chemists and c1inical pharmacists. There is indeed enough to do for all ofus.

The progress in this field is easy to recognize for those of us who during the sixties

found ourselves in the situation of prescribing drugs that later were found to be

poorly absorbed or using dosage-schedules resulting in homeopathic drug concentrations in body fluids. Studies that until very recently were considered to be original

contributions to the understanding of pharmacokinetics are now part of the routine

work up ofthe fate ofnew drugs in man. Editors of c1inical pharmacological journals

are today 'begging' for papers eIucidating the c1inical relevance of all this new kinetic

information - an area where c1inical pharmacology should play a key-role by

designing better methods for assessing drug response .

There are many other challenges to c1inical pharmacological research.

Recent advances in pharmacogenetics with the demonstration of polymorphisms

in drug hydroxylation may open up new possibilities to identify rare individuals at

risk during long-term drug therapy. Future utilization oftissue banks in metabolism

studies may become feasible. Biochemical assessment ofdrug action in man is still in

its infancy c1inically. Receptor pharmacology has already provided new tools for

c1inical investigations. New methods have to be developed to assess the overall

effects, beneficial as weil as adverse, of drugs in general use outside the controlIed

academic environment. Drug epidemiology will probably develop into an indispensable asset ofclinical pharmacology.

These examples and past history show that c1inical pharmacology has a vigorous

research profile of its own, focussing on general drug problems and on drugs as

molecules rather than as remedies for particular diseases. The future of c1inical

pharmacological research has never been brighter.

Due to the complexities of drug research in man, particularly from the ethical

point of view, there is an increasing demand for a constructive and international

dialogue between scientists in drug control, pharmaceutical industry and academia.

The recent worries both in the US and UK about bureaucracy becoming a rate

limiting factor in drug development underline the key role that a strong, independent

academic c1inical pharmacology could play in trying to resolve this dilemma. As an

example should we not along with the tremendous developments in human kinetics

and drug metabolism require less extensive studies ofthe same kind in animals that at

the best reconfirm the existence of species differences in metabolism, binding and

distribution.

While the contributions of c1inical pharmacology in research are evident, our

efforts in teaching therapeutic principles, in drug information and in drug consultation seem to have had Iittle impact on drug utilization among doctors and patients.

Data are now being collected from various countries in the world notably Australia,

Northern Europe, UK and the USA which show marked deviations in some drug

prescribing patterns from accepted pharrnacological principles (Turner, 1979). There

is a peculiar combination of under- and over-utilization of drugs. As an example, in

1976, drug problems (intoxications and drug abuse being excluded) were the main

reasons for over 10% of consecutive admissions to the medical wards of a university

hospital in Stockholm. Adverse drug reactions and inadequate treatment with drugs

including poor patient compliance with the regimen were equally important

(Bergman & Wiholm, 1980).

International comparisons of drug prescribing have revealed astounding

differences that are difficult to explain in terms of geographical differences in

morbidity or other rational factors. The number of drug products available to the

prescribing physician may vary IO-fold between European countries with similar

PREFACE xv

socioeconomic and medical conditions. In an interview with Elliott (1979), William

G. Anlyan who directed the compilation ofthe Institute ofMedicine report on drug

prescribing in USA states, "T he medical profession needs to clean up its own house.

The most important goal we must set is to improve the quality of education in

clinical pharmacology at the med icaI school and residence level . . .' With the

constant manpower problem in clinical pharmacology we are not likel y to be able to

do much of this education ourselves. Rather we will have to develop the means of

getting our messages across through other educators in clinical medicine and other

health personnel. It is high time that drug prescribing and therapeutic principles

become dependent variables; and equally important, drug therapy must be assessed

in a general medical perspective and balanced against other therapeutic approaches

such as dietary treatment and psychotherapy.

The clinical pharmacologist has, therefore, an equally important role in informing

the medical community when drugs are needed and when they are not. A majority of

physicians in the developed countries and almost all in the developing countries still

get their prime continued drug education solely from industry, particularly from

detail men . This is far from satisfactory. Clinical pharmacology should get into the

medical curriculum in all medical schools not as an elective extravagance but as a

tough required course. And I cannot see why the pedagogic influence of clinical

pharmacology on ph ysicians should stop when they start prescribing drugs . Clinical

pharmacologists cannot afford to close their doors to the real world of drug

prescribi ng.

For a new discipline, striving for international recognition, collaboration within

and between countries is essential. We have a forum for international collaboration

through the section of Clinical Pharmacology in IUPHAR. Founded in 1975 this

Section has as its key priorities to facilitate training and to develop training

programmes in clinical pharmacology and to serve as a consulting body for

individuals and organizations interested in the discipline.

On behalf of the Section of Clinical Pharmacology I wish to extend our cordial

thanks to the organizers of thi s conference, particularly Professor Colin Dollery and

Professor Alasdair Breckenridge and their close associates, for having given us the

opportunity to come to London for th is meeting. It will be an important milestone in

establishing clinical pharmacology world wide as an indispensable branch of

medicine.

References

Bergman, U. & Wiholm, B. E. (1980). Drug related problems causing admission to a medical

c1inic. Eur. J. clin. Pharmac.. in press.

Breckenridge, A. M. (1980). Assessment of new drugs: a c1inical pharmacologist's view. Brit.

med. J.. 1, \303-\305.

Carr, E. A. (1963). Proposed aims, organization, and activities of a division of c1inical

pharmacology. Clin. Pharmac. Ther.. 4, 587-595.

Crout, J. R. (1965). Academic clinical pharmacology and the university medical center.

Pharrnacologist, 10,82-85.

Dollery, C. T. (1966). C1inical pharmacology.Lancet. 1, 359-360.

Elliott,J. (1979). Physician prescribingpracticescriticized; solutions in question. Medica! Ne ws.

J. Am. med. Ass.. 241,2353-2360.

Gross, F. (1978). The thorny path ofclinical pharmacology. Clin. Pharmac. Ther.. 24, 383-394.

Lasagna, L. (1962). Profession, governmentand drug industry. Lancet,1,580-581.

Lasagna, L. (1966). C1inical pharmacology: present status and futuredevelopment. Science, 152,

388-391.

Turner, P. (1979). A clinical pharmacologist's view.Lancet, 2, 735.

WardeIl, W. M. (1970). C1inical pharmacology at university medical centers: functions and

organization. J. clin. Pharmac.. 14, 309-324.

xvi PREFACE

Wilson, G. M. (1963). Difficulties and opportunities in clinical pharm acology. Clin. Pharmac.

Ther., 4, 255-260.

World Health Organization (1969). Clinical pharmacology. Scope organization, training.

Report of a WHO study group. World Health Organization Technical Report Series No.

446.

Plenary Lectures

THEETHICAL AND

SCIENTIFIC BASIS

OFREGULATION

L.LASAGNA

Department ofPharmacology and Toxicology,

The University ofRochester.

601 Elmwood Avenue,

Roch ester,

New York 14642.

USA

A cyn ic has described regulation as the substitution of error for chance. Be that as it

may, modern regulation unquestionably is based, to some degree at least , on seientific and ethical considerations .

Both science and ethics are time- and culture-bound. One century's science is

another century's quackery, and what is ethical for one society is not necessarily so

for another society, or even for the sa me society at a later date . Beecher (1966), who

initiated a world-wide revul sion against unethical investigators with his article, had

for years performed experiments on human subjects (both healthy volunteers and

patients) without obtaining informed consent. The evolution in this century of the

randomized controlled trial has made it impossible to return to earlier times, when

medicines could be registered on the basis of uncontrolled observations. The knowledge that in vitro tests are imperfect predictors ofbioavailability means that reliance

on unvalidated in vitro tests is no longer scient ifrcally defen sible .

For many, unfortunately, this concept of cha nge is disturbing. It is more comfortable for them to live with eternal verities. As Cornford said in 1908: 'T hey must

never be troubled with having to think whether this ought to be done or not, it should

be settled by rules . . . . Plainly, the more rules you can invent, the less need there will

be to waste time over fruitless puzzling about right and wrong.' This witty man

added, 'Ever y public action wh ich is not customary, either is wrong, or, ifit is right , is

a dangerous precedent. It follows that nothing should ever be done for the first time. '

The ethical infrastructure of drug development Irrst comes into play when seientists seek justification in the laboratory for ta king a potential new medicine into

humans. The search for such ehernieals is based on unmet medical needs, that is,

illness not optimall y served by available medicines. There is little likelihood that

science will run out of motivation in this regard in the fore seeable future . No

significant disease or symptom now exists for which more effective or less toxic

remedies are not desired by the public or the medical profession.

But predicting what chemical will be a 'better' drug poses serious difficulties, since

our subhuman models and in vitro tests are never completely isomorphic with the

4 L.LASAGNA

human condition. As the years go by, scientists have gotten better and better at predicting performance in humans from preclinical testing, but some surprises always lie

in store. Nevertheless, it behoves us to do the best we can, predicting the predictable, whether it be desired or adverse effect. As Aristotle said, '... it is the mark ofan

educated man to look for precision in each class of things just so far as the nature of

the subject admits ... .'

There are, to be sure, limits imposed upon us by reality. No drug developer can

afford to run all the screening tests that could possibly be imagined, or study

thousands of animals of many different species, searching for extremely rare adverse

effects. The ethical posture, hence, is to perform those tests in numbers and with a

completeness that are consistent with a functional drug development programme.

We can neither cut corners recklessly nor shun the heavy responsibility for fmding

better new drugs .

Once it is decided that cautious exploration of a drug in man is justified, a new

series ofquestions arise :

How shall one ca1culate the first doses to be given to man?

How many subjects are to be studied in Phase I?

Are these first subjects to be healthy volunteers, or sick patients?

At what rate ofprogression shall the dose be increased?

What variables should be measured?

How shall the doses be chosen for the first clinical trials (Phase II)?

How much additional toxicity testing will be required as clinical trials increase in

scope, as drug dosage is increased in amount or in duration ofadministration, or as

new kinds ofsubjects are put at risk?

These questions, it will be noted, involve both ethical and scientific considerations.

For these reasons, protocols proposed for human studies deserve to be scrutinized

critically by review boards with a membership capable of dealing with both levels of

concern. AJatally flawed experimental design is unethical because it puts subjects at

risk with no possibility of'benefit to anyone. Individuals other than thc invcstigators

should decide whether a proposed experiment is too dangerous to be done .

Yolunteers can usually be found somewhere, somehow, to participate in any experiment. While it could be argued that people in a free society who are in command of

their senses should be allowed to make their own decisions about participating in

dangerous experiments, and that censorship of experimentation carries its own risks,

science in general has to beware the harm that can accrue to all investigators by lurid

publicity attendant on research that strikes the public as (or worse, that turns out to

be) needlessly risky . In asense, all scientists owe it to their guild to protect the public

image ofscience by avoiding research that will give it a black eye.:

So far as consent is concerned, I see no great theoretical problems in regard to

providing information to healthy volunteers. Such subjects deserve a full explanation

of what is planned by the investigator, including the purpose of the experiment, its

anticipated benefits, and the possible risks. Some have argued that informed consent

is a sham, either because the risks cannot be fully disclosed or bccause subjects cannot evaluate the information. But this misses the point, which is to play fair with subjects, that is, to do the best one can in explaining what is planned, including the

information that one cannot ever anticipate everything which might possibly happen.

This uncertainty should be appreciated by potential subjects in addition to their

comprehending the available facts. There is a problem with regard to the mode of

presentation. Our own works (Epstein & Lasagna, 1969) has shown that there is

danger of information overload in obtaining consent from subjects. Excessive detail

may actually impede communication, and it cannot be assumed that exposure to

information is tantamount to comprehension.

With the sick subject, the problem of informed consent becomes more complicated. Unlike the situation in the healthy volunteer, total candor is not always in the

ETHICAL AND SCIENTIFIC BASIS OF REGULATION 5

best interest of the sick, although the investigator must be able to justify any deviation from a full explanation. There are special problems when the investigator is also

the physician ofrecord for a patient subject. People who oppose the use ofstudents or

prisoners in research often seem to ignore the much greater likelihood of a 'captive'

status for the patient approached by his doctor to take an investigational drug. It

must be made clear to any patient who is asked to participate in research that medical

care or the patient-doctor relationship will not be jeopardized by failure to volunteer

or by withdrawal at any time after the experiment begins.

Compensation for research subjects is best approached, in my view, by negotiations between subjects and investigators. It is difficult to lay down rigid rules for what

is 'enough' reward for healthy volunteers or patients. Rewards should not, it has been

said, be so extraordinary as to be coercive per se in seducing volunteers who would

otherwise not participate. I propose, however, that intolerably dangerous experiments are best prevented at the review board stage, independent of reward proposed.

An experiment that is 'too dangerous' is not made ethical by a large reward for subjects. In fact, of course, most drug research involves commitment of time and some

inconvenience for subjects, not danger (Cardon, Donnell & Trumble, 1976; Zarafonetis, Riley, Willis, Power, Werbelow, Farhat, Beckwith & Marks, 1978). Once it is

decided that an experiment can be conducted ethically, adequacy of compensation

should be a matter for agreement between the parties involved.

If harmed as a result of research participation, subjects deserve compensation for

medical care, for lost time from work or horne, or for irreparable damage. Society

needs mechanisms for compensation that are equitable but not punitive, barring

negligence on the part ofsponsor or investigator.

It is unethical to perform aseries ofclinical trials that are simply repetitions ofwell

done experiments that have demonstrated beyond question that a chemical produces

a desirable specific therapeutic efTect. Some repetition is a fundamental tenet of

science, but going beyond enough repetition to assure reasonable scientists that the

efTect is real represents regulatory overkill and immorality. Some local experience is

no doubt desirable, but this need not be in the nature of double-blind, randomized

trials.

Recently, Burkhardt 800-Kienle (1978) have proposed that all controlled clinical

trials are unethical, illegal , or both, because the null hypothesis is really not seriously

entertained by the investigator even at the outset. In fact , this statement about the

null hypothesis is generally true: scientists pick drugs for human trial because they

strongly (and often correctly) suspect that the drug will be efTective. It is not clear,

however, that it is automatically unethical to do such trials, provided that patients

are fully informed and not subject to serious risk, since qualitative expectations about

a drug need to be fleshed out quantitatively, as to both benefit and harm. One can,

however, agree that as we deal with increasingly serious disease states or as we

develop increasingly better standard drugs, it is imperative that proposed research be

scrutinized with greater and greater care, not only justifying the choice of control

(placebo or standard drug) but also the decision to perform any experiment with a

drug candidate in humans. In a way, we are trapped by our own ingenuity, As our

standard drugs approach perfection, we will have an ever more difficult task in taking

new ones into clinical trial. It is clear, of course, that we are not even close to this

millenium.

Another ethical dilemma is generated by the choice between placebo-drug comparisons and drug-standard comparisons. There are, for the research patient, fewer

ethical and medical difficulties involved in the latter, but for future generations of

patients it might be argued that superiority over placebo represents a more convincing experiment than a trial which fails to demonstrate a difTerence between a standard

drug and a new medicine, since the latter result might reflect simply a nondiscriminating population. Needless to säy, superiority over a standard drug is

6 L. LASAGNA

impressive and cogent evidence, but useful drugs are not necessarily 'better' on

average than standards. (They may be able to treat certain patients uniquely weil,

either by providing more therapeutic efTect or less adverse effect.) We shall simply

have to devote more energy to devising techniques that will get us off the horns of

these ethical dilemmas. (Some are already available, such as dose-response curves as

areplacement for placebo controls.) We may have to re-examine the utility ofclinical

data, that are not ofthe controlIed, randomized variety.

.Sequential designs allow one to perform c1inical trials with smaller sam pie size, on

average, than fixed-sample size trials. These designs have not become very popular,

for a variety of reasons, but they have much to ofTer in the study of patients who

sufTer from very rare diseases or from diseases where ineffective treatment will be

seriously deleterious. It has been proposed that c1inical trials should not be

terminated until 'administratively significant' difTerences between treatments are

observed (Buncher, 1973). To quote Buncher (1973): '.. . [al fmding should be

declared 'adrninistratively significant' only after .. . the likelihood that cxposure of

additional patients to the trial will persuade many physicians to adopt the better

therapy is so small that it is outweighed by the risks to the patients continuing in the

trials.' The point of this argument is that if a study is terminated prematurely, its

impact will be small or nil and the trial will have been performed in vain. Some

serious problems exist with regard to this approach. One is that it is difficult (if not

impossible) to know what will convince others. 'Statistical significance' not only

poscs the need to pick some level of confidence, but ignores the fact that statistically

significant and biologically significant difTerences are not necessarily the same. A

second problem is that an investigator should terminate a trial when he is personally

convinced that it would be unethical to continue. One's own level of'ethical significance' might or might not be congruent with the judgments ofothers.

Arecent ethical concern has arisen with regard to prescribing information intended for patients. There is evidcnce (Joubert & Lasagna, 1975a,b) that many patients

desire more information than they have received in the past about their diseases and

their treatments. At least some patients want to be involved in health care decisionshence the emphasis on 'patient package inserts' and other informational materials. It

is not easy to come up with such materials that are adequate for all patients and all

occasions. Patients difTer in age, intelligence, education, language skills, attitudes,

and many other factors. There is disagreement about the purpose ofsuch materials:

should the y be primarily warnings, or balanced presentations of benefrts and risks?

Who should administer the materials? Should there be flexibility in both materials

and mode of presentation? How does one keep these materials up to date? The area

cries for research before widespread implementation of the use of new educational

materials.

Scientifically, regulation needs a data base that will allow the responsible agency to

make a judgement that, with proper directions, physicians will be able to use the drug

so that treated patients will reap more benefrt than harm. This global or group

judgment is really the only one that can be made. It is fruitless to search for more

precision than this, although we must insure that enough c1inical and preclinical data

exist to make this importantjudgment, and that we havc a reasonable idea as to how

to write directions for the physician that will encourage flexibility of treatment

depending on c1inical variables ofhuman importance.

In my view , it does not take a tremendous amount of c1inical work to identify a

useful drug or to delineate its more common side efTects. There are , however,

formidable problems if we try to an swer in advance of marketing all the questions

that might possibly be raised:

What special adverse efTects occur after years ofcontinued treatment for medicines

intended for chronic illness?

ETHI CAL AN D SCIE NTIFI C BASIS OF REG ULATIO N 7

What very rare but serious toxic effects will be seen?

What interactions (beneficial or harmful) will occur between the drug and other

prescription or proprietary remed ies?

What interactions will be seen between dru g and disease, such as renal or hepat ic

or coronary insufficiency?

What special hazards will exist for the foetuses oftreated pregnant women? For the

aged? For children?

Will the new medicine be abu sed by segments of the popul ation , either for addi ction or suicide?

In thinking about these que stion s, society must balance costs against benetits.

Demanding deta iled answers to all conceivable questions of interest before registration will mean interminable delays in getting useful drugs to their target populations.

Not onl y will pati ents suffer thereby, but economic incentives will disappear. New

chemical entities already cost many mill ions of dollars and years ofwork to reach the

market. Effective pat ient life shrinks yearly, on average , as a consequence of the

delay s.

Th e most rational policy, I submit, is to plan carefully for many ofthese quest ions

to be an swered with appropriate speed after registration. Society must judge which

question s must be answered befor e marketing, as weil as tho se that must be attacked

afi erwards.

Thi s will mean post-marketing studies that are more than simply attempts to study

poorly in large numbers ofpatients what has been weil studied in more modest trials.

It is, however, extremely pertinent to search for new information, both good and bad,

about a new medicine , including very rare or long dela yed toxicity, although the

latt er concerns are just tho se that science is least weil equipped to identify. We also

need to know whether a drug is being overprescribed, underprescribed, or misprescribed (such as the wrong dose or regimen), and which adverse effects are

preventable , as opposed to tho se that ar e not.

Post-registration respon sibilit y also include s the application oftechn iques that will

detect qual ity control problems at the manufacturing level, and the use of registers

(for example, of congen ital defects or cancers) that may l1ag a probl em becau se of

temporal changes in prevalenc e ofabnormalities that might possibly be drug-related.

Let me, fmally, return to my first theme - the need to keep reviewing our scientific

and eth ical bases for regulation as the world change s. The skyrocketing costs of drug

development have, for example, decreased the likelihood that unprofitable drugs (for

patients suffering from rare diseases) will reach the market. Society must come to

grips with this issue, and decide either that it reall y does not care about 'orphan drug'

problems, or that it will take regulatory and economic steps to allev iate it. One

approach may be a radically different level of demands for data on a drug intended

for a very few individuals who might weil be willing to take a unique remed y for a

serious disea se even without the safeguards usually demanded ofa new drug .

A second example is the proliferation of laboratory tests for carcinogenicity,

mutagenicity, and teratogenicity. These tests are in some cases easier to perform than

to interpret. How do we deal, scientifically and ethically, with an Ames test result ?

This test is far quicker and easier to perform reliably than an epidemiological study

to see whether saccharin use is associated with bladder cancer, but the latter may be

more to the point. What sort of in vitro or animal toxicity should justify 'killing' a

therapeutic agent? How much such data does one requ ire for action earl y in the

histor y of a drug's development? How much data after apparently satisfactory use of

a drug for years? How do we balance evidence ofmutagenicity in bacteria again st the

potential of a chem ical for treating the sick, realizing that the earlier a drug is ' killed',

the less we will know about its potential for good as weil as for harm?

00 our statistical ground rules need changing as our technology and science

8 L. LASAGNA

increase in sophistication? There never has been an ything magical about 'P = 0.05 '

values, yet regulators tend to look upon this barrier as the statistical equivalent ofthe

Magie Fire that Wotan threw up around Brunhilde. Should the level ofconfidence be

set flexibl y, to deal with the very different sorts of conditions for which med icines are

intended? Is a two-tailed test justifiable in view of the overwhelming likelihood, for

mo st rernedies, that a new drug will 'outperforrn' the placebo, both therapeutically

and to xicologically? I am weil aware ofthe arguments for and against one-tailed tests.

There are relatively few case s where I personally feel so cocky that I am unwilling to

accept as real a very large difference in the unexpected direction. As Burke (1953)so

wittily pointed out: 'We counsel an yone who contemplates a on e-tailed test to ask of

himself (before the data are gathered): 'If my results are in the wrong direction and

significant at the one-billionth of 1% level, can I publicly defend the proposition that

this is evidence of no difference?' Ifthe answer is affirmative, we shall not impugn his

accuracy in choosing a one-tailed test. We may, however, question his scientific

wisdorn.' .

I do not expect agreement on the answers to those questions, but I respectfully

submit that if we are to defend our scientific and ethical postures toward drug

regulation, we must keep asking ourselves about them. Science and morality have not

remained static in centuries past; it is difficult to see how they can avoid change in the

years ahead.

References

Beecher,H. K. (1966). Ethics and c1inical research. New Eng. J. Med.. 274, 1354-1360.

Buncher, C. R. (1973). Administratively significant. New Eng. J. Med., 289,155.

Burke, C.J. (1953). Abriefnote on one-tailed tests. Psych. Bull., 50, 384-387.

Burkhardt, R. & Kienle, G. (1978). Controlled c1inical trials and medical ethics. Lancet, 2,