How good is the evidence for medical marijuana?

Wayne Haythorn
6 min readFeb 26, 2019

People are making extravagant claims for medical marijuana. In first-person stories, patients tell of using cannabis extract to cure stage 4 cancers and childhood epilepsy. The cures are miraculous and enthusiasts seem to think it works for everything.

Most of these claims are not accepted by established medicine. In 2017 a committee of the National Academy of Sciences reviewed the scientific research on cannabis as medicine. They studied 24 conditions and found conclusive evidence that cannabis is effective for only one (chemotherapy-induced nausea). This was enough to repudiate federal policy, which is based on the claim that marijuana has no medical use, but it is a tiny fraction of what medical marijuana is being used to treat.

The “Gold Standard”

The Academy defines “conclusive evidence” as multiple, successful, randomized controlled trials on humans. Controlled experiments on humans — clinical trials — are “the gold standard” for medical knowledge. The evidence for medical marijuana is mostly “preclinical” — things like patient reports, cell biology, experiments on mice, and statistical studies comparing marijuana users with nonusers. This is not because marijuana failed to work in clinical trials. The trials have not been done, due to a decision made 45 years ago.

In 1974, the Journal of the National Cancer Institute reported that chemicals from marijuana suppress cancer in mice. These results were reported in the Washington Post:

After this research was completed, but before it appeared in the Post, Congress put marijuana research under control of the National Institute on Drug Abuse. To study cannabis medicine on humans, a researcher must get a variety of permits from the FDA, DEA and NIDA, and get the marijuana from a NIDA-licensed supplier. The DEA and NIDA took the position that by law marijuana is a schedule 1 drug, meaning it has no medical use, so NIDA wouldn’t approve studies of its medical use.

Research yes, medicine no

This didn’t stop research — it just meant no trials on humans. Laboratory science has been legal all along. In fact NIDA was mandated to conduct research into marijuana, so NIH, NIDA, and other government agencies have funded marijuana research for all of these years. They funded a lot of fraudulent scare mongering, but also a lot of good basic research. You can give marijuana to mice, or drop THC onto cells in a petri dish, or examine mothers in Jamaica who smoke while they’re pregnant. The only thing you can’t do is have an FDA trial to develop a medicine for humans, in other words, you can’t do clinical trials.

This has created a situation that scientifically trained people are not used to thinking about. Behind the dam created by the prohibition on FDA trials, a vast lake of preclinical evidence has built up. The National Academy of Sciences committee found 10,759 articles related to cannabis medicine, but almost no clinical trials. Normal science requires controlled experiments to establish reliable knowledge, so despite over 10,000 articles the Academy concluded there isn’t much high-quality evidence for cannabis medicine.

Without clinical trials, the evidence for cannabis medicine remains inadequate. At the same time, the mass of preclinical data is overwhelming. Let’s look, for example, at traumatic brain injury (TBI).

Cannabinoids and Traumatic Brain Injury

In the body, natural cannabinoid levels rise after brain damage. Is this part of the damage, or part of the body’s protective response? Three recent survey articles describe the actions of cannabinoids for TBI and stroke (here, here, and here). Together these reviews summarize nearly 400 original studies showing protective effects of cannabinoids.

Many people who are killed by brain trauma don’t die right away, instead they die days or weeks later from what is called secondary injury. After the primary trauma, damaged neurons spill glutamate, free radicals, and inflammatory signals into the intercellular space. This sets off a chain reaction that causes more damage than the original injury, involving oxidation, inflammation, loss of blood vessel integrity and cell death. Cannabinoids have been shown to mitigate all of the processes of secondary damage. All of them. Cannabinoids protect against:

These results have been shown in cell cultures, in mice, rats, guinea pigs, and pig pigs. They’ve been studied in mice that are genetically engineered not to have cannabis receptors. The treatments tested include plant cannabinoids, animal endocannabinoids, a variety of synthetics, and blocking three different enzymes that break cannabinoids down. This variety of treatments has been tested and found to work for severe mechanical trauma, repetitive mild trauma, fluid percussion injury, occlusion of the carotid artery, and neurological damage caused by HIV virus.

How does all of this relate to quality of life for the animals? Most of the animal studies measured post-trauma outcome variables. Treatment with cannabinoids produced smaller infarcts, better learning and memory, less motor deficit, less anxiety, and fewer seizures. In recovery, cannabinoids promote neurogenesis — that is, healing — around the infarct and in the hippocampus.

Normally, if a handful of studies show that a drug impacts a single mechanism of injury, clinical trials will follow. In the case of marijuana, hundreds of studies show that for TBI and stroke cannabinoids regulate all of the mechanisms of secondary injury and produce better outcomes. These effects have been shown multiple times using a variety of experimental treatments.

So we know it’s good for mice. Does it work on humans? Patients say it does, but suppose we ignore them by chanting “anecdotal” or “they just want to get high”? What about science, with numbers?

Many hospitals screen emergency room patients for illicit drugs. Two studies have examined hospital records to see how marijuana in the blood affects the course of brain trauma. Among patients admitted for intracerebral hemorrhage, those who tested positive for cannabis presented milder symptoms and had less disability on discharge. For traumatic brain injury, patients who screened positive for cannabinoids were significantly more likely to survive the injury:

Overall mortality was 9.9 per cent (44); however, mortality in the THC(+) group (2.4% [two]) was significantly decreased compared with the THC(-) group (11.5% [42]; P = 0.012).

So used recreationally, without any planning or medical supervision, THC is associated with milder symptoms, less disability, and an 80% reduction in mortality rate. According to this data, we would save 4500 lives a year in the U.S. if medical science could deploy the drug as effectively as the average stoner.

But neither of these studies meet the gold standard. They are based on observation rather than a controlled experiment, so they show correlation and not causation. According to normal science, marijuana users might have something else in common that is causing their better outcomes. But hundreds of preclinical studies tell us that the cannabinoids are responsible for the improved outcomes. With all due respect for the gold standard, to say “there is no evidence that cannabis is medicine” is a willful failure to think about the actual evidence.

The ban on clinical trials certainly hurts us. With this mass of preclinical evidence, we know that some cannabis therapy will be a lifesaver for TBI, but without clinical trials we don’t know what that therapy is. The body’s cannabinoid system is complex — it has been evolving for more than 540 million years. The plant is a pharmaceutical laboratory in itself. THC and CBD, the two main plant cannabinoids, have shown both opposing and synergistic effects. We need clinical trials to sort out dosage, which cannabinoids to use, and how they interact with other drugs.

At this time, the only practical knowledge we have is in the experience of doctors and nurses who took the risk of practicing with this medicine. Practitioner knowledge in medical marijuana states is far ahead of the published science. Trials are coming. In Israel, where cannabis science is legal, drug companies are capitalizing on their advantage. In 2017, 120 trials were underway for cannabis medicine. Scientifically validated knowledge will be here in a few years. Until then, I suggest you find an experienced practitioner, and use the only approach that’s known to be better than medical marijuana: don’t hit your head.