Contaminated soils represent a potential health risk for the human population. Risk assessment for humans requires specific methods, which must reflect the peculiarities of human behaviour, physiology and biochemistry with respect to contaminant uptake and processing. Biomarkers of effect or exposure have become an appropriate tool. Organic pollutants influence the expression profile of cytochromes P450 (CYP), and CYP1A1 has been shown to be a suitable biomarker for polycyclic aromatic hydrocarbons (PAH). The latter are widely distributed in soils and constitute an important soil contamination. Upon intake of PAH-contaminated soils, CYP1A1 is induced in various organs of rats and minipigs. Increased CYP1A1-levels in lung, kidney and spleen, after oral soil intake, indicate that contaminants escape the primary duodenal and hepatic metabolism and reach further organs. Dose-response relationships reveal that induction effects are to be expected in children based on known exposure conditions. Generally, CYP1A1-induction does not correlate with results of toxicity tests with lower organisms, performed with the same soils. The organic carbon content is largely responsible for this discrepancy. It severely affects the toxicity of soil bound PAH for microorganisms, but obviously affects the mobilization efficiency for PAH in the gastro-intestinal tract of mammals to a minor extent. Soil remediation by different methods may result in a significant reduction of the PAH content and of toxicity. Ingestion of remediated soils by rats shows, however, that the induction potential for CYP1A1 is only slightly decreased after remediation. This means that the major inducing components resist biological remediation or soil washing and remain in the soil. Because data obtained with experimental animals form the guiding principle for in vitro tests to be developed, the suitability of the animal model used for extrapolations to humans has to be proven. Upon soil ingestion, minipigs show a tissue-specific response pattern, which substantially differs from that of rats, which are widely used as animal models. It is not known which response pattern resembles that of man. In summary, cytochromes P450, in particular CYP1A1, are suitable biomarkers to assess the bioavailability of soil bound contaminants and their effects on mammalian species. There are, however, a number of questions to be answered in order to develop an in vitro test for human risk assessment. This concerns, for example, the identification of the suitable animal model, the identification of biomarkers for other contaminants and concepts to transpose the in vivo data to in vitro technologies or to mathematical modelling. (C) 2004 Elsevier Ireland Ltd. All rights reserved.