROLD BURN, M.D., Hon.Sc.D., F.R.S.
IN attempting the task of giving the Stephen Paget Memorial Lecture I am very sensible of the honour I receive in being invited to give it, and at the same time am conscious of the responsibility placed upon me. This is the one occasion in the course of the year when an attempt is made to consider the challenge of the antivivisectionists and to explain the methods and results of those experiments on animals which are permitted by the law of the land. I have read the account of Stephen Paget's work with care, and have been deeply impressed by the efforts he made, and by the feeling that if the same dangers rose again today we might not have a champion so distinguished and so devoted who would state the case on our behalf so persistently. How many of us to-day would be willing as Paget was lo give as many as 24 lectures in the last three months of 1911, and to tour the country as he did in 1914? He showed himself to be a man of very great public spirit who advocated a cause in the face of fanatical hostility, seeking for no reward, but sacrificing his health, because he considered it right to do so. Let us hope that if the need were to arise again, there would be found others who would attempt, as well as they could, to follow his lead and to persist as he did in winning the support of those who were not blinded by prejudice but who were willing to listen to an explanation of the facts. It has always been the neutral layman who has saved the day for us and the great majority of these have always been convinced by a clear statement of the position.
I myself have held a licence now for forty-six years, and in that time have carried out a great many experiments on both dogs and cats. In these the dissection was made on the fully anaesthetized animal and the animal was finally killed while still fully anaesthetized. I have therefore practised vivisection of the animal when fully anaesthetized. But this is not what the anti-vivisection societies represent me as doing. When they speak of "the thrusting of tortuous experimentation" upon animals, when they refer to "the vast total of animal suffering in the laboratories", they imply and intend readers of their words to understand that dissection is practised upon the animal which is not anaesthetized at all. The anti-vivisection appeal is expressed in words which indicate that while the process of dissection goes on, the animal is conscious all the time, and is suffering.
Yet the anti-vivisectionists who imply these things, whose whole activity is directed to implying these things, are aware that the dissection of the unanaesthetized animal is forbidden by law inGreat Britain, and that only those forms of experimentation are permitted which are allowed by the Act of 1876. The word vivisector is indeed an accusation that the man performing experiments within the limits allowed by law is guilty of a cruelty which the law forbids. It is, I submit, slanderous and libellous for the anti-vivisection societies to say that dissection of the unanaesthetized animal is practised when it is in fact illegal and it should not be impossible to compel those societies to change their very name by process in the courts. They could not maintain that the surgeon who performs a surgical operation under anaesthesia inflicts pain on his patient. Therefore they could not maintain that the dissection of an animal under anaesthesia causes pain to the animal. Therefore they must admit that since the animal dies while still anaesthetized it suffers no pain at all. Therefore it is slanderous and libellous to say that an animal dying in that way has suffered from vivisection.
I for one am grateful for the existence of the Act of 1876 for more than one reason, as I shall explain. But for the reason that it makes nonsense of the accusations of the anti-vivisectionists it is a great protection to those carrying out experimental work.
In my experience the Act has been administered with great wisdom. This certainly does not mean that the decisions have always been welcomed by the experimental worker. I can remember occasions when an Inspector has complained about the size of the cages in which rabbits were kept and has complained about cats being kept in cages at all. Although these complaints seemed irritating at the time, the effect of the Inspectors' complaints over the years has been to raise the standard of animal care in this country very materially. How much it has been raised I have realized when visiting animal houses in other countries where there is no inspection. It is true that very often the care of small animals is excellent, but it is also true that occasionally the care of larger animals is deplorable. This happens when the head of the department, always a busy man, does not visit the animal houses, and when those responsible for looking after them are lazy or indifferent. The inspection we have in this country is a safeguard against any form of neglect, and demands thought for the welfare of all animals.
On one point the provisions of the Act, laid down in 1876, seem to me out of date, namely when they refer to the use of curare. The anaesthetics used at that time were principally the volatile anaesthetics, chloroform and ether, and the absence of reflexes always indicated that the anaesthesia was sufficiently deep. However, if curare was injected this indication was no longer effective, and presumably for this reason the very severe rules about the use of curare were laid down. These rules are. however, surely unnecessary if an anaesthetic like chloralose, which is given by injection and which is not volatile, is used, because once full anaesthesia has been produced with chloralose, it remains for many hours, and there is no possibility of anaesthesia becoming lighter. In these circumstances it would seem reasonable to allow the use of curare without express permission being required, and without notification to the Home Office of the day on which the experiment is to be made. The present rules which are enforced are the more onerous because they refer not only to the use of curare but to the use of any substance having properties similar to it.
Another point on which the rules seem to me to be too severe is that they do not allow a medical student to learn to insert a glass cannula in the vein or the artery of an anaesthetized animal, even though he does so under the eyes of a holder of a licence. It is true that the medical student may do this provided the brain of an animal has first been destroyed, but many experiments on the heart and circulation cannot then be successfully carried out. I have often felt that it should be possible for medical students to expose the vein or an artery of a fully anaesthetized animal provided that a person who holds a licence is actually watching while the dissection is being done. It seems unfortunate that medical students have to make their first attempts on patients, who not being anaesthetized may suffer discomfort and sometimes pain. I have never understood why the unconscious anaesthetized animal should be given more consideration than the conscious patient.
The choice of a main topic for the Paget Memorial Lecture is not easy if it is to be one which has not been considered before and one which is of general interest. In the Paget Memorial Lecture of 1954 Sir Geoffrey Jefferson chose the subject of "Man as an experimental animal". I wish to-day to say something, which may at first sight seem similar, on the testing of drugs on man.
The number of new drugs which are introduced for the treatment of disease steadily increases. In 1958 Modell and Houde gave the number as about 550 per year in theUnited States. The question of testing these drugs on man is therefore formidable, in the first place because there are so many to be tested and in the second place because tests on man present difficulties of their own which arise because of the peculiarities of human behaviour. I will begin by referring to a paper written by Feldman (1956) whose task it was to put together the reports of a good many doctors who had been asked to test the action of chlorpromazine in their patients. Chlorpromazine as you no doubt know is one of the tranquillizing drugs, and the tests with it were made in mental hospitals on those who for the most part were suffering from schizophrenia.
There is no question that chlorpromazine is a very powerful drug because its tranquillizing action can be demonstrated in mice. There is a stimulant drug, amphetamine, which makes it possible to keep awake for long periods. If some of this is given to each of the mice, which are together in a box, the mice become very excited and appear to threaten one another, though they are not actually observed to fight. When chlorpromazine is given before the amphetamine even in a dose which, per unit of body weight, is a small dose for a man, the mice remain unexcited, and their mortality due to the amphetamine is reduced more than ten fold. The effect is striking because the dose of chlorpromazine is so small.
Despite the certainty that chlorpromazine is potent. Feldman found that there was little unanimity among the reports on the patients. The reports were submitted on a form which required appraisal of changes in 25 different aspects of clinical behaviour: these could be graded on a 4-point scale. There were 37 physicians who between them made observations on subjects, and it seemed clear that the personality of the physician played an important part in determining what results were obtained. Feldman found that the doctors could be divided o four groups. In the first group were those o received the new drug with enthusiasm, e usual treatment for schizophrenia was by electric shock. The attitude of the physicians in; first group was that chlorpromazine was the drug "we have been looking for. Now we can row away the shock machine ". As an example of this class was a doctor who tested the effect of chlorpromazine in 44 patients, and observed improvement varying from moderate to marked in 34 of them. Thus chlorpromazine provided effective treatment for 77 per cent of his patients. Now schizophrenia is a chronic affection which has proved refractory to most other forms of therapy, and in view of the many factors which make the outlook gloomy, the result obtained was no small accomplishment. The report, of course, might be attributed to the doctor's personal bias, but from among his 44 patients. 12 were released by the social service to return home, and 13 others were in process of release.
The second group of physicians contained the majority, and their attitude was calmer. They were willing to try chlorpromazine in order to see how it worked. They concluded that it had some value and they attempted to define the circumstances in which it should be used. They were willing to include it as one of their methods of treatment. The attitude of the third group was "characterized by open, but only partial rejection of drug therapy". They took the view expressed by one of them that the use of chlorpromazine "is contrary to our dynamic concepts, detrimental to our teaching program; our residents will waste their time with drugs, instead of observing and learning dynamic psychiatry". One member of this third group who tested chlorpromazine on 30 patients reported slight improvement in 26; he saw moderate improvement in only three. Only two of his patients were released from the hospital, and one other was in process of release.
The fourth group contained doctors who refused to test chlorpromazine. Feldman came to the conclusion that "conscious and unconscious factors determine how any given individual will report the efficacy of a drug". It also seemed to follow from his paper that the actual number of patients benefited depended a great deal on the doctor's readiness to give the drug a proper trial. Even when the patients are suffering from a similar affection, and are tested in reasonable numbers, and when the drug has a clearly demonstrable action, the reports of different doctors may vary greatly, and the results they obtain may also vary greatly.
Having considered the variations in doctors, we may now consider variations in patients, which as you can imagine throw the testing of drugs into still greater confusion. We all know that when a patient visits his doctor he expects and likes to receive some form of treatment. This used to take the form of a bottle of medicine, and may often do so still, though it is now usual for tablets to be prescribed or for some course of injections to be given. Often, however, the doctor cannot quickly discover what is wrong and therefore does not know what medicine to choose. In this situation the doctor gives a placebo.
There have been many tests carried out which demonstrate that placebos, which should of course have no therapeutic value, often seem to have one, and that this is particularly true in relation to the relief of pain. Professor Beecher ofHarvardUniversityhas made many studies on this subject. In 1953 he described the results of observations on patients which began 48 hours after they had had a surgical operation: these patients were chosen because they were still suffering from persistent pain. To a first group of 40 patients he gave at two-hour intervals alternately a capsule containing 10 mg. morphine and a capsule containing lactose. Both capsules, of course, were given by mouth. Of the patients no less than 32 per cent said that their pain was relieved by lactose, and no more than 41 percent said that their pain was relieved by morphine. Then he took a second group of 36 patients: in these he compared the effect of 10 grams of aspirin in a capsule, with the effect of lactose in a capsule. Of these patients 26 per cent found that their pain was relieved by lactose, while 55 per cent found that their pain was relieved by aspirin. Thus in both these tests lactose was effective in one-third to one-quarter of the patients, and surprisingly also it appeared that aspirin was more effective than morphine when given by mouth.
In a second studyBeecherused injections, comparing the effect of morphine with the effect of sterile saline. The patients were again those whose post-operative pain still continued after 48 hours, and thus was persistent. On various occasions an injection of sterile saline was given instead of morphine, the patient not being told. When asked at the end of 45 minutes, and again at the end of 90 minutes, whether the pain was less, only 21 out of 69 said that the injection of saline gave no relief. That is to say only 21 out of 69 were not deceived by saline. There were 38 patients who were sometimes relieved by an injection of saline, and sometimes were not relieved. Finally there were 10 patients who were always relieved by the injection of saline. Beecher and his colleagues later interviewed all 69 patients to see if there were detectable differences in their personalities. Differences were indeed observed. Thus they noted that those patients whose pain was relieved by saline were all patients who thought that treatment in hospital was wonderful! Of those whose pain was not relieved by saline, only a minority took this enthusiastic view. It was further noted that those who got relief from saline were more talkative, more concerned about themselves and more inclined to weep during the interview.
Another study was made by Wolf and Pinsky (1954) to test the claim that the drug mephenesin exerts a specific effect on subjective anxiety and feeling of tension, and also on objective manifestations of this state. They compared the effect of 0-5 gramme mephenesin given in tablets with the effect of lactose given in similar tablets. The comparison was made at theNew YorkHospitalin 31 patients during several periods of two weeks each. The tablets were prepared in six batches with three different identification labels, though three batches contained mephenesin and three batches lactose only. At the end of the investigation there appeared to be no difference between the mephenesin and the lactose. Whether taking mephenesin or taking lactose, between 20 and 30 per cent of the patients felt better, 50 to 70 per cent were unchanged and 10 to 20 per cent felt worse. For our purpose the important point was that from 20 to 30 per cent felt better when taking lactose.
A more objective research was that described by Hillis (1952) who studied the effect of drugs in preventing cough. He worked inGlasgowand all his observations were made on one volunteer. Hillis inserted a long naso-pharyngeal sprayer as far down as the lower pharynx, which was tolerated for long periods without difficulty by the man. This was found to be a reliable method of introducing into the larynx small quantities of fluid sufficient to produce coughing. Hillis used ether and also peppermint water, which were both very effective. A comparison was made of heroin, morphine, amidone, codeine and saline, all given by injection. The comparison was thorough and was made on many occasions over several months. As might be expected, heroin was the best substance to prevent cough, and morphine and amidone were next best.
What was of great interest was that Hillis was unable to observe a significant difference between the effect of codeine given subcutaneously in 60 mg. doses and the effect of physiological saline : both prevented cough to some extent, though less than the other three drugs. We are left to decide for ourselves whether physiological saline is as good for stopping cough as is codeine, or whether codeine is as useless as physiological saline.
Several other investigations have been carried out in which tablets containing only lactose have been substituted for tablets containing a drug of known therapeutic value. In 15 different investigations, seven of them from Beecher's laboratory, and eight from other laboratories, it was established that tablets containing only lactose were able to arrest cough, were able to relieve the pain of angina pectoris, were able to relieve sea-sickness, were able to relieve anxiety and tension, and finally were able to relieve the common cold. When all the observations were taken together, they were made on more than 1,000 patients, and the proportion of patients in whom the tablet of lactose gave relief was about 35 per cent. That is to say a dummy tablet was effective in about one patient in three in these different ways.
The Double-blind Trial
The difficulty of knowing whether a new drug is really active in man or whether its effects are those of a placebo, has led to the introduction of the " double blind '' trial. In such a trial there are, so to speak, two blind men. One is the patient who is unaware whether a tablet is a genuine tablet or a dummy, and the other is the doctor who gives the tablet and judges its effects. He also does not know whether a tablet is a genuine tablet or a dummy. The tablets are handed to the doctor by a third person who alone has the key to their composition.
This method is ideal for determining the action of a new drug when its effects are sufficiently objective. This was well illustrated in the test of streptomycin carried out on patients with tuberculosis for the Medical Research Council. The appeal of the double-blind trial has indeed been so great that the proposal was made inDenmarkto set up an institute for the testing of new medical remedies by the double-blind method. About 70 new drugs are introduced inDenmarkevery year. The cost of the institute would have been great, but it was suggested that the saving effected by discarding useless remedies would also be great.
The double-blind trial can. however, be misapplied, and when it is misapplied it can give disquieting results. Thus Batterman and Grossman (1955) were very disturbed when they tested well established analgesic substances by this method. They said " It is with surprise that we note the same effectiveness for placebo medicaments, known effective analgesics and unknown drugs. Regardless of the medicament administered we always achieve the same result. There appears to be an incidence of effectiveness that occurs with all medicaments with the double-blindfold technique regardless of what preparations are being investigated". They further said that placebos which are by usual methods 40 per cent effective are by the double-blind method 60 per cent effective. They came to the conclusion that no difference between tablets of aspirin and tablets of lactose could be detected unless at least the physician knew what he was giving.
Modell and Houde (1958) have discussed the findings of Batterman and Grossman and have made some observations of their own. They first of all made tests very much in the same way as Batterman and Grossman. They studied 25 patients with arthralgic pain, and treated them with aspirin in doses of 250 mg, 500 mg and 1,000 mg. Each of these doses was given to a patient three times a day. Each patient was also given tablets of lactose three times a day. The results as recorded by the patients at the end of two weeks indicated no difference between aspirin and tablets of lactose. Thus when taking tablets of lactose, 38 per cent were improved, 38 per cent were not affected and 24 per cent were worse. When taking 1 gramme of aspirin, 44 per cent were improved. 31 per cent were unaffected and 25 per cent were worse. However, Modell and Houde said that the patients did not keep trustworthy records, some of them only filling up their cards when they came to the clinic at the end of two weeks, by which time they had forgotten what degree of pain relief they had had. Batterman and Grossman had obtained their results in this way. Having given bottles of tablets labelled SI, S2 and S3 to each patient, they had told him to return to the clinic after one to three weeks and had then attempted to find out from him which bottle had produced 'he most relief. Modell and Houde then made further observations in which the patients were again given tablets, but this time, the patients communicated what relief of pain they feltdirectly to an observer during the five-hour Period after taking the tablets. In this study a vicar difference was obtained, aspirin producing a much greater and more prolonged effect than lactose. Modell and Houde emphasize that the difficulties some observers have had in applying the double-blind test to substances of known efficacy can be explained by insufficient care in the collection of results, and by too much reliance on the patient's memory.
Effect of Placebos on Side-effects
There is still another side to the giving of a placebo, for not only may it have curative effects, but it may also exert side-effects or unwanted effects, as for example does morphine when it causes vomiting. It is not on reflexion wholly surprising that, when a patient who is suffering from pain is given a tablet in circumstances which suggest that it may relieve his pain, he should feel some relief. It does, however, seem strange that he should have side effects. Yet the following symptoms are on solemn record. Out of 92 persons given a placebo. 23 complained that they suffered from a headache, 9 complained of nausea and 14 complained of difficulty in concentrating. On another occasion, out of 72, 14 complained of a feeling of heaviness and 7 complained of a dry mouth: 6 said they felt a warm glow. On a third occasion, out of 72, no less than half complained of drowsiness, and 7 of them went to sleep. On a fourth occasion, out of 57, 10 complained of fatigue, and 5 said they felt relaxed. All these effects were a consequence of taking a tablet containing lactose.
The effects in some individuals were more serious. In the course of the observations made by Wolf and Pinsky already described, on a fifth occasion when 31 persons took a tablet of lactose, three of them suffered as follows. One had epigastric pain followed by watery diarrhoea; there was also an urticarial rash and swelling of the lips. These symptoms were produced on three occasions in this subject. A second subject developed a diffuse skin rash described as itchy erythematous and maculopapular. The rash was diagnosed by a dermatologist as dermatitis medicamentosa. A third subject experienced a feeling of great weakness, palpitations and nausea.
A different problem created by the placebo has been demonstrated by Modell and Garrett (1960). All persons have a certain degree of finger tremor of high amplitude. This finger tremor was measured in 12 male students, and found to be at the rate of 60 per minute. The tremor was increased by mental work when the students were competing with one another in performing a particular task, and the figure rose from 60 to an average of 152. Now the point of interest was that when the students were given a dummy tablet beforehand (not knowing that the tablet was a dummy), and then repeated the exercise, the finger tremor rose to an average of 198. Thus the taking of a dummy tablet had made the degree of finger tremor worse. It was clear that the taking of a tablet had created apprehension in spite of the fact that the subjects were aware that the purpose of the investigation was to test tablets for their power to allay apprehension. When, instead of a dummy tablet, a tablet containing 100 mg phenobarbi-tone was taken, the finger tremor rose to 162 only. Thus the phenobarbitone reduced the apprehension created by the taking of a tablet, but not so much that the result was better than no tablet at all.
The Flood of New Drugs Today
The evidence that has been described makes it clear that the testing of new drugs in man is not easy. As a result there must always be uncertainty whether a new drug has a useful action which is peculiar to it and not possessed by other drugs already in use. The new drugs which have been introduced during the last eight or nine years as tranquillizers are very expensive, and therefore when prescribed in this country add appreciably to the taxes we all have to pay.
There are doubtless large numbers of chemical compounds to be found which will tranquillize a fierce monkey or dog, and some of these when tested clinically may appear from results in the first one or two hundred cases to be freer from side effects than their predecessors. These are the drugs which are selected for advertisement to the medical profession and. as can be readily imagined, it often remains uncertain whether they have any special properties which are useful. There are two tranquillizing agents about which there is no doubt, and these, reserpine and chlorpromazine, both have actions which are sufficiently well defined to be studied in animals in many different ways. Concerning most others this is not true. For some of them effects have been discovered in animals, but these effects are not obviously useful in man.
To me it seems at least a tenable view that drugs which are effective in the brain will always be found to have properties which can be observed in the rest of the body. This is true, for example, of reserpine, chlorpromazine, nicotine and also of cocaine. Unless the manufacturer discovers a new drug which produces clearly understood peripheral and central effects in animals. 1 would be inclined to discourage him from selling it if his only ground for doing so is that it has a general tranquillizing action and is not obviously harmful in man. We are at the moment I think retreading a path which we have trodden before. Some thirty years ago the battle was being waged about the sale of medicaments which had not been tested on animals. I remember very well the sale of Hormotone, a mixture of dried animal glands to be taken by mouth. It was devoid of activity except that due to the presence of a small amount of thyroid. There were large numbers of other preparations. But the battle was won. and it became the rule for pharmaceutical houses to subject their products to animal tests both for potency and toxicity. That was a considerable advance. Now so far as concerns the tranquillizing drugs we are back again where we were, and having established that a new substance will tranquillize a monkey or a dog, a manufacturer relies on clinical reports. I have tried to show by the examples given of the effects of placebos, how misleading these clinical reports can be. The accident that French workers discovered the usefulness of chlorpromazine in patients before it was shown to exert its tranquillizing properties in animals is no doubt responsible for this. It seems to me, however, that reliance on clinical reports is not enough. Clinical trials will of course be needed at the end, but better devised tests on animals, as the result of a much more extensive study of the animal brain, are needed. Feldberg has shown us how to produce tremor in cats, and. still more important, how to cure it. This surely ought to encourage those manufacturers who are concerned to make a valuable contribution to medicine and should help them to believe that it is possible to make advances in brain pharmacology by animal observations. I am convinced that it is only in this way that the advances will come.
Batterman, R. C, and Grossman, A. J. (1955). J. Amer. Med. Ass., 159, 1619.
Beecher, H. K. (1953), J. Pharmacol., 109, 393. (1956) Amer. J. Physiol., 187. 163.
Feidberg, W., and Malcolm, J. L. (1959), J. Physiol., 149, 58.
Feidman, P. E. 0956), Am. J. Psychiat., 113, 52.
Hillis, B..R. (1952), Lancet 1, 1232.
Modell, W., and Garrett, M. (1960), Nature, 185, 538.
Modell, W., and Houde, R. W. (1958), J. Amer. Med. Ass.. 167,2190.
Wolf, S., and Pinskv, R. H. (1954), J. Amer. Med. Ass. 155, 339.