Frequent accumulation of diarrheic shellfish toxins by different bivalve species in a shallow subtropical estuary

Abstract Harmful blooms of Dinophysis spp. have been frequently reported along the southern Brazilian coast. This study intended to determine the accumulation of diarrheic shellfish toxins (DST) in co-occurring wild Anomalocardia brasiliana clams and farmed Perna perna mussels, and relate it to the mesoscale distribution of Dinophysis spp., Mesodinium rubrum and toxins over 13 months in a shallow and vertically homogeneous estuary. Bivalves and depth-integrated water samples were collected from two regions (shallower and deeper areas) representative of the outer section of Babitonga Bay estuary. Cells belonging to the Dinophysis acuminata complex (99% of the total Dinophysis abundance) and those of M. rubrum occurred frequently over the studied period. The DST okadaic acid (OA) was detected both in the particulate fraction of the seawater and in the soft tissues of suspension-feeding bivalves, even when Dinophysis cell density was extremely low (e.g. 30 cell.L−1). There was a marked seasonal variation in both the abundance of D. acuminata complex cells in suspension and the DST levels in bivalve tissues, with higher values recorded from late autumn to spring. Farmed mussels accumulated 3- to 13-fold greater levels of DST than co-occurring wild clams, even though no significant variations in Dinophysis abundance or particulate toxin concentration were observed between the areas where each bivalve species was sampled. Non-esterified OA levels in hanged mussels were not linked to Dinophysis cell density, suggesting that part of the toxin recently incorporated by these bivalves could be adsorbed onto organic and/or inorganic particles. Considering the potential chronic effects of OA, the frequent accumulation of moderate toxin levels (up to 100 μ g.Kg−1 in mussels) may thus represent a health risk for frequent consumers of bivalves in this area. Additionally, the risks for episodes of acute human intoxication in this area should be monitored at shorter sampling intervals over multiple years.