Document By3vGm7xy95OMQG44rLv8xZJj

JUN 2 7 1967 TELEPHONE HUDSON 3 6126 Manufacturing Chemists' Association, Inc. (FOUNDED 18721 1825 Connecticut Avenue, N. W. Washington, D. C. 20009 June 14, 1967 TO: Members of the Food, Drug, and Cosmetic Chemicals Committee SUBJECT: PMA Paper Gentlemen: I am enclosing a copy of Dr. Lehman's paper given at last fall's PMA meeting. It has been suggested that you will be interested in it. Sincerely yours, i MMH:Sjg Enclosure ASI 00002132 THE FUILOSOPHY EEI-TKD rr.D-CIJUTCAL TOXICITY STUDIES A. J. Lchmsn, M.D. Division of Pharmacology, Bureau of Science, Food and Drug Administration S. Department of Health, Education, end Welfare, Washington, D. C. 20204 The underlying principles involved in animal experimentation for predicting safety of drugs, Intentional and unintentional food additives, and other chemicals to which man may be exposed, can be of great value. However laudable studies in animals may be, the approach to the problem is often questionable. Twenty or thirty years ago, animal data servea fairly well for assessing the kind of toxic responses that might occur in man. j The chemicals tested were substances to which the human race had been exposed for many years. Thus, the toxic responses could be extrapoi plated to man with a fair degree of certainty. The rapid development in food and drug technology has changed all this. Introduction of hundreds of new synthetic chemicals has brought with it a host of new problems. The impact of these new substances with which man had no previous experi ence began to be recognised by the lack of correlation of animat data with data obtained in humans. As the apprehen3iveness of safety testing grew, the toxicologist increased the numbers of animals used in his investiga tions and prolonged the duration of his observations. Toxicological studies tended to assume a standard minimum one-shot deal* This eventually led to what some regarded as playing a numbers game. Since it is not feasible to monitor large groups cf animals with any degree in depth, reliance was placed on results of prolonged toxicity trials. Readily detectable changes were recorded, such a3 effects on weight, growth, and mortality rates, occasional urinalysis and hemograms, followed by autopsy *--Treaented at the meeting of the Research & Development Section, Pharmaceutical Manufacturers Association, Hot Springe, Va., Oct. 25, 1906. ASI 00002133 -2nnd histopothology at termination of the experiment. Actually, these parameters cannot be assured to have much relevance nor can they be directly transposed to man. The ease with which this once-and-for-all type of toxicity trial con be carried out is probably responsible for the predominant interest in this procedure. However, modern scientific technology is rapidly changing V our concept of what constitutes toxicology. Since continuous and rapid change i3 inevitable and normal, the task is to remain open-minded to accommodate these rapid changes. With present-day technology, it is now possible to correlate functional change with morphological state as an V early indication of prospective, not retrospective events. A careful V study of alterations in physiological functions in the living animal greatly strengthens the validity of toxicity trials. The need to innovate and develop-, new approaches to narrow the gap between the predictive value of animal data and experience in man was pointed out recently by the President's Science Advisory Conmittee^. The National Academy of Sciences' Subcommittee on"Some Considerations in the Use of Human Subjects (2) in Safety Evaluation of Pesticides and Food Chemicals" also emphasized the need for more certain methods for translating animal data to man. The World Health Organizations* Scientific Group on Principles for Preclinical Testing of Drug Safety (317 will issue a report laying great stress on biochemical studies. Very few systematic studies have been designed to meet these objec tives, Present methodology has been criticized as empirical, mechanical. ASX 02i34 -3rnd even pointless, and not based on any fundamental scientific principles. It is imperative then that animal testing procedures be modified end thor oughly developed so that their technical performance will yield results reflecting more closely human characteristics and reactions. This sophis tication in methodology, having its beginning about 10 years ago, is now well on the way. Modem safety evaluation of drugs and other chemicals in man's environ ment calls for studies of specific organs and tissues. This involves a multidisciplinary approach utilizing the principles of biochemistry, pharmacology, and pathology. New instrumentation and technics are employed to give an insight into early alterations In structure and function at the subcellulor level. The major aim is to develop clearly defined bio chemical or pharmacological responses in the live, intact, unanesthetized, animal by analytical procedures applicable to man. Much has been written on the subject of species variation. However, when the interpretative stage cf a safety test is reached, the animal data are extrapolated to man, while quite frequently the investigator is fully aware such application is not completely reliable in predicting responses in humans because of interspecics differences. At the same time, the welldocumented pharmacological phenomenon of species specificity is rigidly rpplied in any evaluation of efficacy. Only evidence in man is acceptable. Modem toxicologists are fully aware of species difference and have learned a creat deal about how chemicals, whether drugs, pesticides, or food additives, are handled in the body. Thus, the conventional practice of slavish devotion to old procedures la being rapidly supplanted by not ASI 00002135 -4concentrating on rata and dogs but intensively studying the test substance in four or five species and developing evidence of the events occurring at the subcellular level. The objective is to seek measurable parameters which may be useful for a logical prediction of toxicity in man. Since the most useful species for further testing will be that species which has predicted certain events in man, It is desirable to develop as early as possible how the substance Is handled in man. The early and cautious clinical' trial data will make the choice of the proper animal species for further study less difficult. The toxicologist now has a greater assur ance that the responses in the species he has chosen bear a reasonable resemblance to responses in man. In the initial clinical trial of a new drug, the patient or volunteer is treated as an individual and great attention is given to details. The average hospital employs 242 persons to care for every 100 patients. In most hospitals only 20 per cent of space is allotted to beds. The remaining 80 per cent is occupied by equipment for tests and procedures, administration, engineering, and maintenance departments necessary to keep the hospital functioning 24 hours a day (4) . An animal on experiment is performing a very important function. If anything more than a few routine responses are to be recorded, the animal, too, should receive the full attention of the investigator. Dr. Karl Beyer once stated that he would not entrust a study to a toxicologist whose only other qualifications were a compound and a roomful of animals. Although conventional safety testing methods have been useful, there has been too much trading on convention to the point of losing sight of the serious purpose of animal studies. An Inherent problem in older 002l36 toxicological investigations hoc been on over-dependence on animal tests. A few examples will point this out. The IP-50 Thi3 procedure has developed into an empirical screening test for toxicity. Very little attention is clven as to whether or not solvent, vehicle, or suspending agent bears any relation to the ultimate formula tion in which the test substance may be administered to man. At this early stage of testing, probably little is known about the physical state of the drug, less about its biochemistry, and many other circumstances which are different between animal experimentation and clinical trial. Data are usually obtained in a standardized animal as against an vnstandordized human. Animals arc in good health, whereas the drug may be tried first in ill patients. This does not mean that tests should bn done in mongrels or sick animals* It dce3 point out, however, the limitations that can be placed on a statistically calculated LD-50 as a measure of safety in man. Since animals on an LD-50 test are often kept under obser vation for a few days to a few weeks, much can be learned about absorp tion, transport, metabolism, distribution, and excretion during the obser vation period. Instead, for all practical purposes, only death in taken as the endpoint. Absolute toxicity of any compound is of relatively little interest. An LD-50 offers little more than a guide for dosage to be em ployed in a trultiple-dose test, Long-Term Studies Rats are almost routinely chosen for-this purpose. Dogs ore fre quently used. In either case there is usually little or no knowledge ASI 00002137 f -6ns to whether there is any close resemblance to man in the species response to the test substance. The large groups of animals on experiment preclude eny detailed studies in depth, and thu3 criteria of effects are limited to gross observations. A prodigious mass of data results, and great difficulty is often encountered in assessing its value. Interspecies differences ore apt to complicate assessment. Even if some detailed studies were incorporated into the design, this effort would be a waste of time and biological material without knowledge of similarities and diff erences between the rat, dog, and man. If the investigator decides it is necessary to conduct a long-term study, then the importance of choosing the species closely resembling man becomes obvious. There may be some justification for long-term studies in testing for carcinogenicity. However, the predictive value of animal data has been and still is the subject of much discussion^. Histopathology Although the design of a long-term test and its validity in pre dicting toxicity in man has been frequently discussed, surprisingly little attention has been given to the value of histopathological examination of tissues. As a rule, histopathological reports are presented in great detail, listing ratios of organ weight to body weight, naturally occurring diseases and senility changes, statistical calculations of significance or insignificance of spontaneous or other neoplasms, and some variations from normal which.the pathologist cannot ascribe to treatment. The rele vance of these observations, many of which may have been produced by large and possibly overwhelming doses, is not mentioned. The highest dose level ASI 00002138 -7producing no histopathology io generally taken aa the "no-effcct" nr safety level. Hlstopathology lo important and certainly has ita place in final evaluation of safety; nevertheless, the purpose of a toxicological investigation is to offer some predictive evidence of possible harm. . This retrospective approach seldom meets the objective. Even in clinical practice the value of the autopsy is a controversial issuev ' . Margin of Safety An arbitrary mathematical calculation has been used to assess safety of a chemical. The "100-fold" margin of safety has been widely employed and extensively applied to food additives and pesticides. Two defini tions are applied to "100-fold margin of safety." When applied to a food additive, it should be possible to feed to animals, without effect, 100 times the amount proposed for the human diet. For pesticides, the , safe intake for humans is one-hundredth of the dose producing no effect in animals. Neither the animal species nor the criteria for "no-offect" are specifically named or defined. This irrational application of a mathematically calculated safe dose is exemplified by variations which have been introduced. These variations range all the way from about 10 to 500. Obviously, .mathematical rules for margins of safety cannot be applied to drugs. It is more appropriate to assign a safety index, the ratio between minimal effective dose and the maximal tolerated dose based on clinical experience. Simple Methods Expressions such as "the single laboratory test," "the easy and simple test," or "a simple screening method" are rather common in labora tory parlance. No doubt an inveatigator. In attempting to monitor a lnrge ASI 00002139 -8 group of cnimals, ia faced with the fact that the conventional indices usually applied offer little proof of safety of the chemical being investi gated. He may attempt to fill the gap by using "quick and easy" tests for added confidence in his final conclusions. A chromatographic paper has been developed and described as "an accurate* simple, reproducible micro method for the determination of urea nitrogen in either serum or plasma." The accuracy and precision of these paper strips were carefully evaluated by comparison with two acceptable quantitative procedures. Results showed that some strips were not accurate, lacked precision, and gave no or only partial migration. The migration process was also quite temperature-sensitive, and the strips were not stable even when stored in accordance with the manufacturer's instructions'1 . Hemoglobinometry, introduced almost 90 years ago, has been and still is an important clini cal laboratory test. The test appears simple enough to perform. Many methods have been devised to improve accuracy and uniformity of results, and yet surveys have shown discrepancies and inaccuracies in this"simple" (9) analytical procedure . An editorial comnenting on the bromsulphalein test lists many pitfalls in carrying out this "extremely simple" proce dure^1^. Use of certain detergents in cleaning glassware has been sus pect, and even method of transportation of specimens' from bedside to laboratory has created problems. If a simple laboratory test actually exists, it must be an extremely elusive one. Re-use of Animal's An animal in which various normal baselines have been established represents a considerable expenditure of time and effort. This investment ASI 00002140 -9increases as the animal is placed on test. Modern toxicology makes every attempt to avoid death, since efforts are directed towards developing evidence of early signs and symptoms of toxicity without sacrifice or death of animals. Previous reference to the need for individual monitoring indicates that only a few, generally no more than six of larger species, may be placed on experiment. An animal or two may be lost during an experiment,bum the majority may survive. A well-designed experiment should have developed a fairly comprehensive index of effects of the test sub stance. There Is actually no need to sacrifice the survivors for an ex haustive histopathological examination. When baselines have returned to normal, or have been re-established, there la no reason why these valuable animals cannot be re-used for other toxicity triala. This has its counter part in clinical practice. A patient may have several Illnesses in his , lifetime and be treated with a variety of drugs. In every instance appro priate baselines are established before any new treatment is begun. The Placating Petition Certain features of toxicity trials for evaluating safety of food additives are somewhat different than for drugs. A drug is a medicament developed for a specific use. Eventually the drug will be subjected to a contrdled clinical trial. Thus, clinical experience supports evidence of safety. Although the ultimate consumer rarely objects, he does have the right to choose If he does not wish to take the drug. The consuming public does not have this choice with food additives, and only on rare occasions has there been any attempt to obtain clinical data on safety. The toxicologist relies almost entirely on animal work for this evidence. He fully realizes the pitfalls in transposing animal data to man, but even ASI 00002141 - 10 knowing thio, ha makes every effort: to support HIb conclusions by testing the additive for months or years in two and even three species, different ctrains of the came species, end short-term studies In other species. The lcrge numbers of animals--as many es 2400 have been used--are reported In great detail, species differences, strain differences, and details of certain tests and other observations, which hold no allegiance to any species except the onea used, . The ponderous report seems to ingtill a placating effect on the toxicologist that safety has been established at soma level of use for man. More assurance may be attempted by invoking a mathematical margin of safety. Even further security may be sought by comparing the results of the test substance with an imitative compound whose safety has been reasonably established--for instance, applying safety data on formic ecid and ethyl alcohol to ethyl formate on the theoretical basin that ethyl formate hydrolyzes to ethyl alcohol and formic acid in the body. Obviously (his is not justified. The toxicologist's apprehensivencss also becomes apparent in his final assessment, which states that the level of proposed use in foods is safe. Any reference to human safety is omitted because human trial is rarely attempted or even considered. The creation of on exhaustive compilation of charts, graphs, tables, and written material, without knowledge of the extent to which these may support the conclusions, has been blamed on the legialator^\ some administra tive fcody^\ bureaucrats^**^, authors of marrow-curdling books^1^, and so on. There is but one answer to the placating petition--a full knowledge of the compound's toxicity. New methods of modern toxicology must be applied to learn how to deal with a passible hazard or how to avoid it. ASJ 000021 42 Tf the toxicologist doen not keep in tune with developments in modern toxicology, the opportunity io apt to be grasped by someone else who may incorporate features which ere most undesirable. The toxicologist cannot afford to be remis3 in his responsibilities* Summary Toxicological evaluation of a substance intended for human use must be based on methods which ore adaptable to the first human clinical trials* They must be prospective and not retrospective in purpose, or else the concept of preclinical toxicity ctudiea becomes a distorted effort, it is conceded that evidence on absorption, blood levels, distribution, metab* olism, and excretion may be difficult to attain, and for this reason is generally negleeted. Nevertheless, greater use should be made of the molecular biological approach employing histochemlcal, cytochemicnl, ( ultractructural end lnnunohietochcmical methods* The specific purpose is not to determine cumulative anatomical changes, but cumulative physiological effects correlated with ultra3tructural changes where the effects are taking plr.ce--at the cellular level. By such methods, early warring signs cf toxicity should be demonstrable* The test substance should be capable of inducing similar changes in men if the animal model is to ser-s a useful purpose. Since .there are more than 200 analytical procedures of prospective value, it la unreasonable to expect to apply all of these tests in every toxicity trial* The appropriate choice of parameters, not the number, must be left to the experienced investigator* Obviously, the number will vary with the stags of the investigation. In the beginning phase, hori zontal coverage of a few animals representing 4 or 5 species may be required ASI 00002143 - 12 to uncover leads. The more observations the Investigator makes on one subject, the more confidence he has in his conclusions. Choice of species and parameters for studies in greater depth rests largely on results from early clinical trial. Finally, the investigations should be placed in the hands of an experienced toxicologist who should be free to design, conduct, and interpret the toxicological problem. Although there is a lack of methods for detecting responses, and for differentiating a beneficial effect from an injurious one, nevertheless, if we are inept enough not to incorporate more precise indicators into the research program,* we deserve the consequences. ASI 00002144 REFERENCES 1. Use of Pesticides. A report c The Trcc'dcnt's Science Advisory Committee, May 15, 1963. 2. Some Considerations in the Uca c Human Subjects in Safety Evalua tion of Pesticides end Food Chemicals. Publication 1270, National Academy of Sciences, NationrA Research Council, 1965. 3. Scientific Group on Principles for Prc-Clinical Testing of Drug Safety. Report to the Dircctor-Gcncral cf WHO. To be published. A. Part 2 of the Guide Issue of Hospitals. J. A. llosp., August 1, 1965. 5. Symposium on The Risk of Carcinogenic Effects Due to the Administra tion of Some Medical Drugs, sponsored by the Union Internationale Contre le Cancer (U.I.C.C.), Paris, 1965. To be publls'hed.6. Value of the Autopsy. J.A.M.A. 1_93, 805, 1965. 7. Wilson, R. R. In Defense of the A.utopgy. J.A.M.A. 196, 1011, 1966. 8. Logan, J. F. Canad, Med, Ass. J, 89, 3A1, 1963, 9. Sunderman, F. V, Am. J. Clin, rath. A3, 9, 1965. 10. McGath, T. B.Am. J. Clin, rath,- 39, 630, 1953 11. Gaultier, M., and Fournier, C. Proceedings of the European Society for the Study of Drug Toxicity VI, 223, 1955. 12. Barnes, J. M., and Denz, F. A. Fharmnc. Pcv. 6, 191, 1954 13. Long, D. A. Second International Pharmacological Meeting 8, 68, 1965. 1A, Information Bulletin. British Industrial Biological Research Associa tion 5, 150, 1966. WTO Publications (Ref. 3) may be obtained from Columbia University press. International Document Service, 2960 Ercadway, M.Y.C. 10027. UICC Publications may be obtained from Prof. R. Truhaut, University of Paris, Facultie de Pharmacie, A Avenue da I'Observatoire, Taris, France (Ref. 5) ASI 00002145