Proficiency test for simultaneous determination of up to eleven mycotoxins in maize by using LC-MS(MS) methodology.

A recent survey on the use and application of methods for the determination of mycotoxins in food and feed revealed that 42% of participant laboratories routinely use LC-MS(MS) methodology for their single or simultaneous determination. However, within both CEN and AOAC contexts there are no LC-MS(MS) official methods for the measurement of this group of contaminants. A proficiency test (PT) for the simultaneous determination of up to 11 mycotoxins in maize using LC-MS(MS) methodology was conducted to obtain information on currently used methodologies and method-related performances. Each participant received a mixed mycotoxins calibrant solution, a spiking solution, a spiking protocol, instructions, a comprehensive questionnaire and two test materials (a naturally contaminated maize sample and a blank maize sample to be spiked with a mixture of 11 mycotoxins (aflatoxins B1, B2, G1 and G2, ochratoxin A, deoxynivalenol, T-2 toxin, HT-2 toxin, zearalenone, fumonisins B1 and B2). Of the 64 laboratories enrolled in the PT, 41 laboratories from 14 countries returned results for various combinations of analytes. The percentage of satisfactory z-scores (below an absolute value of 2) for naturally contaminated maize ranged from 50-52% for fumonisins to 68-84% for aflatoxins; comparable results were also obtained for spiked maize. In terms of the methodologies used, the majority of laboratories (54%) used sample extract cleanup (immunoaffinity chromatography, solid phase extraction, liquid/liquid extraction), a group of 39% analyzed a crude extract without purification whereas no information were obtained from few laboratories (7%). Electrospray ionization was used by all laboratories whereas triple quadrupole and ion trap detectors were used by 88% and 12% laboratories, respectively. For quantification, the majority of laboratories (76%) prepared calibration curves in pure solvents (with or without addition of internal standards to the sample extracts) whereas the other laboratories prepared matrix-assisted calibration curves to compensate for matrix effects.