Water-soluble CBD and nanoemulsions: the science and the rules

Cannabidiol is lipophilic: it dissolves in fats, not in water. Swallowed, it crosses the intestinal wall poorly and undergoes intense first-pass metabolism in the liver. The systematic review by Millar and colleagues, published in Frontiers in Pharmacology in 2018 after examining 792 articles, notes moreover that the absolute oral bioavailability of CBD has never been measured in humans: the figures in circulation come from earlier work, largely in animals. The only absolute bioavailability documented in that review concerns inhalation, at 31 %.

The scale of the problem reads better in regulatory data on the medicine. The summary of product characteristics for Epidyolex puts figures on the effect of a meal: taken with a meal rich in fat and calories, cannabidiol sees its peak concentration multiplied by five and its total exposure by four, with reduced variability. Even a low-fat meal still multiplies peak concentration by four. Cow's milk multiplies peak concentration by three.

The molecule is therefore poorly absorbed on an empty stomach, and the most powerful lever known is dietary. Promises about formulation should be read in that context. On the sublingual route, which partly sidesteps the problem, our article on the sublingual method and bioavailability takes stock.

CBD never becomes soluble in water. These products disperse it in droplets fine enough to form a stable, translucent liquid.

A nanoemulsion is a dispersion of droplets generally below 100 nanometres, even though the literature itself is not unanimous on the boundary, some authors using 20 to 200 nanometres. A nanoemulsion is kinetically stable, not thermodynamically stable: it tends spontaneously to separate over time and requires an input of energy to be formed. Self-emulsifying systems, often designated by the acronym SNEDDS, are anhydrous mixtures of oil, surfactant and co-emulsifier that form a nanoemulsion on contact with water.

These three families, plus micellar dispersions and gum arabic emulsions, are sold under the same label. The word says nothing about the technology used, nor about its performance.

Three published trials give figures, to be read with their conditions.

A double-blind crossover trial in sixteen healthy volunteers, published in Molecules in 2019, compares a self-emulsifying system at 40 or 50 nanometres with the same extract in a fractionated coconut oil: peak concentration multiplied by 4.4, total exposure by 2.85. The study was funded by the owner of the technology.

A crossover trial in fourteen participants, published in 2025, tests a self-emulsifying powder at 158 nanometres: relative bioavailability of CBD multiplied by 2.3. The study was funded by the manufacturer, whose chief executive is among the authors.

The third is the most instructive for a reader. Published in Pharmaceuticals in 2021, it compares five oral preparations at an identical dose of 30 mg in fifteen adults. The best reaches a peak concentration of 5.57 ng/mL. The worst, at 1.29 ng/mL, is precisely a CBD powder dispersed in water, that is to say the product one would spontaneously call water-soluble. A conventional tincture in an oil sits between the two, at 2.20 ng/mL. That trial too was funded by a manufacturer.

The best formulations claim a factor of 2.3 to 2.85 on total exposure and up to 4.4 on peak concentration, measured in fourteen to sixteen people by teams paid by the seller. These pharmacokinetic endpoints, measured in plasma, are not a clinical effect.

The effect of a meal rich in fat, for its part, is a factor of five on peak concentration and four on exposure, documented in a marketing authorisation dossier, over a far wider development programme.

The authors of the 2024 review on strategies for improving CBD bioavailability conclude that it is not known whether these changes to the pharmacokinetic profile translate into greater efficacy.

Everything above is academic in France. Foodstuffs containing CBD as an ingredient are not authorised, under European regulation 2015/2283 on novel foods, for want of any demonstration of their safety. They have to be withdrawn from the market. The Ministry of Agriculture states this for every sales channel: specialist shops, supermarkets and online sales. The only things escaping novel food status are hemp seeds and their derivatives, along with leaves intended exclusively for an aqueous infusion with no added cannabinoid extract.

The ban is not new: CBD meant to be swallowed has never been authorised under that regulation. What changed in 2026 is enforcement, with a national plan presented to the sector on 15 April and rolled out from mid-May. Our tracking of the subject is in the review of the Novel Food regulation and in the article on the withdrawal of gummies.

In February 2026 EFSA published an update adopting a provisionally safe level of 0.0275 mg per kilogram of body weight per day, that is about 2 mg a day for an adult of 70 kg. That value applies only to food supplements of a purity greater than or equal to 98 %, and expressly without nanoparticles. The one regulatory opening for CBD meant to be swallowed therefore excludes precisely the nanoemulsified products. Our dose reference points are set out in how much CBD a day, and our batch analyses are published before purchase.