When cannabis becomes your sleep aid

A Cambridge study on cannabis and stroke has been circulating again this week. It was first published in January, so the numbers people are seeing now are the same as those from March.

What?

Megan Ritson and colleagues looked at data from 32 studies, covering more than 100 million people. Of these, 19 studies gave 24 estimates about cannabis. When all the data was combined, cannabis was linked to 37% higher odds of stroke, cocaine to 96%, and amphetamines to 122%. In their main analysis, opioids did not show a significant link overall.

The cannabis studies had very mixed results. On the standard measure of how much studies disagree with each other, they came out at 99%, which is about as inconsistent as a meta-analysis can get. The authors found that smaller studies influenced the results, and after adjusting for this, the cannabis figure dropped from 37% to 12%. Only 14 of the 32 studies clearly showed that drug use happened before the stroke.

So what?

The study does not show how the risk changes with different doses or how often someone uses cannabis. Its strongest evidence is about cannabis use disorder, which means problematic or dependent use. This is very different from using cannabis occasionally, but the paper also cannot estimate the risk for people who use it only sometimes.

The genetic analysis suggests that cannabis use disorder might cause stroke, but it focused on genetic risk for cannabis use disorder, not just general cannabis use. For people under 55, the figure was 14% higher odds, but this is similar to the result for all ages. The authors say there was no significant difference by age.

For me, this study is a reminder to watch for patterns. If you find yourself using cannabis every night to help you sleep, it’s worth paying attention. In the 2024 National Health Interview Survey of 31,509 adults, 3.7% said they used marijuana or CBD most days or every day for sleep. A National Sleep Foundation survey of 1,367 adults found that current users still slept worse than nonusers, though it didn’t show whether cannabis helped. A 2025 review looked at 18 sleep-lab studies, and in the 9 that could be combined, there was no consistent change in sleep duration, how long it took to fall asleep, how much of the night was spent asleep, or sleep stages.

Now what?

Five questions are more helpful than any number this study could provide.

• How many nights out of the last 30?

• Has the dose gone up?

• Can you sleep without it?

• Have you tried to cut down and failed?

• Are you carrying on despite worse sleep, daytime fog or something else?

Then three things worth doing regardless of your answers.

• If you have ongoing insomnia, address it directly. CBT-I is the best first treatment, even before trying medication. Two key parts of CBT-I are waking up at the same time every day and getting out of bed if you can’t sleep, but the full program includes more steps.

• Check your average blood pressure with proper home readings. No matter how cannabis fits in, blood pressure is the main stroke risk factor you can control.

• If you decide to take a break, expect the first few days to be the hardest, and remember that withdrawal can last several weeks. In a national sample of people who used cannabis at least three times a week, 12% had withdrawal syndrome, and 68% of those had trouble sleeping. This is a self-test, not a result from the stroke study.

If you use cannabis just once a month, this study cannot tell you your risk. The genetic evidence is about cannabis use disorder, and the other studies did not track use often enough to set a safe limit.

I wrote in detail about sleep problems in July, including why most effective solutions are not for sale. You can read it here: [archive URL needed]

This paper is a warning about problematic cannabis use. It’s more helpful to ask yourself the five questions than to focus on a number the study never actually tested.

Stay healthy.

Andre