Mycotoxin Testing for Spices and Grain-Based Foods: What FDA Action Levels Actually Mean for Your Brand

Aflatoxin contamination in imported spices isn’t a rare edge case — it’s a documented, recurring pattern. FDA surveillance data consistently shows chili powder, paprika, and red pepper among the commodity types most frequently flagged for aflatoxin exceedances, with some detained lots testing at 40 to 100 ppb — two to five times the 20 ppb action level for human food. The mold responsible, primarily Aspergillus flavus, does its damage in the field, under the warm, humid conditions where spice crops are grown. By the time those ingredients reach a US manufacturer’s warehouse, the mold itself may be long gone. The toxin isn’t.

That’s the thing about mycotoxins that catches food and supplement brands off guard: they’re chemically stable. They survive grinding, drying, and the low-moisture conditions we associate with safe, shelf-stable dry goods. And under FSMA’s Preventive Controls for Human Food rule (21 CFR Part 117), naturally occurring chemical hazards — including mycotoxins — must be evaluated in your written food safety plan if they’re reasonably foreseeable for your ingredient set. For most spice blends and grain-based formulations, they clearly are.

Why Spices and Grains Carry Disproportionate Mycotoxin Risk

The risk profile for mycotoxins depends heavily on crop type and growing conditions. Aspergillus species — responsible for aflatoxin B1, B2, G1, and G2 — thrive in warm, drought-stressed crops: corn, peanuts, tree nuts, cottonseed, and spices like cumin, chili, turmeric, and paprika. Fusarium species, which produce fumonisins, deoxynivalenol (DON), and zearalenone, are associated primarily with cereal grains — corn and wheat especially — and tend to infect crops during cool, wet flowering periods.

Most of the high-risk source regions for aflatoxin-prone spices are in South Asia, Latin America, and parts of Africa, where inconsistent post-harvest drying infrastructure and variable storage conditions allow mold activity to continue long after harvest. That doesn’t mean you can’t source from those regions. But it does mean that supplier qualification alone is not a substitute for incoming raw material testing. A COA from a supplier-side laboratory you haven’t vetted against ISO 17025 accreditation standards is not the same as an independent third-party result.

For supplement brands, the risk is easy to underestimate. Botanical extracts — 95% curcuminoid extract from turmeric, capsaicin from chili, piperine from black pepper — concentrate the spice matrix. If the source material carries aflatoxin at 15 ppb, extraction and concentration processes can push a finished extract above the 20 ppb threshold. This is worth testing for. Many brands don’t, and it’s a gap FDA auditors reviewing 21 CFR Part 111 cGMP compliance are increasingly looking at.

FDA’s Mycotoxin Limits — Action Levels, Advisory Levels, and the Gaps

FDA uses two frameworks for mycotoxin limits, and treating them as interchangeable creates real compliance risk.

Action levels are formal FDA thresholds above which a product is considered adulterated under the Federal Food, Drug, and Cosmetic Act. Exceeding an action level means potential seizure, import detention, and enforcement action. The most significant is the 20 ppb total aflatoxin limit (B1+B2+G1+G2) for human food — applied broadly and without commodity-specific carve-outs for spices. FDA also maintains a 0.5 ppb action level for aflatoxin M1 in fluid milk, and a 50 ppb action level for patulin in apple juice and apple juice-containing products.

Fumonisin limits exist as guidance documents rather than formal action levels, which creates a false impression of flexibility. FDA’s 2001 guidance recommends 2 ppm for corn masa-based foods (tortillas, chips, tamale products) and 4 ppm for dry-milled corn bran intended for human consumption. In practice, exceedances still trigger regulatory action — the “guidance” framing doesn’t insulate a brand from enforcement.

Advisory levels govern DON (deoxynivalenol, also known as vomitoxin). FDA advises a limit of 1 ppm (1,000 ppb) in finished wheat-based products for human consumption. “Advisory” sounds softer than “action level,” but DON-positive products have driven voluntary market withdrawals and import alerts.

Ochratoxin A (OTA) is where US regulation creates the largest blind spot. FDA has no established action level for OTA in food or supplements. But the EU limits OTA to 3 µg/kg (ppb) in processed cereal products and 10 µg/kg in raw cereals — limits that apply to any brand exporting to European distributors, retailers, or private label customers. We see this come up regularly: a US brand assumes domestic-only sales exempt them from OTA requirements, then a retail buyer in Germany or the Netherlands requests an OTA result on the COA and it becomes an urgent problem. Testing for OTA proactively, even without a US regulatory mandate, is increasingly a market access requirement.

What a Food Safety Testing Lab Actually Checks for Mycotoxins

The analytical method matters, and matching the right approach to your regulatory purpose prevents both gaps and unnecessary spend.

ELISA (enzyme-linked immunosorbent assay) is the workhorse of rapid mycotoxin screening. It’s fast — aflatoxin ELISA kits can return results in under two hours — and cost-effective for high-volume incoming raw material checks. The limitations are real, though. ELISA is matrix-dependent, prone to cross-reactivity, and typically targets a single toxin class per run. Using ELISA for aflatoxin alone means fumonisins, DON, OTA, and zearalenone go undetected. That’s not a complete mycotoxin program; it’s a partial screen.

HPLC with fluorescence detection (HPLC-FLD) has been the standard confirmatory method for aflatoxins for decades. It’s specific, sensitive at low ppb concentrations, and accepted by FDA and international trading partners. Most ISO 17025-accredited food safety testing labs run aflatoxin confirmation by HPLC-FLD after an ELISA screen flags a positive — it’s the appropriate tool for a defensible, reportable result.

LC-MS/MS (liquid chromatography-tandem mass spectrometry) is where the field has moved for comprehensive panels. A validated LC-MS/MS method can simultaneously quantify aflatoxins, fumonisins, DON, OTA, zearalenone, and T-2/HT-2 toxins in a single injection. What used to require four separate assays now runs in one. For food brands that need a complete mycotoxin profile — for a regulatory submission, retail buyer qualification, or a finished-product release standard — LC-MS/MS is increasingly the expected baseline, not a premium upgrade.

The practical calibration: ELISA for high-frequency screening of lower-risk incoming materials; HPLC-FLD for aflatoxin confirmation; LC-MS/MS multi-toxin panels for finished products, high-risk ingredient categories, or any situation where the full contamination picture matters.

Building a Mycotoxin Testing Protocol That Holds Up to FDA Scrutiny

Sampling strategy matters more than most brands account for. Mycotoxin contamination in bulk commodities is highly heterogeneous — a single contaminated pocket in a 500-pound lot can produce a clean composite result if sub-sampling is done carelessly. Composite sampling across a lot, drawing multiple incremental sub-samples and blending them before analysis, is standard practice for a reason. Following FDA’s commodity-specific sampling guidance for aflatoxin, where it exists, is not optional when you’re making compliance hold/release decisions.

Source traceability should be integrated into the hazard analysis, not just the supplier qualification program. Ingredients from high-risk growing regions warrant higher testing frequency and potentially stricter incoming specifications than a supplier’s default COA provides. Corn from drought-affected growing regions in a bad harvest year carries a different aflatoxin risk profile than the same corn variety grown with adequate moisture. That context belongs in your written hazard analysis — with supporting rationale.

Document your mycotoxin hazard assessment decisions with genuine supporting logic. If an FDA inspector reviews a food safety plan for a chili-spice blend product and finds mycotoxins assessed as “not reasonably likely to occur” with no substantive justification, that’s a Form 483 observation waiting to happen. The assessment should reflect what you actually know about your supply chain — including the geographic origin of ingredients, harvest conditions, and historical surveillance data.

Retain test results with lot traceability. Under 21 CFR Part 117, records supporting preventive control monitoring must be maintained for at least two years minimum. If you’re using contract manufacturing, confirm that your CMO maintains testing records in a transferable format — including COAs from accredited third-party labs, not just internal results that may not stand up to scrutiny in an audit context.

The regulatory standard isn’t zero mycotoxin contamination. That’s agronomically unrealistic in commodity supply chains, and FDA doesn’t expect it. What’s expected is that you’ve characterized the risk, established testing frequencies appropriate to your ingredient risk profile, and can demonstrate finished product compliance with applicable action levels. That’s achievable — with a structured testing program and a food safety testing lab that understands both the analytical science and the regulatory context behind the numbers.

---

Written by Nour Abochama, Vice President of Operations, Qalitex Laboratories. Learn more about our team

Talk to our team about your testing needs. Contact us

Related from our network

• FDA Audit Readiness and FSMA Food Safety Plan Development — Aurora TIC’s regulatory consulting team helps US food and supplement brands build audit-ready compliance documentation, including written food safety plans under 21 CFR Part 117.
• Raw Material Mycotoxin Screening and Supplier COA Verification — Ayah Labs provides global B2B raw material testing services, including multi-toxin LC-MS/MS panels for botanical ingredients and grain-based commodity inputs.