Relevance to human health is the importance of factors affecting the well-being and health outcomes of individuals or populations. In toxicology, it concerns whether, and how, hazardous effects seen in test animals may also affect humans. This paper illustrates how human relevance is used in classifying cancer and reproductive toxicity hazards under the EU CLP Regulation, noting that the Regulation does not clearly define the concept. Different hazard endpoints use varying terminology and evidentiary standards, and ECHA's CLP Guidance does not always match the Regulation's wording. Different assessors can also reach different conclusions from the same data. These factors create uncertainty about how human relevance should be interpreted and applied. To improve consistency with the CLP Regulation's aim of protecting workers and consumers, a clear definition and more practical interpretation of human relevance are needed. Human relevance is not binary and is inherently subjective, so both qualitative and quantitative approaches should be combined. The threshold for human relevance should not be so low or vague that classification for human health would only rarely be considered unnecessary.
Occupational case reports and workplace outbreaks can reveal delayed toxic phenotypes that are not readily anticipated from standard exposure metrics. This structured critical review examined human occupational or work-like reports of hepatocellular injury associated with chloroform and 2,2-dichloro-1,1,1-trifluoroethane (HCFC-123), using Pubmed/MEDLINE searching, citation tracking, reference-list screening, and targeted searches. The aim was not to estimate incidence or derive exposure thresholds, but to evaluate how human sentinel reports may inform occupational toxicology and risk management. Across reports, liver injury was often delayed or recognized only after repeated exposure, occurred in only a subset of described or investigated workers, improved after exposure withdrawal, or worsened after continued exposure or return to work. Available exposure estimates were important but often insufficient to characterize timing, apparent selectivity, and recurrence patterns. HCFC-123 reports provided more visible immune-compatible evidence, whereas chloroform reports primarily supported clinical phenotype recognition through workplace clustering, viral hepatitis mimicry, sentinel investigation, and selected host-susceptibility contexts. Idiosyncratic solvent-induced liver injury is proposed as a cautious interpretive framework, not as a validated diagnosis or proven mechanism. This framework may support latency-aware monitoring, early exposure cessation, cautious return-to-work decision, re-exposure avoidance, and future mechanistic evaluation.
New approach methodologies (NAMs) originated in chemical hazard assessment, where two forces operate together: the scale of characterizing many substances, and a long-standing ethical and legal commitment to replacing animal testing. Drug development is weighted differently. It is a translational and decision problem applied to small sets of compounds at pharmacologically active exposures. As NAMs capabilities have expanded, the two domains have been conflated, and framing NAMs against animal studies as a binary choice risks undermining both. The so-called translatability crisis is not a single problem with a single solution. Clinical attrition reflects gaps in mechanistic understanding, exposure-irrelevant study designs, underpowered analyses, and survivorship bias that obscures the screening value of nonclinical studies. Addressing these challenges requires an integrated strategy in which the choice of platform, the design of endpoints, and the rigor of validation are governed by a clearly defined context of use. A biologically relevant model is defined not by its type but by its fitness for the purpose it serves. We argue for a patient-centered nonclinical paradigm that combines animal and non-animal approaches, aligned with recent FDA, NIH, and EMA direction, and note that both domains converge on the same non-dogmatic conclusion.
In 1927, Hermann Muller announced that X-rays induced inheritable gene mutations. Edgar Altenburg indicated that his high-dose ionizing radiation probably "punched structural holes" into chromosomes. To defend his claim Muller attempted to induce "reverse mutations" at the classical white eye-color locus in Drosophila melanogaster, reasoning that a physical deletion could not reconstruct itself. He failed to establish a reproducible model of X-ray-induced reversion. Muller's reverse mutation hypothesis was doomed due to a mechanistic mismatch and a phenotypic "black box". Toxicologically, ionizing radiation predominantly induces double-strand DNA breaks and large-scale deletions rather than the single-nucleotide transitions required for precise back-mutation. Biochemically, the wild-type red eye color of Drosophila is a complex, multigenic trait requiring two parallel biosynthetic pathways (ommochromes and pteridines) feeding into a single bottleneck: an ATP-binding cassette (ABC) transporter encoded by the white gene. Mutating any component of this complex transport machinery disrupts pigment deposition, yielding an identical, white-eyed phenotype. Attempting to restore the precise structural folding of this broken macromolecular pump using random, destructive ionizing radiation was a statistical/structural impossibility. While Altenburg's critique was ultimately validated, Muller's framework of mutational reversibility shaped the evolution of genetic toxicology and the linear non-threshold dose-response model for radiation/chemical carcinogenesis.
Human milk oligosaccharides (HMOs) have gained attention owing to their diverse composition and potential functions, including prebiotic effects, pathogen protection, immune system development, and cognitive development. We produced Disialyllacto-N-neotetraose (DSLNnT) disodium salt, an HMO with beneficial effects such as anti-inflammatory activity, using a unique one-pot enzyme reaction system. However, its safety remains unclear. HMO safety is critical for commercial use in food and pharmaceutical industries. We evaluated the safety of DSLNnT disodium salt using genotoxicity tests (bacterial reverse mutation, in vitro mammalian chromosomal aberration, and in vivo micronucleus tests), a 28-day repeated-dose oral toxicity dose range-finding study in rats, and a 90-day repeated-dose oral toxicity study in rats. DSLNnT disodium salt showed no genotoxic effects in the three genotoxicity tests. In the 90-day repeated-dose oral toxicity study, no DSLNnT disodium salt-related effects were observed in clinical signs, body weight, food intake, water intake, hematology, clinical chemistry, organ weight, necropsy, or histopathological findings at concentrations up to 2000 mg/kg/day. The no-observed-adverse-effect level of DSLNnT disodium salt was 2000 mg/kg/day for males and females. These results provide strong evidence that DSLNnT disodium salt is safe for consumption, supporting its potential commercial application as a functional ingredient in food and pharmaceutical industries.
Read-across is one of the most widely used New Approach Methods (NAMs) for evaluating the safety of a chemical with insufficient toxicological data. However, the identification of analogs suitable for read-across is not always straightforward as it requires understanding of critical chemical and biological features that drive toxicity. The OECD QSAR Toolbox is recommended as a valuable tool to support analog identification. Here, a case study is presented on Benzenesulfonic acid, comparing two different analog identification approaches within the Toolbox. The suitability of analogs identified from each method was evaluated and the approach that retrieved the highest suitability analogs served as criteria for determining the best analog identification approach. The selected approach was then applied to a second case study to assess the systemic toxicity data that will predict the safety of Tropolone. Additionally, a third case study,on 5-Dodecanoylsalicylic acid analyzed analogs identified for skin sensitization using the Defined Approaches for Skin Sensitization (DASS) Automated Workflow (AW) and a new approach developed internally. Through these case studies, this paper aims to analyze and define the steps involved in analog identification and evaluation within the Toolbox as well as critically assessing the corresponding data supporting the read-across prediction of a target chemical toxicity potential.
This work presents a read-across case study to predict the genotoxicity of two previously untested groundwater metabolites of dimethenamid-P (α-chloroacetamide herbicide). Due to synthesis challenges, direct testing of the metabolites was not feasible. Structural space alerts and matched molecular pairs analysis were employed to identify analogues and assess structural features. These analyses indicated that these features were not associated with genotoxicity. However, neither method provided evidence regarding the genotoxic potential of the lactone/hemiacetal moiety unique to the metabolites. To address this data gap, a surrogate compound containing the relevant moiety was synthesised and tested negative for genotoxicity. Computational profiling and metabolic simulation further demonstrated structural and metabolic similarity between the surrogate and the target metabolites. This work shows that structural data gaps can be resolved through the synthesis of a surrogate compound combined with experimental toxicological data. The integrated weight of evidence from synthetic and computational chemistry supports the conclusion that the two dimethenamid-P groundwater metabolites are unlikely to be genotoxic. Overall, this study demonstrates that the use of synthetic chemistry and computational toxicology can provide a robust strategy to address structural moiety data gaps in the genotoxicity evaluation of plant protection products, their active ingredients, and their metabolites.
The zebrafish is formally incorporated into several OECD Test Guidelines for chemical hazard and risk assessment, including TG 203, 210, 212, 215, 229, 230, 234, 236 (FET), and 250 (EASZY), with zebrafish-specific guidelines (ZEOGRT, iFEDT) under development. Related non-zebrafish guidelines addressing endocrine disruption, TG 240 (Japanese medaka) and TG 248/XETA (Xenopus), offer complementary but non-zebrafish coverage. Despite this expanding TG portfolio, significant research and methodological gaps continue to limit the regulatory utility, mechanistic depth, and predictive accuracy of zebrafish-based testing. This critical review systematically evaluates the principal gaps across existing zebrafish-relevant OECD TGs, including the chorion permeability barrier in TG 236, insensitivity to neurotoxicants and metabolically activated compounds, absence of standardized behavioral and molecular endpoints, incomplete thyroid axis coverage, unresolved mixture toxicity guidance, lack of validated omics interpretation frameworks, high inter-laboratory variability, and the non-mandatory status of developmental neurotoxicity testing. It further critiques the slow integration of AOP logic into TG structure, the absent regulatory framework for nanotoxicology and epigenetic endpoints, and methodological shortfalls in chronic and multigenerational testing. Collectively, these gaps constrain the capacity of OECD zebrafish TGs to meet the demands of contemporary chemical safety regulation under REACH, the EU Chemicals Strategy for Sustainability, and global harmonization efforts. Targeted recommendations are provided for each gap category to position zebrafish as a fully integrated, mechanistically credible regulatory test organism.
Permanent dermal fillers based on poly (methyl methacrylate) (PMMA) offer durable aesthetic correction but introduce lifelong patient risk due to their non-resorbable nature. This risk profile raises the question of whether existing regulatory frameworks adequately address the material determinants of long-term biocompatibility. This study presents a comparative analysis of publicly accessible regulatory documentation governing PMMA dermal fillers across three major jurisdictions: the U.S. Food and Drug Administration (FDA), the European Union under the Medical Device Regulation (MDR), and Brazil's National Health Regulatory Agency (ANVISA). Across the public-facing documentation examined, we did not identify consistently accessible, class-wide physicochemical acceptance criteria for key material attributes such as microsphere size distribution, morphology, or residual monomer content. Instead, regulatory frameworks rely primarily on manufacturer-defined specifications evaluated within confidential premarket submissions. This limits the extent to which material safety parameters are externally verifiable and comparable across products. While this approach is well established and effective for many medical devices, its suitability for permanent, micro-particulate implants remains uncertain. Using Brazil, where publicly available documentation enabled direct comparison of two registered PMMA dermal fillers, as an empirical illustration, this study shows how variability in publicly disclosed information may complicate clinician-level interpretation and cross-product comparison. In response, we propose a complementary "material-first" regulatory perspective in which physicochemical parameters are defined through publicly accessible standards, such as pharmacopeial-style monographs. These standards are presented as a conceptual framework to enhance transparency and consistency in risk assessment, rather than as a replacement for existing process-based regulatory systems. For permanent implantable biomaterials, integrating material-level benchmarks with current regulatory practices may improve transparency, comparability, and long-term risk assessment.
The biocidal products regulation (BPR) increases the need for timely assessment of existing biocidal active substances under constraints on animal use, time, and resources. We used tebuconazole, a triazole fungicide relevant to both Korean and European BPR frameworks, as a case study to examine how publicly accessible new approach methodologies (NAMs) data can complement legacy regulatory hazard assessment. We integrated three evidence streams: the historical EU product-type 8 assessment as a legacy benchmark, a structured review of tebuconazole-specific NAM studies published from January 2013 to 2025, and mechanistically annotated ToxCast/Tox21 assay activity. Relative to the legacy benchmark, NAM evidence indicated steroidogenic and thyroid-related perturbations for endocrine endpoints, genotoxicity-related DNA-damage and stress-response signals, and mechanistic support for reproductive/developmental effects. These findings do not justify direct regulatory reclassification. Instead, they illustrate three practical uses of NAM data: prioritizing further evaluation when legacy information is insufficient; identifying supporting information for residual uncertainty and guiding targeted clarification when mechanistic signals emerge; and strengthening mechanistic interpretation.
This commentary advocates for the inclusion of in silico acute oral toxicity (AOT) models within the classification criteria of the Globally Harmonized System (GHS) and Dangerous Goods (DG) Model Regulations. AOT data, traditionally derived from animal studies, are central to hazard classification and regulatory compliance. However, these studies involve severe outcomes including significant toxicity and death as an endpoint and are conducted primarily to support categorical classification rather than for mechanistic understanding. Recent cross-sector evaluations demonstrate that current in silico models provide a high degree of predictivity with accurate or more conservative classification predictions across thousands of chemicals. Incorporating use of these models along with expert review can eliminate the need for routine animal AOT studies while maintaining robust hazard identification and communication practices.
Twenty years ago the European Union introduced regulations known as REACH (Registration, Evaluation, Assessment and restriction of Chemicals). Amongst other things, these required a defined minimum effect level (DMEL) for respiratory allergens. A DMEL of 60 ng/m3 was established for industrial enzymes as the starting point for risk management/occupational hygiene. In the present work, an 11-year set of airborne exposure measurements and employee health information was harvested from three enzyme-producing factory locations in Denmark which use this DMEL as their starting point/action standard. Air sampling data (approximately 45000 points) demonstrated good exposure control. More than 42000 measurements met the standard; only 6.0% exceeded the 60 ng/m3 limit, many by only a modest margin, but which served as a trigger for action. The associated occupational health information confirmed employee health was well controlled, with rates of enzyme sensitisation at 1.2%, and expression of symptomatic allergy kept to a minimal 0.3%. Of the 8 employees with allergy symptoms, 7 continued in their employment, mostly for several years, free of disease. This outcome is consistent with previous reporting and with proper occupational health management, designed to achieve "minimum effect" and thus substantiates the longstanding adoption of the industrial enzyme DMEL of 60 ng/m3.
Traditional smoking and grilling in Benin generate high polycyclic aromatic hydrocarbon (PAH) levels. This study aims to evaluate the effect of an improved indirect-heat kiln (QualiSani) on PAHs exposure and health risk. Fifteen PAHs were measured in 46 samples of grilled pork, smoked fish and smoked-dried fish. Deterministic and 10,000-iteration Monte Carlo simulations estimated margin of exposure (MOE) distributions were used to assess the risk for 300 pork consumers and 186 fish consumers. For carcinogenic compounds such as PAHs, in line with European Food Safety Agency guidance, MOE ≥ 10,000 indicates low concern. Thanks to the QualiSani kiln, PAHs levels were reduced to about 98% (for grilled pork) and 99% (for smoked fish and smoked-dried fish). In both deterministic and probabilistic risk assessments, all the MOEs were well above 10,000 except for ∼2% of high-end consumers showing low health concern for human health. In conclusion, the use of the QualiSani kiln in replacement of traditional techniques allows to lower to safe levels the Beninese consumer exposure to PAHs through grilled and smoked food consumption.
This study aims to evaluate the prenatal developmental toxicity potential of D-allulose produced by genetically engineered Escherichia coli in rats. Given the increasing use of D-allulose in food products, a thorough toxicological assessment is essential to ensure its safety. A prenatal developmental toxicity test in Sprague-Dawley rats was conducted in accordance with the OECD Guidelines Test No. 414 (prenatal developmental toxicity study) and GB 15193.14-2015 (National food safety standards teratogenicity test of the People's Republic of China). Pregnant female rats received repeated doses of 1250, 2500, and 5000 mg/kg bw D-allulose by gavage on Gestation Days (GD) 6-15. A vehicle control (distilled water) and a positive control (cyclophosphamide) were also included. On GD 20, pregnant rats were euthanized, and fetuses were examined for external, soft tissue and skeletal abnormalities. No treatment-related anomalies were observed, indicating that D-allulose did not exhibit prenatal developmental toxic effects. The No-Observed-Adverse-Effect-Level (NOAEL) was determined to be 5000 mg/kg/day.
International standards describing extractables and leachables (E&L) testing for chemical analyses are not prescriptive allowing much scope for interpretation. Balancing the requirements of regulatory agencies with issues of laboratory-based practicality and feasibility is challenging in the E&L analysis of advanced wound dressings, specifically the expectation of testing under ‘exhaustive’ conditions in three solvents (polar, semi-polar, non-polar) when not all of the test conditions/solvents are compatible with common wound dressing materials (e.g. silicone adhesives, polyurethanes). The aim of this study is to describe a rigorous E&L testing framework, aligning with the clinical usage of wound dressings whilst remaining strongly aligned with regulatory standards. Five major practical challenges were identified relating to the selection of suitable E&L solvents, including: selection of extraction conditions, inappropriate swelling, loss of device integrity, formation of particulates, and choice of appropriate analytical techniques. A systematic process was developed to identify acceptable/compatible solvents and justify the omission of solvents shown to be incompatible. A test case was developed to demonstrate how the framework can be adopted in support of real-life regulatory submissions. This systematic approach to design of E&L analysis of wound dressings may be useful in streamlining future regulatory submissions for medical devices including advanced wound dressings.
Insecticidal proteins expressed in genetically modified (GM) crops represent a sustainable and effective alternative for controlling insect and nematode pests and constitute a significant portion of the GM crop portfolio. Proteins derived from Bacillus thuringiensis (Bt) are currently the most extensively studied and widely used, particularly those belonging to the Cry class. In addition to Cry proteins, other insecticidal proteins from different classes have been developed. However, for many of these, information on their mode of action (MoA) and host receptors remains limited. Over the past two decades, the European Food Safety Authority (EFSA) has evaluated more than 70 applications for GM crops expressing 24 different insecticidal proteins. These proteins belong to various classes, and the level of available information regarding their MoA varies considerably. This review aims to summarise the evidence currently available to EFSA on the MoA of insecticidal proteins expressed in GM crops. It also seeks to identify areas where further understanding of receptor binding and protein function could support the continued development of safety assessment strategies, with the goal of reducing reliance on animal studies as part of a more refined risk assessment approach.
Permitted Daily Exposure (PDE) limits established by ICH guidelines and Occupational Exposure Limits (OELs) adopted in the European Union are central to controlling human exposure to pharmaceutical impurities and industrial chemicals, yet they are derived for different exposure scenarios, target populations, and protection goals. We compiled data for 90 chemicals with both PDEs (Q3C, Q3D, Q3E, M7) and EU OELs, normalized PDEs to a 70 kg worker, and converted them to airborne equivalents assuming 10 m3 inhalation over an 8-h workday. OEL/PDE ratios ranged from 0.04 to 8,929, with an overall geometric mean of 57.1, indicating that OELs are generally several-fold higher (less stringent) than airborne PDE equivalents; eleven substances exceeded ratios of 1,000, most of them M7 mutagenic impurities. Five substances, however, showed more stringent OELs than the corresponding PDE-derived values. These patterns reflect systematic differences in risk-assessment objectives and support structured, Threshold of Toxicological Concern (TTC)- and hazard-banding-based approaches for deriving default OELs where none exist.
Petroleum substances (PS) are complex hydrocarbon substances that present major challenges for toxicity assessment under REACH regulations. In vivo prenatal developmental toxicity (PNDT) testing is often mandated but remains resource-intensive, underscoring the need for New Approach Methodologies (NAMs) such as mechanistic read-across and in vitro models. The toxicity of PS is largely attributed to their polycyclic aromatic hydrocarbon (PAH) content, known to act through activation of the aryl hydrocarbon receptor (AhR). AhR activation induces cytochrome P450 enzymes that metabolize PAHs into reactive intermediates capable of eliciting developmental toxicity. To investigate the role of AhR in this adverse outcome pathway, we orally exposed wild-type and AhR-knockout (AhR-KO) Sprague-Dawley rats to Benzo[a]Pyrene (BaP), a prototypical PAH. BaP exposure caused marked embryonic resorptions and delayed ossification in wild-type rats, effects absent in AhR-KO animals, demonstrating that BaP-induced developmental toxicity is predominantly AhR-dependent. Transcriptomic profiling of blood, kidney, liver, and thymus revealed distinct AhR-mediated expression of Cyp1a1/1a2 in wild-type rats, whereas AhR-KO rats upregulated alternative detoxification pathways (Cyp2b2, Cyp3a23-3a1). These findings confirm AhR as the molecular initiating event in PAH-driven developmental toxicity and support the application of mechanistic data to enable efficient, less animal-intensive hazard assessment of petroleum substances.
Acute oral toxicity (AOT) testing is widely used for hazard identification in chemical safety assessment, but traditional animal methods are still implemented despite their scientific limitations and ethical implications. Despite the availability of non-animal methods, regulatory uptake has been inconsistent, also due to a lack of structured guidance. To address this, the Animal-Free Safety Assessment Collaboration has developed a science-based Decision Tree (DT) aligned with the Organisation for Economic Co-operation and Development (OECD) Integrated Approaches to Testing and Assessment (IATA) framework. IATA integrate existing information in a Weight of Evidence (WoE) assessment with the possible generation of new data appropriate to the decision context. The DT is built as a specific IATA to provide a transparent, modular workflow integrating in silico tools, in vitro assays, and WoE strategies to support AOT hazard identification for regulatory decision-making, such as European Union's Registration, Evaluation, Authorisation, and Restriction of Chemicals (REACH) Regulation, ensuring that animal testing is used only as a last resort. The DT aims to inform regulatory guidance updates and promote acceptance of non-animal approaches by formalising decision-making to bridge the gap between legal mandates to reduce animal testing and current regulatory practice, still heavily reliant on animals, supporting the transition toward improved chemical safety evaluations.
