A Precise Mycotoxin Test Can Still Start with the Wrong Sample

A feed mill can use a validated rapid test or an accredited laboratory and still make the wrong receiving decision before the analysis begins. The weak link may be the sample.

Mycotoxins are not necessarily spread evenly through grain or feed. They can occur in localized pockets. A scoop from a convenient surface location may miss those pockets and produce a false sense of security. A scoop taken directly from a pocket can create the opposite error, making the entire lot appear to contain the concentration found at that point. Visible mould is not a dependable shortcut either. Nebraska Extension notes that visibly mouldy material may have no detectable mycotoxin, while material without visible mould may contain a high concentration.

This is why “send a sample to the lab” is incomplete. The real task is to design a reduction chain that preserves the lot’s variability.

The reduction chain

Nebraska Extension describes a typical submitted specimen of 5–10 lb, or about 2.3–4.5 kg. Yet only 10–50 g of that material may become the analytical portion that is actually tested. Before that small portion is removed, the laboratory mills and thoroughly mixes the submitted material. The Canadian Food Inspection Agency similarly describes grinding, mixing and, for large samples, representative reduction with a riffle splitter before analysis.

The size relationship is easy to underestimate. At the extremes of those published ranges, 10–50 g represents only about 0.2–2.2% of a 2.3–4.5 kg submitted specimen. That specimen, in turn, may represent a truckload measured in tonnes. The laboratory instrument can measure its test portion precisely, but precision is not the same as representativeness.

Coverage matters more than one large scoop

Kansas State University recommends collecting from at least 10 evenly spaced locations in a bulk carrier, or at least 10 times at regular intervals during loading or unloading. For bagged material, it recommends random collection from at least 10 bags. Those numbers are useful practical starting points, not a universal official plan for every commodity, hazard, contract or jurisdiction.

The principle is stronger than the number: distribute incremental samples across space or time, then combine them into one composite representing the defined lot. Sampling a moving stream at regular intervals can often provide better coverage than reaching only the accessible surface of a stationary load. Any probe or automated sampler must be appropriate for the ingredient and used safely.

Grinding and mixing are not administrative details. If contaminated kernels or particles are rare and unevenly distributed, splitting an unground composite can simply move the original bias into a smaller container. Proper comminution, homogenization and representative splitting give each part of the composite a better chance of reaching the analytical portion. Research on standardized mycotoxin sampling identifies sampling as usually the largest source of error in the overall test procedure.

A receiving control the mill can audit

For each high-risk ingredient or condition, the mill’s written plan should answer six questions:

1. What defines the lot: one truck, one railcar, one supplier batch or a timed receiving window?
2. Where or when will increments be collected?
3. How many increments and what total composite mass are required?
4. How will the composite be ground, mixed and reduced without contamination or bias?
5. How will the sample identity, date, supplier, lot and collection points be documented?
6. What happens when a result is near a decision limit or conflicts with another result?

A surprising result should trigger a review of the sampling record before arguments begin over the test kit or laboratory. Reanalysis of the same poorly collected material may repeat the same mistake more precisely.