Pyrrolizidine alkaloids (PA) and their N-oxides (PANO) are a group of toxic secondary plant metabolites occurring predominantly as contaminants in (herbal) teas, honeys and food supplements, as well as in spices and culinary herbs. Depending on the botanical origin of the contaminating plant, the pattern of PA/PANO can strongly vary within a sample. The current study aimed to broaden the existing data on the occurrence of PA/PANO in spices and culinary herbs. For this, 305 authentic samples covering 15 different matrices mainly harvested in 2016 or 2017 and originating from 36 countries were investigated for the presence of 44 PA/PANO. Fifty-eight percent of the samples contained at least one PA/PANO. The average sum content over all samples was 323 µg/kg (median of 0.9 µg/kg, 95% percentile of 665 µg/kg). The highest amount of 24.6 mg/kg was detected in an oregano sample. Additionally, conspicuous analyte patterns were discovered in samples from similar cultivation regions, indicating related botanical sources of PA/PANO contaminations. Particularly, oregano and cumin from Turkey often contained high amounts of PA/PANO. The results were used to assess the acute and chronic health risks related to PA/PANO intake via spices and culinary herbs, indicating a potential health risk in particular for adults and children with high consumption or when considering worst-case contamination scenarios of a sum content of 5500 µg/kg.

A comprehensive definition introducing the term “modified mycotoxins” to encompass all possible forms in which mycotoxins and their modifications can occur was recently proposed and has rapidly gained wide acceptance within the scientific community. It is becoming increasingly evident that exposure to such modified mycotoxins due to their presence in food and feed has the potential to pose a substantial additional risk to human and animal health. Zearalenone (ZEN) is a well-characterized Fusarium toxin. Considering the diversity of modified forms of ZEN occurring in food and feed, the toxicologically relevant endocrine activity of many of these metabolites, and the fact that modified forms add to a dietary exposure which approaches the tolerable daily intake by free ZEN alone, modified forms of ZEN present an ideal case study for critical evaluation of modified mycotoxins in food safety. Following a summary of recent scientific opinions of EFSA dealing with health risk assessment of ZEN alone or in combination with its modified forms, uncertainties and data gaps are highlighted. Issues essential for evaluation and prioritization of modified mycotoxins in health risk assessment are identified and discussed, including opportunities to improve exposure assessment using biomonitoring data. Further issues such as future consideration of combinatory effects of the parent toxin with its modified forms and also other compounds co-occurring in food and feed are addressed. With a particular focus on ZEN, the most pressing challenges associated with health risk assessment of modified mycotoxins are identified and recommendations for further research to fill data gaps and reduce uncertainties are made.

1,2-Unsaturated pyrrolizidine alkaloids (PA), their corresponding N-oxides (PANO), and tropane alkaloids (TA) are toxic secondary plant metabolites. Their possible transfer into the milk of dairy cows has been studied in feeding trials; however, only few data on the occurrence of these toxins in milk are available. In this study, the development of a sensitive analytical approach for the simultaneous detection and quantification of a broad range of 54 PA/PANO as well as of the TA atropine and scopolamine in milk of dairy cows is presented. The method optimisation focused on sensitivity and separation of PA/PANO isomers. Milk samples were extracted using liquid–liquid extraction with aqueous formic acid and n-hexane, followed by a cation-exchange solid-phase extraction for purification. Reversed phase liquid chromatography tandem mass spectrometry (LC–MS/MS) analysis was performed using alkaline solvent conditions. Validation proved low limits of detection and quantification of 0.005 to 0.054 µg/L and of 0.009 to 0.123 µg/L, respectively. For 51 of the 54 tested PA/PANO and both TA, the recovery rates ranged from 64 to 127% with repeatability (RSD_r) values below 15% at concentration levels of 0.05 and 0.50 µg/L and below 8% at a concentration level of 3.00 µg/L. Only three PANO did not match the validation criteria and were therefore regarded as semiquantitative. The final method was applied to 15 milk samples obtained from milk vending stations at farms and from local marketers in Bavaria, Germany. In three of the milk samples, traces of PA were detected. Graphical abstract