Patlewicz G, Mihalchik A, Thompson CM, East A, Borghoff S. Investigating the feasibility of deriving a new environmentally relevant Thresholds of Toxicological Concern (TTCs) from estimated human equivalent benchmark dose (eBMDHED) values. Abstract 4013, Society of Toxicology 65th Annual Meeting, San Diego, CA, March 2026.
Abstract
Background and Purpose: The Threshold of Toxicological Concern (TTC) is a pragmatic approach used to establish safe thresholds below which there is likely no appreciable risk to human health. Kroes et al. (2004) presented a tiered TTC approach that established several human exposure thresholds over several orders of magnitude, ranging from 0.0025μg/kg-day to 30 μg/kg-day. The lowest tier is for substances that raise a concern of genotoxicity based on structural alerts for genotoxicity. For substances without structural alerts for genotoxicity, there are a series of non-cancer TTC tiers, which are based on the Cramer et al. (1978) decision tree. Derivation of TTC values for each of these Cramer classes originates from work by Munro et al. (1996) who compiled a dataset of NOELs for 613 substances that had been tested in subchronic, chronic and reproductive-developmental repeat dose oral toxicity studies. In Nelms et al. (2019), an evaluation was performed to explore whether TTC values derived from toxicity data on environmentally-relevant substances as compiled in the EPA Toxicity Values DB (ToxValDB) would give rise to comparable TTC values as those derived by Munro et al. (1996). In that study, statistical outliers were removed, but no effort was made to review the quality of the individual studies. TTC values are calculated by taking the 5th percentile of the NOAEL values and then adjusted 100-fold by default intra and interspecies uncertainty factors to derive a final value. More recently, efforts by Aurisano et al. (2023) and Harrill et al. (2025) have attempted to develop curated datasets restricted to effect levels, exposure routes, study designs and species relevant for deriving toxicity values. The effect levels are then adjusted to chronic human equivalent benchmark doses using the World Health Organization (WHO) framework which considers different conceptual models and magnitudes of effects for different types of endpoints within toxicity studies. Methods: In this study, the feasibility of deriving new oral TTCs was investigated using estimated human equivalent benchmark dose (eBMDHED) values from EPA’s ToxValDB v96_1. Study level eBMDHED values were generated using the R code from https://github.com/USEPA/toxvaldb_dcap_prep. Chemical level values (i.e., those representative of the study level eBMDHED values) were derived by taking the 25th percentile of the fitted normal distribution. Chemical structural information was pulled from the EPA CompTox Chemicals Dashboard resulting in a dataset of 4076 substances. The chemical set was processed through a computational implementation of Kroes et al. (2004) workflow. The 1271 substances assigned to the Cramer structural classes were then processed through the Toxtree Cramer module. The 5th percentile of the empirical cumulative distribution function for each Cramer structural class was then derived. Results: The Cramer I class 5th percentile was estimated as 4.24 mg/kg-day. Applying a 10-fold intraspecies factor, resulted in a TTC of 0.424 mg/kg-day which was 14 times less conservative than the current TTC for Cramer I (0.03 mg/kg-day). The TTC derived for Cramer III gave rise to a value of 0.014 mg/kg-day, which was 9.4 times higher than the default Cramer Class III TTC value (0.0015 mg/kg-day) derived from the Munro dataset. Given there were only 56 substances in Cramer II, no TTC value was derived. Conclusions: Ongoing work will consider other ways to subcategorize the dataset beyond reliance on existing Cramer structural classes. TTC values derived from human equivalent chronic values (eBMDHED) offer promise of establishing exposure thresholds that are more human relevant while still protective of human health.
