Application of a Tenax model to assess bioavailability of PCBs in field sediments

Abstract

Bioavailability has been estimated in the past using equilibrium partitioning-based biota-sediment accumulation factors. These values are not always reliable using field-collected sediments, however, likely due to varying amounts of different organic carbons, particularly black carbon, in sediments. Therefore, improving approaches to better evaluate contaminant bioavailability in sediment are needed. In the present study, a literature-based model was constructed that relied on both laboratory-exposed and field-collected oligochaete bioaccumulation data. The model system used 24-h Tenax extraction data paired with bioaccumulation tests using oligochaetes to establish the ability and utility of the biomimetic extraction. The model was then tested to confirm its utility and reliability to estimate bioavailability of oligochaetes exposed to polychlorinated biphenyl (PCB) contaminated sediments from the Ottawa River in Ohio, USA. The model correctly identified the bioaccumulation of PCBs for 94.9% of the data and 97% of the dioxin-like congeners. The mono- and di-substituted chlorinated biphenyls were the least well described, but the estimates were conservative, for example, the model overpredicted bioaccumulation. Thus, the Tenax model was robust and reliable across a wide range of sediment characteristics for estimating PCB bioaccumulation in oligochaetes.

Tenax extraction as a simple approach to improve environmental risk assessments

“…Although the ability of Tenax extractions to predict bioaccumulation of pyrethroids, despite biotransformation, has previously been demonstrated [22], it was unclear how biotransformation would impact the ability of the bioaccumulation Tenax model to predict bioaccumulation. The results of the comparison showed that 100% of the pyrethroid data fell within the 95% confidence interval of the bioaccumulation Tenax model, suggesting that this model could account for the bioaccumulation of a variety of chemical classes by multiple species [22,25]. The pyrethroid data considered bioaccumulation of permethrin and bifenthrin by L. variegatus and Hexagenia sp.…”

“…In a log-log linear regression, the bioaccumulation Tenax model correlates the concentration in the organism on a lipid-normalized basis against the amount recovered from Tenax normalized to the amount of sediment organic carbon in the sample. Where the model was less successful at estimating bioaccumulation within the confidence interval, it overestimated, rather than underestimated, exposure [25]. Specifically, 95% of the bioaccumulation data fell within the 95% confidence limits of the model.…”

“…Furthermore, Tenax-extractable concentrations tend to have very low variation among replicates [22,24,25,[35][36][37]. The solvent necessary for the extraction of the Tenax beads is generally only 15 mL to 30 mL per sample, depending on the Tenax mass used in the extraction.…”

“…Correlations between Tenax-extractable concentrations and tissue concentrations are uniform across compounds and species in some instances [22] but not others [34]. However, correlations can often be improved by further separating compounds within a class (e.g., PCBs by degree of chlorination) or by test species [25,34]. Although it was initially hypothesized that Tenax-extractable concentrations may not be able to estimate bioaccumulation for compounds and/or species where significant biotransformation occurred [34], significant linear relationships between Tenaxextractable concentrations of pyrethroid insecticides and tissue concentrations of these compounds in Lumbriculus variegatus and Hexagenia sp.…”

“…These data were used to create a single correlation, the bioaccumulation Tenax model, for all compounds ( Figure 1) [25,34]. The model demonstrated that bioaccumulation for exposures to sediments from an independent system, the Ottawa River, Ohio, USA, were well estimated from the model (Figure 1) [25]. The model demonstrated that bioaccumulation for exposures to sediments from an independent system, the Ottawa River, Ohio, USA, were well estimated from the model (Figure 1) [25].…”