The Chemical Safety Report (CSR) is required for substances registered at levels above 10 t and every registrant will need to attach their own personalised CSR to their IUCLID dossier, even if part of a joint registration.  In many cases, it is hoped that consortia and the SIEF will agree on a generic CSR that can be applied to all registrants, but this may not always be possible in cases where some registrants have specialised uses that require special consideration.

In all cases, the hazard input sections of the CSR, including DNEL, PNEC and classification, will be agreed by the SIEF (or dictated by the lead registrant or consortium) and many ‘secondary' joint registrant will not need to get too involved in the technical input. 

However, the whole process does appear quire scary for many registrants and this short guide is intended to put the requirements into context; as with many parts of REACH, once the concepts are understood, the rest makes more sense.

Data input

REACH requires that the hazard data are collated in the form of a Chemical Safety Assessment (CSA).  This needs to include hazard data relating to physico-chemical data, human health and environment and conclude with classification and choice of hazard phrases.  As part of this process, every substance will need a ‘derived no (or minimal) effect level' or DN(M)EL to be assigned.  This is described in more detail below.

The data input itself needs to be considered carefully and any testing needs to reflect exposure.  It is impossible not to isolate hazard from exposure at any stage of the process as high exposure may meant the that there is a greater need for a high level of certainty in the data; for low exposure uses, read across or modelling may suffice, but if high exposure, good quality testing may be necessary.  This is especially the case for environmental effects and if there is not exposure predicted for the soil, earthworm and higher plant testing may not be needed.

Once the process of risk assessment broken down to the simple elements, it appears less complicated - however, the need to repeat this process for different substances being used in different scenarios or when considering different mixtures suggests that there is considerable work to be done.

The SDS remains the key communication tool and is supported by the other documents, the CSA, ES and CSR; if any of the sources of hazard or exposure data or the CSR itself are incomplete, the SDS will possibly fail to provide sufficient information.

Hazard Assessment

The terms ‘CSA' and ‘CSR' are often used interchangeably, but the CSA ‘Assessment' is best described as the process in which all input data are considered and the CSR ‘report' is the document that writes it up for submission.

The types of hazard data considered necessary for the CSA are prescribed in the legal text of the REACH Regulation and depend on the volume of supply and on preliminary test data.  If it is not possible to conclude the classification and labelling requirements or it is not possible to determine safety limits for exposure to man or the environment, further testing will need to be considered.

If new data are required, it is important to consult the official guidance.  This helps describe how to use published data, use models, conduct read-across and as a last resort, conduct new laboratory testing.  If new testing is required, strategies are described to ensure that time and money, and especially animals' are not wasted.

It should be noted that it is not possible to make a vPvB or PBT assessment on a mixture - these end-points are for substances and the presence of such materials in mixtures needs to be identified.

The Derived No (Minimal) Effect Level or DN(M)EL

The DNEL, or for materials without a no effect level (eg certain carcinogens), a Derived Minimal Effect Level (DMEL), is calculated by taking the toxicity end points, including acute, sub-acute or chronic toxicity and applying ‘safety factors' depending in the species, the duration of the study, the dose route etc.  The methods to make this calculation and the science that went into the process are too complex to discuss in this report, but it is described in Chapter r-8 of the technical guidance. 

As a rough rule, the table on page 38 of the guidance can be used (reproduced below), but it is possible to assign different factors if these are known form comparing real data with these assumptions.  In reality, only a few substances that have been examined in considerable detail will be possible to assess under more specific factors.  The most confusing part is the derivation  of dermal and inhalation DN(M)ELs from oral data - the scientific validity is questionable and the units required for completion of IUCLID are difficult to relate to.

To compound these problems, the technical guidance on DNEL and DMEL is not that easy to follow and most of it is not going to be relevant to most registrants. For experts in the field, the source documents are worth reading, but for those not experts, the unfortunate advice is perhaps to find one.  However, the guidance does describe the concepts of hazard assessment for health, but some of the detail is unclear; it does not help matters that many of those involved in the high levels of toxicity assessment disagree with each other on how this process should be conducted.  However, even if the DNEL does turn out to be a fairly random number, it does provide a system to compare the relative toxicity of substances that may have been tested by different methods.

The DNEL or DMEL is reported in the summary of toxicity in IUCLID and makes up part of the lead registration dossier. These numbers will be public and those supplying hazardous substances are expected to report the figures on the SDS.  The problem is that many readers of the SDS will not understand the DNEL and may confuse this with the exposure limits (eg WEL).  Unfortunately, there is likely to be considerable confusion as a result of publishing these rather meaningless numbers.

As an example, a no observed adverse effect level of 300 mg/kg/day in the 28 day erpeat dose oral study would result in a DNEL of 1 mg/kg/day in the workplace, or 0.5 mg/kg/day for the consumer.  This is based on Allometric Scaling (extrapolation from rat to human) of 4, multiplied by 2.5 for interspecies systemic effects, multiplied by 5 for work (or 10 for consumer, which doubles the factor) and then multiplying the factor by 6 for sub-acute to chronic.  The total safety factor is then 300 for workers and 600 for consumers.  The result of the 28 day study is then divided by this, so 300 mg/kg/day becomes  1 mg/kg/day for workers exposed over long periods and for consumers, half this figure. 

Table R. 8-6 Default assessment factors (reproduced from Chapter r-8 of technical guidance)

Assessment factor - accounting for differences in:		Default value systemic effects	Default value local effects
Interspecies	- correction for differences in metabolic rate per body weight - remaining differences	ASa, b 2.5	- 1f 2.5g
Intraspecies	- worker	5	5
	- general population	10c	10c
Exposure duration	- subacute to sub-chronic	3	3 h
	- sub-chronic to chronic	2	2 h
	- subacute to chronic	6	6 h
Dose-response	- issues related to reliability of the dose-response, incl. LOAEL/NAEL extrapolation and severity of effect	1d	1d
Quality of whole database	- issues related to completeness and consistency of the available data - issues related to reliability of the alternative data	1d 1e	1d 1e

a AS = factor for allometric scaling which is ‘4' for rat to human

b Caution should be taken when the starting point is an inhalation or diet study

c Not always covering for very young children; see text for deviations from default

d See text for deviations from default

e Special consideration needed on a case-by-case basis

f for effects on skin, eye and GI tract via simple destruction of membranes

g for effects on skin, eye and GI tract via local metabolism; for effects on respiratory tract

h for effects on respiratory tract.

To convert between different dose routes (oral to dermal, for example) needs special consideration and guidance on this is complicated with no clear recommendations. A case-by-case approach is recommended.

A simple example is given in the appendices of the Chapter r-8 guidance suggesting that oral to inhalation derivitisation can be done on a 1 to 1 basis and if considering human breathing to be 20 m³ day for a 70 kg ‘standard European person', and NOAEL of 50 mg/kg/day oral rat equates to a NAEL of 44 mg/m³ to humans (24 hr).  A bit of a leap of faith, but unfortunately, DNELs are expected for different dose routes and between workers and general population.

PNEC

The Environmental equivalent of the DNEL is the PNEC - this is a more simple process as there is a standard safety factors applied according to the type of tests performed, such as toxicity to fish, Daphnia or algae.  For acute studies, the safety factor of 1000 is applied to the EC₅₀ value; ie. a Daphnia EC₅₀ following 48 hours exposure of 50 mg/l would lead to a PNEC of 0.05 mg/l.  Longer-term studies require a smaller safety factor, as indicated below and if there are multiple studies, for example more than one algal study, the results can be refined further.

If there is no effect at the limit of solubility, this limit needs to be used as a starting point.

PBT / vPvB

To determine if a substance is PBT or vPvB, it is necessary to understand the persistence (stability) in the environment, the potential to bioaccumulate and its toxicity.   Initial assessment can be made with only three sets of test data; the ready biodegradability, partition coefficient and an acute toxicity test on an aquatic organism (eg. fish, Daphnia or algae).  Annex XIII defines PBT and vPvB, including rates of degradation in different environments; although these can be partly predicted by standard biodegradation methods, if your substance is borderline, certain expensive specialist studies may be worth performing if it keeps your substance on the market without restrictions (or  Authorisation).

Bioaccumulation is initially assessed through solubility characteristics, including partition coefficient between water and octanol; substances that show a preferential solubility in octanol to water will in turn be more likely to partition into fatty tissues in animals and not be excreted in urine.  Bioaccumulation potential is more complex than simple partitioning and takes into account factors such as metabolism and ability to cross biological membranes - however, the simple and quick partition coefficient test is a good guide and if the Log Pow if less than 3 (ie. less than 1000 times more soluble in octanol than in water), the risk of bioaccumulation is considered negligible and further testing would not normally be required.

The following text is taken from Annex XIII of the REACH Regulation 2006/190

Definitions: vPvB

A substance that fulfils the criteria of the sections below is a vPvB substance.

2.1. Persistence

A substance fulfils the very persistence criterion (vP-) when:

- the half-life in marine, fresh- or estuarine water is higher than 60 days, or

- the half-life in marine, fresh- or estuarine water sediment is higher than 180 days, or

- the half-life in soil is higher than 180.

2.2. Bioaccumulation

A substance fulfils the very bioaccumulative criterion (vB-) when:

- the bioconcentration factor is greater than 5 000.

Definition: PBT

A substance that fulfils all three of the criteria of the sections below is a PBT substance.

1.1. Persistence: A substance fulfils the persistence criterion (P-) when:

- the half-life in marine water is higher than 60 days, or

- the half-life in fresh- or estuarine water is higher than 40 days, or

- the half-life in marine sediment is higher than 180 days, or

- the half-life in fresh- or estuarine water sediment is higher than 120 days, or

- the half-life in soil is higher than 120 days.

The assessment of the persistency in the environment shall be based on available half-life

data collected under the adequate conditions, which shall be described by the registrant.

1.2. Bioaccumulation: A substance fulfils the bioaccumulation criterion (B-) when:

- the bioconcentration factor (BCF) is higher than 2 000.

The assessment of bioaccumulation shall be based on measured data on bioconcentration in aquatic species. Data from freshwater as well as marine water species can be used.

1.3. Toxicity: A substance fulfils the toxicity criterion (T-) when:

- the long-term no-observed effect concentration (NOEC) for marine or freshwater

organisms is less than 0,01 mg/l, or

- the substance is classified as carcinogenic (category 1 or 2), mutagenic (category 1

or 2), or toxic for reproduction (category 1, 2, or 3), or

- there is other evidence of chronic toxicity, as identified by the classifications:

T, R48, or Xn, R48 according to Directive 67/548/EEC.

Exposure Assessment

The term ‘Exposure Scenario' (ES) is part of the new REACH terminology and although it is only a formal Registration requirement for hazardous substances, the concept of the ES should be part of every risk assessment, whether for regulatory submission or as part of a well constructed SDS. 

The ES needs to include details of exposure to workers, the general public and the environment that are specific to expected use patterns of the substance or mixture (see below for mixtures).  Generic scenarios can be used (eg. Lubricants, paints, chemical intermediates) and although should be suitably comprehensive to cover generic locations, may be site specific for Authorised substances.

It is accepted that suppliers cannot know every detail of use by their customer and likewise, customers, or Downstream Users (DU) may not want to divulge their specific uses to suppliers.  Therefore, it is possible for DUs to prepare their own ES and subsequent CSR.  This is especially applicable to mixtures in cases where such formulation details are commercially confidential.

Exposure to the environment

The Predicted Environmental Concentration (PEC) is the end-point of the environmental exposure assessment and can be calculated using data for the degradation or distribution of the substance in the environment (between water, air and solids) using physico-chemical and biodegradation data.  As well as the test data, other key factors include how the substance is manufactured, formulated or used and the dilution factors from use. 

The distribution of chemicals discharged to waste water treatment plants is described in the aptly named ‘SimpleTreat' model.  This is a simplistic model that considers the volatility (Henry's constant, H), the partition coefficient, adsorption coefficient and biodegradation.

Volatility                           Log H > 3       = Significant loss to air

Partition coefficient              Log Kow > 3   = Accumulation threat

Adsorption Coefficient          Log Koc > 3    = Adsorption to soil / sediment

Biodegradation                     > 60%           = Biodegradable

The model, presented in tabular form in the TGD or incorporated into the software of the risk assessment model EUSES (see below), compares each of these factors in determining the distribution of the substance in the environment.  A water soluble substance with Kow = 0, that is biodegradable, for example, is predicted to be 76% degraded with 24% lost to surface water.  But a non-biodegradable poorly water soluble material with a Log Kow of 4, may have 56% to water and 44% to sewage sludge.

Default figures are provided in the Technical Guidance Documents (TGD), describing estimated concentrations of waste in effluent, standard dilution factors, sizes of water treatment plants etc. These default values consider worse-case scenarios with, for example, 2% of material produced being lost to waste water, the position of the production unit in a small town with a small treatment works, and with final discharge going into a small river. However, where only limited sites are involved in production, formulation or use, location-specific factors can be used, such as the size of the waste treatment works, river flow rates etc. 

The conclusions of an environmental risk assessment based, on the ratio of predicted environmental concentration (PEC) and The requirements for risk assessment for substances reviewed under REACH follow the same concepts and the level of further control, or the need to submit more test data, may rest on the ratio between predicted exposure and the predicted no effect levels; ie. the Risk Characterisation Ratio. 

The conclusion of the Risk Assessment is the recommended Risk Management Measures - these describe what must be done to ensure Adequate Control.

Human exposure

A risk assessment based on human exposure should also be considered.  This will depend on the type of exposure, whether deliberate or accidental, whether repeat or one-off or whether direct (eg. factory workers) or indirect (eg. in food or drinking water).  

Although the Technical Guidance Document for Directive 93/67/EEC does devote a large proportion of its pages to human health risk assessment, there is less direct support for those preparing risk assessments and there are no simple quantitative assessments as found with environmental data.  Instead, it is necessary to make a full review of all parts of the life-cycle in which the substance can come into contact with people, including manufacture, transport, storage, formulation, use and disposal.  Parts of the life cycle may not be obvious, such as the exposure to pigments caused by the degradation of paints. 

Chapters r-13 and r-14 provide good overview of the exposure to workers and how this exposure can be reduced by risk management measures.  Much of the information provided is quite simplistic and makes assumptions over skin area and inhalation rates, but this does provide a good starting point for risk evaluation.  It must be remembered that for the vast majority of substances that are not major health or environmental hazards, the CSR is going to be a largely academic exercise using these defaults and simple assumptions.

Risk assessment template and software

To help users of chemicals, the CSR template can be downloaded in Microsoft Word from the ECHA web-site together with over 2000 pages of guidance on testing and data evaluation.  The template is about 50 pages log, but the good news is that there will be many sections that are not considered relevant and can be omitted.  Tables are also embedded to help with Exposure Scenario preparation. 

EUSES software can be obtained for free from the European Chemical Bureau (ECB).  EUSES must be considered as a tool for risk assessment, but must not be relied upon to give definitive answers; unfortunately, some regulatory agencies rely too heavily on what is a very clever, but limited, tool.  It is advised that EUSES is used to confirm you own manual assessment and if there is a major discrepancy in conclusions, reconsider your work.

The free ECETOC Targeted Risk Assessment tool (TRA) is based on the same defaults and maths as EUSES and although intended to be easier to follow as a more simple process, it is indeed perhaps more difficult to use than EUSES.  The choice of model is often down to personal preference and experience.

The greatest danger in using any model is when this is done without the technical expertise to consider if the result is actually what you expect.  A quick back-of-envelope manual calculation may help models to be checked and one of the problems with the ECETOC TRA tool is that the ‘black-box' approach in which numbers are entered and results presented.  EUSES is a little better in that it provide intermediary results, even though these intermediary pages can cause confusion (and terror) for users.

These models are tools and nothing more; they cannot provide answers, but can be used to help identify areas of concern and help back-up scientific assessment. 

ECHA will be issuing a new tool in the near future that will help users, but again the danger here is that there may be over reliance on such a tool and common sense will be relegated.

Risk Characterisation Ratio (RCR)

The ratio between PEC and PNEC, or between workplace exposure and the DNEL, is ultimately used as an indicator of risk, allowing it to be quantitatively labelled.  If the PEC is greater than the PNEC (ie. ratio > 1 ), then it can be assumed that there is a risk of effects to the environment and likewise, if humans are exposed to concentrations greater than the DNEL, there may be a health risk. 

The scale of the risk can therefore be crudely measured by considering this ratio - a figure of 1 to 10 is of low concern, but over 100 is of major concern, and limitation of supply could be required.  Putting a number on risk may appear inappropriate (with the fundamental question of whether risk can be quantified numerically), but this does allow comparisons to be made between exposure scenarios or between different substances.

The CSR

If you need to register under REACH for over 10 t per annum of a substance (or less, if the substance is of very high concern), or if you are a Downstream User with a use of a material that is not covered by your supplier's Registration, you will need to prepare a formal risk assessment - the CSR.  If you are supplying less than 10 t or even if your supplier has done the formal part of Registration for you, knowing how to perform a risk assessment is vital to enable the provision of a good SDS.

The CSR is formatted into two sections; section A is the conclusions and declarations and section B is the technical part.  The total length of the CSR may run into many pages, depending on how many exposure scenarios need considering.  The guidelines do however suggest that only the main hazards are considered and that only the scenarios with the highest level of exposure need covering in detail.  In other words, if the scenarios with high exposure are not a cause of concern, the other scenarios should be OK.  Note that if there is concern, risk management measures may need to include restricting supply to avoid the scenarios of highest exposure.

Part A

1Summary of risk management measures

2Declaration that risk management measures are being implemented

3Declaration that risk management measures are communicated

This is very important in view of communication as it demonstrates that risk management measures (RMM) have been determined (the conclusion) and have been implemented by the organisation preparing the CSR.  It also confirms that you have informed your customer about the RMM (by way of SDS, for example).  If your customer chooses to ignore your recommendations, they are then responsible for the safe use of the product.

Part B

1.Identity

2.Manufacture, use patterns

3.Classification / labelling

4.Environmental fate assessment

5.Health hazard assessment

6.Physical hazard assessment (eg explosivity)

7.Environmental hazard assessment

8.Persistent / Bioaccumulative / Toxic  (PBT) assessment

9.Exposure assessment

10.Risk characterisation

Under the communication requirements of REACH, it will be acceptable to prepare CSR or ES for mixtures of substances (preparations) where it is considered that as a result, the DU is provided with better and more relevant information.  This is especially applicable for dangerous mixtures where specific hazards of the mixture need to be considered.  However, there is still the option to prepare individual documents for each specific substance and this is recommended when only one or two of the components are  dangerous.

It is worth noting that many of the essential tests for hazard assessment cannot be performed on mixtures, including solubility, partition coefficient, biodegradation, bioaccumulation etc.  it is not possible to assign PBT or vPvB to mixtures, only their components.  Likewise, the EUSES model does not apply to mixtures.  The documents used to support the Registration process must be for the substance Registered.

Conclusions

A CSR will need to be provided by all registrants of over 10 t and although the lead registrant or consortium concerned with the substance may do much of the input, it is essential that all those who have their name on the report understand the implications.

Some of the difficult parts such as DNEL, PNEC and PBT / vPvB assessment will be done by experts working for the lead registrant, but a basic understanding should be sought by all registrants.  In some cases, DUs will rely on these derived figures and care is needed when putting these numbers together.

Risk assessment is the final comparison of the hazard data with exposure and the simple rule is to use precautions to ensure that exposure remains below the level that is predicted to cause health or environmental problems.  Ultimately, we have a responsibility to ensure use of chemicals is within limits of safety.

Home page