Assessing the bioavailability of complex petroleum hydrocarbon mixtures in sediments

Several experimental methods have been developed to assess the bioavailability of individual organic compounds. So far none of them has however been applied to complex mixtures, such as oil (petroleum hydrocarbons), which is an ubiquitous pollutant. In the present study, we tested the potential of five of these experimental methods and that of a model approach to predict bioaccumulation of oil in the aquatic worm Lumbriculus variegatus exposed to 14 field-contaminated sediments. Actual and predicted bioaccumulation were compared in terms of both total bioaccumulative petroleum hydrocarbon concentrations and the relative distribution pattern of separate boiling point fractions (hydrocarbon blocks). None of the experimental methods was able to directly assess bioaccumulation in L. variegatus and correction factors were needed to match predicted and actual concentrations. These factors appeared concentration-dependent for solid phase micro extraction (SPME) and extractions with Tenax and cyclodextrin, most probably due to artifacts. Moreover, the hydrocarbon block pattern produced by these methods considerably differed from the pattern observed for worms; an additional reason for disqualification also applying to headspace-SPME. In contrast, the pattern produced by polyoxymethylene solid phase extraction (POM-SPE) closely mimicked the worm pattern and a sediment, hydrocarbon block, and concentration-independent correction factor (17) could be derived, based on which actual bioaccumulation could be predicted within a factor of 3. Finally, the model predicted bioaccumulation directly within a factor of 2. The accompanying hydrocarbon block pattern however deviated significantly more from the worm pattern than the POM-SPE pattern did. We therefore conclude that POM-SPE may be the overall best approach for predicting bioaccumulation of complex hydrocarbon mixtures in aquatic worms, all the more since an experimental approach will implicitly capture all factors determining bioavailability, which may prove difficult through a modeling approach.