A weekly read on the gut, the brain, and the wiring in between. One clinical pearl, one paper, one product, one person. No hype, no selling, just the stuff I actually think about.
1. Practice
The line that stopped being a measurement and became a diagnosis.
In 1569 Gerardus Mercator published a map that solved a problem sailors had been losing to for a century. Plot a course on it and a line of constant compass bearing comes out straight. Set the heading, hold it, arrive.
The bargain was area. His projection stretches the world as it climbs away from the equator, and at the top it is grotesque. Greenland looks about the size of Africa. Africa is fourteen times larger. Mercator knew this. He was solving a navigation problem, he solved it completely, and four and a half centuries later his map still hangs in classrooms where nobody is plotting a course.
Something close to that happened to the ECG.
Nobody explains this part to patients. When you arrive with chest pain, the tracing gets measured against a threshold: one millimeter of ST elevation in two neighboring leads. Clear it and you are a STEMI, the alarm goes off, and a team gets called in from home to open the artery now. Miss it and you are an NSTEMI, which is treated as the serious-but-not-emergent kind, and you get admitted, medicated, and scheduled.
That millimeter was not handed down from biology. It was drawn in the 1980s to decide who should receive a clot-dissolving drug that caused bleeding into the brain in close to one percent of the people who got it. With a risk like that, the diagnosis had to be near-certain before anyone reached for the syringe. The threshold was set high on purpose, and for that question it was the right answer.
Then we hung it on the wall. STEMI and NSTEMI stopped describing a tracing and started describing an artery.
The artery did not agree.
Researchers pooled 25 studies covering 60,898 patients who carried an NSTEMI diagnosis and looked at what the catheterization actually found. Thirty-four percent had a completely blocked culprit artery, and those patients had roughly 70 percent higher odds of dying and 66 percent higher odds of going into cardiogenic shock than the NSTEMI patients whose arteries were open (Hung et al., Crit Care 2018;22:34). One in three patients we file under not-blocked is blocked.
It replicates. An independent cohort of 1,128 NSTEMI patients in China found complete occlusion in 25.4 percent, and over 3.6 years those patients ran a 46 percent higher risk of death, another heart attack, stroke, or unplanned repeat procedure, after adjustment (Am J Cardiol 2025, published November).
Then there is the part that should be hard to read. In a study of 808 patients with suspected heart attack, the ones whose arteries were blocked but whose ECG never crossed the line had the same infarct size and the same mortality as the ones whose ECG did cross it. What they did not have was the same speed. They waited longer to get to the catheterization lab (Meyers et al., Int J Cardiol Heart Vasc 2021;33:100767). Same damage, same risk of dying, slower response, and the only thing separating the two groups was whether the injury happened to register above an arbitrary line.
The same paper showed the information was on the tracing the whole time. Blinded readers trained to look for occlusion patterns rather than millimeters caught 86 percent of the blocked arteries. The millimeter criteria caught 41 percent.
Hospitals measure door-to-balloon time. It is reported, benchmarked, and enforced, and it is the reason STEMI care in this country got dramatically faster over twenty years. There is no clock at all for the patient whose artery is shut and whose ECG stayed under the line. That delay produces no dashboard entry and no missed-metric report. To every quality system in the building, it looks like an appropriately managed NSTEMI.
The harm is not hidden. It is unmeasured, which is a far more durable condition, and it is how this survived two decades of quality machinery aimed directly at chest pain.
What to do with this.
"It's not a STEMI" does not mean "your artery is open." It means the tracing did not clear a threshold. Those are different sentences, and the second one is the one you want answered.
Ongoing pain is information, and it expires. Say it out loud, say it again when it does not stop, and ask for a repeat tracing. A single ECG is a still photograph of a moving process. If you or someone you are with is still hurting, that is worth saying every time somebody new walks in.
There is a question you can ask, and it is a fair one. Is anyone considering that the artery could be completely blocked even though the ECG doesn't meet criteria? In 2025, the American College of Cardiology, the American Heart Association, and the American College of Emergency Physicians jointly published a guideline naming several ECG patterns that call for emergency artery-opening even though they never reach the millimeter (Rao et al., Circulation 2025;151:e771-e862). The obligation has changed.
And one thing in fairness. The objection from the cath lab is legitimate. Every unnecessary activation means contrast, radiation, and a team driving in at three in the morning for something that turns out not to be a heart attack. The reading that caught 86 percent of blocked arteries gave up a few points of accuracy in the other direction to do it, and at national volume that is not free. The criteria were built to guarantee a floor, so that no reasonable cardiologist would argue with the call. Treating a floor as the ceiling is the error, and it belongs to those of us reading the tracing, not to the people answering the phone.
2. Publication
One paper worth reading, in three sentences.
Researchers searched three databases without language restriction, pulled every study of adults diagnosed with NSTEMI that reported what the angiogram showed, and pooled 25 studies and 60,898 patients to ask a question the diagnostic category had made it easy not to ask: how many of these people had a closed artery (Hung CS, Chen YH, Huang CC, et al., Crit Care 2018;22:34, doi 10.1186/s13054-018-1944-x). The answer was 17,212 of them, 34 percent, with a 95 percent confidence interval of 30 to 37, most often in the left circumflex, the vessel whose territory the standard twelve-lead ECG sees worst. Hold it honestly, because the authors do: these are observational cohorts rather than randomized trials, the individual studies varied in how they defined a culprit lesion, and the paper's own conclusion stops at whether these patients should be treated more aggressively "warrants further study" rather than claiming the case is closed. What makes it my pick is that the finding was always available. Nobody needed a new technology or a new trial. Somebody had to ask a question the category discouraged.
3. Product
This week's pick: a blood pressure cuff that has actually been tested. Unsponsored, no affiliate relationship, no brand, and obviously not a BellyMD product.
This issue is about a measurement determining a decision. The measurement you control at home is your blood pressure, and most of the devices sold to consumers have never been through a formal accuracy validation protocol. Not failed one. Never taken one.
There is a free public list. The American Medical Association runs the US Blood Pressure Validated Device Listing at validatebp.org, reviewed by an independent committee of physician experts. Search the model before you buy it. If it is not on the list, it has not been checked.
Use an upper-arm cuff. Wrist devices are not recommended for routine measurement except when an upper-arm cuff will not fit or cannot be used (Picone DS, Padwal R, Stergiou GS, et al., J Hum Hypertens 2022;37:108-114). And cuffless wearables, including the watch on your wrist right now, sit outside most validation programs entirely.
4. Person
One person worth following in this space.
Stephen W. Smith, MD, emergency physician at Hennepin County Medical Center in Minneapolis, and the author of Dr. Smith's ECG Blog.
He has spent well over a decade posting real tracings from real shifts, with the angiogram result at the bottom, for free, in public, to anybody who wants to get better at this. No paywall and no institutional imprimatur. He also put his name on the revised criteria for reading a heart attack through a left bundle branch block, which is the pattern that used to be treated as unreadable (Smith et al., Ann Emerg Med 2012;60:766-776).
What I want you to notice is the direction the change traveled. The occlusion argument did not come down from a guideline committee and get distributed to the bedside. It came up from clinicians reading tracings, publishing what they found, arguing about it in public for fifteen years, and eventually pulling the guideline toward them. The 2025 document is downstream of that, not upstream.
Medicine does not usually work that way. It is worth watching when it does.
Every specialty has a Mercator projection. Somewhere in its history, somebody drew a line for a specific and reasonable purpose, and everybody after them inherited the line without inheriting the purpose.
Mine is the one I work on now. The conditions I spend my time on, the disorders of gut-brain interaction, are diagnosed by symptom criteria that were built substantially so that researchers in different countries could agree on who belonged in which study. That was a real problem and the criteria solved it. Then they migrated into clinics, and a threshold designed to make studies comparable started being used to decide whether a person's suffering counted. You have met the result if you have ever been told your workup was normal, as though normal and fine were the same word.
That is why we built MeNome, our free tracking app at menome.belly-md.com. It does not replace a diagnosis. It keeps the part of you the threshold was never designed to capture.
The artery does not care what we call it.
That's the four. See you next week.
Rick
From BellyMD
I build MGB+, a small supplement line for gut-brain support: Clear, Cool, and Calm, each a magnesium and B-vitamin based formula meant to support the gut-brain axis as part of a broader routine. Structure-function support, not a treatment for any disease. If that is relevant to you, it lives at belly-md.com.
If you want to talk through your own gut-brain picture with me directly, I offer a gut-brain consultation. Details at belly-md.com.
Not medical advice. This newsletter is for education and does not create a physician-patient relationship. Nothing here is a diagnosis or a prescription. If you are having chest pain, call 911. Talk to your own physician before starting, stopping, or combining any medication or supplement, including over-the-counter ones.