Goal of EHRA
The aim of EHRA is to identify the health outcome and the exposure level that will lead to this outcome. Then, based on these two factors, we decide on a relatively "safe" exposure level or "acceptable" exposure level. The word safe here means that excess health risks will not accrue, and acceptable exposure level means that even if we know that an exposure is not safe and will result in health effects or adverse outcomes, and we cannot prevent it, we need to decide how much of it we can live with. You can see that with this approach, some people will like it and some people will resent it. Some would say that we should not allow any level of health harm to anyone, and therefore, this exercise assumes that healthcare authorities are OK with the idea that some people are exposed to harm; other people would claim that if this exercise results in too conservative estimates of exposure, then because of precautionary principles, people have to be restricted in how much exposure they are allowed but there is no real data to believe that such exposure will actually cause harm. For example, if emission standards from household wood burners are kept at very low levels that are allowed and only certain classes of wood burners are allowed even with no real data to support that such low levels of emissions will actually lead to lower health risks, then some people may get upset that they will have to invest in high cost technologies even when somewhat lower cost conventional equipments could be used.
Safety Assessment versus Risk Assessment
Depending on the health effect, an environmental health risk assessment can be a safety assessment or a risk assessment. When the health effect is one that does not cause cancer but something like heart disease or asthma, then EHRA essentially defaults to a safety assessment. In that form of assessment, EHRA exercise results in characterising a level of risk and a level of maximum allowable exposure so that no additional disease over and above a certain baseline results. The assessor accepts a certain level of exposure. On the other hand, when the health effect is one of cancer, then no theoretical upper limit of accepted level of exposure is possible, as cancer is a multi-step disease. At some point during the development of cancer, a molecular mechanism triggers the cancerous process and as this happens at molecular levels, therefore a theoretical low acceptable limit is impossible to define. Therefore in cancer, we usually do not define a threshold level but instead we define a level of exposure where certain amount of excess cancer cases accrue.
Steps of EHRA
EHRA consist of four steps: hazard identification, dose response assessment, exposure assessment and risk characterisation (Figure 1) Figure 1. EHRA four step process
Hazard Identification
Hazard identification is based on literature review and in hazard identification, we identify what health effects will result for humans from a given exposure. For example, if we want to conduct a hazard identification for something like a wood burner in the house, we would have to review what health effects will result from burning wood inside a house would result. This could be based on reviews of outdoor air pollution and associated health effects in winter, or health effects of burning wood inside the house and indoor air pollution that results from this activity and associated health effects. As this is a desktop exercise, we conduct hazard identification by conducting a review of the toxicological, and epidemiological literature. A toxicological literature review basically means that we review animal experiment data and data from studies on human tissue based samples and studies on health effects. Epidemiological literature are based on human health outcomes. When we conduct such literature review, we must pay attention to the following features:
- The quality of the review
- The specific exposure or the mixture of the exposure variables
- The health outcome studied
- If an animal model is studied, how does that animal model relate to the humans?
- What are the toxicological mechanisms?
- What are the potential human routes of exposure? Some exposures are through inhalation; others are through ingestion; how are they comparable with human studies?
- Is it possible to extrapolate quantitative dose response studies?
According to the California Environmental Protection Agency, occupational epidemiological studies are limited because the studies are conducted for healthy adults, and excludes children and older adults who are outside of workforce age \cite{Agency:2010rm}. Animal studies are of limited use too unless the animal models are similar to humans or unless animal diseases or health states can be replicated in humans. Epidemiological studies also suffer from limited range of exposure and limited number of people on whom these studies are conducted. Therefore, while no study is foolproof, we need to use caution and a checklist to check the limitations and usefulness of specific types of studies. What health effects shall we study? In general, we should focus on one particular health effect, even if a number of different health effects are listed. So, for example, when we are studying indoor radon gas, a gas often given off in places where uranium mines are located and found in basements of the houses, we should look for one particular health effect -- generally the one that occurs that occurs frequently or something that is significant. The word significant can be misleading here as what is significant to one group of people may not be significant to the others. In general, cancers such as lung cancer in case of radon gas exposure would be considered as an important health outcome to focus on. Another way to look for health effect is to focus on what is the most sensitive health effect and occurs at low doses of exposure or critical, a health effect that is so important that it must be addressed as it affects a large number of people who live in the environment where such exposures occur.
Dose-Response Analysis
After we have completed the hazard identification process, we will next conduct two steps in parallel. One of the two steps is to conduct a dose-response analysis. Depending on the nature of the health outcome, we will have different approaches to conducting the dose-response analysis. If I am studying a non-cancer health effect, then my dose-response assessment will be something like a safety assessment where I will identify if there is a "safe" level of exposure or dose for which there is no additional worse effect, or very low levels of adverse health effects. On the other hand, if my health outcome is a cancer, then it is no more a matter of "acceptable" risk. At all levels of exposure, we should strive to bring down the risk of cancer as we do not know at what molecular level the induction of tumour takes place. Hence, a "threshold" model will not work for us.
Difference between exposure and dose
We need to distinguish between exposure and dose. We know that we are exposed to the toxicants in the environment through various pathways. We can inhale a toxin (such as a toxic gas), or we can ingest a toxin in two ways - we can drink a toxin dissolved in a liquid (such as tea or another beverage), or water (dissolved in water as microbes and chemicals such as arsenic); we can also be exposed by eating (toxins that are dissolved or in the form of food; food poisoning bacteria, or heavy metals, or children often put soil in their mouth, or pica, for example). Once the toxin enters the body, then it does something to the body (toxicokinetics) and the body does something to it (toxicodynamics). The toxin can reach liver carried by the blood stream where it is metabolised. After metabolism in the liver, the metabolites are carried by the blood-stream to the kidneys where they are excreted through urine. Exposure refers to the concentration of toxin in the environment and dose refers to the amount or concentration of the toxin at the site in the body where it is going to act in the form in which it is going to act. For example, inorgnanic arsenic dissolved in drinking water is associated with cancer and non-cancer effects; inorganic arsenic exposure is associated with bladder cancer \cite{Tsuji:2014cz}. But it is not inorganic arsenic that is associated with such effects, they are caused arsenic metabolite, notably MMA3 \cite{Melak:2014hl}. For those exposures that are not associated with cancers (non-carcinogens), EHRA is essentially a safety assessment. Here, the aim of the EHRA process is to identify the lowest level of exposure that will result in an acceptable level of health risk. In order to do this, for the exposure (chemicals or other agents), the level of exposure at which health effects manifest are studied. Some health effects are significant and serious; others are non-serious. Some health effects manifest at very low levels of exposure: these health effects are sensitive health effects. Some of the health effects are very sensitive; the most sensitive health effect such as skin lesions may or may not be serious. On the other hand, there may be serious health effects such as cardiovascular effects that are manifest at doses higher than that lead to the manifestation of the highly sensitive yet trivial or non-serious health effects. These effects can be critical health effects. The critical health effects are identified from the hazard identification steps and then these significant or critical health effects are studied with respect to the dosage in which they manifest. The lowest dose at which the critical yet non-cancers health effects that manifest and the ones that are not beyond their baseline risks is referred to as no observed adverse effects level (NOAEL). For some combinations of exposure and disease conditions, a NOAEL is not possible. There are other situations, where the health effects are manifest but these manifest at certain levels of exposure. These are referred to as lowest observed adverse effect level (LOAEL). The NOAEL and LOAEL are threshold levels. Thus, for non-cancerous health effects, the focus is on finding the apparent safe levels of exposure rather than finding out conducting a risk assessment. With cancerous health effects, the aim is to eliminate risks at all levels. In risk assessments that involve cancerous health effects, the dose response graph is studied and the specific levels of the dosage in which the effects are manifested are charted. As many dose response charts are studied in the contexts of animal studies where very high doses are involved, these doses are unsustainable or impractical in human beings. As a result, most dose-response studies in humans for carcinogens are studied using extrapolation from large dose studies to low or ultra low dose studies. The goal of these extrapolation exercises is to see what would be the result in terms of health effects if the dosage is at the level of human scale (this is in the form of mg or micrograms/Kg of body weight or for inhalation agents, on a scale of mg or weight per L of air). Because of this, dose response curves for cancerous health effects are studied assuming that there is a linear dose response function between the exposure or dose and the health effect (see Figure 2 to view a linear dose response association). Figure 2. Linear dose-response association Another thing about dose-response assessment for carcinogens is that, there is no concept of a threshold dose limit. This is because cancerous processes are thought to start at a molecular level and therefore in theory, there cannot be a lower limit of the concentration or amount of an agent that can lead to the develop of cancer. For many chemicals such as chloroform and arsenic can have both carcinogenic and non-carcinogenic effects. Therefore, there is an increasing trend to identify non-carcinogenic effects that can in turn lead to cancers, for example change in methylation patterns, or formation of DNA adducts. Also, in a bid to bring together the concepts of non-cancerous and cancerous effects into a common framework, a concept of using reference dose or reference concentrations has emerged.
Exposure Assessment
Exposure assessment is the other step that happens in parallel with dose-response assessment. Exposure assessment can be done either through surveys, monitoring, or statistical modelling. Surveys and monitoring are two tools where direct measurements as to the exposure are conducted. The limitations of this approach is that, these may not be accurate and may result in under assessment. Surveys and monitoring are also expensive. An alternative is to use statistical modelling of the exposure. This requires that the person who is conducting the statistical modelling must know the source and dispersion of the chemical or the exposure variable through the media through which it spreads. The process is relatively inexpensive and can be used to develop possible exposure scenario. One can use this to develop high exposure scenarios that are preventive. Dose response analysis is a step where we establish what would be a safe dose from assessments of dose versus health outcomes. This differs between cancer health effects and non-cancer health effects. In case of cancers, there is no upper theoretical level, and dose responses are usually read off from either animal studies or epidemiological studies assuming a linear statistical model. For non-cancer health effects, we usually identify a threshold level at which the lowest possible health effects or least possible health effects at or above the expected frequencies. Such levels are referred to as no observed adverse exposure level or low observed adverse exposure levels. In exposure assessment, we measure the amount of the environmental toxin to which a person is exposed.
Risk Characterisation
In the step of risk characterisation, we bring together the concepts of exposure assessment and dose response assessment and characterise for the most critical or sensitive or significant outcome, estimate the risk for each level of the dose that we estimated for the level of exposure. Risk characterisation is a step where information from the exposure assesment and dose response assesment or analysis are combined to determined a reference dose category. Again, with respect to cancer and non-cancer health effects, these estimates are somewhat different. In case of cancer health effects, we plug in the exposure value and read off from the dose response curve to estimate what is referred to as the excess number of cases for a given exposure. This excess is given for a reference dosage. For cancer health effects, we target or express an acceptable standard of about 1 per million excess cases. Depending on the technological capability, sometimes teh acceptable risk is set at 1 per 100, 000 or 1 per 10, 000 person-years of exposure. For non-cancerous effetcts, we use hte concept of reference dose, that is, a dose at which animals develop the disease or health conditions, and then we further divide that dose by 10, 000 or 100, 000 to estimate the acceptable human dosage. The results of the environmental health risk assessment guide policy setting for controlling of environmental hazards and taking public health action. These data are then used for setting policies or setting action agenda for environmental health risk management. In the following section, we will conduct an EHRA exercise to illustrate these points.