Juvenile Senegalese soles (Solea senegalensis) were exposed to estuarine sediments through 28-day laboratory and in situ (field) bioassays. The sediments, collected from three distinct sites (a reference plus two contaminated) of the Sado Estuary (W Portugal) were characterized for total organic matter, redox potential, fine fraction and for the levels of metals, polycyclic aromatic hydrocarbons (PAHs) and organochlorines, namely polychlorinated biphenyls (PCBs) and dichloro diphenyl tricholoethane plus its main metabolites (DDTs). Genotoxicity was determined in whole peripheral blood by the single-cell gel electrophoresis (SCGE or "comet") assay and by scoring erythrocytic nuclear abnormalities (ENA). Analysis was complemented with the determination of lipid peroxidation in blood plasma by the thiobarbituric acid reactive substances (TBARS) protocol and cell type sorting. The results showed that exposure to contaminated sediments induced DNA fragmentation and clastogenesis. Still, laboratory exposure to the most contaminated sediment revealed a possible antagonistic effect between metallic and organic contaminants that might have been enhanced by increased bioavailability. The laboratory assay caused a more pronounced increase in ENA whereas a very significant increase in DNA fragmentation was observed in field-tested fish exposed to the reference sediment, which is likely linked to increased lipid peroxidation that probably occurred due to impaired access to food. Influence of natural pathogens was ruled out by unaltered leukocyte counts. The statistical integration of data correlated lipid peroxidation with biological variables such as fish length and weight, whereas the genotoxicity biomarkers were more correlated to sediment contamination. It was demonstrated that laboratory and field bioassays for the risk assessment of sediment contamination may yield different genotoxicity profiles although both provided results that are in overall accordance with sediment contamination levels. While field assays may provide more ecologically relevant data, the multiple environmental variables may produce sufficient background noise to mask the true effects of contamination.