CBD and anxiety: what the trials actually measured

What has actually been tested

Almost everything written about CBD and anxiety rests on three small trials, and all three use the same experimental model: a simulated public-speaking test, which produces anxiety reliably in a laboratory and can be scored on a scale while it happens.

That is 118 people in total. It is a genuine signal — three separate groups, randomised and placebo-controlled, pointing the same way — and it is a very small evidence base for a compound sold in every petrol station. Two of the three tested a single dose in a laboratory on a specific kind of anxiety. Only one ran for as long as four weeks.

The best dose was not the biggest

The Linares trial is the one worth understanding properly, because it is the only one that tested several doses side by side, and the result is not the one people expect. Anxiety fell at 300 mg. It did not fall at 150 mg, and it did not fall at 600 mg either. Plotted, the response is an inverted U: too little does nothing, and so does too much.

It also makes the usual advice — start low and increase until you notice something — a poor fit for this particular compound, because the thing you are looking for may sit in the middle of the range and disappear again above it.

Depression is where the evidence stops

The anxiety and depression pairing is an inheritance from the way these products are marketed, not from the literature. There is no randomised controlled trial of CBD for depression in humans of the kind described above. The case for it runs through animal work, where CBD produces antidepressant-like behaviour in rodent models, and through the observation that anxiety and depression frequently occur together.

Both of those are reasons to run a trial. Neither is a result. The honest summary is that nobody has measured it, and a page that tells you otherwise is describing a hypothesis as though it were a finding.

What a case series can and cannot show

The most quoted number in this field comes from a 2019 case series in The Permanente Journal: 72 adults given CBD in a psychiatric clinic, with anxiety scores falling in 79.2% of them within the first month. It is real data and it is worth reading. It is also not a trial.

• There was no control group. Nobody received a placebo, so there is nothing to compare the 79.2% against — and placebo response in anxiety is famously large.
• People arrive at a clinic at their worst. Scores taken at the worst point tend to improve afterwards whatever is done, an effect with its own name: regression to the mean.
• Everyone knew what they were taking, including the clinicians scoring them.

None of that makes the series worthless. It makes it the kind of evidence that justifies a randomised trial rather than the kind that replaces one.

Questions for a prescriber first

This is the part of the topic where the answer is specific to you rather than general, because anxiety and depression are commonly medicated and CBD interacts with the machinery that clears those medicines.

CBD inhibits cytochrome P450 enzymes in the liver — the same enzymes that metabolise a long list of antidepressants and sedatives. Inhibiting them raises the blood level of the other medicine without anything else changing, which is not a theoretical problem when the other medicine is an SSRI or a benzodiazepine.

Reading a label without being taken in

A 2017 analysis in JAMA bought 84 CBD extracts online and tested them. Around seven in ten were inaccurately labelled: some held substantially less CBD than stated, some substantially more, and some held measurable THC the label did not mention. On a topic where the trials used precise doses, that is the difference between running the experiment and guessing at it.

That last row does more work here than on any other topic. Calm blends are the norm in this category, and a product that pairs CBD with a botanical and then seems to help tells you nothing about which half did it.

What the trials reported going wrong

CBD is usually described as well tolerated, and at the doses sold in a shop that is largely fair. The doses in the anxiety trials were not those doses, and the fullest side-effect picture comes from the epilepsy programme that put purified CBD through regulatory trials at 10 to 20 mg per kilogram of body weight a day — several times what any tincture delivers.

None of that is an argument against trying a shop-strength product. It is the reason the phrase "CBD is safe" needs a dose attached to it, and the reason the interaction question above is not pedantry.

The products this page covers