Okadaic acid as a major problem for the seafood safety (Mytilus galloprovincialis) and the dynamics of toxic phytoplankton in the Slovenian coastal sea (Gulf of Trieste, Adriatic Sea)

This article presents the first results on shellfish toxicity in the Slovenian sea (Gulf of Trieste, Adriatic Sea) since the analytical methods for the detection of biotoxins (PSP, ASP, DSP and other lipophilic toxins) in bivalve molluscs were included in the national monitoring program in 2013. In addition to toxins, the composition and abundance of toxic phytoplankton and general environmental characteristics of the seawater (surface temperature and salinity) were also monitored.

During the 2014–2019 study period, only lipophilic toxins were detected (78 positive tests out of 446 runs), of which okadaic acid (OA) predominated in 97 % of cases, while dinophysistoxin-2 and yessotoxins only gave a positive result in one sampling event each. The number of samples that did not comply with the EC Regulation for the OA group was 17 or 3.8 % of all tests performed, all of which took place from September to November, while a few positive OA tests were also recorded in December, April, and May. This toxicity pattern was consistent with the occurrence pattern of the five most common DSP-producing dinoflagellates, which was supported by the development of warm and thermohaline stratified waters: Dinophysis caudata, D. fortii, D. sacculus, D. tripos and Phalacroma rotundatum. The strong correlation (r = 0.611, p < 0.001) between D. fortii, reaching abundances of up to 950 cells L⁻¹, and OA suggests that D. fortii is the main cause of OA production in Slovenian waters.

Strong interannual variations in OA and phytoplankton dynamics, exacerbated by the effects of anthropogenic impacts in this coastal ecosystem, reduce the predictability of toxicity events and require continuous and efficient monitoring. Our results also show that the introduction of the LC-MS/MS method for lipophilic toxins has improved the management of aquaculture activities, which was not as accurate based on mouse bioassays.