Assessment and Management of Cancer Risks From Radiological and Chemical Hazards

4. Risk Management

Risk management is the process in which the results of risk assessment along with other considerations are used to select and implement one or more strategies for controlling a risk. Standards used for risk management are set on the basis of not only the magnitude of risk, but also technical, economic, and socio-political factors. In Canada, there is no consensus on acceptable levels of risk for radiological, chemical, and micro-biological hazards. The acceptability of risk is often determined by judgements on such factors as the weight of scientific evidence, the nature, extent and severity of the hazard based on risk assessment evaluations, the degree of public concern, the benefits associated with the substance, product, or process, the cost and feasibility of reducing exposures, and regulatory agency policies defining limits on acceptable risks. In addition, consultation with stakeholders is necessary to determine de facto the acceptable residual risk after implementing risk reduction controls or measures. Judgements of acceptable or tolerable levels of risk are often difficult as a result of polarized positions which exist within society. In Canada, the weight given to various factors in risk management decision-making depends on the context in which risk management decisions are to be taken, including consideration of the legislation which applies.

If a cancer risk is judged to be significant or unacceptable, then it is generally expected that some action will be taken to reduce or eliminate the risk. In contrast, a de minimis or essentially negligible risk is one that is so small that no action needs to be taken. If a risk is judged to be insignificant or acceptable, however, this does not necessarily mean that it is de minimis or negligible [Sadowitz and Graham 1994].

The federal government has jurisdiction over activities that are considered to be of national interest, or of inter-provincial or international concern, and in setting minimum standards for protecting the health of Canadians and the environment. Provincial and territorial governments are responsible for the health and safety of their citizens, and have jurisdiction over industries within their borders, and in setting and enforcing provincial standards for health. In general, these standards cannot be less stringent than federal standards.

Approaches to risk management generally take the form of source control, point-of-use control, or educational strategies, and are the responsibility of both the federal and provincial governments. The application of legislative, technical and procedural controls may be different for source control and point-of-use control. Management tools include legally enforceable limits, regulations, and standards, as well as non- enforceable operating targets, guidelines, or goals for source and point-of-use control. Source control strategies limit human health risk by imposing regulations and operating criteria on the industry or process in question. Criteria governing the release of chemical contaminants into the environment are primarily set by provincial authorities, although the federal government has some jurisdiction under CEPA and its regulations. The regulation of radioactive emissions from nuclear facilities is entirely a federal responsibility.

Whereas source control strategies are source-specific, point-of-use control strategies are concerned with the levels of contaminants in air, food, and drinking water from all pollutant sources. Maximum allowable levels for contaminants are set through cooperation between the federal and provincial governments, and apply in addition to, but independently of, controls at the source. Such standards are not permits to contaminate up to the maximum values, but rather limits below which actual levels should be kept. This approach to risk management will be discussed later in this report in the context of the Guidelines for Canadian Drinking Water Quality [Health Canada 1996].

Education strategies can be used to inform the public when situations of potentially higher risk may exist. Such information may include air pollution advisories or suggestions on limiting consumption of particular types of foods, for example, sport fish [OMEE 1995a]. In general, source control and point-of-use control are the most important strategies in radiological and chemical risk management.

An overview of the responsibilities of the AECB and Health Canada in risk management is provided in Appendix C.

4.1 Ionizing Radiation

4.1.1 Philosophy

Although the usefulness of radiation in medicine and science was recognized soon after the discovery of X-rays in 1895, accumulating reports of harmful effects created a need for basic safety rules. Thus, from the very beginning of the use of these sources, radiation risk-reduction strategies were developed in parallel under the assumption that a balance of the risks and benefits of radiation and radiation-producing technologies was necessary. These strategies have evolved over the last century in light of an increasing knowledge of dose-response characteristics, risks and benefits resulting from radiation practices, and an environment that includes unavoidable natural background radiation. Under normal situations, radiation protection practices are concerned primarily with control at the source.

In Canada, as in most countries of the world, the system of radiological protection is based on the recommendations of the International Commission on Radiological Protection (ICRP). This body was first established in 1928 to focus on safety aspects of medical radiology. Its scope was expanded in 1950 with the widespread use of radiation outside the sphere of medicine. Members of the ICRP and its committees are chosen on the basis of their recognized expertise in the fields of medical radiology, radiation protection, physics, health physics, biology, genetics, biochemistry and biophysics. The risk management philosophy recommended by the ICRP focuses on controlling hazards from nuclear facilities at their source.

Initially, ICRP recommendations were based on the prevention of observable harmful effects, such as skin reddening, among medical radiologists. Tolerance doses were recommended based on the concept of a threshold value for these effects. Late effects were not immediately recognized because of the long latency period between radiation exposure and expression of a cancer.

A major change in radiation protection philosophy occurred in ICRP Publication 2 [ICRP 1960], in which genetic damage was assumed to be the main effect to be prevented. The assumption of a zero threshold for genetic and carcinogenic effects resulted in the basic precept that there should be no man-made exposure without the expectation of benefit. By 1977, continuing observations of radiation effects in the Japanese atomic bomb survivors, including the absence of observable genetic effects, led the ICRP to update its radiation safety recommendations. Publication 26 [ICRP 1977] recognized cancer as the main effect to be avoided. It also recognized that the various tissues and organs of the body have different susceptibilities to radiation-induced cancer. This led to the concept of effective dose, and the recommendation of a maximum effective dose expressed as an annual dose limit that included the sum of external radiation dose and the dose from internally deposited radionuclides [Cember 1996]. Criteria for maximum effective dose were based on quantitative risk estimates and comp ari sons with non-radiological risks considered acceptable by society.

New dose limits for occupational and public exposure were recommended in ICRP Publication 60 [1991], based on continued study of the Japanese bomb survivors. Publication 60 sets forth a comprehensive framework for radiation protection, with the goals of preventing early effects, and controlling the risk of radiation-induced cancers and serious genetic disorders to levels deemed to be acceptable to society. The three basic principles of radiation protection recommended by the ICRP [1977 and 1991] may be summarized as follows: Justification: No practice involving exposures to radiation should be adopted unless it produces sufficient benefit to the exposed individuals or to society to offset the radiation detriment it causes.

Optimization: In relation to any particular source within a practice, the magnitude of individual doses, the number of people exposed, and the likelihood of incurring exposures should all be kept as low as reasonably achievable, economic and social factors being taken into account (ALARA principle). Dose limitation: The exposure of individuals resulting from the combination of all the relevant practices should be subject to dose limits. ICRP dose limits are set such that continued exposure at a dose just above the limit would be unacceptable on any reasonable basis.

The system of dose limitation applies to all exposures arising from all regulated practices. Recommended dose limits do not apply to radiation exposures received by patients in the course of medical diagnosis or treatment, by persons carrying out lifesaving procedures in an emergency, or by the public from natural sources.

Prior to its 1990 recommendations, the ICRP set dose limits for public exposure arising from regulated radiation practices based on an acceptability of fatal risk that was 10-50 times lower than that for occupational risks, depending on whether public exposures were considered for a single year or for a complete lifetime. The ICRP indicated that a risk in the range of one in a million (10^-6) to one in one hundred thousand (10^-5) per year would likely be acceptable to any individual member of the public [ICRP 1977]. Occupational limits were set such that the corresponding radiation risk would be no greater than the risk of accidental death in other industries not associated with radiation, or no greater than one in one thousand (10^-3) per year.

This approach to dose limitation was modified in 1990 to incorporate not only fatal risk, but also non-fatal conditions. In addition, the ICRP felt that it was difficult to assess the acceptability of risk for public exposures in the same manner as for occupational exposures. Therefore, in choosing the new dose limits for the public, the ICRP has taken into account both the concept of acceptable risk, and the variations in the existing level of dose from natural sources. The fact that a man-made radiation practice causes doses which are small in comparison to background does not necessarily imply that the practice is justified, but it does imply that the total radiological risk to the exposed individual is not significantly changed [ICRP 1991]. The annual dose from natural sources, including radon exposure, is about 2 mSv.

On the basis of these judgements, the ICRP recommends limits of 20 mSv per year averaged over 5 years for occupational exposure, and 1 mSv per year for public exposure arising from all regulated radiation practices. The public dose limit is roughly half the average exposure to radiation from natural sources and considerably smaller than the normal variation in exposures to radiation from natural sources.

At an exposure of 1 mSv, the total risk of excess, radiation-induced fatal cancers, weighted non-fatal cancers and hereditary disorders summed over all future generations would be about seven per hundred thousand. The ICRP has indicated that continued exposures at or near the recommended limit for many years is not acceptable.

The use of dose limits is aimed at ensuring that no individual is exposed to radiation risks that are judged to be unacceptable in any normal circumstance. In Canada, design, manufacturing and operating practices have kept the actual maximum exposures of the public well below the legal dose limits set by the AECB. These practices have been interpreted as fulfilling the intent of the ALARA principle [AECB C-129 1994], even though rigorous ALARA processes such as those recommended by the AECB Advisory Committees [AC-2 1991] were not applied.

Consideration of relative costs and health benefits using the ALARA principle has been recommended by the ICRP, the U.S. EPA Science Advisory Board [US EPA SAB 1992], the Joint Committee on Health and Safety of the Royal Society of Canada and the Canadian Academy of Engineering [JCHS 1993], the AECB Advisory Committees [AC-2, 1991], and other organizations including the American Medical Association. The AECB Advisory Committees also believe that all risks arising from a nuclear facility should be included in the ALARA analysis; the AECB, however, may not have the legal authority to control non-radiological risk. In spite of any differences of approach and the difficulties that will be associated with its application, the Advisory Committees believe that the ALARA principle offers the best approach for establishing acceptable risk by balancing risks, costs and benefits.

Finally, the Advisory Committees have recommended that individual exposures of 0.01 mSv per year or less to members of the public from regulated practices could be regarded as a de minimis level, as it carries negligible risk to human health [AC-1 1990, ACNS-20 1995]. Exposures at this level would be regarded as safe, requiring no further mitigative action. Radiation protection groups in the U.S. and other countries have recommended similar values.

4.1.2 Regulatory Control

Laws governing the use of radioactive materials, radiation emitting devices and ionizing radiation exposures exist in Canada at both the federal and provincial levels, and are generally applied at the source. The principal legal instruments at the federal level are the Atomic Energy Control Act¹ and Regulations, and the Radiation Emitting Devices Act and Regulations. The Atomic Energy Control Act regulates, among other things, the use of radioactive materials and fissile material or processes which could be used in a nuclear chain reaction. This Act is administered by the AECB, which has the lead role in the regulation of nuclear facilities and the use of nuclear materials. Discussion of the Act will focus on the nuclear industry. The Radiation Emitting Devices Act, administered by Health Canada, pertains to specific classes of radiation emitting devices used both occupationally (e.g. X-ray equipment, lasers, ultrasound therapy devices) and residentially (e.g. microwave ovens, television receivers). Background radiation from natural sources is not covered under either Act.

Federal Legislation

The Atomic Energy Control Act

Nuclear facilities regulated by the AECB under the Atomic Energy Control Act include power and research reactors, uranium mines, mills and refineries, nuclear fuel fabrication plants, high energy particle accelerators, heavy water plants, and radioactive waste management facilities. The AECB is also responsible for the regulation of radioisotopes and the transport of radioactive materials (together with Transport Canada). The three major stages of licensing for all nuclear facilities are site acceptance, construction approval, and issuance of an operating licence. The applicant is required at each stage to show that its facility can be built and operated without undue risk to workers, the public, and the environment. The AECB monitors the facility and carries out inspections throughout the facility's lifespan to ensure that licensing criteria are met. In addition, the International Atomic Energy Agency inspects the nuclear generating stations of all member states, including Canada, to ensure compliance with the Nuclear Non-Proliferation Treaty. In the specific case of Ontario Hydro, the utility also participates in the Peer Evaluation Program adopted from the Institute of Nuclear Power Op e rators in the U.S. At the end of the useful life of a facility, the AECB must approve all plans for decommissioning.

Limits on both occupational and public exposures are stated in the AECB Regulations and are written into licences issued by the AECB. The current occupational dose limit for radiation workers is 50 mSv per year but the AECB is in the process of adopting the 1990 ICRP recommendation of 20 mSv per year [AECB 1991]. All radiation workers in Canada are required to wear dosimeters to monitor both annual and cumulative radiation exposures. These are currently recorded in the National Dose Registry maintained by Health Canada. According to the Registry, the average occupational exposures at nuclear generating stations in Ontario, for example, were about 0.6 mSv in 1994 [Ontario Hydro Nuclear 1995a, Myers 1996]. This average occupational exposure of 0.6 mSv per year is approximately 30 percent of the average annual exposure of all Canadians to radiation from natural sources (about 2 mSv per year).This occupational exposure corresponds to a total theoretical risk of about 0.12 fatal cancers per year [ICRP 1991a] for the total of all Ontario Hydro radiation workers, with the fatal cancers possibly developing some 20 or so years after the radiation exposure, in contrast to fatal accidents which result in an immediate loss of life.

The National Dose Registry, together with the accompanying mortality database and cancer incidence files maintained by Statistics Canada, provide a wealth of useful information. Epidemiological studies on correlations between recorded radiation doses in the National Dose Registry and causes of death and cancer incidence are currently in progress. Similar registries do not exist for workers exposed to carcinogenic chemicals.

Within the operating licence, the AECB sets annual maximum release limits for radioactive emissions from the facility in order to limit exposures to the general population. Operation at these limits for a full year would result in a maximum estimated dose to individuals in the population group at greatest risk, the critical group, equal to the AECB legal dose limit for public exposure. The nature of the critical group depends on the facility. For example, the critical group might be assumed to reside at the site boundary, and to derive all of their food and water from local sources. Release limits for individual radio-nuclides are derived based on a multi-pathway approach. Currently, the legal dose limit governing all radioactive emissions from nuclear facilities in Canada is 5 mSv per year. The AECB is currently in the process of incorporating the 1990 ICRP recommendations of 1 mSv per year into their regulations.

In practice, the AECB specifies that annual releases from nuclear generating stations must be a small fraction of the maximum annual release limits for each of several groups of radionuclides. The dose to an individual in the critical group from all radioactive emissions must be less than 0.05 mSv per year for each of several groups of radioactive substances, with a total of about 0.3 mSv per year for all radionuclides combined. The actual maximum doses to the most exposed population from nuclear generating stations are about 30 times lower. Such operational limits are first referenced in AECB News Release 73-1 [AECB1973], which mentions "... the intention of the major licensee to take any steps necessary to keep effluents below 1 percent of the license limit";. While justification for the target was based on the ALARA principle, no cost-benefit analysis was undertaken. Rather the target was set by reviewing the operating records for the Pickering-A generating station. According to these records radioactive releases were generally below 1% of the limit, so that 0.05 mSv per year for each group of radionuclides seemed to be readily achievable and was therefore selected as the target. Later, a requirement of 0.05 mSv per year was set as an operating target on emissions, and has since become a de facto limit [AECB 1994b]. The system of licensing requirements which are set at a fraction of the legal dose limit is supported by the most recent recommendations of the ICRP [1991] on dose constraints.

Monitoring to demonstrate compliance with the Act and license conditions is the responsibility of the licensee. Emissions are continually monitored. Should those for a given week or month exceed the specified operating emission levels, examination of procedures and facility design by the licensee is required to determine what actions, if any, are necessary to ensure that the annual emission requirements can be achieved. Independent monitoring is performed by other agencies including provincial ministries and Health Canada. In estimating doses to demonstrate compliance with their license conditions, licensees must consider and sum all potential pathways of exposure to a radioactive material in the environment (e.g. inhalation from air, absorption through skin, ingestion from food or water), based on assumptions that likely over-estimate actual exposures. The dose received from each of the different radioactive materials is summed to obtain the total individual dose or collective dose to members of the public.

In addition to the management of risk associated with the normal operations of nuclear power plants, the AECB requires that the public be adequately protected in the event of a radiological emergency. Nuclear power plants are equipped with special safety systems whose sole function is to prevent or mitigate serious accidents that could result in radiological releases from the plants. The design of these and other safety related systems is based on a defence-in-depth approach. The AECB requires that the performance of these systems during serious accidents be analyzed on a conservative basis during the design of the facility to demonstrate with a high degree of confidence that the resulting doses to the public will be at acceptable levels [AECB 1995d]. These analyses must be updated as required by new information during the life of the facility.

The major licensees are also required to have an effective on-site emergency response plan and coordination with the province, in conformity with provincial requirements. Off-site nuclear emergency plans are a provincial responsibility as outlined, for example, in Ontario's Nuclear Emergency Plan and the Ontario Emergency Plans Act. These are supported by the Federal Nuclear Emergency Response Plan, administered by Health Canada.

Finally, licensees are required to conform to all other relevant federal and provincial regulations with regard to non-radioactive emissions. For example, in Ontario, the main authorities are the Ontario Water Resources Act and Regulations, and the Ontario Environmental Protection Act and Regulations. These Acts do not apply to radioactive emissions from facilities regulated by the AECB; however, emission limits or ambient environmental quality limits for non- radioactive substances under these Acts are included when appropriate in AECB licences, and are therefore under the authority of the AECB.

The Radiation Emitting Devices Act

The Radiation Emitting Devices Act (RED Act) applies to all devices that emit X-rays or non-ionizing radiation in occupational or clinical settings, or in personal use. Regulations written under the RED Act specify minimum safety standards for the design, construction, labelling, and advertisement of the devices or their components. The standards apply to devices at the point-of-sale, and are concerned with the performance of a device with regard to its intended function and manner of operation.

Provincial Legislation

The responsibility for controlling the use of radiation emitting devices belongs to the provinces. Provinces regulate and monitor exposure that may result from radiation emitting devices (but not radioactive material), as well as non-nuclear fuel cycle activities which give rise to occupational exposure to radionuclides. Some provinces, such as Saskatchewan, have prepared their own legislation for the control of ionizing radiation exposure [AECB 1995b].

The provinces set general environmental quality standards for radiation which are not used to regulate emissions from federally-regulated facilities. For example, Ontario Drinking Water Objectives for radionuclides are used to evaluate the acceptability of water supplied to the Ontario public and are legally enforceable on agencies supplying communal water. These drinking water objectives cannot be used to control emissions from federally licensed facilities.

The Ontario Health Protection and Promotion Act gives Medical Officers of Health (MOHs) the authority to close water supplies immediately when public health is threatened. MOHs use Ontario Drinking Water Objectives to evaluate public health risks. The Ontario Ministry of Health under the same Act also controls patient therapeutic exposures to radiation.

The AECB regulates allowable radiation exposures for miners in uranium mines, while the provinces regulate radiation exposures (primarily radon and its progeny) to miners in non-uranium mines (e.g., gold mines). The allowable limits for non-uranium miners in Ontario are roughly one quarter of those allowed for uranium miners under AECB regulations, or approximately one-third of the dose limit recommended by a Canadian federal-provincial committee as an allowable limit for radon in the air in homes [ICRP 1993]. Some other non-nuclear industries have the potential for radiation exposure to workers from naturally-occurring radioactive materials, such as the manufacturing of phosphate fertilizer. Under the provisions of the Ontario Occupational Health and Safety Act, employers are required to prescribe what precautions will be taken to protect workers from harm. In practice, exposures are low.

4.1.3 Public Exposures

Background levels of radiation exposure from natural sources have been extensively documented by various scientific committees, including UNSCEAR, BEIR, and the NCRP. In Canada the average dose from natural background radiation is about 2 mSv per year which includes a population-weighted average for the inhalation dose from radon gas. However, doses from background radiation vary extensively. This is partly due to the wide range of radon levels in homes measured across Canada, which give doses from 0.2 - 3.5 mSv or more per year from exposure to radon and radon progeny [UNSCEAR 1982, NCRP 1987]. There are also some data indicating that some individuals living in northern Canada may receive higher total doses from natural radioactivity due to elevated levels of polo-nium-210 in foods such as caribou meat.

A radon guideline for homeowners was established in 1988 by a federal-provincial working group under the Conference of Deputy Ministers of Health. The guideline recommends that remedial measures be taken where the level of radon in a home is found to exceed 800 Bq/m³ (or about 14 mSv per year) as the annual average concentration in the normal living area. This would be equivalent to a risk of fatal cancer of about 1 in 15 for lifetime exposure at this level, using recent ICRP estimates [ICRP 1993]. As there is some theoretical risk at any level of radon exposure, the guideline suggests that homeowners may wish to reduce levels of radon as low as practicable. The guideline was reviewed and re-adopted in 1995. Interpretation of measurements in homes and advice to homeowners is generally the responsibility of the provinces.

Radiation doses from medical diagnoses and therapy, which are of considerable benefit to patients, may represent a significant source of exposure for the individual and appear to average about 1 mSv per year [ACRP 1996a]. Such exposures when averaged across the population, are less than background, and considerably greater than those from industrial sources, as will be noted later in this report.

On average, radiation exposures of the public from regulated sources represent a minor increase above exposures from natural sources, and are substantially less than the variations in the background dose across the country. The major contributors to the total dose arising from CANDU reactors are radioactive noble gases (krypton and xenon) and iodines emitted into the air, carbon-14 emitted into the air and subsequently incorporated into food, and tritium emitted into air and water. Based on environmental models and actual monitoring data, estimates of maximum annual doses received by members of the public living near Ontario nuclear generating stations in 1994 were on the order of 0.01 mSv [Ontario Hydro 1995]. Because of the cautious nature of the assumptions used in these models, actual doses received will be lower, and those received by populations living further from reactor sites will be substantially less.

Doses to the most exposed individual members of the public near AECB-licensed facilities have been calculated using radionuclide concentrations in various environmental media obtained either directly from monitoring data or from environmental transfer models. In general, estimated maximum annual doses to a hypothetical individual residing near various nuclear fuel cycle facilities, based on conservative environmental transfer models, are in the range of:

• 0.0025 - 0.2 mSv for uranium refining and conversion facilities
• 0 - 0.17 mSv for fuel fabrication plants
• 0.002 - 0.02 mSv for nuclear power plants.

Estimates of calculated maximum doses to members of the public living near uranium refining and fuel fabrication plants [AECB 1995] may be too high, since no measurements of actual maximum doses are available, as they are for nuclear power plants in Ontario. Although dose information is not available for people living in the vicinity of uranium mines, the AECB requires mine operators to impose limits on effluent contaminant concentrations and implement adequate environmental monitoring programs.

Based on ICRP risk coefficients, a theoretical fatal cancer risk can be calculated for the various exposure levels experienced by the public. For example, the hypothetical number of fatal cancers associated with the estimated maximum annual dose from nuclear power plant emissions, 0.01 mSv, is about 1 in two million. As another example, a natural background radiation dose of 2 mSv per year is about 1 in 10,000 per year, or 7 in 1,000 for a 70 year lifetime exposure. This is 2.5% of the total risk of fatal cancer observed in the Canadian population in 1991 and 1992 [Statistics Canada 1993, 1995].

4.1.4 Summary

In summary, the risks associated with ionizing radiation exposure from regulated practices are limited through the system of radiological protection recommended by the ICRP, implemented by the licensee and regulated by the AECB. All regulated practices must produce a net benefit to society, must be optimized with respect to benefits versus risks, and must include a system of individual dose limitation. Dose limits recommended by the ICRP and AECB are viewed as the lower limit of unacceptable levels. They must not be exceeded under normal circumstances, and actual doses should be as low as reasonably achievable, economic and social factors taken into consideration. Public dose limits apply to the sum of all exposures from all regulated practices, and are based on both a level of risk and on variations in natural background radiation. In practice, maximum doses to individuals in the general public from nuclear generating stations in Ontario are about 0.01 mSv per year or about 100 times lower than the recommended legal limit, which in turn, is lower than the variation in background radiation levels across Canada.

4.2 Chemical Hazards

4.2.1 Philosophy

Chemical risk management began with an assumption that public health could be completely protected. This assumption developed in the U.S. early in the century for food additives. By the 1960s, an approach which balances costs and benefits became well established for chemicals, although, in retrospect, it was aimed at reducing risk to levels that would be considered low by almost any criterion. Currently, the presence of natural sources of carcinogens is being given increasing attention in risk management strategies, although for many synthetic chemicals, significant natural sources are absent.

The idea that the dose-response characteristics for some chemicals might have no threshold resulted in a 1958 amendment to the U.S. Food, Drug, and Cosmetics Act prohibiting the addition of any chemical that can cause cancer in humans or animals to the human food supply. Almost immediately, however, it was realized that assuring the complete absence of carcinogens from the food supply was impossible, particularly in view of the rapidly advancing ability to detect ever lower levels of chemicals in food, and the abundance of naturally occurring carcinogens. As a result, the U.S. Food and Drug Administration proposed that if risks calculated under the no-threshold assumption were below some small value, the carcinogen was effectively absent in the food.

The first U.S. proposal for a virtually safe dose was to limit cancer risk to one in one hundred million (10^-8) over a lifetime of exposure [cf. Rodricks et al 1987]. This idea was tied to the notion that if the entire United States population was exposed at or near the virtually safe dose, only one or two of the then-current U.S. population of approximately 150 million would be affected. Shortly thereafter, it was realized that this criterion was an almost impossible burden on regulators for assuring the safety of food additives with considerable benefits. It was then proposed that a lifetime risk of one in a million would be considered negligible by most people. At this level, only about three excess cancer cases per year would result if everyone in the U.S. were exposed.

The one in a million criterion for acceptable risk became institutionalized over the next several years and, when cancer risks from environmental exposures became recognized in the late 1960s and early 1970s, the concept of negligible lifetime risk at one in a million (10^-6) was often applied, predominantly in the U.S. [Kelly and Cardon 1994]. Initially of greatest concern were widespread risks such as exposures to PCBs or pesticide residues in the environment. Later the same risk criterion began to be applied to much less widespread risks such as those which existed in the vicinity of industrial facilities or hazardous waste disposal areas.

Eventually, it became evident that one in a million (10^-6) was a very stringent criterion when relatively few people were exposed [US EPA SAB 1992]. Risks levels at or above one in ten thousand are accepted in setting U.S. Environmental Protection Agency (EPA) Maximum Contaminant Levels for carcinogens in drinking water when further limitation is not technically or economically feasible. In general, however, risk levels above one in ten thousand, even to very few individuals, are viewed as excessive and therefore require action to reduce exposure and risk [US EPA SAB 1992].

The U.S. EPA has set a lifetime risk goal of one per million (10^-6) for the regulation of individual genotoxic chemicals, particularly when the exposed population is large. Prior to the establishment of the Clean Air Act Amendments of 1990, Section 112 of the Clean Air Act required the EPA to set emission standards for hazardous air pollutants "to protect the public health with an ample margin of safety";. This was interpreted to mean that the EPA must first determine a safe emissions level (representing an acceptable degree of risk), and then add a margin of safety in view of uncertainties in scientific knowledge about the pollutant in question. Thus, the EPA adopted a general policy that a lifetime cancer risk of one in ten thousand (10^-4) for the most exposed person may constitute acceptable risk and that the margin of safety should reduce the risk for the greatest possible number of persons to an individual lifetime risk no higher than one in one million (10^-6) [NRC 1994].

Review of relevant decisions by the U.S. EPA and other U.S. government agencies shows that the levels of lifetime risks deemed acceptable for the public by different U.S. agencies under different circumstances vary over a range of ten thousand fold from about one in a million to one in a hundred [Sadowitz and Graham 1994]. The levels of lifetime risk associated with Health Canada guidelines for drinking water vary under different circumstances from about one in ten million for dichlo-romethane to about one in one thousand for arsenic (see Table 3).

The U.S. Office of Management and Budget reviewed the cost of compliance with EPA regulations. The Office found that the cost in millions of 1990 U.S. dollars per potential premature death avoided varied considerably as a result of compliance with EPA regulations [U.S. Office of Management and Budget 1991]. For example, the cost of setting drinking water standards for trichloromethane (chloroform) was about $200,000, while that for disposal of wood preserving chemicals as hazardous waste was about $5.7 trillion (10¹²). The Office concluded that further consideration of the balance between health risks and benefits, in terms of lives lost and lives saved in society, would be appropriate before the promulgation of such regulations. When total societal resources are limited, excessive societal expenditures on reduction of minimal risks, rather than on more severe risks, are expected to be detrimental to societal health. By way of comparison, the AECB has suggested that total expenditures to reduce industrial radiation exposures should not exceed about $2 million 1994 Canadian dollars per fatal cancer avoided [derived from AECB 1994].

Regulatory authorities in Canada do not recommend any single legal dose limit or level of acceptable risk at which to regulate chemical carcinogens. Risk management decisions concerning control are made following consultation with affected parties, and involve judicious balancing of the estimated risks against the associated costs, feasibility of controls, and benefits to society. For example, management strategies pertaining to exposure from Priority Substances found to be toxic under the Canadian Environmental Protection Act (CEPA) vary as a function of different cost/ benefit profiles, based on best-available control technologies economically achievable.

In establishing point-of-consumption standards for chemicals, regulatory agencies generally set generic values based on the risk-reduction potential in terms of costs, benefits, achievability, existing background sources, and societal values. In general, the majority of regulatory controls are for synthetic chemicals. However, in those cases where the contaminants occur naturally, as for example trace metals in drinking water, the extent of exposure from natural sources is often considered in the development of standards or controls (e.g. arsenic in drinking water).

Exposure to a single type of chemical through many pathways has, in the past, often not been taken into account in the development of controls for chemical contaminants; it is, however, being increasingly considered in this regard. For example, for Priority Substances under CEPA, the relative magnitude of the contribution of each pathway of exposure (e.g. air, food, water and consumer products) to total intake is estimated for each of five age groups in the population. These contributions are then considered in the risk management process to ensure that the total estimated risk from all sources is controlled. Probably the greatest barrier to consideration of multimedia exposure in establishing risk management strategies results from jurisdictional constraints. At present, different agencies effectively deal with different media.

In addition to controlling industrial emissions of chemicals into the environment, point-of-source risk management includes controls on occupational exposures. A number of chemicals and industrial processes have been identified as carcinogenic in occupationally exposed workers [Doll and Peto 1981; IARC 1995]. However, cancer risks associated with particular chemicals in the workplace are difficult to assess due to the lack of an adequate database on both chemical hazards and exposure levels. Though it is possible to estimate the risk for some substances (e.g. benzene, arsenic and asbestos) these represent only a small proportion of chemicals commonly used in industry.

The estimated risks associated with occupational exposure limits for different carcinogens differ considerably, due to the weighting of various risk-management factors. Gold et al [1987] compared the legally permissible dose limits for workers in the U.S. to the chronic dose level that induces cancer in 50% of laboratory animals. For 41 chemicals on which reasonable data existed, this ratio differed by more than 100 000 fold. Although this ratio does not take into account actual exposure levels or the number of exposed workers, it nevertheless suggests that more attention should be given in risk reduction strategies for occupational exposures to chemical substances that appear most hazardous to animals.

In Ontario, occupational exposure levels are set by the Ontario Ministry of Labour based both on health studies, and impact studies in terms of cost and benefit. For specific chemical carcinogens, the ALARA approach is used to establish regulatory limits. Actual limits exist only for those chemicals that are in most common usage, or for which health effects are known. However, all reasonable precautions must be taken by an employer to protect the health and safety of workers from all potential exposures.

4.2.2 Regulatory Control

Risk management for chemicals is carried out under a number of different Acts and Regulations, most notably: the Food and Drugs Act and associated Regulations, the Canadian Environmental Protection Act (CEPA), the Pest Control Products Act, the Drinking Water Materials Safety Act, and the Hazardous Products Act. Control mechanisms for chemicals are complex, involving the responsibility of several levels of government. Consequently, the information provided below is not exhaustive, but is intended to highlight some of the major legislation. Some further information is provided in Appendix C.

In assessing and managing chemical risks, Canada interacts with many international organizations such as the International Agency for Research on Cancer (IARC), the Food and Agricultural Organization (FAO), the International Program on Chemical Safety (IPCS), the World Health Organization (WHO), the Organization for Economic and Cooperative Development (OECD), and the Codex Alimentarius Commission. However, in general, there are no international bodies that recommend standard chemical risk management approaches. In North America, the International Joint Commission of Canada and the U.S. serves as an advisory body to both national governments on pollution management in trans-national boundary waters.

Source Control

The management of chemical contaminants in the environment generally involves the use of source controls.

Canadian Environmental Protection Act (CEPA)

The assessment and management of chemical risks is carried out at the federal level under the Canadian Environmental Protection Act (CEPA). The Act sets out broad federal policies on pollution prevention and control, and contains provisions for dealing with toxic substances, nutrients, ocean dumping, environmental research, guidelines and codes of practice, as well as agreements with provinces and territories. In general, controls are implemented where substances are not addressed under other federal legislation and, in particular, where there are international or transboundary implications. For substances deemed to present a risk to health or the environment, controls may be instituted in consultation with the provinces, and polluters can be fined for failure to comply with the regulations. Twenty-five regulations have been enacted under CEPA.

CEPA's mandate covers toxic substances throughout the ecosystem and may control any stage of a product's life cycle. The primary focus of the Act is the prevention of environmental problems before they occur. Preventive measures include regulation and enforcement mechanisms, non-regulatory approaches such as incentives with industry, as well as the development and transfer of pollution measurement and control technologies. Environment Canada and Health Canada develop CEPA regulations and guidelines, and Environment Canada administers the Act on behalf of the federal government.

The Toxic Substances Management Policy (TSMP) is a new federal policy for managing toxic substances. Under the TSMP, any substance that results from human activity, takes a long time to break down, accumulates in biological tissues, and is CEPA toxic or equivalent will be designated as a Track I substance and targeted for virtual elimination. For substances that meet some but not all of these criteria (Track II substances), the objective is to prevent or minimize their release throughout their life cycles (during their manufacture, use, transport and disposal), using pollution prevention approaches [Environment Canada 1995c].

Fisheries Act

The protection of waters frequented by fish is covered under the Fisheries Act, which is the legal responsibility of the federal Department of Fisheries and Oceans. Contained in the Act are provisions related to the implementation of pollution prevention, inspections, enforcement, and civil remedies. Environment Canada is also responsible for the administration of these pollution prevention provisions.

Example of Source Control

In addition to provisions contained under CEPA and the Fisheries Act, the control and regulation of industries producing or using chemicals is subject to provincial regulation, although actual regulatory approaches may vary between provinces and territories. The following describes the situation in Ontario.

The Ontario Ministry of Environment and Energy (OMEE) has the legal mandate under the Ontario Water Resources Act and the Ontario Environmental Protection Act to regulate industrial discharges to water, air, and land which may be harmful to human health, non-human biota, and commercial or private uses of water and air. The powers under these Acts include the requirement for Certificates of Approval prior to construction and operation of industrial facilities. Such certificates include limits on the amount of allowable discharges of chemicals and other harmful agents into the environment.

The Ministry has established standards, guidelines and objectives for assessing and setting emission limits and ambient environmental quality for contaminants in such media as air, surface water, drinking water, soil, and hazardous wastes. Criteria for general environmental quality are established to protect against the most sensitive effects in the most sensitive populations. Regulated criteria, such as for air and hazardous waste, are directly enforceable. Other criteria become legally-enforceable when included in a legal instrument such as a Certificate of Approval or control order.

Under the Ontario Water Resources Act, the discharge of non-radioactive substances from industrial and municipal sources into provincial waterways is controlled using a two-pronged approach. These are the Treatment Technology-Based Effluent Requirements and the Receiving Water-Based Effluent Requirements [OMEE 1994]. Treatment-based water effluent requirements were developed under the Municipal Industrial Strategy for Abatement (MISA) program for several industrial sectors, including the electric power generation sector. The goal of MISA is to protect the environment through the elimination of persistent toxic substances from wastewater discharged into Ontario's waterways. Receiving water-based effluent requirements are developed on a site-specific basis, and are based on OMEE's surface water quality objectives. OMEE policy requires that the more stringent of the two approaches be applied.

Under the MISA program, industrial discharges cannot be lethally toxic to aquatic life before dilution. After dilution in the watercourse, contaminant levels must meet provincial water quality objectives for both toxic and carcinogenic chemicals. Objectives are set to prevent toxic effects in aquatic life at all stages of development. In rare cases where bioaccumulation of a specific contaminant, such as PCBs or dioxins, may occur in species consumed by humans (e.g. sport fish), objectives for that contaminant will be based on human health.

The MISA effluent requirements have been promulgated as regulations and will be legally enforceable in 1998 for the electric power generating sector [OMEE 1995]. The regulations require that industries comply with discharge limits that have been set based on both loading (i.e. kilograms discharged per day) and effluent concentration for specific substances. These limits were developed based on both the results of an effluent monitoring program and the best available technology economically achievable (BATEA) for pollutant reduction. BATEA is defined as the combination of demonstrated treatment technologies and i ndust rial process changes that can reduce or eliminate the discharge of contaminants and that the industry can afford.

Total daily loading and monthly-average loading and concentration limits have been established for several substances for the electric power generation sector. To show compliance with these limits, facilities are required to: establish sampling points for effluent collection; monitor the effluent on a daily and weekly basis; and calculate the loading and concentration values for the two time periods. Furthermore, the facilities must measure the pH of the effluent; conduct acute lethality tests for rainbow trout and Daphnia magna; conduct chronic toxicity testing using fathead minnow (i.e., 7-day growth inhibition test) and Ceriodaphnia dubia (i.e., 7-day reproduction inhibition and survivability test); conduct quality control tests; and determine the volume of the effluent. Finally, the facilities are required to keep records of the data and analytic procedures, and prepare and submit reports to the OMEE. These reports will also be available to the public [OMEE 1995].

The Ministry has not established a generic policy on acceptable risk levels for carcinogens; rather they are evaluated on a case-by-case basis taking into account the scientific information and the implementation implications. Releases of carcinogens in liquid effluent must meet OMEE surface water quality objectives after dilution in the watercourse. Emissions of carcinogens into air must not exceed Point of Impact (POI) standards. These POI standards apply to short-term (30 minutes) releases and are set at a factor of 15 times the annual ambient air quality criteria (AAQC) for the specific contaminant. In setting AAQCs for individual carcinogens, a lifetime risk of ten per million to one per million (10^-5 to 10^-6) is generally used for specific chemicals, in the absence of significant technical and economic limitations, but is case-specific.

As a condition of licensing, facilities are responsible for monitoring and reporting to the responsible authority to ensure compliance with their licensing conditions, and are subject to compliance inspection and monitoring. The regional offices of the OMEE are responsible for working in conjunction with the industry when abatement actions are required to address potential or actual harmful effects. Non-compliance with MISA regulations is a violation and is grounds for enforcement and prosecution. The non-compliance event is documented and investigated further by OMEE. Non-compliance does not automatically lead to prosecution, as abatement activity is also considered to bring about compliance [OMEE 1994]. In situations where serious harm or breaches of conditions have occurred, investigations and legal prosecution are pursued by the Investigations and Prosecutions Branch of the OMEE. For infractions under CEPA, options range from negotiations with the licensee, to prosecution.

Finally, as a requirement for licensing, industries are required to have effective emergency response plans in conformity with provincial requirements. The Major Industrial Accidents Council of Canada, a non-governmental organization, sets standards and guidelines to aid industry in emergency preparedness.

Point-of-Use Control

In general, chemical risk management strategies involving food and drinking water rely heavily on point-of-use controls. Major legislation relating to the control of food additives and contaminants, pesticides and herbicides and drinking water is discussed below. The Guidelines for Canadian Drinking Water Quality are discussed in section 4.4.

Food and Drugs Act and Associated Regulations

Risks from food additives and food contaminants are managed through the Food and Drugs Act and associated Regulations, the Pest Control Products Act, the Fisheries Act, and the Meat Inspection Act. Food additives are chemical substances deliberately added to food with a view to achieving an intended beneficial effect. Food contaminants are chemical substances that are found in food but not deliberately added. They can occur as a result of human activity, industrial or otherwise, or because of their natural occurrence in the environment. The approach to evaluating the risk to humans from deliberately-added or not-deliberately-added chemicals in food is similar, but the management of risk in each of these instances is different.

The evaluation of food additives is based on complete toxicological data supplied by a petitioner before their usage is approved. If a food additive is shown to be a carcinogen in any species, it will not be approved; if previously approved, it would be removed from the Food Additive Tables. In all cases, the use of such additives must be justified, and the minimum level which achieves the desirable effect must be established. If, in taking into account all existing uses of an additive, a proposed additional use does not cause the estimated or probable daily intake to exceed the acceptable daily intake, then that extension of use can be approved and written into the regulations. Monitoring and compliance activities ensure that the level of acceptable risk established at the time of evaluation is not exceeded.

Sodium and potassium salts of nitrite are used to cure meats and fish. Nitrite has several effects on food including colour preservation, flavour enhancement and antioxidant effects; however, the most important function is inhibition of the growth of bacteria, particularly Clostridium botulinum. Growth of C. botulinum results in production of botulinum neurotoxin and leads to foodborne botulism, a potentially fatal intoxication.

While nitrite itself is not carcinogenic, use of nitrite in foods may lead to formation of nitrosamines by reaction of nitrous acid with secondary amines [Kim and Foegeding 1993]. The carcinogenic and mutagenic properties of nitrosamines have been well documented [Lijinsky 1976], and the occurrence of nitrosamines in foods has been demonstrated [Gray and Randall 1979].

The risk of botulism arising from removal of nitrite as a preservative in cured meats has been suggested to be in the same range as that calculated for cancer deaths resulting from its inclusion in foods [Miller 1980]. Any changes in the regulations reducing permitted levels of nitrite in cured meats are likely to affect microbiological stability of the product [Gibson et al 1984]. A reduction in levels of nitrite in foods would require substitution with another suitable preservative to keep the risk of botulism poisoning within acceptable levels.

Unlike food additives, chemical contaminants are usually evaluated after their potential or actual presence in food is recognized. Since no proponent submits toxicological data as in the case of food additives, the necessary data are obtained from published scientific literature. The toxicological database for chemical contaminants is therefore often incomplete. The probable daily intake of the contaminant is then estimated based on the identification of all foods that may contain the contaminant, the intake of those foods by general and target populations, and consideration of other routes of exposure, for example in air or water. If the probable daily intake exceeds the tolerable daily intake derived from the database, various risk management options may be considered. Options include establishing guidelines or legally-binding tolerances for the contaminant, restricting the sale or distribution of foods obtained from the source locality, and recommending or issuing advisory notices about consumption of contaminated foods. Consideration is also given to whether the nutritional benefit of a food outweighs any measure to restrict consumption of a staple in the diet. If the contaminant is proven to be a carcinogen, exposures would be reduced to a level as low as reasonably achievable, social and economic factors being taken into account.

Pest Control Products Act

The Food and Drugs Act and the Pest Control Products (PCP) Act are the relevant federal instruments for the assessment and management of risks from pesticides. Under the PCP Act, safety, merit and value have to be considered in the assessment of potential risks from pesticides. This fundamental principle focuses specifically on the protection of human health and the environment, and product performance. Standard risk assessment procedures as outlined in the Health Protection Branch risk management booklet [HPB 1990] are followed.

The maximum legally permitted residue levels for pesticides which have undergone detailed risk assessment typically range from 1-5 parts per million (ppm). Other pesticides are subject to a maximum residue limit of 0.1 ppm. Actual residue levels of pesticides in food are generally lower than the regulatory limit. Food market-basket surveys indicate that most pesticides are generally not detected.

Monitoring for compliance of pesticide residues in food under the Food and Drugs Act and associated Regulations is conducted by laboratories within Health Canada including the newly formed Canadian Food Inspection Agency.

Drinking Water Material Safety Act

In Canada, it is the responsibility of municipal water authorities to decide how to adapt treatment processes in order to implement provincial-territorial drinking water limits. To assist municipalities - and individuals who rely on private water supplies - the federal Minister of Health introduced the Drinking Water Materials Safety Act in December 1996. The purpose of the Act is to protect the health of Canadians by preventing unsafe drinking water materials from being sold or imported into Canada. The Act would provide for the certification (by accredited third-party certification organizations) of water treatment devices, water treatment additives and water system components to which health-based performance standards have been established. For example, chemical additives such as chlorine-based disinfectants and fluoride would be regulated, as well as materials that come in contact with treated drinking water and household drinking water treatment devices [Bureau of Chemical Hazards 1995; Health Protection Branch 1995a]. In 1996 and 1997, Health Canada held a series of public consultations designed to elicit feedback on this initiative.

Canadian Environmental Protection Act (CEPA)

In addition to a number of source control measures, CEPA includes measures for point-of-use control, including environmental guidelines and codes of practice (further information about CEPA may be found in the "Source Control"; section and in Appendix C).

4.2.3 Public Exposures

Some carcinogenic chemicals exist naturally in the environment; exposure to these substances varies primarily as a function of proximity to the sources. For example, although the level of arsenic in drinking water supplies is generally less than 5 micrograms per litre (µg/L), levels range from 50-500 µg/L in the vicinity of natural sources. The interim maximum acceptable concentration of arsenic in drinking water supplies is set at 25 µg/L, on the basis primarily of achievability at reasonable cost. It has been designated as interim to be reviewed periodically in light of developments of treatment technology and additional data on health risks primarily due to the high estimated lifetime skin cancer risk of one in one thousand at this level.

Ames et al [1990] have pointed out that natural sources of carcinogens can be significant. They have suggested that the public is ingesting, on average, about 10,000 times more pesticides from natural sources than from industrial sources, and have argued that the carcinogenic risks of these may be greater than those of synthetic pesticide residues in food.

Natural carcinogenic pesticides and chemicals found in fruit and vegetables include methoxypsoralen, limonene, caffeic acid, and aflatoxin. A recent review by the U.S. National Research Council [1996] supported the hypothesis that risks from natural carcinogens found in the food supply may outweigh the risks from synthetic chemical contaminants, although additional research was called for to establish support for this conclusion. The potential health effects associated with ingested food contaminants does not imply that individuals should avoid certain foods. Rather, it is known that natural foods also contain protective factors which tend to decrease the carcinogenic effects of chemicals from natural and industrial sources [Doll 1992].

In general, exposure levels to synthetic chemicals are well below regulatory standards and guidelines. Doll and Peto [1981] have calculated that about 80% of all cancer deaths in North America are due to factors such as dietary habits, cigarette smoking, infections, and reproductive or sexual behaviour (Table 2). Although the values in Table 2 are uncertain, industrial products and food additives together are believed to account for less than 2% of all fatal cancers in the general public. The corresponding values calculated by Travis et al [1991] and by Gough [1990] are in the range of 0.25% to 2%. In general, the values calculated by Doll and Peto [1981], listed in Table 2, have stood up remarkably well in light of more recent scientific reviews of available evidence [Krewski 1987, Hen-derson et al 1991, Ames et al 1995, Trichopoulus et al 1996, Willett et al 1996]. Miller [1992] has suggested that the proportion of cancers attributed to occupation might be underestimated by a factor of about 2.

It is important to recognize that although there has been no evidence of cancer associated with exposure to chemical contaminants in the general environment, there have been serious non-cancer effects resulting from releases of pollutants to the environment. Studies have linked hospitalizations due to respiratory illness to summertime concentrations of ozone and particulate matter as well as elevated ambient levels of carbon monoxide and the coefficient of haze in regions across Canada [Burnett et al 1995; Thurston et al 1994; Burnett et al 1996; Stieb et al 1996; Delfino et al, 1996]. In addition, particulate matter and carbon monoxide have been linked to cardiac disease and cardiovascular mortality [Burnett et al 1996; Ozkanyak et al 1995]. Based on the results from these studies and a number of similar investigations conducted worldwide, it now appears clear that more adverse cardio-respiratory health events occur on days when ambient air pollution is elevated. However, none of these studies has been able to demonstrate a statistically significant association between ambient concentrations of ozone and either deaths or hospitalizations for cardiac diseases.

Table 2. Estimated Proportion of U.S. Cancer Deaths Attributed to Various Factors

Factor		Best Estimate of Percent of all Cancer Deaths
Diet (including fat intake, meat intake, obesity)		35
Table 2 derived from Table 20 in Doll and Peto (1981) * Data on the ultraviolet component of sunlight are derived from the text in Doll and Peto (1981); data on the ionizing radiation component were increased on the basis of recent ICRP 1991-1993 risk estimates and recent data on background exposures to radiation from natural sources (UNSCEAR 1993).
Tobacco (primarily cigarette smoking)		30
Infection (including certain viruses)		~10
Reproductive and sexual behaviour(including number of sexual partners, number of children)		7
Occupation		4
Alcohol		3
Geophysical factors:	ultraviolet light	1 - 2*
	ionizing radiation	2.5*
Pollution (including combustion products in air, chlorinated water supplies)		2
Medicine and medical procedures		1
Industrial products		<1
Food additives		<1

Considerable efforts have been made in Western countries to reduce pollution from the combustion of fossil fuels. Several countries, including Canada and the United States, established new stringent guidelines and standards for air pollutants such as sulphur dioxide, nitrogen dioxide, carbon dioxide, ozone, and particulate matter. In Canada, the National Ambient Air Quality Objectives [Environment Canada 1994] for these pollutants are rarely exceeded.

4.2.4 Summary

In summary, the risks associated with chemical hazards are controlled primarily through implementation of various federal and provincial regulations and standards. Source controls are used primarily for environmental contaminants, while point-of-use standards are used for contamination in food and drinking water. Point-of-use controls are also implemented through, for example, the Guidelines for Drinking Water Quality and the

Food and Drugs Act. Regulation of industries producing or using chemicals is the responsibility of each province. Criteria and standards governing industrial discharges to water are based on the prevention of toxic effects in aquatic life; emissions to air are based on the prevention of harmful effects in humans and vegetation. Limits for releases of carcinogens into the environment are established on a case-specific basis. In Ontario, short-term (30 minutes) atmospheric emissions of carcinogens must not exceed by more than a factor of 15 the ambient air quality criteria for specific contaminants, which are generally based on a lifetime cancer risk of one in a hundred thousand (10^-5) to one in a million (10^-6). Regulations and standards are legally enforceable by the province.

4.3 Microbiological Hazards

In Canada, regulations for microbiological contamination in food are established under the Food and Drugs Act. Unlike chemical or radiological hazards, microbiological hazards are highly sensitive to environmental conditions such as temperature changes. Thus, they may increase or decrease in municipal water supplies, or during production, processing, storage, retailing, and home preparation of foods. Therefore, risk management strategies do not usually aim at achieving a defined level of risk. Instead, risks are reduced by first subjecting water or food to treatments to bring about a specific number of ten-fold reductions of specific target organisms, and then preventing recontamination by, for example, the use of appropriate packaging, limiting the growth of organisms by such means as refrigeration, dehydration or curing. Examples of treated food include pasteurized milk, canned goods, and refrigerated foods.

Point-of-consumption approaches are used in the management of microbiological hazards, and are carried out at the both the federal and provincial levels. These strategies have been developed on an ad hoc basis, and have not yet been derived from microbiological risk assessments. However, it is expected that the basic control procedures for pathogens will apply. In addition, microbiological hazards are managed by surveillance of human infections and disease, monitoring of microbiological pathogens and their indicator organisms in the environment (i.e. food, water, soil, feed, farm animals), voluntary guidelines for industry, training of employees, and education of both those responsible for manufacturing or selling safe products, and the general population.

4.4 Drinking Water

The Guidelines for Canadian Drinking Water Quality are described in this report as an example of how radiological, chemical, and microbiological risk assessment and management practices are combined within a flexible risk reduction strategy. In Canada, the quality of drinking water is primarily the responsibility of the provinces and municipalities. Health Canada works in collaboration with provincial health and environment ministries to establish national guidelines for drinking water quality under the auspices of the Federal-Provincial-Territorial Committee on Environmental and Occupational Health. The Guidelines are intended to facilitate the delivery of high quality drinking water to Canadians [Health Canada 1995, Krewski et al 1996].

Development of the drinking water guidelines is a flexible process designed to accommodate the needs of the various jurisdictions involved. The steps of risk assessment and risk management are clearly delineated in the development of Canadian Drinking Water Guidelines, with Health Canada recommendations for genotoxic carcinogens being as low as possible. Maximum acceptable concentrations for these compounds are then established by a Federal-Provincial Subcommittee, taking into account feasibility and costs. These include identification of substances to be reviewed, assessment, evaluation, decision making and approval, announcement and publication of decisions, and re-evaluation of findings as required. Certain steps may be modified in order to satisfy the needs of the jurisdictions involved. Through this consensus-based development process, a guideline is established, and the associated health risk assessment is modified to create a criteria summary that reflects the risk management decisions involved in the guideline's development.

Although not mandatory, the Guidelines may be used by the provinces and territories as a basis for setting maximum permissible levels for radionuclides, chemicals, and microbio-logical hazards. Provinces may adopt the Guidelines in whole or in part, or may establish their own criteria.

Radionuclides

Guidelines for radionuclides in drinking water conform to international radiation protection methodologies, including recommendations of the World Health Organization [WHO 1993]. As a result of the method of dose limitation recommended by the WHO, the levels of risk associated with the guideline dose, although low, are somewhat higher than the basic criteria for most individual chemical carcinogens in water. However, the guideline dose for radionuclides applies to the total dose received from all radionuclides in the water supply.

Maximum acceptable concentrations (MAC) for radionu-clides in drinking water are based on a committed effective dose of 0.1 mSv from one year's consumption of drinking water, consumed at the rate of two litres per day, or one-tenth of the ICRP's recommendation on total public exposure from regulated sources. The guideline reference dose is based on the total activity in a water sample, whether the radionuclides appear singly or in combination, and includes the dose due to both natural and artificial radionuclides. Individual MACs therefore apply only in the event that a single radionuclide is found in the water supply. If multiple radionuclides are detected, the dose received from all radionuclides should not exceed the guideline dose of 0.1 mSv per year. The guideline reference dose corresponds to a lifetime risk of fatal and we ighted non- fatal cancer of about four in ten thousand.

Because radionuclide guidelines are based on a reference dose, rather than actual concentrations in water, MACs are orders of magnitude greater than concentrations currently observed in, for example, the Great Lakes [Ahier and Tracy 1995]. The estimated average annual dose from drinking water from all radionuclides in the Great Lakes is on the order of 0.001 mSv per year, which corresponds to a 70-year lifetime risk of fatal and weighted non-fatal cancer of about four per million. This dose represents about 1% of the Health Canada guideline for radionuclides in drinking water or about 0.05% of the average annual dose attributable to radiation from natural sources. Specific cases where doses are higher occur as a result of natural radionuclides, such as radium-226, in ground and well water.

Water supplies which would result in a total radiation dose below the reference level are considered acceptable for consumption based on radiological considerations. However, treatment of water supplies for radionuclides should be governed by the ALARA principle of keeping exposures as low as reasonably achievable, economic and social factors taken into consideration, and levels may be reduced further if justified. In cases where a single sample does not meet the guideline, the reference dose would be exceeded only if exposure to the same measured concentration were continued for a full year. Hence, such a sample does not in itself imply that the water is unsuitable for consumption, and should be regarded only as a level at which further investigation, including additional sampling, is needed. Guidelines do not constitute an approval to permit the increase of radionuclide concentrations to the MAC; any facility contributing radionuclides to a drinking water source must meet the regulatory requirements of the AECB.

As noted in the introduction, the Ontario Ministry of Environment and Energy has established an interim objective of 7000 Bq/L for tritium in drinking water. Ontario Hydro Nuclear Plants, the only significant industrial source of tritium in Ontario, have agreed to keep average annual concentrations of tritium in drinking water at nearby pumping stations to less than 100 Bq/L. It might be noted that the average 1994 concentration of tritium in drinking water at the Ajax pumping station was 15 Bq/l, and about half of this concentration was due to the nearby Ontario Hydro Pickering Nuclear Generating Station with the remainder due to residual fallout from nuclear weapons testing and to natural sources. Tritium in drinking water accounted for about one percent of the total industrial radiation dose received by all persons living within 30 km of the Pickering station in 1994 [Ontario Hydro Nuclear 1995].

Chemicals

Different approaches to guidelines are adopted for carcinogenic versus non-carcinogenic chemical contaminants. In the case of non-carcinogenic chemicals, it is generally assumed that the dose-response relationship demonstrates a threshold below which no adverse health effects are observed. For carcinogenic chemicals, it is generally assumed that carcinogene-sis is a non-threshold phenomenon. Consequently, carcinogenic chemicals should ideally be absent from drinking water. However, the incremental risks associated with exposure to low levels of these chemicals in drinking water may be sufficiently small so as to be essentially negligible compared with other risks commonly encountered in society.

Maximum acceptable concentrations (MACs) for substances not known to be carcinogenic are based on a tolerable daily intake (TDI) for organ-specific neurological/behavioural, reproductive or teratological effects. Where possible, the TDI is derived by dividing the lowest no-observed-adverse-effect level (NOAEL) obtained from long-term ingestion studies by an uncertainty factor. Uncertainty factors are derived on a case-by-case basis; in general, however, a factor of 1 to 10 times is used to account for various elements of uncertainty including intra-species variation, inter-species variation, nature and severity of the effect, adequacy of the study and the use of a lowest-observed-adverse-effect-level (LOAEL) versus a NOAEL. An additional factor of 1 to 5 times can be incorporated where there is information indicating a potential for interaction with other chemicals. If the chemical in question is an essential nutrient at low concentrations, dietary requirements can also be taken into consideration in deriving an uncertainty factor. Finally, in certain cases, an additional factor of 1 to 10 times can be applied to account for limited evidence of carcinogenicity.

Where appropriate, the MAC is based on intake in the most sensitive subpopulation (e.g. pregnant women, children). When appropriate data exist on other sources of exposure (e.g. air, food, soil), a proportion of the TDI can be allocated to drinking water in the calculation of the MAC. In the absence of such data, a default allocation of 20% is used. When a MAC is less than levels considered to be reliably measurable or achievable, an interim MAC (IMAC) is established, and improvements in methods of measurement or treatment are recommended.

MACs are set as close to zero as reasonably practical, on the basis of consideration of the following factors:

• the MAC must be achievable by available water treatment methods at reasonable cost;
• where possible, the upper 95% confidence limit for the lifetime cancer risk associated with the MAC is less than ten in a million to one in a million (10^-5to 10^-6), a range that is generally considered to be essentially negligible. In cases where intake from sources other than drinking water is significant, the upper 95 % confidence limit for the lifetime cancer risk associated with the MAC is less than or equal to one in a million (10^-6);
• the MAC must be reliably measurable by available analytical methods.

The guidelines recommend that where estimated lifetime cancer risks associated with the MAC are greater than one in a hundred thousand (10^-5) to one in a million (10^-6), an interim MAC be established, and improvements be made in methods of measurement and/or treatment.

The levels of lifetime risk associated with MAC levels of various chemical carcinogens vary under different circumstances from about one in ten million for dichloromethane in drinking water to about one in one thousand for arsenic in drinking water (Table 3. In general, guidelines exceeding risks of one-in-a-hundred thousand (10^-5) to one-in-a-million (10^-6) are associated only with natural chemical contaminants that may have a significant background, such as arsenic. In cases where actual exposure levels approach or exceed the guideline, usually only a small population is exposed.

Microbiological Hazards

Microbiological considerations are based on disease-causing or pathogenic microorganisms that commonly occur in polluted water. Pathogenic microorganisms present in surface water include certain protozoa, bacteria and viruses; protozoa are not commonly found in ground water. The most common illnesses attributable to waterborne pathogenic microorganisms are gastrointestinal illness and diarrhea, although more serious health effects may occur, including death. For some waterborne pathogenic microorganisms, notably the protozoan Giardia lamblia or hepatitis A virus, one infectious unit of virus or a single protozoan can cause illness.

While the desired goal in terms of public health protection is zero risk of illness from waterborne pathogens, it is rarely technically and economically feasible. Instead, acceptable microbial risks are derived and used in risk assessment. The MAC for total coliform bacteria in drinking water is zero organisms per standard sample, although some samples may contain very low numbers to account for the non-uniform distribution of coliforms in water. No consecutive samples from the same site should contain coliform organisms. The Federal-Provincial Subcommittee on Drinking Water is considering the U.S. Surface Water Treatment Rule which has set a risk of one infection per 10,000 people per year as a health goal for exposure to one particular type of protozoa in treated drinking water. Thus most public water systems must disinfect as well as provide filtration. Using this approach, treatment must achieve at least 99.9% removal or inactivation, or both, of Giardia, and 99.99% removal or inactivation, or both, of viruses.

Disinfection and other treatment methods are recommended to prevent waterborne diseases and ensure good quality drinking water. The Federal-Provincial Subcommittee on Drinking Water emphasizes that the health risk, such as carcinogenicity, associated with the use of disinfectants must also be considered.

Summary

In summary, exposure limits based on lifetime risk for most individual chemical carcinogens are more stringent than corresponding limits for all radionuclides combined. With the exception of arsenic, the lifetime cancer risks associated with the MACs for individual chemical carcinogens in drinking water are significantly lower than the risk for all radionuclides (Table 3). However, according to internationally accepted practice, the total dose from all radionuclides is evaluated and compared with the guideline reference dose. No attempt is made to evaluate the potential risk of all chemical contaminants combined because of the large number of chemical contaminants that may be present in drinking water, not all of which have been assessed. Unlike radionuclides, background exposures are generally not considered in establishing chemical exposure guidelines. In addition, many of these act by different mechanisms leading to the development of different types of tumours. Information on the potential interactions between chemical contaminants in drinking water, or between radio-nuclides and chemicals [UNSCEAR 1982], is also generally lacking.

Harmonizing drinking water guidelines for chemicals and radionuclides would require consideration of a number of fundamental risk assessment and risk management issues. Guidelines for radionuclides are set to protect public health at a given risk level; actual exposures are generally much lower. Chemical exposure guidelines are set at the lowest achievable level that is both protective of human health and cost effective. To achieve harmonization, there are technical, regulatory and jurisdictional issues that would have to be resolved, as well as the basic question of whether harmonization would result in public health benefits. Future discussions on harmonization should take place in a broader context in which all relevant public health concerns are addressed. For example, in addition to chemicals and radionuclides in drinking water, the impact of microbial agents on public health needs to be assessed.

Table 3. Estimated Lifetime Cancer Risks from Selected Carcinogens in Canadian Drinking Water^*

Agent	Risk per Million People based on Continuous Exposure at Maximum or Interim Maximum Acceptable Concentrations
* As derived from Guidelines for Canadian Drinking Water Quality, Supporting Documentation, 1995 Revision, developed jointly by Health Canada and the provinces. ** Because radionuclide guidelines are based on a reference dose, rather than actual concentrations in water, Maximum Acceptable Concentrations are orders of magnitude greater than concentrations currently observed in, for example, the Great Lakes.
Arsenic	890
Benzene	3.1 - 34
Benzo(a)pyrene	0.5
Carbon tetrachloride	1.7 - 5.2
1,2-Dichloroethane	8
1,4-Dichlorobenzene	0.6 - 2.2
Dichloromethane	0.085
2,4,6-Trichlorophenol	2.2
Trihalomethanes (chloroform)	3.6
Vinyl chloride	10
All radioactive materials combined in drinking water**	400