Bioassay screening for toxicants in oil samples from the toxic-oil syndrome outbreak in Spain

A bioassay screening procedure was developed using an agar-plate inhibition assay and applied to a set of 32 oil samples obtained during investigations of the toxic-oil syndrome in Spain. After blind-decoding, it was noted that the assay had singled out the case-associated sample as the most toxic oil from the 32-sample set. Five other oil samples were also selected by the assay. Further evaluation showed that these case-associated, suspect case-associated, and miscellaneous oils had significantly more peroxidation than the controls, as indicated by the thiobarbituric acid assay. Products reactive with dinitrophenylhydrazine were present in greater amounts in the assay-selected oils, also indicating larger amounts of aldehydic and ketonic peroxidation products. A volatile peroxidation product, 4-hydroxy-2,3-nonenal (4-hydroxynonenal), was identified and was very toxic to the bioassay screen. The assay-selected oils had significantly greater amounts of this compound. Heating the toxic oils either removed or destroyed the toxic components as revealed by bioassay testing. The 4-hydroxynonenal and other unidentified toxic products were present in the assay-selected oils before heating and absent after heating. Toxic peroxidation products apparently caused the observed response in the assay-selected oils. These findings may indicate that peroxidation, or some type of free radical process, was associated with the toxicity of these oils to humans. The bioassay procedure should be applied to a larger set of epidemiologically documented oils to determine if this screening process can contribute to a better understanding of the toxic-oil syndrome.