In this paper, the results obtained by five independent methods for the quantification of fumonisins B(1), B(2), and B(3) in raw maize are reported. Five naturally contaminated maize samples and a reference material were analyzed in three different laboratories. Although each method was validated and common calibrants were used, a poor agreement about fumonisin contamination levels was obtained. In order to investigate the interactions among analyte and matrix leading to this lack of consistency, the occurrence of fumonisin derivatives was checked. Significant amounts of hidden fumonisins were detected for all the considered samples. Furthermore, the application of an in vitro digestion protocol to raw maize allowed for a higher recovery of native fumonisins, suggesting that the interaction occurring among analytes and matrix macromolecules is associative rather than covalent. Depending on the analytical method as well as the maize sample, only 37-68% of the total fumonisin concentrations were found to be extractable from the samples. These results are particularly impressive and significant in the case of the certified reference material, underlying the actual difficulties in ascertaining the trueness of a method for fumonisin determination, opening thus an important issue for risk assessment.