Introduction to risk assessment

 

A risk assessment is the result of combining a hazard assessment with exposure scenarios – risk is a function of hazard and exposure.  To put it very simply:

 

Non-hazardous and low exposure = low risk

Non-hazardous and high exposure = good hazard assessment required

Hazardous and low exposure = risk management required

 

A hazard assessment needs to consider the physico-chemical, toxicological and ecological properties of a chemical substance and the result of this assessment leads to Classification with assignment of ‘H Statements’; however, this does not reflect the risk of the materials to workers, consumers or the environment and to take the hazard data to the next stage of assessment, it is necessary to understand potential exposure.

 

Risk assessment principles

REACH requires information on exposure to be gathered for all registered substances. For substances in the 1-10 tonnage band, this is a simplified set of information. For all substances manufactured or imported in quantities of 10 tonnes per year or more, a formal Chemical Safety Report (CSR) must be produced.

 

Even without the need to meet regulatory requirements, ‘risk assessment’ is an essential part of responsible care during the production, supply or use of any chemical material.  A basic risk assessment will help determine what conditions to avoid or what protective equipment to recommend to meet statutory requirements for handling chemicals – such as the communication or risk to customers, protection of workers and also to minimum impact on the environment through loss or accident.  The Safety Data Sheet (SDS) itself is a product of a risk assessment in that it describes how to reduce exposure in accordance with known properties of the material. Under REACH, the SDS needs to take into account elements of the CSR.

 

Any type of risk assessment should include consideration of:

 

·         The life cycle of the product

·         Potential hazards to health and the environment

·         Physical form (solid, powder, liquid, gas etc)

·         How is the product handled (manually, remotely)

·         Is there contact with workers / consumers

·         Is there a chance of loss to the environment

·         What happens to the product after use (disposal)

 

The conclusion of a risk assessment is to compare the estimated exposure with the estimated effects to derive a risk characterisation ratio (RCR); for the environment, this is expressed as a PEC / PNEC - the ratio of predicted environmental concentration (PEC) over predicted no effect level (PNEC).  Human health elements to the risk assessment are less clear-cut in their conclusions, but the principles remain the same with a comparison between the predicted level of exposure and the Derived No-Effect Level (DNEL), which is estimated from toxicity data with the application of safety factors.

 

The environment

The Predicted Environmental Concentration (PEC) is based on models 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 risk

Adsorption Coefficient

Log Koc > 3

= Adsorption to soil / sediment

Biodegradation

> 60%

= Significantly biodegradable

 

The model, described in the technical guidance or incorporated into the software of the risk assessment models such as  EUSES, ECETOC TRA or Chesar (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.

 

These models work well for organic, mono-constituent substances with no surface activity properties, but for UVCB or multi-constituent substances (eg mixed molecular weight or mixed isomers), special attention is needed.

 

Default figures are provided in the guidance documents provided by ECHA (based on the old TGD for Directive 93/67/EEC that is itself based on much earlier science) 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 during formulating activities, the position of the production unit in a small town with a small treatment works, and with final discharge going into a small stream.  Chapter 16 of the ECHA guidance provides these defaults that are built into the TRA and Chesar models.  EUSES is more basic and allows easier user input.

 

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.  and this can make a big difference to the final conclusion.

 

The Predicted No Effect Concentration (PNEC) is based on environmental effect data, such as toxicity to fish, Daphnia or algae and is determined by applying a safety factor. 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; 

 

Acute EC50

÷ 1000

(Acute = short term, eg 4 days fish)

Sub-acute EC50

÷   100

(Sub-acute = eg 21 days fish)

Chronic EC50

÷    10

(Chronic = long term, pond work etc)

 

In the absence of effects with acute studies, the PNEC is set at 1/1000 of the limit of solubility; likewise if there are no effects in longer-term studies, the PNEC will be 1/100 or even 1/10 of water solubility.  The models will estimate these figures and also set marine and terrestrial estimates as needed.

 

The ratio between PEC and PNEC is ultimately used as an indicator of risk, allowing it to be quantitatively labelled. For REACH the aim is to achieve a RCR of less than 1, i.e. exposure levels less than the PNEC.  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 through the uses leading to exposure.  The scale of the risk can therefore be crudely measured by considering this ratio – a figure of < 1 is of low concern, but if over 1,  risk management to reduce environmental release is needed and limitation of supply could be required. 

 

For industrial use (local, site specific), risk management can be possible through tight controls on discharge, collection of waste and washings for authorised disposal etc, but management for widespread use (regional or continental) risk management is not generally possible other than limitation of supply (ie 'uses advised against').

 

Human exposure

A risk assessment based on human exposure should consider 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).   The physico chemical properties such as dusts, vapours or liquid inhalation, splashing must also be considered as this can effect exposure routes.  Physical hazards, such as flammability or explosivity are also important for overall risk considerations. 

 

Quantifying exposure is very difficult and models attempting to make this easier (such as the EASE which is incorporated into TRA and Chesar models) rely on inputs such as vapour pressure, temperature of the process leading to exposure, dust content etc.  This is a rather simplistic model, but is suitable for targeted risk assessment (this is what 'TRA' stands for !) and is otherwise known as 'Tier 1'.  If the basic models and assumptions fail to provide assurance that exposure is likely to be below levels of concern, higher level assessments beyond those of these models are needed and  should be left in the hands of an expert. Site-specific worker exposure monitoring is certainly one approach.

 

Exposure needs to be considered in detail in the Exposure Scenario (ES) that is described in Section 9 of the CSR and several exposure scenarios may be needed to cover different processes or typesof  use.  Full exposure scenarios may take up several pages of the CSR, but summary scenarios (perhaps no more than 2 pages each) need to be used in the e-SDS to ensure that recipients of the SDS are not engulfed with volumes of paper and in-penetrable calculations.

 

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.  The exposure during manufacture of a pigment or its formulation into paint may seem obvious, as is the exposure to the wet paint when applying it to a surface.  However, the exposure to the paint from weathered surfaces is less obvious, but it all needs considering up to the point where there is sufficient dilution or degradation to show that the hazard is sufficiently low to avoid classification of the diluted material.

 

In an attempt to quantify ‘safe’ levels for human exposure, it is necessary to calculate a Derived No-Effect Level (DNEL) that is based on safety factors being applied to toxicity data endpoints such as the lowest observed adverse effect level (LOAEL) or no observed adverse effect level (NOAEL). The DNEL effectively follows the same principle as the PNEC for environmental effects.

 

To estimate the DNEL (or 'minimal' DMEL for substances with no no-effect level, such as carcinogens), many factors must be taken into account such as inter-species differences (rat to human), dose route variation (absorption by ingestion or through the skin) or the impact or chronic life-time exposure vs. relatively short term animal tests.  In many cases, there is insufficient data to make a full assessment and DNELs need to be estimated using generic safety margins; these are described in Chapter 8 of ECHA guidance and reproduced in the Table below;

 

Default assessment factors taken from REACH Guidance on Information Requirements, Table R. 8-6

 

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

 

AS a, b

 

2.5

 

1 f

2.5 g

Intraspecies

-          worker

5

5

-          general population

10 c

10 c

Exposure duration

-          subacute to sub-chronic

3

3 h

-          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

1 d

1 d

Quality of whole database

-          issues related to completeness and consistency of the available data

-          issues related to reliability of the alternative data

1 d

 

 

1 e

1d

 

 

1e

 

a              AS = factor for allometric scaling (rat to human = 4)

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.

 

 Note that the methods for preparing DNELs can vary and the idea is to provide adequate scientific justification for decisions taken

Risk assessment software

To help users of chemicals, software can be obtained from European Competent Authorities (such as the UK HSE).  Programs include ECETOC TRA and EASE for making a simple assessment of worker exposure and EUSES for predicting environmental impact. 

 

EUSES is quite complex to use and interpret and can be manipulated to give a more realistic answer than if it is used at its most basic level, using pre-set defaults.  For example, simple changes can be made to change the size of waste water treatment works or to put in the correct local dilution rate when effluent finally reaches surface water in the environment.  It is also possible to make changes to more subtle parameters that only environmental chemists can fully understand – or at least, they claim to be able to.

 

More recently, ECHA have released a new tool. CHESAR that brings together the models EUSES and ECETOC TRA.

 

Communicating Risk

The communication of risk to customers is through the SDS and where appropriate, the Exposure Scenario.   Even though substances over 1 tonne supply will need to be considered with an exposure scenario, lower volume hazardous substances will still require an SDS in the same way as currently applies.

 

The CSR does not need to be given to customers, but the ‘enhanced’ SDS must take into account the conclusions in considering risk reduction recommendations. 

 

Preparing a CSR

The CSR should be prepared using the templates provided by ECHA. The CSR consists of two parts; Part A is the conclusions and declarations that risk management measures are implemented and communicated and Part B that is the main technical component.  However, extraction from IUCLID (report generator) and merging with Chesar report produces a very good looking document in minutes.  This may take final editing, but once the IUCLID dossier is complete and Chesar populated (with justifications and comments added), the CSR itself can be produced electronically in minutes; the author then has to ensure that is it scientifically valid

 

 

Part B covers:

 

  1. Identity
  2. Manufacture and use patterns
  3. Classification and labelling
  4. Environmental fate assessment
  5. Health hazard assessment
  6. Physical hazard assessment
  7. Environmental hazard assessment
  8. Assessment of whether PBT or vPvB
  9. Exposure assessment
  10. Risk characterisation

 

The main source of hazard data will be the Chemical Safety Assessment (CSA) and the format for this will consider the environmental hazard, the human health hazard and physico-chemical health hazard data as well as the vPvB potential.  If vPvB, additional elements will need to consider exposure assessments and risk characterisation; if vPvB, close control over supply is expected and it is possible that Authorisation will not be given.

 

SDS Communication

The format of the SDS   has been updated under REACH to include the findings of the ES and CSR. These findings need to be communicated in various section of the SDS and it is important that the risk reduction recommendations are consistent with the ES and CSR. 

 

The ES may also need to be supplied to the customer to provide more detailed information on safe handling and control.  Sections relating to physico-chemical properties, toxicology and ecotoxicology must also be consistent with the CSA that in turn has been used to prepare the CSR.

 

Effectively, the job of the SDS remains the same, but under REACH, it is expected that there will more information to communicate at a greater level.  Chemical supply may also be linked to specific uses only and the SDS must obviously make this clear.

 

For the vast majority of non-hazardous products, there may be little real change in communication requirements.

 

Control and Authorisation

Dangerous substances may be Restricted through Annex XVII of REACH, or Authorised only for specific uses that have been considered to represent an acceptable level of risk or are considered to be of socio-economic benefit for a specific function.  Restrictions and Authorisation can be expected on substance considered to be CMR (Carcinogenic, Mutagenic or Toxic for Reproduction) or that are PBT or vPvB.  

 

For all hazardous substances, some level of risk management will be required and this is to be proposed in the CSR and communicated through the SDS and ES.

 

Conclusions

Good risk management of all chemicals is reliant on good communication – by informing those in contact with or responsible for the use of chemical of hazards and then proposing mechanisms for control of exposure, risk reduction is possible.  Exposure control must be proposed to reflect the hazard and good communication will allow judgement to be made on the level of control required. 

 

Denehurst Chemical Safety Ltd

Chemical Safety Report (CSR) and Chesar - Training and support

Technical Guidance Documents

Comprehensive Guidance on Information Requirements and Chemical Safety Assessment is available from ECHA.

 

http://echa.europa.eu/web/guest/guidance-documents/guidance-on-information-requirements-and-chemical-safety-assessment  

 

(opens in new window)

The Chemical Safety Report (CSR) is the end point in REACH for hazard, exposure and risk assessment and the findings of the report will dictate the levels of risk management to recommend in the SDS and to determine acceptable uses of a substance and to be able to advise against certain uses.

 

For Substances of Very High Concern (SVHC), the CSR will be a critical step in determining continued use in certain applications and if Authorisation is needed, then this is a central part of the justification for use.

 

A CSR is required for substances registered under REACH at over 10 t per annum and will include in Sections 1 - 8 a summary of substance identity, data assessment and justification for classification.  The most challenging part of the report is  in Section 9 (exposure scenario and estimated exposure) and Section 10 (risk characterisation).  A summary of Section 9 will need to be added to the SDS of hazardous substances. 

 

The level of detail required should be in relation to the known hazards and to the level of exposure and for low-hazard substances, the exposure and risk characterisation sections can be reduced to short statements.  The data summaries can be extracted directly from IUCLID (IUCLID CSR tool).

 

There is no prescribed format other than being split into two parts (imaginatively called Part A and Part B) and with three sub-headings in Part A and the ten Section headings in Part B.

 

Experience to date has been that many CSRs were prepared in a hurry to meet the deadline by consortia and consultants who were perhaps not fully briefed by the registrants and using early drafts of guidance.  Consequently, many CSRs need to be checked and perhaps updated; as the CSR is part of the registration, updates to the CSR will need to be re-submitted to ECHA as a spontaneous update (attached to IUCLID). 

 

Many of the issues coming to light are as a result of information being extracted to prepare the extended-SDS and some pitfalls are noted below.


Chesar is a free software tool that works with IUCLID.  It is available from ECHA using the following link (opens new page)  https://chesar.echa.europa.eu/


The software itself is not hard to use and is a major time saver in terms of generating the finished chemical safety report and re-importing Section 3.7 back into IUCLID.  However, it has many flaws and unless properly understood, can give misleading advice.


 

 

 

 

 

Common issues found when reviewing CSRs

 

 

 

 

Service offered by Denehurst

 

Although extraction from IUCLID using the CSR tool is easy and the TRA model / Chesar can be run with moderate IT and technical skills by most lead registrants, Denehurst can provide a reality-check either for lead registrants seeking independent outside opinion, or by joint registrants needing to confirm that the CSR prepared on their behalf by consortia or the lead registrant will not impact adversely on their intended uses.