Collaborative study for evaluating performances of a multiplex dipstick immunoassay for Fusarium mycotoxin screening in wheat and maize

Abstract

An experimental protocol for evaluating performances of qualitative methods through interlaboratory validation has been designed and applied to a multiplex dipstick kit for the determination of major Fusarium toxins, namely zearalenone, T-2 and HT-2 toxins, deoxynivalenol, and fumonisins B1 and B2 in wheat and maize. The test is intended for screening of cereals for the presence/absence of these mycotoxins at maximum permitted levels established by European legislation. The response of the measurement is determined with a reader device. Decision on whether a sample is classified as positive or negative is taken by comparing the measured response with a cut off value which has been established based on the results from the method validation. The collaborative study involved 12 laboratories and included a training phase enabling the participants to familiarise with the protocol. The interlaboratory validation design consisted of three steps, namely (1) estimating the precision under reproducibility conditions; (2) establishing robust and laboratory independent cut off values for the dipstick response at target mycotoxin levels, assuming an acceptable rate of false negative results of 5%; and (3) assessment of the rate of false positive results of compliant samples. The total standard deviation of the response varied from 10 to 27% for the analyte/concentration/matrix combinations included in the study, indicating the assay ruggedness between different laboratories. The test resulted to be able to differentiate blank samples from samples contaminated at target mycotoxin levels with a false positive rate lower than 10% for zearalenone, deoxynivalenol and fumonisins, whereas unsatisfactory results were obtained for the sum of T-2 and HT-2 toxins. Finally, a critical evaluation of the outcomes of the whole validation study enabled to identify key issues for improving method transferability and reliability of the results. © 2014 Wageningen Academic Publishers.