Abstract

A validated method in liquid chromatography tandem mass spectrometry was applied to assess the contamination profile of 17 pyrrolizidine alkaloids and their N-oxide metabolites (PAs/PANOs) in 775 honey samples of different origins, including from local beekeepers (579, LB), large retail and online stores (Italian (LRI) and foreign (LRF) products). Fifty-five percent of the honey samples contained the 17 monitored compounds at below the limit of quantification. The LB samples were characterised by a total mean concentration of PAs/PANOs at 6.7 and 3.3 times lower than those of the LRF and LRI samples, respectively. The PA echimidine characterised mainly honey samples from the large retail stores (75.8 % and 72.5 % of LRI- and LRF-contaminated samples, respectively). The LB samples showed a senecionine-type contamination profile, providing beekeepers clues to the plants responsible for contamination. The results highlighted that the consumption of the local honey samples analysed does not represent a double risk for human health compared to the consumption of products from large retail stores. Careful attention must be paid to honey imported from areas where lycopsamine-type and senecionine-type PAs/PANOs-producing plants are invasive. A machine learning approach could predict the origin of the sample given its contaminant profile at an accuracy of 0.74, which is useful for detecting fraud in the declared honey origin.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call