Sampling and sample preparation methods for determining concentrations of mycotoxins in foods and feeds.

Sample variation is often the largest error in determining concentrations of mycotoxins in food commodities. The worldwide safety evaluation of mycotoxins requires sampling plans that give acceptably accurate values for the levels of contamination in specific batches or lots of a commodity. Mycotoxin concentrations show a skewed or uneven distribution in foods and feeds, especially in whole kernels (or nuts), so it is extremely difficult to collect a sample that accurately represents the mean batch concentration. Sample variance studies and sampling plans have been published for select mycotoxins such as aflatoxin, fumonisin, and deoxynivalenol, emphasizing the importance of sample selection, sample size, and the number of incremental samples. For meaningful data to be generated from surveillance studies, representative samples should be collected from carefully selected populations (batches or lots) of food that, in turn, should be representative of clearly defined locations (e.g. a country, a region within a country). Although sampling variability is unavoidable, it is essential that the precision of the sampling plan be clearly defined and be considered acceptable by those responsible for interpreting and reporting the surveillance data. The factors influencing variability are detailed here, with reference to both major mycotoxins and major commodities. Sampling of large bag stacks, bulk shipments, and domestic supplies are all discussed. Sampling plans currently accepted in international trade are outlined. Acceptance sampling plans and the variabilities that affect operating characteristic curves of such plans are also detailed. The constraints and issues related to the sampling of harvested crops within subsistence farming areas are also discussed in this chapter, as are the essential rules of sample labelling and storage. The chapter concludes with a short section on sample preparation methods.