Alan was disoriented and his words were not making sense. His wife thought he might be having a stroke, so she took him to the emergency room where he was seen by the on-call neurologist. When asked, Alan admitted to using cannabis on a regular basis for many years. The neurologist then brought him a printout with the title: “Marijuana Use Associated with Increased Risk of Stroke, Heart Failure.” That was when I got the call asking me if this was for real.
I have not seen this dangerous trend in my clinical practice, but many of my patients have used cannabis for many years so I was motivated to track down the referenced article and review it. If this was a valid concern, I wanted to know so I could inform patients about the risk.
Bias Baked In
A valid study can inform, enrich, and save lives. Poor studies can create fear and ignorance. The bias can lean either pro or con. Either way detracts from our understanding of cannabis and our ability to provide patients with the best care. I obtained a copy of the original article and reviewed it carefully.
The first paragraph gave me a clue. “…cannabis is…the most widely cultivated, trafficked, and abused drug…” (emphasis added)
I had read similar statements in other scientific articles:
- “Cannabis remains one of the world’s most widely used substances of abuse amongst pregnant women.” (emphasis added)
- “Despite increasing public health concerns, cannabis remains the most commonly used illicit drug…” (emphasis added)
What do all these first paragraphs in published scientific articles have in common? Each one reveals a prejudice that makes the rest of the data that follows less trustworthy. Because cannabis has been illegal and vilified for so many years, many publications assume harm even before they are written.
Patients Are Looking for Answers
The scientific literature is teeming with new publications every week reporting on cannabis, cannabinoids, and other medicinal uses for the plant. Some of these studies are well done, but how do you know which are valuable and which are faulty? Scary headlines like “Cannabis Use Predicts Risk of Heart Failure” are dramatic, and often circulate widely in the press and on social media.
Most health care providers know little about the medical use of cannabis; they are not taught the endocannabinoid system in medical schools, and many avoid this sensitive topic altogether. Patients are educating themselves the best they can by reading news articles and reviewing scientific studies made available online, but not everything we read is accurate and not every study is well-designed.
How to Read a Scientific Study
Here are a few guidelines to help you tell the difference between valid information and that which should be taken with a grain of salt:
Animal vs. Human Studies
A study of value to real people reports on a sample that is representative of most human beings. Humans are not mice, so valid conclusions shaping clinical care cannot be reliably based on how mice respond to cannabis. This does not mean that information derived from mice, rats, pigs or other animals is not useful, but the best we can say about animal studies is that further research may be indicated.
“Associations between regular cannabis use and both mental illness and lung cancer have been well established.”  This is an untrue statement. Dr. Donald Tashkin at UCLA designed a study intending to prove that smoking cannabis was associated with increased cases of lung cancer.
To his great surprise he found that this was not true, and eventually published an article that indicated just the opposite.  False statements based on poorly designed studies are sometimes referenced as fact, leading to further poor conclusions. Supporting a hypothesis with weak science is often an indication of prejudgment.
Opinions Stated as Facts
“Mortality post [myocardial infarction] may additionally be increased in cannabis users…”  This statement was used in a scientific article reporting on the damaging effects of cannabis in the cardiovascular system. The use of the word ‘may’ here makes this an opinion, not a fact.
It is also contradicted by other data. Look out for words like “may” or “could” as they indicate a guess, assumption, or opinion rather than a fact backed by observation. The accuracy of the above statement is questionable. In 2018 Johnson-Sasso  published a well-done study concluding: “(Our) results suggest that, contrary to our hypothesis, marijuana use was not associated with increased risk of adverse short-term outcomes following AMI. Furthermore, marijuana use was associated with decreased in-hospital mortality post-acute myocardial infarction.”
Science studies humans when possible, but selection of subjects is difficult, especially when studying the effects of cannabis. As long as the plant is federally illegal and socially suspect, most individuals will be apprehensive about disclosing information related to their use.
In many studies, information is gathered by asking patients if they use cannabis, or any illicit substances (self-reporting). The substances are often listed: “Have you used any of the following: amphetamine, marijuana, methadone, heroin, LSD, PCP, cocaine, other?” Not everyone is going to admit to using a substance included in that list. Would you?
This problem was clearly illustrated in a study done in 1995.  This research collected data from both self-reports and blood tests. When tested, 585 women tested positive for THC, but only 31% of these women had self-reported use of cannabis. As expected, self-reporting clearly carries the risk of under-reporting. If data is collected only on those who disclose the personal use of an illegal substance, that data will be skewed.
Lab testing to select subjects has limitations as well. Serum drug tests may underestimate the use of cannabis because the THC metabolite they test for is only present for a short period of time. A study subject could have used cannabis last week, or a few days ago, and no longer test positive.
Selection skewing leads to statements like, “Compared with non-cannabis users, cannabis users were older and predominantly men [and]…had an increased prevalence of most risk factors including hypertension, tobacco use, and alcohol use.”  This is most likely true for that study’s selection, but not accurate for the general population.
Controlling for Variables
Many, but certainly not all, who use cannabis also use other substances that include tobacco and alcohol. Separating out the subjects who are only affected by cannabis is difficult but must be done accurately for good data on the effects of cannabis alone. Because this task is so challenging, many study results are weakened by confusing the effects of more than one substance.
We Need Facts—Well-Collected and Well-Stated
It is important to review scientific publications carefully and consider any weaknesses stated or implied. The risks and benefits of cannabis as medicine need to be known so the plant can be used safely to everyone’s best advantage. Fear and social attitudes have no place in well-done scientific studies. Unscientific enthusiasm for a widely used herb has no place in the science either. For cannabis to be trusted and appropriately used as medicine, we need impartial facts — well-collected and well-stated.
Thankfully, Alan had not had a stroke. It appeared he had a ‘TIA,’ which is a transient loss of blood flow to the brain with no long-term damage done. But they kept him overnight for tests and to make sure he was safe to discharge. He went home the next day and continued to use cannabis, knowing that the information shared with him by a well-meaning neurologist was not necessarily valid. For him, the personal benefits were worth the possible risks.
Stacey Kerr, M.D. is a teacher, physician, and author living and working in Northern California. After several years working with the Society of Cannabis Clinicians, and co-developing the first comprehensive online course in cannabinoid medicine, she now serves as Medical Director for Hawaiian Ethos. Dr. Kerr is a Project CBD contributing writer.
This article was originally published by Hawaiian Ethos. May not be reprinted without permission from the source.
- Kalla et al. Cannabis use predicts risks of heart failure and cerebrovascular accidents. J Cardiovasc Med, 2018, 19:000-000 doi:10.2459/JCM.0000000000000681
- Tashkin. Effects of Marijuana Smoking on the Lung. Ann Am Thorac Soc Vol 10, No 3, pp 239–247, Jun 2013
- Johnson-Sasso CP et al. Marijuana use and short-term outcomes in patients hospitalized for acute myocardial infarction. PLoS ONE 13(7): e0199705. 2018.
- Shiono et al. The Impact of cocaine and marijuana use on low birth weight and preterm birth: A multicenter study. Am J Obstet Gynecol. 1995 Jan;172(1 Pt 1):19-27.
Preclinical studies are supposed to provide precise, controllable, and translational models for human diseases. But the manipulations that researchers use — like injecting a precise dose of THC into a rat — may sometimes miss important points. What if smoking and injection have different effects? How might this bias results? University of Florida researchers asked just that question, studying the effects of injected vs. smoked THC in rats. Injecting low doses of THC impaired memory.
Researchers from eight different countries recently collaborated on an article warning about the dangers of teens falling in love.
Scientists from the Netherlands recently published the protocol for an upcoming study examining how CBD might be helpful in treating specific phobias. The endocannabinoid system is required for eliminating fearful memories, or at least making them palatable. But why have they only published their methods? It’s part of an ongoing movement to make science more accurate. Many statistical methods assume that the researchers plan all of their data analysis before looking at any of the data.