Commission Regulation (EC) N° 2074/2005 recognises the biological method as the reference method for Paralytic Shellfish Poisoning (PSP) toxins detection in molluscs. It was amended by Commission Regulation (EC) N° 1664/2006 that accepted the so-called Lawrence method as an alternative to the reference method. The goal of this study was to compare AOAC Official Methods of Analysis 959.08 (Biological method) and 2005.06 (Prechromatographic Oxidation and Liquid Chromatography with fluorescence detection) in samples with different toxin profiles. The influence of extraction solvent in the total samples toxicity was also evaluated. A total of 40 samples including mussels, clams, scallops, razor-clams, cockles, oysters and barnacles were analysed by both official methods. Samples were selected with Alexandrium and Gymnodinium toxic profiles, from different origin and including several presentations: fresh, frozen, canned and boiled. Acetic and hydrochloric acid extractions were performed in all samples and the extracts were simultaneously analysed by both methods. Most samples were naturally contaminated and two samples were spiked. Comparison of both official methods, mouse bioassay (MBA) with HCl extraction and Liquid Chromatography with fluorescence detection (HPLC-FLD) with acetic acid extraction, led to an 85% of consistent results regarding compliance with legal limit, including samples below and above it. The linear correlation coefficient was r² = 0.69 and the paired t test (two tails, α = 0.05) indicated that there were not significant differences among both sets of data. Nevertheless, toxicity differences were found in several samples. In 15 out of 18 shellfish with a Gymnodinium toxic profile, higher toxicity levels were obtained by MBA. This fact was more evident in 7 samples, partially related to the lack of standards and the impossibility of analysing dc-NEO, C1, 2 and GTX6 at the beginning of the study. However, other factors concerning the extraction and SPE clean-ups steps may also contribute. By contrast, 9 samples presented a much higher total toxicity by HPLC-FLD than by MBA. These higher results obtained by HPLC-FLD could not only be due to the use of the highest toxicity equivalency factor (TEF) for isomers oxidated into products that coelute when total toxicity of these samples were calculated. Further analyses of results obtained by HPLC-FLD and by MBA with both extracts were done separately.