International proficiency testing as a tool to evaluate the state-of-art of LC-MS methods for multi-mycotoxin determination

Mycotoxin contamination of agricultural food commodities and beverages poses a risk to human and animal health due to their toxic effects. Over 100 mycotoxins have been identified, although only a few of them present a significant source of food-borne illnesses and are of major concern worldwide. They are: aflatoxins B1 (AFB1), B2 (AFB2), G1 (AFG1) and G2 (AFG2), ochratoxin A (OTA), fumonisin B1 (FB1) and B2 (FB2), deoxynivalenol (DON), zearalenone (ZEA), T-2 and HT-2 toxins. In the European Union, harmonized maximum levels for mycotoxins in foodstuffs have been specified in the Commission Regulation EC 1881/2006, and further amendments. Effective and efficient analytical methods are therefore required to identify and determine mycotoxins at legislated levels and to enforce regulatory limits. Within this context the application of LC-MS(MS) techniques is being largely explored since it enables the simultaneous monitoring of different mycotoxins in one run. Even though LC-MS(MS) methodologies for single or multiple mycotoxin determination are routinely used in control laboratories, to date none of official or standard methods approved by AOAC International or CEN (European Standardization Committee) is based on LC-MS. Proficiency Testing (PT) is an effective procedure to determine the performance of individual laboratories for specific measurements, providing a clear and a straightforward way of evaluating the accuracy (trueness and precision) of results obtained by different laboratories. An international PT was organized in 2014 to check, next to the laboratory performance, the state-of-art of currently used multi-mycotoxin methods and their implementation in the respective laboratory. The PT was free of charge and was organized by ISPA-CNR in the framework of the Italian project S.I.Mi.S.A. (PON02_00186_3417512) and promoted by the MoniQA Association (www.moniqa.org). Within the EU Network of Excellence MoniQA several efforts have been made for method comparison and deeper understanding of performances of the available LC-MS(MS) methodologies for multiple-mycotoxin analysis [1,2]. Participants were asked to determine DON, FB1, FB2, ZEA, T-2, HT-2, OTA and aflatoxins (AFB1, AFB2, AFG1, AFG2) in maize, and DON, ZEA, T-2, HT-2 and OTA in wheat. The contaminated test materials were produced and characterized by ISPA-CNR. The use of LC-MS(MS) methods, although not strictly required, was highly recommended, while the use of multi-mycotoxin methods was mandatory. Participants were not obliged to determine all toxins in each material, and let free to report only on those mycotoxins that were simultaneously determined with their multi-mycotoxin methodology. Twenty-two participants from 10 countries registered for the exercise. Nineteen laboratories returned 22 sets of results for various combinations of analytes. Three laboratories returned two sets of results obtained by using two different methods for both contaminated maize and wheat. The assigned values (consensus values) were calculated according to ISO 13528:2005 whereas the target standard deviation was derived from the truncated Horwitz equation. No statistical evaluation was reported for AFB2, AFG2 in maize due to lack of sufficient quantitative data. The assigned values for maize test materials were 1264 µg/kg for DON, 1305 µg/kg for FB1, 350 µg/kg for FB2, 2.73 µg/kg for OTA, 54.4 µg/kg for T-2, 30.7 µg/kg for HT-2, 21.7 µg/kg for ZEA, 1.35 µg/kg for AFB1 and 0.63 µg/kg for AFG1. The assigned values for wheat test materials were 1298 µg/kg for DON, 7.21 µg/kg for OTA, 8.26 µg/kg for T-2, 58.8 µg/kg for HT-2 and 148 µg/kg for ZEA. All laboratory results were rated with z-scores in accordance with ISO 13528:2005 and the International Harmonized Protocol for the Proficiency Testing of Analytical Chemistry Laboratories. Some general conclusions could be drawn: a.The participation of the laboratories was regarded as satisfactory concerning the number of received results (86% of participants) b.Fifty-five percent of laboratories analysed all the 11 targeted mycotoxins in maize, whereas 73% of laboratories analysed all the 5 mycotoxins in wheat. The remaining laboratories reported results for a different combination of mycotoxins (from 2 to 10 in maize and from 1 to 4 in wheat). c.The assessment of laboratories on the base of their z-scores indicated that only 23% of participants were considered successful for the whole interlaboratory test. d.The majority of laboratories used mixtures of acetonitrile-water (73%) or methanol-water (18%) for mycotoxins extraction. e.Fifty percent of laboratories analysed the crude extract; the others cleaned-up the extract by solid phase extraction (23%) or immunoaffinity columns (14%) prior to the analysis. f.The majority of laboratories (55%) used the internal standard calibration mode using stable isotope internal standards (13C) for mycotoxins determination; the other laboratories used external calibration mode using native standard mycotoxins.