Skeptical Raptor's Blog hunting pseudoscience in the internet jungle

Marijuana and cancer–Sanjay Gupta’s anecdotes are not science

Because Google is real research.

Because Google is real research.

Unless you were hiking in the Amazon River jungles, with no access to the internet or American TV, you probably have heard that CNN’s chief medical correspondent, Dr. Sanjay Gupta, changed his mind about marijuana (or “weed” as he keeps saying). Of course, this has become big news, because he’s such a “respected doctor” (why is that? Because he’s on TV?), and because a few years ago he was vociferously anti-cannabis.

I have no doubt that Dr. Gupta’s “conversion” to being pro-weed is genuine (and that his previous stance of anti-weed was similarly authentic), but we need to weed out what is real and what’s just smoke about his comments. His first major point about cannabis* was that the United States Drug Enforcement Agency (DEA) considers marijuana to be a Schedule 1 drug, which is defined as “drugs with no currently accepted medical use and a high potential for abuse.”

Dr. Gupta thinks this classification is ridiculous, and on the surface, many people, even those who are not devoted pot smokers, would probably agree. However, this is a political discussion, at least in the USA, and it is hardly a medical/scientific one. The chances of any political party having the fortitude to correct this classification is about as close to 0 as you can get, without actually stating that there is a 0 chance. But if Gupta wants to make a big deal of this, or that he’s so self-centered that he thinks he’ll change the mind of politicians, more power to him. But for me as a skeptic, it is not the most important thing he says.

In his article, he mentions a young girl who “started having seizures soon after birth. By age 3, she was having 300 a week, despite being on seven different medications. Medical marijuana has calmed her brain, limiting her seizures to 2 or 3 per month.” This is simply an anecdote of no quality whatsoever. Did he thoroughly investigate her case to determine if the number of seizures actually went down? Do we know that cannabis has anything to do with the change? Is this nothing more than a Post hoc ergo propter hoc fallacy, that just because she consumed cannabis and the seizures decreased does not mean anything about cannabis’ causative properties with regards to this type of seizure. And then, Dr. Gupta continues with the anecdotes by stating, “I have seen more patients like Charlotte first hand, spent time with them and come to the realization that it is irresponsible not to provide the best care we can as a medical community, care that could involve marijuana.” Why do these TV doctors (like Dr. Oz) think that their anecdotes are better than anyone else’s.

Anecdotes are useless because they aren’t controlled, because they are subject to all levels of bias, and because these stories aren’t peer-reviewed. In other words, anecdotes have no value in science-based medicine. Anecdotes do have value in formulating testable scientific hypotheses, but assuming that anecdote=data, and more anecdotes=more data is simply pseudoscientific. I don’t care what Sanjay Gupta writes or says publicly, but providing these stories as “evidence” that marijuana has a medical benefit is essential like telling me that he observed homeopathy (which is just water) working. It’s laughable.

However, the most important part of Dr. Gupta’s statements appears to be his belief that marijuana can be used in treating cancer. If you spend any amount of time on Twitter, Facebook, or just researching cancer treatments on the internet, you will run across someone claiming that smoking or eating marijuana, or injecting hemp oil (which is manufactured from the seeds of cannabis plants and don’t contain much THC, or tetrahydrocannabinol, the active hallucinogenic agent of cannabis) will cure or prevent cancer.

There is a common belief, again based on a boatload of anecdotes, that cannabis is effective in reducing nausea and vomiting in chemotherapy. In a large scale systematic review of the peer-reviewed literature, which included 30 randomized clinical trials and nearly 1800 patients, there was not much positive evidence that the use of marijuana reduced the incidence of vomiting. In fact, cannabis was only 1.38X better than other types of anti-emetics (medications that inhibit vomiting), which isn’t that much more than unity. However, patients who use cannabis are 4.67X more likely to suspend the use of the drug as a result of the side effects. The authors of the British Medical Journal article (one of the top medical journals) concluded that, “In selected patients, the cannabinoids tested in these trials may be useful as mood enhancing adjuvants for controlling chemotherapy related sickness. Potentially serious adverse effects, even when taken short term orally or intramuscularly, are likely to limit their widespread use.”

In a more recent meta-review of cannabis’ effectiveness in reducing nausea and vomiting in HIV/AIDS patients, published by Cochrane Reviews, the leading publisher of high quality meta-reviews of medical literature, the authors could find no evidence that cannabis had any effect on nausea and vomiting. The best study they could find in their analysis showed only a slight improvement in HIV/AIDS patients, and even then, the error was so large that it could be argued that cannabis might actually have a negative effect, causing more nausea (and less weight gain). The Cochrane authors concluded that:

Despite dronabinol (editor’s note: this is the pure form of THC, the active ingredient of cannabis) in being registered by at least some medicines regulatory authorities for the treatment of AIDS-associated anorexia, and some jurisdictions making allowances for the “medical” use of marijuana by patients with HIV/AIDS, evidence for the efficacy and safety of cannabis and cannabinoids in this setting is lacking. Such studies as have been performed have been of short duration, in small numbers of patients, and have focused on short-term measures of efficacy. Long-term data, showing a sustained effect on AIDS-related morbidity and mortality and safety in patients on effective antiretroviral therapy, has yet to be presented. Whether the available evidence is sufficient to justify a wide-ranging revisiting of medicines regulatory practice remains unclear. 

Since cannabis’ big claim to medical fame is that it is useful as an anti-nausea treatment, it’s surprising that the literature is so lacking in supporting evidence, even when we go looking for it. But this happens a lot in junk medicine. People make claims that are so-often repeated that people accept it as a fact. Then when real scientific studies are performed, especially high quality meta-reviews, we find nothing.** However, because this claim is so profoundly accepted, there will undoubtedly be more data to review in the near and distant future.

However, Dr. Gupta’s most egregious belief is that he appears to be convinced that cannabis has a role in treating cancer. But, is there any evidence out there that actual cannabis or its byproducts have any effect on cancers? Before we start, let’s remember that there are 100 to over 200 different types of cancer (the actual number depends on how some researchers subdivide some cancers) in humans. And each of these different cancers have different pathophysiologies, different genetics, different prognoses, different causes, and different treatments. In other words, it is not one singular disease with one unified course of treatment. Always be skeptical when someone makes some claim that “XYZ cures or prevents cancer”, because that’s going to be nearly impossible. Every cancer is so different with such different physiology, there is just never going to be a magic pill.

Let’s see what’s going on recently in cannabis and cancer research.*** 

  • Breast cancer. A recent review of the relevant research regarding the use of cannabinoids (the essential active compounds of marijuana) in the war on breast cancer concludes that “our current knowledge on the anti-tumor potential of cannabinoids in breast cancer, which suggests that cannabinoid-based medicines may be useful for the treatment of most breast tumor subtypes.” All of this research is based on rodent studies, and the joke in scientific research circles is that we’ve cured cancer in mice for decades. It’s important to understand that only a tiny percentage of therapeutic cancer drugs make it from an animal study clinical trial (about 5%), and even then, only about  less than 8% (pdf) of oncology drugs that enter clinical trials actually end up being approved for use in humans. In other words, there’s only a 0.4% chance of any drug that’s being tested on cells or mice is ever going to end up being approved for human use. There is no nefarious conspiracy going on to block these drugs, it’s that in clinical trials the vast majority of these compounds fail to show effectiveness beyond their safety issues. And in cancer therapy, sometimes drugs that have a 51:49 benefit to risk ratio, and only keep a person alive for a few months get approved. So, the ones that don’t make are remarkably ineffective or unsafe. There is just no evidence that cannabis will treat or prevent breast cancer in humans. None. But let’s hope that it will fall in the 0.4% category. 
  • Colon cancer. In a recent article, researchers examined the effect of cannabidiol, a non-psychotropic ingredient of cannabis, on chemically induced cancerous colon cells in cell culture. The authors concluded that it could prevent colon cancer. Now, if I was harsh about mouse research as being an indicator of clinical success, I’d be harsher about cell cultures. This was tested on chemically induced colon cancers, which may or may not have the same pathophysiology of in situ colon carcinomas. And once again, we lack any clinical evidence that it might work in humans, but I guess this is another bet into the 0.4% chance column.
  • Glioma. Junk medicine hawkers have been pushing hash oil to “cure” gliomas, a type of malignant brain and central nervous system cancer. It is based upon some preliminary research on just 9 patients in Spain. Only two of the patients survived more than a year, while the seven others had a course of the disease not different than what is experienced in typical treatments of gliomas. In fact, in the two patients who survived the longest (yet still died), the effect can be attributable to spontaneous (but temporary) regression of the disease, a fairly common event. In a review of these studies and claims, the author stated, “I didn’t find anything I would call “earth-shattering” or even anything that could be considered credible evidence that hash oil can cure advanced gliomas.” At this point, there is no supporting evidence that hash oil should be considered the first, second, third, or fourth line of treatment for gliomas. Right now, we have observations (which are just barely above anecdotes on the scale of scientific evidence) that hash oil might work. And maybe those observations can be used to create an hypothesis which can be tested in a controlled clinical trial. But until that point, there is no evidence that hash oil works to treat glioma. None.
  • Lung cancer. In this case, it’s not about whether cannabis can treat lung cancer, but whether it actually causes it. In a recent retrospective epidemiological study, the authors stated, “in conclusion, the results of the present study indicate that long-term cannabis use increases the risk of lung cancer in young adults.” There is a “belief” that smoking pot is less problematic than smoking other types of plant material, such as tobacco. That is a false assertion, because the reason why smoking is so tied to lung cancer is that the epithelial lining of the lung is susceptible to carcinogens; moreover, as the authors state, part of gaining effects of the pot is to inhale deeply and hold the smoke in the lungs for a greater period of time than in cigarettes (and certain cigars), it might actually increase the exposure of lung cells to the smoke and any associated carcinogens. To be fair, the evidence is conflicting, but even in one large study that shows no conclusive evidence that there is an increased risk of lung cancer in non-cigarette smoking individuals, there appears to be increased risks for some types of cancer amongst marijuana smokers, including prostate cancer.
  • Clinical trials involving cannabis and cancer. There are currently 8000 clinical trials that are recruiting patients for anti-cancer drugs (and over 19000 that are closed to patients), and as far as I can tell, there are none registered for marijuana, or isolated cannibinoids, dronabinol and nabilone.  On the clinicaltrials.gov website, which tracks every clinical trial registered across the world (and in case you’re going to ask, no one would have a real clinical trial and not register it, since it shows you have regulatory approval to begin a trial), there is precisely one cannibinoid being used to treat cancer, and that is Sativex, which is a patented drug of nabixomols isolated from the marijuana plant. This study is just a Phase 1 trial, it hasn’t recruited any patients, and is years from providing us with any meaningful results. There might appear to be a lot of research into cannabis using cell culture and animal models, but none have been transferred over to human clinical trials. This is not unusual, because even though it seems that there is a lot of research into cannabis and cancer, the total mass of research into other compounds with respect to cancer is substantially larger, because the evidence for both mechanisms of the treatment and clinical successes for these other products are much higher. Look at it this way–there are currently over . Cannabis research is a tiny speck (though not an insignificant speck) of the cancer research world, compared to the vast body of research currently ongoing. Cancer research isn’t randomly throwing compounds at cells and seeing if they work or not, it’s a logical process to determine if a compound has a reasonable chance of inhibiting some part of the cancer growth or development. There is some potential for cannabis, but it probably doesn’t compare to what is currently in the 27,000 compounds in current clinical research, and, in a business sense, the return on investment for researching compounds that have better understood pharmacologic mechanisms of action on cancer are higher.
Let’s assume for a moment that you actually buy into smoking a joint is going to prevent or cure cancer, because you somehow think that an animal or cell culture study is all that you need to justify what you’re going to do. You read the breast cancer study, and you decide that you’re going to kill your nasty breast cancer cells. Let’s see if it’s at all possible to smoke enough marijuana to actually do what is seen in some of these early animal cell studies.
 
First, we need to determine how much THC actually would kill most breast cancer cells. In one study, the researchers determined that it would take a concentration of cannibinoids of approximately 10µmol/L to cause the death breast cancer cells in cell culture. This converts to around 3.14mg/L of THC. So, you’d have to assume that to kill any breast cancer cells, you’d need at least a blood level of 3.14 mg/L to achieve breast cancer cell death.
 
The second piece of information we need is how much THC you could get into your bloodstream by smoking just one joint. According to research, smoking one joint will give you a blood level of around 1.3-6.4 ng THC/mL serum, or about .00013-.00064 mg/L. In other words, to get an anti-cancer effect, you need to light up around 1000 joints per day. Every day. Until every single breast cancer cell would die. Of course, the lungs couldn’t tolerate that, nor probably your ability to function in any “normal” manner. Of course, you could consume this in other ways, for example ingesting it, but again, you’d need to eat more than 1000 joints (because less digestion is less efficient in absorbing THC than the lungs, which is why it is smoked). And these levels may be more generally toxic to the body than just breast cancer cells.
 
If cannibinoids are going to one day be used to treat breast cancer, it’s not going to be through smoking it (or frankly ingesting that amount). A pharmaceutical company is going to extract the exact active cannabinoid (and probably patent it, as it is their right to do****), then put it in a form that allows it to be delivered directly to the breast cancer cells. They are going to rigorously test it for safety and efficacy. They are going to show what precise dose will have the highest effect along with the lowest adverse effects. And they are going to publish that data for public analysis, possibly subject to scientific scrutiny, especially if the results are anything like the glioma research. Given the state of cancer research using cannabinoids, you could expect the first cancer treatment on the market, if all goes well, in 10-15 years. Yes, that’s how long it takes.
 
At this point, you might understand that 1000 joints a day would be expensive and harmful, but you think that you’ll just smoke one joint a day and that would be enough to at least kill some of those evil cancer cells. Except that’s not how it works. Those cancer cells are not going to be affected by lower dosages. In fact, given that there are some increased risks from smoking marijuana, you might actually have a net positive risk of cancer, rather than preventing anything.
 
I have heard the Strawman Arguments that Big Pharma, the FDA, the National Cancer Institute (if the cannabis supporters know it exists), and the US Government suppress all the research that show how great cannabis is for cancer because those groups don’t want pot to be legal. As amusing as that argument might be (and it’s a fairly bad one), if cannabis or any of its components actually could show efficacy against any of the 200 or so cancers, Big Pharma would be all over it. Because, they would not be selling joints, they would be isolating the active ingredient, defining the exact dose, determining how to deliver it to the local cancer site in the body, funding clinical trials, filing documentation with the FDA, then getting it into physician’s hands. This is not an easy process, but it would be a profitable one if it worked. Big Pharma and the National Cancer Institute are looking at everything, and they ignore nothing for potential in treating cancers. If cannabis works (and it might), they are all over it. Big Pharma is providing a lot of the funding for it.
 
Now, I’m only focusing on cancer and cannabis, because Dr. Gupta made a big point about it. Marijuana may, because of its neurological effects, have a lot of possibilities in treating mental disorders and pain. There are some case reports where cannabis may relieve the symptoms of post-traumatic stress disorder (PTSD), although there is a lot of conflicting evidence. Again, Dr. Gupta makes a huge point about it, when a simple PubMed search gives results that are both positive and negative with regards to PTSD.
 
The one area where it appears that there actually might be some near-term clinical trials which could lead to ground-breaking research is in pain treatment. There are a number of studies involving nabiximols, one of the cannabinoids, in treating pain, which could be breakthrough research. But again, we’re at the point of early clinical trials, which often fail, so we are far from considering it clinically relevant.
 
Dr. Gupta does make some valid points. It is ridiculous that marijuana is considered by the DEA to be a Schedule 1 drug, even if the medical evidence of its benefits are right now more intriguing than solid. There are some interesting areas where cannabis might be an invaluable tool in medicine (pain treatment is the closest to reality). I personally dismissed most of the cancer research because it’s too early and many of the studies I reviewed to write this article were laughable quality at best. The glioma “clinical trial” had just nine patients, and it just wasn’t powered in a way to get any information whatsoever, other than, “yeah, something happened, we don’t know why.” 
 
But my biggest point is that Dr. Gupta is not using real science when discussing the medical benefits of marijuana, he’s relying on anecdotes, speculation, preliminary research, and no high-powered clinical trials. In other words, he’s no different than the other TV doctor charlatans. His credibility, at least with me, dropped to nothing.
 
Notes:
 
*Cannabis is the term most used by medical researchers, with cannabinoids or THC used to describe the chemical(s) that are most responsible for its effects. 
 
**Now this article is not going to go down the road of an argument from ignorance. If there isn’t any evidence that cannabis can cure a particular cancer, then that means there’s no evidence. I’m not going to think “oh, they just haven’t tested it on THAT cancer, so how can you not be sure that pot doesn’t cure THAT cancer.” If I read something that asserts that cannabis “cures” breast cancer in women, then I’m going to hunt that research down and read it. If I can’t find it, because it’s not indexed in medical research indices, or it isn’t published at all, then we’re going to conclude that it didn’t happen, because again, we’re not falling for any logical fallacies.
 
***The first step I take before I investigate any internet claim is check with the Cochrane Reviews, which is my primary research tool to find systematic reviews of primary research in human health care, and to find the best information for evidence based medicine. A search of the Cochrane Reviews shows not one systematic review of THC or cannabis in cancer therapies (although several in some mental health issues, which may be a topic for some other article). There are reasons why Cochrane may not have a review can be many, but mainly it’s because there just aren’t enough studies of high enough quality to roll up into a systematic review. That’s a clue, but it’s more a lack of evidence rather than solid evidence. Let me remind you that the quality of the source used matters, and cherry picking primary studies to support a Confirmation Bias is not not how real skepticism works. Remember, if some article was published 10 years ago, and there’s not one single follow-up study, it’s dead on the vine, meaning that no one was able to repeat the data.
 
****There is a US Patent that has, as one of its statements, “This invention relates to the use of phytocannabinoids, either in an isolated form or in the form of a botanical drug substance (BDS), as a prophylactic or in the treatment of cancer.” There are memes that state that Big Pharma knows that cannabinoids cure cancer or else they wouldn’t have patented it. However, patents do not represent peer-reviewed science, and merely conjecture on the part of the patent holder, so that they may potentially block anyone from manufacturing the drug for the use that it claims. There is no evidence that this cannabinoid has any real potential of doing anything until such time that there are randomized clinical trials that support these claims.
 
 
Key citations:
 
 
Comments (79)
Powered by WordPress 4.0
Don't forget to subscribe to this blog through any the services in the right sidebar.