Chest pain: gas, or the heart?

Why chest pain gets called gas here

This is the observation that made me write the page, and I want to be straightforward about its status. In this practice, chest discomfort is very frequently presented as “gas” or acidity. Patients arrive having taken an antacid, sometimes for days, and describe the problem in gastric terms even when the story sounds cardiac. That is a clinical observation from a Cooch Behar outpatient practice, not published epidemiology.

What is published is the consequence. Indian patients with a heart attack reach hospital a median of roughly five to six hours after symptoms begin — documented across 20,937 admissions in 89 centres in the CREATE registry, and described there as several times longer than in high-income countries.^[1] A 2022 study from north-east India, the nearest comparable geography, found a mean of 355 minutes, with only 54% of patients arriving inside the window where treatment works best; the principal cause the authors identified was patients not recognising their symptoms at onset.^[2]

On the specific question of what people mistake it for, there is a genuine gap in the Indian literature — I could not find an Indian study quantifying misattribution of cardiac pain to gas or acidity. The nearest quantified evidence is from Karachi: of 300 patients with acute myocardial infarction, 36.3% had interpreted their symptoms as non-cardiac, and of those, 57.5% had attributed them to “gastric” problems — and misinterpretation significantly predicted delay, with over half of that group arriving beyond six hours.^[3] Internationally, a systematic review of 25 studies and 14,843 patients found that only 45.1% of people having an acute coronary syndrome correctly attributed their symptoms to the heart.^[4]

So roughly half of everyone, everywhere, gets this wrong. That is worth knowing, because it means the mistake is not ignorance. It is what the sensation actually feels like.

First, the reassuring part: most chest pain is not the heart

In general-practice populations, only about 1.5 to 3.6% of people presenting with chest pain are diagnosed with an acute coronary syndrome, across a systematic review of 14 studies.^[5] Widen it to coronary heart disease of any kind and a study of 1,199 primary-care chest pain patients found 14.9%.^[6] The commoner causes, by a distance, are chest wall and musculoskeletal pain, reflux, anxiety, and pain from the cervical spine.

Those figures are European and North American. I could not find an Indian primary-care study of what chest pain turns out to be, so treat the proportions as indicative rather than local.

The logic of assessment follows from the arithmetic, though. When something is uncommon but dangerous and the alternatives are common but benign, you exclude the dangerous one first — not because it is likely, but because it is the only one where being wrong for two hours matters. That is why an ECG gets done for a pain that will probably turn out to be muscular.

Why an antacid making it better proves nothing

This is the single most important idea on the page, and it runs directly against intuition.

If a pain settles after an antacid, it feels like proof the pain came from the stomach. It is not. The evidence here is cleanest for nitroglycerin — the drug that ought to be specific to cardiac pain, because it works by dilating coronary arteries. In 459 patients presenting to an emergency department, nitroglycerin relieved chest pain in 35% of those with active coronary disease and 41% of those without, a difference that was not statistically significant.^[7] If the drug that targets the coronary arteries cannot tell you whether the problem is the coronary arteries, an antacid certainly cannot.

The “GI cocktail” — antacid given to see whether the pain settles — has never been shown to distinguish the two, and the major review of chest pain history features does not list antacid response among the findings with usable diagnostic value at all.^[8] That absence is the point: it is not that the test performs poorly, it is that it has never been shown to perform.

There is also a simple reason why relief is unreliable. Cardiac pain waxes and wanes on its own. Anything taken during a waning phase looks like it worked.

What actually shifts the probability

Individual features do carry real information, and it is worth knowing which. The figures below are likelihood ratios — how much a finding multiplies or divides the odds of a heart attack. A ratio above 1 raises the probability, below 1 lowers it.

Two things to notice. Nothing in that table is anywhere near zero — not one feature excludes a heart attack on its own. And the strongest single predictor is not in the table at all: it is whether the person already has known vascular disease, which in a primary-care study carried a likelihood ratio of 4.51.^[6]

The chest-wall point deserves its own sentence, because it is the one most often over-read. Pain reproducible by pressing does make a cardiac cause considerably less likely. It does not exclude it: in one series, 6% of emergency department patients labelled with costochondritis were in fact having a heart attack.^[9]

Why a normal ECG does not settle it either

An ECG is done in the consultation because it is fast, free of risk and occasionally decisive. What it is not is a rule-out test.

The numbers are stark. Standard ECG criteria for a STEMI — the kind of heart attack caused by a completely blocked artery — have a pooled sensitivity of about 63.5% for occlusive infarction across 9 studies and 11,757 patients, meaning roughly a third of people needing urgent intervention do not meet them.^[10] And among 1,675 patients presenting with suspected heart attack, 78.9% had no ischaemic changes on either the admission ECG or a repeat three hours later.^[11]

So an ECG that looks normal narrows nothing very much. What it does is catch the cases where it is abnormal — and in those, catching it in the first hour rather than the sixth is the difference the whole delay literature is about.

What can actually be done in Cooch Behar

Being plain about this, because it has changed and because most people assume otherwise.

An ECG is done here, during the consultation. And troponin T on a one-hour protocol is available locally through the Apollo laboratory — a quantitative result, reported as a number, with a second sample an hour after the first. That matters more than it sounds. India's National Essential Diagnostics List places troponin at district-hospital level rather than clinic level,^[12] and much of what is available at clinic level across the country is a qualitative card that reads positive or negative — which cannot support the one-hour comparison at all, because there is no number to compare. Having the quantitative assay within reach means a proper rule-out is possible in this district rather than only after a journey.

And here is the limit, which is the important half. The one-hour algorithm was designed for a patient who stays under observation between the two samples, in a place with a defibrillator and staff. It is not designed for someone who goes home for an hour and comes back. So the way it is used here is specific: it is for the patient who presents stable, some hours after symptoms that have settled, where the question is whether anything happened at all. That patient can be worked up locally.

The patient with ongoing chest pain, or an abnormal ECG, or who looks unwell, does not wait in a clinic for a laboratory result. They go to hospital, and getting them there fast is the single most useful thing this consultation does. A troponin result an hour from now is no use to someone who needs an artery opened in the next ninety minutes.

When a heart attack does not hurt much

This is the part that most changes what you should do, and it applies particularly to people who read this site.

Across 434,877 patients in a national registry, 33% of heart attacks presented without chest pain at all, and their in-hospital mortality was 23.3% against 9.3% for those with pain.^[13] A later analysis of over 1.1 million patients put the figure at 35.4% overall, with 42.0% of women and 30.7% of men presenting without chest pain — and the sex difference was widest in younger patients.^[14] People with diabetes were substantially over-represented in the no-pain group.

What it looks like instead: breathlessness, sudden profound fatigue, sweating, nausea, a vague sense of something being badly wrong. If you have diabetes — and a large share of the people this clinic sees do — the absence of severe pain is not reassurance. It is a known feature of the disease you already have.

Worth noting that current guidance has abandoned the word “atypical” for these presentations, precisely because calling them atypical made clinicians treat them as unlikely. The 2021 American guideline recommends describing pain as cardiac, possibly cardiac or non-cardiac instead.^[15] They are not atypical. They are one third of all heart attacks.

What it usually is instead

Having spent this page on the cardiac question, the honest resolution is that most of the time the answer is one of these.

Chest wall and musculoskeletal pain is the commonest, including costochondritis at roughly 13% of primary-care chest pain. It is tender to press, worse on certain movements, and settles with time and simple analgesia.

Reflux and acid-related pain is genuinely common, and often genuinely is what the patient thought. It typically worsens after meals and on lying flat, and a trial of a proton pump inhibitor that clearly helps carries a likelihood ratio of about 5.5 for reflux disease.^[16] The distinction from cardiac pain is made by the whole picture, not by whether an antacid helped once.

Anxiety and panic account for 18 to 26% of chest pain in emergency departments and about 25% in family practice — and are badly under-recognised, correctly identified in only around 15% of family-practice cases in one review.^[17] This is a real diagnosis with effective treatment, not a way of saying nothing is wrong, and it is worth asking about directly.

Pulmonary embolism deserves naming because it is missed. In 1,000 consecutive adult medical autopsies at an Indian tertiary centre, pulmonary embolism was present in 15.9% and clinically significant in 12.6% — and 79.9% of those cases were in people under 50.^[18] That is an autopsy series and so biased towards severity, but it is a reasonable warning against assuming that a young patient with sudden breathlessness and chest pain has anxiety.

Also on the list: pleurisy, shingles before the rash appears, and pain referred from the cervical spine.

Danger signs: call 108, do not book an appointment

If any of the following applies, the correct destination is a hospital emergency department, not this clinic.

Crushing or pressing central chest pain, particularly if it spreads to the left arm, both arms, the jaw or the back, and especially with sweating, nausea or breathlessness. Pain that comes on with exertion and settles with rest, appearing for the first time. Chest pain with fainting, or with a very fast or very irregular pulse. Sudden severe breathlessness with chest pain. Any of these in someone with diabetes, even if the pain is mild — see the section above.

Two practical points that matter more than they sound. Do not drive yourself, and do not be driven by a relative who is panicking; call 108. And note the time the symptoms started, because it is the first question the hospital will ask and the answer determines which treatments are still available.

The nearest emergency department is at MJN Medical College & Hospital, Cooch Behar.