The wordings below are drawn from the EU Register as established by Commission Regulation (EU) No 432/2012 and subsequently amending regulations. The right-hand column is the point of the table: an authorised claim is a dose commitment before it is a marketing asset. Conditions of use are summarised; the Register entry governs.

• The botanical carries the brand story. The nutrient carries the claim. A liver formula can be built around a silybin-phosphatidylcholine complex as its hero ingredient and still make a lawful EU liver claim — because choline is in the formula at 82.5 mg and the claim is attributed to choline.
• This is not a workaround. It is how the Regulation is designed to operate, and it is the single most useful thing an R&D partner can bring to a detox brief.
  It changed in 2025, and much of the detox marketing material has not yet caught up.
• Since 2010, the Commission has suspended the evaluation of health claims on botanical substances. It has accumulated a backlog of over two thousand applications and a list on hold that had claims covered by the transitional provisions of Article 28(5) of Regulation (EC) 1924/2006. Operators have interpreted it as a tolerant grey zone.

• This changed in 2025 and a great deal of detox marketing copy has not caught up.
• Since 2010 the Commission has suspended its evaluation of health claims relating to botanical substances, leaving a backlog of over two thousand applications and an “on-hold” list whose claims were tolerated under the transitional provisions of Article 28(5) of Regulation (EC) 1924/2006. Many operators read that suspension as a permissive grey zone.
• In the decision in Case C-386/23 (Novel Nutriology), issued on 30 April 2025, the Court of Justice held that claims relating to the health benefits of botanicals cannot be made unless authorized, despite the suspension of the evaluation procedure. The ban of Article 10(1) still applies, along with the Article 10(3) obligation to provide, in the case of general references to benefit, an authorized claim as well. The transition path remains valid only for the cases in which an application was made to the relevant national competent authorities before 19 January 2008 and the claim is on the on-hold list.
• Consequence in terms of a detox brand. Milk thistle, dandelion, artichoke, turmeric and schisandra cannot make a health claim in the EU, unless there is a specific on-hold application and all the criteria of the transition path are fulfilled. It must be derived from the nutrients in the formula. It is not detrimental to the product but just means that the benefit is attributed to another ingredient in the label, and it is a straightforward design constraint once it is known at concept stage rather than discovered at label review.

• Process flow diagrams, HACCP plan and validated processing parameters
• Quantitative ingredient declaration, finished-product specification and certificate-of-analysis templates
• Botanical identity documentation — species, plant part, extraction solvent, native-to-extract ratio, standardisation certificate and, where warranted, DNA-barcode confirmation
• GMO status, allergen assessment and cross-contact controls
• Contaminant testing packages: elemental impurities, pesticide residues, mycotoxins, residual solvents, polycyclic aromatic hydrocarbons
• Microbiological specification and testing protocol
• Safety data sheets and toxicological assessment, including herb–drug interaction review
• Stability protocol and shelf-life justification with overage rationale
• Claim substantiation file mapped to each target market’s permitted wording, with the Register or monograph entry cited per claim
• Novel food, NDI, GRAS or monograph submissions as the market requires

The most common failure in this category is a product whose evidence supports something the label is not allowed to say, or whose label says something the evidence does not support. Both are avoidable at design stage.

Evidence tiers

1. Mechanistic in vitro — Nrf2–ARE reporter assays, CYP inhibition and induction panels, GST and NQO1 activity, β-glucuronidase inhibition. Establishes plausibility and guides dose ranging. Not sufficient for a claim.
2. Bioaccessibility and permeability — INFOGEST 2.0 digestion, Caco-2 apparent permeability. Confirms the active reaches circulation in a relevant form.
3. Human pharmacokinetics — randomised crossover, Cmax / Tmax / AUC. Required to defend any bioavailability claim.
4. Human efficacy — randomised, double-blind, placebo-controlled parallel or crossover design, pre-registered primary endpoint, powered sample size, per-protocol and intention-to-treat analysis. This is the tier that supports a market claim, and the tier FTC expects for a US structure/function claim.

Choosing the correct endpoint will determine whether a study is capable of supporting any claims at all. Defensible, quantifiable endpoints in this category may be:

• The frequency and form (by Bristol Stool Form Scale [14]) of bowel movements, whole gut transit time
• Plasma total glutathione and GSH:GSSG ratio; erythrocyte glutathione peroxidase activity
• Urinary D-Glucaric Acid – a well-established measure of glucuronidation activity of the liver and one of the very few direct measures of Phase II activity [15]
• Markers of oxidative damage: urinary 8-OHdG, plasma F2-isoprostanes and malondialdehyde
• Serum ALT, AST and GGT tracked within the reference range as safety and function indicators
• The urinary ratio of 2-hydroxyoestrone to 16α-hydroxyoestrone to assess estrogen metabolite handling
• Faecal calprotectin, serum zonulin and lipopolysaccharide-binding protein for barrier function
• The profile of SCFA’s in faeces measured by GC-FID; microbiota composition by 16S rRNA sequencing or shotgun metagenomics
• Patient-reported outcome tools for assessing bloating, digestion quality and fatigue

A statistically significant change in the biomarker alone does not constitute a legally permitted claim. A study is designed such that its primary endpoint correlates with wording which is currently approved by the regulatory body: claim language, monograph indication, Permitted Indication or catalogue function.

• FRL uses ORAC, DPPH and FRAP as raw-material screening and batch-consistency tools only, and does not present them as evidence of a physiological effect.
• The USDA withdrew its ORAC database in 2012 on the basis that in vitro antioxidant capacity does not predict in vivo effect, and neither FDA nor EFSA accepts chemical antioxidant capacity as substantiation for a health claim. An EU antioxidant claim comes from vitamin C, vitamin E, selenium, zinc, copper, manganese or riboflavin at source level — not from an ORAC number.