Eleanor Crane: Ben, I need to tell you something embarrassing — I genuinely hesitated this morning reaching for my second coffee.
Ben Okonkwo: Because of the headline.
Eleanor Crane: Eleven times more cardiac arrest reports. It just sat there in my head. And I know enough to know that's not the same as eleven times the risk — but knowing that didn't fully stop the hesitation.
Ben Okonkwo: That response is actually — hm, that's the most honest illustration of the problem I could ask for. Because the number is real. It comes from a preliminary analysis of twenty-eight thousand plus FDA adverse-event reports from 2017 through 2025, presented at the American Heart Association Scientific Sessions 2026. Cardiac arrest appeared roughly eleven times more often in reports tied to caffeine or stimulant products versus controls. ScienceDaily covered it. And now it's in your morning routine.
Eleanor Crane: And what I couldn't answer standing in my kitchen is whether that number belongs to me or to a twenty-six-year-old stacking energy drinks on top of a pre-workout before a gym session.
Ben Okonkwo: That's — yeah, that's exactly the fracture. The statistic lives in an FDA adverse-event reporting database. It doesn't know your dose. It doesn't know that guy's pre-existing conditions. Gregory Marcus, who literally chairs the AHA's 2026 scientific statement on caffeine and cardiovascular disease, was careful to say this does not prove that caffeine products cause cardiac arrest. The eleven-fold figure is a signal about how often something shows up in a voluntary database, not about how often it shows up in a body.
Eleanor Crane: So what does it actually tell us?
Ben Okonkwo: It tells us something worth investigating. Which is different from something worth fearing — but that distinction doesn't survive a phone notification.
Eleanor Crane: But I think that's the part most people miss — when they hear 'eleven times more,' they hear risk. And the actual machinery here is something different. It's not counting bodies. It's counting complaints.
Ben Okonkwo: Exactly — and the method has a name: reporting-odds ratio, or disproportionality analysis. The core idea is this — it asks how often a word like 'cardiac arrest' appears in reports about one product category versus a control group. That's the whole thing. Not how often cardiac arrest happens to actual consumers.
Eleanor Crane: Think of it like a complaint box at a restaurant. If everyone who had a terrible night writes in, the box fills up — but you have no idea how many people ate there and had a perfectly fine meal.
Ben Okonkwo: Oh — that's actually it. That's the whole structure of the database.
Eleanor Crane: And the box for energy drinks filled up faster. That's what the eleven-fold number is measuring.
Ben Okonkwo: Right — and the FDA adverse-event reporting system is voluntary, so the box itself is — I mean, it's self-selected by design. If a product gets coverage, more people write in. The complaints spike before the risk does.
Eleanor Crane: Which is why Gregory Marcus was so precise about this. He chairs the AHA's 2026 scientific statement on caffeine — and even he said it 'raises concerns' but 'does not prove that caffeine products cause cardiac arrest.' That language is doing a lot of work.
Ben Okonkwo: It is. And the other numbers matter too — heart attacks showed up roughly 4.5 times more often, ventricular events about four times more often. So it's not one anomalous spike. There's a pattern across cardiovascular categories. That's what makes it a signal worth taking seriously, even without causation.
Eleanor Crane: So the complaint box isn't random noise. It's patterned noise — and that's the thing that changes what we do next.
Ben Okonkwo: Patterned noise — that pattern might not be about what the drinks do. It might be about who drinks them. And those are very different claims.
Eleanor Crane: Meaning the eighteen-to-thirty-four signal.
Ben Okonkwo: Right. That's the sharpest number in the whole analysis — cardiac arrest reports skewing hardest toward people who statistically almost never have cardiac arrests. And the instinct is to say: pharmacology. Stimulants. Young heart. But — actually, wait — the cardiologist who pushed back hardest on this, posting on X, afshineemrani, eighteen-thousand-plus patients, called it a safety signal, not a verdict, specifically because we have zero confounder data. No information on pre-existing arrhythmias, no alcohol co-use, no sleep deprivation, no exercise pattern. Those aren't minor omissions.
Eleanor Crane: So the eighteen-to-thirty-four clustering — it might just be describing who gets marketed to, not who's at pharmacological risk.
Ben Okonkwo: That's exactly the gap. Think of a college junior cramming at two a.m. — she's had four hours of sleep, shared a beer with a friend at dinner, then opens a pre-workout supplement with over twelve times the caffeine concentration of a standard coffee. That's not an energy drink. The study bundled those products together. We can't tell whether the signal belongs to her supplement or a canned drink with eighty milligrams of caffeine.
Eleanor Crane: Twelve times. That number is — I hadn't clocked the gap being that large.
Ben Okonkwo: Some pre-workout supplements exceed twelve times standard coffee concentrations per ounce. And the AHA's own 2026 scientific statement says four hundred milligrams daily is generally considered safe for most adults. But if you're not reading labels — and most people aren't — dose becomes invisible. The bundling collapsed exactly the variable that matters most.
Eleanor Crane: And that's before we ask whether the EFSA warning labels — they recommended those for children and pregnant women — whether any of that dose context is reaching the person actually buying the product.
Ben Okonkwo: It mostly isn't. And the part that makes all of this more tangled — we'll get to it — is that media coverage of a database study like this one can actually inflate the database it's describing. Which is — I mean, that's a feedback loop with real regulatory consequences.
Eleanor Crane: And that loop — the coverage inflating the database it's describing — that's the part I can't stop pulling on. Because the ScienceDaily piece runs, someone reads it Saturday morning, they had a Red Bull two weeks ago and a weird heartbeat, and they file a report. That report goes into the FDA adverse-event system. And next year's analysis finds a larger signal.
Ben Okonkwo: The study and the coverage of it are not independent systems. That's the thing.
Eleanor Crane: So what does a regulator actually do with that? Because EFSA has already moved — warning labels for children and pregnant women. That benchmark exists. The U.S. has nothing comparable. And now we have a preliminary, not-yet-peer-reviewed presentation from the American Heart Association's 2026 Scientific Sessions that's already reshaping public perception of a legal product.
Ben Okonkwo: And some countries have gone further than labels — bans, age-restricted sales. So globally it's a patchwork. No consistent standard for when a signal becomes sufficient.
Eleanor Crane: Which brings the actual unanswerable question into focus — what has to be true before you act? A voluntary database signal, or controlled data establishing causation and dose-response?
Ben Okonkwo: And those are not close to equivalent evidentiary standards. I mean — a reporting-odds ratio from a self-selected database versus a randomized trial tracking actual exposure. The gap between those is enormous.
Eleanor Crane: But the recursive part won't wait for the trial. Picture a sixteen-year-old in Portugal — a country with age restrictions — posting a TikTok about the AHA findings. His American cousin reads it, files a report about chest tightness after a workout. That report enters the system. Perfectly legitimately.
Ben Okonkwo: Legitimate report, inflated signal, future study cites the growth. And nobody lied anywhere in that chain.
Eleanor Crane: That's — I think that's the thing that actually unsettles me. The system isn't broken. It's working exactly as designed. And it still can't tell us what we need to know.
Ben Okonkwo: And that guy at the convenience store cooler keeps coming back to me. Twenty-six years old. He read the headline — Marcus said the study 'raises concerns,' not a verdict, not an exoneration, just concerns — and now he's standing there with a can in his hand and no idea whether eleven times means eleven times more likely today, or eleven times more likely to show up in a database he'll never read.
Eleanor Crane: And that gap — between the signal and what it means for him specifically — that's not a failure of science. It's just where science actually lives right now.
Ben Okonkwo: He buys it or he doesn't. And either way, the data can't tell him if he was right.