Comparação da eficiência de dois planos de amostragem de milho para análise de micotoxinas

Abstract

The aim of this research was to evaluate variability and performance of two maize sampling plans for mycotoxins quantification. Eleven maize lots were sampled using two distinct sampling methods: manual, using a sampling spear, and automatic, using a continuous flow sampling. Total variability and variability from each step of the analysis procedure (sampling, sample preparation, and analysis) for results of aflatoxins, fumonisin, and zearalenone were determined in 8, 11, and 5 lots respectively for each sampling plan. Those variances were compared, for each mycotoxin, between the two sampling plans using a multifactor analysis of variance. Distribution of the quantification results for aflatoxins and zearalenone were compared among four theoretical distribution models. The acceptance/rejection probabilities of a lot with determined aflatoxins or zearalenone concentration were determined using calculated variances and the information on the selected distribution to plot an operation characteristic curve (OC). Sampling and total variances were lower on the automatic sampling plan for aflatoxins, fumonisins, and zearalenone quantification. The OC curve from automatic sampling plan for aflatoxins and zearalenone quantification reduces the consumer end seller risks when compared with the OC curve from the manual sampling plan. The automatic sampling plan is more efficient than the manual one for mycotoxins quantification in maize, reducing lot classification error, and providing more safety in commercial transactions.