An international research team has built a blood test that can tell “broken heart syndrome” apart from a real heart attack before a patient is taken for an invasive artery scan. The study, published in the European Heart Journal on August 31, 2026, analyzed 3,615 patients from two overlapping registries and found that the new five-marker BioTAK score sorted almost 90% of them into a likely diagnosis before invasive coronary angiography.
Key takeaways
- The BioTAK score uses four blood markers plus a patient’s sex to separate Takotsubo syndrome from a heart attack.
- In an independent group of 1,792 patients, the score was correct in about 93 cases out of 100 when asked to pick which of two patients had Takotsubo (AUC 0.93).
- Only 146 of the 3,615 patients in the study actually had Takotsubo syndrome, so the score needs testing in real emergency departments before it changes care.
What the study found
Takotsubo syndrome is a sudden heart condition, usually triggered by intense emotional or physical stress, that most often hits postmenopausal women. According to the study authors, it looks almost exactly like a heart attack at first: chest pain, an abnormal electrocardiogram, and rising cardiac enzymes. The difference is that the coronary arteries are not blocked.
That is why the diagnosis is usually made backwards. As the researchers write, identifying it currently relies on history, repeated cardiac imaging, and “exclusion of a culprit epicardial coronary artery lesion, generally with invasive coronary angiography.” In other words, doctors thread a catheter into the heart, find clean arteries, and only then name the condition.
The researchers measured cardiovascular blood markers before angiography in every patient. Machine learning picked the most useful combination, and the final score came down to five items: NT-proBNP (a marker of strain on the heart muscle), PAM (tied to blood vessel constriction and anxiety regulation), soluble LOX-1 (tied to unstable artery plaque), LDL cholesterol, and sex.
The score was built in 1,823 patients and then tested in a separate group of 1,792. Its accuracy, measured as area under the curve, was 0.97 in the development group and 0.93 in validation. A score of 0.5 is a coin flip and 1.0 is perfect, so 0.93 means that given one Takotsubo patient and one heart attack patient, the score would rank them correctly about 93 times out of 100. The statistical range around the validation figure runs from 0.90 to 0.96, so the true accuracy is very likely somewhere in that band.
Dr. Kumar’s take
Two things get corrected here at once. Broken heart syndrome is not a poetic curiosity, it is an acute cardiac emergency that puts people in a hospital bed with a failing pump. And until now it has been diagnosed only after the invasive procedure it is not supposed to need.
The score’s weak point is buried in the numbers. The development group held just 69 Takotsubo cases against 1,754 heart attacks, and validation held 77 against 1,715. An AUC of 0.93 built on 77 events is a strong signal from a small number of people, and rare-condition scores tend to look worse once they meet an unselected emergency department population. The authors are careful about this themselves, saying the score “might reduce potentially avoidable invasive procedures” rather than claiming it will.
The biology is the more interesting part. The score works because Takotsubo and a heart attack leave different fingerprints: stress signaling and vessel tone on one side, plaque instability and lipids on the other. That a cholesterol number helps sort two chest pain patients says something about what LDL actually is, a marker of one specific process rather than a verdict on a person, which is worth holding onto when reading your own LDL alongside a calcium score.
What it means for you
Nothing about this changes what you should do during chest pain. A score designed to avoid a catheterization is dangerous the moment a patient uses it to avoid an ambulance. This test lives in the cardiologist’s workflow, not yours, and it runs on hospital assays most labs do not offer yet.
What it may change, in a few years, is the experience of chest pain patients who go through an invasive procedure to rule out something they never had. If you have ever been told your heart event happened with clean arteries, that was a real cardiac event with real biology behind it, and it is finally being measured. The idea that a routine blood panel can carry that much signal is where cardiology is heading.