The CBD sector has had a turbulent few weeks on both sides of the Channel, with the European Food Safety Authority (EFSA) terminating four novel food applications, just days after announcing incredibly restrictive new safety limits. 

The terminations were issued on February 9 and 12, adding to the dozens of terminated CBD novel food applications since the regulator brought the process to a standstill in 2022, citing persistent data gaps. 

Yet, despite the continued frustration on the part of the industry, much of which has collapsed or pivoted away from CBD since the pause on assessments, the European regulator is finally signalling plans to move this forward. 

EFSA’s February 9 publication of a provisional safe intake level, the first time the authority has established any safety benchmark for CBD, signals long-awaited movement. Seventeen dossiers are now under active scientific review, and EFSA has announced a follow-up webinar for April to help applicants address outstanding data requirements.

The catch is that the benchmark EFSA has set, just 2mg of CBD per day for a 70kg adult, would not only see the vast majority of existing applications rendered moot, but make any CBD approved for sale entirely ineffective for the consumer. 

 As Business of Cannabis reported yesterday, the UK’s Food Standards Agency is pushing toward an autumn 2026 ministerial approval at a 10mg threshold, an already controversially restrictive ceiling. 

Four more applications rejected 

The four terminated applications cover a range of CBD product types including cannabidiol isolate, broad-spectrum CBD, and broad-spectrum CBD distillate, all submitted between 2021 and 2022. 

Three of the four applicants were US-based companies. Sunflora Inc., the Florida-based parent company of Your CBD Store and Sunmed, which describes itself as the largest hemp retailer in the United States with more than 250 locations across 39 states, had two applications terminated, one for a CBD isolate and one for a broad-spectrum extract. 

Elsewhere, KND Labs, a Denver-based cannabinoid manufacturer that counts Europe among its active marketsnhad its application for a broad-spectrum distillate terminated. The identity of the fourth applicant could not be confirmed at the time of publication.

According to EFSA, in each case it identified significant deficiencies in the original dossiers and issued requests for additional information, and extended deadlines for both companies to respond. In KND Labs’ case, EFSA contacted the company on seven separate occasions between April 2023 and February 2024 without receiving a response. Sunflora was chased three times on each of its two applications, again without reply, according to the regulator. 

Both remain active and well-resourced businesses. However, as with the UK’s protracted process, a significant portion of companies sat in regulatory limbo waiting for regulatory clarity have simply disengaged, cut their losses, pivoted to other markets, or shut down altogether. 

For smaller operators, the financial burden of generating the toxicological data EFSA requires was prohibitive. For larger ones, the commercial calculus of pursuing a market with no guaranteed outcome and an increasingly restrictive regulatory environment is equally as discouraging. 

READ MORE…

Why Now?

Four years after first declaring that the safety of cannabidiol (CBD) as a novel food could not be established, EFSA has returned with a long-anticipated update. 

At first glance, the 2026 statement appears to offer movement, a provisional safe intake level, fresh modelling, and a structured review of new literature. However, this is more of a recalibration than meaningful progress. 

Last year, EFSA was formally requested to revisit its 2022 position and ‘prepare an updated Statement to transparently review the current data gaps status identified in the available scientific literature regarding the safety of CBD as a NF’ . 

Since the original opinion, the number of studies on CBD has grown substantially, but more importantly the regulatory pressure has intensified. More than 200 CBD novel food applications have now been submitted to the European Commission, with 17 undergoing risk assessment at EFSA . 

While acknowledging that significant amounts of new literature has been published since it’s original opinion, EFSA reiterated its previous position that it still was ‘not sufficient to address the data gaps and uncertainties previously identified’ . 

In fact, EFSA notes that requests for additional animal and human data ‘have not been provided by the applicants’ . 

Those uncertainties remain wide-ranging. Liver toxicity continues to be the primary concern, with consistent findings in animal studies and aminotransferase elevations observed in human trials. Questions about long-term exposure, potential accumulation, reproductive and neurodevelopmental effects, immune system impacts, and drug–drug interactions remain unresolved. Across multiple sections of the report, the regulator reiterates that the original ‘data gap remains’.

That said, the report is still significant. For the first time, EFSA has derived a toxicological reference point using benchmark dose modelling from GLP-compliant 90-day rat studies. From this, applying a total uncertainty factor of 400, the Panel calculated a provisional safe intake of 0.0275 mg/kg body weight per day,  approximately 2 mg per day for a 70 kg adult . Yet even this figure is framed cautiously. 

EFSA states that ‘the safety of CBD as a novel food cannot be established… until the relevant safety data become available’ . The 2 mg threshold is not an authorisation level; it is a conservative reference point, limited to highly purified (≥98%) CBD supplements, excluding nanoparticles, and not considered applicable to individuals under 25, pregnant or lactating women, or those on concurrent medication .

New literature and its findings

Among the most robust recent studies is the Phase IIb, double-blind, placebo-controlled MedCan1 trial conducted across five Australian tertiary centres. The study randomised 144 patients with advanced cancer to receive escalating doses of synthetic CBD oil, up to 600 mg per day, or placebo over 28 days. Its primary aim was symptom control, and on that measure CBD failed to outperform placebo. 

Total symptom distress scores improved in both groups, with no statistically significant difference between arms. Secondary outcomes, including anxiety, depression, quality of life and opioid reduction, likewise showed no meaningful benefit.

For regulators, however, the most relevant findings were not in efficacy but in tolerability. Participants received a median daily dose of 400 mg, with some reaching 600 mg. Across the 28-day period, there was no signal of clinically meaningful hepatotoxicity. No participant exceeded predefined ALT or AST thresholds (three times the upper limit of normal, or five times in those with liver metastases), and serious adverse events were not considered related to CBD.

A subsequent liver-focused substudy, published in late 2025, pooling data from MedCan1 and the related MedCan2 THC/CBD trial (287 patients in total), examined ALT and AST trajectories specifically. It reached a similar conclusion: over four weeks, medicinal cannabis did not produce clinically significant elevations in liver enzymes, and no difference was observed between CBD-only and THC/CBD formulations.

At first glance, this appears to sit uneasily beside EFSA’s 2026 update, which reaffirmed the liver as a ‘major target’ of CBD exposure and derived a conservative provisional intake of approximately 2 mg per day for a 70 kg adult based on subchronic rat modelling and a 400-fold uncertainty factor . The agency also concluded that key safety data gaps remain unresolved, particularly around long-term exposure and human variability .

Yet the apparent contradiction narrows under closer examination. The MedCan studies evaluated short-term biochemical tolerability in patients with advanced malignancy under specialist supervision. Exposure lasted just 28 days. Individuals with severe hepatic dysfunction were excluded. Monitoring was intensive. The question being asked was whether CBD was tolerable in a defined clinical population over weeks.

EFSA’s remit is fundamentally different. Its modelling is not based on acute enzyme spikes in monitored patients, but on subchronic animal toxicology, inter-species extrapolation, human variability and uncertainty around cumulative exposure . The agency explicitly notes that long-term liver effects and pharmacokinetic behaviour in humans remain insufficiently characterised.

In that sense, the clinical data does not directly refute EFSA’s position, it does complicate the narrative. They suggest that, at least in the short term and under medical oversight, moderate to high doses of pharmaceutical-grade CBD do not trigger overt hepatotoxicity. What they do not establish is long-term safety in healthy consumers, or resolve questions around chronic exposure, accumulation, reproductive endpoints and population-wide variability.

The MedCan liver findings may weaken claims of acute hepatotoxicity at moderate doses. They do not, on their own, close the evidentiary gaps that underpin EFSA’s cautious assessment.

The divergence between clinical medicine and food law now defines the CBD debate. Trials may demonstrate tolerability under supervision, but EFSA’s mandate is population-wide, long-term safety. The 2 mg benchmark signals movement, but it also leaves almost no room for a functioning industry.