Adherence to regulatory limits for mycotoxins in agricultural commodities is important to safeguard consumers and to permit trade in affected commodities across international borders. Reliable estimates of mycotoxin concentrations are required to implement regulatory decisions on the suitability of lots of produce for consumption or trade. Effective schemes to test for mycotoxins depend not only upon sound analytical methods, but also on well designed and implemented sampling plans. This manual provides information to food analysts and regulatory officials on effective sampling plans to detect mycotoxins in food. The concepts of uncertainty and variability in mycotoxin test procedures are discussed as well as the importance of ensuring that samples are representative of the lot being sampled, and the consequences of a poorly designed sampling plan on the reliability of the measured levels of mycotoxins, possibly resulting in legal disputes and barriers to trade.

1. Introduction 1.1 Regulatory Limits 1.2 Mycotoxin Units 1.3 The Rationale of Accurate Sampling Procedures 2. Definition of a Sampling Plan 3. Uncertainty 3.1 Accuracy 3.2 Precision 4. Sample Selection 4.1 Static Lots 4.2 Dynamic Lots 4.3 Aggregate versus Laboratory sample 5. Sample Preparation 6. Analysis 7. Accept/reject Limit 8. Random Variation 8.1 Sampling Variability 8.2 Sample Preparation Variability 8.3 Analytical Variability 8.4 Total Variability 9. Reducing Variability of a Mycotoxin Test Procedure 9.1 Laboratory Sample Size 9.2 Test portion Size 9.3 Number Aliquots Quantified 10 Designing Mycotoxin Sampling Plans 10.1 Calculation of Acceptance Probability 10.2 Effect of Uncertainty (Sample Size) on Risks 10.3 Accept/Reject Effects on Risks 11. Conclusions 12. References 13. Appendix I - Evaluating the Performance of Sampling Plans to Detect Fumonisin B1 (FB1) in Maize Lots Marketed in Nigeria