A discussion panel was held after each of the three main plenary sessions to (i) understand the different stakeholder perspectives of persistence (ii) highlight innovation, best practice and challenges in assigning persistency at the screening stage, and (iii) share experiences in higher tiered assessments of persistence.
Due to the number of common themes that were raised across these discussions, the issues discussed in the three sessions have been collated around a number of key themes. These issues are outlined in more detail below. Many of these issues are explored in more detail and given further consideration in the syndicate sessions described later within this workshop report.
Some questions highlighted problems associated with variability in the way REACH registrants report persistence and other data, and asked what compliance or checking mechanisms in REACH can be used to deal with this issue. It was also unclear what the proportion of chemical substances that require simulation or higher tier testing at tonnages > 10 tpa under REACH was and the relative weight given to conflicting screening and higher tiered studies. Other questions focused on the different data requirements required under different regulatory frameworks and the potential for a given chemical to be classified differently within different regulatory schemes. Finally, one attendee questioned whether the data contained within the publically available REACH database was sufficient to conclude on PBT properties, and whether the current REACH instruments were fit for purpose. On the question whether data requirements may be insufficient for persistence assessment, ECHA responded that the understanding of the quality of reporting and of the use of data tends to differ a lot between registrants and authorities.
Validation and reference chemicals
The CEFIC-funded ECO 12 report was considered useful. The inclusion of the ECO 12 chemicals in any future method development and guideline development was seen as important. Testing some of these reference chemicals within the OECD 314 studies in order to increase stakeholder confidence in lab-to-field interpretation from these studies was also seen as a priority.
However, it was noted that the reference chemicals from different persistence bins do not provide any mechanistic information to help predict whether chemicals will be degradable / persistent. It was unclear whether the biodegradability of ECO 12 chemicals could be explained by a mechanistic understanding, e.g. of growth, co-metabolism or the biochemical pathways for biodegradation. It was agreed that a more mechanistic approach to identifying an appropriate validation or reference chemical list could assist the development of next generation QSBRs and structure / degradation rule bases. It was noted that the latest version of Biowin is based on empirical data rather than expert knowledge of degradation pathways.
Enhanced screening studies for persistence
It was suggested that the current RBTs were a ‘bacterial lottery’ and there was a critical need to improve understanding on a number of factors including what concentrations of chemicals should be tested and how to obtain an appropriate level of diversity of bacteria in the inoculum used. There was also some consensus that enhancements to existing RBTs were required to improve persistence assessments but there were different views about the extent to which some of these enhancements could be made. Some participants were keen to point out that the fact that biodegradability can occur is very positive and an indication that a substance is unlikely to persist. However, the regulatory concern was that if tests were ‘over enhanced’ these would underestimate persistence in certain environments (i.e. poorly degradable substances easily passing enhanced RBTs and then being found in environmental monitoring programmes such as those conducted under the auspices of the Water Framework Directive) (EC, 2000).
A participant stated that the discussion was mainly based on deriving a relevant inoculum size sufficient to ensure that it would include relevant degraders from the environment. A member of the panel said that the number of microorganisms is in the order of 10 6 per ml of water, 10 9 per gram of soil (Whitman et al. 1998), and 10 9 per ml activated sludge (Goodfellow et al, 1996). The issue is finding a better screening test for persistence, since in his opinion the available ready tests are tests of inocula capability, not substance biodegradability.
It was also questioned whether the Oslo/Paris Convention (for the Protection of the Marine Environment of the North-East Atlantic) (OSPAR) marine degradation test was a good candidate for targeted field flow fractionation. The response was that it would be; however the variability in marine inocula is high and test durations need to exceed 60 days for many positive reference chemicals. This prompted some discussion about how realistic the biodegradation test durations are with respect to the lack of reference conditions. In response to that, it was indicated that ideally one needs to know the rules that govern inocula behaviour, certainly the relationship between diversity and degradation and the kinetics involved.
It was asked, from a pragmatic point of view, at what point testing should stop and when it is decided that no degradation is going to occur. A member of the panel replied that they would try one of the inherent biodegradation tests (OECD 302 tests) (OECD, 1981a,b; 1992a) for several months, then a closed bottle test and if that does not provide any evidence of primary degradation it would be sufficient to conclude that the chemical is persistent.
A comment was made on the need to see some form of comparison of new enhanced tests with existing ready tests (OECD 301) (OECD, 1992a-h),. It was also mentioned that microbial density and endpoint (e.g. time) need to go hand-in-hand as it becomes a kinetic issue for simulation type tests. The response was that we need to use reference chemicals, with an established range of persistency profiles (such as those identified in ECO 12), against a range of inoculum densities that includes those encountered within the existing ready tests (OECD 301) and (OECD, 310) (OECD, 2006). The work presented by Russell Davenport demonstrated that slight increases in inoculum density reduced variation between studies and increased confidence in the persistency conclusion.
There was also some discussion on the value of microbial profiling which, coupled with information on substance composition, could provide a greater insight to the potential persistence of substances. Currently, there is limited experience with these tools but their value could be enhanced as metagenomic libraries develop and grow and more is known about the catabolic potential of microbial communities.
Use of higher tiered tests in persistence assessments
Regulatory participants indicated that they would not accept the use of wastewater treatment plant (WWTP) simulation studies to simulate degradation in the environment and were clear that the ECHA and Member States have rejected the OECD 314 and OECD 303 tests. Their justification is that:
- the inoculum is derived from WWTPs and not the environment;
- there are concerns that not all chemicals, including down the drain chemicals, pass through WWTPs prior to entering the natural environment;
- there is a large variability in the WWTP removal efficiency for micropollutants depending amongst others on the treatment technology, operation mode and weather conditions;
- the duration of the OECD 314B study (OECD, 2008b), which at 28 days, is substantially longer than typical hydraulic sludge retention times in full scale WWTPs and this could lead to overly favourable assessment of degradation in the OECD 314B.
The arguments against accepting the OECD 314 test were challenged by some participants who re-emphasised that there are studies showing that the vast majority of ‘down the drain’ chemicals pass through a WWTP before being discharged into the environment and that experience with the OECD 314B indicates that most of the degradation takes place over the first few days, i.e. within the hydraulic and sludge retention time of most WWTPs. It was emphasised that the additional duration of the OECD 314B was to enable slowly degrading intermediates to be identified. From a scientific perspective it was pointed out that WWTPs contain natural assemblages of bacteria and other biota that are also present in the receiving waters. As such the tests assess whether substances will have potential to be degraded or will persist in the environment.
Furthermore, it was questioned why the regulators would accept the OECD 301 tests, that also contain inocula derived from sewage, when this was even more unrealistic than the OECD 314B test (which uses more relevant test concentrations, no nutrient solutions, etc.). The response was that a lot more was known about the OECD 301 test and it had more stakeholder confidence. In support of the OECD 314B test, participants indicated that the study is data- and knowledge- rich providing information on elimination pathway, transformation products and degradation kinetics. It was clear that there was a divergent view between industry proponents and regulators. This could potentially be resolved by generating more data on a reference set of chemicals. One other option that was identified within these discussions was that the persistency criteria applied to the OECD 302B (OECD, 1992a) inherent biodegradation test for the length of the lag and log phase could be applied to the OECD 314B.
There was a general concern that the higher tier tests required to support product registrations (e.g. the OECD 308 and 309) (OECD, 2002a, 2004) were expensive to undertake considering the lack of validation and consensus on the value of the data they provide. The discussion then moved on to the OECD 309 test which is being requested by regulatory agencies although there was a consensus that this test had not been subject to extensive validation and there were different levels of relative stakeholder experience of this study versus alternative test systems such as the OECD 314D and 314E (OECD, 2008c,d).
Overall persistence (P OV )
The question why overall persistence (P OV ) was not used in P assessment was asked. P ov describes the average lifetime of a chemical in the environment. It therefore encapsulates the net effect of the reactivity of a chemical in individual media (usually in terms of single-media half-lives), residence time, partitioning behaviour and the distribution of emissions (ECETOC, 2003). A response was made that P OV depends very much on use profile (not just partitioning behaviour) so if use pattern changes, P OV might change too. However, it was widely accepted that P OV was useful in identifying the most appropriate, or relevant environmental matrix or compartments, to generate the environmental half-life data. This approach is advocated by ECETOC in its persistence report (ECETOC, 2003). Since the overall persistence is the weighted average of the residence time in all media, it may sometimes be greater than any of the individual medium-specific half-lives for some chemicals (depending on their transport properties). There is an ongoing debate in the scientific community on the best method for using the overall persistence. At this point in time, however, there is no agreement on the criteria that should be applied to the overall persistence.
Degradability versus bioavailability
A question was asked about the relationship between bioavailability and persistence i.e. when is a bound residue available and when is it not? Specific examples (e.g. alkyl benzene sulfonates) identified by the participants considered the impact of sorption and partitioning on biodegradation kinetics e.g. how to deal with an adsorptive readily degradable chemical that is slowly released from sediment / soil, such that half-lives may be longer than P/vP criteria and low level exposure could occur. This was followed by a discussion on how bioavailability affects persistence assessment, for instance sorptive chemicals which are slowly released.
This raised the issue about PBT risk assessment because the substance may be categorised as persistent but risks may be mitigated if it is released slowly and once released it is then biodegraded. As a result of these discussions, several participants highlighted the need for a more holistic approach that considers bioavailability (to microorganisms and higher organisms), bioaccumulation and toxicity, rather than focusing attention on a single property such as persistence or environmental half-life. One participant highlighted that the re-suspension of particulates may increase bioavailability of adsorbed chemical materials and the need to account for this phenomenon within test designs.
A member of the group questioned the presenter Charles Eadsforth on the subject of bioavailability and bioaccessibility of NERs. He said that an ISO standard addressing extraction is available, and that this was not covered in his talk. However, this ISO standard was developed with neutral non-polar chemicals in mind and may not be applicable to all classes of chemicals. A specific example highlighted that some antibiotics behave in a different way to agrochemical products in that they tend to stick in the soil but there is evidence that these are mineralised further once in the soil (a figure of 20% or more being degraded within 100 days). It was noted in the discussions that instantaneous formation of large percentages of NER, as seen with various pharmaceuticals, seems to be a different mechanism than slow steady NER formation (Müller et al, 2013).
Given the difficulties in characterising UVCBs, it was suggested that an overall dissipation half-life is a practical measure of the persistence of a UVCB and can be related from laboratory to the field. However, some argue this is an unsatisfactory measure for assessing risks since the ultimate fate of the components are unknown. Furthermore, it was pointed out that there was a legal requirement under REACH to provide information on biodegradation rates. A risk assessment will need to distinguish between loss processes such as leaching, volatilisation or formation of NERs. There are significant concerns with NERs because with respect to legal interpretation this is a ‘black box’ under REACH Annex 13.
Further discussions on standardised extraction schemes to characterise NER continued. Many thought this would be very difficult given the diversity of chemicals and variety of matrices found in the environment. It was suggested that using all the available information, including functional groups present in the test substance, knowledge of binding mechanisms, knowledge of the types of sorbents used and other available information could support the development of ‘extraction models’ useful in defining the best extraction approach for any given chemical. This was identified as a topic for further research.
The use of sterile controls and different extraction procedures may help to differentiate between biodegradation and physical losses and assess potential availability of NERs. This was considered essential for assessing complex substances although some practitioners had experienced problems maintaining the sterility of soils and sediments.
There was discussion on the importance of a ‘thinking phase’ before initiating any higher tier studies. Basically this was to look at possible degradation pathways before starting the study, write down SMILES of potential transformation products and run prediction models to gain an insight on their physico-chemical properties. It should then be possible to optimise extraction scheme and analytical methodology as well as gaining information on predicted toxicity and log K ow of the transformation products. There was consensus that there should be increased use of in silico screening to help focus and adapt the experimental strategy. This was particularly relevant to pharmaceuticals where there should be a lot of potentially relevant information prior to the need to conduct higher tier testing.
Where studies were conducted at 20°C, one participant indicated that temperature should be normalised using the Q 10 value of 2.8 to account for a temperature difference of 10°C (EFSA Journal, 2007). However, it was pointed out that the Arrhenius equation temperature correction is not always applicable to biotic transformation as different environmental species have different temperature optima and that the situation seems to be substance and inoculum-specific. It was also pointed out that some bacteria in the inoculum would die if the experimental incubation temperature exceeded their inherent temperature range. Work to date appears to indicate that some degraders perform better at low temperatures e.g. some microbes have temperature ranges between –15 and 10°C (which are classified as psychrophiles) and could be excluded from higher tiered studies at 20°C. Whilst no agreement could be made about routinely normalizing studies with respect to temperature, it was agreed that more realism may be provided by testing at environmentally relevant temperatures or the temperature at which the inocula were collected from the environment. One participant disagreed because if the tests were not made at standardised temperatures like 20°C the outcome of the studies will not be comparable. However, it was also noted that as different inocula are used for different studies then they will never be comparable and truly standardised.
In response to the presentation from Kees van Ginkel, one participant asked a question about how relevant laboratory observations showing that growth-linked biodegradation can occur are to environmental conditions where chemical concentrations are too low to support microbial growth or that the relative abundance of the competent microbial population may be too low to measure removal. The presenter, Kees van Ginkel, responded that in his studies he added some glucose into test systems not to promote cometabolism but to stimulate inocula growth (including competent degraders). This provides some realism where continuous nutrient input is occurring in natural habitats. In combination with similar substances, levels of degradation can increase in the environment through growth-linked degradation, not always through cometabolism. Another response suggested that we need to bear in mind the complexity of the community where some microorganisms may be cometabolising in the environment while others grow as they degrade substances.
Several of the talks had referred to the importance of photodegradation (either direct or indirect) in aquatic persistence testing, with one talk describing modified simulation studies that included a defined photoperiod that resulted in shorter dissipation rates. In his talk, Robin Oliver had described aquatic simulation studies in which algae had been included in an attempt to more realistically mimic environmentally relevant conditions. It was thought that generation of hydroxyl and other oxygen radical species by virtue of algal presence might contribute to the increased degradation seen in these studies. A discussion focused on the relevance for the environment of photodegradation in aquatic environments; where suspended matter, depth and shading will all decrease the intensity of light available for photolysis. A comment was raised that the issue of photodegradation was irrelevant for groundwater. In the discussions a recent report (Jiménez and van de Meent, 2011) on the relevance of direct and indirect photolysis in aquatic environments by RIVM was referred to, in which the conclusion had been drawn that photolysis was unlikely to contribute to rates of degradation compared to other abiotic and biotic mechanisms. However, some other members of the workshop disagreed with this view, citing that primary productivity by algae and cyanobacteria in oceans was global and occurs at depths of up to 200m. No consensus was reached with respect to the relevance of light in persistence assessments in aquatic environments.