Bisphenol A (BPA) is widely used in the manufacturing of polycarbonate plastic food and drink packaging. Possessing a weak estrogenic activity, BPA is listed among a growing list of endocrine disrupting compounds. In this study, a polyclonal anti-BPA antibody was obtained by immunization with BPA-monocarboxymethylether covalently linked to BSA. The antibody demonstrates negligible cross-reactivity with most analogous BPA phenolic structures, and no cross-reactivity with endogenous steroids. An extraction step with ethyl acetate minimized matrix effects and allowed the BPA measurement in plasma and other biological samples. Recovery after loading test was 96 +/- 4% and dilution tests had a linear profile (r2 > 0.93). The limit of detection of the BPA RIA was 0.08 microg L(-1) with an IC50 of 1.25 microg L(-1). The intra- and inter-assay coefficients of variation were 5.6 and 8.6%, respectively at a BPA concentration of 0.7 microg L(-1) and 6.9 and 5.7% at a BPA concentration of 1.3 microg L(-1). A significant correlation was found between the values obtained by the RIA and HPLC-MS (r2 = 0.92) or HPLC coupled to a fluorescence detector (r2 = 0.80). In conclusion, we described a BPA-RIA that is a suitable tool for evaluating human exposure to BPA.