Water-soluble CBD: what the trials actually measure

Cannabidiol is lipophilic, meaning it dissolves in fat and does not dissolve in water. Once swallowed, it is poorly absorbed by the gut, and the liver then breaks much of it down before it reaches the circulation. This is known as the first-pass effect.

The systematic review by Millar and colleagues, which appeared in Frontiers in Pharmacology in 2018, went through 792 articles. The absolute oral bioavailability of CBD, meaning the share of the dose that reaches the blood, had never been measured in humans. The one absolute value the review records is 31%, for inhalation.

For oral absorption, the most solid data come from the medicine. According to the Epidyolex summary of product characteristics, a high-fat meal in healthy volunteers multiplies peak concentration by 5 and total exposure by 4 compared with a fasted dose. With a low-fat meal the factors are still 4 and 3, and with cow's milk about 3 and 2.5.

These products disperse CBD into droplets small enough to make a stable liquid, sometimes a clear one, but the molecule itself never becomes soluble in water. Several technologies share the same word:

• classic emulsions, in which a natural agent such as gum arabic or quillaja coats the oil droplets;
• nanoemulsions, with droplets from a few tens of nanometres to about 200, depending on where each author sets the cut-off;
• self-emulsifying systems (SEDDS or SNEDDS), mixtures of oil and surfactants that produce these droplets when they meet water or digestive fluids;
• powders, made by drying the emulsion onto a carrier such as starch.

A nanoemulsion is not stable at equilibrium. It tends to separate over time, and the formulation is designed to slow that process. Nothing in the word "water-soluble" tells you which technology was used or how well it works. Our article on the sublingual method explains the options for a route that partly bypasses the gut.

Figures come from three published trials, each giving a single dose to healthy volunteers.

The company that owns the technology funded the first trial, published in Molecules. Its most quoted figure, 4.4, concerns the peak; over 24 hours, exposure is only 1.70 times higher. The powder's maker funded the third trial, published in the Journal of Cannabis Research, and its chief executive is one of the authors.

For a buyer, the second trial is the most informative. This Pharmaceuticals study compares five preparations at the same 30 mg dose in 15 adults, with blood sampled over four hours.

All three emulsions outperformed the oil tincture, by a factor of 1.3 to 2.5. CBD powder simply suspended in water came last. Between the best and the worst formula, the gap exceeds a factor of 4, although every one of these products could be labelled "water-soluble". The maker of the preparations tested also funded this trial.

The best formulas claim 2.3 to 2.85 times the exposure and up to 4.4 times the peak, measured in 14 to 16 people. In a medicine's authorisation file, a high-fat meal achieves 4 times the exposure and 5 times the peak, at no cost.

All these figures are plasma concentrations, and none of them describes an effect. A 2024 review in Pharmaceuticals on strategies to improve CBD bioavailability notes that there is little evidence that a changed pharmacokinetic profile makes any practical difference. Five of its six authors work for industrial groups, Fertin Pharma and Philip Morris.

In France, all of this stays theoretical. The French Ministry of Agriculture recalled in a notice of 20 May 2026 that foods containing CBD are not authorised under EU Regulation 2015/2283 on novel foods. Their safety has not been shown, so they must be withdrawn from the market, in shops and online, food supplements included. The only exceptions are hemp seeds and their derivatives, and leaves intended solely for infusion with no added cannabinoid extract.

No CBD to swallow has ever been authorised under this regulation, so the ban predates 2026. That year brought more intensive checks. Our update on CBD regulation in 2026 follows the file, and our article on the withdrawal of gummies gives a concrete case.

EFSA published a provisional safe level of 0.0275 mg per kilo per day on 9 February 2026, which comes to about 2 mg a day for a 70 kg adult. The level applies only to food supplements with CBD of at least 98% purity and without nanoparticles, made by a process judged safe, with genotoxicity ruled out. People under 25, pregnant or breastfeeding women and people on medication are outside its scope. The one regulatory opening for swallowed CBD thus leaves nanoemulsified products out. For dose benchmarks, see how much CBD per day.