Ultra-performance hydrophilic interaction liquid chromatography tandem mass spectrometry system (UP-HILIC–MS/MS) was used in multi-toxin analysis of paralytic shellfish toxins (PSTs) and tetrodotoxins (TTXs) in sample matrices from bivalve molluscan species commercially produced for human consumption in Sweden. The method validation includes 17 toxins of which GTX6 and two TTX analogues, TTX and 4,9-anhydroTTX, were previously not analyzed together with hydrophilic PSTs. 11-deoxyTTX was monitored qualitatively with a non-certified reference standard. The performance of the method was evaluated for selectivity, repeatability, and linearity by analyzing spiked samples which generated linear calibration curves across the concentration ranges used (R2 > 0.99). The in-house reproducibility (RSD) was satisfactory including the LOD and LOQ for both PST and TTX toxins being far below their regulatory action limits. The major advantage of the method is that it allows direct confirmation of the toxin identity and specific toxin quantification using a derivatization-free approach. Unlike the PST-chemical methods used in routine regulatory monitoring until now for food control, the UP-HILIC-MS/MS approach enables the calibration set-up for each of the toxin analogs separately, thereby providing the essential flexibility and specificity in analysis of this challenging group of toxins. The method is suitable to implement in food monitoring for PSTs and TTXs in bivalves, and can serve as a fast and cost-efficient screening method. However, positive samples would, for regulatory reasons still need to be confirmed using the AOAC official method (2005.06).

In this paper, an LC-MS/MS method for the simultaneous identification and quantification of cyanotoxins with hydrophilic and lipophilic properties in edible bivalves is presented. The method includes 17 cyanotoxins comprising 13 microcystins (MCs), nodularin (NOD), anatoxin-a (ATX-a), homoanatoxin (h-ATX) and cylindrospermopsin (CYN). A benefit to the presented method is the possibility for the MS detection of MC-LR-[Dha7] and MC-LR-[Asp3] as separately identified and MS-resolved MRM signals, two congeners which were earlier detected together. The performance of the method was evaluated by in-house validation using spiked mussel samples in the quantification range of 3.12–200 µg/kg. The method was found to be linear over the full calibration range for all included cyanotoxins except CYN for which a quadratic regression was used. The method showed limitations for MC-LF (R2 = 0.94), MC-LA (R2 ≤ 0.98) and MC-LW (R2 ≤ 0.98). The recoveries for ATX-a, h-ATX, CYN, NOD, MC-LF and MC-LW were lower than desired (<70%), but stable. Despite the given limitations, the validation results showed that the method was specific and robust for the investigated parameters. The results demonstrate the suitability of the method to be applied as a reliable monitoring tool for the presented group of cyanotoxins, as well as highlight the compromises that need to be included if multi-toxin methods are to be used for the analysis of cyanotoxins with a broader range of chemical properties. Furthermore, the method was used to analyze 13 samples of mussels (Mytilus edulis) and oysters (Magallana gigas) collected in the 2020–2022 summers along the coast of Bohuslän (Sweden). A complementary qualitative analysis for the presence of cyanotoxins in phytoplankton samples collected from marine waters around southern Sweden was performed with the method. Nodularin was identified in all samples and quantified in bivalve samples in the range of 7–397 µg/kg. Toxins produced by cyanobacteria are not included in the European Union regulatory monitoring of bivalves; thus, the results presented in this study can be useful in providing the basis for future work including cyanotoxins within the frame of regulatory monitoring to increase seafood safety.