CBD and reproductive toxicity: where the 1B listing stands

ANSES proposed a harmonised EU classification of cannabidiol in the class "reproductive toxicity, category 1B", with two hazard statements attached:

• H360FD: "May damage fertility. May damage the unborn child."
• H362: "May cause harm to breast-fed children."

The file draws on studies in monkeys, rats and mice, which show effects of CBD on sperm production, fertility, deaths around birth and brain development. The agency stresses that the process concerns "only" the assessment of hazards. At this stage, it sets no measures for products.

The CLP Regulation ranks reproductive toxicants on three levels. Category 1A is for substances known to affect human reproduction, chiefly on the basis of human data. Category 1B is for presumed reproductive toxicants, chiefly on the basis of animal studies. Category 2 holds suspected ones, where the evidence is weaker.

The proposal thus puts CBD at the middle level: the animal studies were judged clear enough, although there is no direct proof in humans. H362 belongs to a separate category, which covers effects on or via breastfeeding.

A classification describes what a substance can do. What it does at a given dose is another question, and the gap between the two is the difference between hazard and risk. A classified substance may stay in use if exposure is low and controlled, a point that each sector's rules then settle.

Article 37 of the CLP Regulation sets the route for a harmonised classification. The ANSES file was open to public consultation on the website of the European Chemicals Agency (ECHA) from 17 March to 16 May 2025. ECHA's Risk Assessment Committee, known as RAC, then has 18 months to deliver its opinion.

RAC has now delivered it. The Belgian federal health service reported on 5 May 2026 that the committee had adopted an opinion to classify CBD as toxic for reproduction. That opinion covers fertility, the unborn child and breast-fed children.

The political step is still ahead. If the European Commission considers the classification appropriate, it prepares without undue delay a draft entry in Annex VI of the CLP Regulation, which is then put to the member states. The regulation sets no deadline for this decision, and adoption is not automatic.

The CLP Regulation itself imposes no hazard labels on finished consumer products, since its Article 1 excludes medicines, cosmetics and foods, among others. Its effects pass through the rules of each sector, which refer to the classification.

Cosmetics are the plainest case. Article 15 of Regulation (EC) No 1223/2009 bans from cosmetics any substance classified as carcinogenic, mutagenic or toxic for reproduction in category 1A or 1B. An exemption is still possible if four conditions are all met: compliance with food safety requirements, no suitable alternative, a specific use with known exposure, and a favourable opinion from the Scientific Committee on Consumer Safety. The current framework is described in our article on claims for CBD cosmetics.

In practice, nothing changes for foods. No CBD food has obtained the authorisation that the novel food regulation requires, as our overview of CBD regulation in 2026 explains.

The classification makes no claim that a CBD flower or oil makes people infertile. What it says is that animal studies show effects on reproduction clear enough to presume a hazard in humans. The real risk depends on the dose, the duration and the person.

For exposed groups, two opinions converge. EFSA concluded on 9 February 2026 that the safety of CBD cannot be established for pregnant and breastfeeding women, people under 25 and people on medication. For other adults, it sets a provisional safe level of about 2 mg a day for 70 kg. The agencies' positions are gathered in our article on CBD, pregnancy and breastfeeding.

If you are pregnant, breastfeeding or trying to conceive, the careful choice is to avoid CBD and talk to a doctor or midwife. Otherwise, the file is one to keep watching on the regulatory side, and our page on CBD side effects sums up the known data.