You are here:
Meta-analysis of aquatic chronic chemical toxicity data
Citation:
Hoff, D., G. Elonen, T. Dawson, AND Dave Mount. Meta-analysis of aquatic chronic chemical toxicity data. SETAC North America, Salt Lake City, UT, November 01 - 05, 2015.
Impact/Purpose:
not applicable
Description:
Chronic toxicity data from the open literature and from tests submitted for pesticide registration were extracted and assembled into a database, AquaChronTox, with a flexible search interface. Data were captured at a treatment and, when available, replicate level to support concentration response modeling and a variety of meta-analysis. An initial meta-analysis looked at chronic toxicity data specifically for acetylcholinesterase inhibiting (AChEI) pesticides in the organophosphate and carbamate groups. The purpose for this analysis was to determine whether acute-chronic ratios (ACRs) or other mechanism-specific parameters could be developed that would allow more robust extrapolation of toxicity data. When compiled across all AChEI chemicals and all species, ACRs varied widely, by almost four orders of magnitude. Invertebrates showed higher relative sensitivity to AChEI chemicals than did fish, even though ACRs for invertebrates were generally much smaller for invertebrates than for vertebrates. This apparent dichotomy arises from the much lower acute sensitivity of fish compared to invertebrates, resulting in lower chronic sensitivity even with higher ACRs. Another aspect of the analysis evaluated whether ACRs generated using acute values that were “matched” to the chronic data set (e.g., conducted in the same laboratory) had different distribution than those that were not matched. Interestingly, ACRs having unmatched acute values were not any more variable, but were generally larger than those for matched acute and chronic data sets; the reasons for this difference are not immediately clear. Regression analysis conducted on these data sets indicated that EC20 values generally fell between NOEC and LOEC values as determined by hypothesis testing; few EC20 values fell below NOEC values, and generally only in data sets with unusual features (e.g., high control variability).