Questions About The New Study Alleging 6-Fold Risk of Heart Attack in Heavy Cannabis Users Under Age 50 (edited 3/23/2025)

confounding variables

(Updated 1515ET, 03/23/2025, with additional research and commentary attached to the end of the original post.)

Headline news from USA Today that just dropped into my feed:

Cannabis users under 50 are 6 times more likely to have a heart attack, new study shows

(Greta Cross, writer; published 3:48pm ET; updated 4:22pm ET, March 21, 2025)

"Until we have more solid data, I advise users to try to somehow put some regulation in the using of cannabis," researcher Ahmed Mahmoud said.

Well, yeah. That’s what I say, too. The part I wonder about is the “until we get more solid data” assertion. It isn’t as if cannabis effects on the cardiovascular system and related physical effects haven’t been studied already. Studies go back decades. Cannabis, pot, reefer, marijuana, weed, hashish, you name it, has been studied for many years. More studies all the time. But this is the first one I’ve read alleging that

people under the age of 50 who consume marijuana are about 6.2 times more likely to experience a myocardial infarction, commonly known as a heart attack, than non-marijuana users. Young marijuana users are also 4.3 times more likely to experience an ischemic stroke and 2 times more likely to experience heart failure, the study shows.

I’ve read many studies on cannabis effects on physical health over the years, and this one brings a lot of drama. “Huge if True”, as the social media jet set would put it.

So I thought I’d have a look at the original study, recently accepted and set for publication by the Journal of the American College of Cardiology, March 18, 2025. https://www.sciencedirect.com/science/article/pii/S2772963X25001152?via%3Dihub

Before I get into reviewing the details, I’ll re-state the position on the legal status of cannabis that I’ve held for decades: my non-negotiable bottom line is that personal possession and cultivation of cannabis in small household quantities by adults should be Federally legal and unregulated, similar to the provisions for moderate quantities of home brewing of beer and wine. That’s an absolute minimum restoration of a Constitutional right that should never have been at issue.

I’m ambivalent about the existence of the commercial market, but inclined to support it. I oppose some aspects of it. In the states where legal retail sales are presently ongoing, the currently continuing Federal prohibition on cannabis—with its restrictive effects on banking and leasing, and consequent requirement for a large amount of front-end cash investment—has indisputably worked to the advantage of wealthy venture capitalists and wealthy investors at the expense of small businesses and cultivators. I don’t think that’s a fair result. But current Federal government policy practically ordains it. Furthermore, I’m opposed to profiteering and any move to permit the advertising of cannabis in any television or radio medium. (I’d also support a ban on the TV advertising of alcohol and prescription pharmaceutical drugs.) I’ve written more extensive Substack posts on the subject earlier. Suffice it to say that my views are not informed by any financial interest connected with the cannabis business.

I’ve also written in detail about my personal views on the increasing potency of cannabis over the decades. In the the early 1990s I began hearing claims that increasing THC concentration was intensifying the effect of marijuana to the point where it was practically an entirely different substance bearing little resemblance to the mild, sensual mind altering properties of earlier eras. At first, I was dismissive. By the mid-1990s, I found that I had to agree; the effect, the high, from smoking pot had changed markedly. For one thing, where I had once regarded higher potency pot as a means of lowering my exposure to amount of smoke ingested into my throat and lungs, I now found that even one hit was too much. (Shades of tobacco! where one deep-lunged hit of tobacco and the vertigo can put me on the floor.) The new superpot contained so much THC that one inhalation was sufficient to trigger my receptors almost instantly. Not a bongload, either- a few little flakes were sufficient. The amounts were so small that I actually began mixing it with tobacco in order to have enough combustible substance in my pipe. Routine ingestion of tobacco, how ignominious. Anyway, I eventually found that even using very small amounts of super-engineered THC-laden weed had an onset that was just a little too abrupt for me to really find the lift enjoyable until it had worn off a bit. I think the medical term is “bolus effect”—the saturation effect resulting from the use of a potent product that hits the relevant receptors all at once, due to the means of ingestion. Bolus effects can be produced either by inhalation or injection. The front-loaded saturation works to produce an effect that comes on strong and wears off fast, and it’s most prominently associated with cocaine, in the form of crack. The effect of a THC bolus bears no similarity to crack or any other substance; it’s still THC, the cannabinoid realm. But it does hit roughly as fast as making the circuit in a revolving door, and there’s enough of a shift in consciousness that it can be experienced as disorienting, especially for new initiates and those who have little experience with cannabis. I have ample experience with cannabis. I can deal with the fast onset, but I don’t care for it.

The problem is that ultrahigh THC cannabis began taking over many of the markets 25 years ago. That effect is what’s considered a “normal pot high” nowadays, and it’s been that way for some years. I shifted from regularly using cannabis to seldom using it 20 years ago; there have been years when I haven’t used it at all. I was once a member of the heaviest user cohort, as a male in my late teens and early 20s. I can recall what it was to experience a high from smoking milder—but also more subtle and distinctive—varieties in the 1970s, and that’s a perspective the kids lack. I also know what the superweed effect is like. It’s my fieldwork opinion that in order to enjoy using large amounts of the New Cannabis on a regular basis, one has to be pretty much constantly using it. Otherwise, the effect is experienced too intensely to feel entirely comfortable. This encourages regular use to the point where the altered state becomes more or less the new baseline. And it has to be noted here that between the intensity of high-THC pot onset and the subjective advantages of using it all the time in order to counter that effect, the high-test product appears to have resulted in a side effect profile that any pothead I know of would have laughed out of the room in the 1970s or 1980s.

No one I know ever experienced cannabis hyperemesis syndrome, for example. And while I’ve known regular cannabis users who have told me that they found it difficult to quit, I never had the idea that they found it as difficult as quitting tobacco, or alcohol, or heroin. I was able to use cannabis without craving it when deprived; I often abstained voluntarily for a few weeks every year. It’s difficult for me to set aside my personal bias and skepticism when reading the recent studies on the “addictive” properties of cannabis. But it’s definitely being increasingly reported. So are acute negative effects like anxiety, paranoia, disorientation, and even more serious reactions like derealization and psychotic episodes. Those effects were not unknown in the 1970s and 1980s. But they were rare side effects, and practically never associated with violence or aggression. Now negative reactions are much more commonly reported. Admittedly, most of my research findings have been found in Internet social media discussions and chats: reddit, Erowid, various other websites and social media platforms. I think social media conversations have a way of encouraging the momentum of content in the direction of concentrating on drama, dysfunction, and dilemmas; the weather on the Internet always seems to be bad. But even given the force multiplier illusion, I’m willing to grant that marijuana in the 20%-30% THC zone—or even higher, as with the resin concentrates—turns into a product with a much increased side effect profile, when compared with lower-test product.

That lengthy introduction is my way of saying that I’m open to reviewing the evidence that the reported increase in negative effects from high-potency cannabis might even include an increased risk of serious negative cardiac effects—including high blood pressure, heart attack, and stroke.

With that in mind, it’s my conclusion that the JACC research study referenced in the USA Today article has some glaring flaws, and the claims that it makes are wildly exaggerated. I have some serious reservations about its methodology, and also some about claims to effectively counteract the possibility of confounding factors interfering with the results by using weighted comparisons of the cannabis user population with the non-using population.

Excerpted from the text of the study:

Methods

Data Source: This retrospective cohort study utilized the TriNetX health research network, which aggregates de-identified electronic medical records from healthcare organizations worldwide. The specific network used for this study was Research, which contains data from 53 healthcare organizations (HCO) throughout the United States. Given the nature of the study and included date, IRB approval was waived. (Italics added)

“IRB approval” refers to the Institutional Review Board that crafts guidelines for best practices in medical studies.

Study Population: Focused on adults aged ≤50 years between 2010 and 2018, divided into two cohorts. 1) The cannabis-user group with cannabis use diagnoses (ICD-10: F12.1, F12.9, F12.90). 2) The non-cannabis-user group. Both cohorts were free of significant comorbidities at baseline, including hypertension, hyperlipidemia or LDL>100 mg/dL, coronary artery disease including prior MI or history of coronary interventions, diabetes mellitus (DM) or HbA1c >7, and tobacco use. The relatively healthy status of these patients was a key aspect of this study design, ensuring that comparisons were not influenced by pre-existing high-risk conditions. Propensity score matching (PSM) was performed, covariates were matched by 1:1 PSM using the greedy nearest neighbor matching algorithm with a cutoff of 0.1 pooled standardized mean difference (SMD). The following covariates were included in the PSM: age at index, sex, race, chronic kidney disease, depressive episode (DD), family history of ischemic heart disease, pregnancy, cancer history, creatinine, LDL, HbA1c, BMI, systolic blood pressure.

(“Propensity score matching” refers to the weighting used in conjunction with the factors indicating increased health risks, in order to provide a more accurate comparison of the cannabis user and nonuser cohorts.)

Note the assertion made by the study: “The relatively healthy status of these patients was a key aspect of this study design, ensuring that comparisons were not influenced by pre-existing high-risk conditions.”

That statement implicitly assumes that all such “pre-existing high-risk conditions” were accounted for. However, in my opinion, the list provided is plainly not comprehensive. I realize that a complete inventory is too much to ask; still, there are some other conditions that aren’t included as possible confounders, even though they have a long track record of association with cardiac problems and stroke—most notably alcohol and illicit stimulants. I’d also include statistics on a record of incarceration, because confinement environments induce significant stress and PTSD in some of the population subjected to it. So I’m not assured that “both cohorts were free of significant co-morbidities”—especially given that the cannabis users in question were not simply a random sample of users, or even self-reported daily users; the >90,000 “users” had all been officially diagnosed with Cannabis Use Disorder, which would seem to indicate that they’ve been specifically flagged on the basis of details found in their medical records; they aren’t merely self-reporting daily use, they’ve been defined as dysfunctional for it. That’s a subset of a subset. Curiously, it’s also a group that may not necessarily be actively using cannabis—precisely because they’re in a treatment regime.

It’s also worth noting that a large percentage of referrals to substance use disorder treatment originate from legal entanglements. From a study on the subject:

Referrals to specialty SUD treatment from the legal system made up 24.2% of all 8.3 M treatment episodes between 2015–2019, with only the self/family/friend category of referrals making up a larger proportion (36.4%; Fig. 1) The number and proportion of referrals from the legal system to SUD treatment declined by more than 25,000 episodes or 4.4 percentage points (26.2 vs. 21.8%; p < 0.001) between 2015–2019…

Cannabis was implicated in almost half of all referrals from legal sources but only a third of non-legal sources (45.5 vs. 33.1%; p < 0.001).

It’s difficult to know how that situation might have played out in the gathering of the data, but the sample—already not even close to a random assemblage—may be even more biased than is already plain to see. The numbers indicate that of the 8.3 million total SUD referrals between the years 2015-2019, at least 2 million involved legal referral, with approximately 1 million referred specifically for Cannabis Use Disorder.

There’s another factor to be considered, in regard to a Cannabis Use Disorder diagnosis: of all the forbidden substances subject to identification on the basis of a urine drug screen, cannabis is by far the most frequently detected. Not only is it the most commonly used illicit substance, the fat soluble nature of THC metabolite means that it’s also the one that stays detectable for a long period of time—in stark contrast to the other substances, which can be washed out of the body by drinking enough water and taking mild diuretics like B and C vitamins. What this fact implies is that even patients subject to random drug tests are often able to pass clean urine with a couple of days advance notice. So those substances are less likely to be detected, and unless their use is habitual, it may not be detected in a drug screen.

The JACC study, again:

A total of 4,636,628 relatively healthy adults aged ≤50 years: 93,267 (2.01%) cannabis-users and 4,543,361 (97.99%) non-users.

What we have here is "a “Big Data” study—or anyway one with the pretense of relying on Big Data. It isn’t a clinical study, using identified patient case histories, with all the details those imply.

Cannabis-users were older (26 ± 8 vs 21 ± 9.5 years, P < 0.0001) and had higher comorbidities, including a nearly 15-fold higher prevalence of DD (30.63% vs 1.88%, P < 0.01) and BMI >30 (18.72% vs 3.25%, P < 0.0001). After PSM, each group had 89,776 patients with balanced demographics and baseline health characteristics.

“DD” = depressive disorder. “BMI” = body mass index. “PSM” is propensity score matching, a weighting score intended to select a group from the nonuser cohort that’s suitably matched with the CUD group. The group that’s already estimated as having a >30% prevalence of depressive disorder, and almost 19% prevalence of stage 1 obesity (at minimum.) So if I’m reading the methodology right, the end result of PSM resulted in a comparison to be drawn from two equally age matched, equally unhealthy* cohorts of equal numerical size: 89,776.

There’s no way I’m going to refer to any group with such a high prevalence of diagnosed clinical depression as “healthy.” Or “average”, or “normal.” As the study points out, that’s 15 times the average of DD incidence in the general population. So that’s the matched baseline for the two samples of roughly 90,000 that were compared against each other through years 2010-2018.

A note on Methodology: that comparison was not a longitudinal study, along the lines of the long-term clinical studies done by pulmonologist Donald Tashkin with in-person assessments of the participants active in the study. The comparison in this study is retrospective, and also entirely statistical, based on harvesting metadata from a huge number of anonymized patient records. To revisit:

Data Source: This retrospective cohort study utilized the TriNetX health research network, which aggregates de-identified electronic medical records from healthcare organizations worldwide. The specific network used for this study was Research, which contains data from 53 healthcare organizations (HCO) throughout the United States. Given the nature of the study and included date, IRB approval was waived.

This is the Trinex Health Network. https://trinetx.com/solutions/live-platform/#s_1

From the home page:

Our web-based platform, TriNetX LIVE™, puts you at the center of our real-world data and evidence ecosystem. With an interface that’s as powerful as it is easy to use, you can build and analyze cohorts drawn from around the world in just minutes. When you’re ready to invite these patients into a study, connect with their healthcare providers with a few clicks.

The researchers note that they didn’t actually look up any patients and invite them into an experimental study. That’s what the phrase “the TriNetX health research network…aggregates de-identified electronic medical records” means.

Instead, they harnessed the awesome! power! of “Big Data”! to formulate their researches and inform their conclusions.

Statistical analysis: The analysis was performed with TriNetX Live built-in analytics.

This “research study” was built entirely out of keyword searches of metadata, crunched with a statistics program. Studies of this sort have become enormously popular in all sorts of fields, most notably in fields like evolutionary biology, social psychology, and epidemiology. Because Now We Have The Technology. But—do we really have Big Data? I’d allow that sometimes we do. But not this study.

The authors say as much, in their closing remarks. In so many words, at least:

This study has limitations due to lack of detailed cannabis consumption data and potential misclassification. The inherent limitations of real-world data often result from inconsistent patient reporting in electronic medical records. Future research should investigate the dose-response relationship, and the effects of synthetic cannabinoids.

The bland phrase “potential misclassification” means that the data might be inaccurate, to some unknown extent.

The sentence “The inherent limitations of real-world data often result from inconsistent patient reporting in electronic medical records” discreetly implies what many primary care physicians already know, even though the conversation is still mostly carried out informally, without the complications of bringing it up in official channels: many people lie to their physicians about their illicit substance habits. Including the illicit substance to which 80% of surveyed illicit drugs users confine their consumption: Cannabis. Weed. Mary Jane. People lie about their healthy habits, too. But they really lie about weed consumption. Because if Insurance Company Big Brother finds out, Shit invariably gets more Complicated. It’s a sleeping dog that just about everyone would prefer to let lie, unless they’re forced into admitting it.

I realize that post-hipster multimillionaires get to say anything they want to their personal private practice physicians; some of them may even party together. But those of us in the rest of the world get processed through gigantic actuarial mainframes of HBOs and insurance plans and programs, and we don’t have that luxury. Just saying.

For that reason alone—and there are other possible confounders to be brought up, as unanswered questions that might be that quick and dirty keyword search—this is not a probative study. It arguably isn’t even a serious study. It’s a preliminary skim that’s easily performed, hardly more complicated than a set of advanced keyword searches. And this is what I can’t stand about so many of these statistical metastudies: they’re reported as authoritative and conclusive, until you notice how threadbare they are. Then they begin to resemble a product that’s similar to a collateralized debt obligation: laundered junk data presented as AA certified.

That does not mean that cannabis has been cleared of all suspicion as far as the possibility of presenting increased cardiac risks. But I need to note a disturbing tendency that I’ve noticed in recent years, in regard to medical research on cannabis. There’s been a massive cascade of it in the aftermath of Colorado’s legalization of the commercial market. And at least some of it appears to me to have an axe-grinding agenda, to put cannabis in the worst possible light. I’m absolutely on board with following every research trail impartially. But speaking as someone who’s tried to keep up with the reading for the past 50 years, my hunch is that Politics and Law Enforcement continue to have a thumb on the scale. Where it doesn’t belong. And there’s no way to really get truthful assessments until they cut it out. I notice. I notice who these studies get funded, and who features them most prominently and uncritically on their web pages, and how they’re reported in the press. It isn’t just cannabis, of course. The manufacture of bogus fake “knowledge” by Power is all about who gets their first with the biggest headline banners. I detest that game, and all I can hope to do is reverse the process to what it needs to be, the only way with integrity: so that knowledge generates power.

Back 50, 60 years ago, research studies on the effects of cannabis were relatively scarce, at least those published in English. There were no sweeping surveys and estimates of hundreds of millions of users worldwide, either. But the studies that did exist were often quite comprehensive, such as the 1975 book by medical anthropologist Vera Rubin, Ganja In Jamaica. and the 1980 book edited by William E. Carter, Cannabis In Costa Rica. and the extensive writings of Tod Mikurya, MD, an oeuvre documented on this National Institutes of Health web page https://circulatingnow.nlm.nih.gov/2020/02/25/tod-mikuriya-papers-now-available-for-research/ (Access that link while you still can. It may not be available tomorrow.)

I’ll also note that while recent studies of cannabis report that using it increases blood pressure, the studies I’ve read from earlier than the 1990s were inclined to agree that cannabis use lowered blood pressure by 10-15 points, along with slightly lowering body temperature by about 1 degree.

I can’t help thinking that Big Data meta-analysis lacks the capability to offset that raging epidemic, Generational Amnesia—at least when it comes to the topic of Marijuana, for some reason. A textbook case of how the Sociology of Knowledge—the Politics of Knowledge—can influence scientific research with its taboos and dogmas, its profit motives and private ends. Sometimes shutting it down, other times warping it. Especially when Criminalization gets involved, and so much of the free exchange of knowledge is driven into the shadows.

There's a history of exaggeration--or confabulation--of the dangers of forbidden substances, with spurious studies purporting to rely on the protocols associated with the "modern medical research" of the 20th and 21st century to support their legal prohibition.

I've recently been reading reports from the 1800s by physicians who examined the effects of alcohol, tobacco, cannabis, opiates, and cocaine in the era before clinical studies were done on patients. Their writings were based entirely on empirical observations of users, and their opinions contained at least as much insight as the findings of clinical studies in the 20th and 21 centuries. Their scrutiny was typically dispassionate, and their conclusions about behavioral and public health impacts were not much different than modern medical consensus. Drug addiction was recognized as a liability, but rarely one of catastrophic proportions. The substance most widely viewed by physicians as a public health problem was alcohol. Overwhelmingly so.

It's worth noting that 19th century American and European physicians had relatively little to say about cannabis use and effects. Tincture of Cannabis indica--"hash oil"--was part of the medical apothecary and freely available both by prescription or over the counter for general purchase. But unlike alcohol, tobacco, opium, and the newly synthesized opiates and cocaine, I’ve found no reports of long-term cannabis habituation dating from the 1800s. "Recreational use" in 19th century Europe and the US consisted almost entirely of experiments by a self-selecting group of artistic and literary types. It just wasn't an issue. Even chloral hydrate and ether were more common intoxicants than cannabis. So the only 19th century medical reports of chronic cannabis use and habituation came from India--most notably the 1896 Chopra report from British colonial India, drawn up by a panel of British administrators and Western-aligned Indian Brahmins. I found the Chopra report conclusions to be a bit exaggerated in some respects, and open to criticism about some aspects of the methodology. But I found the basic conclusion to have merit: that cannabis preparations are used in various strengths (bhang, ganja, and charas or hashish), and that continuous heavy use of the most potent concentrated product, charas (hashish,) sometimes produced deleterious mental effects and triggered psychosis. It's now commonly known that serious mental breaks often occur in the teens and early 20s, and that extra levels of induced stress can induce the first onset, including as the result of substance abuse. There have been clinical observational studies that also show direct correlations with alcohol, tobacco, and coffee ingestion. I view cannabis as a nonspecific sensitization agent and amplifier of perceptions and cognition; that increased sensitivity can be experienced as disorienting or dysphoric, especially in people predisposed to mental fragility. As such, the experience probably is more liable to induce feelings of anxiety, panic, or paranoia than alcohol, nicotine, or caffeine, which are experienced by most (although not all) users as stimulus blockers. And emotional states like anxiety, panic, and paranoia are liable to be internally boosted by mentally unstable temperaments, sometimes to the extent of inducing acute psychosis. The exterior setting often plays a rule as crucial as the interior mental set; some settings are more stressful than others. In the general sense, the social Zeitgeist of the year 2025 is inherently more stressful than that of 1975. Indisputably.

To return to the topic of the study's claim of possible increased risk of heart attacks and strokes from cannabis use, I think that the research should have been prefaced by an epidemiological survey of incidence of heart attack and stroke in the US over the previous 75 years or so--because cannabis use skyrocketed between 1965 and 1975, and remained relatively common throughout the 1980s and 1990s and into the 21st century. If there's a pronounced direct correlation between cannabis use and deleterious physical effects like heart attack, stroke, and early mortality, its additive impacts should logically have shown up long ago as a statistical trend. I would expect that physicians would have noted such an association in the course of their practice, too—except that the use of cannabis, a forbidden and taboo substance, has most often been concealed from medical practitioners by almost the entire user population. The situation is further complicated by the contrast with the case of other forbidden substances, where users may be similarly inclined to conceal their use but are betrayed by the pronounced negative consequences of dysfunction and addiction. It’s just an empirical fact that even daily cannabis use doesn’t lead to nearly the same incidence of blatant dysfunction found with other other drug habits, so it’s more easily overlooked by physicians. So it’s possible that incidents like myocardial infarction and ischemic stroke in patients under age 50 might not have been subjected to scrutiny in regard to the possible role played by cannabis use.

Several other behavioral health trends have also markedly increased within the span of the previous 60 years, notably the increase in overweight and obesity, and the increasing amount of street drug abuse of all sorts, not just cannabis. But a rigorous study would have begun by reviewing overall epidemiological trends in the under-50 population for shifts from year to year, and decade to decade, before proposing a 600% increased risk of heart attack and stroke associated with cannabis use in people under age 50. The percentage of regular cannabis users in the US population is quite large; it’s a cohort sufficiently numerous to have shifted the wider epidemiological trend significantly in the wider population. Without a baseline depicting an increasing frequency of heart attacks and strokes in the under-50 age demographic over time within the general population, it’s difficult to provide a rationale for the recent JACC study in the first place.

---

Edit for additional review commentary, 03/23/2025:

The JACC study also seems to me to present a problem in regard to its choice of references, particularly Endnote 4, which is used to support this claim:

This analysis provides evidence linking cannabis-use to adverse cardiovascular events, including MI, ischemic stroke, HF and mortality. Notably, cannabis use appears to pose a substantial and independent risk for these outcomes, even in a population without traditional cardiovascular risk factors. These findings suggest cannabis as a novel and underrecognized risk factor for cardiovascular diseases. Our study results are consistent with prior research documenting acute coronary syndrome following cannabis use (4).

My preliminary keyword search was not able to locate the full text of the linked study referenced in endnote 4, which was titled Ramphul K, Kumar N, Dhaliwal JS et al. Prevalence of cannabis consumption among acute myocardial infarction patients across various age categories over a two-decade span in the United States. Curr Probl Cardiol 2024;49:102555

The search was able to find the first page of the study, with a reference to the survey data used in the research:

An estimated 20,041,408 AMI cases were studied and overall, the use of cannabis rose from 0.20 % in 2001-2005 to reach 1.76 % in2016-2020 (ptrend <0.01) (Table 1). The prevalence of cannabis use among AMI cases of patients <18 years fluctuated rising from 5.67 % in 2001-2005 to 7.41 % in 2006-2010. However, a drop was seen in 2011-2015 as it reached 5.26 %, followed by a surge in 2016-2020 to 8.09 %. This trend was not statistically significant (ptrend =0.19). Patients of ages 18-34 had the highest overall prevalence of cannabis with 9.44 %, and their use constantly increased from 5.23 % in 2001-2005 to 12.49 % in 2006-2020 (ptrend<0.01). Similar sharp rise was noted in patients aged 34-49 (0.98 % in 2001-2005 to 6.09 % in 2006-2020, ptrend<0.01), 50-60 (0.28 % in 2001-2005 to 3.42 % in 2006-2020, ptrend<0.01), and >60(0.01 % in 2001-2005 to 0.68 % in 2006-2020, ptrend<0.01)

That paper appears to have focused entirely on the changing percentages of cannabis-using myocardial infaction patients between the years 2000-2020, the data showing a marked increase in the percentage of identified cannabis users. (Note the ambitious size of the study: over 20 million cases in all age groups, compiled over 20 years.)

The results of my keyword search also turned up a related paper by the same authors published in the same month as the Current Problems in Cardiology article, June 2024, appearing in the Archives of Medical Science for Atherosclerotic Diseases. The full text of the study is available, and linked here https://amsad.termedia.pl/Unmasking-the-cannabis-paradox-in-hospital-outcomes-of-cannabis-users-admitted-with,189731,0,2.html

This study also reviewed a massive number of cases- nearly 10 million. The AMS study also goes into considerable detail in investigating data correlations between cannabis use and heart attacks and ischemic stroke, with detailed tables of possible confounding factors in order to produce a properly weighted comparative result. The co-morbidities examined are measured slightly differently, but most of them overlap. The JACC study includes factors like pregnancy and history of cancer, and the AMS study does not. Perhaps the most significant difference is that the JACC study makes no mention of screening for alcohol abuse.

Crucially, unlike the JACC data set, the cannabis using cohort does not consist exclusively of those identified in medical records as being diagnosed with Cannabis Use Disorder.

The abstract of the June 2024 AMS study:

Introduction:
Cannabis is increasingly becoming a socially acceptable substance, with multiple countries having legalised its consumption. Epidemiological studies have demonstrated an association between cannabis use and an increased risk of developing coronary artery disease. However, there is a lack of studies about the influence of cannabis consumption on the outcomes following acute myocardial infarction (AMI).

Material and methods:
We retrospectively analysed hospitalised patients with a primary diagnosis of AMI from the 2001 to 2020 National Inpatient Sample (NIS). Pearson’s X² tests were applied to categorical variables, and t-tests for continuous variables. We conducted a 1:1 propensity score matching (PSM). Multivariate regression models were deployed on the PSM sample to estimate the differences in several events and all-cause mortality.

Results:
A total of 9,930,007 AMI patients were studied, of whom 117,641 (1.2%) reported cannabis use. Cannabis users had lower odds of atrial fibrillation (aOR = 0.902, p < 0.01), ventricular fibrillation (aOR = 0.919, p < 0.01), cardiogenic shock (aOR = 0.730, p < 0.01), acute ischaemic stroke (aOR = 0.825, p < 0.01), cardiac arrest (aOR = 0.936, p = 0.010), undergoing PCI (aOR = 0.826, p < 0.01), using IABP (aOR = 0.835, p < 0.01), and all-cause mortality (aOR = 0.640, p < 0.01), but with higher odds of supraventricular tachycardia (aOR = 1.104, p < 0.01), ventricular tachycardia (aOR = 1.054, p < 0.01), CABG use (aOR = 1.040, p = 0.010), and acute kidney injury (aOR = 1.103, p < 0.01).

Conclusions:
Among patients aged 18–80 years admitted to hospital with AMI between 2001 and 2020 in the United States, cannabis use was associated with lower risks of cardiogenic shock, acute ischaemic stroke, cardiac arrest, PCI use, and in-hospital mortality.

The last passage was set in bold type by myself. The words are those of the study authors: lower risks of cardiogenic shock, acute ischaemic stroke, cardiac arrest, PCI use, and in-hospital mortality.

I have to wonder why the authors of the JACC study did not include this study as one of their references. Granted, it undercuts their own claims entirely.

---