Publications : 2022

Bates C, Russell A, Dotson GS, Vincent M, Lotter J, Maier MA. Establishing OELs for sensory irritants with limited data using predictive and in silico models. Poster presentation P126 at the Society of Toxicology (SOT) Annual Meeting & ToxExpo, San Diego, CA. Toxicologist Late-Breaking Supp. 168(1):17. Abstract 5023. March 2022.


Sensory irritation is one of the most common health endpoints that serve as the basis for OELs. Numerous research efforts have been conducted to derive OELs based on predictive methods, various data types, and mathematical models for sensory irritation. One study by Schaper (1993) described a linear relationship with high correlation between measured RD50s and ACGIH TLVs for airborne chemical irritants and demonstrated the use of the RD50 as a benchmark for occupational exposures. There is a need to develop a predictive model for deriving short-term exposure limits (STELs) for sensory irritants. The aim of our study was to establish a model capable of correlating the relationship between RD50 and STELs in order to derive OELs for sensory irritants. A focused review of scientific literature was conducted. The results of this review provided evidence that the potency and onset of sensory irritation is not governed by Haber’s Law, but instead is governed primarily by concentration. The review also resulted in the identification of a NTP database based on Schaper (1993) that included chemicals with both RD50 and established STEL. Some of the chemicals included in the NTP database had multiple RD50s and STELs. In such cases, each RD50 and STEL was analyzed separately. Based on these criteria, a total of 136 observations associated with 47 unique chemicals were eligible for inclusion in the analysis. Further, the data underwent log transformation because it was determined that RD50s were log normally distributed. The correlation between LnSTELs and LnRD50 values were determined via a linear model using R software. A strong correlation between RD50s and STELs was identified, with a predictive equation of Ln (STEL) = 0.77 * Ln (RD50) – 1.49 and an R2 value of 0.80. Sensitivity analyses, including evaluation of outliers, showed that the model was robust and independent from extreme values. This model supports the use of RD50s to derive STELs for chemicals without existing exposure recommendations. Further, for chemicals that are sensory irritants with no data, predicted RD50s from in silico QSAR models could be used to derive STELs. For an example, a case study using acetone demonstrated the application of the model (predicted STEL: 758 ppm; actual STEL: 750 ppm). Hence, in silico methods and statistical modeling can present a path forward for establishing reliable OELs and improving worker safety and healt