OELs to carcinogenic substances: the additional risk at current TLVs.
P. Valente*, MD, F. Cavariani**, PhD

* Health Local Unit Roma G, Department of Prevention (Italy).
** Health Local Unit Viterbo, Industrial Occupational Laboratory (Italy).

ABSTRACT
Background: At the present there is no consensus on the levels of exposure to carcinogenic agents which can be considered "acceptable". Different countries and international agencies have proposed or have adopted different limits, on the basis of different criteria, for those exposed in the occupational setting and for the general population.
Objective: To illustrate the magnitude of the additional risk to which occupationally exposed groups could undergo on the basis of the current TLVs for a number of carcinogenic substances.
Methods: The additional risk of cancer to workers (A.R.(w)) has been derived from the Unit Risk (U.R.) produced by US EPA as follows: A.R.(w) = (U.R. X C) / 8, where U.R. = P_O (R-1)/X; P_O = background lifetime risk for the general population; R = relative risk; X = lifetime average exposure; C = exposure concentration equal to ACGIH TLV-TWA (expressed in µg/m³).
The formula takes into account the time spent at the occupational setting (1/8 of the lifetime). These calculations were carried out for all substances in the R45, R49 and R40 category for which U.R. (by the USA-EPA) and C (by ACGIH-TLV-TWA) were available.
Results: The results show that the additional risk for the occupationally exposed groups (A.R.w) at the current ACGIH TLVs-TWA, is always higher than the risk considered "acceptable" for the general population (set by USA-EPA, at 1/100.000). The ratio A.R.(w) / 10-5 is in the order of 1.000 times higher for the majority of 18 substances investigated.
Conclusions: Despite the many uncertainties involved in risk assessment and in risk management, the excess of cancer risk to workers estimated in this study shows that present day prevention policies are inequitable from the ethical point of view and necessitate further study and new solutions.

Introduction
The control of occupational exposure to carcinogens represents an important action in the primary prevention of cancer. Chemical exposure standards at workplace appear under different names, such as Threshold Limit Values (TLVs) or Maximum Acceptable Concentration (MAC) or Maximum Concentration at the Workplace (MAK) .
In 1977 the ILO proposed the generic term occupational exposure limit (OEL) which also accepted by WHO (1981).
OELs are set to protect workersí health. However , the OELs depend highly upon the definition of health that was chosen as the basic criterion.
Exposure values differ from country to country, mainly because of the use of differing basic concepts. The fundamental issue that underlies the different practices world wide is the lack of consensus on the criteria to be applied in setting values.

Historically, according to Robinson (1989) there are several approaches to setting occupational exposure values:

1. the exposure value is to be such that, if not exceeded, the health of the average person will be protected; a variation is to select the value to protect the whole population, including a sensitive subpopulation;
2. the exposure value is to be set to take account of economic factors and to provide a compromise between benefits and the costs associated with some (defined) risk to those so exposed;
3. the exposure value is a number used primarily in relation to engineering design consideration and without any detailed assessment of potential adverse effects upon health.

A standard provides a description or definition of a set of conditions that define quality, in this case, of occupational and environmental quality, and which are intended to protect human health and safety. It is clear that more approaches for standard setting need to be applied.
The OSH Act defines a PEL as the level that ensures " ...that no employee will suffer material impairment of health..." (OSH Act, 1970).
The OSHA is the organism, in the USA, specifically charged with the task of setting exposure limits for the workplace. Initially, in 1970, the OSHA adopted the TLVs of the ACGIH for all industrial chemical including carcinogens. The TLVs in fact were not established on the basis of carcinogenicity and have not recommended human monitoring requirements.
Recognizing that its PELs were significantly outdated, this organization promulgated new standards in 1989. That action resulted in the addition of 164 new substances and in the lowering of 212 existing PELs (Rappaport, 1993). In 1992 the Eleventh Circuit Court of Appeals overturned the new PELs hearing the criticisms of different nature, coming both from industry and labour , according to the interest of each party.
The court indicated that OSHA needed to perform quantitative analysis of risk for noncancer effects where possible, that more extensive discussion of the evidence for adverse effects for each substance was needed that feasibility analysis should be more detailed . Consequently, the Agency was obligated to revert back to enforcing the limits set in early 1970s and a total halt of the regulatory process resulted. The present day standards prevailing in the USA, with few exceptions are those of the ACGIH, as set on the experiments of several decads ago (Paustenbach, 1997).

The ACGIH defines TLVs for chemical substances as follows ì Threshold limit values refer to airborne concentrations of substances and represent conditions under which it is believed that nearly all workers may be repeatedly exposed day after day without adverse health effects. Because of the wide variation in individual susceptibility, however, a small percentage of workers may experience discomfort from some substances at concentrations at or below the threshold limit: a smaller percentage may be affected more seriously by aggravation of a pre-existing condition or by development of an occupational illnessî.
The TLVs are expressed as a time- weighted average (TLV-TWA), or as a Short-Term Exposure Limit (TLV-STEL) or as a Ceiling (TLV-C) Limit.
These limits are developed as guidelines to assist in the control of health hazards. These recommendations or guidelines are intended for the industrial hygiene professional to assist in control of exposures in the workplace.They are not developed for use as legal standards.
In 1979, the WHO Office of Occupational Health established a project on Internationally Recommended Health-based Limits in Occupational Exposure (WHO, 1995).
The WHO Study Group (1979) has proposed to use the term ërecommended health-based occupational exposure limitsí. This term was in accordance with the International Convention N. 148 adopted by the International Labour Conference. This term represents level of harmful substances in workplace air at which there is no significant risk of adverse health effects; this does not take into account technological and economic considerations and thus should be distinguished from operational exposure limits.

Occupational Exposure Limits for Carcinogens

The knowledge base about many substances and their adverse effects is often not complete or even very poor. For setting health based occupational standards the toxicological knowledge base requested is about the:

1. toxicokinetics
2. identification of the critical toxic effect (s)
3. mechanism of the critical effect
4. NOAEL or LOAEL
5. exposure/dose-response relationship for a specific effect or syndrome
6. indication of risk associated with a given dose, especially when the effect noted is not reversible
7. mutagenicity
8. carcinogenicity
9. reproductive toxicity
10. immunotoxicity
11. an assessment of the magnitude of the risk to groups of workers with hypersusceptibility.

These data basically coincide with the ones used in the risk assessment to general population.
The critical effect is the adverse effect that appears at the lowest exposure level and it is of particular interest , since an OEL that protects against the critical effect of a substance protects against all its adverse effects (Hansson, 1997).
The WHO health-based occupational exposure limit values have been derived from the integrated information of health risk assessment (WHO,1995). Acute, subacute and chronic toxic effects, metabolism and other toxicokinetic criteria, exposure-response relationship, critical adverse effects have been taken into consideration as the main criteria for the health-based OELs development.

A number of the OECD countries have introduced legislation concerning carcinogens.
Some of the legislation concerning carcinogens covers both the general population and occupational population settings, whereas some countries (e.g., U.S.A.) have separate legislation for the general population and the occupational population, and Canada and Australia have general legislation only for the occupational population.

The EU directive on carcinogens defined that the exposure limits will be published in an appendix to the directive at a later date, but for genotoxic or carcinogenic substances, in the EU no standards are yet developed, except for asbestos, vinyl chloride monomer and benzene for those specific directives have provided. Indicative limits has been proposed also for chloroform and carbon tetrachloride.

In The Netherlands, the Arboraad, an advisory board for the Dutch Minister of Social Affairs and Employment, proposed to separate the standards for genotoxic carcinogens from current traditional OELs , because a mortality risk is accepted when a standard for a genotoxic carcinogenic substance exceeds zero. So the Arboraad introduced the term technical limit concentrations (TLC values) for genotoxic carcinogens. Preconditions for using TLCs are (a) the availability of measuring methods and strategies; (b) the principle to keep exposure as low as reasonably achievable; and (c) justification for using a genotoxic agent (Stijkel and Reijnders, 1995).

In Germany standards for this category of substances have already been introduced in 1976 establishing ìtechnical guidance concentrationsî for occupational carcinogens (Technische Richt Konzentrationen - TRK ). The TRK represents the concentration in the workplace air (TWA 8h) that, according to the present state of practically feasible technology, can be achieved on the shop floor. The actual exposure levels should be as much as possible below the TRKs. They are principally based on technical, analytical and economic criteria.
It is considered that there is not enough information on effects of these substances to provide a legally enforced limit but that it is of value to establish a TRK to serve as a directive for necessary protective measures and surveillance through monitoring of existing exposures.
In 1972 ACGIH TLV Committee made a clear distinction between animal and human carcinogens. Human carcinogens were listed in two groups: (1) those with an assigned TLV and (2) those without an assigned TLV. In 1976 the TLV Committee published guidelines for the classification of experimental animal carcinogens. In essence this guidelines divided experimental carcinogens into three groups, according to the high, low and intermediate carcinogenci potency.
In 1987 the ACGIH TLV Committee categorized carcinogenic substances in two groups: A1 - Confirmed Human Carcinogens and A2 - Suspected Human Carcinogens. A1 carcinogens were defined as ìsubstances, or substances associated with industrial processes recognized to have carcinogenic potentialî, while A2 carcinogens were described as "chemical substances or substances associated with industrial processes, which are suspect of inducing cancer, based on either limited epidemiological evidence or demonstration of carcinogenesis in one or more animal species by appropriate methods"
A more recent attempt was made to create a third category (A3 - Animal Carcinogen) for experimental carcinogens with negative evidence of human carcinogenicity.
The ACGIH - TLV Committee recommended that exposure to carcinogens must be kept to a minimum. "Workers exposed to A1 carcinogens without a TLV should be properly equipped to eliminate to the fullest extent possible all exposureto the carcinogen. A1 carcinogens with a TLV and for A2 and A3 carcinogens, worker exposure by all routes should be carefully controlled to levels as low as possible below the TLV" (ACGIH, 1996).

The OSHA (1989) uses a weight-of- evidence approach to assess the carcinogenic potential of chemical substances. This approach involves examining all available human epidemiological studies, clinical and case studies, animal studies, mutagenicity studies and metabolic studies, combined with a quantitative assessment of cancer risk, to make determinations regarding the potential that occupational exposure to a substance increases the risk of cancer. OSHA relies most heavily on epidemiological studies of worker populations and well-conducted animal bioassays to make these determinations.

For those substances for which data are suitable for estimating quantitative cancer risks, OSHA relies on these estimates, in part for making significant findings.
In 1989, OSHA lowered the PELs - TWA or established new limits for a group of carcinogenic compounds (16 substances), based upon the quantitative carcinogenic risk assessment approach, using the linearized multistage model.

Risk Assessment Approach
The development of chemically induced cancer in humans and animals is a complex and multistep process that is not completely understood. It is currently believed that the mechanism by which cancer develops requires at least two stages: initiation and promotion.
Several mathematical models have been developed to estimate the cancer risk that is associated with exposure to low doses of carcinogenic substances.

The carcinogenic risk estimate is generally based on data derived from experimental studies and
from epidemiological studies. In addiction, physiologically based pharmacokinetic models, regarding saturability of activating and detoxifying metabolic pathways, have been developed to better define the exposure-dose response relationship.


The Netherlands
Since 1978 an approach to quantitative risk assessment of carcinogens has been used in The Netherlands for deriving health-based recommended exposure limits for human exposure to environmental carcinogenic substances (Health Council of The Netherlands, 1994).
The Dutch approach to risk assessment adopted the multistage model of carcinogenesis. The multistage model is compatible with an increase of the cancer incidence in an exposed population with a power of the dose of carcinogen.
Based on this model two categories of carcinogens were distinguished:

1. iniziators and complete carcinogens, that both modify DNA and act by a stochastic mechanism,
2. promoters and other substances acting as cocarcinogens that do not modify DNA and act by nonstochastic mechanism (HCN, 1978).

So, they estimate the cancer risk of initiator carcinogens on the basis of the dose-response curve: the excess cancer incidence at actual exposure levels is estimated by linear extrapolation through the origin (background tumor incidence subtracted) using dose-response data of the lowest carcinogenic ìlifetimeî dose whenever possible. Although the multistage model implies multihit kinetics, at very low doses the tumor incidence will be a linear function of the dose.
Generally, for promoters a threshold below which no adverse effects occur is assumed, so the quantitative risk assessment, in the Dutch approach, is based on no-observed-adverse-effect-level (NOAEL) and include a safety margin in which uncertainties regarding interspecies variation, sensitive human subpopulations, and the type of adverse effects observed are taken into consideration: the result of this assessment is the health-based recommended exposure limit.
The Dutch approach differs from the method that has been employed in the USA, by the Environmental Protection Agency (EPA), that has not made a distinction between initiating and promoting agents in its carcinogenic risk assessment (EPA, 1986) and has applied linear extrapolation at low doses without a threshold to all carcinogens.

United States
The U.S. Environmental Protection Agency' approach to evaluating carcinogenicity data follows the general format defined by the National Academy of Sciences (NAS, 1983) in its description of the risk assessment process.

The NAS report defines the risk assessment process in four elements:

1. hazard identification
2. dose-response assessment
3. exposure assessment
4. risk characterization

The EPA's risk extrapolation procedure allows to estimate an upper bound to human risk and not probable risk (or expected disease incidence). All substance identified as categories A, B1,B2, and some C are subjected to the same risk extrapolation.
Numerical estimates of risk can be presented in one or more of the following ways: (1) unit risk (2) individual risk (3) population risk (4) reference dose or reference concentration.
The confidence in the numerical risk estimate depends on

1. appropriateness of data to estimation of human carcinogenic risk
2. quality of study design
3. strenght of study results
4. appropriateness of model application to the data
5. support of risk estimate by data from collateral studies

The EPA has established guidelines for selecting the appropriate set of data for the assessment: first, the tumor incidence data are separeted according to the organ and site and type of tumor; second, all biologically and statistically acceptable data are presented; third the range of the risk estimates are presented with consideration of the biological relevance and appropriateness of route of exposure. Finally, because it is possible that humans may be more sensitive than the other species, the data set from long-term animal studies showing the greatest sensitivity (highest potency) is given the greatest credence.
Central in the EPA philosophy is the concept of "default". In essence, this concept implies that the protection of the population is the primary objective of all risk assessment. Therefore, when the evidence is not sufficient, gaps in knowledge should be filled with conservative ("default") assumptions.
This approach assumes that: 1) a substance found to be carcinogenic for one group of humans, may be assumed to be carcinogenic for all humans; 2) a substance which proves carcinogenic for animals, should also be assumed to potentially cause tumors to humans, unless proved otherwise; 3) when the dose-response correlation data are not available, a linear default approach should be assumed. This means that the risk becomes null, only at zero exposure level (EPA, 1996).

"Since the dominant view of the carcinogenic process holds that most cancer initiators cause irreversible damage to DNA, there is reason to assume that the dose-response of most carcinogens will follow a linear, nonthreshold relationship. The Office of Science and Technology Policy (OSTP,1985) recommends the use of models that incorporate low-dose linearity when the data are limited and the uncertainty exists regarding the mechanisms of carcinogenic action. In conducting
risk assessments for prior rulemakings, OSHA has generally relied on the linearized multistage model" (OSHA, 1989).

Denmark and United Kingdom
Denmark and United Kingdom divide carcinogens into genotoxic and non genotoxic agents and use different extrapolation procedures for each.
Denmark for genotoxic carcinogens uses the MLE from one and two hit models and a judgement made concerning the best outcome.
For non genotoxic carcinogens may be used the simplified biologically- based model of Thorslund; the Mantel-Bryan model may also be used as a conservative model for tumor promoters that do not interact with cellular receptors involved in growth control, whereas a linear model may be most appropriate for receptor binding promoters.
In UK, for genotoxic carcinogens, or for carcinogens that do not appear to be genotoxic but for which no mechanism of action has been established, a prudent approach is taken that assume no threshold. Actual dose-response estimation left to scientific experts on a case-by-case basis.
United Kingdom treats nongenotoxic carcinogens as threshold toxicants. A NOAEL and safety factor are used to set ADIs.

Nordic Council
A recommendation from Nordic Council of Ministers (1986) on guidelines has been adopted in classification of carcinogens in Norway and Sweden. It was recommended to allocate carcinogens in three potency groups : high, medium and low potency carcinogens. The allocation involves a comprehensive evaluation where epidemiological studies, the TDx values (the lowest dose in mg/kg body weight/day which has induced a significantly increased number of tumors in long- term animal experiments), dose-response relationship, information on mechanisms including genotoxicity and toxicokinetics are considered (Sanner, 1996).

World Health Organization (WHO)
A linear model was used by the WHO experts considering carcinogens in the Air Quality Guidelines (WHO,1987). The air quality standards were derived from appropriate data, calculating the Unit Risk (UR) and the risk of developing cancer on exposure to environmental levels of carcinogen.
"Efforts should be made to harmonize approaches to the quantitative assessment of carcinogenic risk. The quantitative evaluation of carcinogen risk to humans should rely on observed induction of cancer in humans and in experimental systems, an evaluation of the mechanisms involved under human and experimental conditions, and evaluation of likely incidence of cancer in humans (with attendant uncertainty) in the light of mechanisms involved under human exposure conditions" (Moolenaar, 1994).

Methods

The Unit Risk is "the additional lifetime cancer risk occurring in an hypothetical population in which all individuals are exposed continuously from birth troughout their lifetimes to a concentration of 1ug/mc of the agent in the air they breathe" (US EPA, 1985).
So, it is a quantitative assessment of carcinogenic risk, based on human data, associated with lifetime exposure to a certain concentration of a carcinogen in the air.

WHO and US EPA calculate the Unit Risk (U.R.) on the basis of the following formula:

UR = P₀ (R-1) / X

Where P_O = background lifetime risk; this is taken from age/cause-specific death or incidence rates found in national vital statistics tables;
R = relative risk, being the ratio between the observed (O) and expected (E) number of cancer cases in the exposed population; the relative risk is sometimes expressed as the standardized mortality ratio SMR = (O/E) x 100);
X = lifetime average exposure.

An estimate of incremental risk of cancer to workers has been derived from the US EPA Unit Risk simply applying a correction factor that take into account the difference in duration of exposure between workers and general population.
Considered that the fraction of lifetime spent by people in working activity may be estimated equal to 1/8 of lifetime (8/24 x 240/365 x 40/70), it is plausible to hypothesize for the workers exposed at concentration levels 8 times higher than the general population the same additional risk for cancer (a).

(a) A.R. (w) = (U.R. x C) /8

where A.R. (w) is the additional risk calculated on the basis of exposure level ( C) for 1/8 of lifetime ( 8/24 hours, 240/365 days, 40/70 years) at the carcinogenic substances and UR is the Unit Risk developed by EPA.

C = ACGIH TLV - TWA (1998) expressed in µ/m³
The formula takes into account the time spent at the occupational setting (1/8 of the lifetime) These calculations were carried out for all substances classified Carcinogens Cat 1-2-3 (R 45, R 49, R 40) by European Union, for which the Unit Risk (by Us-EPA) and C (by ACGIH TLV-TWA) were available. The A.R. (w) is simply derived from the EPA Unit Risk.
Except for the US EPA Unit Risk, the data reported in table 1 were extracted from the TLVs - BEIs ACGIH (1998). The Units Risk were selected from the Integrated Risk Information System (IRIS) file on-line by US EPA.


Table 1. ACGIH data set about 18 carcinogenic substances which are classified by EU in Carc. Cat. 1-2-3 (R 45 - R49 - R40) and for which the Unit Risk by US-EPA is available.

Results

On the basis of these considerations, a comparison of incremental risk of cancer between workers and general population is shown in table 2.
In the last column of table 1 a ratio AR (w) / acceptable risk (1 in 100.000) has been reported.
The Additional Risk / acceptable risk (1 in 100.000) ratio is: >1000 in 11/18 substances, > 100 in 14/18 substances and > 10 in 17/18 substances.
Of course these results strictly depend on the unit risk value developed by US EPA, that may not be appropriate in the case of very elevated exposures.

Table 2. Additional Risk to workers (AR w) derived from E.P.A. Unit Risk.

This comparison, anyway, permits to develop more specific considerations on the degree of health protection guaranteed by the respect of TLVs in the workplaces and also on the acceptability of the levels of risk experienced by workers, confronted to the general population.
The results depend on the Unit Risk value estimated by US EPA that may not be appropriate for elevated exposures and indicate an excess of cancer risk to workers generally 1000 times higher than the level of risk considered acceptable for the general population .
This risk is significant at exposure levels equal to the ACGIH - TLVs ) and could hypothesize over-exposed categories of workers to some carcinogenic substances and demonstrates that the TLVs considered are outdated and not sufficiently protective.

Discussion

In formal risk assessment it is attempted to use explicit and replicable risk criteria, defining the assumptions made in each step. Nevertheless, the process of risk assessment even when formalized, remains speculative, with a large margin of uncertainty.

ncertainty arises from many different sources: the quality of data used, the extrapolation from animal to humans, the shape of the dose-response relationship curve are all related to the risk evaluation. Another important source of uncertainty relates to the adjustment of workplace exposures (from epidemiological studies), to equivalent lifetime exposure. It's implicitly assumed, for example, that someone exposed to a cumulative dose X over 40 years, has the same risk as someone with the same cumulative dose over only in few years, absorbed in a more concentrated way in the workplace. These uncertainties are further corroborated by the lack of a well established theory of the mechanism of environmental carcinogenesis.

In his comparative analysis of risk assessment procedures used by ten different bodies, Moolenar (1994) showed considerable differences both in the criteria used for assessing the carcinogenic risk, and in the final risk characterization phase. The main reason of these differences is to be attributed to the different objectives that each organization set out in undertaking the risk assessment, and the final use to which the information is destined.

n some countries the risk assessment procedure may be undertaken for the purpose of labeling of products, in others for the protection of working populations, and still other for the protection of the environment and the general population.

n order to bridge these gaps and to harmonize policies, at least for what concerns the workplaces, the EU has recently institued a formal scientific group (Scientific Committee for Occupational Exposure Limits to Chemical Agent - SCOEL), with the mandate of producing recommendations for limits to be adopted by the European Commission as a directive ( Hunter et al. , 1997).

ntil now the EU Commission has established OELs binding limit only for vinyl chloride, asbestos and benzene. It will be interesting to see how the EU Commission will proceed on this field.

n an recent analysis of the situation in Sweden Hansson (1997) showed that even in this country, often cited as model of correct social policy for occupational health, the coordination between the scientific body and the regulating body is not satisfactory, and the decision about exposure limits to carcinogens are not sufficiently protective .

n particular, in Sweden, the scientific body has been traditionally concerned with setting limits to chemicals on the basis of the evidence of the "critical effect" of a given substance. The critical effect is the adverse effect that appears at the lowest exposure level and involves, in most cases, eye irritation. This criterion is obviously not adequate for setting limits for carcinogenic substances. In fact the list of 330 substances with an occupational exposure limit contains 35 carcinogens. Of these only 15 were assigned carcinogenicity as their "critical effect", whereas the others have exposure limits on the basis of eye irritation or other critical effects, which are much less serious than the potential carcinogenic effect.

In the Usa the problem of classification and risk assessment of carcinogenic substances is being dealt with by a number of federal and state agencies, each using a different approach, and each having a different objective. In addition, in the last two decades, there have been changes in the methodologies and the evaluation criteria of carcinogens adopted by some agencies and not others.

he National Toxicology Programme, for example, has been active since 1978 in classifying carcinogenic substances, initially in two classes (carcinogenic and non ) and since 1982 in five classes according to the level of evidence. The agency produces its own experimental evidence and is therefore in a position to undertake specific experiments in order to resolve doubtful cases. The data produced by the NTP are used by other bodies who undertake risk assessment activities, mainly OSHA and EPA (Huff, 1992).

The differences in the limits set by ACGIH and EPA have very different origin, objectives and evolution in the last decades. Substantially, the former are expression of the concern to maintain the status quo in the occupational health settings irrespectively of the scientific and social developments of the recent decades.

The EPA standards are an expression of the political will to reduce the risks of the general population even at the cost of being excessively cautious.

Independently of the origin of the differences in objectives and methodology of risk assessment procedures however, the fact remains that at the present two different standards are used for the protection of citizens of the same country from carcinogenic substances. One standard, more stringent for the general population, and another, more tolerant, for the working population. The ethical issues raised by this inequity are not difficult to see.

Dilemmas arise in dealing with a number of issues in the regulatory process: the notification on the part of scientist - (risk analyst) of the risk to the interested parties, the partecipation of the scientific community in the risk assessment process, the proposal of solutions to prevent the technological hazards, the testimony of the scientist in court for the compensation of harm caused by occupational exposures.

What's the margin of uncertainty, and the margin of safety to be used for the protection of the long term health of workers?

The problem became more complex at the individual level, where individual susceptibility, but also need to keep ones job come in conflict.

In his role the risk assessor is faced with still other dilemmas. Despite the popular idea that science is "neutral", establishing the "Truth" in the biomedical sciences, and in risk assessment in particular, is an approximation process which is never 100% certain . The way data are interpreted is clearly influenced by the values held by the individual scientist, and by anticipated use of the data (Ashofd, 1988).

Often scientists are in the unhappy position to be constrained to chose between to possible errors: the error of overestimating the risk (type I error) and wasting resources by unnecessary regulation, and the error of underestimating the risk (type II error) and failing to regulate a risk which later turns out to be harmful. Striking the correct balance between these two positions is obviously determined not only by scientific considerations, but also by value judgements. In addiction, past experiences have often shown that in occupational health type II errors are more common than type I errors. It is enough to recall the history of regulating carcinogenic agents such as ionizing radiation, asbestos, VCM among many others.

The ethical principle behind the risk assessment and the risk management procedures is based on the theory of utilitarianism, which has influenced much the social thinking of the modern western world . In contrast to the previously prevalent intuitionism, which views actions as "good" or "bad", independently of their consequences, the theory of utilitarianism asserts that an action should be judged as good or bad, on the basis of its consequences. In a very simplified form this ethical approach asserts that the best action is the one which provides the greatest benefit for the greatest number of people.

Unfortunately, this approach has sometimes been interpreted in occupational health to imply that the unit of measuring the "benefit" is money, rather than the health and the life of workers and the general population (Lee, 1977).

In fact cost-benefit analysis , translating costs and benefits into monetary units is deceptively attractive and "scientific". The basic, conceptual fallacy in applying this reasoning in occupational health, however is that the costs are supported by social groups ( and individuals) that are different from those enjoying the benefits. This creates a situation of inequity an example of which we have seen in the different exposure limits for workers and for the rest of the population.

In addition , maximizing profits irrespective of risks to the heath of the workers, has been now shown to be a short sighted policy with long term negative effects not only for the workers, but a also for the health of the general population and the environment.

[Thanks to Mrs Ida Marcello for her precious aid for the collection of data and documents and thanks to Mrs Francesca Costamagna, Miss D'Angiolini Antonella and Mr Giovanni Riva for their patience and collaboration].

References

American Council of Governmental Industrial Hygienists (ACGIH). Documentation of the Threshold Limit Values and Biological Exposure Indices, 6^th Ed. (1996).
Asford NA. Science and values in the regulatory process. Statistical Sciences 3 (3) : 377-383 (1988)

merican Council of Governmental Industrial Hygienists (ACGIH). Threshold Limit Values and Biological Exposure Indices (1998)

Carter V.L. ACGIH Threshold Limit Values for carcinogens. Medicina del Lavoro 77:4; 445-448.

Cook, W. A. Occupational Exposure Limits for Carcinogens - Variant Approaches by Different Countries. American Industrial Hyg. Association Journal 50: A680-A684 (1989).

E.U. (1996). "Technical Guidance Documents in support of the Commission Directive 93/67/EEC on Risk Assessment for New notified Substances and the Commission Regulation (EC) 1488/94 on Risk Assessment for Existing Substances", E.U., Brussels.

Hansson S.O. Critical effects and exposure limits. Risk Analysis 17, 2:227-236 (1997)

Health Council of The Netherlands: Committee on the Evaluation of the Carcinogenicity of Chemical Substances. Risk Assessment of Carcinogenic Chemicals in The Netherlands. Regulatory Toxicology and Pharmacology 19: 14-30 (1994).

Henschler D. The concept of occupational exposure limits. The Science of Total Environment. 101: 9-16 (1991).

Holmberg Bo and Lundberg P. Assessment and Managment of Occupational Risks in the Nordic (Scandinavian) Countries. American Journal of Industrial Medicine 15: 615-626 (1989).

Hunter W.J., Aresini G., Haigh R., Papadopoulos P., Von der Hude W. Occupational exposure limits for chemicals in the European Union. Occupational and Environmental Medicine 54: 217-222 (1997)

Mikheev M.I. Toward WHO-recommended occupational exposure limits. Toxixology Letters 77: 183-187 (1995)

Moolenaar R.J. Carcinogen Risk Assessment: International Comparison. Regulatory Toxicology and Pharmacology 20: 302-336 (1994).

National Academy of Sciences (NAS) (1982): "Risk and decision making: Perspectives and research", National Council Committee on Risk and Decision Making, NAS, Washington, D.C.

ational Academy of Sciences (NAS) (1983): "Risk assessment in the federal government: Managing the process", National Council Committee on Risk and Decision Making, NAS, Washington, D.C.

ational Academy of Sciences (NAS) (1994): "Committee on Risk Assessment of Hazardous Air Pollutants, Science and Judgement in Risk Assessment, NAS Press, Washington D.C.

SHA. Air Contaminants, final rule. Federal Register 54: 2332-2893(1989).

Paustenbach DJ. Updating OSHA's Permissible Exposure Limits: Putting Politics Aside. AIHA Journal 58: 845-849 (1997)

Rappaport S.M. Threshold limit values, Permissible Exposure Limits, and Feasibility: The Bases for Exposure Limits in the United States. American Journal of Industrial Medicine 23: 683-694 (1993).

Robinson A.E. Standard setting and protection of human health and safety. American Journal of Industrial Medicine 15: 213-231 (1989).

Sanner T., Dybing E., Krese D., Roelfzema H and Hardeng S. Carcinogen classification systems: similarities and differences. Regulatory Toxicology and Pharmacology 23: 128-138 (1996)

Stijkel A. and Reijnders L. Implementation of the precautionary principle in standards for the workplace. Occupational and Environmental Medicine 52: 304-312 (1995)

US EPA. EPA's Approach for Assessing the risks Associated with Chronic Exposures to Carcinogens.Background Document 2 January 17, 1992. http://www.epa.gov/ngispgm3/iris/carcin

US EPA (1993): "The integrated Risk Information System (IRIS)", US EPA Office of health and Environmnetal Assessment, Environmental Criteria and Assessment Office, Cincinnati, Ohio, U.S.A.

S EPA (1996): "Proposed Guidelines for Carcinogen Risk Assessment", U.S. Federal Register, 61, No. 79, April 23, 1996.

HO (1987): "Air Quality Guidelines for Europe", WHO, Geneva.

ielhuis R.L. and Wibowo A.E. Standard setting in Occupational Health: "Philosophical" Issues. American JournaL of Industrial Medicine 16: 569-598 (1989).