Cannabis and Parkinson's Disease: Tremors, Sleep and What the Trials Found

About a quarter of people living with Parkinson's in the United States use cannabis. Most of them report that it helps. Roughly two thirds say it improves their tremor.

Controlled trials have found almost no effect on motor symptoms at all. Not a small effect. Not a borderline one. Repeatedly, statistically, nothing.

Both of those things are true, and the distance between them is the actual story. This guide covers what the trials measured, where the weak positive signals genuinely sit, and why the gap between patient experience and clinical data is wider here than almost anywhere else in medicine.

What people with Parkinson's report

The survey data is large and consistent. A study of 1,064 people across 49 states found that 24.5% had used cannabis in the previous six months, with substantial self-reported improvement across the board.

Anxiety improved for 78%. Sleep for 76%. Pain for 72%. Stiffness for 64%. Tremor for 63%.

The same survey found something else worth holding onto. Just over half received no dosing guidance at all, 47% did not know their own dosage, and 64% had no recommendation from a licensed doctor. Nearly a quarter had stopped using in the previous six months, most commonly because it was not helping.

A separate survey of nearly 1,900 people through the Michael J. Fox Foundation's research platform found mild benefits reported for sleep, mood and pain, alongside common side effects including dizziness and cognitive changes. A third of respondents did not know their dose. Over 30% had never mentioned cannabis use to their physician.

What the controlled trials found

The best-designed test in Parkinson's randomised 61 people to a cannabidiol and THC combination or to placebo, measuring change in the standard motor examination scale. The treatment group improved by 4.57 points and the placebo group by 2.77, and the difference between them was not statistically significant, at p equal to 0.379.

A 2021 meta-analysis pooling three randomised trials found a mean difference of minus 0.21 points on the motor scale, with a p value of 0.92, which is about as close to no effect as a number can get. Those authors also noted that every randomised trial in their set carried a high risk of bias.

A 2026 meta-analysis covering eleven trials reached the same place across every domain it measured, including overall disease severity, motor examination, motor activities of daily living and non-motor symptoms. None reached significance, and the authors concluded that cannabinoids have not shown significant benefit in Parkinson's disease.

The classic negative result is older and blunter. A 2004 crossover trial testing oral cannabis extract for levodopa-induced dyskinesia found no evidence of a treatment effect on dyskinesia or on any secondary outcome measure, and reported no objective or subjective improvement in dyskinesias or parkinsonism.

Where the weak positives are

Sifting the trials carefully, there are places where something might be happening. They are not the places people expect.

Of three randomised trials with dyskinesia as the primary outcome, one showed a reduction, and it used nabilone, a synthetic cannabinoid and not plant cannabis. In nabilone groups, quality of life improved. In CBD groups, anxiety and anxiety-induced tremor were reduced.

Read that last phrase precisely. Anxiety-induced tremor is not parkinsonian tremor. Reducing anxiety in someone whose tremor worsens under stress will reduce visible shaking without touching the underlying disease mechanism. That may well be the whole explanation for the 63% who report tremor improvement, and it is a real benefit to the person experiencing it. It is simply not what a motor scale is measuring.

Sleep, anxiety, pain and quality of life are where the signal lives. Motor symptoms are where it does not.

The trial that flipped

There is one sequence in this literature that explains the whole problem, and it is worth walking through.

An open-label study gave 13 people with Parkinson's CBD at 20 to 25 milligrams per kilogram per day. The results looked strong: a 24.7% improvement on the motor subscale and a 17.8% improvement overall, both statistically significant.

It also produced adverse effects in every single participant. Diarrhoea in 85%, somnolence in 69%, fatigue in 62%. Nearly 40% showed elevated liver enzymes at the higher doses, and 23% dropped out because they could not tolerate it.

Then the same research group ran the placebo-controlled version. That is the trial that came back at p equal to 0.38.

Nothing about the open-label study was dishonest. It was a dose-escalation safety study doing exactly what it was designed to do. But it demonstrates cleanly what happens when you remove the control group: expectation, regression to the mean and clinician enthusiasm produce a 24.7% improvement that vanishes the moment somebody gets a placebo.

The unexamined day is a wasted opportunity. Reflect on what you did, what you learned, and how you can improve.

John Dewey

What specialists say

The Parkinson's Foundation's position is unambiguous. Its national medical advisor states that marijuana should never be thought of as a replacement for dopaminergic and other approved therapies, and the organisation writes that there is no conclusive scientific evidence that marijuana is beneficial in Parkinson's.

The harms it lists deserve more attention than they get. Impaired executive function, dizziness, blurred vision, loss of balance and hallucinations. Every one of those is a problem in a population already at elevated risk of falls and of medication-induced psychosis. Parkinson's itself impairs the planning and risk-limiting functions that cannabis further impairs.

Only about 10% of surveyed physicians recommended cannabis use, while 93% wanted cannabis education included in medical training. That combination describes doctors who feel unequipped, not opposed.

Why the two pictures disagree

It is tempting to pick a side here. The more useful move is to understand why both datasets look the way they do.

Surveys recruit people who are currently using cannabis. Anyone who tried it and found it useless mostly stopped, and stopped being available to survey. That alone tilts self-reported benefit upward, and the Parkinson's survey data shows the mechanism directly: nearly a quarter had quit within six months, mostly because it was not working.

Parkinson's is also unusually placebo-responsive. Placebo in this disease produces measurable dopamine release in the striatum, which is not imagination, it is a physiological response visible on a scan. Any uncontrolled study of a Parkinson's treatment is measuring that on top of whatever the treatment does.

And the symptoms people report improving are the subjective ones. Anxiety, sleep and pain are assessed by asking the person. Motor examination is scored by a trained rater watching them move. Those two categories of measurement disagree in almost every field of medicine, and cannabis is not a special case.

None of that means people are wrong about feeling better. It means feeling better and having less parkinsonism are separate claims, and only one of them has been tested properly.

Why we are telling you the negative result

A seed company has an obvious commercial interest in cannabis being good for things. We would rather be straight about this one.

The evidence for Parkinson's motor symptoms is not weak or mixed. It is repeatedly null in the trials designed to detect it. Anyone selling you a strain on a tremor claim is selling you something the data does not support.

What the data does support is narrower and still worth something: sleep, anxiety, pain and general quality of life, in a condition where all four are commonly degraded and where the approved treatments have their own significant costs. That is a legitimate reason someone might use cannabis, discussed with a neurologist who knows about it.

If someone with a movement disorder does grow, there is a practical point worth making that has nothing to do with efficacy. Photoperiod cultivation demands regular precise intervention: switching light schedules, topping, training, tying. Autoflowering plants remove most of that. They flower on age instead of on light, need no schedule changes, and tolerate a hands-off approach that suits anyone for whom fine motor tasks are difficult or unpredictable. Thin Mint x Sour Pinot Auto runs 70 to 75 days seed to harvest at 90 to 120 cm indoors, and asks very little of the grower along the way.

For the evening end of things, where the sleep evidence sits, London Pound Cake is a 70% indica from Sunset Sherbet and a GSC indica phenotype, finishing in 55 to 65 days with the heavy full-body character that makes sense last thing at night. 

The short version

A quarter of people with Parkinson's use cannabis and most report improvement, particularly in anxiety, sleep, pain and tremor. Controlled trials have consistently found no significant effect on motor symptoms, including in the best-designed CBD trial and in two separate meta-analyses.

The positive signals that survive scrutiny sit in sleep, anxiety, pain and quality of life. The tremor improvement people report is plausibly anxiety reduction, which is real but is not disease modification.

Cannabis is not a substitute for dopaminergic therapy, and the balance, cognition and hallucination risks land hardest on exactly this population. Anyone considering it should be having that conversation with a neurologist, not with a budtender.

Barney's Farm has been developing premium cannabis genetics since the 1980s, with over 40 Cannabis Cup wins. Explore our full seed catalog and find strains bred for every climate and skill level.