Toxicology has a long and distinguished history as a branch of science concerned with protecting people against the unwanted effects of chemicals and in extending scientific and medical knowledge. The development during the twentieth century of standardized approaches to toxicological testing of pharmaceutical products, pesticides, food additives etc., has reduced the risk of unwanted effects occurring in the general population to a low level. Toxicology has also led to strict controls of exposure to chemicals used in industry and the virtual disappearance of many occupational diseases is another cause for satisfaction.
    The twenty-first century offers new and difficult challenges to toxicologists. It is remarkable that public concern about the possible effects of chemicals on health has risen steadily as toxicologists have striven to reduce the likelihood of such effects. Pressure groups calling for tighter regulation of chemicals are more active today than at any other time. This is due, in part, to the public becoming better, or perhaps more extensively informed, about the possible effects of chemicals on health. Access to information of variable quality via the Internet continues to expand and anyone wishing to compile a dossier on the toxicology of a chemical substance or group of substances can do so today from a home computer. This would have been difficult, if not impossible, for the vast majority 25 years ago. All those involved in dealing with public enquiries about chemicals will have come into contact with enthusiasts who have taken up the toxicology of single compounds or groups of compounds as a major interest, or who have concluded that a particular disorder must be due to a certain exposure. An increasingly interested and better informed public should be welcomed and ought not to be seen as a problem, but the pressure they engender does pose a serious challenge to the role of toxicologists in this new century.
    A number of other challenges are also appearing. There is an increasing interest in the need to protect individuals and susceptible groups. The public perceives that though exposure to a compound at an ‘acceptable level’ may, in general, do little harm, some people may be affected. That this is possible is, of course, admitted by all toxicologists: it is implicit in the use of a statistical approach to exposure_/response curves. The perception that protecting the majority from harm is insufficient appears to be growing and media attention is, inevitably, focused on individual cases for whom effective protection may lie in personal more than public measures. Recent work on the human genome has led the public to expect that prediction of individual susceptibility to chemicals may not be far off.1,2 Assertions that ‘we’ know of genotypes that lead to susceptibility to the effects of certain chemicals lead, inevitably, to the assumption that susceptibility to all chemicals is genetically determined and to demands for identification of such susceptibilities before exposure, at any level, is permitted. The strength of the claimed associations is rarely considered.
    Many toxicologists would agree _/ as a general point _/ that toxicological testing is better at identifying the effects of short-term or acute exposure to chemicals than the effects of chronic exposure. This point has been noted by the public. Of course, in some areas, for example carcinogenicity, toxicologists have always been interested in the effects of long-term exposure. In other areas, this has not been the case. Recent work which has suggested that chronic exposure to concentrations of carbon monoxide that produce only mild symptoms leads to damage to the central nervous system, is a case in point.3
    Increasing expectations of ‘complete’ health have led people to pay greater attention to what might be called infractions of well-being. This leads to the attribution of minor symptoms to external causes and, of these, exposure to chemicals is an obvious choice. Similarly, attention has focused on mixtures of chemicals. Toxicological testing focuses on single compounds, but people are increasingly aware that we are all exposed, daily, to mixtures and may assert that unexpected interactions occur. Toxicologists tend to call upon mechanistic arguments to deny that this is likely. Such a defence can be convincing if the mechanisms of the effects of the individual compounds are known in detail. Of course, this is unlikely and doubt, at least in the public mind, remains, and is sometimes exploited by those seeking to blame chemicals for many current ailments. Toxicologists sometimes speak of the plausibility or implausibility of an interaction. This is a dangerous line of argument, as plausibility can change rapidly as biological knowledge advances. It should be noted, Bradford Hill counseled against putting too much weight on ‘biological plausibility’ in his classic account of features of causal associations.4
    The growing pressure to limit the use of animals in toxicological testing makes ensuring public safety more difficult. All toxicologists accept that using animals to investigate the effects of chemicals, to predict effects on man and to identify levels of exposure which provide a basis for setting exposure standards is a regrettable necessity given the current state of development of in vitro methods. Toxicologists often deplore the contrary desires expressed by the public: safer chemicals and less animal experimentation. This pressure to reduce work in vivo has been particularly felt by those who use primates in their work and this will, inevitably, limit the exploration of the effects of chemicals on the more subtle aspects of the functions of the central nervous system. The last point makes it more difficult to deal with the perception and assertion that the outcome measures of standard toxicology are inadequate to predict effects on the central nervous system. Yet, complaints about possible effects of chemicals often include assertions of effects on memory, on mood, on the quality of cognition and on motivation. This has led to a call for an integrated approach and portmanteau terms, such as ‘neuro-immuno-endocrino-toxicology’, devised to reflect perceptions of the need to investigate complex and interactive systems are appearing. To the physiologist, calls for investigation of responses in integrated systems imply the need for in vivo methods. All these procedures are made more difficult by the litigious nature of society which precludes, too often, an open and ‘without prejudice’ discussion of possible toxicological effects. Additionally, calls for de-regulation run counter to many public perceptions of risk.
    These are some of the problems currently facing toxicologists. Are there any solutions? I think there are, but they are by no means easy and will call for something of a revolution in thinking amongst toxicologists and those using toxicological data to protect public safety. The following list contains both accepted and controversial ideas.
(i) Regulatory authorities should move away from the concept of ‘safety’ and embrace the concept of ‘estimated and controlled risk’. This will involve saying, repeatedly, that no guarantee of absolute safety can be provided. Efforts to predict risk in quantitative terms should be increased. The risk should be set in context and balanced, overtly, against benefit. ‘Risk’ and its converse ‘safety’ must be seen to be probabilistic and not absolute concepts.
(ii)The much misquoted ‘Precautionary Principle’ must be seen as a warning to consider potential risks and benefits ever more carefully as we enter areas of decreasing knowledge, but not as a complete barrier to change until there is ‘complete’ knowledge. The original purpose of the ‘Precautionary Principle’ as a tool for risk management should be stressed.
(iii)Regulators should also stress that benefit is as difficult to quantify as risk and that, in some cases, no obvious benefit of an exposure may be identified. That risks and benefits may be unequally distributed needs to be considered.
(iv)Regulators should move from an approach based on providing such complete protection that the public do not need to think for themselves, to one of providing the public with the information needed to make informed choices. This is likely to be opposed by some of the public, but seems to me to be the ‘other side of the coin’ of the explosion of access to information. Accepting that information technology is not a toy or designed merely for entertainment, is a challenge that the public will face and need to meet in the coming century.
(v)Toxicologists should acknowledge openly that unpredictable effects will occur. Currently, dealing with cases of unpredicted or unpredictable effects often involves litigation and the question of fault or liability is foremost in the arguments. This polarizes views and often prevents, in my view, proper investigation. Building into development costs an allocation for no fault liability payments would ease this problem.
(vi)Companies supplying chemicals should be under obligation to fund studies of alleged toxicological reactions. Research should be undertaken by independent contractors and all the evidence should be openly available. This will not be possible unless a no fault approach to compensation is adopted.
(vii)Governments should accept that a more developed approach to toxicological safety will have benefits and disadvantages. The benefits will include greater public confidence in the process and greater involvement in making decisions about their own risk profile. The disadvantages may include a slowing of development of new products.
 (viii)Governments should accept that toxicology as a scientific discipline will need to be significantly strengthened during the coming century. The need for independent toxicological investigation of problems will increase and steps to ensure that this is available should be taken. The establishment of the Health Protection Agency is an important move in this direction. In the UK, toxicology as a subject seems to be in decline and this needs to be addressed. As a subject, toxicology suffers from having its failures paraded and its successes ignored. This is inevitable in a subject that seeks and sometimes fails to prevent harm. 
    There are many challenges facing toxicology today. These will increase and great efforts will be needed to meet them. Toxicologists may ask what can they do to reverse the trends against them. Perhaps the most important point is the need to embrace public concerns and not to discard them as uninformed. This will require an unusually broad approach by toxicologists to potential problems and has serious implications for the training of toxicologists, something that has almost ended in the UK.
Acknowledgements  The author would like to thank Professor Anthony Dayan for his helpful comments on an early draft of this paper.
RL Maynard
Department of Health,
133-155 Wellington House,
Waterloo Road, London SE1 8UG, UK
E-mail: robert.maynard@dh.gsi.gov.uk