7. What should be the framework for nanomaterials risk assessment?
A suitable framework for the assessment of all engineered nanomaterials requires exposure and hazard data on a wide range of products. At present there have been insufficient published studies to establish a detailed framework. Nonetheless in the previous CSRSEN opinion an outline for such a framework, in the form of an algorithm, was presented (CSRSEN 2007a). This framework remains appropriate although a few further details can be added in the light of recent publications.
Relevant physicochemical properties
The most important properties of a nanomaterial to characterise, from a risk assessment viewpoint, are:
- Size and size distribution of free particles and fibres/rods/tubes. These may be produced during the manufacture, use (including wear) and/or disposal/recycling of the nanoproduct.
- Specific surface area
- Stability in relevant media (including the ability to aggregate and disaggregate)
- Surface adsorption properties
- Water solubility
In addition, suitable measurements of chemical reactivity are needed although at present the most relevant ones for a particular nanoparticle/nanofibre are best judged on a case by case basis bearing in mind the likely applications of the product (see read across).
Depending on the nature of the nanoparticle/nanofibre it may also be appropriate to consider:
- Photoactivation. Recent data have indicated that some nanoparticles may, by virtue of their relatively large surface area and reactive potential, become activated by light. This is relevant both in considering their stability and their potential to be photo activated when in contact with the skin/external surfaces of other species.
- Potential to generate active oxygen. Production of active oxygen is one accepted general mechanism for the adverse effects of nanoparticle/nanofibre. Thus the in vitro measurement of the ability of a particular nanoparticle/nanofibre to generate active oxygen species may be considered.
There is insufficient information to identify opportunities for read-across based on the general chemical composition of nanomaterials. Nonetheless there are some properties for which read-across is appropriate in determining the experimental studies that need to be considered.
i) Fibres, rods and tubes. In the light of experience with asbestos, and the recent studies on carbon nanotubes (reported above) if there is a potential exposure to free fibres, rods and tubes that are chemically/ biologically persistent, are rigid and have a high aspect ratio (i.e. micrometres in length and nanometres in diameter) the possibility should be considered that they may have similar properties to asbestos.
ii) Particles. There is a large amount of data on airborne fine particles generated as a result of combustion to indicate that comparable particles may cause respiratory and cardiovascular effects. Situations in which there is production of fine nanoparticle/nanofibre with reactive surfaces could cause comparable effects.
iii) Nanomaterials of comparable dimensions and surface properties.The database for extrapolation is very limited, however, for the assessment of a specific nanomaterial it may be possible to highlight relevant properties that require particular assessment.
iv) Bulk material. For any nanomaterial for which significant exposure of man or other species could occur, it is appropriate to consider the toxicological/ ecotoxicological properties of the material in other physical forms, unless there is good evidence that no bulk material will be released in biological systems.
Development of the risk assessment framework
In the previous opinion of the CSRSEN (2007a) a four tier algorithm was presented in which the initial consideration was the potential for exposure of man and/or other environmental species to the nanomaterial. There are no new data that would suggest a significant change in the CSRSEN exposure driven framework to be appropriate, other than the aspects discussed above.
This four stage algorithm offered a framework for the case by case evaluation of the potential risks due to exposure of humans and other species to nanomaterials. The algorithm is exposure–driven. To be of practical value a thorough assessment of the potential exposure of humans and other environmental species during the entire life cycle is vital. This must include not only the current use but also possible further applications. In addition, it must take into account the potential for nanomaterials to be released during use (e.g. as a consequence of wear and tear), and the range of end-use fates, that may occur (e.g. waste disposal or waste recycling options).
It is anticipated that as the scientific knowledge improves, it may be possible to classify nanomaterials into specific risk categories that might become the subject of category specific risk assessments. However, at present this categorisation is not possible. New significant developments have come to light (as identified above), which allows further work on Stage III (Hazard identification and characterisation). These include:
i) Cell/tissue uptake tests;
ii) Bioaccumulation tests to assess prolonged exposure;
iii) Selection of a test system(s) for nanofibers/tubes that are biopersistent, are rigid and have a high aspect ratio (HAR – L>20μm); \
iv) Ability of nanomaterials to trigger one or more of the putative mechanisms of toxicity (e.g. generation of reactive oxygen species).
Addressing deficiencies in the data base
As discussed above, one of the major challenges for nanotechnologies (and indeed for other emerging issues) is how to characterise the risks where the database is very limited.
The traditional approach to address such a situation is to adopt a traditional risk assessment framework and either:
– introduce a default value for each major data gap
– compare the new/potential nano product with the conventional (non-nano) product
However, other approaches to the risk assessment of nanomatertials exist. There have been several recent reviews of the emerging health issues from products of nanotechnologies (Hannah and Thompson 2008, Hoyt and Mason 2008, Linkov et al. 2007, Sweet and Strohm 2006, Wardak et al. 2008,). Two approaches for coping with the large gaps in the data have emerged from these publications:
– Application of lifecycle methodology currently used to evaluate sustainability that has a lesser data requirement.
Sweet and Strohm (2006) have proposed a structured approach that combines risk assessment and risk management approaches viz:
– Is it likely that the product system contributes to actual harm in the life cycle? – How much does each product or stage contribute?
– Do relevant toxicity data exist for risk assessment?
– What is the potential upstream and downstream technology units impacted?
– What opportunities are available for upstream or downstream improvements (e.g. environmental quality, emission reductions)?
– What opportunities to control the risks by selecting less risky options or by restricting access to the hazard or life cycle stage of concern?
– Use of expert judgment to fill the critical gaps.
One interesting aspect of this approach is its use to identify measures that should be taken for containment or environmental control in the workplace. The approach uses a matrix to characterise the level of concern and the consequent action that should be taken (See Table 2). It involves a two dimensional matrix of the likelihood of exposure and the likelihood of effects and severity. With modification it could be applied to identify the potential impacts of individual nanomaterials on human health and on other species along the following lines: Table 2: Expert Judgement matrix
[Insert Table 2 – not accessible from PDF] Such a methodology could prove useful as a first step, for example, in achieving consensus among experts and risk managers on priorities. It may also be of value in a matrix or a graphic form to compare products/ exposure situations in a transparent manner.
Von Gleich et al. (2008) have proposed that the assessment of the potential impacts of nanomaterials should run closely in parallel with the research and technical work to develop them. They refer to this process as ‘leitbuilder’. This is a logical concept that is worthy of further elucidation. A more structured way of selecting which nanoproducts to be developed has been proposed by Linkov et al. (2007) using multi-criteria decision analysis. Their approach is that, prior to comparison of individual nanomaterials; expert judgment is used to set weighted values for relevant parameters of health and ecological effects (e.g. public health effects, effects related to occupational exposure, environmental effects), importance to society and stakeholder preference. The challenge with such an approach is to set the most appropriate weighting, in a manner that is both transparent and acceptable to all the key stakeholders.
The Swiss Federal Office of Public Health (2008) has proposed a structured scoring system for the categorisation of risks posed by nanomaterials into two classes A and B. The approach is compatible with the above approach. The key parameters incorporated into this system are:
- Exposure of human beings / release into the environment
- Potential effects (e.g. stability in biological systems (biopersistence), redox activity)
- Nano-relevance (e.g. physico-chemical properties) and
- Information about the life-cycle (e.g. is the future life-cycle known?).
The approach is sound but needs to be evaluated using actual case histories. One possible concern is the use of a single total numerical score to assign nanomaterials as either low risk or possible risk.
What was the Committee’s conclusion on risk assessment?
The development of a widely accepted and robust methodology that would be used at the R&D stages to identify and mitigate potential human health (including occupational health) and environmental risks, associated with individual nanomaterials should be given high priority. For this purpose it is vital to develop a data bank of case histories to assess its validity.