Cannabis, Cardiovascular Health, and Scientific Misfeasance

As Fourth of July weekend approaches, marijuana sales are booming. Should we be concerned?

Somewhere between the extremes of Reefer Madness ("Women cry for it, men will die for it") and Bob Marley ("Herb is the healing of the nation..."), we can find research on the health risks of marijuana use, though the data isn't as conclusive yet as anyone would like.

For the 420-friendly, this newsletter starts on a grim note but ends up cautiously optimistic. I'll discuss new evidence that marijuana use undermines cardiovascular health, then comment on the health risks more broadly.

The new evidence – a review of studies – turns out to be one of the worst pieces of scholarship I've encountered in a reputable journal. I had planned to call it scientific malfeasance, but I learned a new word this week.

Unlike malfeasance, where intentional wrongdoing causes harm, "misfeasance" refers to a lawful act that causes harm because it's done improperly.

Suppose a contractor accepts money to build a house. If he pockets the money and moves to Florida, that would be malfeasance. If he builds the house but uses substandard materials that quickly fall apart, it's misfeasance.

The new review is shockingly misfeasant. The authors may have good intentions, but their materials are substandard, and so their conclusions fall apart at the slightest touch of critical analysis.

Some historical context

Notice anything disturbing in this chart?

The chart shows that even though evidence of marijuana's harmfulness is accumulating, Americans are worrying less and partaking more.

Specifically, the percentage of adult users has risen almost every year since 2002 to just over 20% now. During this time, the percentage of users over 65 rose from 1 to 7%. (Does grandma seem more cheerful lately?)

These changes have been driven by growing legal tolerance. Since 2006, at least one state per year has decriminalized or legalized some form of marijuana use. Recreational use is now legal in 24 states and the District of Columbia; another 15 states permit medicinal use. Some studies suggest the following sequence of events:

(This is just the beginning of the story. 12 states are currently considering legislation that would increase medical access or legalize recreational use. The DEA is in the process of reclassifying marijuana from Schedule I to Schedule III, meaning, in effect, that the federal government will finally acknowledge that marijuana has accepted medical value and a relatively low potential for dependency or abuse.)

All this to say that marijuana consumption in the U.S. will probably continue to increase – and, to the extent that using it is harmful, more people will suffer. But how harmful is the drug?

A new paper on the cardiovascular effects offers a seemingly grim answer.

A new review

This review, led by Wilhelm Storck (University of Toulouse) and a team of French researchers, was published in the journal Heart two weeks ago. (Heart is the official journal of the British Cardiovascular Society and prestigious in its field.)

Storck and colleagues reviewed 24 observational studies on the relationship between cannabis intake and cardiovascular health.

Their conclusions? Taken together, the 24 studies, including a total of more than 432 million participants, linked marijuana use to non-fatal heart attacks, non-fatal strokes, and cardiovascular deaths overall. Most dramatically, marijuana users were just over twice as likely to experience cardiovascular death compared to nonusers, a finding that made national news.

24 studies. Nearly half a million participants. We can't argue with that, right?

Wrong. The shoddiness of the review is egregious and easily explained. I'll share just two of the most damming examples.

1. Biased definition of marijuana use.

Each one of the 24 studies compared marijuana users to non-users. Amount of use was not recorded.

In some of these studies, a "user" was defined as someone who reported using marijuana as little as once in their life. (Sample question: "Have you ever, even once, used marijuana or hashish?")

In other studies, "users" were people diagnosed with Cannabis Use Disorder. They'd been consuming heavy doses of marijuana, more or less daily, for years.

Treating both types of people as "users" obviously biases the data toward finding marijuana harmful.

Oh, and 14 of the studies were drawn from the same three databases. The researchers acknowledged the overlap but didn't manage it.

It's bad enough to mix Cheech and Chong with people who smoked weed once in college and label them all "users". Continuing to sample from that mix magnifies the bias.

What's missing here is...well...everything. Do cardiovascular effects depend on how long a person has been using and how often they partake? Does the potency of the product matter? What about method of delivery (smoking, vaping, eating)?

Bottom line, the review tells us nothing about the conditions under which marijuana use increases the risk of cardiovascular problems.

This is an example of scientific misfeasance: The authors did the right thing by reviewing peer-reviewed studies on a related topic, but their materials (i.e., the studies) were substandard, thus rendering their conclusions virtually meaningless.

Now for a deeper problem.

2. Unaddressed confounds.

Not all studies are equally rigorous or trustworthy. This poses a problem for reviewers attempting to combine data from multiple studies and draw conclusions.

A common solution is to rate the quality or potential bias of each study, and then exclude (or in some way set aside) the weakest or most biased ones.

Storck and colleagues used a relatively new system, the ROBINS-E, to do this. How did their 24 studies fare? Well, 20 of them were rated as "highly biased". The other four reflected "some concerns".

Talk about substandard materials. The review shouldn't have been carried out in the first place.

Does it help that the reviewers had data on more than 240 million people?

No, not at all. Sorry to be crude, but if the data are garbage, more data just constitutes more garbage. And, once you start sifting through this particular dumpster, it gets even worse.

According to Storck and his team, the main source of bias across studies was a failure to rule out confounds. (A confound is a variable that distorts the relationship between variables of interest.)

For instance, in some of the studies marijuana users were compared to nonusers, but data on tobacco use was not recorded.

Marijuana users might show poorer cardiovascular health because they smoke cigarettes, not because of the marijuana. In other words, smoking may be a confound here.

Studies do show strong correlations between marijuana and tobacco use – many people who use one also use the other. Thus, we can't just compare marijuana users to nonusers without measuring tobacco use. We need to sample tobacco-free members of each group, or match them on cigarette consumption, or find some other way to manage this particular confound.

Bottom line: If marijuana users show poorer health outcomes than nonusers, we can't blame marijuana unless other plausible explanations are ruled out.

Recreational marijuana use and health

"I think people need to be educated to the fact that marijuana is not a drug. Marijuana is an herb and a flower."

That's Willie Nelson, in a 2016 interview. Marijuana is a drug of course, as consumption almost immediately increases heart rate and blood pressure, affects perception and mood, and causes people to misspeak during interviews.

We know that marijuana is particularly hazardous for certain groups (e.g., children and drivers) and that excessive use, to the point of being diagnosed with Cannabis Use Disorder, is linked to poor mental and physical health.

This leaves a lot of unanswered questions about the impacts of different types, quantities, and frequencies of less extreme marijuana use.

As the saying goes, the dose makes the poison. Anything, including water, can hurt you if you ingest too much of it, or too much too quickly. Should ordinary recreational users of marijuana be concerned about health risks?

Unfortunately, the long-term impact of marijuana use is hard to gauge, because it's a moving target. Legal use increases every year, the products have become more potent, and there are new or newly popular methods of delivery (vaping, edibles, etc.). The science struggles to keep up, in spite of ongoing calls for more research to better understand the risks.

In the meantime, here's some tentative good news for marijuana users: Several decades of studies and reviews have failed to show a clear link between marijuana use and cardiovascular health, cancer, or a variety of other health problems.

Smoking marijuana can damage lung tissue and promote respiratory issues such as bronchitis. However, even the impacts on lung health remain poorly understood.

It's not hard to find reports of other health risks – the news and social media are eager to alert us – but the data comes from individual studies. One by one, these studies are undermined by small or biased samples, simplistic comparisons between users and nonusers, crude measurements of marijuana intake (e.g., using self-report data on one month's use to estimate use over a period of decades), a failure to adjust for confounds (tobacco use, alcohol intake, activity levels), and so on. In short, the usual suspects.

I wish I could close with something concrete, like "you can consume up to X amount of marijuana per week without increasing your risks of health problems", but the data doesn't permit such concreteness. Not even close. All I can say is that for now, low to moderate levels of consumption (whatever that means exactly), even on a daily basis, have not been clearly linked to increased health risks.

Thanks for reading!