A new rapid assay for the okadaic acid group of toxins, based on lateral flow immunochromatographic (LFIC) test strips developed by Jellett Rapid Testing Ltd., was assessed on naturally contaminated bivalves from the Portuguese coast. One prototype was evaluated using samples harvested during 2005, extracted with 80% methanol, followed by dilution with the running buffer of a methanolic extract after alkaline hydrolysis for esters. The second prototype was assessed using samples harvested during 2006, extracted with 100% methanol and, after alkaline hydrolysis, the methanol was evaporated by a nitrogen stream followed by re-suspension with the running buffer. The first prototype failed to detect 20% of samples that were positive by LC-MS in the range 160-480 microg kg(-1), and were classified as negative or trace level by LFIC. The presence of methanol in the extracts made correct detection of toxins more difficult. The second prototype classified as positive all samples above 160 microg kg(-1), as confirmed by LC-MS. However, in the second prototype, matrix effects were a major drawback and led to 45% false positives, particularly for mussels, due to compounds in shellfish extracts interfering with the antibodies and reducing the test line intensity. Extraction with a higher percentage of methanol was thought responsible for these matrix effects. Regarding sample migration, both prototypes needed one hour before reading. In an attempt to speed-up sample preparation, a direct digestion of bivalve tissues with sodium hydroxide was evaluated. Low recoveries for esters were found by LC-MS with this hydrolysis technique compared to conventional hydrolysis of methanolic extracts. While prototype A was not sensitive enough, prototype B had too many false positives to be of use to the shellfish industry or in a monitoring program.