Pyrrolizidine alkaloids (PAs) and their N-oxides (PANOs) are natural toxins, exclusively biosynthesized by a wide variety of plant species (>6000). They are secondary plant metabolites against herbivores and are believed to be one of the most widely spread natural toxins. Human poisoning cases have been documented; they are characterized by acute and chronic liver damage, pulmonary hypertension, cardiac or kidneys injuries, and can lead to death. Therefore, the development of efficient analytical methods is required to detect PAs/PANOs even at low levels.
The Belgian Scientific Institute of Public Health has directed a project aiming to report occurrence levels of PAs and PANOs in targeted food items to perform a dietary exposure assessment to these natural contaminants and evaluate if they can pose a health problem, filling a data gap at Belgium’s level. Analytical methods based on UPLC separation and MS/MS detection were developed for the analysis of 16 PAs and 14 PANOs in honey and honey based snacks, meat and meat products, milk and dairy commodities, plant based products and food supplements. A structured sampling plan has been established for more than 1,000 samples available on the Belgian market. The presentation will focus on the interesting case of PAs and PANOs in plant based products: dry (herbal) teas and “as consumed” teas (brewed), but also less targeted matrices such as salads, spices and food supplements.
On one hand, the sample preparation for dry plant based products (freeze dried salads, spices, (herbal) teas and food supplements) involved organic acidic extraction of the analytes followed by Envi-Carb SPE purification to avoid loss of sensitivity due to matrix effect. On the other hand, tea samples were brewed following ISO3103 standardized method, the infusion was then purified using C18 SPE cartridges. Several analytical steps were optimized in order to obtain the highest sensitivity. The separation of PAs/PANOs was performed on a C18 stationary phase LC column. The analytical methods were validated in-house and the method’s performances were within the criteria allowed by Directive 2002/657/CE and SANTE/11945/2015. The LOQs achieved (0.1 ng.g-1 for dry products, 0.01 ng.mL-1 for brewed teas) were lower than those proposed by EFSA (2011).
Surprisingly, high levels of heliotrine-type PAs/PANOs (>500 ng.g-1) were detected in 2 Mediterranean spice mixes and certain salad mixes were also contaminated with significant levels of senecionine-type compounds. More than 85% of plant based food supplements contained PAs/PANOs, with levels up to 557,000 ng.g-1. An original approach enabled us to demonstrate that only a part of the PAs/PANOs contamination is transferred during the brewing process of (herbal) teas (15-25%). This transfer rate contrasts with the conventional strategy consisting of applying a dilution factor to the PAs concentration found in the dry plant product to estimate the contamination level of the infusion.
