Cannabis "6x heart-attack risk" hits HN — and the guy who read the study defends it both ways
The scary number draws the predictable wave of "they didn't control for tobacco." Then someone who actually opened the cohort paper points out it did — and, separately, that the odds ratio is no bigger than simply being low-income.
An American College of Cardiology release pairs a 4.6-million-patient cohort (≈6x heart-attack risk in healthy under-50s) with a meta-analysis of 75M+ people showing a milder 1.5x. Researchers admit the studies are retrospective and can't fully control dose, route, or other drugs.
The thread's best correction is contrarian in both directions: yes it controlled for tobacco, no the effect size isn't remarkable next to income, and nobody split smoking from edibles.
Two studies, two very different numbers
The ACC press release bundles two studies. A retrospective cohort of 4.6 million TriNetX records — under-50, cardiovascularly healthy at baseline — found roughly 6x higher heart-attack risk, 4x ischemic stroke and 2x heart failure over about three years. A separate meta-analysis of 12 studies and 75M+ people (average age 41) found a far more modest 1.5x heart-attack risk, with 7 of 12 studies showing significant associations. The researchers themselves cautioned the work is retrospective and couldn't account for cannabis dose and duration, tobacco, or other drug use; the lead author suggested clinicians ask about cannabis the way they ask about cigarettes.
The 223-comment thread reflexively reached for the standard confounder — and then someone who'd read the actual paper complicated everyone.
The correction that cuts both ways
The crowd's opening move was "tobacco invalidates this." A reader who opened the cohort study pushed back on the crowd:
"Agree on ROA, but the retrospective cohort study (one of two discussed in this article) did account for many other factors including tobacco use." — a commenter who read the underlying cohort study
But the same reading offered a deflation the headline never mentions — scale:
"the odds-ratio reported for cannabis use (4-6) is not so different from the odds-ratio associated when studying high vs. low income cohorts (3-5) … it's reporting a trend present in ~5 million electronic health records" — a commenter reasoning about odds ratios and EHR data
And the genuinely unaddressed gap — route of administration — got named bluntly by a reader who flagged the post:
"Literally no mention of ROA. It matters whether they smoke plant matter, vape, use it orally, etc." — a reader who flagged the submission on methodology
The honest takeaway
The "6x" is real as a correlation and shaky as a scare. One of the two studies did control for tobacco, so the reflexive rebuttal is wrong; but the effect sits in the same range as being poor, no one separated a joint from a gummy, and it's all retrospective. Ask your doctor about cannabis the way you'd ask about cigarettes — and treat a single relative-risk number, in either direction, as the start of a question, not the answer.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
Jim Keller's startup wants to mass-produce chip fabs from Texas — a semiconductor expert calls it smoke and mirrors
Atomic Semi rebrands as Fab2 and moves to Texas with a wild pitch: a factory that builds small semiconductor factories. In the thread, a self-described semiconductor expert reads the same news and sees another wunderkind startup that dissolves on contact with reality.