Syndicate discussions
Four parallel syndicate sessions were held in each of the two afternoon sessions to discuss specific challenges associated with degradation and persistence assessments. The Organising Committee identified a number of key questions for members of the four syndicate groups to address in each afternoon session. The four syndicates discussed questions under the following themes: (i) the challenges with the persistence assessment of difficult to test substances, (ii) improved screening approaches for persistence assessment, (iii) interpretation of higher tiered studies, and (iv) enhanced realism within persistence assessment. A plenary was held at the end of each of the syndicate sessions. Each rapporteur of the four syndicate groups provided feedback to all participants; an opportunity to clarify any of the points that were raised and for open discussion was also provided.
Syndicate 1: Challenges with the persistence assessment of difficult-to-test substances
Moderator: Graham Whale
Rapporteur: Daniel Merckel
Rolf-Alexander Düring
Charles Eadsforth *
Malyka Galay Burgos
Bruno Hubesch
Anu Kapanen
Georg Kreutzer
Jacques l'Haridon *
Marie-Chantal Huet *
Laurence Libelo *
Dan Salvito
Markus Seyfried
Kees van Ginkel
* Participated in Syndicate session 1B.
1. When should higher tiered tests on complex substances be waived?
2. If degradation data is required, under what circumstances can this be provided on the basis of predicted data for representative molecules?
3. From an analytical perspective, what should be measured and how?
4. Can endpoints be based on fate of whole substance as well as components?
5. How should complex substances be dosed into the test bearing in mind potential differences in physico-chemical properties of components?
6. Transformation products and non-extractable residues – do they matter?
Syndicate session report
The syndicate began with a tour de table on what participants understood by ‘difficult-to-test substances’. The responses can be grouped under two headings (with potential crossover): (i) substances with difficulties mainly around their characterisation and analysis, and (ii) substances whose properties mean that practical aspects of testing are difficult. Examples given were petroleum-like products (UVCB examples, i.e. not deliberate mixtures, with components with various physico-chemical properties), natural complex substances (with compositions that can vary each year according to different natural product crops), single substances that are poorly soluble and/or volatile, amphiphilics, and other substances for which analysis / characterisation is difficult (e.g. substances subject to major interferences from background contamination).
1. When should higher tiered tests on complex substances be waived?
The group thought that no general conclusion was possible on situations where higher tier testing could be waived; it is dependent on specific examples. For complex substances, use of read across approaches between components (interpolation and extrapolation) were thought to be important in cases where ready biodegradability could be used as the basis for waiving higher tier testing when ready biodegradability test (RBT) data were available on the substance itself or major components of the substance. The group recognised that in such cases residue analysis was important to check if some components were not degrading in a substance that ‘passed’ a ready test (ideally specific analysis for parent disappearance and metabolite formation – i.e. more detail than secondary measures of biodegradation like CO 2 production and O 2 uptake). This would especially be the case for more poorly characterised materials where structural similarity of components was less certain. One recommendation was a non-target approach to look for both non-degrading components and metabolites. This would require high-resolution mass spectroscopy (MS) combined with ambient ionisation techniques.
There was also some discussion on the role of microbial profiling to give an indication of the presence and growth of different types of degraders to provide weight of evidence but the general consensus was this was not sufficiently developed to be reliable. Another consideration was that increase in biomass could be a useful measure but there was no consensus on the value for complex substances, the concern being that this does not give any information on substances that do not degrade.
2. If degradation data is required, under what circumstances can this be provided on the basis of predicted data for representative molecules? From an analytical perspective, what should be measured and how?
A recurring theme in the group was the need to characterise substances (especially complex ones) as far as possible to allow robust conclusions to be drawn for regulatory purposes. Weight of evidence approaches are important, but the ability to identify the majority or nearly all constituents in a substance was thought essential to drawn meaningful conclusions using read across approaches. The same is true of predictive (QSAR) approaches where modelling is dependent on representative molecular structures (and associated physico-chemical properties). This discussion included approaches like the hydrocarbon block, and the dangers of splitting complex substances into too many sub-groups – delineation of physico-chemical properties and knowing where to place the thresholds. It was recognised that although future developments in analytical techniques may allow more straightforward detection and characterisation of the constituents and degradants within multi-component substances, analysis remains one of the major challenges and is extremely resource-intensive.
The idea to use available REACH registration data to train existing biodegradation QSAR models further was also mooted.
One of the concerns discussed was that experience suggests the ECHA take a precautionary approach regarding read across of data between endpoints. If a similar attitude prevails with read across to components of complex substances this could become a significant issue. An alternative perspective is that Annex 13 allows more freedom than before and the critical point is to build up a robust case. The key point regarding persistence assessments here is how can the reliability and suitability of component data (either measured or predicted) be demonstrated such that it gains regulatory acceptance.
4. Can endpoints be based on fate of whole substance as well as components?
The group discussed the question at what point one stops attempting to identify a complex substance’s constituents and focuses on testing the ‘whole’ substance. Again it was felt no general rules could be elaborated, that a ‘case-by-case’ approach would be needed. The experience of one group member was that a ‘constituent approach’ can be successful for many higher volume (more economically important) products, but that for some substances they were forced to assess the complex substance itself as the number of constituents and structures and properties of the constituents diverged and became unmanageable in a constituent approach. It seemed probable that lower volume (less economically important) substances would be likely to undergo whole substance assessment on cost grounds.
This led onto a discussion on the use of chemical profiling to follow (ready) biodegradation as a pragmatic means to give more information than the usual secondary measures of ready biodegradation, and as a way to inform the hazard assessment (mainly PBT) for the parent substance and degradation products. Ideas put forward included: following a complex substance’s average molecular weight change (i.e. decline) over the course of a test using MS; from a hazard assessment point of view, upon completion of an RBT running a log K ow test (e.g. HPLC or shake flask) on the residue and comparing the log K ow against that of the parent substance to inform on the degradation product’s / residue’s potential for bioaccumulation (i.e. does the resulting mixture’s log K ow fall below the threshold for PBT concern, or, if present, can components in the ‘boundary of concern’ be identified and put forward for further testing?); and use of surrogate testing e.g. change in toxicity over time for discharges from SCAS type systems. It was noted there will be exceptions to the supposition that toxicity of degradants will be lower than the parent (e.g. alkyl phenol ethoxylate degradation).
There was a comment made that the issues encountered with assessing complex substances were in many ways analogous to those encountered in whole effluent assessments (WEAs) developed to assess whether PBT substances were present in effluent discharges. As such, the approach could follow / adapt some of the guidance which has been developed by OSPAR in their practical guide for WEA (OSPAR, 2007). This was considered relevant because the OSPAR WEA approach had considered schemes for not just assessing the persistency of effluent components but the persistence of potentially bioaccumulating and toxic substances. For example, solid-phase micro extraction (SPME) run before and after degradation assessments can give information on the persistence of potentially bioaccumulating substances.
5. How should complex substances be dosed into the test bearing in mind potential differences in physico-chemical properties of components?
One of the problems with the whole substance approach was that this could contain components with considerably different properties which could not only cause problems with the way that the test was set up but also how the results should be interpreted. For example, petroleum products can contain components which are volatile whilst also containing very poorly soluble components. This leads to problems that some components will be lost via volatilisation before others have had any chance to reach equilibrium within the soil / sediment. There was also debate on how poorly water soluble substances should be dosed. From an ecotoxicity perspective the use of solvents for preparation of test media may affect the expected toxicity. Therefore it is likely that similar issues will apply to degradation tests (i.e. solvents may affect the availability of the substance to competent degraders).
6. Transformation products and non-extractable residues – do they matter?
The idea of specifically investigating a modified OECD 301C (“modified MITI I test”), in which degradants are characterised, with conditions more suited to biodegradation (lower substance concentration, more viable inoculum, larger test vessels) was put forward.
Following on from the idea to screen for changes in hydrophobicity following a ready test, one of the group members described a project that is being planned between the Research Institute for Fragrance Materials (RIFM) and University of Stockholm (Michael McLachlan). A wastewater treatment simulation study will be followed by the McLachlan ‘better BCF’ approach (Adolfsson-Erici et al, 2012) for selected (6 – 10) fragrance chemicals, followed by ‘simple’ natural products if successful, thereby linking the assessment of P and B (for degradants / persistent parent constituents). It was suggested that ECETOC could be involved in this project, as it is very relevant for other types of complex substance. Links / co-funding and the possibility of a CEFIC/LRI RfP could be explored.
Another idea that came out of the group was for a project entitled “Methods to characterise the poorly defined fraction of complex materials and assess through a screening approach P, B and T endpoints of these complex materials”. This would involve literature searching and potentially a subsequent ring test of complex mixtures.
Syndicate 2: Improved screening approaches for persistence assessment
Moderator: Jason Snape
Rapporteur: Russell Davenport
François Brillet
Juha Einola
Tom Fisher
Marie-Chantal Huet
Ulrich Jöhncke
Thomas Junker
Anu Kapanen *
Georg Kreutzer *
Laurence Libelo
Kathleen McDonough
Markus Seyfried *
Eric Verbruggen
* Participated in Syndicate session 2B.
1. Is there value in current RBTs in P assessment?
2. What are the views on extending the duration from 28 to 60 day tests (impact of test duration)?
3. What are the opinions on modified tests (how confident are we in the current modified approaches)?
4. Do OECD 314 tests have value in persistence assessments?
5. How to incorporate pre-adaptation in screening-level persistence assessments?
6. How confident are we in enhanced approaches?
7. Can we get more out of inherent biodegradation studies (e.g. MITI II)?
8. How can we reliably assess marine persistence?
9. What are useful endpoints for persistence assessment (primary versus ultimate)?
The order of the questions provided at the workshop was changed and further questions were posed. The questions above reflect those used during the break-out session.
1. Is there value in current RBTs in P assessment?
It was widely agreed that there is still a role for ready biodegradation tests (RBTs) in persistence assessment. This role is limited to the initial screening assessment stage where there is a high degree of confidence that chemicals that pass the RBT will be non-persistent in the environment.
Currently, there are seven RBTs (OECD 301A-F and OECD 310) (OECD, 1992c-h;2006) assessing ready biodegradability using a variety of semi-specific endpoints such as Biochemical Oxygen Demand (BOD), Dissolved Organic Carbon (DOC) removal and carbon dioxide evolution. It was recognised that there was significant room to consolidate these RBTs. In reality, the most frequently used RBTs are the OECD 301B, D and F and the OECD 310. The OECD 301 series of tests that rely on DOC removal (OECD 301 A and E) are not widely used as they are not suitable for adsorbing, volatile or poorly water soluble test substances. Based on some experience of enhanced biodegradation screening under REACH where the test durations had been extended to 60 days there was some concern that the fixed 28 day test duration of the RBT was missing the degradation phase for some chemicals with extended lag periods.
Finally, some technical innovations were also discussed. These included: (i) the availability of new instrumentation that increased analytical sensitivity (ii) the use of tests with combined analytes (e.g. BOD and carbon dioxide evolution), and (iii) the introduction of a test to screen for biodegradability in water-sediment systems (Junker et al, 2010).
2. What are the views on extending the duration from 28 to 60 day tests
The issue of prolonging test duration is already written into REACH technical guidance and in principle it was agreed that this was a good idea but not following any pre-adaptation approaches. It was also highlighted that in some cases lag periods extended to beyond 60 days and so a case could be made to extend the test duration further; this is particularly true for marine biodegradability assessments using OECD 306 and 309 tests (OECD, 1992i;2004a) (see below). The use of these data in persistency assessments would be subject to expert judgement based on the nature of the biodegradation curves and the behaviour of any reference compounds.
3. What are opinions on modified tests?
It is important to recognise the nature of chemicals being tested. It was felt that the current REACH guidance allowing modified tests to assess chemicals with (i) poor water solubility and/or (ii) toxicity overcame some of the limitations associated with the RBTs. These were accepted as a good idea and should be allowed.
4. Do OECD 314 tests have value in P assessment?
There was a difference of opinion between regulators and practitioners with respect to the use of OECD 314 tests, which comprise five separate tests simulating the fate of down-the-drain chemicals in wastewater treatment plants and surface waters. The greatest concern was associated with the use of the OECD 314B within persistence assessments due to its high level of activated sludge solids. One participant highlighted that the OECD 314B was a more up-to-date version of the OCED 302B Zahn-Wellens test and that the test should be able to be used in persistence assessments if the strict criteria applied to the OECD 302B test were applied to the OECD 314B test. Some regulatory stakeholders appear to have more confidence with the OECD 309 Surface Water Transformation test. However, it was recognised that there are few data for both the OECD 309 and 314 tests at present. The OECD 314 series are better described than OECD 309 tests, and it was pointed out that the OECD 309 test also allows the inclusion of more solids and provides an option for some pre-exposure. It was felt that the OECD 314D and E tests may have value in P assessment, but some of the modification in Annex of 309 may be preferable e.g. pre-adaptation. It was agreed that more data was required for both studies to increase the confidence in the studies with regulators and practitioners.
5. How to incorporate pre-adaptation in screening-level P assessments?
There was a good discussion about the advantages and limitations / concerns of pre-adaptation. Most of these discussions focused on (i) concerns over ecological significance of adaptation, (ii) the timescales required for adaptation to occur and (iii) its relevance and importance for different habitats.
The various ways in which adaption or pre-exposure could take place were discussed and it was agreed that there was a need to (i) review and define adaptation (ii) describe what is happening in extended lag periods, (iii) identify relevant experimental adaptation regimes, (iv) validate these approaches with appropriate benchmark chemicals and, (iv) define appropriate pass criteria (e.g. pass / fail or half-lives).
6. How confident are we in enhanced approaches in P assessment?
Following on from the presentations earlier in the day, there was general all round support for the use of (i) extended test durations and (ii) enhanced inocula concentrations but particularly when used in conjunction with a reference validation set of chemicals.
It was felt that progress in this area was such that a large ring test for enhanced tests in comparison with other current screening tests and with a reference set of validation chemicals would be necessary. It was also felt that more validation around adaptation methods was required.
7. Can we get more out of the current inherent biodegradation studies?
Inherent biodegradation studies were designed to establish whether the potential for degradation existed (OECD, 2006). It was agreed within the syndicate that the failure to observe any biodegradation within an Inherent biodegradation study could be used as evidence of environmental persistence. However, there was some concern that where inherent biodegradability studies provide evidence of degradation potential the knowledge generated has limited use under REACH within a persistence assessment. A major limiting factor is associated with the short lag and log period criteria defined within REACH that would not be assessed within the sampling regime of a standard OECD 302B study. The OECD 302A (modified Semi-Continuous Activated Sludge Test; SCAS) cannot be used within persistence and exposure assessments due to concerns over its infinite sludge retention time and strong adaptation potential. However, subtle changes to the 302A and 302B studies addressing test compound concentrations, test durations, sample intervals and the inclusion of a sludge retention time within the SCAS could improve the relevance and reliability of these studies. In part the more recent OECD 314B is a more rigorous and widely applicable test than the OECD 302B test that addresses many of these refinements and can meet the specific criteria laid out under REACH (see question 4 above). It was also recognised that the enhanced biodegradation screening tests, with increased inocula levels, might also provide a more robust alternative to the OECD 302B and OECD 302C tests.
8. How can we reliably assess marine persistence?
Marine persistence is usually assessed by the OECD 306 and/or 309 test guidelines (OECD, 1992i; 2004a). Both these studies have a test duration that ends at day 60. Marine studies presented on Day 1 clearly demonstrated that lag periods for chemicals known to be non-persistent can exceed these test durations and result in a false persistency assignment. In many cases positive reference chemicals (e.g. aniline) can fail marine OECD 309 tests, and lag periods of > 70-80 days can be observed before the onset of rapid degradation even with enhanced marine inocula at 100X cell concentrations. Data also presented on Day 1 also indicated that the greatest sources of variation in inocula were also associated with microbes derived from marine sources (Davenport et al, presentation). The current level of experience and confidence with marine biodegradability assessments remains limited and further research work is required. One option suggested included the use of marine water-sediment systems or a semi-continuous test regime with low level adaptation as described within the Annexes of the OECD 309 test guideline.
9. What are useful endpoints for persistence assessment (primary versus ultimate)?
Whilst persistence assessment is focused predominantly on the parent compound it was widely recognised that the mineralisation endpoint (CO 2 and BOD) can be more robust as it accounts for losses that can be attributed to both parent compound and degradation products. The group did recognise that (i) the mineralisation pass criteria for RBTs should not be assigned to higher-tiered biodegradation studies, (ii) the rates and extents of degradation observed differ for each endpoint (e.g. parent removal, DOC removal, carbon dioxide evolution), and (iii) the levels of degradation observed change with substance and inocula concentrations.
The group also recognised that mineralisation data would not always be available for higher tiered tests. A radiolabel was required and this is not always feasible (e.g. complex mixtures and some chemical moieties). In such circumstances many higher tiered tests will only provide evidence of removal or dissipation and not necessarily biodegradation. However, it was agreed that the more endpoints that you could assess the higher the level of confidence in the study. It was also agreed that every effort should be made to assess degradation or transformation products formed at > 10% of the applied parent or 14 C material (e.g. OECD, 2002b). Within REACH an obligation exists to look at transformation products formed at >0.1% of the applied parent or 14 C material. However, this poses some significant practical issues as most higher tiered biodegradation studies can only resolve 2% of the applied radioactivity due to (i) the specific activity associated with the radiolabelled test material and (ii) dosing near an environmentally relevant concentration.
Recommendations
- The OECD should consider convening an Expert Working Group to consolidate and update the RBTs to reflect (i) the availability of new instrumentation with increased analytical sensitivity (ii) the use of tests with combined analytes (e.g. BOD and carbon dioxide evolution), and (iii) the need to screen for biodegradability in water-sediment systems.
- A laboratory-based study to compare the performance of the OECD 309 and relevant OECD 314 tests, using appropriate benchmark chemicals with known biodegradability / persistency is required to increase industrial and regulatory stakeholder confidence in these studies. There are a growing number of OECD 309 studies being performed for the AIR process of plant protection products. Up to now there is only limited confidence in this test especially with respect to the biomass variation during the experiment. It was also recommended that the enhanced biomass studies could be included as part of this exercise. It was also recommended that this exercise includes a ring test for enhanced biomass studies to compare their performance against other current screening tests with a reference set of validation chemicals.
- Research is conducted to demonstrate the ecological significance of adaptation and get a better understanding of adaptation processes, the probability of these to occur under environmental conditions, including the mechanisms of adaptation and suitable test systems to allow provision for these mechanisms to occur. This approach should also develop appropriate marine studies with a reference set of validation chemicals.
- Test durations should be made more flexible and be allowed to extend to beyond 60 days as marine biodegradation studies often have lengthy lag periods before the onset of degradation.
- It is important to recognise the physico-chemical properties of chemicals being tested (e.g. solubility and volatility) and the limitations of each biodegradation test system. It was recommended that the current REACH guidance allowing modified tests, to assess chemicals with (i) poor water solubility and/or (ii) toxicity, should be allowed as they overcame some of the limitations associated with the existing RBTs.
Syndicate 3: Interpretation of higher tiered studies
Moderator: Caren Rauert (Jon Ericson 3B)
Rapporteur: Jon Ericson (Caren Rauert 3B)
Silvia Berkner
Michiel Claessens
Rolf-Alexander Düring *
Kathrin Fenner
Bernhard Gottesbüren
Silvia Jacobi
Andreas Kaune
Jörg Klasmeier
Arnaud Lagriffoul
Andreas Schäffer
Henrik Tyle
* Participated in Syndicate session 3B.
1. Non-extractable residue (NER) characterisation and interpretation.
2. DT 50 versus DegT 50 .
3. Impact of temperature and normalisation of results.
4. Transformation products (TP) identity and need for Environmental Risk Assessment (ERA) risk.
5. Endpoints for persistence assessment (primary versus ultimate).
6. Dealing with half-life differences between soils and sediments.
1. NER characterisation and interpretation
It was noted early in the syndicate discussions that it would be helpful to have updates, or reports from the two ECETOC Task Forces (TF) (ECETOC, 2013a,b) heading up discussions on the characterisation and assessment of non-extractable residues. It was suggested that a review of these TF activities on the ECETOC web page would be good follow-up for interested parties. In light of this, the following represents the discussion of this syndicate realising that many of the issues may have already been discussed and/or may already have recommendations in place. It was also noted that NER will likely be part of other syndicate discussions at this workshop.
In general, a review of terms was helpful to ensure everyone has the same understanding. Non-extractable residue (NER) has always been an operational definition, defined by the steps taken to extract the residues from the various environmental matrices without destroying the matrix itself or altering the nature of the residue in question. Anything remaining is considered NER.
NER is no more and no less than the fraction that cannot be extracted under the analytical constraints of the method used. This does not include any information on processes or binding mechanisms that may be responsible for the observed ‘loss of extractability’ (compared to the desired 100% yield).
The term bound residue (BR) has sometimes been used synonymously to NER, but this is obviously not useful. The term bound is an indication that this fraction should be considered to only encompass molecules that are irreversibly bound to the matrix. Therefore, BR constitute a sub-could be clarified by defining BR as covalently or irreversibly bound residues (or incorporated into biomass, see below). In this case, BR formation would be an ultimate loss process for the substance in question and contribute to its relief in terms of persistence.
This is not justified, however, for NER. The reason for this statement is two-fold:
a) Since the NER fraction depends on the method used; a substance could be described as an NER based on the use of an inappropriate extraction procedure. Consequently, the methods of extraction require careful scrutiny and judgement.
b) It has been unequivocally demonstrated for some compounds (atrazine, sulfadiazine) that there is the possibility of long-term remobilisation of the parent compound from solid matrices (sediment, soil) on a time scale of years although the extractability of the compounds was low on shorter time scales (high amounts of NER). Obviously, there is a fraction that is not bound strong enough to ensure its ultimate ‘disappearance’ from the system. This sub-fraction of NER should be considered at least in long-term risk assessment.
One opinion was that what may have been characterised as ‘remobilisation’ of parent may be due to a change in redox conditions from aerobic to anaerobic. Anaerobic processes may be responsible for reappearance of the parent (or an active metabolite) that has been either immobilised or biotransformed by aerobic processes - anaerobic conditions do exist in ‘aerobic’ water-sediment test systems. The group felt it is necessary to generally understand remobilisation in order to define to what extent it may be an issue in risk assessment. There is also a need to define suitable (case-specific?) metrics for this risk.
In recent months, UBA has introduced other terms (type 1 and type 2 NER) that result in a similar distinction of NER: Type 1 NER is the fraction that may be remobilised as parent substance (i.e. sulfadiazine) or active metabolite due to environmental events and may thus be considered as a hidden hazard. Type 2 is the NER fraction that is incorporated as parent molecule and/or metabolite into humic substances by covalent binding and thus will not be remobilised. Some syndicate participants felt that the Type 2 NER was the same as what has been previously defined as ‘Bound Residue’. Yet another group additionally introduced type 3 NER, separating those that are (truly degraded and) incorporated into biomass as biogenic residues, e.g. amino acids, phospholipids, etc. (no environmental risk). The further development of extraction schemes for separating type 1 NER and type 2 NER (and type 3 NER) was suggested, although it seems necessary to distinguish different schemes for different substance classes. Another way forward could be the identification and quantification of the fraction of biogenic residues within the NER, which is what Andreas Schäffer et al are working on. (e.g. Kaestner et al, 2013)
It was remarked that instantaneous formation of large percentages of NER, as seen with various pharmaceuticals, seems to be a different mechanism than slow steady NER formation observed for other types of compound (Müller et al, 2013).
The formation of NER leads to the additional problem that allows for a wider scope of interpretation. As discussed above, a significant emphasis has focused on the interpretation of NER in terms of persistence assessment and it is still under debate i.e. should NER be regarded as degraded or as potential reservoir that has to be excluded from the persistence evaluation. Regardless of which interpretation is agreed upon, it is necessary to distinguish the different processes and pathways. In water-sediment studies (OECD 308) where estimating a total system half-life is achievable, the derivation of valid DegT 50 values for the separate water and sediment compartments is often difficult, due to competing processes of the test substance dissipating from the water phase into the sediment, and the test substance forming NER. Separate DegT 50 for the water and the sediment phase can often not be estimated because the models used parameter estimation and include too many unknown parameters and the degrees of freedom becomes too high. By using an inverse modelling approach transformation of the chemical and formation of NER can be mathematically separated. It may be possible to differentiate between the fraction of chemical that is non-extractable, but may be remobilised (equivalent to type I NER, entrapped NER, as proposed by UBA and Kästner et al, 2013) and the fraction that is permanently removed (equivalent to type 2 NER, bound residues), if sufficient data were available. Such an inverse modelling approach has been applied to data from degradation tests in soil for a number of pesticides (Matthies et al, 2008; Loos et al, 2012) resulting in half-lives for total dissipation (DT 50 ) and degradation only (DegT 50 ).
How should additional information on NER, such as differentiation between NER and BR, be used? What implication could that have on risk assessment? Several examples of bioassays were discussed to assess the potential ecotoxicity of NER. Such approaches are needed to better understand the potential risk posed by NER and provide additional tools for risk assessors to enable more definite action with respect to having no risk, or having a risk that requires further follow-up. The concept of a ‘soup’ test has been recommended by the ECETOC TF on assessing NERs where sediment ecotoxicity data is developed in real sediments where NERs are present as well as potential transformation products (ECETOC, 2013b). Such a test would help identify where no risk exists, and/or identify where further follow-up is needed either in terms of the bound residue or potential TPs that may have formed. Also discussed is the assessment of NER found in manure, where soil germination and growth studies are conducted in soil amended with manure.
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
2. DT 50 versus DegT 50
The syndicate discussed the current definitions of DT 50 and DegT 50 . Typically, DT 50 and DegT 50 are determined from the remaining concentrations of extractable substance in the test system. DT 50 is usually defined as the time for 50% of the substance to dissipate (i.e. by degradation or other loss processes such as sorption, leaching, volatilisation, etc.) as determined by observation or extrapolation of substance disappearance, whereas DegT 50 is considered as the time for 50% degradation of the substance. There are cases where DegT 50 values are difficult to derive, i.e. for field study data, if appropriate abiotic controls are not possible or lacking, or experimental systems where several compartments are involved, e.g. for OECD 308 studies. While estimating a total system half-life is readily achievable in water-sediment studies (OECD 308), the derivation of valid DegT 50 values for the separate water and sediment compartments is often difficult, due to competing processes of the test substance dissipating from the water phase into the sediment, and the test substance forming NER.
The scope of this procedure is the mathematical derivation of characteristic numbers to be used in persistence assessment or multimedia fate modelling. Thereby, it has to be regarded that multimedia fate models implicitly assume first-order kinetics of all loss processes and that regulatory frameworks use half-lives in environmental compartments as persistence criteria. Therefore, it is inevitably necessary to ensure that the derived characteristic “half-lives” fulfil the basic assumption of representing first-order processes.
As described above, by definition DT 50 does not necessarily obey first-order kinetics, but constitutes the observed time needed to decrease the concentration of a chemical by 50% within an environmental system or compartment – be it in the field or in a controlled experiment. Such DT 50 values depend on the initial concentration which can be seen from the fact that even within one dataset multiple DT 50 may occur (often the loss of the first 50% is faster than the subsequent loss of another 50% of the remains). This is due to the fact that observed DT 50 values are the result of several different overlying processes such as degradation, phase transfer, adsorption/desorption or binding to humic matter. Even though the single processes are of first-order, this does not hold for the observed decrease of the chemicals’ concentration. This is especially true for one of the standard outputs of the OECD 308 test - the total system DT 50 . Therefore, observed DT 50 values, which have not been clearly shown to describe a first-order process, are inappropriate as input parameters for multimedia fate models and at least questionable to be used as half-life equivalent in persistence assessment.
To circumvent this problem, model approaches that consider the individual first-order processes can be used to estimate rate constants which then can be transformed to real half-lives. However, this is often difficult because of too few data points for reliable estimation of the many parameters (low degrees of freedom). Nevertheless, inverse modelling could be applied to results from standard tests such as OECD 307 or 308 and allows for estimation of first-order rate constants (and therefore true half-lives). Unfortunately, the terms half-life and DT 50 are sometimes used interchangeably. DT 50 is then defined as the true half-life as a result of all first-order loss processes (including NER formation) and DegT 50 being the true half-life of all first-order degradation processes. Deviations from first order kinetics must be addressed, and a possible lag phase should be described. Biphasic kinetics may be due to sorption and desorption processes. A discussion of these deviations and an overall description of the important processes that describe biotransformation in test systems may be found in FOCUS (2006).
It should be noted that there are studies where DT 50 and DegT 50 have been defined differently than above (Matthies et al, 2008; Loos et al, 2012). Where DT 50 in those studies is defined as the time to 50% disappearance due to NER and metabolite formation, DegT 50 is defined as the time to 50% metabolite formation only. Both studies show that DT 50 and DegT 50 values typically do not differ by more than a factor of 2 for the limited number of chemicals studied. However, at present there is no clear simple relationship between DegT 50 and DT 50 , since the relative contribution of NER formation to DT 50 is compound-specific and not (yet) predictable.
More guidance is needed on persistence assessment with data from multiple tests, as it is unclear how the natural variability of DegT 50 data should be taken into consideration when comparing against a single trigger value. As a first approach, with multiple (valid) data, the range of DegT 50 values should be considered along with all available information on soil types and any other sources of variability. Generally, if a log-normal distribution is given, it was suggested that the geometric mean should be used for comparison against the trigger value, but a weight of evidence approach should be used, as suggested in the revised annex XIII of REACH (EC, 2011).
3. Impact of temperature and normalisation of results
The syndicate thought it was logical to use the same reference temperature both for studies and for trigger values, e.g. 12°C for fresh water and soil, and 9°C for marine water, or 20°C alternatively as what is typically used in standard lab persistence tests. However, a need is seen nonetheless to discuss the relevance of temperature normalisation, as opposed to the uncertainty of single DegT 50 values. Normalised DegT 50 values may not reflect reality, as this remains to be proven (see research needs). For temperature normalisation, the Q 10 factor of 2.2 used in plant protection products (PPP) assessment (FOCUS, 2000), or the new Q 10 of 2.58 as developed by the European Food Safety Authority (EFSA) (EFSA, 2007) which is based on a literature review could be used, or the Arrhenius equation as suggested in REACH. It was discussed that this normalisation may have originated from its use in chemical degradation processes such as hydrolysis where the overall rate is both temperature and pH dependant. Over time, it has been applied to biotic processes as well, though whether it is correct to apply simple temperature correction factors to biological data is still debated today. On the other hand, the diversity of inocula may have a bigger influence on degradation rates than normalisation from different temperature environments. An evaluation from Helbling et al (2012) on degradation tests in 10 different sewage treatment plants resulted in a difference of 4 orders of magnitude in half-lives. So here, it seems vital to understand the different processes first.
Benchmarking with reference chemicals was also suggested by the syndicate as an alternative to temperature normalisation, though not necessarily a new concept. Many fate tests (OECD 308, 314, 303) do not require reference chemicals as part of the study protocol. This leaves a challenge of interpreting study results especially when samples for any given study may have unique seasonal and geographical characteristics that may influence study outcomes in a positive (greater transformation rate) or negative way (slower transformation rate). Including such reference chemicals will help normalise study outcomes when such diversity in samples or sample conditions exist and facilitate the use of the data in determining the regulatory outcome. This may be especially important for borderline persistence / non-persistence cases where the estimated half-life approximates the criteria used between one regulatory outcome and the next more persistent one (e.g. non-persistent, persistent or very persistent).
For persistence assessment, all available information should be assessed, using expert judgement such that a more holistic review may be made. This information may include but not be limited to physico-chemical properties (solubility, log D, log P, vapour pressure), metabolism data, other fate data such as hydrolysis, photolysis and sorption data; route(s) of introduction into the environment and volume of substance applied or introduced per a specific period. Each piece of information is considered together with the other data, e.g. data on hydrolysis and sorption: rapid sorption to sediment may prevent the substance from hydrolysing.
4. Transformation products identity and need for ERA risk
The syndicate group had a discussion on whether the persistence of transformation products was a relevant issue in persistence assessment. It is not included in all regulatory frameworks, but it was agreed that generally transformation products tend to be more polar, i.e. more water soluble and less toxic, but it is important not to miss those that deviate from that rule, such as DDT and DDE, and to understand the underlying processes (see research needs). So there is a need to understand when transformation products should be identified, under what circumstances, and for what endpoints (environmental or human safety) they are relevant in the risk assessment (Escher and Fenner, 2011). It was also noted, for example, that more polar transformation products are more mobile and thus may pose a risk to ground water. As such, identification would be necessary, e.g. for estimation of an acceptable daily intake if drinking water was a concern.
In terms of better defining the issue, more examples of when transformation products are a concern would be helpful, especially when defining / developing action criteria for additional testing. And, as risk reduction measures differ between commodity groups, different information may be needed. Regulatory agencies are always looking for information on transformation products.
There were different viewpoints on how to determine what may be relevant transformation products. Some suggested degradation studies at 30°C to ‘screen’ for potential products, while others preferred simulation studies conducted at environmentally realistic temperatures. While one approach may allow for a quick turnaround time, the other identifies those more closely linked to standard OECD methods.
Depending on the test sampling intervals, test duration and the test temperature, it was noted that it may be possible to miss the formation of some transformation products. Discussion in the group asked whether these transient transformation products are a real risk, as they are transient in nature, or are the ones that tend to accumulate over the study period more of a concern. While accumulating TPs were generally agreed as a potential concern, no consensus was reached over the concern of transient TPs.
For determining transformation rates, especially when there is more than one subsequent transformation product, more sampling points than what is typically required (6) may be needed to appropriately determine the kinetics for each of the transformation products. It is sometimes very challenging to determine how many extra samples could be needed and maintained for such, while not impacting the overall cost of the study. It would be helpful if one could rapidly screen for TPs prior to the study such that this could be determined ahead of time.
5. Endpoints for persistence assessment (primary versus ultimate)
The syndicate discussed that using primary biodegradation endpoints such as biotransformation is useful and needed in the persistence assessment as well as the risk assessment. But there was general consensus that it was not a major issue. In general, information about both endpoints is always needed.
Several technical points were discussed in how fate studies are conducted with respect to the need for more positive controls to confirm the viability of soil / sediment microbial community; and sterile controls to assess abiotic transformation. It was noted however, for abiotic controls, that it is difficult to achieve 100% sterile samples even with the combination of different techniques (chemical and physical) such as is done with activated sludge. A=Although there are other very effective alternatives to sterilisation, such as y-irradiation, it is also a technique that is not easily available to everyone. Most sterilisation techniques offer a 99% kill at best; therefore with a starting population of 10 billion bacteria a 99% kill would still result in there being 100 million viable bacteria remaining. No consensus was established as to whether sterile controls would be useful and significant regulatory questions could be raised about the validity of a study if sterility could not be maintained in an abiotic control.
A discussion on the use of the data from the OECD 314 test for classification purposes was also held. Although it was argued that the OECD 314 may be useful for risk assessment purposes (refinement of the PEC), the authorities participating in the discussion felt it could not be used for PBT assessment as it informs about degradation in the technosphere (STP or STP-influenced receiving stream) and not the natural environment. It was noted by industry that there have been some correlations established (non-published data) between what has been observed in STP degradation to the mixing zone, receiving waters and subsequent surface waters and also that more work is needed in this area. There is a need to bring forth what is known about the bacteria used in this STP test, including but not limited to the profile of micro-organisms present, occurrence of similar micro-organisms or enzyme systems in the natural environment, transformation half-lives of reference compounds and formation of subsequent TPs such that we can determine whether this is a potential model for the natural environment (surface waters) or not. Though the current regulatory position seems to be that such data can’t be used in classification schemes; it is also fair to say that without further research and disclosure of what is learned there is no means of furthering the debate or of advancing the science. This was identified as an area of research during the reporting of the syndicate in the session overview. One participant highlighted that the study may be able to have some immediate use in persistence assessments if the OECD 314 test was interpreted using the strict criteria applied under REACH to the OECD 302B test.
6. Dealing with half-life differences between soils and sediments
How does the fate in one compartment impact the other; after all the compartments are interconnected? The syndicate did not add much to this question other than to say that it seems the overall exposure is quite often driven by the compartment with the slowest half-life and that in general this needs to be approached on a case by case analysis. Further discussion on this topic is also found under the DT 50 vs. DegT 50 question. In general, this review requires all the information that is available including route of introduction into the environment and rate of introduction into the environment in order to provide the best assessment possible.
Syndicate 4: Enhanced realism within persistence assessment
Moderator: Stuart Marshall
Rapporteur: Gary Bending
Charles Eadsforth
Juha Einola *
Tom Federle
Kathrin Fenner *
Dieter Hennecke
Jacques L’Haridon
Robin Oliver
Ed Schaefer
Georg Streck
Markus Telscher
Karoline Wallner
* Participated in Syndicate session 4B.
1. Importance of adaptation.
2. Field relevance of lab data.
3. Can the OECD 314 be used to assess the persistence of down the drain chemicals?
4. Impact and significance of light on degradation studies.
5. Impact of temperature and normalisation of results.
6. Batch versus continuous methods to assess degradation.
7. Interpretation of BR / NER from persistence perspective; why should we worry about bound residues?
8. Assessment of Transformation products.
1. Importance of adaptation
The importance of considering adaptation in biodegradation assessments was seen largely as a problem for aquatic systems, mostly as a result of high variability in dynamics and heterogeneity in time and space. It was felt that in soils, inocula were larger and more diverse, so that laboratory tests using soil provided closer fit to field degradation rates. It was recognised that there are currently no specific protocols or recommendations available to test for adaptation, although there are a range of published studies which could provide guidance. Studies including an adaptation phase (e.g. Annex V of the OECD 309) would need to provide environmental degradation rates rather than providing a pass/fail assessment of the inherent potential of a chemical to degrade under non-relevant conditions. In other words, adaptation should not lead to over-estimation of the extent or rate of degradation in the environment. There was discussion about the Japanese MITI (Ministry of International Trade and Industry) test and whether a modified test of this nature could prove useful. Guidance would be required about the appropriate duration of tests, which may even require year-long timescales. Rather than assessment against persistence criteria, any such test could be interpreted in terms of pass – i.e. adaptation can occur – or fail, in which case higher tier tests would need to be conducted. Alternatively, discussions centred on the need for benchmarking against reference chemicals to allow calibration of test results in terms of persistence. Bins 2-4 proposed in the ECO 12 report (Comber and Holt, 2010) could provide a starting set of chemicals to allow this, although there is a need to increase the number of reference chemicals available.
2. Field relevance of lab data
The extent to which the results of laboratory tests need to be relevant to the field situation was discussed. It was suggested that screening tests do not necessarily need to be field relevant if they are sufficiently conservative with enough benchmarking. Generally, it was considered that ideally screening tests should be indicators of field dissipation rates for the appropriate scenario. OECD 307/308 were considered to have reasonable relevance to field rates for which they were originally designed, and for intermediate tests, such as the OECD 314, it was considered to be important that the test reflects the appropriate emission scenario and thereby processes occurring in the environment. The group felt that more confidence and data were needed for the OECD 309 to determine its relevance. With appropriate calibration they could potentially work well in terms of predicting rates, although it was considered that currently evidence is limited. There was discussion that in order for tests to have field relevance then all microbial processes which could act in specific environmental compartments should be considered within tests. In particular, following Robin Oliver’s talk, the importance of considering degradation by phototrophs, in addition to bacterial pathways, was considered. Soil testing in particular proves problematic, considering sieved soil alters the biology of the system.
3. Can the OECD 314 be used to assess the persistence of down the drain chemicals?
Scientifically it was considered that the test is robust, it may be a good predictor for the degradation of down the drain chemicals in wastewater treatment plants or in the immediate vicinity of effluents, and it was considered that there were no problems with the test which make it unsuitable for use in a regulatory setting. It was considered that if a substance is persistent in the OECD 314, then it will be also persistent in the aquatic environment. However, discussions within Syndicate 3 described above were more cautious about the use of this study in persistence assessments without any further investigation.
It was considered that the relevance of including of light in tests depends on the testing tier; in particular it could be useful in higher tier tests in which greater environmental realism is valued. Additionally, it will depend on the specific protection goal and scenario. It was felt that the registrant should have the option of including light. It should be possible to develop high throughput screening approaches which would work for many compounds, providing the potential for assessment of direct and indirect photolysis, and degradation by phototrophs, and allowing integration of all key processes which are important in specific environmental compartments.
5. Impact of temperature and normalisation of results
There was limited discussion of this topic. It was considered that the need for normalisation of degradation data for temperature was scenario specific. There was considered to be some evidence that temperature adjustment should be made on the basis of Q 10 . However, generally it was felt that there was currently insufficient knowledge about the extent to which temperature correction would make a difference to the test outcome.
6. Batch versus continuous methods to assess degradation
The longevity of inoculum in batch tests was considered. There appears to be good evidence that inoculum viability declines very quickly. Therefore, it is important to add the test chemical as soon as possible. In WWTPs a typical hydraulic residence time (HRT) is less than 24 hours. The 28 day duration of OECD 314 tests for this scenario is to assess the dynamics of intermediates and many sample points are clustered to assess removal during both the HRT and sludge residence times. The only continuous test is the OECD 303A test. However the frequent need for radioactively labelled chemical and manpower needed to run the units, makes this test expensive. It was considered that from a regulator’s point of view, there would be concern about extrapolating data from this test to the environment under direct discharge conditions, i.e. WWTP simulation tests would not be acceptable for assessing persistence per se. Generally it was considered that if a chemical fails to degrade in an OECD 314 test it will not degrade in a WWTP and it will be persistent in aquatic systems and possibly will accumulate in the environment.
7. Interpretation of BR / NER from persistence perspective; why should we worry about bound residues?
It was felt that there was still a need to clarify definitions of bound residues, and the need to integrate chemical and biological assessment of bioavailability. There was consensus that on a scientific basis type II and type III non-extractable residues (see Syndicate 3) represent a negligible risk. No likely scenarios have been outlined which could result in concern over release. It was felt that should a bound chemical or metabolite subsequently become available by slow desorption, risk would be covered by the original assessment. Although there could be an issue regarding differences between acute and chronic toxicity there was felt to be no evidence this could be an issue. However it was recognised that the regulatory position was at odds with this view. A key gap was considered to be understanding and prediction of covalent binding and desorption processes.
8. Assessment of transformation products
It was felt that considering the parent and transformation product together in testing schemes was the best approach. Where there was a shift in toxic mode of action between parent and transformation product this could be a problem, although formation of intermediates more toxic than the parent was considered to be rare. There were discussions about a range of scenarios in which transformation products could prove problematic, such as when polar parent compounds degrade to non-polar transformation products which have potential for bioaccumulation, when transformation products are more mobile than parents, posing a risk of leaching or contamination of raw drinking water, and when the parent and metabolite have different water solubilities and therefore pose risks to different compartments. It was considered that although tools are available to predict the properties of transformation products, such as QSAR, there is insufficient guidance on regulatory requirements, particularly in REACH and the accuracy and confidence of predictions for more complex molecules is limited.