Indications
What medical cannabis is used for
This list is sorted by the strength of the evidence, not by how often the indication is prescribed — and the two orders are almost the reverse of each other. At the top are two indications with an approved medicine behind them. By far the most frequently prescribed indication, chronic pain, sits in the middle, because serious professional institutions openly contradict one another there.
“Approved medicine” means that an authority has examined the studies and authorised a specific preparation for this indication — not that cannabis flower works for it. “Contested” means: the same studies, different conclusions. “Weak” means exactly that — not refuted, but no basis for confident promises either.
Spasticity in multiple sclerosis
Approved medicine
Nabiximols (Sativex), a THC:CBD spray, is authorised across Europe as an add-on therapy where other treatments are not sufficient. In controlled trials the benefit is real but moderate.
For the appointment: This is the indication with the clearest route — an approved medicine exists. That does not make reimbursement automatic: in Austria Sativex sits in the yellow section of the Erstattungskodex and needs chief-physician approval, and in Switzerland payment only ever runs through a case-by-case decision under Art. 71a–71d KVV.
What 2024–2026 added — Two 2026 meta-analyses quantify what Cochrane 2022 said qualitatively: nabiximols reduces spasticity (standardised mean change −1.29) and pain (−0.88) substantially, disability (EDSS) hardly (−0.17) — symptomatic, not disease-modifying. Signals beyond spasticity (bladder, sleep, gait) are exploratory, certainty low to very low. Rating unchanged.
Severe childhood epilepsies (Dravet, Lennox-Gastaut)
Approved medicine
Purified cannabidiol (Epidyolex) reduced seizure frequency in placebo-controlled trials and holds an EMA authorisation. It is the strongest cannabis-related evidence there is — and it works without THC.
For the appointment: What matters in that conversation: what works here is CBD, not THC. Anyone who concludes from this that CBD oil from a shop helps in epilepsy is confusing an approved medicine with a food supplement.
What 2024–2026 added — Confirmed and sharpened: a 2026 meta-analysis of 7 RCTs (1,154 patients) finds a reproducible reduction in seizure frequency with adjunctive oral CBD (relative risk 0.72; Dravet 43 % fewer seizures at 20 mg/kg/day) with an acceptable safety profile — liver enzyme elevations with valproate, somnolence with clobazam. A syndrome- and age-stratified analysis (46 studies, 2,592 patients, 2026) sees ≥ 50 % seizure reduction in about half, seizure freedom in 5.7 %; which subgroup benefits most cannot be said with confidence. Rating unchanged: the strongest evidence in the field, for purified CBD.
Nausea and vomiting under chemotherapy
Moderate evidence, approved medicine
Nabilone is authorised as an add-on where standard antiemetics fail: older trials, a consistent effect, noticeable side effects.
For the appointment: In Germany nabilone stays on the narcotics prescription form even after cannabis was moved into the MedCanG in 2024 — which matters for how a pharmacy may dispense it.
What 2024–2026 added — A 2025 systematic review of 32 RCTs (1,889 patients) shows cannabinoids beat placebo, but almost all comparisons ran against outdated antiemetics; only one trial compared with today's standard (5-HT3 antagonist + NK1 antagonist + dexamethasone) — with benefit, but notable side effects. No high-quality study, 29 of 32 poor. The authors deem the evidence insufficient for a recommendation. Rating unchanged: a reserve option where the standard fails.
Chronic pain, including neuropathic pain
Contested — professional institutions disagree
The most honest sentence in the whole field: serious institutions read the same studies and reach different conclusions. NASEM found substantial evidence, the IASP declines to make a recommendation, NICE advises against use outside trials, and Cochrane finds small benefits against relevant harms in neuropathic pain.
For the appointment: This is by far the most frequently prescribed indication — and the one where you are most likely to meet marketing that does not mention the dispute. Knowing that the disagreement exists makes for a better conversation with a doctor.
What 2024–2026 added — The two weightiest updates confirm the weak picture. Cochrane 2026 (16 RCTs, 1,750 participants, neuropathic pain): ≥ 50 % pain relief in 21 % vs 17 % on placebo (NNTB 20), ≥ 30 % in 39 % vs 33 % (NNTB 11); against that, nervous-system adverse events in 61 % vs 29 % (NNTH 3) and psychiatric ones in 17 % vs 5 % (NNTH 10); benefits and harms remain "unclear", certainty low to very low. The AHRQ living review (last update July 2025): balanced THC:CBD spray probably a small improvement in pain and function (moderate certainty), THC-dominant possibly a small improvement (low), CBD-dominant probably none — each with markedly more dizziness, sedation and nausea. Rating unchanged.
Sleep disorders
Weak evidence
Small short-term effects in some studies, mostly as a secondary endpoint of pain trials. Tolerance develops, and stopping can make sleep temporarily worse. No cannabinoid is authorised for insomnia anywhere.
For the appointment: The most effective treatment for chronic insomnia is cognitive behavioural therapy (CBT-I) — the guidelines place it ahead of any drug option. Ask about it before you talk about cannabis.
What 2024–2026 added — This is where most has moved — upward, but not far. For the first time there are RCT meta-analyses with sleep as the primary endpoint: 10 RCTs (2,134 participants, 2026) find lower insomnia severity (−3.7 points), better sleep quality and about 34 minutes more sleep; 6 RCTs (1,077, 2025) confirm better subjective sleep quality — only for THC-containing products, not CBD alone. Against that, polysomnography (18 studies, 2025) finds no consistent effect on sleep architecture, and withdrawal after chronic use reliably worsens sleep; the 54-RCT Lancet Psychiatry meta-analysis of 2026 sees more sleep time but no improvement in insomnia symptoms. Rating stays "weak": measurable subjectively, not objectively, long-term data missing.
Anxiety disorders
Weak evidence
Barely any controlled trials of the products that are actually prescribed; the largest meta-analysis found sparse evidence of low quality. THC works in opposite directions depending on the dose: at a low dose it can relieve anxiety, at a higher one it can provoke it.
For the appointment: What is called for here is particular caution, not particular hope — and a history of psychosis is a hard exclusion criterion.
What 2024–2026 added — The direct update of Black 2019 — 54 RCTs, 2,477 participants, Lancet Psychiatry 2026 — finds no significant effect on anxiety symptoms and one additional adverse event per seven treated; the authors' verdict: routine use in mental disorders is "currently rarely justified". A CBD-specific signal exists (8 studies, 316 participants, 2024: large effect) but is small and heterogeneous. Rating unchanged.
Post-traumatic stress disorder
Insufficient — guidelines advise against it
A great deal of observational hype, almost no controlled evidence. The US VA/DoD guideline explicitly advises against cannabis in PTSD, and cannabis use disorders are more common in this group.
For the appointment: If a provider offers you PTSD as an indication without mentioning where the guidelines stand, that says more about the provider than about the treatment.
What 2024–2026 added — Four 2024–2026 papers back the VA/DoD recommendation against cannabis: Lancet Psychiatry 2026 (3 RCTs) finds no effect on PTSD symptoms; a scoping review (26 studies, 3,598 patients, 2026) sees a single positive signal across 7 RCTs — nabilone for nightmares in a 10-person trial — while the largest smoked-cannabis RCT (80 people) shows no difference from placebo; a 2024 review finds no overall benefit and consistent worsening with comorbid cannabis use disorder; the Utah Center for Medical Cannabis guideline (April 2025) calls the evidence "insufficient". Rating unchanged.
Palliative care, appetite and cachexia
Weak evidence
Older evidence for increased appetite in HIV-associated weight loss; trials in tumour cachexia have been largely disappointing. An individual palliative benefit is possible; at group level it has not been shown.
For the appointment: In a palliative situation different standards apply than in long-term therapy — the question there is not the state of the trials alone, but what helps the person now.
What 2024–2026 added — Two 2024 papers confirm NASEM 2017: in cancer, THC increases appetite but yields less weight gain than megestrol alone (3 % vs 11 %), and dronabinol with or without CBD improves neither appetite nor weight, nausea or quality of life versus placebo; an expert panel (J Pain Symptom Manage 2024) finds the evidence insufficient to support or refute cannabinoids for anorexia-cachexia — outside trials they are rarely prescribed for it. For HIV-related weight loss the old dronabinol authorisation remains the basis. Rating unchanged.
And in your country?
Whether an indication can be treated at all depends not only on the evidence but on what the law of your country allows and on who pays. In Austria, for instance, there is no flower — there is dronabinol and three finished medicines; in Germany the statutory insurers have paid for no flower since 30 July 2026.
What is missing from this list
A good deal of what cannabis is actually prescribed for — ADHD, migraine, fibromyalgia, inflammatory bowel disease, endometriosis, Tourette's. These indications are missing not because they are unimportant, but because we have no sources solid enough to carry a statement of our own. As soon as we have, they will be added — by the same standard as all the others.
The full text, with the disagreement over pain, the figures and the primary sources: Where medical cannabis works — and where it does not. If you are right at the beginning: What is medical cannabis?
The evidence guide behind these links is published in German only; the links open the German page.