What is the evidence for medical use of cannabis?

“Results reflect the … relative dearth of high-quality evidence related to cannabis for therapeutic purposes”

“The level of evidence for the use of medical marijuana among people with disabilities varies greatly and has a clear lack of methodologically sound studies”

“Evidence is gathering … but further research is required to declare cannabinoids a medicine”

“These studies suffered from multiple limitations, including selection bias, lack of standardized dose and route of administration, absence of blinding, recall bias and lack of a control population”

“The long-term safety profile of chronic cannabis use has not been well defined, mainly due to the heterogeneity of preparations, varying routes of administration, and the lack of controlled studies addressing safety”

“Evaluation of these low-quality trials … was challenged by methodological issues such as inadequate description of allocation concealment, blinding and underpowered sample size”

These are some of the things the experts are saying about the research into medical uses of cannabis to date. To prepare this article, I read 11 “review” articles — articles that pull together all of the research out there and reach conclusions.

The authors are unanimous in saying there is little research, and many of the studies that have been done are really small — less than two dozen patients. Some lack another treatment for comparison — either a placebo or another drug. Other studies can’t be blinded, because patients can tell whether they are getting cannabis, either because they are smoking it or they feel stoned. Others are “observational” studies (some patients are given cannabis and the researchers see what happens), considered the poorest type of evidence.

But there is some good research, and it shows that cannabis might be useful in a few health conditions:

  • Pain
  • Epilepsy
  • Inflammatory bowel disease (Crohn’s disease and ulcerative colitis)

These medical uses of cannabis are not just based on folklore (or, in this case, folk-singer lore). Science has shown that some of the chemicals in cannabis (cannabinoids) mimic natural chemicals in the human body that are important in the nervous system and the immune system. This is why its use is being taken seriously as a potential addition to the drugs we already have.


The toughest pain to control is cancer pain. A review of five randomized controlled trials looked at cannabis to treat this type of pain. Although one of the studies didn’t find any difference between cannabis preparations containing both THC and CBD (nabiximols) and placebo, there was definite effect in two larger studies with more than 150 patients each. Two earlier studies that had looked at short-term effects in small numbers of patients also found an effect. This review was funded by a Canadian medical cannabis company but was conducted with Sunnybrook Health Sciences Centre and the University of Toronto, so the oversight, ethics and evidence should be solid.


I have known someone with very serious epilepsy who must cope with several seizures each day. This type of disability limits a person’s life and can even pose a danger. In some cases, the well-established drugs that control epilepsy in most people don’t work or cause serious side effects.

Four reviews of the research found evidence that cannabis preparations can reduce or even stop seizures in people with severe seizure disorders. But the effectiveness seems to vary from patient to patient.

A review that looked at cannabis research for many different disabilities found robust evidence only for epilepsy and pain. Another review that looked specifically at the non-intoxicating cannabinoid CBD for epilepsy showed that it was effective but also showed high rates of side effects, including sleepiness, loss of appetite and diarrhea. A third review found positive effects of CBD in the worst types of epilepsy: Dravet syndrome and Lennox-Gastaut syndrome.

Inflammatory bowel disease

There was a lot of hope that cannabis could help with IBD through its effects on the immune and gastrointestinal systems, but findings are mixed.

A review of five studies (only two of which were high-quality and all of which were small) found positive and remarkable effects on symptoms of IBD. However, two systematic reviews (one for Crohn’s disease and one for ulcerative colitis) by the Cochrane Collaboration — well known for its high-standard reviews — found that the effects were uncertain, the studies small, and conclusions unreachable. However, because of the positive findings of some studies, treatment of IBD needs to be tested in larger groups.

Why not more, better, larger studies?

That’s the question. When I attended a panel on medical cannabis earlier this year (2018), one of the panellists representing a cannabis company buttonholed me afterward about the difficulty in conducting studies. He pointed out that the best level of evidence is provided by randomized controlled trials involving hundreds or thousands of patients. But running those trials is incredibly expensive. He also argued that it is difficult to use placebos when testing cannabis, since those taking cannabis may be smoking it and it may be unethical to let patients go without some kind of treatment. This also means it might be impossible for the trials to be double-blinded — when neither patients nor researchers know who is getting the drug and who the placebo or comparison drug.

These are good points, but they sound like companies are throwing up barriers to running studies.

First, many randomized controlled trials compare the test drug with the standard treatment or best treatment (“gold standard”) rather than with a placebo. This is certainly the case for drugs for mental health problems, as it would be unethical to let a serious mental health problem go untreated. So, trials do not need to use a placebo.

Also, many trials are single-blinded rather than double-blinded. That is, the patients can figure out what they are getting, but the researchers don’t know who is getting what and aren’t influenced by that knowledge. In many trials, this is necessary, if not optimal.

The real issue is money. Running the kind of studies that you need to submit to Health Canada to have your drug recognized as a safe and effective treatment for a particular condition takes years and millions of dollars. Right now, cannabis companies are not that rich.

But in Canada the cannabis companies have formed a consortium that could fund trials. As well, disease foundations in Canada and other countries (like, the one to the south) are interested in the promise of cannabis, and they put money into research. With its open regime, Canada is a good place to carry out international studies of medical cannabis, which would not face the legal hurdles that exist in many other countries.

There are many conditions for which safe and effective drugs are needed. If cannabis can help, it should be prescribed and monitored by doctors to ensure it works for patients, and it should be covered by drug plans. It should not be bought at a dispensary for self-treatment, often by patients who can’t afford it.

As I mentioned to the cannabis company representative, doctors were burned on opioids. The manufacturers claimed opioids weren’t addictive, which wasn’t true. Studies would give doctors full information on cannabis’ effects and risks, so they could inform their patients. Doctors and patients deserve no less.


Medical cannabis use carries a risk of dependence and of lung cancer

I like to make my position clear from the outset: I support the legalization of cannabis (marijuana and hashish). I think criminalizing it makes criminals out of otherwise upstanding people who use it recreationally, either harmlessly or harmfully (of more, anon). And criminalizing it is a latter-day prohibition, often putting its sale in the hands of organized (and, frankly, disorganized) crime. The fact that cannabis has been illegal has never stopped anyone in Canada from using it and has probably stopped users from admitting their use to their doctor, social worker, parents, etc.

But does it have legitimate medical benefits? Should it be used to help with symptom control in pain, chronic neurodegenerative diseases and so on? Should it be prescribed and monitored, rather than just leaving patients to buy it at Shoppers Drug Mart and experiment with the dosage?

This spring, I attended a panel discussion by cannabis producers as part of the Science Writers and Communicators of Canada conference in Vancouver. Naturally, the producers were in favour of the medical use of cannabis, as this is a potential market for them. It also gives the leaf a patina of credibility and respectability.

I am all in favour of treating cannabis as a potential drug, including research into its safety and effectiveness and licensing for proven uses.

But I heard a few things said that were untrue, skirting serious issues in the medical use of cannabis. Sometimes I wondered whether the speakers had been smoking something.

I want to review a few things that were said.

  • Only 3% of cannabis users become dependent.
  • Cannabis doesn’t cause lung cancer, or does so at lower rates than tobacco.

Let’s look at these claims. The best study I have found on rates of cannabis dependence is a 2012 review of data to date [Degenhardt and Hall (2012). Extent of illicit drug use and dependence, and their contribution to the global burden of disease. The Lancet 379(9810): 55-70].

It found a 9% rate of dependence among users.

A later study by the same authors estimated the rate of cannabis dependence in Canada at 0.6% of the high-income population in 2010 [Degenhardt et al. (2013). The global epidemiology and contribution of cannabis use and dependence to the global burden of disease: results from the GBD 2010 study. PLoS One 8(10): e76635, Table S4]. The rate is undoubtedly higher for the low-income population, as the authors’ 2012 study found.

This is in line with the recent Statistics Canada survey that found that 14% of Canadians over 15 had used cannabis recently (previous three months), with 56% of that 14% saying they used cannabis daily or weekly.

I have seen people become dependent on cannabis, and my observations are in line with the higher estimates.

So, definitely more than 3% of users.

But whether it causes dependence shouldn’t matter anyway, because, evidently, smoking cannabis doesn’t cause lung cancer. Except that it certainly does.

In fact, it increases the overall risk by five times, according to a case–control study carried out in New Zealand and published in 2008 (Eur Respir J. 2008 Feb; 31(2): 280–286. doi: 10.1183/09031936.00065707. For each year of cannabis smoking, the risk increased 8%.

For every study, there is another that contradicts it. Some other studies have not found this association. One of the problems is that a lot of people who smoke cannabis also smoke tobacco, or smoke both together. The New Zealand study used statistical analysis to separate the effects of cannabis smoking from tobacco smoking, but some other studies have not found a separate effect.

An analysis that combined data from several studies had mixed conclusions [Int J Cancer. 2015 Feb 15; 136(4): 894–903]. On one hand, the authors found that having ever smoked cannabis did not alter the overall risk of lung cancer. On the other hand, risk was clearly higher when they looked at a particular kind of lung cancer – adenocarcinoma – especially if the patient smoked a joint or more per day, and especially if they had smoked for more than 10 years. The difference also showed up in people who had smoked cannabis for 20 years or more, for all types of lung cancer.

OK, the risk of lung cancer associated with cannabis smoking appears to be lower than the 23-fold higher risk linked to tobacco smoking (US estimate in men only). But yes, you could get lung cancer.

There is a growing body of evidence on medical use of cannabis, and more trials are needed. I’ll try to get to this in future blog posts.

But arguments for medical use are not helped by unsupported statements and claims. Risks of dependence and lung disease need to be addressed in medical use. Downplaying them or wishing them away doesn’t help anybody.

BC provincial government site publishes unsupported lifestyle advice on hot flashes

At the risk of being a one-note Johnny, I am very concerned when lifestyle advice to menopausal women from normally credible, reputable sources is unsupported by evidence, at best, or is basically old-wives’ tales, at worst. If other health information were based on such equivocal and weak evidence, it would not be considered acceptable.

The Globe and Mail recently published a lifestyle article on hormone therapy for hot flashes, which I read with interest. It mentioned that HealthLink BC, a patient health information site run by the government of British Columbia, was recommending lifestyle advice that my own literature review indicates is unsupported. By contrast, its advice on medical therapy is generally referenced and based on evidence.

This is what HealthLinkBC says:

  • Avoid using tobacco or drinking a lot of alcohol. They tend to make hot flashes worse.

The evidence on both of these is equivocal (see my previous post). There is some limited evidence from well-designed observational studies that never smoking and quitting smoking are associated with fewer hot flashes. The evidence on alcohol is mixed, with some studies showing alcohol decreases hot flashes.

  • Manage stress. Stress can make hot flashes worse.

Well, there is evidence showing that being happy and contented actually increases hot flashes and stress decreases them. There is little good evidence about effects of stress.

  • Exercise regularly, and eat a healthy diet.

As I indicated in my earlier post, the evidence on exercise is mixed and shows only modest changes. There is evidence that obesity (but not healthy weight) is associated with hot flashes, so losing weight is advisable as a way to decrease hot flashes only if you are obese.

  • Try rhythmic breathing exercises. This is called paced respiration. It can help you meditate and relax, and it may reduce your hot flashes.

OK, the first randomized controlled trial on this recently reported. There was a modest (less than 20%) decrease in hot flashes in the group that used this paced respiration method in which you slow your resting breathing rate (kind of conflicts with the exercise advice). But (get this) the control group, which listened to music instead, had a 44% decrease in hot flashes. Keeping in mind that placebo effect is huge in hot flashes, we should all be listening to music rather than breathing slowly.

  • Drink cold liquids rather than hot ones.

Hot liquids (and anything hot) can bring on a hot flash, but there is no evidence that they increase the overall frequency or severity of hot flashes.

  • Eat smaller, more frequent meals. Digesting a lot of food can make you feel hotter.

No evidence. Also, while the immediate trigger of a hot flash may be warm ambient temperature and other heat sources, the problem is not that women affected are overly hot. Menopausal women in Canada have hot flashes in minus 20 degree Celsius weather. Hot flashes do not feel like being hot from ambient temperature or exercise, and body temperature before a hot flash is normal and then actually rises measurably during the hot flash.

  • Stay cool
  • Keep your area cool. Use a fan.
  • Dress in layers. Then you can remove clothes as needed.
  • Wear natural fabrics, such as cotton and silk.
  • Sleep with fewer blankets.

Staying cool does help prevent the immediate onset of hot flashes and helps women recover from them quickly when they occur. However, from personal experience, I’m not sure it reduces the overall frequency.

No evidence.

From HealthLinkBC’s medical therapy section:

  • Black cohosh may reduce or prevent hot flashes, depression, and anxiety.

There is a reference for the paragraph containing this statement to a Health Canada page on traditional herbal use of black cohosh, which includes menopausal symptoms. There is no reference to the medical literature, which has plenty on black cohosh. There have been some small randomized controlled trials (around 300 patients) showing benefit of black cohosh in hot flashes and published in a respected, peer-reviewed journal (Obstetrics and Gynecology). However, a Cochrane systematic review pooling data from 16 randomized controlled trials involving more than 2000 women with menopausal symptoms found “insufficient evidence to support the use of black cohosh for menopausal symptoms.” However, since some studies found some benefit, the review called for more research on this traditional herbal remedy.

I find a lot of information for patients tries to get around the fact that the evidence is poor by using words like “may” or “might.” As in, “black cohost may reduce or prevent hot flashes.” Or it may not. I suspect many patient-readers looking for something to try are not going to pick up on the subtle distinction of a remedy that may or will provide some benefit. And this is concerning, as many herbal remedies are not risk-free. Black cohosh has been linked with liver-damage events in people who have taken it, although a meta-analysis of black cohosh studies found no evidence of liver problems. Black cohosh may or may not have risky side effects.

But why not just be upfront with readers about the state of the research evidence on various lifestyle advice and herbal remedies? And why tell women something is going to help when it isn’t? I also find that a lot of patient advice is good, general health advice. But it’s like health authorities are trying to trick readers into following advice by selling it as something that’s going to help with a particular problem.

There are lots of good reasons to eat properly, quit smoking and limit alcohol. Like heart disease, cancer and diabetes. Don’t tell women healthy living will help their hot flashes when it won’t.