ToxicologyLetters, 64&i (1992) 149-155 0 1992 Elsevier Science Publishers B.V., All rights reserved 03784274/92/$5.00
149
Are newer scientific concepts in regulatory toxicology used timely and appropriately? C. Hodel” and R. Bassb Y$terling Winthrop Pharmaceuticals Research Department, Longvic @‘rance) and bInstitute for Drug Toxicology, Bundesgesundheitsamt IBGA), Berlin (Germany) Key words: Regulatory toxicology; Toxicity testing; Testing safety; Test procedures; Risk assessment SUMMARY Laws regulating toxicology (e.g. toxic thresholds allowed, poison classes or definition of necessary preclinical testing) might improve health and save lives. Scientific facts will always serve as a mandatory base for political decision-making, but there will also be additional influences (perception and acceptance of risks, possible benefits, economic considerations etc.). These latter factors may vary considerably from one society to another. The Delaney clause prohibited the marketing of any product which was found to be carcinogenic in animals. Due to their benefits, exceptions were made for drugs. In other countries, too, other chemicals could be an exception due to a different perception of the risk or different scientific evaluation. Clear cases of major events always trigger changes in legislation. When in 1937 a newly-marketed sulfanilamide elixir led to severe kidney damage and 70 deaths, the FDA quickly endorsed the propositions of the investigation team set up by the American Medical Association: animal testing in two species with histopathologic examination before a marketing authorization could be granted became mandatory. A similarly rapid reaction followed in Europe when it was detected that Thalidomide was responsible for malformations in the offspring of mothers who had taken the drug in early pregnancy. When the effects are more difficult to link to a chemical, there may be time delays in regulatory actions. However, a sophisticated evaluation system was introduced for better monitoring of drug and chemical hazards. Some examples will be given in order to discuss the difficulties of timely and appropriate use of scientific findings.
INTRODUCTION
There are peculiar aspects in toxicology. It is therefore not surprising that, at some time, the phrase “regulatory toxicology” came into existence a phrase which does not exist in biology, pharmacology, or even in medicine Correspondence to: C. Hodel, Sterling Winthrop Pharmaceuticals Research Department, Bd. Eiffel, 21600 Longvic, France.
150
and is not in common use. Regulatory toxicology probably underlines the subconscious wish that something dangerous (toxic) should be regulated to be not dangerous. It would be very reassuring if, by simple regulation (guidelines or laws), any toxic action or reaction could be eliminated. Enormous amounts of money are spent on testing safety and defining risks of new chemicals. The possibility of measuring minute concentrations of exposure and the impossibility of proving absolute safety when exposure is likely, tend to increase the legal requirements of test procedures. The scientist may find himself in a position where he has to perform experiments which he feels are unnecessary or absurd. A couple of years ago we were asked to run a pharmacokinetic study in dogs with a slightly new formulation of a suppository of an anti-inflammatory agent despite the fact that, in another country, the kinetics were already well known in humans. For chemicals with exclusively topical exposure, an oral carcinogenicity study might be demanded although with no resorption through the skin being measured. The regulatory toxicologist is often exposed to findings (positive or negative) which are difficult to evaluate. There might be sloppy experimentation, technical flaws or sometimes wrong tests (not validated, no absorption, etc.). The temptation just to follow the legal guidelines and checklist is evident. The university toxicologist is often so highly specialized that he seeks unnecessary one-sided perfectionism. In the 1970s a statistician asked for an acute toxicity test with 900 animals in order to be able to calculate mathematically an exact value with minimal standard deviation. In this complex field of influences (and sometimes also pressures), a way must be found to define the real issue - in our case, the toxic action - and define the risk quantitatively and qualitatively. Scientific methods are needed for this purpose; the experimental models should be sensitive and validated. Because of the complexity of the issue, there is sometimes the feeling that new findings are not incorporated into decision-making fast enough or appropriately, with well-accepted test models answering the relevant question. CASE STUDIES
The sulfanilamide story [lJ During the months of September and October 1937, at least 76 humans in various localities in the USA died as a result of poisoning by elixir of sulfanilamide. After analyzing this preparation, the AMA Chemical Laboratory found it to be essentially a 10% solution of sulfanilamide in about 72% diethylene glycol, together with some coloring and flavoring agents. There
151
were no contaminants, such as mercury, which might have had similar clinical symptoms produced by the elixir. Patients usually complained first of a sensation of heartburn, followed by nausea, abdominal cramps, dizziness, malaise and vomiting. The symptoms occurred about 24 h after medication was started. They were followed by oliguria and anuria. Oedema developed. Creatinine rose steeply. Later on, patients became comatose and died within 2-7 days after first medication. The main pathologic changes were tubular nephrosis and necrosis and centrilobular liver cell swelling. Apparently the makers of this product were unaware of its possible toxicity and distributed it without having tested it adequately on animals. It had not been submitted to the Council on Pharmacy and Chemistry or the American Medical Association nor did the Food and Drug Administration of the US Department of Agriculture know of its composition. Newspaper reports of deaths from this new medicine created consternation among many physicians who were using sulfanilamides. The elixir, its acting components (pure diethylene glycol and sulfanilamide) and synthetic composite were tested separately on groups of animals. Rats, rabbits and dogs were given the various drugs three times daily in graded doses. The experiments were devised with the object of developing conditions which might simulate those in human cases. The drug was, of course, immediately withdrawn from the market. In this case good toxicology work would have helped to detect that the solvent was the culprit and the sulfanilamide could have been saved as a medicine. Even a good literature search could have given a clue, although little had been published at that time on diethylene glycol. Not surprisingly, this disaster had consequences. Geiling and Cannon [ll presented nine postulates which shaped toxicology work for the following decades (Table I). TABLE I NECESSARY INFORMATION FOR DRUGS TO BE FURNISHED BY PHARMACEUTICAL COMPANIES IN THEIR NDAs, AS PROPOSED BY GEILING AND CANNON IN 1938 [l]. 1.
Chemical composition
2.
Acute toxicity studies in two species
3.
Chronic toxicity studies with different doses in different species
4.
Observation of animals
5.
Pathologic examination of animals
6.
Studies in animals with experimental lesions
7.
Studies on the absorption and elimination of the chemical
8.
Study of interactions
9.
Knowledge of idiosyncrasies and untoward reactions
152
The aforementioned example shows that animal experimentation could have helped in a scientifically appropriate way to define and avoid a major risk. Of course, such knowledge needs to be disseminated and must ultimately find its way into regulatory decision making. This can take days, weeks, months or years, depending on the manner of dissemination, the language barrier and different risk perception. In 1986, almost fifty years after the original diethylene disaster, Carne et al. 121published a report from a Burns Unit of a general hospital in Spain where, within some months, a couple of patients died with acute renal failure and metabolic acidosis. Apparently, some time before the first death, the brand of a widely used Ag-sulfadiazine cream had been changed. The cream contained 15% diethylene glycol which, in view of the typical syndrome, was to blame for the deaths. After Spanish legislation the ingredients had not to be disclosed on the package. This is an obvious case where scientific concepts have not been or were not used timely or appropriately, for whatever reason. A widely used pharmacological agent (caffeine) Table II gives an overview of toxicity data of a widely-used product [3)l Although it is impossible to make a sound risk assessment from these summary data, there could, for instance, be some doubt as to whether this would be a suitable drug. It would be neither timely nor appropriate to make TABLE II SUMMARY FINDINGS OF CAFFEINE TOXICITYa Acute toxicity
LD 50 in rats:lOO-200 mg/kg; restlessness, diarrhoea, anuria; anorexia, convulsions, respiratory or cardiac arrest
Chronic toxicity
Man more sensitive to non-lethal doses than animals. Above 500 mg per day influence on dopaminergic system, tremor. In rats dosedependent decrease of food intake, weight loss and diminished weight gain. Excitation leading to auto mutilation. Increase of relative organ weights, especially of the adrenals. Glycosuria, elevation of free fatty acids. No organ toxicity.
Toxicity of the reproductive system
Rats: increased neonatal death (500 u/ml in water), increase of plasma cholesterol, bilirubin, calcium and retardation of ossification in newborn (2550 or 100% in water), testicular weight decrease in males (25-30 mg/kg) Chicken embryo: potentiates teratogenic effect of ephedrine at 0.5 umol per embryo Man: inhibition of spermatogenesis at high chronic doses, low weight of newborn and decreased cranial perimeter (300 mg/day)
Mutagenesis
Most tests negative, occasionally chromosomal breaks described
Carcinogenesis
Negative, except in one experiment colon tumors epidemiology in man has suggested increase in the rate of pancreatic tumors
‘After “Note de synthese, SNIP, 1984 131
153
a definitive decision. Data on daily exposure, clinical benefits, pharmacokinetics, and reliability of the data, amongst other things, need to be taken into account. Decisions and recommendations would be different if this substance was at the development stage or in use. Given the above data and knowing that human use is in the range of 4-10 mg/kg/day, most risk assessors would classify this substance as having moderate or low toxic risks. Would they remain convinced if some spectacular media news on malformations in children or stillbirths put the blame on the use of this compound by the mothers? The product described above is caffeine. It might be said that the argumentation used is far-fetched because this product has been used for centuries with no major toxicity problem. The discussion on herbal teas in Switzerland proves the opposite. Because some of the teas have definite pharmacological actions, there were serious plans not to allow restaurants to sell such products freely to the public. Some herbal infusions contain very toxic substances. Therefore all pyrrolizidine-alkaloid containing herbal preparations are strongly regulated in Germany. Risk assessment should not be restricted to one single moment of a product’s lifetime. It should be repeated whenever new risk factors become known 141. CONCEPTS
IN CARCINOGENICITY
TESTING
An important part of the modern toxicity testing is the carcinogenicity study and the mutagenicity tests. They are one of the hottest battlefields of toxicology today and their value for prediction for humans and their necessity are debated and challenged. On both sides, experienced scientists argue their case [5,61. A carcinogenicity study is such a complex study that any simplification could make prediction easier but would, at the same time, make the value questionable. Since human carcinogens have been shown to be carcinogenic in animal bioassays, such a test might be appropriate. Indeed, those who have undertaken a study with a genotoxic cancer-inducing agent, e.g. dimethylbenzanthracene or nitrosamine, and have seen the tumours appearing within weeks and metastasizing, believe in the power of such a study. Is the test therefore appropriate? Cancer, already difficult to understand for the scientist, has a demonic aspect for the public at large. If a chemical induces cancer, it should therefore be banned. Nowadays, however, the Delaney clause seems to be, in many cases, an inappropriate concept. New concepts should take into account an approach where acceptable risks are defined and not absolutely safe doses. Two problem complexes
154
should be mentioned here: the impact of diet and the MTD (maximum tolerated dose). When using the MTD it must be realized that many chemicals are metabolized by different pathways than at lower doses; the high dose may inhibit normal defence mechanisms or stimulate all kinds of additional, (nonlphysiological responses. Are then the high dose exposures predictive for a target population exposed to lower doses? Apparently, more than half of the chemicals that were found to be carcinogens in the NTP program would not have been so classified if the results obtained with the MTD were excluded 171. Diet plays a major - if not the major - role in cancer incidence both in rodents and in humans 181. There are calculations that cigarette smoking, alcohol and widely-used foods together are associated with almost all excess cancer. The combination from foods is about as great as that from smoking. If such findings can be substantiated and proven they might trigger a new view of cancer testing and cancer risk perception by chemicals, especially drugs. However, for such scientific concepts to be used in a timely and appropriate way, they also need to be comprehended by the public and by politicians. In mutagenicity tests - originally thought to be reliable predictors of carcinogenicity - it is even more difficult to understand what their results (often obtained in extreme test conditions) mean for man who has powerful defence mechanisms. Many chemicals show occasional positive results which either cannot be repeated or are, on repetition, negative. All the same, there are real mutagens. In this field and in other areas we need new concepts, e.g. to discover the mechanism of action of a toxic finding instead of meaningless tests 191,or to find ways to elucidate and predict allergies. CONCLUSION
The public seems to demand risk-free chemicals. In order to avoid censure, the regulatory agencies are extremely cautious and establish test procedures that may take many years after the initial development or discovery of a new product. Much of this time is used for toxicity testing. Of the examples, mentioned it can be concluded that in many cases newer scientific findings and concepts have found their way in a timely and appropriate manner into regulatory toxicology. The reason that this is not always the case lies in the complexity of the issue: the ambiguity of scientific results in many cases; legal constraints, ignorance or negligence or the dream of absolute safety. Since toxicity issues have become common problems, regulations are necessary. Improvements are still possible in the application of new scientific concepts. The toxicology societies can play an
155
important role in this complex field: results are presented and discussed at their scientific meetings. Working groups are validating experimental models. Information gaps are covered by publication and teaching. In addition, the complexity of risk assessment should be made more widely known and debated publicly. REFERENCES Geiling, E.M.K. and Cannon, P.R. (1938) Pathological Effects of Elixir of Sulfanilamide (Diethylene Glycol) Poisoning. JAMA 111,919-926. Carne, X., Costa, J., Arnau, J.M. and Laporte, J.R. (1986) Subacute lethal intoxication attributed to diethylene glycol stereate contained as an excipient in a topical cream. Abstract 1121 V Congress of Clinical Pharmacology, Stockholm. Anonymous (1984) Cafeine, note de synthese du syndicat de l’industrie pharmaceutique. Bass, R. (1986) Drugs and Risk Assessment. In: FEST, Toxicology in Europe in the Year 2000. Elsevier Publications, Cambridge, pp. 25-28. Ames, B.N. and Gold, L.S. (1990) T oo many rodent carcinogens: mitogenesis increases mutagenesis; Science 249,970-971. Weinstein, I.B. (1991) Mitogenesis is only one factor in carcinogenesis. Science 251, 387388. Doull, J. (1991) Historical Perspectives in Toxicology. In: New Horizons in Molecular Toxicology, Lilly Research Laboratories Symposium, Eli Lilly and Company, Bethesda, Maryland, pp. 5-8. Abelson, P.H. (1992) Diets and cancer in humans and rodents, Science, 255,141. Aldridge, W.N. (1981) Mechanisms of toxicity. TIPS 3,228-231.
149
Are newer scientific concepts in regulatory toxicology used timely and appropriately? C. Hodel” and R. Bassb Y$terling Winthrop Pharmaceuticals Research Department, Longvic @‘rance) and bInstitute for Drug Toxicology, Bundesgesundheitsamt IBGA), Berlin (Germany) Key words: Regulatory toxicology; Toxicity testing; Testing safety; Test procedures; Risk assessment SUMMARY Laws regulating toxicology (e.g. toxic thresholds allowed, poison classes or definition of necessary preclinical testing) might improve health and save lives. Scientific facts will always serve as a mandatory base for political decision-making, but there will also be additional influences (perception and acceptance of risks, possible benefits, economic considerations etc.). These latter factors may vary considerably from one society to another. The Delaney clause prohibited the marketing of any product which was found to be carcinogenic in animals. Due to their benefits, exceptions were made for drugs. In other countries, too, other chemicals could be an exception due to a different perception of the risk or different scientific evaluation. Clear cases of major events always trigger changes in legislation. When in 1937 a newly-marketed sulfanilamide elixir led to severe kidney damage and 70 deaths, the FDA quickly endorsed the propositions of the investigation team set up by the American Medical Association: animal testing in two species with histopathologic examination before a marketing authorization could be granted became mandatory. A similarly rapid reaction followed in Europe when it was detected that Thalidomide was responsible for malformations in the offspring of mothers who had taken the drug in early pregnancy. When the effects are more difficult to link to a chemical, there may be time delays in regulatory actions. However, a sophisticated evaluation system was introduced for better monitoring of drug and chemical hazards. Some examples will be given in order to discuss the difficulties of timely and appropriate use of scientific findings.
INTRODUCTION
There are peculiar aspects in toxicology. It is therefore not surprising that, at some time, the phrase “regulatory toxicology” came into existence a phrase which does not exist in biology, pharmacology, or even in medicine Correspondence to: C. Hodel, Sterling Winthrop Pharmaceuticals Research Department, Bd. Eiffel, 21600 Longvic, France.
150
and is not in common use. Regulatory toxicology probably underlines the subconscious wish that something dangerous (toxic) should be regulated to be not dangerous. It would be very reassuring if, by simple regulation (guidelines or laws), any toxic action or reaction could be eliminated. Enormous amounts of money are spent on testing safety and defining risks of new chemicals. The possibility of measuring minute concentrations of exposure and the impossibility of proving absolute safety when exposure is likely, tend to increase the legal requirements of test procedures. The scientist may find himself in a position where he has to perform experiments which he feels are unnecessary or absurd. A couple of years ago we were asked to run a pharmacokinetic study in dogs with a slightly new formulation of a suppository of an anti-inflammatory agent despite the fact that, in another country, the kinetics were already well known in humans. For chemicals with exclusively topical exposure, an oral carcinogenicity study might be demanded although with no resorption through the skin being measured. The regulatory toxicologist is often exposed to findings (positive or negative) which are difficult to evaluate. There might be sloppy experimentation, technical flaws or sometimes wrong tests (not validated, no absorption, etc.). The temptation just to follow the legal guidelines and checklist is evident. The university toxicologist is often so highly specialized that he seeks unnecessary one-sided perfectionism. In the 1970s a statistician asked for an acute toxicity test with 900 animals in order to be able to calculate mathematically an exact value with minimal standard deviation. In this complex field of influences (and sometimes also pressures), a way must be found to define the real issue - in our case, the toxic action - and define the risk quantitatively and qualitatively. Scientific methods are needed for this purpose; the experimental models should be sensitive and validated. Because of the complexity of the issue, there is sometimes the feeling that new findings are not incorporated into decision-making fast enough or appropriately, with well-accepted test models answering the relevant question. CASE STUDIES
The sulfanilamide story [lJ During the months of September and October 1937, at least 76 humans in various localities in the USA died as a result of poisoning by elixir of sulfanilamide. After analyzing this preparation, the AMA Chemical Laboratory found it to be essentially a 10% solution of sulfanilamide in about 72% diethylene glycol, together with some coloring and flavoring agents. There
151
were no contaminants, such as mercury, which might have had similar clinical symptoms produced by the elixir. Patients usually complained first of a sensation of heartburn, followed by nausea, abdominal cramps, dizziness, malaise and vomiting. The symptoms occurred about 24 h after medication was started. They were followed by oliguria and anuria. Oedema developed. Creatinine rose steeply. Later on, patients became comatose and died within 2-7 days after first medication. The main pathologic changes were tubular nephrosis and necrosis and centrilobular liver cell swelling. Apparently the makers of this product were unaware of its possible toxicity and distributed it without having tested it adequately on animals. It had not been submitted to the Council on Pharmacy and Chemistry or the American Medical Association nor did the Food and Drug Administration of the US Department of Agriculture know of its composition. Newspaper reports of deaths from this new medicine created consternation among many physicians who were using sulfanilamides. The elixir, its acting components (pure diethylene glycol and sulfanilamide) and synthetic composite were tested separately on groups of animals. Rats, rabbits and dogs were given the various drugs three times daily in graded doses. The experiments were devised with the object of developing conditions which might simulate those in human cases. The drug was, of course, immediately withdrawn from the market. In this case good toxicology work would have helped to detect that the solvent was the culprit and the sulfanilamide could have been saved as a medicine. Even a good literature search could have given a clue, although little had been published at that time on diethylene glycol. Not surprisingly, this disaster had consequences. Geiling and Cannon [ll presented nine postulates which shaped toxicology work for the following decades (Table I). TABLE I NECESSARY INFORMATION FOR DRUGS TO BE FURNISHED BY PHARMACEUTICAL COMPANIES IN THEIR NDAs, AS PROPOSED BY GEILING AND CANNON IN 1938 [l]. 1.
Chemical composition
2.
Acute toxicity studies in two species
3.
Chronic toxicity studies with different doses in different species
4.
Observation of animals
5.
Pathologic examination of animals
6.
Studies in animals with experimental lesions
7.
Studies on the absorption and elimination of the chemical
8.
Study of interactions
9.
Knowledge of idiosyncrasies and untoward reactions
152
The aforementioned example shows that animal experimentation could have helped in a scientifically appropriate way to define and avoid a major risk. Of course, such knowledge needs to be disseminated and must ultimately find its way into regulatory decision making. This can take days, weeks, months or years, depending on the manner of dissemination, the language barrier and different risk perception. In 1986, almost fifty years after the original diethylene disaster, Carne et al. 121published a report from a Burns Unit of a general hospital in Spain where, within some months, a couple of patients died with acute renal failure and metabolic acidosis. Apparently, some time before the first death, the brand of a widely used Ag-sulfadiazine cream had been changed. The cream contained 15% diethylene glycol which, in view of the typical syndrome, was to blame for the deaths. After Spanish legislation the ingredients had not to be disclosed on the package. This is an obvious case where scientific concepts have not been or were not used timely or appropriately, for whatever reason. A widely used pharmacological agent (caffeine) Table II gives an overview of toxicity data of a widely-used product [3)l Although it is impossible to make a sound risk assessment from these summary data, there could, for instance, be some doubt as to whether this would be a suitable drug. It would be neither timely nor appropriate to make TABLE II SUMMARY FINDINGS OF CAFFEINE TOXICITYa Acute toxicity
LD 50 in rats:lOO-200 mg/kg; restlessness, diarrhoea, anuria; anorexia, convulsions, respiratory or cardiac arrest
Chronic toxicity
Man more sensitive to non-lethal doses than animals. Above 500 mg per day influence on dopaminergic system, tremor. In rats dosedependent decrease of food intake, weight loss and diminished weight gain. Excitation leading to auto mutilation. Increase of relative organ weights, especially of the adrenals. Glycosuria, elevation of free fatty acids. No organ toxicity.
Toxicity of the reproductive system
Rats: increased neonatal death (500 u/ml in water), increase of plasma cholesterol, bilirubin, calcium and retardation of ossification in newborn (2550 or 100% in water), testicular weight decrease in males (25-30 mg/kg) Chicken embryo: potentiates teratogenic effect of ephedrine at 0.5 umol per embryo Man: inhibition of spermatogenesis at high chronic doses, low weight of newborn and decreased cranial perimeter (300 mg/day)
Mutagenesis
Most tests negative, occasionally chromosomal breaks described
Carcinogenesis
Negative, except in one experiment colon tumors epidemiology in man has suggested increase in the rate of pancreatic tumors
‘After “Note de synthese, SNIP, 1984 131
153
a definitive decision. Data on daily exposure, clinical benefits, pharmacokinetics, and reliability of the data, amongst other things, need to be taken into account. Decisions and recommendations would be different if this substance was at the development stage or in use. Given the above data and knowing that human use is in the range of 4-10 mg/kg/day, most risk assessors would classify this substance as having moderate or low toxic risks. Would they remain convinced if some spectacular media news on malformations in children or stillbirths put the blame on the use of this compound by the mothers? The product described above is caffeine. It might be said that the argumentation used is far-fetched because this product has been used for centuries with no major toxicity problem. The discussion on herbal teas in Switzerland proves the opposite. Because some of the teas have definite pharmacological actions, there were serious plans not to allow restaurants to sell such products freely to the public. Some herbal infusions contain very toxic substances. Therefore all pyrrolizidine-alkaloid containing herbal preparations are strongly regulated in Germany. Risk assessment should not be restricted to one single moment of a product’s lifetime. It should be repeated whenever new risk factors become known 141. CONCEPTS
IN CARCINOGENICITY
TESTING
An important part of the modern toxicity testing is the carcinogenicity study and the mutagenicity tests. They are one of the hottest battlefields of toxicology today and their value for prediction for humans and their necessity are debated and challenged. On both sides, experienced scientists argue their case [5,61. A carcinogenicity study is such a complex study that any simplification could make prediction easier but would, at the same time, make the value questionable. Since human carcinogens have been shown to be carcinogenic in animal bioassays, such a test might be appropriate. Indeed, those who have undertaken a study with a genotoxic cancer-inducing agent, e.g. dimethylbenzanthracene or nitrosamine, and have seen the tumours appearing within weeks and metastasizing, believe in the power of such a study. Is the test therefore appropriate? Cancer, already difficult to understand for the scientist, has a demonic aspect for the public at large. If a chemical induces cancer, it should therefore be banned. Nowadays, however, the Delaney clause seems to be, in many cases, an inappropriate concept. New concepts should take into account an approach where acceptable risks are defined and not absolutely safe doses. Two problem complexes
154
should be mentioned here: the impact of diet and the MTD (maximum tolerated dose). When using the MTD it must be realized that many chemicals are metabolized by different pathways than at lower doses; the high dose may inhibit normal defence mechanisms or stimulate all kinds of additional, (nonlphysiological responses. Are then the high dose exposures predictive for a target population exposed to lower doses? Apparently, more than half of the chemicals that were found to be carcinogens in the NTP program would not have been so classified if the results obtained with the MTD were excluded 171. Diet plays a major - if not the major - role in cancer incidence both in rodents and in humans 181. There are calculations that cigarette smoking, alcohol and widely-used foods together are associated with almost all excess cancer. The combination from foods is about as great as that from smoking. If such findings can be substantiated and proven they might trigger a new view of cancer testing and cancer risk perception by chemicals, especially drugs. However, for such scientific concepts to be used in a timely and appropriate way, they also need to be comprehended by the public and by politicians. In mutagenicity tests - originally thought to be reliable predictors of carcinogenicity - it is even more difficult to understand what their results (often obtained in extreme test conditions) mean for man who has powerful defence mechanisms. Many chemicals show occasional positive results which either cannot be repeated or are, on repetition, negative. All the same, there are real mutagens. In this field and in other areas we need new concepts, e.g. to discover the mechanism of action of a toxic finding instead of meaningless tests 191,or to find ways to elucidate and predict allergies. CONCLUSION
The public seems to demand risk-free chemicals. In order to avoid censure, the regulatory agencies are extremely cautious and establish test procedures that may take many years after the initial development or discovery of a new product. Much of this time is used for toxicity testing. Of the examples, mentioned it can be concluded that in many cases newer scientific findings and concepts have found their way in a timely and appropriate manner into regulatory toxicology. The reason that this is not always the case lies in the complexity of the issue: the ambiguity of scientific results in many cases; legal constraints, ignorance or negligence or the dream of absolute safety. Since toxicity issues have become common problems, regulations are necessary. Improvements are still possible in the application of new scientific concepts. The toxicology societies can play an
155
important role in this complex field: results are presented and discussed at their scientific meetings. Working groups are validating experimental models. Information gaps are covered by publication and teaching. In addition, the complexity of risk assessment should be made more widely known and debated publicly. REFERENCES Geiling, E.M.K. and Cannon, P.R. (1938) Pathological Effects of Elixir of Sulfanilamide (Diethylene Glycol) Poisoning. JAMA 111,919-926. Carne, X., Costa, J., Arnau, J.M. and Laporte, J.R. (1986) Subacute lethal intoxication attributed to diethylene glycol stereate contained as an excipient in a topical cream. Abstract 1121 V Congress of Clinical Pharmacology, Stockholm. Anonymous (1984) Cafeine, note de synthese du syndicat de l’industrie pharmaceutique. Bass, R. (1986) Drugs and Risk Assessment. In: FEST, Toxicology in Europe in the Year 2000. Elsevier Publications, Cambridge, pp. 25-28. Ames, B.N. and Gold, L.S. (1990) T oo many rodent carcinogens: mitogenesis increases mutagenesis; Science 249,970-971. Weinstein, I.B. (1991) Mitogenesis is only one factor in carcinogenesis. Science 251, 387388. Doull, J. (1991) Historical Perspectives in Toxicology. In: New Horizons in Molecular Toxicology, Lilly Research Laboratories Symposium, Eli Lilly and Company, Bethesda, Maryland, pp. 5-8. Abelson, P.H. (1992) Diets and cancer in humans and rodents, Science, 255,141. Aldridge, W.N. (1981) Mechanisms of toxicity. TIPS 3,228-231.
