What is a silent heart attack, and how often does one go unnoticed?

A silent heart attack is exactly what it sounds like: damage to the heart muscle that meets the definition of a myocardial infarction but was never recognised as one at the time. It shows up later — as scar on an electrocardiogram or a cardiac MRI — in someone who never sought emergency care, because the event either caused no symptoms or caused ones vague enough to be blamed on indigestion, fatigue or a pulled muscle. It is not a milder disease. On the evidence, it is a heart attack that skipped the alarm.

More common than most people assume

The Atherosclerosis Risk in Communities study followed 9,498 people who were free of cardiovascular disease at enrolment in 1987 to 1989 [s1]. Silent, or unrecognised, myocardial infarction was defined as new ECG evidence of a heart attack in someone with no clinically documented event [s1]. Over a median 8.9 years, 317 participants (3.3%) developed a silent infarction and 386 (4.1%) had a clinically recognised one [s1]. Put another way, silent events made up more than 45% of all incident heart attacks in the cohort [s1] — close to half, going undiagnosed as they happened.

Incidence differed by sex. Silent infarctions occurred at 5.08 per 1,000 person-years in men and 2.93 in women, while recognised infarctions ran at 7.96 and 2.25 respectively [s1]. Men had more heart attacks of both kinds; the sex difference was statistically significant (P<0.0001) [s1].

Silent does not mean harmless

The reason this matters is prognosis. In ARIC, a silent infarction — compared with no heart attack — was associated with roughly triple the risk of death from coronary heart disease (hazard ratio 3.06, 95% CI 1.88 to 4.99) and a 34% higher risk of death from any cause (hazard ratio 1.34, 95% CI 1.09 to 1.65) [s1]. That is lower than the risk after a recognised heart attack (hazard ratios 4.74 and 1.55) but far from benign [s1].

Longer follow-up sharpens the point. The ICELAND MI study used cardiac MRI to detect unrecognised infarction in 935 older adults, then tracked them for up to 13.3 years [s2]. Early on, survival looked reassuring: at three years, mortality after an unrecognised infarction (3%) matched those with no infarction and was well below the 9% seen after a recognised one [s2]. But the gap closed. By ten years, mortality was 49% after an unrecognised infarction and 51% after a recognised one — statistically indistinguishable — against 30% in people with no infarction (P<0.001) [s2]. After adjustment, an unrecognised infarction carried a 61% higher risk of death than no infarction (hazard ratio 1.61, 95% CI 1.27 to 2.04) [s2]. Over a long enough horizon, the "silent" event caught up with the loud one.

Why they get missed

By definition, a silent heart attack produces no symptom clear enough to send someone to hospital. That does not always mean no symptom. The recognised warning signs — chest discomfort, shortness of breath (sometimes the only symptom), and discomfort in the arms, back, shoulders, neck, jaw or upper stomach, along with nausea, lightheadedness or a cold sweat — can be mild, fleeting, or easy to attribute to something else [s3]. A silent infarction is often only the extreme end of that spectrum: symptoms so faint or atypical that neither the person nor, sometimes, a clinician connects them to the heart. Many are found incidentally, years later, on an ECG done for an unrelated reason.

What to take from this

The practical message is not to hunt for a hidden heart attack, and certainly not to self-diagnose from an ECG printout. It is narrower: symptoms do not have to be dramatic to be cardiac, and unusual, unexplained breathlessness, fatigue or discomfort deserves medical attention rather than dismissal. Whether an ECG, imaging or a change in risk management is warranted is a judgement for a clinician who has the whole picture.

The limits

Both studies are observational. ARIC detected silent infarction by ECG, which misses some events and can flag changes that are not infarctions; ICELAND MI used more sensitive MRI in an older, mostly white population aged 67 to 93, so its numbers do not transfer cleanly to everyone [s1] [s2]. Neither trial tested whether treating a silent infarction once found improves outcomes — the ICELAND authors note explicitly that this still requires prospective testing [s2]. What the data establish is prevalence and prognosis, not a screening programme.

And if symptoms are happening now, none of the above applies: anyone with signs of a heart attack should call 911, even if unsure it is one [s3].