- Chemical safety assessments (CSA)
- chronic effects
- chronic exposure
- default AF
- derived no-effect levels (DNEL)
- exposure scenarios (ES)
- general population
- Informed AF
- local effect
- R.8 guidance
- REACH TGD
- safety assessment
- skin contact
- systemic effect
- targeted risk assessment (TRA)
TR 110 : Guidance on Assessment Factors to Derive a DNEL | Dec 2010
The European Regulation on the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) requires from chemical manufacturers and importers detailed registration dossiers including chemical safety assessments (CSA) for all chemicals produced or imported in amounts of > 10 tons/year. The CSA includes exposure scenarios (ES) and assessments of exposure for all supported uses and requires these to be justified against derived no-effect levels (DNEL) which are in turn based upon hazard assessments. The DNEL developed must address acute or repeated exposure, different exposure routes (such as inhalation or skin contact), differentiate between systemic and local effects, and between workplace and general population exposure as appropriate for the intended use pattern. Thus, in theory, up to 15 DNEL may be defined for each compound although in practice fewer will normally be required. The DNEL that the registrant derives will depend upon the points of departure (POD) selected and the assessment factors (AF) subsequently applied.
Since the publication of Chapter R.8 of the “Guidance on information requirements and chemical safety assessment” (REACH TGD) chemical companies have been preparing and submitting registrations under REACH. Thereby, valuable experience has been gained on deriving DNEL using the R.8 guidance and on balancing these with the exposure predictions using the ECETOC targeted risk assessment (TRA) tools and other models. It is becoming clear that even for relatively data-rich chemicals submitted in the first tier of registrations (by December 2010) the multiplication of AF result in DNEL that are relatively low. These DNEL are extremely difficult to balance with the conservative exposure predictions, derived using screening tools such as the ECETOC TRA, under the risk characterisation ratio (RCR) recommended under the REACH TGD. Furthermore, it is anticipated that for the chemicals comprising the subsequent tiers of registrations (2013 and 2018) the compounding of more individual AF to account for data limitations will make balancing of the RCR more difficult, if not impossible. While it is recognised that REACH is a precautionary regulation and the guidance anticipates that if the RCR cannot be balanced the exposure prediction should be refined, the extra burden that this will impose on the registrants may not be justified. This may especially be the case, if the compounding of individual AF leads to unnecessary conservatism that can justifiably be avoided.
ECETOC and the R.8 guidance recognise that the use of “informed” AF is preferred over “default” AF wherever possible, whether supported by substance-specific data or, for example, by read-across to other chemicals or mechanisms of action. The use of informed AF for hazard and risk assessment is well-established and has been used for many years by organisations such as the Scientific Committee on Occupational Exposure Limits (SCOEL) and national competent authorities to set occupational exposure limits.
In reviewing the individual default AF recommended in the R.8 guidance, ECETOC referred to its previous publication on AF, Technical Report 86*, and supplemented this with an updated review of the literature published on this topic during the intervening years. This review revealed that the available scientific data, while supporting several of the AF being recommended in the R.8 guidance, did not support all of them. Specifically, no AF for exposure duration are necessary for local effects, and lower AF for intraspecies adjustment for workers and general population are indicated versus the ones proposed by the REACH TGD. In addition, the scientific justification for the additional interspecies AF of 2.5 for “remaining differences” is questionable as it involves aspects of “science policy”. Consequently, it is recommended that this factor should be further investigated with data that will become available from the first tier chemicals. As some of the R.8 guidance default AF appear to be unjustified by the current state of scientific knowledge it is recommended that, in the absence of substance-specific data, the ECETOC AF are used preferentially.
Although for most chemicals DNEL will solely be based upon animal data, for some health effects data derived in humans will be an additional and important source of information on effects. In this regard, the ECETOC Technical Report 104** provides a guide for an integrative framework for human and animal data that assesses the quality of each data set with respect to a given chemical or exposure scenario. This report supplements the newly issued REACH technical guidance on “Characterisation of dose[concentration]-response for human health, DNEL/DMEL derivation from human data” and provides further guidance on the selection of appropriate AF for human data.
In contrast to data generated on experimental animals, data on human exposure and effects are less controlled and therefore require greater expert interpretation. The AF recommended are typical maximum values that may be considered appropriate on a case-by-case basis to account for study deficiencies and are not intended to be arbitrarily multiplied together.
The guidance in this report is illustrated by a number of case studies drawn from SCOEL documentation in which the outcome of assessments based on default (REACH TGD, Chapter R.8) versus ECETOC recommended AF has been compared.