Risk Assessment of LMOs - Training Manual

Module 3: Conducting the Risk Assessment

---

Introduction

Risk assessments are intended to calculate or estimate the risk to a given target organism, system, or (sub)population, including the identification of uncertainties, following exposure to a particular agent, taking into account the inherent characteristics of the agent of concern as well as the characteristics of the specific target system (WHO, 2004). In the context of biosafety, risk assessment can be defined as the process of estimating risks that may be associated with an LMO on the basis of what adverse effects may be caused, how likely the adverse effects are to occur, and the consequences should they occur.

The risk assessment process involves a critical review of available data for the purpose of identifying and possibly quantifying the risks resulting from, for example, natural events (flooding, extreme weather events, etc.), technology, agricultural practices, processes, products, agents (chemical, biological, radiological, etc.) and any activity that may pose threats to ecosystems, animals and/or people.

The objective of a risk assessment under the Cartagena Protocol “is to identify and evaluate the potential adverse effects of living modified organisms on the conservation and sustainable use of biological diversity in the likely potential receiving environment, taking also into account risks to human health” (Annex III).

The results of risk assessments of living modified organisms (LMOs) are typically used by decision-makers to make informed decisions regarding the approval, with or without conditions (e.g. requirements for risk management and monitoring strategies), or prohibition of a certain use of the LMO. This module provides an introduction to risk assessment and considerations that may assist risk assessors in conducting risk assessments of LMOs that are consistent with Article 15 and Annex III of the Protocol ¹.

Overview of the risk assessment methodology

In order to understand what is meant by risk assessment it is important to be familiar with the concepts of risk and hazard, and how these terms differ. The term “risk” does not have a single unambiguous definition but it is often defined as “the probability of harm”. This is broadly understood as the likelihood that a harmful consequence will occur as the result of an action or condition.

	Figure 4 – Assessing risks Source:// http://www.scienceinthebox.com/en_UK/safety/riskassessment_en.html

Risk is often assessed through the combined evaluation of hazard and exposure.

• “Hazard”, in the context of LMO risk assessment, is defined as the potential of an organism to cause harm to human health and/or the environment (UNEP, 1995).
• “Exposure” means the contact between a hazard and a receptor. Contact takes place at an exposure surface over an exposure period (WHO, 2004). In the risk assessment of LMOs, “exposure” can be understood as the route and level of contact between the likely potential receiving environment and the LMO or its products.

The exposure pathway from the hazard to the receptor and the possible exposure scenarios ² form important additional elements in understanding risk. Ascribing the probability and consequences of exposure of a receptor to the hazard characterizes the risk. All these elements must be evaluated to form an effective and useful risk assessment for specific scenarios (UNEP Division of Technology, Industry and Economics).

A simple example can be used to distinguish hazard from risk: acids may be corrosive or irritant (i.e. a hazard) to human beings. The same acid is a risk to human health only if humans are exposed to it without protection. Thus, the degree of harm caused by the exposure will depend on the specific exposure scenario. If a human only comes into contact with the acid after it has been heavily diluted, the risk of harm will be minimal but the hazardous property of the chemical will remain unchanged (EEA, 1998)..

	Example 11 – What is risk? What is risk assessment? Risk: the combination of the magnitude of the consequences of a hazard, if it occurs, and the likelihood that the consequences will occur. Risk assessment: the measures to estimate what harm might be caused, how likely it would be to occur and the scale of the estimated damage. Source: UNEP (1995).

Risk assessment of LMOs can be divided into four main phases (WHO, 2004):

1. Hazard identification – The identification of the type and nature of adverse effects that an LMO could cause to an organism, system, or (sub)population.
2. Hazard characterization – The qualitative and/or quantitative evaluation of the nature of the adverse effects associated with an LMO.
3. Exposure assessment – Evaluation of the exposure of the environment, including organisms, to an LMO or products thereof.
4. Risk characterization – The qualitative and/or quantitative estimation, including attendant uncertainties, of the overall risk.

If risks are identified during the risk characterization step above, risk management strategies may be identified which may effectively prevent, control or mitigate the consequences of the adverse effects. As such, the risk assessment process often includes an additional step to identify a range of possible risk management strategies that could reduce the level of risk.

It is worth noting, however, that it is only during the decision-making process that a choice is made as to whether an identified risk is acceptable and whether or not risk management strategies are to be implemented (see more details on the identification of risk management strategies under step 5).

As a whole the risk assessment process can be highly iterative; meaning that one or several steps may need to be re-evaluated when, for instance, new information becomes available in an attempt to increase the level of certainty.

The methodologies for risk assessment of LMOs have evolved over the past few decades. At a conceptual level, the methodologies have been adapted from the existing paradigms for environmental risk assessment developed for chemicals and other types of environmental stressors (Hill, 2005). As a result, the terminology used within each methodology may vary.

Familiarity with the different terms used in risk assessment enables a more direct comparison between the terminology used in Annex III and different risk assessment frameworks. It will also facilitate the interpretation of results from different risk assessments, for instance, for the same LMO.

	Figure 5 – Variation in terminology used to describe methodological components common to many risk assessment frameworks Source: Hill (2005).

Overarching issues

Risk assessors need to identify the information needed to conduct a risk assessment and understand how it will be used. Using and interpreting existing information, as well as identifying information gaps and understanding how to deal with scientific uncertainty are important factors during the risk assessment.

Quality and relevance of information

Considerations of the quality and relevance of information available for the risk assessment are important throughout the risk assessment process. Relevant information may be derived from a variety of sources such as existing scientific literature, experience and outcomes from previous risk assessments, in particular for the same or similar LMOs introduced in similar receiving environments, as well as new experimental data such as laboratory experiments (e.g. early tier toxicology testing), confined field experiments or other scientific observations. The relevance and level of detail of the information needed may vary from case to case depending on the nature of the modification of the LMO, on its intended use, and on the scale and duration of the environmental introduction.

Scientifically sound methodologies should be determined and documented for testing any identified risk scenario. When assessment methods are well described, risk assessors and subsequent reviewers are better equipped to determine whether the information used was adequate and sufficient for characterizing the risk.

	Example 12 – Data acquisition, verification, and monitoring “The importance of the data acquisition, verification, and monitoring process in the development of accurate risk assessments has been emphasized. Models, no matter how sophisticated, are simply attempts to understand processes and codify relationships. Only the reiteration of the predictive (risk assessment) and experimental (data acquisition, verification, and monitoring) process can bring models close to being a true picture of reality.” Source: UNEP/IPCS (1994).

Identification and consideration of uncertainty

Uncertainty is an inherent and integral element of scientific analysis, and its consideration is undertaken throughout the whole risk assessment process. The risk assessment methodology as set out by the Cartagena Protocol states that “where there is uncertainty regarding the level of risk, it may be addressed by requesting further information on the specific issues of concern or by implementing appropriate risk management strategies and/or monitoring the living modified organism in the receiving environment”. ³

Although uncertainty may, in some cases, be addressed by requesting additional information, the necessary information may not always be available or new uncertainties may arise as a result of the provision of additional experimental data. The golden rule during the risk assessment of an LMO is to request additional information that is relevant to the overall evaluation of risk and will facilitate the decision making. Thus, it is important to consider and analyze, in a systematic way, the various forms of uncertainty (e.g. types and sources) that can arise at each step of the risk assessment process.

Uncertainties may arise from: (i) lack of information, (ii) incomplete knowledge, and (iii) biological or experimental variability, for example, due to inherent heterogeneity in the population being studied or to variations in the analytical assays. Uncertainty resulting from lack of information includes, for example, information that is missing and data that is imprecise or inaccurate (e.g., due to study designs, model systems and analytical methods used to generate, evaluate and analyse the information) (SCBD, 2012).

If the level of uncertainty changes during the risk assessment process (e.g. by provision of new information), an iteration of parts or the entire risk assessment process may be needed.

It is important to note that while scientific uncertainty is considered during the risk assessment process and the results of uncertainty considerations may be reported it is, ultimately, the responsibility of the decision-makers to decide how to use the information in conjunction with the principals of the precautionary approach when making a decision on an LMO.

	Example 13 – Scientific uncertainty “There is no internationally agreed definition of ‘scientific uncertainty’, nor are there internationally agreed general rules or guidelines to determine its occurrence. Those matters are thus dealt with – sometimes differently – in each international instrument incorporating precautionary measures.” Source: IUCN (2003).

Planning phase

Establishing the context and scope

When the regulatory process of a country triggers the need for a risk assessment, it usually results in a request from the competent authority to the risk assessor(s). This request includes the scope of the risk assessment to be carried out as well as some important elements that will set the context of the risk assessment. In a typical case-by-case scenario, in accordance with the Cartagena Protocol, these elements will include at a minimum: the LMO(s), its(their) specific use(s) and, in cases of introduction into the environment, the likely potential receiving environment(s) where the LMO may be released and establish itself. As such, the case-by-case approach does not allow an existing risk assessment to be applied “as is” to different LMOs, uses or receiving environments. Nevertheless, a risk assessment carried out on a case-by-case basis most often takes into account relevant knowledge and experience gained in previous risk assessments.

In practice, if a risk assessor is faced with a request by the Competent National Authorities (CNA) to conduct or review a risk assessment that does not follow the case-by-case principle, the risk assessor recommends to the CNA that a new risk assessment be carried out with a scope that is specific to the case under consideration (i.e. the LMO, its specific use and the likely potential receiving environment).

Protection goals for the conservation and sustainable use of biodiversity, may be defined in national, regional and international policies. In setting the context of a risk assessment, these goals may be relevant to the identification and selection of appropriate assessment endpoints and to determining which methodology will be used in the risk assessment process. Understanding the contribution of national, regional and regulatory policies in setting the context of the risk assessment is part of the preparatory work for a risk assessment as seen in Module 2.

After consideration of the protection goals, the risk assessment of a particular LMO proceeds to establishing the scope in order to define the extent and the limits of the risk assessment process. This phase usually consists of at least three main actions: (i) selecting relevant assessment endpoints or representative species on which to assess potential adverse effects; (ii) establishing baseline information; and (iii) when possible, establishing the appropriate comparator(s).

Although these actions are described here as separate activities, in practical terms, this is an iterative process where the risk assessors will usually draw on the results of each action to inform the subsequent actions until all their elements have been considered sufficiently enough to enable the risk assessment to proceed.

Selecting relevant assessment endpoints or representative species

The purpose of an assessment endpoint or of representative species is to provide a measure that will indicate whether or not the LMO may cause an adverse impact on a protection goal. In order to be useful, the selected assessment endpoints or characteristics of the representative species should be specific and measureable.

Assessment endpoints or species representative of important ecological functions ⁴ or roles should be selected on a case-by-case basis. The complexity of ecosystems and the large number of potential candidates add to the challenges in selecting the appropriate assessment endpoints in ecological systems. Some important criteria for the selection of assessment endpoints to be used in the risk assessment of LMOs may include, for example: (i) their relevance to the protection goals; (ii) a well-defined ecological function; (iii) accessibility to measurement; and (iv) level of potential exposure to the LMO.

Identifying assessment endpoints or representative species that are relevant within the context of the likely potential receiving environment allows the risk assessor(s) to focus on interactions that are likely to occur. Moreover, risk scenarios may be also formulated to include assessment endpoints or representative species that are not present in the likely potential receiving environment but may, nevertheless, be indirectly exposed to the LMOs. This could occur, for example, if a third species, which is sexually compatible with the LMO and the representative species, has a distribution area that overlaps with the distribution areas of the former two providing an indirect exposure pathway between them.

	Example 14 – Common problems in selecting assessment endpoints Endpoint is a goal (e.g. maintain and restore endemic populations); Endpoint is vague (e.g. estuarine integrity instead of abundance and distribution of a species); Ecological entity may not be as sensitive to the stressor; Ecological entity is not exposed to the stressor (e.g. using insectivorous birds for avian risk of pesticide application to seeds); Ecological entities are irrelevant to the assessment (e.g. lake fish in salmon stream); Importance of a species or attributes of an ecosystem are not fully considered; Attribute is not sufficiently sensitive for detecting important effects (e.g. survival compared with recruitment for endangered species). Source: US Environmental Protection Agency (1998).

	Example 15 – Questions asked when selecting representative species for assessing effects of Bt plants on non-target organisms Which variant of the Bt protein are we dealing with? Where is it expressed (in the leaves, pollen or only in the roots)? Is it produced in the plant throughout its life or only during particular growth phases? Which insects come into contact with the Bt protein? Is this contact direct and long-term or only occasional? Which insects ingest the Bt protein through their prey? Source: GMO Safety (website).

Establishing the baseline

In risk assessment, the baseline is a description or a measurement of existing conditions of an environment, or its attributes or components without the LMO under consideration and taking into account different practices in use (e.g., agricultural practices). The baseline description or measurement may provide quantitative (e.g., number of organisms, variability of abundance) and/or qualitative information about the receiving environment as a reference for estimating effects of the LMO or its use including, if applicable, information on the assessment endpoints. Baselines can refer to, for instance, a particular environment or health conditions of a population.

Baselines are established with the aim of having descriptive and/or measurable information on any element of the likely potential receiving environment that is considered relevant in assessing the impacts from the introduction of the LMO, including considerations on possible impacts on human health.

In practice, if relevant assessment endpoints or representative species are selected, the baseline data will consist of data that establishes the conditions of these endpoints or species before the introduction of the LMO in question.

The choice of comparators

As seen above, a comparative approach is one of the general principles of risk assessment as set out in Annex III to the Protocol, where risks associated with the LMO “should be considered in the context of the risks posed by the non-modified recipients or parental organisms in the likely potential receiving environment”.

Using a comparator, i.e. non-modified recipients or parental organisms of the LMOs used as an element to establish the basis for a comparative assessment in accordance with Annex III, helps a risk assessor to identify the novel characteristics of the LMO and assess if the LMO presents a greater, lesser or equivalent risk compared to the non-modified recipient organism that is used in a similar way and in the same environment.

The ideal comparator is the closest non-modified genotype to the LMO, i.e. (near-)isogenic lines ⁵. However, (near-)isogenic lines are not always available and the choice of appropriate comparators may be guided by policies or guidelines adopted by the country undertaking the risk assessment (e.g. EFSA, 2011). Moreover, depending on the context, the step of the risk assessment and question being asked, a risk assessor may also choose to consider similar or related non-modified genotypes as useful comparators. Related management practices and experience with similar non-modified organisms may also be helpful. For example, when considering the risk assessment for an insect resistant LM crop, a risk assessor may wish to consider, amongst other things, the available experience with pest control practices applied to non-modified organisms of the same species (e.g. use of spores from Bacillus thuringiensis as pesticides).

In some circumstances, choosing an appropriate comparator(s) can be a challenge. This may happen, for example, in the case of LM crops that are tolerant to abiotic stress if the non-modified recipient or parental organisms are not capable of growing in the receiving environment. In extreme situations, when a suitable comparator cannot be grown under the same conditions and in the same or similar receiving environment as the LMO, it may be necessary to treat the LMO as a novel species in that environment (i.e. introduced species). This means that the characterization of the LMO (see below) will focus not only in the novel genotypic and phenotypic characteristics ⁶ resulting from the genetic modification, but rather on the characterization of an entire new genotype in the particular receiving environment.

Conducting the risk assessment

Conducting the risk assessment involves synthesizing what is known about the LMO, its intended use and the likely potential receiving environment to establish the likelihood and consequences of potential adverse effects to biodiversity, taking into account human health, that result from the introduction of an LMO.

Neither the Protocol nor this Manual makes a distinction between the various types of introductions into the environment, such as releases for experimental purposes or releases for commercial purposes. However, the nature and level of detail of the information needed to conduct the risk assessment will vary depending on the intended use of the LMO (e.g. type of release), the LMO itself and the likely potential receiving environment.

The following sections will address the steps of the risk assessment methodology described in paragraph 8 of Annex III to the Protocol. These steps describe a structured and integrated process whereby the results of one step are relevant to subsequent steps. Additionally, the risk assessment process may need to be conducted in an iterative manner, whereby certain steps may be repeated or re-examined to increase or re-evaluate the reliability of the risk assessment. If during the process, new information arises that could change the outcome of a step, the risk assessment may need to be re-examined accordingly.

Step 1: Identification of any novel genotypic and phenotypic characteristics associated with the LMO that may have adverse effects

The first step of the risk assessment is “an identification of any novel genotypic and phenotypic characteristics associated with the LMO that may have adverse effects on biological diversity in the likely potential receiving environment, taking into account risks to human health”. ⁷

What constitutes an “adverse effect” (also referred to as “damage” or “harm”) will depend on the context and scope of the risk assessment taking into account, as appropriate, the specific protection goals as seen above.

	Example 16 – Potential adverse effects “Harm adverse effect reflects an undesirable condition involving damage or injury. This includes change in the morphology, physiology, growth, development, reproduction or life span of an organism or group of organisms that results in an impairment of functional capacity, an impairment of the capacity to compensate for additional stress or an increase in susceptibility to other influences. The perception of harm can vary between people. It can also change over time and differ according to other factors such as variations in the vulnerability of individuals or type of land use. For example, a cold medication may be considered harmful if it causes severe side-effects. However, if a cancer drug causes the same type of side-effects, it may not be considered harmful. Similarly, a plant producing large amounts of biomass in a pasture may be considered desirable whereas the same plant may be considered harmful (weedy) in a nature conservation area as it may end up displacing a native species. In addition, one harmful outcome can sometimes give rise to further downstream harms. For example, increased harms from weeds, pests or pathogens can lead to loss of biodiversity.” Source: OGTR (2013).

	Example 17 – Potential risks With every new emerging technology, there are potential risks. These include: The danger of unintentionally introducing allergens and other anti-nutrition factors in foods; The likelihood of transgenes escaping from cultivated GM crops into wild relatives; The potential for pests to evolve resistance to the toxins produced by GM crops; The risk of these toxins affecting non-target organisms. Source: GMAC Singapore (website).

	Example 18 – Potential adverse effects from weediness in plants Competitive exclusion of other plants; Reduction in yield/biomass of other plants; Reduction in quality of products/services; Restriction of physical movement (e.g. of water, people, animals); Harm to human and/or animal health; Altered ecosystem processes (e.g. levels of nitrogen fixation, water supply and use, soil sedimentation or erosion and salt accumulation). Source: FAO (2011a).

	Example 19 – Topics of concern According to the International Centre for Genetic Engineering and Biotechnology (ICGEB), the main issues of concern derived from the deliberate introduction of LM crops (and their derived products) into the environment or onto the market are classified as: Risks for animal and human health – Toxicity & food/feed quality/safety; allergies; pathogen drug resistance (antibiotic resistance), impact of selectable marker; Risks for the environment – Persistency of gene or transgene (volunteers, increased fitness of LM crop, invasiveness) or of transgene products (accumulative effects); susceptibility of non-target organisms; change in use of chemicals in agriculture; unpredictable gene expression or transgene instability (gene silencing); environmentally-induced (abiotic) changes in transgene expression; ecological fitness; changes to biodiversity (interference of tri-trophic interactions); impact on soil fertility/soil degradation of organic material; Gene transfer – Through pollen or seed dispersal & horizontal gene transfer (transgene or promoter dispersion); transfer of foreign gene to micro-organisms (DNA uptake) or generation of new live viruses by recombination (transcapsidation, complementation, etc.); Risks for agriculture – Resistance/tolerance of target organisms; weeds or superweeds; alteration of nutritional value (attractiveness of the organism to pests); change in cost of agriculture; pest/weed management; unpredictable variation in active product availability; loss of familiarity/changes in agricultural practice.” Source:// ICGEB (website).

The genotypic and phenotypic characterization of an LMO provides the basis for identifying differences, both intended and unintended, between the LMO and its recipient or parental organism(s). Molecular analyses may be performed to characterize the products of the modified genetic elements, as well as of other genes that may have been affected by the modification. Data on specific expression patterns may be relevant for risk assessment in order to determine exposure, and may also include data confirming the absence of gene products, such as RNA and proteins, different from those originally intended. For example, in the case where the gene product (i.e. the RNA or protein that results from the expression of a gene) is intended to function only in a specific tissue, data may be used to confirm its specificity in that tissue and demonstrate its absence in other tissues.

Other phenotypic data are often presented to indicate that the LMO is behaving as anticipated. This could include data on reproductive characteristics, alterations in susceptibility to pests and diseases or tolerance to abiotic stressors, etc.

Once the potential adverse effects have been identified, the risk assessment proceeds to estimating the likelihood and consequences of these effects. To this end, developing risk scenarios may in some cases provide a useful tool.

A risk scenario may be defined as a sequence of events with an associated probability and consequence. In the context of risk assessment of LMOs, a risk scenario may be explained as a scientifically supportable chain of events through which the LMO might have an adverse effect on an assessment endpoint.

	Example 20 – A Risk scenario “The possibility that growing Bt corn may kill ladybird beetles due to ingestion of the Bt protein when preying on insects feeding on the GM corn, thereby reducing the abundance of coccinellids in the agroecosystem and increasing the incidence of pests.” Source:// Hokanson and Quemada (2009).

A well-defined risk scenario should be scientifically plausible and allow the assessor to identify information that is necessary for the assessment of risks.

Although some risk scenarios may appear as obvious (e.g. potential for insect resistant plants to affect insect herbivore populations), it is always useful to identify the risk scenarios fully. Clear and well-defined risk scenarios can also contribute to the transparency of a risk assessment because they allow others to consider whether or not the subsequent steps of the risk assessment have been adequately performed and facilitate the consideration of possible strategies to manage the identified risks.

A common challenge in generating a well-defined risk scenario is to choose representative species that would be exposed to the LMO. This is why an exposure assessment should be considered when selecting assessment endpoints.

When establishing risk scenarios several considerations may be taken into account. These may include: (i) gene flow followed by introgression of the transgene in species of interest; (ii) toxicity to non-target organisms; (iii) allergenicity; (iv) multi-trophic interactions and indirect effects; and (v) resistance development. The following paragraphs explain some of these considerations in more detail:

Gene flow followed by introgression of the transgene in species of interest – “Gene flow” is the transfer of genetic material from one organism to another by vertical or horizontal gene transfer; ⁸ or the movement of an organism from one environment to another. In the case of plants, vertical gene flow may occur even between organisms that are located far apart since pollen can be carried across large distances by the wind or insects, for instance. “Introgression” is the movement of a gene or genetic element from one species into the gene pool of another species or population, which may result in a stable incorporation or some fertile offspring.

Gene flow followed by introgression from an LMO to non-modified organisms may or may not be considered an adverse effect depending on the protection goals.

The potential for gene flow is first evaluated by investigating if sexually compatible species are present in the likely potential receiving environment. If sexually compatible species are present, there is a possibility of gene flow from the LMO to these species. Whether or not the modified genetic elements can potentially introgress into the population of the sexually compatible species will be largely determined by the biology of the recipient organism and of the LMO itself (see considerations regarding the likelihood and consequences of gene flow and introgression in steps 2 and 3).

	Figure 6 - Gene flow to conventional crops and distant relatives through “genetic bridges” Source: Heinemann (2007).

Toxicity to non-target organisms – The potential for an introduced gene product to be toxic to organisms in the environment is typically addressed by controlled exposure in the environment or by direct toxicity testing, or by a combination of the two. Non-target organisms may include, for instance, herbivores, natural enemies (e.g. parasitoids and predators), pollinators and pollen feeders, soil (micro-)organisms and weeds. The need and extent of toxicity tests will depend on characteristics of the LMO and the level of exposure of other organisms to the LMO.

If toxicity testing is needed, it typically follows a sequential series of tiered tests. Early tier studies involve highly controlled laboratory environments where representative or surrogate test species are exposed to high concentrations of the gene product being studied (i.e. worst case exposures) to determine if there are any toxic effects. If toxic effects are observed in early tier tests or if unacceptable uncertainty exists, e.g. regarding effects in multi-throphic interactions (see below), more realistic conditions representative of field-level exposures can be tested to determine the extent of the risk.

The gene products of the modified genetic elements in LMOs may be produced in very small quantities thus may be difficult to isolate in the amounts required for toxicity testing. If this is the case, and it is determined that toxicity tests are required, the risk assessor may consider results from tests using gene products obtained from alternate (surrogate) sources (e.g. bacterial expression systems or the organism from which the transgene was derived) provided that these gene products are chemically and functionally equivalent.

Allergenicity – Allergies are a type of adverse immunological response that affect individuals who are predisposed to certain types of substances (i.e. allergens). Allergens are often proteins or peptides.

In considering allergenicity caused by LMOs, it is important to take into account the exposure to proteins newly expressed by the LMO, including some variants of these proteins (e.g. structural variants of proteins having sometimes very few difference(s) in amino acids composition – or no difference in amino acids composition but carrying slightly different saccharide branches – that may display different allergenic properties through differences in spatial structure) that may be produced uniquely by the LMO. As a consequence, some allergenicity studies must be carried out with proteins isolated from the LMO itself, and not obtained from an alternate (surrogate) source such as a bacterial expression system.

It is also possible that allergens known to exist in the recipient or parental organism(s) are produced in higher amounts, for example by over-expression of the gene that encodes a protein that is known to be a common allergen.

	Figure 7 – Assessment of the allergenic potential of foods derived from modern biotechnology Source: FAO/WHO (2001).

Multi-trophic interactions and indirect effects – “Multi-trophic interactions” involve more than two trophic levels in a food web. They are an important concept in ecology and occur when a change at one trophic level indirectly affects trophic levels which are more than one step away. Consideration of tri-trophic interactions and indirect effects may be relevant to biodiversity protection goals.

	Example 21 – Multi-trophic interactions and indirect effects An important feature of non-target effects is that they can involve knock-on food-web effects, such as effects on predators and parasitoids that are exposed to the transgenic product through their prey or hosts that feed on the GM crop (known as tritrophic exposure), or more complicated linkages. If the prey or host are unaffected by the transgenic product themselves, they may expose their predators or parasitoids over a prolonged period of crop growth, and they may also concentrate the transgenic protein in their bodies to levels higher than those found in the plant tissues. Source: Underwood et al. (2013).

Observations and experimentation to identify such effects are challenging because of the complexity of ecological interactions, the difficulty of establishing causality between observed variation and treatment effects (e.g. the presence of the modified genetic element or its products), and natural variability in populations over time. Moreover, in a food chain (or food web), effects at the trophic levels may become observable only at a later stage.

Resistance development – The extensive use of herbicides and insect resistant LM crops has the potential to result in the emergence of resistant weeds and insects. Similar breakdowns have routinely occurred with conventional crops and pesticides. Several weed species have developed resistance to specific herbicides which are extensively used in combination with herbicide-resistant LM crops. Insect-resistant Bt-crops similarly could lead to the emergence of Bt-resistant insects (FAO, 2004).

The extent of the adverse effect and possible consequences of the insurgence of resistant weeds and insects should be thoroughly considered in a risk assessment. Some regulatory frameworks require that risk management strategies are identified in order lower the risk of resistance development.

Elements of a case-by-case risk assessment of LMOs

The case-by-case approach in risk assessment is based on the premise that risks that may arise from the release of an LMO depend on three main elements: (i) the LMO itself; (ii) the likely potential receiving environment; and (iii) the intended use of the LMO in question. In order to identify and assess risks, each of these elements needs to be characterized in a concerted manner and as appropriate for the specific risk assessment. Moreover, it is important to note that while these three elements may be sufficient to establish the boundaries of a risk assessment, potential adverse effects may extend past these elements, for instance, beyond the likely potential receiving environment and the intended use(s) of the LMO.

The information required for each of these elements in a risk assessment may vary in nature and level of detail from case to case. The following sections provide examples of information that may be relevant for the characterization of each element above. These sections include several of the “points to consider” as indicated in paragraph 9 of Annex III of the Protocol.

A large portion of the information listed here is usually included in the LMO request triggering the risk assessment. The risk assessors can determine whether or not the information provided is sufficient and adequate for conducting a scientifically sound risk assessment. If needed, they can obtain additional information by, for instance, carrying out their own investigation or requesting it from the applicant.

	Example 22 – The case-by-case approach “A risk assessment performed for a particular LMO intended to be introduced to one environment may not be sufficient when assessing the possible adverse effects that may arise if that LMO is to be released under different environmental conditions, or into different receiving environments. A risk assessment performed for a particular use of a particular LMO may not be sufficient when assessing the possible adverse effects that may arise if that LMO is to be used in different ways. Because of this, it is important for each case to be addressed separately, taking into account specific information on the LMO concerned, its intended use, and its potential receiving environment.” Source:// IUCN (2003).

Living modified organism

Characterization of the recipient organism

In order to identify whether or not the LMO possesses characteristics that may cause potential adverse effects (see above), it is first necessary to have information about the non-modified recipient organism (or parental organisms).

For many LMOs, the biology of the recipient organism will strongly influence the potential interactions of the LMO in the receiving environment. Information on the recipient organism is therefore essential as it will help the risk assessor identify the exposure, its scenarios and, ultimately, if any risk is posed by an LMO.

The information that is needed for the characterization of the recipient organism will vary depending on each case. For example, the nature and detail of information about the recipient organism that is required may differ between small-scale releases for experimental purposes and large-scale commercial releases. It normally includes the biological and reproductive characteristics of the recipient organism that can be important for determining the potential exposure of other organisms, such as predators, prey, competitors or pathogens, to the LMO in question in the likely potential receiving environment.

For many species of LMOs, information on the recipient organism can be found in biology documents, such as those published by the Organization for Economic Co-operation and Development (OECD) ⁹ and the Canadian Food Inspection Agency (CFIA). ¹⁰

The LMO will, in most cases, share most of its genetic characteristics with its actual recipient organism (i.e. the one used in the modification) rather than with other genotypes of the same species. Thus, it is also important to consider, whenever possible, comparative data from the actual non-modified recipient organism (see the section on “The choice of comparators”).

Information about recipient organism to be considered may include:

Taxonomic status – This information is useful for identifying the recipient organism and ensuring that information provided and cited during the assessment pertains to the organism for which the assessment is being carried out. Typically, the taxonomic status includes the scientific name (i.e. genus and species, for example, Zea mays) and information about the taxonomic family (e.g. Poaceae). This may also include other information used to further classify (e.g. sub-species, variety, strain) or differentiate the recipient or parental organism(s) (e.g. ploidy level or chromosome number).

Common name – The familiar or colloquial names for the recipient organism that may be commonly used in the country of introduction and in international trade may be useful for finding information relevant to the biology of the organism. Caution is recommended when using information about recipient organism when only common names (versus the scientific name) are used because the same common name can be applied to more than one species.

Biological characteristics – Information on the biological characteristics of the recipient organism, such as the production of endogenous toxins and allergens, its reproductive biology, dispersal of seeds and vegetative propagules, and growth habits, are also important points for consideration.

Origin – The origin of the recipient organism refers to its place of collection and may be important because populations within a species (e.g., variety, strain, isoline, etc.) may have significantly different characteristics. For domesticated species, this may be supplemented with a pedigree map where available.

Centres of origin and centres of genetic diversity – Knowledge of the centre(s) of origin and genetic diversity can provide information on the presence of sexually compatible species and the likelihood of ecological interactions in the receiving environment. In the absence of more specific information, the centre of origin can also offer insight into the biology of the species (e.g. habitats to which the species is adapted).

Habitat where the recipient or parental organism(s) may persist or proliferate – Information about the ecosystems and habitats (e.g. temperature, humidity, altitude, etc) where the recipient organism is known to be native and where it may have been introduced and is now established provides useful baseline information. This allows the risk assessors to understand the range of habitats in which the species exists, the range of behaviours exhibited in those habitats, and how characteristics of the species determine the range of habitats where it can persist or proliferate. This information can be very valuable in determining the likely potential receiving environment and, consequently, the level of exposure to the LMO. Likewise, the ecological characteristics of the recipient organism will help determine which organisms in the likely potential receiving environment are likely to come into contact, either directly or indirectly, with the LMO and will help determine the exposure pathways. For more details on the type of information that may be useful, see the section “Likely potential receiving environment” below.

The history of use can be very valuable as well. If an organism persists in heavily managed environments (e.g. agriculture, sylviculture or recreationally managed land) then this will provide information about the conditions necessary for its survival. It may also provide direct indications of how the LMO will behave in other managed environments.

Description of the genetic modification

Information on the genetic material that was introduced or modified, as well as the method used for the genetic transformation is useful in identifying novel properties of the LMO such as, what new gene products are expressed and which of the endogenous genes of the recipient or parental organism(s) may be affected by the genetic modification.

Typically the description of the genetic modification includes information on (i) the “donor organism(s)” or the source of the inserted genetic element(s); (ii) characteristics of each modified genetic element, including their intended and known biological function(s); (iii) the vector used, if applicable; and (iv) the transformation method. Below is a brief explanation on each of these points:

Donor organism(s) – The relevant information on the donor organism(s) includes its taxonomic status, common name, origin and relevant biological characteristics.

Modified genetic elements – The relevant information on the modified genetic elements encompasses the name, sequence, function and other characteristics of all the nucleic acid sequences that were inserted, deleted or modified in the LMO. These include not only the target gene(s) but also, for example, all marker genes, regulatory sequences, and any non-coding DNA. If available, a history of use may be important with regards to potential toxicity or allergenicity of the gene products derived from the donor organism. If the genetic elements originate from a donor organism that is known to be a pest or pathogen it is also relevant to know if and how these elements contribute to the pest or pathogenic characteristics.

Vector – In molecular biology, a vector is a nucleic acid molecule used as a vehicle to transfer foreign genetic material into a cell. If a vector, for example a plasmid, was used for the transformation, relevant information includes its identity, source or origin, and its host range.

Transformation method – Specifying the method that was used in the transformation (e.g. Agrobacterium mediated, particle gun, etc.) is also relevant when describing the genetic modification. Depending on the transformation method, parts of the vector(s) may also be incorporated into the genome of the newly developed LMO.

Characteristics of the modification – This refers to information about whether or not the inserted or modified genetic elements are present and functioning as expected in the LMO. Normally this involves confirmation that the DNA insert or modified genetic element is stable in the genome of the LMO. Information such as the insertion site in the genome of the recipient or parental organism(s), cellular location of the insert (e.g. chromosomal, extrachromosomal, or chloroplast DNA), its mode of inheritance and copy number may also be relevant.

Identification of the LMO

With regard to the identification of the LMO, the following are important points to consider:

Unique identifiers – A Unique identifier is a code provided by the LMO developer to a transformation event ¹¹ derived from recombinant DNA techniques to enable its unequivocal identification. Each unique identifier is made up of a sequence of 9 alphanumeric digits, for example MON-89788-1, assigned according to the OECD guidance document (OECD, 2006).

Detection and identification methods – The availability of methods for detection and identification of the LMO may be considered as well as their specificity, sensitivity and reliability. This information may be relevant not only for assessing risks but also when considering possible monitoring and risk management strategies (see step 5 below). Some regulatory frameworks require a description of such methods as a condition for regulatory approval in order to ensure the tools to assist with monitoring and risk management are available.

The Biosafety Clearing-House of the Cartagena Protocol maintains an LMO registry ¹² containing, amongst other things, information on unique identifiers, molecular characteristics and available detection methods for the LMOs addressed in countries’ decisions.

	Example 23 – CFIA detection and identification method criteria According to the Canadian Food Inspection Agency, acceptable methods for detection and identification of LMOs must address the following: Test Type - Methods must be suitable and may be protein, RNA or DNA based. Phenotypic based methods will not generally be considered suitable. Limit of detection - Methods must meet the following sensitivity and accuracy requirement: For those methods that are grain based, the method must be able to detect 0.2% modified grain (2 grains in 1000) with a 95% confidence interval. For those methods that are not grain based (e.g. single ingredient feed) the method must be able to detect 0.2% modified material in a sample with a 95% confidence interval. Procedural clarity -The method must be complete and laid out in a step wise fashion that may be easily followed by a person unfamiliar with the method. Detailed descriptions of sample size, replicates, extraction procedure, expected results (figures/sequences), interpretation and acceptance criteria must be included. Cross reactivity - The method must be shown to be specific to the PNT of interest. Any potential for cross reactivity must be clearly stated. Cross reactivity data must be provided demonstrating that the method does not cross-react with other commercially available PNTs of the same species with similar traits/modifications that are currently available in the Canadian marketplace. Reference material - The company must provide appropriate reference materials to the CFIA upon request. Appropriate reference material will be determined by the CFIA based on the method provided. Contact information - The company must provide contact information for a technical support person. Source: CFIA (website).

Likely potential receiving environment(s)

The Protocol calls for the characterisation of the “likely potential receiving environment” of an LMO. According to UNEP (1995), the “potential receiving environment” is the range of environments (ecosystem or habitat, including other organisms) which are likely to come in contact with a released organism due to the conditions of the release or the specific ecological behaviour of the organism. In other words, the likely potential receiving environment of an LMO encompasses both the environments where the LMO will be intentionally introduced as well as other environments which are likely to be exposed to the LMO.

As such, during a risk assessment, in addition to the area where the LMO will be intentionally introduced, the relevant characteristics of the likely potential receiving environment of an LMO should also be thoroughly examined with particular attention given to areas where exposure levels to the LMO will be the highest.

The characterization of the likely potential receiving environment takes into account its ecological characteristics, including physical location/geography, climate, its biological entities and their interactions. The characterization of the likely potential receiving environment will help in selecting appropriate assessment endpoints for the risk assessment (see Module 2) and will also affect the assessment of the potential interactions of the LMO with other organisms.

To determine the likely potential receiving environment, risk assessors may consider potential pathways for dispersal of the LMO as well as the habitats where the recipient/parent organism(s) may persist or proliferate.

An analysis of possible dispersal routes and mechanisms is important when establishing the likely potential receiving environments. Different dispersal mechanisms may exist and could be inherent either to the LMO (e.g. altered seed characteristics), its intended use (e.g. shipment practices) or the receiving environment (e.g. proximity to a river). A scientifically sound risk assessment takes into consideration all possible dispersal mechanisms, keeping in mind the biology of the LMO and non-modified recipient or parental organism(s), in a concerted manner for each case.

Information about the likely potential receiving environment can include considerations on both large scale (e.g. climate) and small scale characteristics (e.g. microclimate) depending on the complexity of the environment. The type of information on the likely potential receiving environment and the level of detail depend on the nature of the LMO and its intended use, in accordance with the case-by-case principle.

It may not be possible or practical to consider every possible interaction between the LMO and the receiving environment. Such challenges and limitations should be acknowledged during the risk assessment process.

Below are descriptions of some physical and biological characteristics of the likely potential receiving environment(s) that can be considered in the risk assessment of LMOs. This is an indicative list thus the information required to satisfy the needs of the assessment will vary depending on the nature of the LMO and its intended use.

The physical or “abiotic” characteristics of the likely potential receiving environment may have a great impact on the ability of an LMO to survive and persist.

Geography and climate – Geography encompasses characteristics such as latitude, which will influence day-length, and altitude. Climate encompasses temperature, precipitation, humidity, wind and other meteorological measures over long periods of time. For the purposes of environmental risk assessment, geography and climate are among the most important factors impacting the ability of an LMO to survive and persist. For LM plants, temperature and precipitation are likely to be key determinants. Seasonality (variations in climate on an annual cycle) can also be an important consideration in the potential survival and persistence of an LMO.

Soil – The type and quality of soil can greatly influence the ability of an LM plant to survive or persist without land management. The type and quality of a soil are heavily influenced by the organisms living in its proximity, but abiotic factors such as climate, geography and topography will also all play a role in determining its characteristics (e.g. mineral content, moisture level, texture etc.).

Management status – The management status of an environment is a measure of how much human intervention takes place in order to maintain a particular condition. A separate but related concept is “disturbance” which can be considered the amount of human activity that affects the environment but without the intention of maintaining a particular condition. Management and disturbance may greatly influence the ability of an LMO to survive and persist in the environment. Likely potential receiving environments can range from highly managed to unmanaged and from highly disturbed to undisturbed.

The biological characteristics of the likely potential receiving environment consist of all the living organisms present in the environment, its biological communities and the interactions among them.

Both managed and unmanaged environments contain complex biological characteristics that pose challenges for environmental risk assessments.

As with any other organism, an LMO released into the environment is expected to have many interactions with other organisms. For the purposes of environmental risk assessment, it is critical to develop verifiable risk scenarios and identify the appropriate species that may be impacted by the presence of the LMO in the environment. For example, gene flow and possibly introgression may occur when sexually compatible species are present in the likely potential receiving environment. The selection of suitable representative species in the likely potential receiving environment is also informative (see section on “Selecting relevant assessment endpoints or representative species”).

Intended use The characteristics of the intended use of an LMO and management practices associated with it, such as tilling and the use of pesticides, can provide valuable information and context for the risk assessment process. Understanding the intended use also helps a risk assessor to perform an exposure assessment starting with the environment where the LMO will be deliberately introduced followed by considering whether or not the LMO is likely to disseminate or persist outside of this environment.

To illustrate how the intended use can affect the likelihood of a risk posed by an LMO, a hypothetical case of an LM tree being used for wood production could be considered, in which the first flowering would occur after 15 years of planting, but logging would take place after only 10 years. As such, the intended use would result in the LM tree being logged before its first flowering. Consequently, in this hypothetical case, the intended use would influence the likelihood of potential outcrossing ¹³ of this LM tree.

Information regarding the intended use of the LMO may also take into account any new or changed use in comparison to the recipient or parental organism(s), for example, in cases where the recipient or parental organism(s) is a crop for human consumption but the intended use of the LMO is the production of a compound for pharmaceutical or industrial use.

The scale and type of the introduction into the environment, for example, field trials versus commercial releases, and whether or not any risk management strategy is being proposed, may also be relevant when considering the intended use. Many regulatory frameworks, for instance, require that submissions for field trials be accompanied by information on risk management strategies to reduce exposure to the LMO.

Considerations on the intended use may also take into account national and regional experiences with similar organisms, their management and exposure to the environment.

Step 2: Evaluation of the likelihood

This step entails an evaluation of the likelihood, i.e. probability, of the adverse effect occurring, taking into account the level and kind of exposure of the likely potential receiving environment to the LMO.

After the potential adverse effects of the LMO have been identified, the risk assessment proceeds to a formal analysis of the likelihood and consequence of these effects with respect to the identified assessment endpoints.

Although the steps of evaluating likelihood and consequences are dealt with separately in Annex III of the Protocol, some risk assessment approaches consider these steps simultaneously or in reverse order.

The likelihood of an adverse effect is dependent upon the probability of one or a series of circumstances actually occurring.

It is difficult to describe in detail an evaluation of likelihood or consequence without using an example because the evaluation is dependent on the nature of the LMO, the receiving environment and, if appropriate, on the risk scenario used. The following are two examples:

• In a case where outcrossing of the transgene with a non-modified organism is determined to be possible (i.e. the two species are sexually compatible), the risk assessment may consider both the likelihood of the outcrossing and, if relevant, the likelihood of the LMO progeny to persist or proliferate. Considerations on the latter may be based, for example, on assessing whether or not the transgene would affect the fitness level of the progeny (i.e. the capability of individuals to compete and reproduce in a given environment). If the transgene induces a positive fitness effect, the likelihood that the population resulting from the outcrossing would increase is high. On the other hand, transgenes that have a negative fitness effect would result in a low likelihood that the resulting population would increase. Transgenes that have a neutral impact on fitness may persist in populations at low levels depending on the rate of outcrossing or introgression as well as the overall population dynamics of the species.
• In a case where the risk scenario involves the potential toxicity of an LM plant (or a substance produced by an LM plant) to a herbivorous insect: the analysis of likelihood may consider the probability that the insect will be present, that the insect will feed on the LMO and that the insect will ingest a sufficient quantity of the LMO to suffer an adverse effect. Likelihood may consider probabilities on an individual level (e.g. what are the chances an individual insect may consume the LM plant) or on a population level (e.g. what percentage of the population of insects will come into contact with the LMO) or both.

	Example 24 – Likelihood of introgression “To evaluate a possible ecological effect of an inserted gene being introgressed into a natural population it is important to estimate the probability of introgression. Such a probability estimate can be obtained from measurements of hybridisation rates, assumed selective advantage of inserted gene, and fitness measurements of parent plants, hybrid plants, and plants from the first and second back-cross generations. If hybrids are formed and it is likely that these hybrids are able to survive the consequences should be discussed.” Source: Ministry of Environment and Energy Denmark (1999).

Step 3: Evaluation of the consequences

The consequence of an adverse effect is the outcome, extent and severity of an adverse effect associated with exposure to an LMO, its handling and use, or its products (in the context of Annex III paragraph 5). Should adverse effects occur, they may be severe, minimal, or anywhere in between. The evaluation of the consequences may consider the effects on individuals (e.g. mortality, reduced or enhanced fitness, etc.) or on populations (e.g. increase or decrease in number, change in demographics, etc.) depending on the adverse effect under evaluation.

The risk assessment should consider the consequences of each adverse effect based on a concerted analysis of what is known about the LMO, the likely potential receiving environment and the assessment endpoints, as well as the likelihood assessment.

	Example 25 – Consequences of effects to non-target organisms When the inserted trait cause the plant to produce potentially toxic compounds, or if flower characteristics are changed, i.e. colour, flowering period, pollen production etc. then effects on pollinators has to be measured. A test of effects on honeybees (Apis melliferae) is obligatory because of the importance of honeybees as pollinators of both wild and crop species and because standardised test protocols testing for effects of conventional pesticides exists for this pollinator. These tests include exposure through nectar and pollen. Source: Ministry of Environment and Energy Denmark (1999).

Also using an example where gene flow and introgression could lead to a potential adverse effect, what impact the presence of a transgene will have on biodiversity will depend on its effect on individual fitness as well as on the importance of that species relative to the protection goals. For instance, if a sexually compatible species, present in the receiving environment, is directly relevant to a biodiversity protection goal (e.g. it is a protected species) then the impact on biodiversity can be assessed by looking directly at the impact of the transgene on the population. If the sexually compatible species is not directly related to a biodiversity management goal, then the impact of the expression of the transgene will be dependent on indirect interactions. Indirect effects may be challenging to assess (see step 1), and are dependent on the ecological importance of the species.

Step 4: Estimation of the overall risk

This step consists of the integration of the likelihood and consequence of each of the individual risks identified through the preceding steps and takes into account any relevant uncertainty that emerged thus far during the process. In some risk assessment approaches, this step is referred to as “risk characterization”.

To date, there is no universally accepted method to estimate the overall risk but a variety of guidance materials are available that address this topic (see for instance, documents under “Scientific and technical issues / risk assessment” in the Biosafety Information Resource Centre, BIRC). ¹⁴

In rare instances, the risk characterization results in a quantitative value (e.g. 6% of a population will be exposed to a stressor, and of that percentage half will experience mortality). More frequently, the risk characterization for an LMO will be qualitative. In such cases, description of the risk characterization may be expressed as, for instance, ‘high’, ‘medium’, ‘low’, ‘negligible’ or ‘indeterminate due to uncertainty or lack of knowledge’.

The outcome of this step is the assessment of the overall risk of the LMO. Once this is achieved, it is helpful to determine, as an internal quality control, whether the risk assessment has met the criteria established at the beginning of the process taking into account also those criteria established in the relevant policies in practice with regard to the protection goals, assessment endpoints and risk thresholds (i.e. the level of tolerance to a certain risk or the level of change in a particular variable beyond which a risk is considered unacceptable).

	Figure 8 – Estimation of overall risk Source: ERMA NZ (1998).

	Figure 9 – Classification of risk Source: FAO (2011b).

Step 5: Acceptability of risk and identification of risk management and monitoring strategies

Annex III of the Protocol states that the risk assessment methodology may entail “a recommendation as to whether or not the risks are acceptable or manageable, including, where necessary, identification of strategies to manage these risks” and “where there is uncertainty regarding the level of risk, it may be addressed by requesting further information on the specific issues of concern or by implementing appropriate risk management strategies and/or monitoring the living modified organism in the receiving environment”. ¹⁵

For the “acceptability” of risks, please refer to the section “Recommendations as to whether or not the risks are acceptable or manageable” below.

Risk management

Risk management strategies refer to measures to ensure that risks identified in the risk assessment are reduced or controlled which may be implemented after the LMO is introduced into the environment (or placed in the market, if applicable). Risk management strategies can be useful to increase confidence when dealing with uncertainty or, in the case where risks have been identified, to reduce the likelihood or impact of the potential adverse effect.

	Example 26 – Application of management strategies for risks from the deliberate release or marketing of LMO(s) “The risk assessment may identify risks that require management and how best to manage them, and a risk management strategy should be defined.” Source: The European Parliament and the Council of the European Union (2001).

Risk management strategies may aim to reduce the likelihood or consequences of potential adverse effects and are referred to as “preventive measures” and “mitigation measures”, respectively. Some approaches to risk assessment may also include the identification of measures to control an adverse effect should it occur.

For LMOs, common risk management strategies have typically been designed to reduce the likelihood of exposure, but depending on the specific case, management options might include a variety of measures that are directly or indirectly related to the LMO. Some examples of risk management strategies for LMOs include: minimum distances from sexually compatible species if there is evidence that gene flow could cause adverse effects, destruction of seeds remaining in the field or of volunteers after harvest, restrictions from introduction into specified receiving environments, etc.

Certain risk assessment steps, particularly the evaluation of likelihood and consequences may need to be re-evaluated to take into account each of the identified risk management strategies since these may affect the estimation of the overall risks.

Monitoring

A risk assessor may identify the need for a strategy to monitor the receiving environment for adverse effects that may arise after the introduction of the LMO and include it as part of the recommendations for the Competent National Authority(ies). This may happen, for instance, when the level of uncertainty could affect the overall conclusions of the risk assessment. Moreover, some biosafety frameworks may have a policy to request a plan for monitoring as part of the risk assessment of all or particular types of LMOs.

Monitoring after the release of the LMO aims at detecting changes (e.g. in the receiving environment(s) or in the LMO) that could lead to adverse effects.

	Example 27 – Post-market monitoring “Post-market monitoring may be an appropriate risk management measure in specific circumstances. Following the safety assessment, the need and utility for post-market monitoring should be considered, on a case-by-case basis, during risk assessment and its practicability should be considered during risk management.” Source: Health Canada (2006).

Monitoring strategies may be designed on the basis of the protection goals identified by national legislation and regulation, if available, and parameters that are relevant to the indication of any increasing risk to the assessment endpoints in a “top-down” approach, or on the basis of specific risks in a “bottom-up” approach.

The strategies may include “general surveillance” that can make use of existing, broader monitoring programs that may identify unexpected effects of the LMOs or traits, such as long-term effects; or be “case-specific” where potential adverse effects identified during the risk assessment are investigated. Monitoring for the development of resistance in insect pests following introduction of pesticide producing LM crops would be an example of a “case-specific” scenario. Monitoring for the abundance of beneficial insect species in an environment would be an example of “general surveillance”.

	Example 28 – Case-specific monitoring and general surveillance of LM plants “The environmental monitoring of the GM plant will have two focuses: (1) the possible effects of the GM plant, identified in the formal risk assessment procedure, and (2) to identify the occurrence of adverse unanticipated effects of the GM plant or its use which were not anticipated in the environmental risk assessment. Appropriate case-specific monitoring measures should be developed on a case-by-case approach depending upon the outcomes of the risk assessment. Possible risks identified in the environmental risk assessment should be studied in hypothesis-driven experiments and tests. The objective of general surveillance is to identify the occurrence of unanticipated adverse effects of GM plants or their use on human health or the environment that were not anticipated in the environmental risk assessment. Since no specific risk is identified, no hypothesis of risk can be tested, so it is difficult to propose specific methods to carry out general surveillance.” Source: EFSA (2006).

Where it is appropriate, other potential adverse effects such as delayed, cumulative, combinatorial ¹⁶or indirect effects resulting from the LMO, the trait or the inserted or modified genes may be considered in the post-release monitoring strategies.

The level of specificity of the monitoring strategies may vary depending on the LMO(s), the intended use(s) and/or the likely potential receiving environment(s). Therefore, it is essential that a detailed methodology for each identified strategy also be identified. The methodology may include, for example, the frequency, locations and methods of sampling, as well as methods of analysis (e.g. laboratory testing).

Preparing a risk assessment report and recommendation

The outcomes of a risk assessment are often presented in the form of a written report prepared by the risk assessor(s). The report is primarily intended to assist the decision makers in making informed decisions regarding the safe use of an LMO.

Presenting the results of a risk assessment could be categorized as a form of risk communication. As in any form of communication, risk assessors should be mindful of the intended recipients, which in addition to decision makers may also include regulators, risk managers, other risk assessors and the general public amongst others.

	Example 29 – Risk communication Risk communication is the interactive exchange of information and opinions among assessors, risk managers, consumers, industry, the academic community and other interested parties throughout the risk analysis process. The information exchange concerns risk related factors and risk perceptions, including the explanation of risk assessment findings and the basis of risk management decisions. It is vitally important that risk communication with the public comes from credible and trusted sources. Source:// FAO (2001).

It is important that the report is presented in a well-structured form, which not only facilitates the deliberations of decision makers, but also allows for an easier exchange of information and experience. The context and scope of the risk assessment should be clearly explained as other institutions (e.g. in the same or in different countries) may have an interest in understanding how the risk of a particular LMO was assessed.

With regard to the sharing of information, a Party to the Protocol is required to submit to the Biosafety-Clearing House (BCH) all “summaries of its risk assessments or environmental reviews of living modified organisms generated by its regulatory process, and carried out in accordance with Article 15, including, where appropriate, relevant information regarding products thereof, namely, processed materials that are of living modified organism origin, containing detectable novel combinations of replicable genetic material obtained through the use of modern biotechnology” (Article 20). This will include all risk assessments generated to support decisions regarding LMOs for intentional introduction into the environment (Articles 8, 10 and 13) or for direct use as food or feed, or for processing (Article 11) whether they are triggered by a transboundary movement or by an internal request.

The required contents and format of a risk assessment report are generally defined by the Competent National Authority(ies) that have the responsibility to make decisions on the LMO(s) in the context of the national biosafety framework.

A risk assessment report typically comprises of an analytic synthesis of all the relevant steps and results of the risk assessment process, including an overview of the context and scope of the risk assessment, methodology used and a detailed summary of the results of the overall risk estimation, including the identification of individual risks, as well as the likelihood and consequences of the potential adverse effects.

The report may also contain an evaluation of the availability and quality of the scientific and technical information that was deemed necessary to perform the assessment and characterize the risks, and whether or not there were gaps in the information.

An analysis of all identifiable uncertainties and how they may impact the overall conclusions of the assessment is also a critical element of the report. This includes uncertainties identified at each step of the risk assessment process as well as those remaining at the end of the risk assessment.

Finally, the risk assessment report often contains a set of recommendations regarding the acceptability and manageability of the risks posed by the LMO and the identification of appropriate risk management and monitoring strategies.

The information above can be organized under five broad topics depending on the requirements of the National Authority that is responsible for the risk assessment:

1. Background, context and scope of the risk assessment;
2. Characterization and estimation of risks;
3. Description of risk management and monitoring strategies identified during the risk assessment;
4. Consideration of remaining uncertainty; and
5. Recommendations as to whether or not the risks are acceptable or manageable.

An overview of the information which may be included under each of these topics may be found in the following sections.

Background, context and scope of the risk assessment

This part of the report focuses on describing the issues that were considered while setting the context and scope of the risk assessment. Basically, this section of the report sets the scene for the reader to follow a clear progression through the subsequent sections of the report.

A risk assessment report usually specifies the mandate that was given to the risk assessor(s) and includes a description of the procedure that was followed in conducting the risk assessment, an indication of which institution has carried out the risk assessment, and which, if any, other institutions were consulted or were part of the process. Any other information that helps in understanding the context in which the risk assessment was carried out is also typically included in this part of the report.

Previous approvals or prohibitions of the same LMO, if any, including the regulatory status of the LMO in the country of export or import as well as in any other country may also be included in this section, if appropriate.

The report describes how the requirements of the national regulatory framework were taken into account including which protection goals were identified as relevant in the context of the risk assessment and how assessment endpoints were selected.

In summary, the following information may be included in this section of the report:

1. Contact details of the LMO developer;
2. Type of approval sought (e.g. introduction into the environment);
3. Contact details of the institution responsible for the risk assessment;
4. Relevant regulation;
5. Relevant protection goals and assessment endpoints;
6. Previous approvals or prohibitions of the same LMO;
7. Overview of the terms of reference for the risk assessment; and
8. Consulted experts or panel of experts, if applicable, and how the involved experts were chosen and how possible conflict of interests were identified and was managed.

In some cases, the bulk of information presented in this section of the report may be extracted from the request triggering the risk assessment, the national regulatory framework, including environmental and biosafety policies or guidelines, and national biosafety-related databases.

Characterization and estimation of risks

This section of the report focuses on the outcomes of the risk assessment steps in accordance with the steps in Annex III of the Protocol and as described above.

Depending on the specific mandate and scope of the risk assessment, the following information may be included in this section of the report:

1. Description of the LMO (e.g. recipient or parental organism(s), transformation method, inserted or modified sequences, novel traits, purpose of the genetic modification), its intended use and the likely potential receiving environment(s), including considerations on how the baselines were established and appropriate comparator(s) chosen;
2. Considerations of the availability and quality of information used during the risk assessment;
3. Methodology used in the risk assessment, explaining, if necessary, the use of terms;
4. Description of the potential adverse effects and risk scenarios arising from the novel characteristics of the LMO;
5. Analyses of the likelihood and consequences of each identified potential adverse effect; and
6. Estimation of the overall risk posed by the LMO.

The information relevant to each of the items above may vary in nature and level of detail on case-by-case basis, depending on the LMO concerned, its intended use and the likely potential receiving environment.

While information related to the description of the LMO and its intended use may be obtained in part from the LMO application, the bulk of information to be presented in this section of the report is obtained through the risk assessment process for the specific case at hand.

Description of risk management and monitoring strategies

If risk management and monitoring strategies were identified during the risk assessment process (see step 5), the risk assessment report should contain a section detailing any strategies to minimize the risks identified.

The risk assessment report may include, for instance:

1. How each identified strategy is expected to contribute to minimizing the likelihood or consequence of potential adverse effects (e.g. by reducing the exposure to the LMO or the consequences of the potential harm);
2. Details of the methodology for each identified risk management or monitoring strategy including, for instance, the frequency, locations and methods of sampling, as well as methods of analysis, including laboratory testing when appropriate;
3. Any uncertainty regarding the effectiveness of any such management or monitoring strategy;
4. An indication as to whether and how different management strategies can be combined to further minimize uncertainty or identified risks; and
5. Considerations on unintentional introduction into the environment and emergency measures as appropriate (see Article 17).

Consideration of remaining uncertainty

As seen in the section on “Overarching issues”, uncertainty is an inherent component of any risk assessment, and should be considered in a systematic manner at each step of the risk assessment process. Nevertheless, at the end of the risk assessment, uncertainties may still remain with regard to one or more specific steps in the process or about the likelihood or consequences of the potential adverse effects.

Annex III of the Protocol addresses this matter by requiring that “Where there is uncertainty regarding the level of risk, it may be addressed by requesting further information on the specific issues of concern or by implementing appropriate risk management strategies and/or monitoring the living modified organism in the receiving environment”. ³

Considerations of remaining uncertainties should be included in the risk assessment report. These considerations may include:

1. Identification of major information gaps and, where appropriate, indication of whether gathering additional data (either before the release or after it by monitoring) would significantly increase the overall confidence in the results of the risk assessment;
2. An analysis of uncertainty, including its types (e.g. gaps in the available information, limitations of the assessment methodology);
3. Discussion on the level of scientific support to issues where there is uncertainty, including an analysis of different scientific views;
4. Discussion of any assumption used in assessing the risks, including its strengths and weaknesses;
5. Discussion of the potential for uncertainties to impact on the overall conclusions of the risk assessment; and
6. Identification of any threats of serious or irreversible damage to the environment (basis for the adoption of the precautionary approach).

	Example 30 – Uncertainty and an approach based on the precautionary principle “The implementation of an approach based on the precautionary principle should start with a scientific evaluation, as complete as possible, and where possible, identifying at each stage the degree of scientific uncertainty. Decision-makers need to be aware of the degree of uncertainty attached to the results of the evaluation of the available scientific information. Judging what is an "acceptable" level of risk for society is an eminently political responsibility. Where possible, a report should be made which indicates the assessment of the existing knowledge and the available information, providing the views of the scientists on the reliability of the assessment as well as on the remaining uncertainties. If necessary, it should also contain the identification of topics for further scientific research.” Source:// Commission for the European Communities (2000).

Recommendations as to whether or not the risks are acceptable or manageable

Recommendations are one of the most important sections of a risk assessment report as they take into account the outcomes of the risk assessment to provide direct science-based advice to the intended recipients of the report. A recommendation as to whether or not the risks are acceptable or manageable should be kept within the scope of the risk assessment and based on its findings.

It is important to note that risk assessor(s) are requested to recommend whether the risks are “acceptable” or not. However, the definition of "acceptability" may not be part of a risk assessment but could be pre-established, for example, in thresholds included in government policies or in the mandate given to the risk assessor. Likewise, the final decision on whether to approve (with or without conditions) or prohibit the specific use of the LMO is taken during the decision-making process, which may take into account, depending of the national regulatory framework and among other things, government policies, public opinion, anticipated benefits, costs of the risk management measures and socio-economic considerations.

In addition to the issues mentioned above, the recommendations section of the report may also include any relevant information to be considered by the decision makers prior to making a decision. Some issues that may be relevant include:

1. A recommendation as to whether or not one or more risk management or monitoring strategies should be implemented and, if so, the specific conditions for each such strategy;
2. Considerations of remaining uncertainties; and
3. A recommendation as to if and when the risk assessment should be re-visited.

References

1. Abrahamson WG (website) The Solidago Eurosta Gall Homepage – A Resource for Teaching and Research. Ecology and Evolution. Available at http://www.facstaff.bucknell.edu/abrahmsn/solidago/gallresearch.html (access June 2010).

2. CFIA (website) Detection and Identification Method Criteria. Canadian Food Inspection Agency (CFIA). Available at http://www.inspection.gc.ca/english/plaveg/bio/detecte.shtml (access May 2010).

3. Commission for the European Communities (2000) Communication from the Commission on the precautionary principle. Commission of the European Communities, Brussels, 29 pp. Available at http://ec.europa.eu/dgs/health_consumer/library/pub/pub07_en.pdf (access May 2010).

4. EEA (1998) Environmental Risk Assessment - Approaches, Experiences and Information Sources. Environmental issue report No 4, European Environmental Agency (EEA). Available at http://www.eea.europa.eu/publications/GH-07-97-595-EN-C2/riskindex.html (access July 2010).

5. EFSA (2006) Opinion of the Scientific Panel on Genetically Modified Organisms on the Post Market Environmental Monitoring (PMEM) of genetically modified plants (Question No EFSA-Q-2004-061). European Food Safety Authority (EFSA). Available at http://www.efsa.europa.eu/en/efsajournal/doc/gmo_op_ej319_pmem_en,0.pdf (access June 2010).

6. EFSA (2011) Guidance document on Selection of Comparators for the Risk Assessment of GM Plants. EFSA Panel on Genetically Modified Organisms (GMO), EFSA Journal 9(5): 2149 doi:10.2903/j.efsa.2011.2149. Available at www.efsa.europa.eu/efsajournal.htm (access December 2013).

7. ERMA NZ (1998) Annotated methodology for the consideration of applications for hazardous substances and new organisms under the HSNO Act 1996. Environmental Risk Management Authority of New Zealand (ERMA NZ), 30 pp. Available at http://www.epa.govt.nz/Publications/Annotated-Methodology.pdf (access August 2012).

8. FAO (2001) Genetically modified organisms, consumers, food safety and the environment. Food and Agriculture Organization (FAO), Rome, 35 pp. Available at http://bch.cbd.int/database/record-v4.shtml?documentid=41952 (access May 2010).

9. FAO (2004) The State of Food and Agriculture: 2003-2004. Part I: Agricultural biotechnology: meeting the needs of the poor? Section B: The evidence so far. Food and Agriculture Organization of the United Nations. Available at http://www.fao.org/docrep/006/Y5160E/y5160e00.htm (access May 2010).

10. FAO (2011a) Procedures for post-border weed risk management. Second Edition. Plant Production and Protection Division, Food and Agriculture Organization of the United Nations, Rome, 2011. Available at http://www.fao.org/fileadmin/templates/agphome/documents/Biodiversity-pollination/Weeds/Docs/ProceduresPostBorderWeedRiskManagement.pdf (accessed December 2013).

11. FAO (2011b) Biosafety Resource Book. Food and Agriculture Organization of the United Nations (FAO), Module C, 90pp. Available at http://bch.cbd.int/database/record-v4.shtml?documentid=102000 (access September 2011).

12. FAO/WHO (2001) Evaluation of Allergenicity of Genetically Modified Foods. Food and Agriculture Organization of the United Nations (FAO) and World Health Organization (WHO), 29 pages. Available at http://bch.cbd.int/database/record-v4.shtml?documentid=41976 (access June 2010).

13. GMAC Singapore (website) Genetically Modified Organisms. Genetic Modification Advisory Committee (GMAC) Singapore. Available at http://www.gmac.gov.sg/Index_FAQs_Genetically_Modified_Organisms.html (access August 2010).

14. Health Canada (2006) Guidelines for the Safety Assessment of Novel Foods. Food Directorate Health Products and Food Branch, Health Canada, June, 2006. Available at http://bch.cbd.int/database/record-v4.shtml?documentid=101221 (access June 2010).

15. Heinemann JA (2007) A typology of the effects of (trans)gene flow on the conservation and sustainable use of genetic resources. Food and Agriculture Organization (FAO), Background Study Paper no. 35 rev. 1, 100 pp. Available at http://bch.cbd.int/database/record.shtml?documentid=48542 (access August 2012).

16. Hill RA (2005) Conceptualizing risk assessment methodology for genetically modified organisms. Environ. Biosafety Res. 4: 67-70. Available at http://bch.cbd.int/database/record-v4.shtml?documentid=41660 (access June 2010).

17. Hokanson K, Quemada H (2009) Improving risk assessment – problem formulation and tiered testing. Presented at SEARCA Agriculture and Development Seminar Series, 28 April 2009. Available at http://bch.cbd.int/database/record-v4.shtml?documentid=101212 (access June 2010).

18. ICGEB (website) International Centre for Genetic Engineering and Biotechnology. Available at http://www.icgeb.org/~bsafesrv/introduction/generalintro.html (access October 2013).

19. IUCN (2003) An Explanatory Guide to the Cartagena Protocol on Biosafety. Available at http://bch.cbd.int/database/record-v4.shtml?documentid=41476 (access June 2010).

20. Ministry of Environment and Energy Denmark (1999) Ecological Risk Assessment of Genetically Modified Higher Plants (GMHP) – Identification of Data Needs. Ministry of Environment and Energy, National Environmental Research Institute (NERI) Denmark. NERI Technical Report, No. 303, 35 pp. Available at http://www2.dmu.dk/1_viden/2_Publikationer/3_fagrapporter/rapporter/fr303.pdf (access July 2010).

21. OECD (2006) Guidance for the Designation of a Unique Identifier for Transgenic Plants. Available at http://bch.cbd.int/database/record-v4.shtml?documentid=101186 (access June 2010).

22. OGTR (2013) Risk Analysis Framework 2013, Australia. Office of the Gene technology Regulator, Department of Health and Aging, Australian Government. Available at http://www.ogtr.gov.au/internet/ogtr/publishing.nsf/content/42D3AAD51452D5ECCA2574550015E69F/$File/raffinal5_2.pdf (accessed December 2013).

23. SCBD (2012) Guidance on Risk Assessment of Living Modified Organisms. Final Report of the Ad Hoc Technical Expert Group on Risk Assessment and Risk Management under the Cartagena Protocol on Biosafety, UNEP/CBD/BS/AHTEG-RA&RM/4/6. Available at http://bch.cbd.int/protocol/meetings/documents.shtml?eventid=5037 (access August 2012).

24. The European Parliament and the Council of the European Union (2001) Directive 2001/18/EC of the European Parliament and of the Council of 12 March 2001 on the deliberate release into the environment of genetically modified organisms and repealing Council Directive 90/220/EEC. Available at http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32001L0018:EN:HTML (access June 2010).

25. Underwood E, Poláková J, Berman S, Dooley E, Frelih-Larsen A, Kretschmer B, Maxted N, McConville AJ, Naumann S, Sarteel M, Tostivint C, Tucker GM, van der Grijp NM (2013) Technology options for feeding 10 billion people. Climate change and agriculture; biodiversity and agriculture. Report prepared for the STOA Panel of the European Parliament. Contract IP/A/STOA/FWC/2008¬096/LOT3/C1/SC5. Institute for European Environmental Policy, BIO Intelligence Service, Ecologic Institute, IVM, Brussels/London. Available at http://www.europarl.europa.eu/RegData/etudes/etudes/join/2013/513514/IPOL-JOIN_ET(2013)513514(ANN01)_EN.pdf (access December 2013).

26. UNEP (1995) International Technical Guidelines for Safety in Biotechnology. United Nations Environment Programme (UNEP). Available at http://bch.cbd.int/database/record-v4.shtml?documentid=42114 (access June 2010).

27. UNEP Division of Technology, Industry and Economics (website) Technical Workbook on Environmental Management Tools for Decision Analysis. Available at http://www.unep.or.jp/ietc/publications/techpublications/techpub-14/1-EnRA3.asp (access June 2010).

28. UNEP/IPCS (1994) Training module No. 3. Section C – Ecological risk assessment. United Nations Environment Programme (UNEP) / International Programme on Chemical Safety (IPCS), pp 177-222. Available at http://www.chem.unep.ch/irptc/Publications/riskasse/C2text.pdf (access July 2010).

29. US Environmental Protection Agency (1998) Guidelines for Ecological Risk Assessment. EPA/630/R-95/002F. Available at http://oaspub.epa.gov/eims/eimscomm.getfile?p_download_id=36512 (access June 2010).

30. WHO (2004) IPCS Risk Assessment Terminology. Part 2: IPCS Glossary of Key Exposure Assessment Terminology. World Health Organization (WHO). Available at http://www.inchem.org/documents/harmproj/harmproj/harmproj1.pdf (access June 2010).

---