Pesticides in CBD flower: no MRL names the flower

An MRL is the highest pesticide residue level tolerated in or on a food after correct use of the product. It does not mark the point where toxicity begins: it is a ceiling set from good agricultural practice, then checked against consumer exposure.

Regulation (EC) No 396/2005 sets these limits across the Union. The Commission page on this legislation gives three orders of magnitude. MRLs apply to 315 fresh products, and to the same products after processing, corrected for dilution or concentration. Around 1,100 pesticides are covered, whether used now or in the past, in the Union or elsewhere. When a pesticide is not specifically mentioned, a default MRL of 0.01 mg/kg applies.

Article 18 of the regulation provides this default value. It covers Annex I products that have no specific limit for a given substance. The next question is whether the hemp flower appears in Annex I at all.

Regulation (EU) 2018/62 rewrote Annex I, and hemp now appears there twice. Hemp seeds have their own code, 0401140, in the oilseeds group. Hemp leaves sit under code 0632010-045, in the group of herbal infusions from leaves and herbs, attached to the main product "strawberry leaves", so they follow the MRLs set for those leaves.

The flower has no entry, either in the group of herbal infusions from flowers or anywhere else in the annex.

The explanation is that Annex I lists food and animal feed. In France, a CBD flower sold in a shop is restricted to adults and is not meant to be eaten, so no MRL is assigned to it by name. A serious certificate therefore has to state the grid it compares its results against: the MRLs for hemp leaves, the default value of 0.01 mg/kg, or another reference.

The Commission runs a pesticides database in which you can look up an active substance or a product. According to the page, it is for information only and has no legal value; the Official Journal of the European Union remains the reference.

The grid changes over time. For 2025 alone, the Commission's legislation page lists about ten regulations amending the annexes of Regulation 396/2005, among them for acetamiprid, deltamethrin and chlorpropham. A certificate from 2024 was read against a different grid from one issued in 2026, so the test date matters as much as the result.

On a report, 0.01 mg/kg often stands for the method's limit of quantification. Readers easily take it for a level that was found, so it helps when the two figures sit in separate columns.

Official control laboratories validate their methods according to document SANTE/11312/2021, in its version applicable from 1 January 2026. The document sets four requirements in figures.

The document gives the limit of quantification (LOQ) a precise definition: the lowest spiking level in the validation that still meets these criteria. A laboratory has to demonstrate its LOQ, matrix by matrix; declaring one is not enough.

Two techniques do most of the work. Liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) covers polar or heat-sensitive molecules, and gas chromatography (GC-MS/MS) covers volatile ones. A broad panel combines the two. On a report, the words "LC-MS/MS" with no number of substances and no LOQ still tell you nothing.

Article 37 of Regulation (EU) 2017/625 sets the rule in paragraph 4, point (e): official laboratories work to the EN ISO/IEC 17025 standard and are accredited against it. Article 34 of the same text requires methods that follow Union rules or recognised performance criteria.

An accreditation has a scope, which is a list of method and matrix pairs, so it never covers everything a laboratory does. A laboratory accredited for cannabinoids by HPLC may hold no accreditation for a pesticide panel on dry plant material. Our guide to a laboratory's scope of accreditation shows where to check it, in five steps.

Take one page from the catalogue. The Super Boof flower, batch PG-26-047F, shows a reference profile of 9.6% CBD and 0.2% Δ⁹-THC, with no laboratory certificate published. Its description mentions cold greenhouse growing without synthetic pesticides, which is a statement about a growing practice. Only a "pesticides" page can turn it into a measured fact, and our product pages do not yet display pesticide results. Four lines make the difference, here as at any other shop.

• The list: how many substances were tested for, and which. A panel of ten molecules and a panel of several hundred answer different questions.
• The limit of quantification, in mg/kg, per substance or per group. Without it, "not detected" has no scale.
• The method and the date: LC-MS/MS, GC-MS/MS or both, with the day of the test.
• The reference and the scope: the comparison grid chosen, and whether the laboratory is accredited for this method on this matrix.

The same reasoning holds for resins, which are trichomes separated from the plant and then pressed (see the hashish entry in the glossary). Colour, smell and price say nothing about residues. An "organic" claim is no substitute for the test either; our article on organic CBD flowers explains what the label guarantees. Each product in the CBD flowers category states its batch. For the other family of contaminants, see our article on heavy metals.

Each year, EFSA publishes the results of residue controls in food. Its report released on 5 May 2026, on 2024 data, draws on more than 125,000 samples.

A sample can exceed the MRL and still escape a non-compliance finding, because measurement uncertainty is taken into account before any conclusion. The 2024 coordinated programme covered twelve foods, including aubergines, bananas, table grapes, virgin olive oil, wheat and hen eggs. Implementing Regulation (EU) 2025/854 of 7 May 2025 set the programme for 2026 to 2028. None of these plans covers CBD flower, since it is not a food. For the THC threshold that does apply to it, see our article on the legal 0.30% limit. For the other lines of a certificate, see our guide on how to read a certificate of analysis.