Introduction: Toward Integrated Safety-Sustainability Assessment

1. Risk assessment (Safety): Using toxicity values following the REACH approach to determine regulatory safety thresholds
2. Sustainability assessment (Sustainability): Using the same data to calculate characterization factors in Life Cycle Assessment (LCA/PEF)

However, the assessment of toxicological impacts of chemicals has long been the Achilles' heel of the LCA method. During the PEF/OEF pilot phase (2013-2018), four recurring problems emerged with the use of the USEtox® model:

1. Systematic dominance of metals in toxicity scores, masking contributions from organic substances
2. Lack of reliability of physico-chemical data for many substances
3. Misalignment between ecotoxicity data used in LCA and those used for regulatory risk assessment
4. The traceability and REACH data valorization challenge: Beyond reliability aspects, a major challenge concerned the traceability and optimal exploitation of available ecotoxicological data. Historically, USEtox® relied primarily on international databases (notably US-EPA's ECOTOX), where documentation on precise origin, experimental conditions, and test quality of the data used to calculate USEtox CF were not completely transp.

An untapped opportunity: In parallel, since REACH entered into force in 2007, Europe has built an unprecedented ecotoxicological data infrastructure. REACH registration dossiers contain thousands of tests performed following standardized OECD protocols, with explicit quality criteria (Klimisch scores, GLP tests, ECHA validation), and have been publicly accessible via REACH-DB since 2008.

This situation created a methodological paradox: on one side, a considerable European investment in generating high-quality, traceable, and standardized toxicological data (REACH, EFSA); on the other, an LCA community using international databases where these new European data were not systematically integrated or updated.

The convergence potential: This desynchronization between the regulatory ecosystem (REACH, EFSA) and the LCA ecosystem represented an improvement opportunity rather than an insurmountable problem.

Facing these limitations, the Joint Research Centre (JRC) of the European Commission developed an innovative approach for Environmental Footprint version 3.0 that aligns both SSbD pillars and seizes the convergence opportunity between regulatory and sustainability. This methodology exploits three major European databases: REACH (ECHA), OpenFoodTox (EFSA), and PPDB (Pesticides), enabling:

✅ A spectacular increase in characterized substances (+140% to +437% depending on category)

✅ Methodological alignment between risk assessment (REACH/EFSA values) and sustainability (PEF factors)

✅ Coherence in the SSbD approach: same data sources, same quality criteria, complementary objectives

The key innovation: Transforming regulatory safety data (REACH, EFSA) into sustainability factors (PEF), thus creating an operational bridge between the two dimensions of Safe and Sustainable by Design.

Data Sources: REACH, EFSA and PPDB

REACH-DB: The Raw Data Reservoir

The REACH database from the European Chemicals Agency (ECHA) constitutes the starting point with 305,068 ecotoxicological test results covering 7,714 substances. However, this data mass presents a major challenge: highly variable quality requiring rigorous curation.

Advantages:

• Considerable volume of physico-chemical and toxicological data
• Broad coverage of substances marketed in Europe
• Data from industrial registration dossiers

Limitations:

• Heterogeneity of experimental conditions
• Absence of systematic peer-review
• Presence of low-quality data (Klimisch scores k3, k4)

OpenFoodTox (EFSA): Guaranteed Quality

EFSA's OpenFoodTox database stands out with already validated data used for food safety risk assessment.

Key characteristics:

• Pre-validated data by EFSA experts
• Assured quality, direct relevance for risk assessment
• Minimal curation need compared to REACH

PPDB: Pesticides Expertise

The Pesticide Properties Database provides specialized data on pesticides, with physico-chemical and toxicological properties already selected and evaluated.

Points of attention:

• Limited documentation on experimental conditions
• Aggregated data requiring cross-validation

JRC Methodology: Six Rigorous Curation Rules

The JRC developed a systematic procedure with 6 main rules to extract REACH data of sufficient quality for calculating characterization factors.

Rule 1: Study Quality and Type

Strict study selection:

• Retention only of "Key study" (46%), "Supporting study" (30%) and "Weight of evidence" (14%)
• Klimisch score k1 ("reliable without restriction", 36%) and k2 ("reliable with restrictions", 50%)
• Total exclusion of k3 ("not reliable", 8%) and k4 ("not assignable", 2%)

Impact: On 305,068 initial results, this first rule already eliminates ~10% of low-quality data.

Rule 2: Aquatic Environment

• Exclusive selection of freshwater tests
• Exclusion of saltwater or brackish water tests (not relevant for "freshwater ecotoxicity" category)

Rule 3: Values Expressed as Intervals

Systematic exclusions:

• Values with ">": 39,602 results (13% of database)
• Values with "<": 4,397 results (1.4%)
• Retention only of exact values or with "≥", "ca.", "≤"

Justification: Censored values (>, <) introduce bias in Species Sensitivity Distributions (SSD).

Rule 4: Acute/Chronic Selection

Criteria based on:

• Measured biological effect
• Measurement endpoint (EC50, NOEC, LOEC)
• Exposure duration

Preference hierarchy:

1. Chronic tests (NOEC, EC10)
2. Semi-chronic tests
3. Acute tests (EC50, LC50) if chronic absent

Extrapolation factors:

• Acute → Chronic (EC50 → EC10): factor 0.1
• Chronic (EC50 → EC10): factor 0.3

These factors are derived empirically from substances with complete acute + chronic data.

Rule 5: Taxonomic Assignment

3 hierarchical assignment levels:

1. Level 1 (preferred): Exact species identified (e.g., Daphnia magna)
2. Level 2: Genus identified (e.g., Daphnia sp.)
3. Level 3: Broad taxonomic group (e.g., Crustacean)

Automatic assignment for unspecified species:

• Crustacean → Daphnia magna (most tested species)
• Fish → Pimephales promelas (US-EPA model species)
• Algae → Raphidocelis subcapitata (formerly Pseudokirchneriella)

Identified risk: Potential bias if actual species are more sensitive than D. magna.

Rule 6: Taxonomic Representativeness

To derive a robust SSD (Species Sensitivity Distribution):

• Minimum 5 species different
• Representing at least 3 taxonomic groups among: 
• Algae / Aquatic plants
• Crustaceans
• Fish
• Insects
• Molluscs
• Amphibians
• Annelids
• Rotifers

If criteria not met: Two options:

1. QSAR (in silico predictive models) if molecular structure available
2. Fixed slope 0.7 (97.5th percentile of observed SSD slopes) as last resort

Results: Spectacular Coverage

Freshwater Ecotoxicity (ECOTOX)

Before (USEtox 2.1): 2,530 substances

After (EF 3.0): 6,038 substances (+140%)

Sources:

• REACH-DB: 68% (4,093 substances)
• USEtox 2.1: 21% (1,253 substances)
• EFSA OpenFoodTox: 5% (322 substances)
• PPDB Pesticides: 2% (141 substances)
• Others: 4% (229 substances)

Distribution by quality:

• 19% of substances: QSe (Ecotox Quality Score) > 1.48 (intermediate-high quality)
• 82% of substances: QSe ≤ 1.48 (low quality, verification recommended)

Interpretation: REACH becomes the dominant source, but with variable quality. LCA users should check QSe for substances contributing significantly to the score.

Human Toxicity Cancer (HTOX_c)

Before (USEtox 2.1): 595 substancesAfter (EF 3.0): 1,024 substances (+72%)

Sources:

• EFSA OpenFoodTox: 47% (482 substances)
• REACH-DB: 32% (327 substances)
• USEtox 2.1: 15% (152 substances)
• PPDB: 6% (63 substances)

Observation: EFSA becomes the majority source, reflecting the quality of food toxicological assessments.

Human Toxicity Non-cancer (HTOX_nc)

Before (USEtox 2.1): 617 substancesAfter (EF 3.0): 3,317 substances (+437%)

Sources:

• REACH-DB: 70% (2,323 substances)
• EFSA OpenFoodTox: 18% (595 substances)
• USEtox 2.1: 7% (226 substances)
• PPDB: 5% (173 substances)

Spectacular increase: REACH exploitation multiplied by 5.4 the number of substances characterized for non-cancer toxicity.

Innovation: Interactive Transparency Tools

The JRC developed innovative web tools to explore source data:

Database_ALL: The Complete Exploratory Base

• 71,605 ecotoxicological results for 9,862 substances
• Filtering interface by species, duration, quality
• Allows detailed audit of raw REACH data

Database_RICH: The Validation Base

• 10,144 results for 670 "data-rich" substances
• Used to validate the fixed slope 0.7 method
• Substances with ≥5 species from 3 taxonomic groups

Access: JRC Platforms - Knowledge Centre for Bioeconomy

Safe and Sustainable by Design: The Dual Value of REACH/EFSA/PPDB Data

The JRC approach perfectly illustrates the Safe and Sustainable by Design (SSbD) principle by demonstrating how the same toxicological data can simultaneously serve two complementary objectives:

1. Risk Assessment (Safety) - REACH Approach

Regulatory use of data:

• DNEL (Derived No-Effect Level) for human toxicity
• PNEC (Predicted No-Effect Concentration) for ecotoxicity
• Safety thresholds based on NOEC, LOEC, EC10
• Exposure assessment → Risk Characterisation Ratio (RCR)

Logic: Protect human health and environment via regulatory thresholds

2. Sustainability Assessment (Sustainability) - PEF/LCA Approach

Use for characterization factors:

• HC20 (Hazardous Concentration affecting 20% species) via SSD
• ED50 (Effect Dose affecting 50% population) for humans
• Characterization factors (CF) in PAF·m³/kg or DALY/kg
• Impact scores → Relative contribution of substances

Logic: Quantify complete life cycle impacts of product/service

SSbD operational benefits:

✅ Data efficiency: Single toxicological data collection for two assessments

✅ Enhanced traceability: Each data documented with origin (REACH dossier), quality (Klimisch), experimental conditions

✅ Decision coherence: "Safe" (REACH) but "unsustainable" (PEF) substance → clear signal for substitution

✅ Cost reduction: No duplication of ecotoxicological tests

✅ EU harmonization: Same regulatory databases (REACH/EFSA) for Safety and Sustainability

✅ Investment valorization: Data generated for regulatory compliance exploited for sustainability

SSbD Perspectives

The JRC approach opens the way to:

1. Early SSbD screening: Assess simultaneously Safety and Sustainability (CF) from substance R&D
2. Alternatives assessment: Compare substitution candidates on both dimensions
3. Structural optimization: Identify molecular modifications improving Safety AND Sustainability
4. Sectoral benchmarking: Compare substances of the same functional class

Checklist for Responsible Use

✅ Identify contributions: Which substances contribute >10% to toxicity score?

✅ Verify QSe: Do key substances have QSe > 1.48?

✅ Consult Database_ALL: How many species? Which sources?

✅ Sensitivity analysis: Test impact of excluding substances QSe < 1

✅ Document: Mention EF version, sources used, QSe in report

Conclusion: Safe and Sustainable by Design in Action

The JRC approach for Environmental Footprint 3.0 represents a major advance for operationalizing Safe and Sustainable by Design. By systematically exploiting REACH, EFSA and PPDB, the JRC has demonstrated that it is possible to align risk assessment (Safety) and sustainability assessment (Sustainability) on a common database, thus responding to the ambition of the EU Chemicals Strategy for Sustainability.

Key achievements:

1. ✅ Spectacular extension: 6,038 ECOTOX substances (+140%), 3,317 HTOX_nc substances (+437%)
2. ✅ Safety-Sustainability alignment: Same REACH/EFSA data for regulatory thresholds AND PEF factors
3. ✅ Traceability and transparency: Database_ALL allows complete audit of source data
4. ✅ Regulatory coherence: Harmonized Klimisch quality criteria between risk and sustainability

The next step? Develop integrated interfaces allowing industry to assess simultaneously Safety and Sustainability from the design stage of new substances and materials.

References

Main publication:Saouter, E., Biganzoli, F., Ceriani, L., Versteeg, D., Crenna, E., Zampori, L., Sala, S., & Pant, R. (2019). Environmental Footprint: Update of Life Cycle Impact Assessment methods – Ecotoxicity freshwater, human toxicity cancer, and non-cancer. EUR 29495 EN, Publications Office of the European Union, Luxembourg. doi:10.2760/611759

Databases:

• ECHA REACH Database
• EFSA OpenFoodTox
• PPDB Pesticide Properties

About the AuthorExpert in environmental assessment and chemical sustainability, specialized in life cycle assessment and toxicological impact methodologies.

Contact: ssbd-expert.eu

Article published on [DATE] | Category: LCA & Environmental Assessment