Meta-Analysis Comparing Nominal and Measured Concentrations of Perfluorooctanoic Acid and Perfluorooctane Sulfonate in Aquatic Toxicity Studies Across Various Experimental Conditions.

Perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) are among the most frequently detected chemicals among the per- and polyfluoroalkyl substances in aquatic environments. Because of their high detection frequency, persistence, and potential toxicity, interest in both PFOA and PFOS has increased in recent years. However, a substantial number of PFOA and PFOS toxicity tests only report nominal, or unmeasured, treatment concentrations, which may complicate the determination of protective values. In addition, previous literature has indicated that differences between nominal and measured concentrations of both PFOA and PFOS could be linked to experimental conditions (e.g., feeding regimes for test organisms, test vessel material [glass or plastic], use of solvent, and the presence of substrate). Therefore, this critical review examined whether nominal and measured concentrations were in close agreement with each other among the current PFOA and PFOS aquatic toxicity literature and if experimental conditions were associated with any observed differences. Nominal and measured concentrations in the current PFOA and PFOS aquatic toxicity literature generally displayed a high degree of linear correlation and relatively low median percent differences. Correlations between measured and nominal concentrations were >0.98 for PFOA and >0.95 for PFOS in freshwater tests across experimental conditions. For saltwater tests, correlations of >0.84 were observed for PFOA and PFOS (separate and combined) across experimental conditions. While measured PFOA and PFOS toxicity tests are generally preferred, the present meta-analysis demonstrated that experimental conditions had little influence on observed discrepancies between nominal and measured concentrations, with the exception of PFOS saltwater tests and PFOA and PFOS freshwater studies that contained substrate. Unmeasured tests with these conditions should be considered carefully based on project needs, with the caveat that the data sets for these two experimental conditions were limited. Environ Toxicol Chem 2023;42:2289-2301. Published 2023. This article is a U.S. Government work and is in the public domain in the USA.