A. How has Risk Assessment Evolved?
Although the overall process of dividing risk assessment into operational steps has been altered to address the nature of the substances or processes being evaluated, the fundamental components of the risk assessment process have remained relatively constant. Thus, for any particular etiologic (causative) agent or process,
- the universe of potential outcomes that may be causally associated with exposure are identified and characterized;
- the relationships between exposure and outcome are described;
- estimates of potential exposure are made; and then,
- the qualitative and quantitative (when available) components are integrated into an estimate of the likelihood of the potential outcomes to occur given that exposure also occurs.
Because information for decision-making is often incomplete, risk characterization also must take into account the degree of uncertainty associated with any of the steps in the overall process, as well as the cumulative contribution(s) that such uncertainties may make to the overall risk estimate.
At various times, the National Academy of Sciences (NAS) has attempted to describe risk analysis in different ways (Table A-1). The 1983 NAS report “Risk Assessment in the Federal Government,” first attempted to consolidate the risk assessment procedures practiced in the US regulatory agencies (primarily FDA’s Bureau of Foods, which subsequently became the Center for Food Safety and Applied Nutrition) into four coherent steps. At that time, these steps were appropriate to the nature of the substances on which risk assessments were performed e.g., radiation and chemical carcinogens.
Chief among the shared characteristics of these substances was the ability to describe dose in discrete units, allowing for the relative precision of exposure and dose-response estimates. By the time of the publication of the NAS’s 2002 report “Animal Biotechnology: Science-Based Concerns” (NAS 2002b), the description of the risk assessment process had evolved to be more accurately suited for the potential risks associated with animal biotechnology. The most important differences reflect the change of etiologic agents from radiation and chemicals to biological agents or processes. These differences are most obviously manifested in the hazard assessment and dose-response sections, where the range of potential adverse outcomes (harms) can differ in kind from radiation and chemical damage, and the concept of dose must accommodate biological potential. Biological potential can be thought of as the ability for the substance or organism being evaluated to either grow, replicate, die, or perform a catalytic function so that dose is no longer a constant (or possibly decreasing) amount.
Table A-1: Risk Analysis Steps as Described by the National Academy of Sciences
|1983 “Red Book”||2002 Animal Biotechnology Report|
B. Thinking About Risk
Qualitatively, risk may be thought of as some function of the combination of exposure and the intrinsic properties of the substance or process under consideration by linking an exposure to the likelihood of an outcome. The “risk equation” was first derived for the condition of carcinogen exposure and written as:
Risk = (exposure) x (potency)
where potency was estimated from an evaluation of the relationship between exposure and outcome (i.e., the dose-response evaluation). More generally, however, the risk equation is best thought of as some function of exposure and some function of the biological properties of the agent causing the outcome:
Risk ? ƒoutcome (exposure, hazard)
In cancer risk assessment, the function of outcomes was often referred to as the “cancer potency” and was derived from the slope of the dose-response curve for tumor formation. For animal cloning, outcomes may be thought of as the adverse health effects resulting from cloning such as Large Offspring Syndrome, or for edible products of clones, a lack of expected nutritional content of milk from animal clones.
Thinking about risk from the perspective of an “equation” is useful, even when performing qualitative analyses, because it allows the equation to be “solved” for any of the variables that have been defined. Often we ask the “forward” or prospective question: given that some process or exposure has occurred, what is the likelihood of a particular outcome (e.g., how likely is exposure to a particular contaminant in milk to cause gastrointestinal distress?). Alternatively, the question can be asked in the “backwards” or retrograde form: given that an outcome has occurred, what etiologic agent under which exposure conditions is responsible for that outcome (e.g., given gastrointestinal distress, did consumption of milk contaminated with x amount of y substance cause that effect? or how much of x do you have to consume before gastrointestinal distress is experienced?).
When performing a risk analysis, it is critically important to distinguish between a hazard and the potential risk(s) that may result from exposure. A hazard can be defined as an act or phenomenon that has the potential to produce an adverse outcome, injury, or some sort of loss or detriment. These are sometimes referred to as harms, and are often identified under laboratory conditions designed to maximize the opportunity to detect adverse outcomes. Thus, such observational summaries are often referred to as “hazard identification” or “hazard characterization.” Risk, as previously discussed, is the conditional probability that estimates the probability of harm given that exposure has occurred. In a qualitative assessment such as this, however, risks can be discussed only within a qualitative context, and no quantitative interpretations should be made.
Another important question to consider is who experiences the risk. At its inception, risk assessment tended to be anthropomorphic; all risks were evaluated in the human sphere, and were expressed in units of the individual, that is, the probability of a person being exposed to a hazard and experiencing a harm over a lifetime. That individual was defined as the receptor. Human risks could also be expressed at the population level, or the probability of x individuals in the population experiencing the harm. For animal cloning issues, the receptor can be considered to be the surrogate dam carrying a fetal clone, the animal clone itself, or humans or other animals consuming edible products of clones (e.g., milk and meat).
C. How Do We Think About Safety?
For purposes of the Draft Risk Assessment, safety may be best thought of as the condition under which risks would be considered unlikely, rather than the condition of no risk (as such conditions do not exist for any scenario). It implies that a risk analysis has been performed, and the “risk equation” is solved for the condition that Risk ? 0 (i.e., the conditions under which risk approaches zero). When considering food from animal clones, this risk assessment has approached the issue of safety from a comparative perspective. Because one of the basic questions that the food consumption portion of this risk assessment asks is whether animal clones are materially different from their conventional counterparts, the risk question that is asked is whether edible products from animal clones or their progeny pose an increased risk relative to the same products from conventional comparators. Likewise, for animal safety, the question that is asked is whether animals involved in the cloning process are at greater risk for any adverse outcome relative to other assisted reproductive technologies.
One of the difficulties with any safety assessment is “proving the negative.” Because in practice the universe of conditions under which some risk may be encountered cannot be explored, there are always some conditions under which the null hypothesis (i.e., exposure to y µg/liter of Substance X will pose no significant risk) will not be disproved. Thus, a careful risk/safety assessment defines the boundaries of its investigation and expresses its conclusions within those particular limits (i.e., clones born after a carefully monitored pregnancy under closely supervised conditions are at a slightly increased risk of dying than animals derived via in vitro fertilization, or artificial insemination, or, for food safety, milk from dairy cow clones that meets existing regulatory standards and is not significantly different from Grade A bulk tank milk is as safe to drink as milk meeting existing regulatory standards from Grade A bulk tank milk derived from non-clone dairy cows).