1. ICCA GPS Hazard Characterization
Workshop: GPS Risk Assessment
and
REACH/GHS Implementation in practice
INT market 45431
27-28 October 2011
Podgorica - Montenegro
Renate Paumann

2. GPS Guidance on Risk Assessment
Section One: Preparation
Step 1: Select chemicals for assessment
Step 2: Gather information
Step 3: Prioritize chemicals into Tiers
Step 4: Develop Tier-relevant information (“Base Set of Information”)
Section 2: Implementation
Step 5: Characterize Hazard
Step 6: Assess Exposure
Step 7: Conduct Risk Characterization
Step 8: Document outcome (GPS Safety Summary)

3. Outline Hazard Identification Hazard Characterization
Physical Chemical Properties
Mammalian Toxicology
Ecological Toxicology and Environmental Fate
Epidemiology Information in Humans
Hazard Characterization
Identify dose descriptors for each endpoint
Modify to the correct point of departure (POD)
Decide on Mode of Action (threshold / non-threshold)
Apply overall assessment factor (AF) to POD
Derive endpoint-specific DNEL or DMEL or MOS / MOE
Risk Characterization
Where does Hazard Characterization fit in the whole picture

4. Gather information
How to obtain the information: sources to access information on GHS information, Phys/Chem, Hazard and Exposure Information
Evaluate the quality of the Information: Reliability, Relevance and Adequacy of data for assessment
Close data gaps: As long as the information is considered reliable, alternative sources are accepted e.g
Route-to-route extrapolation / Read-across from related substances
In vitro methods / (Quantitative) Structure Activity Relationships
Animal tests should always be considered carefully before conduct

5. General principles
A chemical’s potential to cause toxic or adverse effects is known as intrinsic hazard. Hazard characterization is the process of determining if exposure to a chemical can cause adverse effects
Because some hazard effects are limited to the tested animal species, hazard characterization also determines whether the adverse effect is likely to occur in humans.
However, a chemical’s intrinsic hazard will only manifest as an adverse effect if and when a set of conditions are met (a certain level of exposure, threshold of effect, incorrect handling and use).

6. Considerations when analyzing hazard data
The endpoints in the hazard assessment are interrelated
Degradation products and metabolites should be considered
The appropriate route of exposure for toxicity testing should be selected (oral, dermal, inhalation, etc.)
Test System Sensitivity (effect level depends upon the sensitivity of the test system)
Dose-response (the relationship between the amount of exposure and its health effects)
Identification of critical data (key studies)
Select appropriate dose descriptors (e.g., NOAEL, NOEL, BMD, LD50, LC50, and T25)

7. Choose Most Relevant Route of Human Exposure
Ingestion (water and food)
Absorption (through skin)
Inhalation (air)
Injection (bite, puncture, or cut)

8. The Dose Makes the Poison
An apparently nontoxic chemical can be toxic at high doses. (Too much of a good thing can be bad).
Highly toxic chemicals can be life saving when given in appropriate doses. (Poisons are not harmful at a sufficiently low dose).

9. Dose-Response Relationship:
The Degree and Spectra of Responses
depend upon the dose and the organism
Change from Normal State
could be on the molecular, cellular, organ, or organism level--the symptoms
Local vs. Systemic
Reversible vs. Irreversible
Immediate vs. Delayed
Graded vs. Quantitative
degrees of the same damage vs. all or none 2 3 4 1
DOSE
RESPONSE
0-1 NOAEL 2-3 Linear Range
4 Maximum Response

10. Concept of Dose – Response (Generalization)
　Quantification of Toxic potential　
50/50
Lethal Dose
LD(C)50
Safety factor
Frequency
35/50
1/100～(AF1XAF2・・・AFn）
No observed adverse effect level or concentration
(NOAEL or NOAEC)
25/50
　　　DNEL or PNEC　
　 （ADI or BMD:LED10）
20/50
4/50
10％
Detection limit
0.1 1 10
100　
Intake (mg/kg or L)

11. The Hazard Identification Process
The Potential Effects to Human Health or the Environment
Chronic (long term)?
Repeated dose (short-medium term)?
Acute (very short term)?
Species Affected
People, Birds, Fish, Mammals…

12. Hazard Information Chemical & Physical Properties Mammalian Toxicology
Flammability, Reactivity
Persistence in the environment
Mammalian Toxicology
Acute, Irritation and Corrosivity, Sensitization,
Mutagenicity and Genotoxicity, Repeated Dose,
Reproductive and Developmental Toxicity, etc.
Environmental Toxicology and Chemistry:
Aquatic, Terrestrial, Environmental Fate
Epidemiological Information in Humans
Generally taken from industrial exposure settings & not frequently available

13. Physical Chemical Properties
Flammability:
Explosivity:
Reactivity:
Corrosivity:
Persistent Bioaccumulative and Toxic (PBT):

14. Mammalian Toxicology Acute (oral, dermal, inhalation)
Irritation (skin, eye)
Sensitization
Mutagenicity/Genotoxicity
Repeated dose
Carcinogenicity
Developmental and Reproductive
Tests used in hazard ID but not discussed in details here
QSAR
Toxicokinetics
Neurotoxicology Studies
MOA (mechanism of action)

15. Acute Toxicology Tests (LD50 / LC 50)
Species
Rat, mouse
Single exposure, multiple doses within 24 h
Oral, dermal, inhalation
Observed up to 14 days
Endpoint is lethality
Calculate LD50, LC50: calculated dose of a substance that is expected to kill 50 percent of a defined experimental animal population
Modeling and in vitro methods

16. Eye / Skin Irritation Species Single exposure, multiple doses
Rabbit, rat
Single exposure, multiple doses
Observed up to days
Endpoint irritation to eye/skin
Severity of effect (grading of irritation)
Reversibility
Strong acids or alkalis (pH ≤ 2 and ≥ 11.5): no need to test; Severe skin irritant: no need to test on eyes
QSAR and In Vitro Testing
Corrosion
Phototoxicity

17. Dermal Sensitization Endpoint: skin sensitization/allergenicity
Guinea Pig Studies (OECD 406)
Guinea pig maximization test (GPMT) or Buhler Test (BT)
Positive or negative results
No dose response data
Local lymph node assay (LLNA) (OECD 429)
Test species – mouse
Provides dose response data
Calculates EC3: effective concentration that induces a 3-fold increase in cell proliferation (stimulation index)
QSAR Modeling

18. Mutagenicity / Genotoxicity
Endpoints: gene mutation and chromosome aberration
In vitro and in vivo methods
Tiered approach:
Lower tier: gene mutation w/wo chromosome aberration
Higher tiers: in vitro or in vivo MNT, germ cell genotoxicity
Basic study types
Reverse Mutation Test (Ames)
In Vitro Mammalian Cell Gene Mutation (HPRT, TK)
In Vitro Mammalian Chromosome Aberration Test (RLCAT)
In Vivo Chromosome Aberration Test (MNT)
QSAR Modeling

19. Repeated Dose Endpoints: test substance related adverse effects
Larger number of animals
10/sex/dose
Typically performed in 2 species
Rat, dog, mouse, rabbit
28-Day, 90-Day, or major life span
Conducted by most likely route of human exposure
3 doses excluding control
Determine MTD (maximum tolerated dose) or limit dose (1000 mg/kg bw)
Determine thresholds for adverse effects and no effects
NOEL, LOEL, NOAEL, LOAEL

20. Carcinogenicity Endpoints: neoplasm, tumor formation
Often combined with chronic study
Analyze available information:
QSAR models
Genotoxicity test results (In vitro and in vivo)
Largest number of animal
50/sex/dose
Several doses – MTD, 1/2 or 1/4 MTD
Majority of lifetime
Typical species: rat, mouse
18 m (mouse) or 24 m (rat)
Lower doses; subtle effects
Long time, high cost

21. Developmental and Reproductive Test (DART)
Endpoints: developmental and reproductive toxicity:
Focus on in utero development
Birth defects
Growth retardation
Embryonal/fetal lethality
Maternal Toxicity
Fertility
Exposure: The start of implantation to near term
Often required in two species (rat / rabbit)
Endocrine Disruptor Screening Program (EDSP)

22. Aquatic Toxicology Determination of toxicity on aquatic organisms Fish
Algae
Invertebrates such as Daphnia magna (water flea)
Marine
Type of studies
Study populations
Acute to chronic tests
Static, semi-static or flow-through test conditions
Similar endpoints to mammalian acute studies
LD50s, EC50s

23. Terrestrial Toxicology
Majority of tests conducted in birds
Mallard duck
Quail
Chickens (sometimes)
Type of studies
Study populations
Acute to chronic tests
Similar endpoints to mammalian acute studies
LD50s, ED50s

24. Environmental Chemistry
Deals with transport of chemicals (intra- and interphase) in the environment
Relationship of the physical-chemical properties to transport
Their persistence in the biosphere
Their partitioning in the biota and toxicological forecasting based on physical-chemical properties
Fate & Lifetime
What happens to a chemical in the environment?
Where will a chemical reside (soil, water, air, etc.)?
How will a chemical transport between media?
What processes will remove (destroy) a chemical?

25. Environmental Chemistry
Environmental Partitioning and Transport
Water Solubility
Air/Water Partitioning
Octanol/Water Partitioning
Soil Adsorption
Bioconcentration
Fugacity Modeling
Abiotic Degradation
Hydrolysis
Photodegradation

26. Environmental Chemistry
Biodegradation
Ready and Inherent Biodegradability
Aerobic Aquatic Biodegradation
Anaerobic Biodegradability
Metabolite/Pathway Identification
Environmental Fate Simulation Tests
Wastewater Treatment
SCAS, Porous pot, sludge inhibition
Surface Water/Sediment Microcosms
Soil Biodegradation
Aquifer/Groundwater Microcosms

27. Environmental Chemistry
QSAR & Exposure Modeling
Environmental Property Estimation
Atmospheric Chemistry and Transport
Aquatic and Terrestrial Transport and Fate
Subsurface/Groundwater Transport and Fate
Field Studies

28. What is Epidemiology? Risk of Disease
Epidemiology is the study of the patterns of disease occurrence in populations and factors that influence those patterns.
PERSONS (hosts)
Environment
Environment
Risk of
Disease
PLACE
Agent
TIME

29. The Hazard Characterization Process
Two main approaches - both follow the same basic methodology -however, the ways the outcomes presented are different:
MOS/MOE: The classical approach is the derivation of a Margin of Safety (MOS), also termed Margin of Exposure (MOE). Here, assessment factors are considered after deriving the result.
DNEL: In Europe, REACH legislation has established the Derived No Effect Level (DNEL). Assessment factors are accounted for in the process of the DNEL derivation and therefore are included in the result.

30. The Hazard Characterization Process

31. 1. Identify dose descriptors for each endpoint
Identify dose descriptors for each relevant hazard endpoint (e.g. NOAEL, NOAEC, BMD, LD50, LC50, T25).
e.g. NOAELs are derived from sub-acute, sub-chronic, chronic and reproductive toxicity tests – they cannot be derived from acute, irritation or skin sensitization tests.
NOAEL is the highest dose or concentration of the substance used, at which no statistically significant adverse effects were observed.
For example if the dose levels of 400, 100, 50 and 5 mg/kg/ day have been used, and adverse effects were observed at 400, 100 and 50 mg/kg but not at 5 mg/kg , the derived NOAEL will be 5 mg/kg / day.

32. ….When no dose descriptors are available
e.g. no effects are seen even at the highest dose level, then no NOAEL or LOAEL can be derived.
If the dose levels tested are sufficiently high, then it can be concluded that there is no risk for that particular endpoint.
If not, then conduct a DNEL or the MOS/MOE calculation using the highest dose tested as the NOAEL.
If the calculated MOS/MOE is considered sufficiently high, then the conclusion is clear: no concern. If the MOS/MOE is small then the exposure scenarios are likely to show significant human exposures.
The final option is to ask for more data – taking animal welfare issues and conclusions from other endpoints into account.

33. 2. Modify the dose descriptor to the correct point of departure (POD)
In certain situations, the dose descriptor may not be directly comparable to the exposure assessment e.g.
interspecies differences in bioavailability between experimental animals and humans;
the animal dose descriptor might relate to a different exposure route than the human exposure - requiring route-to-route extrapolation;
differences in human and experimental exposure time conditions; or
Differences in respiratory volumes between experimental animals and humans.
In these situations, it is necessary to modify the dose descriptor (e.g. NOAEL) into an appropriate Point of Departure (POD)

34. 3. Decide on Mode of Action (threshold yes / no)
Certain chemicals are thought to impose e.g. a carcinogenic risk without a threshold dose. For these chemicals the conventional NOAEL and DNEL derivation is not appropriate.
The DMEL (derived minimal exposure level) describes the exposure level corresponding to a certain low risk, that appears to be tolerable though it is higher than zero.
However, for carcinogens and mutagens workplace exposures should be avoided / minimized as far as technically feasible.

35. 4. Apply Assessment Factor (AF) to the POD
Uncertainties in the extrapolation of experimental animal test data to real human exposures are addressed by applying Assessment Factors (AF).
Difference in exposure duration between the experimental data and the assumed real-life exposure for humans;
Route of exposure if different for humans
Differences in sensitivity of response between species (inter-species) and within species (intra-species).
After identifying the relevant individual assessment factors, the overall assessment factor is obtained by simple multiplication of the individual AFs.

36. 5. Derive endpoint-specific DNEL
To derive endpoint-specific DNEL(s) - the overall AF is to be applied directly to the POD
Data from more than one valid and relevant studies may be available (e.g. in different species, with different durations), identifying more than one dose descriptor to a given endpoint.
It might be necessary to derive DNELs for more than one dose descriptor, prior to selecting the lowest DNEL for that endpoint.

37. 6. Select leading health effects
After deriving your endpoint-specific DNEL or DMEL, select the leading health effect(s) and the corresponding DNEL/DMEL.
These critical DN/MELs should be the lowest DN/MEL obtained for each exposure pattern.
They will be used to characterize risk in the next step

38. Examples for DNEL calculation
Step 1: Identify dose descriptor
Dermal Irritation (local effect)
Dose descriptor: NOAEL 50 mg/kg bw/ day
Rationale for selection of dose descriptor: Skin irritation observed at higher doses
Adrenal gland changes (systemic effect)
Dose descriptor: NOAEL 10 mg/kg bw/day
Rationale for selection of dose descriptor: Adverse changes to adrenal glands observed at higher doses
Developmental effects (systemic effect)
Dose descriptor: NOAEL 50 mg/kg bw/day
Rationale for selection of dose descriptor: Developmental effects observed at higher doses

39. Examples for DNEL calculation
Step 2: Decide on threshold / non-threshold (Mode of Action)
Dermal Route, Local & Systemic Effects
Irritation - Dose-response information supports threshold
Adrenal - Dose-response information supports threshold
Developmental - Dose-response information supports threshold

40. Examples for DNEL calculation
Step 3: Modify Point of Departure
Irritation (Local)
No modification needed
Adrenal (Systemic) Effects
Substance-specific data indicates dermal absorption is 2x less in humans than rats
Developmental Systemic) Effects

41. Examples for DNEL calculation
Step 4 / 5: Apply assessment factors and derive DNEL
Step 6: Select leading adverse effect
The DNEL-dermal-long term-local is 1.7 mg/kg bw/day
The DNEL-dermal-long term-systemic is 0.1 mg/kg bw/day

42. Situations where no DNEL can be derived
Certain chemicals are considered to impose e.g. a carcinogenic risk without a threshold and their mode of action (MoA) is of concern at the smallest exposure concentration.
For these chemicals the conventional NOAEL and safety factor approach to derive exposure standards is not appropriate.
In these circumstances two options are available
the calculation of a DMEL (derived minimal exposure level), as described in the European REACH legislation; or
the calculation of the MOE (margin of exposure), as described by the European Food Safety Agency for situations where no threshold can be calculated.

43. Derive endpoint-specific DMEL (non threshold)
There are default methodologies which can be applied for deriving a DMEL.
Linear extrapolation from animal bioassay data (quantitative approach): The DMEL is derived by linear extrapolation from the tolerable lifetime cancer risk calculated from a defined POD close to the experimental dose range (e.g. a T25 or a BMD10 cancer incidence in a rodent long-term cancer bioassay).
Threshold Approach: The threshold of toxicological concern (TTC) is a principle which refers to the possibility of establishing a human exposure threshold value, below which there is no appreciable risk to human health (by the oral route) generated in the past.

44. Examples for DMEL calculation (non threshold)
Step 1: Identify dose descriptor
T25 as basis for POD = 250 ppm
Step 2: Decide on threshold / non-threshold (Mode of Action)
Non-threshold carcinogen
Step 3: Modify Point of Departure
No modification needed

45. Examples for DMEL calculation (non threshold)
Step 4 / 5: Apply assessment factors and derive DMEL

46. Environmental hazard characterization
Environmental hazard characterization is conducted in a similar manner as for human health.
Here PNECs (predicted no effect concentration) are used as dose descriptors and derived from the data collected.
PNECs usually result from single species laboratory toxicity tests (e.g. fish, algae, and daphnia). Data are typically reported as the concentrations at which x% (e.g. 50%) mortality or inhibition of function (e.g. growth) is observed.
PNECs are expressed as the lethal concentration (LCx) or the effect concentration (ECx), e.g. LC50 or EC50.
A PNEC must be calculated for each environmental compartment in which exposure is expected (air, water, sediment and soil).

47. Conduct Risk Characterization
A very important concept is the distinction between hazard and risk.
Hazard defines the inherent property of a chemical agent having the potential to cause adverse effects when an organism, system or population is exposed to that agent.
Risk however establishes the probability of the adverse effect in an organism, system or population to occur under specified circumstances
“Risk is the possibility of suffering harm from a hazard”

48. Risk Characterization
Hazard Dose - Equivalent Characterize
Identification Response Dose In Risk to in Animals Humans Humans
Human Exposure
RISK ASSESSMENT PROCESS
RISK
MANAGEMENT
PROCESS*
* Risk/Benefit Decisions
Exposure Controls

49. A Tiered Approach to Risk Characterization
SCREENING RISK ASSESSMENT
(worst-case exposure)
below
above
standard
DONE
REFINE EXPOSURE ASSUMPTIONS
( more realism )
• REFINE ASSUMPTIONS
• INSTALL CONTROLS
DONE !!
(best estimate)

50. Thank you for your attention!