Monitoring programmes for paralytic shellfish poisoning toxins in bivalve molluscs still rely heavily on the use of mouse bioassays (MBA) for consumer protection. A high-performance liquid chromatography (HPLC) methodology (Lawrence method) was implemented in 1996 in the Portuguese monitoring programme as a complementary means of analysis. Comparison between MBA and HPLC was done at the time only by a qualitative approach due to the scarce number of positive samples tested. More quantitative data were obtained recently when studying toxin profiles in Moroccan shellfish, and the correlation found between these two methodologies is reported here for the first time. Two different matrices were studied: blue mussel and the giant cockle Acanthocardia tuberculatum. A good linear correlation was obtained for both matrices. However, a second-degree polynomial best fitted the data at both low and high extremes of toxicity. According to the HPLC quantitative results, 13% of false-negatives could be obtained by MBA due to an underestimation of toxicity near the limit of detection of the MBA. Difficulties on relying solely on HPLC for consumer protection have been aroused with uncommon matrices, such as imported clams or crustaceans, due to the presence of high concentrations of interfering compounds. The solid-phase extraction step of the Lawrence method was implemented to eliminate an unknown compound that could be mistaken for saxitoxin, and an 80% reduction of another common unknown compound eluting close to decarbamoylsaxitoxin. The implementation of the HPLC methodology achieved so far allows a high degree of consumer protection without the need to resource to animal sacrifice.