Pyrrolizidine Alkaloids in Herbal Raw Materials: What the FDA Guidance and EU Limits Mean for Your Testing Plan

Chamomile probably isn’t the first botanical your procurement team flags as a hepatotoxicity risk. It’s a Category 1 botanical under the American Herbal Products Association’s safety classification, it has a well-documented USP monograph, and it sells in enormous volume to reputable finished-goods brands across the Midwest. But in 2020, the FDA specifically cited chamomile — alongside peppermint and other widely trusted herbs — as a botanical that can arrive at your receiving dock carrying pyrrolizidine alkaloid contamination. Not because of what the chamomile plant produces, but because of what was growing next to it at harvest.

That one detail reframes the risk entirely. It means botanical identity testing won’t catch it. Heavy metals panels won’t catch it. Standard microbiological screening won’t catch it. If your raw material testing program doesn’t include a targeted PA screen, you have a gap — and FDA’s 2020 guidance made it clear they expect you to close it.

What Pyrrolizidine Alkaloids Are and Why the Unsaturated Forms Matter

Pyrrolizidine alkaloids are a class of naturally occurring compounds produced by an estimated 6,000 plant species globally. The structure of the molecule is the key variable. Saturated PAs are largely non-toxic. Unsaturated 1,2-dehydropyrrolizidines — the ones with an exocyclic double bond — are a different story entirely. In the liver, cytochrome P450 enzymes metabolize them into highly reactive pyrrolic intermediates that form DNA adducts, cause veno-occlusive disease, and have demonstrated genotoxic and carcinogenic potential in animal models.

Of the 660-plus known PA structures identified in the scientific literature, roughly 100 fall into the hepatotoxic unsaturated category. Comfrey (Symphytum officinale) produces them in concentrations exceeding 1,000 mg/kg in leaf material — which is why FDA banned comfrey-containing dietary supplements intended for oral use in 2001. Borage seed oil (Borago officinalis) is another intentional source, containing amabiline and related PAs. Those two are relatively well-understood risk cases. The subtler problem is cross-contamination.

During commercial-scale harvesting of chamomile, echinacea, peppermint, and lemon balm, PA-producing weeds — particularly species of Senecio, Crotalaria, and Heliotropium — grow in the same fields and get co-harvested with the target botanical. Post-harvest blending across multiple agricultural lots then distributes that contamination evenly through the finished raw material. A supplier COA testing for species identity and heavy metals tells you nothing about this. The contamination is chemically invisible to those methods.

What FDA and the EU Have Actually Required

FDA’s June 2020 draft guidance, Pyrrolizidine Alkaloids in Botanical Dietary Supplements and Teas, was the agency’s first formal acknowledgment that PA contamination in finished supplements and their raw material inputs represents a genuine safety concern warranting manufacturer action. The guidance stopped short of setting enforceable numeric limits — a deliberate regulatory choice that’s frustrated some quality professionals but reflects FDA’s ongoing analytical method validation work.

What the guidance does establish is a clear expectation: manufacturers are on notice. Under 21 CFR Part 111, the GMP purity requirement for dietary ingredients isn’t satisfied by the absence of known contaminants. It requires active testing against reasonably foreseeable risks. PA contamination in high-risk botanicals meets that bar.

The EU moved faster and more prescriptively. EU Commission Regulation 2020/2040, adopted in December 2020, amended the longstanding maximum level framework under Regulation (EC) No 1881/2006 to set enforceable PA limits in food. For dried herbs and herbal spices, the maximum is 1.0 μg/kg, calculated as the sum of 28 specified PAs and their N-oxides. For herb teas and botanical infusions prepared by hot water extraction, the limit drops to 0.35 μg/kg. Those are microgram-per-kilogram thresholds — achievable analytically, but requiring genuinely sensitive LC-MS/MS methodology to measure reliably.

US brands with EU export ambitions are already operating under these limits by necessity. But even domestic-only brands are seeing retailers reference EU PA thresholds in their vendor qualification questionnaires. Whether or not FDA finalizes numeric limits, 1.0 μg/kg is effectively becoming a de facto industry benchmark.

What a Valid PA Panel at an Analytical Testing Laboratory Covers

This is where a lot of internal testing programs fall short, even among brands that have added PA screening to their raw material specs. Not all PA panels are equivalent, and the gap is specifically around N-oxide forms.

PA N-oxides are water-soluble, polar metabolites that co-occur with free-base PAs in plant material. In the gut, gut microflora reduce N-oxides back to their free-base form, which is then absorbed and metabolized hepatotoxically. Analytically, N-oxides are structurally distinct from their free-base counterparts and require different chromatographic conditions to separate and quantify. Labs that run only free-base PA panels — which represent the majority of PA testing offered in the US market — systematically undercount total PA exposure. In some botanical matrices, N-oxide forms make up 60–70% of total PA content. A test that misses them isn’t a conservative estimate; it’s not an estimate at all.

A complete PA screen at a qualified analytical testing laboratory should:

• Cover at minimum 25–30 individual congeners including both free-base PAs and their corresponding N-oxides. The EU’s validated 28-congener panel under Regulation 2020/2040 is a reasonable baseline.
• Use LC-MS/MS as the analytical platform. Liquid chromatography coupled with tandem mass spectrometry delivers detection limits below 0.1 μg/kg in botanical matrices — well below the EU thresholds — with the specificity needed to distinguish individual congeners in complex herbal extracts. GC-MS is not appropriate for this analysis. Standard HPLC-UV lacks the sensitivity and selectivity required.
• Match the congener panel to the source botanical. For chamomile and echinacea, the priority analytes include senecionine, seneciphylline, retrorsine, jacobine, and their N-oxides — these are the most common cross-contamination species from Senecio weeds. For borage and related Boraginaceae-family ingredients, amabiline, lycopsamine, and intermedine take priority.

Turnaround for a 28-congener LC-MS/MS PA screen runs 5–8 business days. Build that into your raw material receiving workflow before the lot is released to in-process inventory.

A Risk-Stratified Approach to PA Testing

Testing every lot of every botanical for a 30-congener panel isn’t financially realistic for most brands, and it isn’t necessary. A tiered approach based on documented contamination risk is how quality programs we work with have made PA screening manageable.

Tier 1 — Test every lot: Botanicals with documented PA co-occurrence risk at the agricultural production level. This list includes chamomile, echinacea, peppermint, lemon balm, valerian, and any herb sourced from growing regions where Senecio and Crotalaria species are prevalent as field weeds — Eastern Europe, India, and parts of North Africa are flagged most consistently in the peer-reviewed contamination literature.

Tier 2 — Test at first supplier qualification, then annually: Botanicals with lower inherent cross-contamination risk but meaningful use volumes in your formulations. Annual testing establishes a trending baseline; if any lot shows detectable PAs, escalate to per-lot screening immediately.

Tier 3 — Require supplier PA certificate with COA: Low-risk botanical forms, or highly refined extracts where PA removal is part of the documented manufacturing process (some standardized botanical extracts specifically include PA content as a controlled specification).

In incoming raw material testing through our Countryside, IL facility, the pattern we see on Tier 1 botanicals is consistent across brands: approximately 12–18% of chamomile lots from Eastern European suppliers register above 0.5 μg/kg for at least one hepatotoxic congener when screened with a full free-base plus N-oxide panel. That’s a meaningful rejection rate. Every one of those lots arrived with a supplier COA that said nothing about PAs — because the supplier wasn’t screening for them.

What Your DSHEA Compliance Records Need to Show

When an FDA investigator reviews your 21 CFR Part 111 component records during an inspection, they’re looking for documented specifications and test results for identity, purity, strength, and composition. PA contamination is a purity issue. The absence of PA screening in your testing specifications for high-risk botanicals — particularly after FDA’s 2020 guidance explicitly named those botanicals — is something that should appear on a 483 observation.

More importantly, if a consumer adverse event involving hepatotoxicity ever traces back to your product, your QC records become central to any enforcement or litigation proceeding. Demonstrating that you assessed PA risk, documented your scientific rationale, and implemented appropriate testing protocols is a fundamentally different position than not having considered it.

The path forward isn’t complicated. Identify your Tier 1 botanicals. Update your raw material specifications to include a PA limit (1.0 μg/kg referenced against EU Commission Regulation 2020/2040 is defensible and widely cited). Confirm your analytical testing laboratory’s PA panel covers N-oxides and uses validated LC-MS/MS methodology — ask specifically for their method validation summary. And make sure your specification documentation reflects the rationale for your tiering decisions.

Regulators on both sides of the Atlantic have made clear that PA risk is a known and manageable problem. The brands that get ahead of it won’t be scrambling when FDA finalizes its numeric limits.

---

Written by Nour Abochama, VP Operations, Qalitex | Quality Consultant, Ayah Labs. Learn more about our team

Ship your sample to our Chicago facility — get a Qalitex CoA in 5–7 days. Contact us

Related from our network

• Supplement Testing & ISO 17025 Accreditation — Qalitex Laboratories performs ISO 17025-accredited analytical testing for dietary supplements, including LC-MS/MS botanical contaminant screening, from their California facilities.