SCOPE 49 - Methods to Assess Adverse Effects of Pesticides on Non-target Organisms

11.1 INTRODUCTION 

`Pesticides' is the umbrella term for chemicals or biologicals effective in the control of pests. This activity may involve unwanted growth and/or infestation of bacteria, fungi, insects, or other organisms in materials. The compounds can be divided into agricultural chemicals, household chemicals, veterinary products, wood preservatives, disinfectants, and preservatives in end-products. Agricultural chemicals may be used on non-nutritional crops such as flowers, bulbs, trees and garden plants, or on nutritional crops used for human or animal consumption. Household chemicals may be used outside or inside buildings, on curtains, or in wallpapers or textiles. Veterinary pesticides are used to prevent or control the growth or existence of unwanted organisms in stables, animal rooms, or on small or large domestic animals. Wood preservatives are used outside or inside buildings and sometimes even in packaging materials used for food products. Disinfectants are mainly used in industrial or medical environments, but they may also be used in materials that come into contact with food. This also holds true for preservatives in end-products, e.g., when applied in paper coatings used for packaging. Furthermore, such preservatives are used in paints, pastes, glues, textiles, or cutting oils. In some countries, some categories of chemicals are not included in regulations; this is particularly true for wood preservatives, disinfectants, and preservatives.

The widespread use of pesticides gives ample possibilities for such substances to come in contact with the environment, man, and animals, either incidentally or intentionally, acutely or chronically, depending on their use, their persistence, and their migrating properties.

Humans can be exposed to pesticides occupationally when producing or applying a substance, as users of the product, or as consumers of food or liquids containing residues. Whereas in the first case people may be aware of a possible exposure to the substance, in the third they may be completely unaware of the exposure, which is usually at low levels. In addition, people may inadvertently and unknowingly be exposed to products used in buildings or equipment.

Since pesticides by nature are used in concentrations sufficiently high to have detrimental effects on biological organisms, they are at the same time a serious threat to living organisms that were not intended to be controlled. In this way, pesticides can be compared with drugs: their biological activity is a prerequisite for use. At the same time, this means that the `margin of safety' between wanted and unwanted adverse effects may be small in contrast with, for example, food additives.

Variations in the use of pesticides forces regulatory bodies to consider such variations carefully with respect to required testing. If a substance is used only on non-nutritional crops, contamination of food products is unlikely and one will focus attention mainly on such acute aspects as acute oral, dermal, or inhalation toxicity, skin and eye irritation, and sensitization.

On the other hand, if the use of such a substance gives rise to residues in food or to contamination of a food product, subchronic oral toxicity testing is certainly indicated. In most cases, chronic toxicity testing (including carcinogenicity, mutagenicity, and reproductive toxicity) is warranted to predict possible effects, especially when humans are exposed repeatedly. This chapter focuses primarily on the assessment of chronic adverse effects of pesticides, although prescreen procedures used to make more appropriate use of chronic toxicity bioassays will be discussed as well.

11.2 STRATEGIES IN CHRONIC TOXICITY TESTING 

Toxicology includes both the gathering of data in biological systems and the interpretation and evaluation of those data to predict possible risk for humans. Currently toxicity testing has become quite a rigid procedure. Standardized protocols inhibit the real goal of toxicity testing: to obtain as complete as possible an understanding of the mechanisms of toxicity at the tissue, cellular and subcellular levels (i.e., `receptor' toxicology) (Kroes and Feron, 1984).

The main goal in toxicity testing is the identification and proper description of adverse effects a pesticide may produce, and in addition, knowledge concerning the doseresponse relationships for such effects. Moreover, information with respect to the mechanism of action should ideally be obtained as well. The two aims in toxicity testingobtaining a toxicity profile and the establishment of a safe doseare complementary and not interchangeable, since different goals and different methodologies may be involved (Doull, 1984).

Among regulatory bodies, a reasonable agreement exists concerning the basic pattern of toxicity testing, although a divergence of philosophy is apparent: for some, a cookbook approach specifies that all substances should be subject to the same pattern of tests; for others, a more flexible approach requires the generation of only those data needed to adequately estimate the potential risks of exposure to a chemical. The latter approach is to be preferred.

Toxicity tests are presently carried out according to consensus guidelines as adopted by the Organization for Economic Cooperation and Development (OECD, 1981). These procedures are helpful in providing an efficient and cost-effective means of testing for toxicity. 

The scope of toxicology has broadened considerably, especially over the past decade (Doull, 1984; Conning, 1986; Kroes and Feron,1990). The firmly rooted conventional approach based on empirical observationsi.e., quantification of biological effects with no elucidation of mechanismshas resulted in studies which are costly and consume many research resources, yet provide no insight into the causal role of chemicals in possible disease.

The basic questions remain how to predict long-term effects on the basis of results of relatively short-term observations, and how to extrapolate data from laboratory animals to humans. These questions can only be answered if more attention is given to pharmacokinetic and pharmacodynamic studies, which provide insight into absorption, distribution, metabolism, and tissue concentration at target site. Ideally, comparative information between humans and laboratory animals on target tissues and pharmacokinetic aspects should be made available; however, such data generally become available only after accidental exposure. Comparative studies on cell or tissue cultures of humans and laboratory animals may add to a better prediction of the possible hazard to humans.

When long-term exposure of humans to pesticides is anticipated, acute and subchronic toxicity testing is not generally considered to be sufficient; in that case, chronic toxicity testing may be necessary, as well as studies for such specific toxic responses as carcinogenicity, reproduction, neurotoxicity, behavioural toxicity, and immunotoxicity. The results of acute and subchronic tests should be thoroughly evaluated to determine which special studies to undertake. 

Chronic oral exposure of humans to pesticides can be expected when residues of agricultural pesticides remain in food after application on crops or reach the food through environmental pathways.

When substances are used as household chemicals inside buildings and longstanding inhalation exposure can be expected, subchronic inhalation studies may be pertinent; likewise, if their use gives rise to detectable residues in food, chronic oral testing may be warranted.

Similarly, subchronic dermal testing may be indicated when the use of chemicals in clothing textiles would lead to longstanding dermal exposure. Chronic oral toxicity testing may also be required for veterinary pesticides in those cases where their use leads to detectable residues in meat, meat products, milk, or eggs. The same holds for wood preservatives, disinfectants, or preservatives that reach food items in finite amounts. Depending on the results of short-term testing studies and the intended use of a pesticide, studies for carcinogenicity, reproductive toxicity, neurotoxicity, or behavioural toxicity may be performed as well.

It will never be possible to describe beforehand which procedure should be followed and which tests should be performed. Toxicologists should be as flexible as possible in their way of thinking and their approach to conducting toxicity studies. Their main goal, however, should be to gather as much information as possible concerning the toxicity profile, the doseresponse relationship, and the mechanisms of action involved to make the most reliable hazard assessment and safety evaluation.

11.2.1 TOXICITY TESTING: CURRENT STATUS AND PERSPECTIVES

The basic package required for toxicity testing generally includes acute, subchronic and chronic toxicity testing and special studies, as noted previously. Considerable changes in the methods used for investigation have been made in the last decades. Parallel with increasing insight into physiology, biochemistry, and morphology, many new parameters have been introduced to study the functional state of the organism, organs, and organelles. An increasing number of parameters is studied, and several parameters are examined frequently in long-term experiments. Likewise, extensive gross and microscopic examinations are being performed on a greater number of organs and tissues. One wonders if this increase in parameters studied provides more relevant information. A lack of balance still seems to exist in the parameters required for general toxicity testing; while much emphasis is being placed on functional tests for liver and kidney, far less attention is being paid to other targets such as the endocrine, cardiovascular, and central nervous systems.

11.2.2 BASIC REQUIREMENTS FOR CHRONIC TOXICITY TESTS 

In toxicity testing, the important factors are the chemical, route of exposure, selection of dose, selection and care of animals, caging, diet, temperature, humidity, parameters studied, data acquisition, presentation of results, and overall evalution and presentation. The Principles and Methods for Evaluating Toxicity of Chemicals (WHO, 1978), Principles for the Safety Assessment of Food Additives and Contaminants in Food (WHO, 1987), Principles for the Toxicological Assessment of Pesticide Residues in Food (WHO, 1990), and the critical appraisal in Long-Term and Short-Term Assays for Carcinogens (Montesano et al., 1986) provide a wealth of relevant information.

For long-term toxicity studies or special toxicity studies, one should determine detailed physico-chemical properties, and the administered dose or exposure concentration of the test chemical. Throughout the study, all efforts should be made to provide accurate evidence of exposure. Selection of animals is extremely important (WHO, 1978; Fox et al., 1979; Montesano et al., 1986), but probably the most important criterion is knowledge of and experience with the test animal used.

The importance of diet as a major environmental variable in toxicity testing has been fully recognized in the last decade. The type, composition, and contaminant concentration may influence the health of animals and thus the effects of chemicals measured in tests. Different levels of macro- and micronutrients may profoundly influence results; high fat may change biotransformation enzyme activities and is associated with high incidence of certain tumours; low fat affects palatability (Kroes and Feron, 1990). High protein may influence the sensitivity to pesticides, whereas it may simultaneously increase spontaneous renal disease (Feldman et al., 1982). The acid-base balance of a diet may also affect the results of a toxicity study, especially when substances involved cause an acidic or basic load to the body (Kroes and Feron, 1990). Reduced food intake may strongly influence toxicity, as demonstrated by paracetamol, whose hepatotoxicity was reduced considerably with the reduction of food intake. Finally, careful monitoring of contaminants in the diet and drinking water (pesticides, mycotoxins, trace minerals, nitrosamines, and other suspect chemicals) is essential and should not be omitted. Other environmental variables should be controlled as adequately as possible.

Unfortunately, insufficient knowledge still exists on the influence of the number of variables such as light/dark cycle and noise, which may change circadian rhythm and biochemical functions and thus modify toxic responses (Fox et al., 1979).

As noted previously, parameters required for chronic toxicity testing may not have the desired balance. The increase in number of parameters has falsely introduced a feeling of confidence. Re-evaluation of the relevance of parameters used and consideration of new parameters to be included will be necessary in the coming decade.

11.2.3 CHRONIC TOXICITY TESTING

Prescreens for chronic toxicity testing are usually the acute and the subchronic repeated-dose studies performed on a substance of interest. These studies may provide information on the dose regimen to be used, the possible targets involved, the parameters to be studied, and the precautions to be taken. Retrospective studies have been performed to investigate whether certain ratios can be estimated in dose patterns between acute and subchronic and between acute and chronic studies (Weil and McCollister, 1963; Layton et al., 1987). These studies suggest the existence of some relationship between the no-observed adverse-effect level (NOAEL) in subchronic studies and chronic studies. If the NOAEL of an adequately performed subchronic study is divided by a factor 510, one has an approximate estimate of the chronic NOAEL. Similarly, retrospective calculations were made comparing even LD₅₀ to a chronic NOAEL, indicating a ratio varying for 1 x 10³ to 1 x 10⁴. Table 11.1 provides estimated conversion factors among the LD₅₀, the NOAEL of a subchronic or chronic study, and the acceptable daily intake (ADI). Such conversion factors are of course arbitrary, especially when derived from values, such as the LD₅₀, obtained from acute toxicity studies. Therefore, they should be taken with great caution when used as a guide to establish an approximate estimate for a dose range in the design of chronic toxicity studies.

Table 11.1. Estimated conversion factors between an LD₅₀, the NOAEL of a subchronic or chronic toxicity study, and the ADI

Chronic toxicity studies today vary in duration (6-24 months) in different countries and incorporate different requirements (Patrick, 1983). For some time, a debate has existed on whether the design for chronic toxicity testing, excluding testing for carcinogenicity, should have a duration of more than 6 months. In a retrospective analysis by Lumley and Walker (1985), the pathological findings (except for carcinogenicity) after 6 months of treatment were not different from those after longer exposures; hence, the safety assessment for these compounds would have been unchanged by the longer duration of exposure. 

This conclusion has been questioned by Frederick (1986). He investigated 11 drugs for which toxicity data from studies of duration of less than one year and greater than one year existed and noted that pathological lesions were observed in the longer duration studies that were not uncovered in the studies of less than one year; these lesions occurred at doses 10200 times the proposed maximal human dose for drugs. It is doubtful whether such prolonged studies would have affected the setting of the NOAEL. Therefore, a duration of 6 months for chronic toxicity tests may be adequate; for some compounds, where toxicokinetic findings indicate that a steady-state will not exist before the end of the treatment, longer duration may be necessary.  

Chronic toxicity testing in the future should focus more on establishing a doseresponse relationship for the mechanism of toxicity. This approach may necessitate the use of additional groups of test animals subject to interim sacrifice to obtain additional relevant information, such as the presence of the parent substance or its metabolites at the target site and in blood, for subsequent comparison with findings from low-dose studies in human volunteers.

With the existence of subchronic toxicity data in both sexes, the value of continuing to test both sexes of experimental animals in chronic tests has been questioned. Restricting a study to one sex when no sex-related differences in toxicity are likely offers the opportunity to add satellite groups to obtain relevant information. Parameters indicative for cardiovascular, endocrine, behavioural, or gastrointestinal function may be highly useful for safety evaluation, whereas some traditional parameters in blood and urine and some morphological parameters should be included only if they are relevant. For instance, morphological examination should perhaps be eliminated for organs that do not show evidence of change by gross examination, function tests, and organ weight at levels of exposure causing other toxicity.

Chronic toxicity testing of the future should be accompanied by in vitro studies using cells and tissues relevant to the target organ toxicity of the compound or its metabolite(s). Such in vitro studies may provide useful insight into the possible mechanisms of toxicity. In conclusion, chronic toxicity testing of pesticides should be made more tailor-made than it is today. The duration of the study may vary according to the pharmacokinetic data available and the design of the study may differ according to the results obtained in subchronic testing. The number of animals used, one or both sexes, the parameters, the number of groups, satellite groups, interim kills, all may differ according to the results obtained in acute and subchronic testing and according to the aim to be achieved. Comparative in vitro studies in both tissues and cells of experimental animals and humans, may complete the studies and provide optimal information for safety assessment and effect prediction.

11.2.4 CARCINOGENICITY TESTING OF PESTICIDES

A highly pertinent issue when pesticide exposure to humans is expected to be long term is carcinogenicity. Recently, the carcinogenicity of pesticides was evaluated by the National Research Council (1987), especially with respect to the health risks of residues in foods. This report has been the subject of considerable debate, since the risk assessment procedures used may have been inappropriate for the substances considered. Nevertheless, the possible continuous exposure of humans to small amounts of pesticides as residues in food is a major reason for conducting carcinogenesis studies. 

The identification of potential carcinogenesis by using in vitro methods has been ameliorated considerably in the last ten years. Recently, Ashby and Tennant (1988) concluded that chemicals that were likely to cause cancer by attacking DNA can be identified with some precision in short-term studies. Their findings that a correlation of more than 90 per cent exists between structural analysis of a chemical's reactive site and the Ames Salmonella assay confirm the existence of at least two groups of chemical carcinogens: those that are genotoxic and those that cause cancer in other ways (non-genotoxic compounds), as was earlier claimed by a number of authors (Kroes, 1979; Weisburger and Williams, 1980; Grice, 1984). 

Thus, in carcinogenicity studies, one can make fruitful use of short-term in vitro tests for genotoxicity (Grice, 1984; Montesano et al., 1986). Such in vitro tests contain two components: the cell, plant, or microbe in which genetic change is expressed; and an appropriate metabolic activation system. Endpoints are point mutations, chromosome breaks, deletions or transpositions, sister chromatid exchanges, and unscheduled DNA repair.

Since genotoxicity tests show significant differences in response for the varous carcinogens, their utility for carcinogen prescreening is restricted to the qualitative detection of genotoxic activity, thus predicting, but certainly not defining, carcinogenicity (Kroes, 1987).

A considerable proportion of genotoxic substances are not carcinogenic in animal assays, due in part to poor absorption of the chemical or in part to efficient detoxification in vivo. Therefore, a strong need exists for appropriate, validated short-term in vivo tests for genotoxicity. Positive findings in such in vivo tests might make long-term assays for the identification of carcinogens superfluous in the future (Ashby and Tennant, 1988). Only limited bioassays (for example, studies only in one sex or one strain of animals) might be needed for further characterization, unless quantitative risk assessment procedures are needed.

Long-term carcinogenicity studies, however, may still be necessary for those substances which have genotoxic properties in vitro, but may not be active in short-term in vivo studies. If the absence of carcinogenicity is confirmed for a large number of compounds, it may be appropriate to rely solely on short-term in vitro and in vivo studies for those substances having genotoxic potential in vitro.

Non-genotoxic carcinogens are characterized by a variety of biological properties that are believed to underlie their carcinogenicity. Examples are non-specific stimulation or inhibition, modulation of immune response, hormone balance, or nutritional factors (Kroes, 1979; Weisburger and Williams, 1980; Grice 1984). When non-genotoxic mechanisms are suspected, attempts should be made to obtain more pertinent information concerning the mechanism of action. Although short-term in vitro tests have been suggested to determine specific promoting properties (Brookes, 1981; Williams, 1981; Trosko et al., 1982; Yamasaki 1984), they still lack sufficient validation to be used in safety assessment. Limited in vivo bioassays (Williams and Weisburger, 1983) may provide adequate quantitative information relevant for safety evaluation.

The non-genotoxic nature of substances has been claimed to be indicative of the existence of thresholds (Weisburger and Williams, 1980; Kroes, 1979, 1987; Ames, 1987). In the future, it might be accepted that non-genotoxic substances do not have to be investigated in long-term bioassays for carcinogenicity, depending on results and anticipated use. If, for example, anticipated use of a non-genotoxic substance is at levels where human exposure is far below (i.e., one thousandth or less) a NOEL, carcinogen bioassays are considered by some to be superfluous (Feron and Kroes, 1986).

In performing carcinogenicity studies, the design and test methods are of primary importance in obtaining results suitable for safety assessment. The conduct of such studies has been extensively described (OECD, 1981; USEPA, 1985; Montesano et al., 1986). The combined chronic toxicity/carcinogenicity study (Test 453, OECD, 1981) has become quite popular in the last decade.

In conclusion, testing of carcinogenic properties of pesticides is important in those cases where continuous exposures to small quantities in humans can be expected. The use of appropriate prescreen procedures is advocated, especially when they might make further testing unnecessary, thereby diminishing the number of animals used. Under some conditions, it might be appropriate not to perform carcinogenicity studies when non-genotoxic substances are involved and when anticipated exposure in humans is at levels far below the NOEL established in adequately performed toxicity tests.

11.3 OTHER SPECIAL STUDIES

Reproduction is a vital function; therefore, in those cases where pesticide exposure can be expected, studies on fertility, reproduction and pre- and post-natal development are needed. For many compounds, these studies have been performed rather infrequently despite the relative importance of this type of toxicity. In a safety evaluation of 51 chemicals it appeared that in 70% of the substances studied reproductive and developmental toxicity were more sensitive than, or equally sensitive to, the conventional measures used in subchronic toxicity studies (Koëter, 1983). In vivo and in vitro pilot studies have been suggested as a prescreen, but better validation should first be obtained. As a cost-effective measure, selective fertility and reproductive parameters should be included in the subchronic study.

Recently, a report became available from a workshop where the so-called Chernoff/Kavlock preliminary developmental toxicity test was evaluated (Hardin et al., 1987). The test was found to be highly reliable in correctly identifying developmentally toxic chemicals for 165 chemicals. It was suggested that a negative finding in a properly conducted Chernoff/Kavlock test could be sufficient basis to determine that conventional teratology tests in the same species were not warranted. Although teratogenic properties may certainly be manifest in those exposed at high levels, it is believed that, at lower levels, thresholds exist for such phenomena (Wilson, 1977; Ames, 1987). Thus, the primary aim in reproduction studies will be the exclusion of teratogenic properties at levels 100 or more times greater than the anticipated human exposure.

Despite voluminous literature on reproductive toxicology, it was recently acknowledged that, although the correlation of reproductive effects among species is imprecise, adverse effects in experimental animals should be presumed to indicate a potential risk to human reproduction (Kimmel and Kimmel, 1986).

With respect to other special toxicity studies, the basis of knowledge about immuno-, neuro- and behavioural toxicity begins with a thorough toxicological characterization of the compound in subchronic studies.

The immune system can be influenced by chemicals in a variety of ways. Careful analysis of lymphoid organs for alterations in weight, morphology, cell count, and cellular function in classical toxicity testing may lead to suspicion of immuno-modulating potential. Such findings can then be confirmed and further characterized using a selected battery of presently available, validated assays, although there still exists a need for further development of assays in the rat, the animal of choice in chronic toxicity testing.

For neurotoxicity and behavioural toxicity, careful evaluation of the results of classic toxicity tests may shed light on such properties. For neurotoxicity, chemical structure and the half-life may prompt investigation of this phenomenon in chronic toxicity assays. Experienced personnel may be able to detect otherwise barely recognizable signs which may be indicative of neurotoxicity or behavioural disturbances. The possible behavioural effects of chemicals may be studied, where appropriate, in chronic toxicity studies by adding satellite groups that can be examined by established, validated methods. 

11.4 ASSESSMENT OF ADVERSE EFFECTS OF MIXTURES  

Pesticides are usually used as formulations in combination with other ingredients, often termed `inert'. When the toxicity of an active ingredient has been characterized, the further testing of formulations is usually restricted to acute studies. Combinations of pesticides in one formulation may necessitate chronic testing. In 1981, the Joint Meeting on Pesticide Residues (JMPR) concluded that, when combinations were used at concentrations or doses equal to the NOAEL, only additive effects may be expected. 

No single approach exists to assess adverse effects of mixtures. Since this issue is very complex and since only limited data are available about the approaches and limitations of testing mixtures, it is of extreme importance to use experience and knowledge in a flexible way. The type of the mixture, known or anticipated interactions among the components, and other information influence how a mixture should be tested and evaluated. 

In general, the toxicity testing of mixtures is not essentially different from that performed for individual compounds, although special considerations should be given to sample selection and collection, storage and stability of the mixture, possible interaction between components, characterization, and bioavailability (USEPA, 1986; Montesano et al., 1986; Vouk et al., 1987).

Short-term assays may well be used to evaluate the potential health hazard of complex mixtures. Special problems in testing may be anticipated, such as changing pH or osmolarity of the media occasioned by the necessary high concentrations, since the effective concentration of any one compound is diluted by the accompanying compounds. Short-term tests may give relevant information to be used if chronic/carcinogenicity tests seem warranted. 

11.5 ASSESSMENT OF ADVERSE EFFECTS OF BIOLOGICALS 

Currently, there is a tendency to replace chemical pesticides with biological agents. Experience already exists with several bacteria, viruses, and fungi  potentially usable as insecticides. No traditional procedures can be used for the assessment of their safety. Besides data on acute and subchronic toxicity, data on the ability of organisms to survive and multiply in the mammalian organism are particularly important. Bacteria can be studied in vivo in subchronic experiments, and viruses in cell culture. Moreover, information about the natural occurrence of these organisms is very important.

Biotechnology products are usually treated as chemicals and are routinely tested. Purity of the products and their immunogenic nature are of primary concern. Focus should be directed at the possibility that contaminants will be introduced inadvertently during fermentation or isolation. Although not pertinent to chemical pesticides, a biotechnology product's homology to native molecules is important, since slight changes in amino acid sequences may result in changes in the three-dimensional structure, thus changing its antigenicity. 

11.6 ADVERSE EFFECTS IN HUMANS  

Ideally humans should never experience adverse effect of pesticides; however, this ideal situation will probably never occur. First, occupational exposure in production or application may take place, not because safety rules are absent, but rather because safety precautions are not perfect. Second, pesticides are effective poisons which can be misused for suicide or homicide. Third, exposure of humans to residues of pesticides considered as safe, takes place every day.

Results from studies in humans exposed occupationally, accidentally, or via residues seem relevant to human risk assessment. Most likely, the best information from acute and subchronic exposure may be obtained from case studies. Epidemiological investigations are difficult to perform, except for those that can be designed in occupational settings; even they, however, have the intrinsic problem that humans in occupational settings are usually exposed to more than one chemical, thus making the interpretation of results rather difficult. Negative epidemiological findings may, however, provide some assurance that precautions have been effective. 

Epidemiological studies in the general population exposed to pesticide residues in food seem infeasible and not worthwhile because their sensitivity is insufficient to find effects, should any be present. More use, however, could be made of information from poison control centres, where patients accidentally exposed to relatively high single doses may provide information concerning uptake and excretion and may in addition provide information on metabolism and blood concentrations.

11.7 EXTRAPOLATION AND SAFETY ASSESSMENT

The evaluation of results of toxicity tests, including extrapolation to humans, is the cornerstone of safety assesment. The result of animal experiments should be weighed against the significance for people, since any animal model may be considered unsatisfactory because of existing species differences. We must rely on animal experimentation as long as results of more advanced methods such as in vitro/in vivo comparisons, physiologically-based pharmacokinetic modelling, and human volunteer studies are not available.

As a prologue to extrapolation and safety assessment, all chemical-biological data derived from toxicity studies should be evaluated and, if available, information on molecular or cellular mechanisms of toxicity is extremely relevant. In the extrapolation from laboratory animals to humans, one is faced with the following aspects: interspecies variation; intraspecies variation; extrapolation from high to low dose and experimental limitations (i.e., failure to detect certain effects in mammals).

In the fifties, the concept of an acceptable daily intake (ADI) was developed. The ADI is the daily intake that appears to be without appreciable risk during an entire lifetime, based on the facts known about a chemical. The ADI is expressed in mg/kg of body weight (b.w.), and is estimated from toxicity data by a WHO panel of experts or by national authorities. In the case of pesticides, this evaluation is done by the WHO/FAO Joint Meeting on Pesticide Residues. Assuming that a biological threshold is likely for non-carcinogenic effects, the safety assessment is made by establishing the NOAEL (expresed in mg/kg b.w.) from animal experimentation. This NOAEL is then divided by a safety factor to compensate for the factors influencing extrapolation, thus predicting a safe level for humans.

Assuming that interspecies variation and intraspecies variation would each not exceed one order of magnitude, a `worst case' approach yields a factor of 100 (i.e., 10 x 10) to be used as a reasonable margin of safety between the NOAEL in animals and the acceptable intake in man. This concept, now in use for almost forty years, has been shown to be sufficiently safe.

Recently, the validity of the use of a safety factor of 100 has been debated. For several reasons the factor of 100 may be high, since no evidence exists that the sensitivities of inter- and intra-species variability are related; thus, the worst-case approach may be overly conservative. Correspondingly, present-day toxicity testing is considerably advanced, and one can justifiably conclude that NOAELs found today will certainly be lower than those found in studies carried out decades ago.

One is, however, faced with an intake of a relatively substantial number of chemicals in small amounts; information on such interactions as inhibition, additivity, and synergism from combined exposures is scarce. Furthermore, a serious drawback in the procedure described above is that it takes no account of the dose-response relationship: where a steep curve may not necessitate the use of a safety factor of 100, a very shallow curve may need a margin of that magnitude. Therefore, better methods for determining allowable daily intakes should be developed. 

For pesticides, however, a conservative approach seems realistic, as long as such a procedure is not applied too rigidly.

11.8 MRLs AND THE PREDICTION OF DIETARY INTAKE 

Proposals for maximum residue limits (MRLs) on food crops are based in part on good agricultural practices. For residues of a pesticide on various crops to be judged acceptable, the total intake should not be higher than the ADI. However, predicting the true intake is often difficult. Calculation of the theoretical maximum daily intake provides a gross overestimation of true intake, because of uncertainties due to the variability in average daily food consumption, unaccounted reductions in pesticide content during food preparation and the assumption that residues are always present at the maximum level. Consequently, a joint FAO/WHO consultation has recently adopted guidelines for estimating dietary intake of pesticide residues (WHO, 19871989).

Using these guidelines, one identifies a theoretical maximum daily intake (TMDI), and an estimated maximum daily intake (EDMI) defined as a prediction of the maximum daily intake of a pesticide residue. The EDMI is derived from the assumption of average daily food consumption per person and the maximum residues on the edible portion of the commodity, corrected for the reduction in concentration of residues from food preparation, commercial processing, and cooking of the commodity. Both the TMDI and the EDMI are expressed as milligrams of the residue per person.

Finally, an estimated daily intake (EDI) is defined as a prediction of the daily intake of a pesticide residue, based on the most realistic estimation of residue levels in food and the best available food consumption data for a specific population.

Whereas TMDIs and EMDIs can be calculated grossly on international levels, EDIs can only be calculated at a national level, thus refining the system of setting maximum residue levels (MRLs). More research is needed to calculate with improved accuracy the likely intake of residues. The establishment of MRLs, as performed by the Codex Committee on Pesticide Residues, is a sensible international approach to this situation. From a public health perspective, it is reassuring to know that actual intakes of all pesticides (as measured in market-basket and total-diet studies) is much lower than the established ADIs.

11.9 REFERENCES

Ames, B. N. (1987) Six common errors relating to environmental pollution. Reg. Toxicol. Pharmacol. 7, 379-383.

Ashby, J. and Tennant, R. W. (1988) Chemical structure, Salmonella mutagenicity and extent of carcinogenicity as indicators of genotoxic carcinogenesis among 222 chemicals testing in rodents by the US NCI/NTP. Mut. Res. 204, 17-115.

Brookes, P. (1981) Critical assessment of the value of in vitro cell transformation for predicting in vivo carcinogenicity of chemicals. Mutat. Res. 86, 233-242.

Conning, D. M. (1986) Toxicologythe next decade. Arch. Toxicol., Suppl. 9, 222-224.  

Doull, J. (1984) The past, present and future of toxicology. Pharmacol. Rev. 36, 15S-18S.

Feldman, D. B., McConnell, E. E. and Knapka, J. J. (1982) Growth, kidney disease, and longevity of Syrian hamsters (Mesocricetus auratus) fed varying levels of protein. Lab. Anim. Sci. 4(32), 613-618.

Feron, V. J. and Kroes, R. (1986) The long-term bioassay for identifying carcinogens: some controversies and suggestions for improvements. J. Appl. Toxicol. 6, 307-311. 

Fox, J. H., Thibert, P., Arnold, D. L., Krewski, D. R. and Grice, H. C. (1979) Toxicology Studies II: the laboratory animal. Food Cosmet. Toxicol. 17, 661-675.

Grice, H. C. (1984) Current Issues in Toxicology: Extrapolation of Toxicity Data, Springer-Verlag, Berlin, Heidelberg, New York, Tokyo.

Hardin, B. D., Becker, R. A., Kavlock, J., Seidenberg, J. M. and Chernoff, N. (1987) Overview and summary: workshop on the Chernoff/Kavlock preliminary developmental toxicity test. Teratogen. Carcinogen. Mutagen. 7, 119-127. 

Kimmel, C. A. and Kimmel, G. L. (1986) Interagency Regulatory Liaison Group workshop on reproductive toxicity risk assessment. Environ. Health Perspect. 66, 193-221.

Koëter, H. W. B. M. (1983) Relevance of parameters related to fertility and reproduction in toxicity testing, Am. J. Ind. Med. 4, 81-86.

Kroes, R. (1979) Animal data, interpretaton and consequences. In: Emmelot, P. and Kriek, E. (Eds) Environmental Carcinogenesis, Elsevier/North Holland, Amsterdam, pp.287-302.

Kroes, R. (1987) Contribution of toxicology towards risk assessment of carcinogens. Arch. Toxicol. 60, 224-228.

Kroes, R. and Feron, V. J. (1984) General toxicity testing: sense and non-sense, science and policy. Fund. Appl. Toxicol. 4, 298-308. 

Kroes, R. and Feron, V. J. (1990) Toxicity testing: strategies and conduct. In: Clayton, D. and Munro, D. (Eds) Progress in Predictive Toxicology, Elsevier, Amsterdam. 

Layton, D. W., Mallon, B. J., Rosenblatt, D. H. and Small, M. J. (1987) Deriving allowable daily intakes for systemic toxicants lacking chronic toxicity data. Reg. Toxicol. Pharmacol. 7, 96-112. 

Lumley, C. E. and Walker, S. R. (1985) The value of chronic animal toxicity studies of pharmaceutical compounds: a retrospective analysis. Fund. Appl. Toxicol. 5, 1007-1024.

Montesano, R., Bartsch, H., Vainio, H., Wilbourn, J. and Yamasaki, H. (Eds) (1986) Long-Term and Short-Term Assays for Carcinogens: A Critical Appraisal, IARC Scientific Publications No. 83, International Agency for Research on Cancer, Lyon. 

National Research Council (1987) Regulating Pesticides in Food: the Delaney paradox, National Academy Press, Washington, DC.

OECD (Organisation for Economical Cooperation and Development) (1981) OECD Guidelines for Testing of Chemicals, OECD, Paris.

Patrick, C. P. (1983) Harmonization of toxicological testing in chemical safety evaluation. Chem. Ind. 1983, 55-59. 

Trosko, J. E., Yotti, L. P., Warren, S. T. Tsushimoto, S. C. and Chang, G. (1982) Inhibition of cell-cell communication by tumor promotors. Carcinog. Compr. Surv. 7, 565-585.

USEPA (US Environmental Protection Agency) (1986) Guidelines for the health risk assessment of chemical mixtures. Fed. Reg. 51, 34014-34025.

Vouk, V. B., Butler, G. C., Upton, A. C., Parke, D. V. and Asher, S. C. (Eds) (1987) Methods of Assessing the Effects of Mixtures of Chemicals, SCOPE 30/SGOMSEC 3, John Wiley & Sons, Chichester.

Weil, C. S. and McCollister, D. D. (1963) Relationship between short- and long-term feeding studies in designing an effective toxicity test. J. Agric. Food Chem. 11, 486-491.

Weisburger, J. H. and Williams, G. M. (1980) Chemical carcinogens. In: Doull, J., Klaassen, C. C. and Amdur, M. O. (Eds) Toxicology: The Basic Science of Poisons (2nd edn), Macmillan, New York, pp. 84-138.

Williams, G. M. (1981) Liver carcinogenesis: the role for some chemicals of an epigenetic mechanism of liver-tumor promotion involving modification of the cell membrane. Food Cosmet. Toxicol. 19, 577-583.

Williams, G. M. and Weisburger, J. H. (1983) New approaches to carcinogen bioassay. In: Holmberger, F. (Ed.) Safety Evaluation and Regulation of Chemicals, Vol. 2, S. Karger, Farmington, Connecticut, pp. 200-209.

Wilson, J. G. (1977) Current status of teratology. General principles and mechanisms derived from animal studies. In: Wilson, J. G. and Clarke Fraser, F. (Eds) Handbook of Teratology. 1. General Principles and Etiology, Plenum Press, New York.

WHO (World Health Organization) (1978) Principles and Methods for Evaluating the Toxicity of Chemicals. Part I. Environmental Health Criteria 6, World Health Organization, Geneva.

WHO (World Health Organization) (1987) Principles for the Safety Assessment of Food Additives and Contaminants in Food, Environmental Health Criteria 70, World Health Organization, Geneva.

WHO (World Health Organization) (1989) Guidelines for Predicting Dietary Intake of Pesticide Residues, World Health Organization, Geneva.

WHO (World Health Organization) (1990) Principles for the toxicological assessment of pesticide residues in food, Environmental Health Criteria 104, World Health Organization, Geneva.

Yamasaki, H. (1984) Modulation of cell differentiation by tumour promotors. In: Slaga, T. J. (Ed.) Mechanisms in Tumour Promotion, Vol. IV, CRC Press, Boca Raton, Florida, pp. 1-26.