Quantitative targeted and retrospective data analysis of relevant pesticides, antibiotics and mycotoxins in bakery products by liquid chromatography-single-stage Orbitrap mass spectrometry.

In addition to 'traditional' multi-residue and multi-contaminant multiple reaction monitoring (MRM) mass spectrometric techniques devoted to quantifying a list of targeted compounds, the global food industry requires non-targeted methods capable of detecting other possible potentially hazardous compounds. Ultra-high-performance liquid chromatography combined with a single-stage Orbitrap high-resolution mass spectrometer (UHPLC-HRMS Exactive™-Orbitrap Technology) was successfully exploited for the complete selective and quantitative determination of 33 target compounds within three major cross categories (pesticides, antibiotics and mycotoxins) in bakery matrices (specifically milk, wheat flour and mini-cakes). Resolution was set at 50 000 full width at half maximum (FWHM) to achieve the right compromise between an adequate scan speed and selectivity, allowing for the limitations related to the necessary generic sample preparation approach. An exact mass with tolerance of 5 ppm and minimum peak threshold of 10 000 units were fixed as the main identification conditions, including retention time and isotopic pattern as additional criteria devoted to greatly reducing the risk of false-positive findings. The full validation for all the target analytes was performed: linearity, intermediate repeatability and recovery (28 analytes within 70-120%) were positively assessed; furthermore, limits of quantification between 5 and 100 µg kg(-1) (with most of the analytes having a limit of detection below 6 µg kg(-1)) indicate good performance, which is compatible with almost all the regulatory needs. Naturally contaminated and fortified mini-cakes, prepared through combined use of industrial and pilot plant production lines, were analysed at two different concentration levels, obtaining good overall quantitative results and providing preliminary indications of the potential of full-scan HRMS cluster analysis. The effectiveness of this analytical approach was also tested in terms of the formulation of hypotheses for the identification of other analytes not initially targeted which can have toxicological implications (e.g. 3-acetyl-deoxynivalenol and deoxynivalenol-3-glucoside), opening a window on retrospective investigation perspectives in food safety laboratories.