Cannabis for chronic pain: studies, guidelines and limits

How often is cannabis used for chronic pain in Germany?

Chronic pain is by far the most common reason for cannabis treatment in Germany. In the accompanying survey by the Federal Institute for Drugs and Medical Devices (BfArM) of around 21,000 reported treatments from 2017 to 2022, 76.4% of the complete data sets concerned chronic pain. This was followed by spasticity (9.6%), loss of appetite and wasting (5.1%), and nausea and vomiting (2.2%).

On average, patients had already been treated for their condition for eight years before a cannabis-based medicine was prescribed. In a third of cases, treatment ended within a year, mostly because it did not work (38.5% of discontinuations) and in 25.9% because of side effects.

The BfArM itself stresses that these observational data cannot replace clinical trials. They show how widespread use is, not how well cannabis works for which type of pain.

What do studies show on cannabis for pain?

The evidence shows a small but measurable effect in some of those treated, with low to moderate certainty. Three reviews are particularly informative.

The Cochrane review by Mücke and colleagues (2018) analysed 16 studies with 1,750 participants on chronic nerve pain. Pain relief of at least 50% was reached by 21% with cannabis-based medicines and 17% with placebo. Relief of at least 30% was reached by 39% compared with 33%. The studies lasted two to 26 weeks; data on long-term risks were missing.

A meta-analysis in the BMJ (Wang and colleagues, 2021) of 32 studies with 5,174 adults came to a similar picture: non-inhaled cannabis-based medicines improved pain on a 10-centimetre scale by 0.5 centimetres on average. The proportion of people with a noticeable improvement was about 10 percentage points higher than with placebo. Physical function and sleep improved very slightly to slightly, emotional and social functioning did not.

There are far fewer controlled studies on inhaled flowers. In the Cochrane review, only two studies came from this area, and the evidence on pain relief was of very low certainty.

Which cannabis-based medicines were studied?

Most studies looked at standardised finished medicines and active substances, not cannabis flowers. In the Cochrane review on nerve pain, ten studies tested a mouth spray with plant-derived THC and CBD (nabiximols), two the synthetic nabilone, two dronabinol (pure THC) and only two inhaled flowers. In the BMJ meta-analysis, the substances were swallowed in 30 of 32 studies and applied to the skin in two.

This matters for everyday practice in Germany: here flowers for vaporising are often prescribed. For this form there are the fewest controlled data. Results from studies with mouth spray or capsules cannot simply be transferred to flowers, because onset, duration of action and variation in the amount absorbed differ.

For cancer pain, the data are even thinner. In the BMJ meta-analysis, only four of the 32 studies came from this area; 28 looked at chronic pain without cancer. For cancer pain, opioids following the WHO approach have been the established standard for decades; cannabis is used there at most as an add-on and in individual cases.

Why do studies only partly reflect everyday life?

The studies were mostly small and short. The Cochrane studies ran for two to 26 weeks, those in the BMJ meta-analysis for one to five and a half months. Chronic pain, by contrast, often lasts for years.

In addition, many studies excluded people with previous substance misuse and with significant other illnesses. Yet these are exactly the groups often affected in everyday life. The Cochrane authors explicitly name this as a reason for their cautious assessment. Long-term data on benefits and risks are largely missing.

What do NNT and NNH mean for cannabis?

The NNT (number needed to treat) shows how many people need to be treated for one more person to benefit than in the placebo group. The NNH (number needed to harm) shows how many are treated before one more person has a particular side effect.

For nerve pain, an NNT of 20 means: out of 20 people treated, one more person achieves clear relief than without the active substance. An NNH of 3 for nervous system side effects means: among just three people treated, one more person has dizziness, tiredness or similar effects. The authors of the Cochrane review conclude that the potential benefits could be outweighed by the potential harms.

These figures are averages. They do not predict how you personally will do. But they explain why medical societies are cautious and why a treatment trial needs clear goals and criteria for stopping.

• Small NNT (for example 3 to 5): many benefit additionally.
• Large NNT (for example 20): only a few benefit additionally.
• Small NNH: side effects occur additionally in many.
• What matters is how the two values relate to each other.

Cannabis for neuropathy: what applies to nerve pain?

Nerve pain (neuropathic pain) is the best studied, for example in polyneuropathy, after shingles, in multiple sclerosis or with nerve root irritation from a slipped disc. Even here the effect is small.

The guideline of the German Society of Neurology on neuropathic pain (S2k, 2019) states: cannabinoids cannot be recommended for the treatment of neuropathic pain of any cause, as their effect tends to be small and the rate of side effects is high. Only in individual cases, when other pain treatments have failed, can use as an off-label therapy within a multimodal concept be considered. This guideline has expired; a more recent version was not available as of 21 October 2026.

The same guideline points out that the German National Disease Management Guideline on diabetic neuropathy does not include cannabinoids for painful diabetic polyneuropathy. Established medicines for nerve pain come from other groups, such as certain antidepressants and anti-epileptic drugs.

Cannabis for fibromyalgia: what is the evidence?

For fibromyalgia, convincing evidence is lacking. A Cochrane review (Walitt and colleagues, 2016) found only two small studies with a total of 72 participants, both with the synthetic cannabinoid nabilone. The authors found no convincing, unbiased evidence of a benefit, and tolerability was poor.

The German S3 guideline on fibromyalgia syndrome (2017) makes a negative recommendation: cannabinoids should not be recommended. According to the LONTS guideline 2026, long-term opioid therapy is also explicitly not recommended for fibromyalgia.

Proven approaches for fibromyalgia are mainly adapted endurance and strength training, information about the condition and psychotherapeutic methods. Medicines play a minor role.

Cannabis for back pain: what is known?

For chronic back pain without nerve involvement, there is no specific German guideline recommendation on cannabinoids. The German National Disease Management Guideline on non-specific low back pain (2nd edition) does not list cannabinoids among its drug options. Its focus is on movement, activation and multimodal programmes.

If the pain radiates into the leg because a nerve root is irritated, there is often a neuropathic component. Then the statements on nerve pain apply. In the large meta-analyses, back pain is usually grouped with other chronic pain; dedicated high-quality studies are rare.

For comparison: for opioids in chronic back pain, the LONTS guideline 2026 suggests offering them as an option only if other indicated measures do not work well enough and a response can be shown over time.

What do guidelines and medical societies say about cannabis for pain?

Medical societies agree on one point: cannabis is not a routine treatment for chronic pain. They differ on whether an individual treatment trial can be offered.

The International Association for the Study of Pain (IASP) stated in 2021 that, given the lack of high-quality studies, it does not endorse the general use of cannabinoids for pain. An international guideline group in the BMJ (Busse and colleagues, 2021), on the other hand, made a weak recommendation to offer people with chronic pain a trial of non-inhaled cannabis-based medicines in addition to standard care if that is not sufficient. The group justifies the weakness of the recommendation with the close balance of benefits and harms and calls for a shared decision.

Where does cannabis sit in the pain treatment ladder?

Cannabis is not part of the classic WHO pain ladder. This scheme sorts painkillers into three steps: non-opioids such as ibuprofen or metamizole, weak opioids such as tramadol or tilidine, and strong opioids such as morphine or oxycodone. In its 2018 guideline, the WHO describes the scheme as a useful teaching tool, not a rigid protocol.

For chronic pain not caused by cancer, the LONTS guideline states: before medicines with potential for dependence, non-drug and other drug options should be exhausted. In practice, cannabis is usually only considered after that, and as an add-on, not a replacement.

Can cannabis reduce the need for opioids?

The hope of saving opioids with cannabis is not scientifically proven. The S3 guideline LONTS (version 4.0, 2026) examined this question specifically for the first time and concludes that, because of very low evidence, no recommendation on accompanying cannabis therapy with the aim of saving opioids is possible.

The underlying work relies mostly on observational studies with a high risk of bias. In studies on cancer pain, nausea and vomiting occurred more often with additional cannabis. The guideline group sees a research gap here.

If you already take opioids, do not reduce them on your own when cannabis is prescribed in addition. Both groups of substances can make you tired and reduce your ability to react, and more so in combination. Every change belongs in a doctor's hands.

What side effects and risks come with long-term use?

Side effects are common, mostly temporary and mainly affect the nervous system. In the BMJ meta-analysis, dizziness, drowsiness, nausea and attention problems occurred more often than with placebo. Dizziness was clearly more common in studies lasting three months or longer than in shorter ones.

Psychiatric side effects such as anxiety, restlessness or psychotic symptoms are less common but more serious. More in the overview Medical cannabis side effects.

With regular use, tolerance and a cannabis use disorder can develop. In a randomised trial from the USA (Gilman and colleagues, 2022) with adults who received cannabis through a medical marijuana card for pain, sleep problems, anxiety or depression, 17.1% developed a cannabis use disorder within twelve weeks, compared with 8.6% in the comparison group. Pain intensity did not improve measurably.

In its accompanying survey, the BfArM points out that the average daily amount of THC with flowers was around 16 times as high as with other cannabis-based medicines, and urges caution with flowers regarding misuse and dependence.

• Interactions, for example with sleeping pills, sedatives or opioids: always tell your doctor and pharmacy (Apotheke) about every medicine
• Heart and circulation, especially with pre-existing conditions
• Fitness to drive and road traffic: Driving with a cannabis prescription

Who needs to be particularly careful?

Certain pre-existing conditions and life circumstances shift the balance clearly against cannabis. They should be discussed openly before any decision.

• Personal or family history of psychosis or bipolar disorder
• Young age, especially under 25
• Cardiovascular disease, for example after a heart attack or with heart rhythm disorders
• Previous or current addiction
• Older age with a risk of falls or many medicines taken at the same time
• Pregnancy, trying to conceive and breastfeeding

What does a sensible treatment trial look like?

A sensible treatment trial has realistic goals set in advance and is honestly assessed after a manageable period. Pain medicine knows this principle from opioid therapy: the LONTS guideline recommends setting treatment goals before starting, monitoring the titration phase closely and ending treatment gradually if the goals are not reached.

Realistic is usually not being pain-free, but somewhat less pain, better sleep or being able to walk further. A treatment diary makes the course, benefits and side effects visible for you and your doctor.

1. Set goals, for example pain intensity, sleep duration or everyday activity.
2. Record baseline values over one to two weeks.
3. Have the amount and form set by your doctor and start slowly.
4. Assess benefits and side effects together regularly.
5. Adjust or stop if there is no benefit or side effects are a burden.

What comes first for chronic pain?

For chronic pain, activating and multimodal treatments come first. These include exercise therapy and physiotherapy, pain psychotherapy, information about pain and, depending on the cause, medicines that have been studied for the particular type of pain.

Multimodal pain therapy combines these components in a coordinated programme, often in specialised pain centres. Medicines are one component of it, never the whole treatment.

Sleep is often affected too in chronic pain. In the BMJ meta-analysis, sleep quality improved only slightly with cannabis-based medicines. Sleep problems therefore need their own assessment and treatment, see Cannabis for insomnia.

When should a treatment trial be stopped?

A treatment trial should be adjusted or stopped if the agreed goals have not been reached after a reasonable time or side effects outweigh the benefits. Stopping is not a failure but part of responsible treatment.

After longer regular use, stopping abruptly can temporarily cause sleep problems, irritability and restlessness. How to stop in an orderly way is explained in Cannabis withdrawal symptoms.

• No noticeable improvement in pain, sleep or everyday functioning
• Burdensome side effects such as dizziness, drowsiness or memory problems
• Mental changes such as anxiety, mistrust or loss of drive
• Increasing need without additional benefit, or loss of control over use

Frequently asked questions