Application of QSTRs in the selection of a surrogate toxicity value for a chemical of concern.

As part of the EPA's mission to protect the environment, chemicals of concern (CoCs) at Superfund or other hazardous waste sites are cleaned up based on their potential toxicity to humans and the surrounding ecosystem. Oftentimes, there is a lack of experimental toxicity data to assess the health effects for a CoC in the literature. This research describes a method using Quantitative Structure Toxicity Relationships (QSTRs) for identifying a surrogate chemical for any given CoC. The toxicity data of the surrogate chemical can then be used to rank hazardous waste-site chemicals prior to cleanup decisions. A commercial QSTR model, TOPKAT, was used to establish structural and descriptor similarity between the CoC and the compounds in the QSTR model database using the Oral Rat Chronic LOAEL model. All database chemicals within a similarity distance of < or = 0.200 from the CoC are considered as potential surrogates. If the CoC fails to satisfy model considerations for the LOAEL model, no surrogate is suggested. Potential surrogates that have toxicity data on Integrated Risk Information System (IRIS), Health Effects Assessment Summary Tables (HEAST), or National Center for Environmental Assessment (NCEA) provisional toxicity value list become candidate surrogates. If more than one candidate surrogate is identified, the chemical with the most conservative RfD is suggested as the surrogate. The procedure was applied to determine an appropriate surrogate for dichlorobenzophenone (DCBP), a metabolite of chlorobenzilate, dichlorodiphenyltrichloroethane, and dicofol. Forty-seven potential surrogates were identified that were within the similarity distance of < or = 0.200, of which only five chemicals had an RfD on IRIS, HEAST, or on the NCEA provisional toxicity value list. Among the five potential surrogates, chlorobenzilate with an RfD of 2 x 10(-2) mg/kg-day was chosen as a surrogate for DCBP as it had the most conservative toxicity value. This compared well with surrogate selection using available metabolic information for DCBP and its metabolites or parent compounds in the literature and the provisional toxicity value of 3 x 10(-2) mg/kg-day that NCEA developed using a subchronic study.