CBD and diabetes: what the one trial measured

What the research has measured

The study people point to is a randomised, double-blind, placebo-controlled trial published in Diabetes Care in 2016. Sixty-two people with type 2 diabetes, none of them on insulin, were randomised across five arms for thirteen weeks: CBD at 100 mg twice daily, THCV at 5 mg twice daily, a 1:1 combination, a 20:1 combination, and a matched placebo.

It is a well-run trial and its result is widely misreported, so it is worth being precise about what each arm showed.

The cannabinoid that moved was not CBD

This is the single most important thing on the page, and it is the thing the old version of it got wrong. Tetrahydrocannabivarin is not cannabidiol. They are different molecules, they behave differently, and a bottle labelled CBD does not contain a meaningful amount of THCV.

So when an article says "a study found cannabinoids improved glycaemic control in type 2 diabetes", the study it is describing is almost certainly this one, and the arm that produced the result is the arm that is not on sale. The CBD arm of that same trial, at a dose far higher than most products deliver, did not separate from placebo.

Why interactions matter more here

On most CBD topics the honest summary is "the evidence is thin, and the risk is mostly to your wallet". Diabetes is different, for two reasons that have nothing to do with whether CBD does anything.

Medicine interactions. CBD inhibits cytochrome P450 enzymes in the liver, the same enzymes that metabolise a long list of commonly prescribed drugs. Inhibiting them can raise the blood level of another medicine without anything else changing. People managing type 2 diabetes are frequently on several medicines at once, which makes this a live question rather than a theoretical one.

What a product is actually made of. The 2017 JAMA analysis of 84 CBD extracts bought online found around seven in ten were inaccurately labelled, and some contained measurable THC that was not declared. Several products aimed at this market are also blends with added sugars or carrier syrups, which is worth reading the ingredients panel for.

Why "anti-inflammatory" does the work

Almost every article arguing for CBD in diabetes runs the same chain of reasoning: type 2 diabetes involves chronic low-grade inflammation, insulin resistance is linked to that inflammation, CBD has shown anti-inflammatory activity in laboratory and animal work, and therefore CBD should improve insulin resistance.

Each link in that chain is defensible on its own. The chain as a whole is not evidence. It is a hypothesis, and it is exactly the hypothesis the 2016 trial was built to test — which is why that trial included a CBD arm at 100 mg twice daily, a substantial dose. The hypothesis was tested and the glycaemic measures did not move.

This is worth naming because the reasoning is persuasive and appears everywhere, including on the previous version of this page. A mechanism that sounds plausible is a reason to run the trial. Once the trial has been run, the result outranks the mechanism.

If you try it anyway, measure it

People will try it regardless, and on this topic they are in an unusually good position to find out something real. Unlike pain, blood glucose is a number you already measure, on a device that does not care what you were hoping for.

• Establish a baseline first. Two to four weeks of your normal readings before changing anything. Without it there is nothing to compare against.
• Change one thing. Not a new supplement and a new diet in the same fortnight.
• Use the measure that moves slowly. HbA1c reflects roughly the previous three months, which is why the trial ran thirteen weeks. Day-to-day finger-prick readings bounce around for a dozen reasons and will happily show you whatever you are looking for.
• Tell your prescriber you are doing it. Both because of the interaction question above, and because they can order the test that settles it.

Run that way, a few months produces an answer specific to you that is worth more than anything on this page. Run any other way, it produces a story.

Reading a label without being taken in

The same four things matter here as anywhere, with one addition specific to this topic.

What the formats change

Format changes how much of a dose reaches your blood and how quickly, which matters for comparing products against each other. It does not change what the evidence says.

The trial that produced the null result used oral capsules, which is the format most people reach for. If you see a product arguing that its delivery method is what makes the difference, the honest position is that nobody has tested that claim in this condition.

What this page will not tell you

It will not give you a dose, because no dose has been established for this purpose, and printing a number would suggest otherwise.

It will not tell you CBD lowers blood sugar. The one randomised trial that tested that directly found it did not, and no larger trial has since found otherwise.

It will not tell you to ignore the topic either. THCV produced a real signal in that pilot and larger work is warranted. If that research arrives and holds up, it will be about a cannabinoid that is currently hard to buy and easy to confuse with the one on the shelf — which is exactly why the distinction is worth holding on to now.

The products this page covers