Is that chest pain your stomach or your heart? Indistinguishable is the guideline's own word, and it decides the order in which you are seen

TL;DR

• The chest pain section is the most important part of this article. The American College of Gastroenterology (ACG) guideline of 2022 states that reflux chest pain and cardiac chest pain are 「indistinguishable」, and that it can be the only symptom of acid reflux. So this article does not offer — and no authoritative body that can be cited offers — a list for telling the stomach from the heart. The order runs one way only: a doctor excludes cardiac disease first, and acid reflux is investigated after that.
• Of the Hong Kong government texts listed at the end of this article, the only sentence telling the public to go to accident and emergency for a chest symptom is on the Department of Health Elderly Health Service's coronary heart disease page: 「如有以上徵象應該立即請醫生診斷，或到急症室求診。」 — if any of the above signs appear, see a doctor at once or attend accident and emergency. Accident and emergency costs 400 dollars per attendance, and patients triaged as category I (critical) and category II (emergency) are exempt from payment (Hospital Authority charges, effective 1 January 2026).
• The proportion of Hong Kong people with "regular" symptoms is not the one in five so often quoted. Across three Hong Kong telephone surveys, the proportion with weekly symptoms was 2.5% (surveyed 2002, 2,209 interviews completed) and 4.2% (published 2019, 2,011 respondents); a third survey in 2011 (2,074 respondents) used the Montreal definition and arrived at 3.8% — which is not a weekly proportion but a different definition altogether.
• There is no universal list of "red flag" symptoms. The gastroscopy route the UK NICE guideline NG12 sets out for reflux has an age threshold of 55, not 40; and reflux on its own never triggers a referral under NG12 at all. But the recommendation immediately after it in NG12 (1.2.2) states that haematemesis is to be considered for gastroscopy at any age, with no age threshold, and the one after that (1.2.3) opens a further route for people over 55 in which upper abdominal pain together with a low haemoglobin stands as a category of its own. ACG's recommendation lists three alarm features and its body text lists five (adding vomiting and anaemia), both written openly; the six items in the Department of Health's Indigestion leaflet are reproduced in full, but the page itself says they are 「一些」 — some — and dysphagia and haematemesis are not among them.
• The safety of proton pump inhibitors (PPIs) has to be read from both sides at once: the associations found in observational studies and the results of the COMPASS randomised controlled trial in 17,598 people point in different directions. Of the dozen or so rows in ACG's table 5, only one has a 95% confidence interval that excludes 1.0.
• Stopping: best practice statement 9 of the American Gastroenterological Association (AGA)'s 2022 de-prescribing advice states plainly that either tapering or stopping outright may be considered — which is not the "you must always come off slowly" of popular accounts. ACG and NICE each set down a third route as well: on-demand use, and a step further down, self-management with an antacid or alginate (NICE CG184 1.2.5 and 1.5.2).
• For anyone taking clopidogrel (an antiplatelet drug) along with a stomach drug: the footnote to ACG's table 5 records the FDA's advice to avoid using clopidogrel together with omeprazole; but the same ACG guideline also states that for a patient who already needs a PPI, the benefit outweighs a cardiovascular risk its own words call 「highly questionable」. That is a question to take to your next appointment, not a reason to stop a drug by yourself.
• What this article will not do: it will not teach you to tell whether chest pain is cardiac, will not recommend which drug to take or at what dose, will not tell you whether you should have a gastroscopy, and will not compare hospitals. All of that is decided by a doctor on your own circumstances.

---

#What is acid reflux, and why do some people have it when the gastroscopy shows nothing?

Acid reflux is not "if it burns, that is what it is". It is a disease defined by objective investigation — and a great many patients have nothing to see on gastroscopy at all.

The ACG guideline of 2022 defines it this way: the objective definition of GERD is characteristic mucosal injury seen at endoscopy and/or abnormal oesophageal acid exposure demonstrated on reflux monitoring. The UK NICE clinical guideline CG184 (published 3 September 2014, updated 18 October 2019) defines it from another angle: in that guideline, GORD means endoscopically determined oesophagitis, or endoscopy-negative reflux disease. [Note 1]

In one sentence: oesophagitis seen on gastroscopy is one thing, and reflux disease with nothing at all to see on gastroscopy is another — and both are called GORD.

Put differently, "having symptoms" and "being objectively shown to have the disease" are two separate things. That distinction matters a great deal to the treatment rules, but it cannot be stated as though there were only one criterion. The presence and severity of oesophageal injury (the Los Angeles classification) does decide some of the rules: as the mechanism section below explains, for LA grade C or D oesophagitis ACG recommends indefinite maintenance PPI treatment (a strong recommendation).

⚠️ But whether to stop a PPI is not decided on that criterion. The section on stopping below reproduces all ten of AGA's 2022 best practice statements in full, and statement 10 states that the decision to stop a PPI should be based only on the absence of an indication, and not on concern about suspected adverse events. That is to say, the question at that step is whether there is still a reason to be on the drug, not how severe the oesophageal injury is.

Typical symptoms. ACG states: heartburn is the most common symptom of GERD, described as a burning sensation behind the breastbone rising from the upper abdomen towards the neck; and regurgitation is the effortless return of gastric contents up into the mouth, usually with a sour or bitter taste. That last clause is usually skipped over, and it bears directly on the conversation at a follow-up appointment: the two have different origins, and the approach to diagnosis and treatment differs according to which symptom predominates. "Mostly I get the acid coming up" and "mostly I get the burning" are, in ACG's own terms, not the same road. [Note 2]

Manifestations outside the oesophagus. ACG uses two lists, and the two are not the same: one gives hoarseness, throat clearing, chronic cough, laryngitis, pharyngitis and pulmonary fibrosis; the other gives chronic cough, dysphonia, asthma, sinusitis, laryngitis and dental erosion. Asthma, sinusitis and dental erosion appear only in the second. Both are introduced by "such as" and both are examples. And the second is immediately followed by a limitation: these symptoms and conditions have low sensitivity and low specificity for diagnosing GERD, and diagnosing GERD on extra-oesophageal symptoms alone, or on the response of those symptoms to a PPI, is unreliable and not recommended. [Note 3]

So can a gastroscopy find it or not? ACG has two recommendations answering that question directly, and both are "do not": do not use upper gastrointestinal endoscopy to establish a diagnosis of GERD-related asthma, chronic cough or laryngopharyngeal reflux; and diagnosing laryngopharyngeal reflux on laryngoscopic findings alone is not recommended, with additional testing to be considered. [Note 4]

In one sentence: "inflammation" seen on laryngoscopy does not establish that reflux caused it — in a study ACG cites, 86% of 105 normal asymptomatic volunteers had laryngoscopic findings associated with reflux.

But this ACG section is not four "do nots". Of its six recommendations, the first three are positive — that is, ACG did not only write down what not to do; it also wrote down how this group of patients should be worked up: in patients with possible extra-oesophageal manifestations, evaluation for non-GERD causes is recommended before attributing the symptoms to GERD (strong recommendation); in patients with extra-oesophageal manifestations but without typical GERD symptoms, reflux testing is recommended before starting PPI treatment (strong recommendation); and in patients with both extra-oesophageal and typical GERD symptoms, a trial of twice-daily PPI treatment for 8 to 12 weeks may be considered before other investigation (conditional, low evidence). [Note 5]

Those three together make a road with a fork in it, and the fork is whether you have heartburn or regurgitation:

• Without typical symptoms (that is, only hoarseness or cough) — ACG recommends reflux testing first, before the PPI. That one is a "strong" recommendation. What it expressly argues against is precisely the practice that is common in Hong Kong: no investigation, just two months of a stomach drug to see what happens.
• With both — a trial of twice-daily PPI for 8 to 12 weeks may be considered before other investigation. That one is "conditional, low evidence", and note that both the frequency (twice daily) and the length (8 to 12 weeks) differ from the typical-symptom route (once daily, 8 weeks).
• And before either of those there is another "strong" recommendation: evaluate the causes other than GERD before attributing the symptoms to GERD. Hoarseness and chronic cough have many causes, and ACG puts "do not assume the stomach at the outset" at the very front of this section.

Why those three have to be set beside the four "do nots": print only the four, and a reader who has seen several doctors about hoarseness concludes that nothing can be found and gives up. ACG's actual position is the opposite — there is a road; it simply is not "have a gastroscopy" or "take a PPI and see". It is to exclude the other causes first, and then fork according to whether the typical symptoms are there.

Who this section is most use to: anyone who has seen several doctors about hoarseness or a chronic cough and been told it might be acid reflux. ACG says in terms that reasoning backwards from those symptoms to GERD is unreliable, that improvement on a PPI does not count as evidence, and that neither gastroscopy nor laryngoscopy alone is the way to make that diagnosis — but the same section also sets down how it should be worked up.

---

#What is the mechanism, and why are the rules drawn the way they are?

Acid reflux is not "too much acid". It is a valve that has failed — and understanding that is what makes it clear why the treatment suppresses acid rather than repairing the valve.

ACG 2022's account of the pathophysiology is that reflux is the result of three lines of defence failing at once: the barrier, clearance and the mucosa. [Note 6] Which is why some people whose acid has been suppressed by a drug still have symptoms — their problem lies in sensitivity or in emptying, not in the concentration of acid.

Why the rules turn so much on the grade of oesophagitis. ACG grades oesophagitis A to D on the Los Angeles classification, and the grade decides directly whether long-term medication is needed: for LA grade C or D oesophagitis, indefinite maintenance PPI treatment or antireflux surgery is recommended. [Note 7] That is one of the few recommendations in the whole ACG guideline about long-term medication rated "strong", and it is also the watershed for the section on stopping below.

Asian and Western practice have already parted company here. The Seoul consensus of 2020 (whose author list includes Wu JCY of the Institute of Digestive Disease at the Chinese University of Hong Kong) takes the view that in the setting of typical symptoms a PPI test can serve as a diagnostic tool; ACG says the PPI response is unreliable. [Note 8] But note that the two are speaking about different settings: ACG is speaking about chest pain and extra-oesophageal symptoms, and the Seoul consensus about typical heartburn and regurgitation. That divergence is a live one and this article does not resolve it for either side; but on either account chest pain does not fall in the "typical symptoms" box, and the order — exclude cardiac disease first — is unaffected.

The upper reference limit for oesophageal acid exposure time in Asians recorded by the Seoul consensus is 3.2% (95% confidence interval 2.7–3.9%), on a denominator of a meta-analysis of 19 estimates. But 3.2% is not the threshold the Seoul consensus itself proposes — immediately afterwards in the same passage it proposes that in Asians an acid exposure time above 4.0% should be judged abnormally high, and the same sentence adds a limitation: that while this is a reasonable threshold for defining the abnormal range, it may not be the threshold at which GERD symptoms are provoked. [Note 9]

⚠️ An inconsistency in the source document itself has to be marked. Statement 13 of the Seoul consensus reads 「≥ 4%」 (4.0% itself already counting as abnormal), while the body text of the same document reads 「more than 4.0%」 (4.0% itself not counting). What separates the two is exactly the point 4.0% itself. This article reproduces both and does not reconcile them — a report reading exactly 4.0% is abnormal by the statement and not abnormal by the body text.

For comparison: the Lyon consensus commonly used in the West takes a distal oesophageal acid exposure time above 6% on 24-hour pH-impedance monitoring as one of the objective criteria for "proven GERD". These three figures are not the same kind of threshold: 3.2% is the upper limit of normal in an Asian population, 4.0% is the abnormal threshold the Seoul consensus proposes on the strength of it, and 6% is the threshold Lyon uses to define "proven GERD". And Lyon does not consist of the 6% alone — there is a band between the two figures, and many Hong Kong readers' reports fall in it: Lyon treats below 4% as normal and above 6% as abnormal, with everything in between as inconclusive. The Porto consensus groups 4–6% together with Los Angeles grades A and B as the range of unproven diagnosis. [Note 9]

Which is to say that a report reading 5% gets a different answer under each of the three: inconclusive under Lyon; the unproven-diagnosis range under Porto; and already above the 4.0% Asian abnormal threshold under Seoul — where the same Seoul sentence goes on to say that being above it still does not mean the symptoms are caused by acid. Those three answers are not a matter of who is right; they are three lines in current use, and how a report is read has to be done by the doctor who ordered the test, together with your symptoms and the other results.

Who this section is most use to: anyone in Hong Kong holding the report of a 24-hour oesophageal pH study and unsure how to read it. Which reference range is used changes the conclusion "normal or not normal" itself. Especially in the band from 4% to 6% — Lyon calls it inconclusive in terms and Seoul treats it as abnormal; with a figure in that range, the question most worth asking the doctor is which set of reference values they are using.

---

#How many people in Hong Kong actually have it? There is a widely circulated wrong number here

Hong Kong has no government-published prevalence for acid reflux. There are only three academic telephone surveys, and all three put the proportion with weekly symptoms between 2.5% and 4.2% — not one in five.

⚠️ The 4.2% and 13.9% from Mak 2019 in the table above come from a paper whose subject is not prevalence, and its own principal conclusion is part of the same set of results and has to be given with them: that study reports that people with monthly GORD symptoms were at higher risk of generalised anxiety disorder (p = 0.01) and of major depressive episode (p < 0.001), with a dose-response relationship between symptom frequency and both. That is an association and not causation, and the paper itself does not say which comes first. But it explains why the UK guideline puts psychological therapy in the lifestyle box (see CG184 1.2.4 in the lifestyle section below).

Do the arithmetic once: why the "one in five" figure appears. The largest figure in the three surveys is Wong 2003's annual prevalence of 29.8% — that is, anyone who has had it once in the past year counts. Among the same 2,209 people, counted on a "weekly" definition, it falls to 2.5%. The same population, the same questionnaire, the same survey, and the two figures differ by a factor of 12 — the whole of the difference is the definition of symptom frequency. So any statement that "X in ten people in Hong Kong have acid reflux" means nothing unless it says whether it is counting weekly or annually.

Tan 2016 concludes that the proportion is rising; and Wong 2003's own concluding sentence, which sits immediately after the figures in the table above, has to be given with them too — the proportion in Hong Kong is lower than in the West, but the same paper also states that these patients carry a significant socioeconomic burden. [Note 10] So this article's correction of the "one in five" is a correction of that number, not a statement that the thing does not matter.

The Asia-Pacific consensus of 2016 (whose authors include Wu J of the Chinese University of Hong Kong) records the same direction, and points out a distribution of phenotype that matters a great deal to Hong Kong readers: Asia is predominantly non-erosive reflux disease (NERD — the kind with no oesophagitis to see on gastroscopy); and its third sentence is the conclusion that follows immediately — in Asia, PPI-refractory reflux disease is considered to be fairly common. [Note 10] The subject of that sentence is a disease category already defined, not anyone at all who still has symptoms after taking a PPI; and the same consensus accordingly treats "refractory symptoms" and "refractory reflux disease" separately.

That phenotype has two consequences, both of them written in ACG: first, what ACG recommends for NERD is on-demand or intermittent PPI use rather than long-term continuous use; and second, the PPI response has least diagnostic value of all in these patients.

As to why Hong Kong has no government figure: the Department of Health health topic pages have no reflux page (the full report of the population health survey is not among the sources at the end of this article, and this paragraph's negative does not cover that report); the Centre for Health Protection has no health topic page for acid reflux (the nearest is 《男士健康資訊 — 消化性潰瘍》, its men's health information page on peptic ulcer, which is about gastric ulcer and does not mention acid reflux); and the Hospital Authority Smart Patient site has no GERD disease page.

Who this section is most use to: anyone who has read an article online and concluded that so many people have it that it need not be attended to. The proportion with weekly symptoms in Hong Kong is 2.5%–4.2%, which is not common enough to be dismissed automatically; and equally, no Hong Kong figure supports "one in five".

---

#Chest pain: this is the most important section in the article

There is no symptom feature that lets a person tell for themselves whether chest pain is acid reflux or a cardiac problem. That is not a cautious formulation; it is what the guideline text that can be cited actually says.

ACG 2022 states: chest pain indistinguishable from cardiac pain can be a manifestation of GERD, and can even be the only symptom. Its recommendation is not "go to accident and emergency if you suspect the heart" but the other way round: the step of investigating GERD is taken after cardiac disease has been excluded. The same recommendation carries one extra qualifier where it appears in the body of the guideline, and this article reproduces both; what the two versions have in common — "after cardiac disease has been excluded" — is the same in both places, and no version permits the order to be reversed. [Note 11]

Getting better on a PPI does not establish that it was acid reflux. That has to be said separately, because so many people use it to diagnose themselves: the systematic review ACG cites found that when endoscopy and pH monitoring are normal there is no significant response to a PPI over placebo, and that chest pain and heartburn do not reliably predict the response to a PPI. [Note 11] Which is to say that "I took the drug and got better, so it is acid reflux" does not hold up on the evidence ACG cites. A treatment trial is not a diagnostic test.

#Which symptoms "may indicate" an acute coronary syndrome

Recommendation 1.2.1.3 of the UK NICE clinical guideline CG95 (published 24 March 2010, updated 30 November 2016) lists four items in full: pain in the chest, the upper abdomen or both, which may radiate elsewhere (the examples given in the original being the arms, back or jaw); chest pain lasting longer than 15 minutes; chest pain with nausea and vomiting, marked sweating or breathlessness (or any combination of these); and new-onset chest pain, or a marked deterioration in the pattern of previously stable angina. [Note 12]

Two things are worth noting. First, the sites of radiation in the first item are the guideline's own examples — arms, back and jaw are not a closed list, and other sites count equally. Second, this list is "may indicate", not "only counts if it matches". A symptom not on it does not mean nothing is wrong. No source that can be cited offers a threshold of the form "so many items and it needs attention".

Recommendation 1.2.1.7 of the same guideline states when to admit as an emergency, and it expressly covers the case of having no electrocardiogram — that is, the ordinary member of the public's situation: where an acute coronary syndrome is suspected and there is currently chest pain, or the pain has now gone but there was chest pain within the past 12 hours and the resting twelve-lead electrocardiogram is abnormal or simply not available, refer to hospital as an emergency. [Note 13]

Note those last three words: "or not available". Not having had an electrocardiogram does not make this rule inapplicable; on the contrary, it is written precisely for that case. In Hong Kong the corresponding route is accident and emergency.

And the recommendations immediately following are where most people actually are — the pain has passed. [Note 14]

• 1.2.1.8: where there is no reason for emergency referral, and there was chest pain within the past 12 hours but the pain has now gone and the electrocardiogram is normal, or the last episode of pain was 12 to 72 hours ago, refer for urgent same-day assessment.
• 1.2.1.9: where the pain has resolved but there are signs of a complication (the original uses "such as" to give its example, which is pulmonary oedema), refer for hospital assessment, the urgency being decided by clinical judgment. That one sits between 1.2.1.8 and 1.2.1.10, and what it covers is precisely the case of "the pain has gone but there is something else" — the pain having subsided does not make it inapplicable.
• 1.2.1.10: where the last episode of pain was more than 72 hours ago and there are no complications, what NICE writes is not merely "assess" — it states expressly that an electrocardiogram and a troponin should be done to confirm the diagnosis, and that how the result is interpreted afterwards must be adjusted for how long it is since the onset. What that means for the reader in practice: if you had pain last week and are only seeing a doctor today, "blood test and electrocardiogram" is what NICE writes down for that box, not the doctor doing more than is needed.

⚠️ There is another recommendation after 1.2.1.10, and it is the one in the whole of section 1.2.1 that speaks most directly to this article's readers — because what it describes is exactly "I had this before, someone saw me, and now it has come back". Recommendation 1.2.1.11: if the previous episode was, or was suspected to be, an acute coronary syndrome and the pain is now recurring, what NICE writes is emergency referral (「as an emergency」), not same-day and not by clinical judgment. Recurrence overrides all three of the time boxes above. [Note 15]

Why that one matters especially to someone reading an article about acid reflux: the group it applies to is exactly the people who went to hospital or saw a doctor about chest pain last time — including anyone told last time that it was probably acid reflux and now in pain again. Not one of the three boxes at 1.2.1.8 to 1.2.1.10 covers that situation.

And the recommendation immediately after that, 1.2.1.12, is about what healthcare staff should be doing at the same time, and contains one sentence about the order of transfer that a reader can use as well: where an acute coronary syndrome is suspected, management should start at once, but transfer to hospital must not be delayed in order to do an electrocardiogram. [Note 15] That runs in the same direction as the "or not available" at 1.2.1.7 above: an electrocardiogram is not a precondition to be assembled before going to hospital.

What the group of them adds up to: having had pain once and not since is not "nothing", only "not the call-an-ambulance-now" tier; but if the previous episode was already suspected to be cardiac and the pain is back, it returns to the "now" tier. And in whichever box you are, once there are signs of a complication such as pulmonary oedema, 1.2.1.9 requires referral to hospital; and once it is recurrent pain after a recent acute coronary syndrome, 1.2.1.11 requires emergency referral.

(For accuracy: section 1.2.1 of CG95 runs from 1.2.1.1 to 1.2.1.13. What this article reproduces item by item are those bearing directly on a member of the public's situation; 1.2.1.1 and 1.2.1.2 are about what healthcare staff must establish at the moment they take over, which is clinical process and is not reproduced here — but they are not a further threshold.)

⚠️ The same guideline has one more recommendation that matters more than all of the above in an article like this one: recommendation 1.2.2.5 — do not exclude an acute coronary syndrome on the basis of a normal resting twelve-lead electrocardiogram. That is to say, "I had an electrocardiogram and they said it was fine" is not a ticket out. Recommendation 1.2.1.7 above says that the case where the electrocardiogram is not available calls for emergency referral; 1.2.2.5 says the reverse — that a normal electrocardiogram cannot be used to exclude. Put the two together and the conclusion is the same: the result of an electrocardiogram is not a tool a layperson uses to decide whether to go to hospital. [Note 16]

And 1.2.2.5 does not stand alone. It sits inside section 1.2.2 of CG95 (headed "Resting 12-lead ECG"), eight recommendations in all, three others of which bear directly on the situation of "the electrocardiogram looks all right": 1.2.2.4 — even in the absence of ST-segment change, other changes (particularly Q wave and T wave changes) should raise suspicion; 1.2.2.6 — where the diagnosis is in doubt, consider serial electrocardiograms, reviewing previous ones and recording additional leads, its own last sentence being that the results may not be conclusive; and 1.2.2.8. [Note 16]

That last sentence is NICE's own, in print: even after several more recordings and several more leads, the result may still be inconclusive. So "the doctor said the electrocardiogram does not look like anything" is not an endpoint even in NICE's own formulation.

⚠️ And in the last recommendation of that section NICE writes down the thing this article's readers most easily overlook: excluding the heart does not leave only a stomach. Recommendation 1.2.2.8: if clinical assessment and the resting twelve-lead electrocardiogram make an acute coronary syndrome less likely, consider other acute causes of chest pain, the three examples given in the original being pulmonary embolism, aortic dissection and pneumonia. [Note 16] All three names are NICE's own in print, introduced with "such as", and are examples.

So the conclusion "it is not the heart" cannot automatically become "it is the stomach". For the reader, what that means in practice is: if accident and emergency or a doctor says the cardiac side does not look like it, and your pain has not improved, or there is something else such as breathlessness, you should not simply take it to be acid reflux and go home to take a stomach drug.

NICE CG95 has a separate set of feature-based scoring for clinicians to categorise stable chest pain. This article does not reproduce it, because it is a categorisation tool for a doctor in a consultation, not a tool for a person to judge themselves with in the middle of an episode; and the risk of using it to judge yourself is exactly that of arriving at a mistaken "no need to worry". And 1.2.1.13 covers the other side: where an acute coronary syndrome is not suspected, other causes of chest pain still have to be considered, the guideline's own words being that some of them may be life-threatening.

#The three ways of telling them apart that NICE expressly forbids

These three are written for clinicians, but every one of them collides with a way members of the public habitually reason: do not use a patient's response to glyceryl trinitrate to make a diagnostic decision; do not assess symptoms of an acute coronary syndrome differently because of gender; and do not assess them differently because of ethnicity. [Note 17]

Put differently: "the tablet under the tongue did not help", "I am a woman so it is unlikely to be the heart", and "Chinese people do not get so much heart disease" are three lines of reasoning that NICE tells doctors in terms not to use. A method doctors are not allowed to use is one a reader should certainly not use on themselves.

#The one sentence the Hong Kong government addresses to the public

Among the Hong Kong government texts listed at the end of this article, only the Department of Health Elderly Health Service's coronary heart disease page actually tells the public what to do about a chest symptom: 「如有以上徵象應該立即請醫生診斷，或到急症室求診。」 [Note 18]

Two things are worth noting. First, that page's lead-in reads 「有些人可能出現下列徵狀」 — some people may develop the following signs — and "some people" and "may" mean the table does not present itself as exhaustive. Second, the same page lists angina before myocardial infarction, and the angina passage describes pain occurring on exertion, under emotional stress, or when walking quickly. That item is easily taken as "not important" — but it is on the same page and inside the same table of signs. The same page also states that about 10% of patients have no signs at all.

The English and Chinese versions of this page do not correspond item by item: the English lists sweating, shortness of breath, nausea and vomiting, and pass out; the Chinese lists 「大量出汗、氣促；翳悶、作嘔；暈眩，甚至不省人事」 — the Chinese adding a feeling of oppression and dizziness, and the English having vomiting where the Chinese has only nausea. Both lists are printed in full in their own version, but neither version states that the list is exhaustive.

Another fact worth knowing: the Centre for Health Protection's own Heart Disease topic page (page dated 31 March 2026) lists symptoms, risk factors and prevention, but contains no instruction to go to accident and emergency or to call an ambulance. Two websites under the same Department of Health, one with an instruction to act and one without — so when you are looking things up, arriving at the Centre for Health Protection's page does not mean you have seen the government's advice on what to do.

What the Centre for Health Protection's version adds is 「心律紊亂」, 「四肢無力」 and 「雙腳水腫」 — disturbance of heart rhythm, weakness of the limbs and swelling of both feet — which the Elderly Health Service version does not have; and on the item of where the pain radiates, the Centre for Health Protection lists 「手臂、肩膀、頸部和下顎」 while the Elderly Health Service lists only 「頸、手臂、下顎」 — the shoulder appears only in the Centre for Health Protection version. Neither table states that it is exhaustive, so the inference that the two put together make up the whole does not hold either. As for the sentence 「但休息後便會有所好轉」 — that it improves with rest: it is the Centre for Health Protection's description of angina, not a rule that improving with rest means nothing is wrong; NICE CG95 expressly forbids diagnosing on symptom response, and this article will not treat that sentence as a basis for exclusion. [Note 18]

⚠️ The largest divergence between the two versions is not in the symptom list but in a basic fact: where heart disease ranks in Hong Kong. The Elderly Health Service version (page carrying no revision date) states 「現時心臟病是僅次於癌症的香港第二號殺手」 — that heart disease is currently Hong Kong's number two killer after cancer; the Centre for Health Protection version (page dated 31 March 2026, using registered deaths for 2024) states that heart disease ranks third among the most common fatal diseases locally, accounting for 12.6% of all registered deaths. Two websites under the same Department of Health, one saying second and one saying third. This article uses the Centre for Health Protection's "third", because it states its page date and the year of its data. That is not a detail: two government web pages saying different things about the same fact is itself something the reader needs to know.

The Centre for Health Protection's page supplies the local figures: in 2024 heart disease caused 6,594 registered deaths, 12.6% of all registered deaths in Hong Kong, making it the number 3 killer; and coronary heart disease accounted for 52.9% of deaths from heart disease. In the 2020–22 population health survey, 1.6% of people aged 15 or above reported having been diagnosed by a doctor with coronary heart disease (men 2.1%, women 1.2%), rising from 0.1% at ages 15–24 to 7.7% at 85 or above.

#Why the local Hong Kong data makes this easier to get wrong

Both Hong Kong telephone surveys found an association between GERD symptoms and non-cardiac chest pain: Wong 2003 (2,209 people), odds ratio 2.3, 95% confidence interval 1.7–3.1; Tan 2016 (2,074 people), odds ratio 1.7, 95% confidence interval 1.034–2.9.

The same sentence in the original has a second half, and the second half is the complete result. In that Hong Kong sample of 2,209, no association was found between GERD symptoms and hoarseness, asthma, bronchitis or pneumonia — so the inference "I am hoarse a lot, it might be acid reflux" has no support in the local Hong Kong data. Just as important: there was no association with dysphagia either — dysphagia is not to be explained away as part of reflux; in NICE NG12 it is an item that triggers referral at any age. [Note 19]

⚠️ And the results paragraph of Wong 2003's abstract has one more sentence immediately after that one — reporting precisely what this article credited to Mak 2019 above, only sixteen years earlier: that 2002 survey of 2,209 people had already reported higher anxiety and depression scores in GERD patients, and one thing more that Mak did not report — days of sick leave, which is the concrete content of the "significant socioeconomic burden" of the conclusion. The same qualification applies: this is association and not causation, and neither study says which comes first. [Note 19]

Those two odds ratios have to be read very carefully, and they carry two meanings. They measure the association between GERD symptoms and a group already defined as having non-cardiac chest pain — they do not say that chest pain in Hong Kong people is mostly acid reflux. The second meaning is statistical: an odds ratio measures the strength of an association, not how often two things occur together — the originals give no absolute co-occurrence rate and no denominator, so it cannot be said that chest pain and reflux together are a common combination. What those two figures are actually for is a warning: GERD symptoms really are substantially associated with this kind of chest pain, which makes the reasoning "this is the same as last time, and last time they said it was the stomach" particularly easy to fall into and particularly dangerous — the danger lies not in how common the combination is but in the fact that an association does not mean this time must have the same cause.

#What accident and emergency costs

Hospital Authority charges (effective 1 January 2026): accident and emergency is 400 dollars per attendance, and patients triaged as category I (critical) and category II (emergency) are exempt from payment. The same fee schedule also sets an annual cap of ten thousand dollars on public healthcare charges (applied for under the Authority's own procedure, not automatic), with no means test.

Who this section is most use to: three sorts of people. First, anyone who believes they "know" that their kind of chest pain is acid reflux — ACG says in terms that the two cannot be told apart. Second, anyone hesitating because they are afraid accident and emergency is expensive — categories I and II are exempt, meaning the most urgent are the ones who do not pay. Third, anyone searching online for "acid reflux chest pain difference" — the table you are looking for has never been published by any authoritative body that can be cited.

---

#Which are the "red flag" symptoms, and why is every list different?

There is no universal red flag list. The three sources that can be cited here produce three different lists, differing in both thresholds and items — so the inference "my symptom is not on the list" fails on every one of them.

NICE NG12's oesophageal cancer referral recommendation 1.2.1 provides: dysphagia on its own calls for referral on a suspected cancer pathway at any age; and so does being aged 55 or over with weight loss together with any one of upper abdominal pain, reflux or dyspepsia. And 1.2.1 is not the whole of that section — 1.2.2 immediately after it, and 1.2.6 on the gastric cancer side, each open a door that 1.2.1 does not cover, and neither has an age threshold: haematemesis at any age calls for non-urgent direct access endoscopy to be considered; and a palpable upper abdominal mass calls for referral on a suspected cancer pathway to be considered. [Note 20]

(Recommendations 1.2.7, 1.2.8 and 1.2.9 of the gastric cancer section repeat the criteria of 1.2.1, 1.2.2 and 1.2.3 respectively, with the same threshold of 55, the same haematemesis rule and the same set of non-urgent direct access conditions.)

And 1.2.3 has to be laid out whole, because it is the only endoscopy route tied directly to reflux, and it contains four boxes, not two. [Note 21] Three things to see clearly:

1. The whole rule is governed by its first line, "aged 55 or over" — all four boxes below hang on that age condition, and not one of them is age-independent.
2. The list paired with the raised platelet count runs to six items (nausea, vomiting, weight loss, reflux, dyspepsia, upper abdominal pain), and the list paired with nausea or vomiting runs to four (weight loss, reflux, dyspepsia, upper abdominal pain) — the two pairing lists are different and cannot be merged into one box.
3. The first two boxes stand on their own: aged 55 or over with treatment-resistant dyspepsia, or with upper abdominal pain and a low haemoglobin, is enough by itself with nothing else to pair it with.

Four conclusions follow directly from those passages:

1. In the endoscopy route NG12 sets out for reflux, the age threshold is 55, not 40. But to be accurate: section 1.2 of NG12 does not have only one age threshold. In the pancreatic cancer part of the same section, 1.2.4 provides for referral on a suspected cancer pathway for pancreatic cancer in people aged 40 or over with jaundice; and 1.2.5 opens a route to an urgent direct access CT scan (or urgent ultrasound where CT is not available) for people aged 60 or over with weight loss together with any one of diarrhoea, back pain, abdominal pain, nausea, vomiting, constipation or new-onset diabetes. So the accurate statement is that this upper gastrointestinal section of NG12 uses three age thresholds at once — 40, 55 and 60 — but the one at 40 is jaundice with pancreatic cancer and the one at 60 is weight loss with CT, and neither is the endoscopy route for reflux. Put differently, "reflux at 40 means you should have a gastroscopy" still has no source in NG12.
2. Reflux on its own never triggers a referral in NG12. From beginning to end it is only ever an additional condition. And 1.2.3 is not just the two reflux boxes — aged 55 or over with treatment-resistant dyspepsia, or with upper abdominal pain and a low haemoglobin, each stands as a box of its own. Which is to say: anaemia does have a route of its own in NG12, only NG12's version adds the age of 55 and the pairing condition of upper abdominal pain.
3. Dysphagia on its own triggers referral at any age (1.2.1 and 1.2.7).
4. Haematemesis on its own calls for endoscopy to be considered at any age (1.2.2 and 1.2.8); and so does a palpable upper abdominal mass (1.2.6). Neither is in 1.2.1, and neither is among the six items of the Department of Health's Indigestion list below — so anyone who reads only one rule, or only one list, will miss them.

⚠️ One more thing about the upper abdominal mass: section 1.2 of NG12 does not stop at 1.2.9 but runs to 1.2.11, and the last two are both about an upper abdominal mass. Recommendations 1.2.10 and 1.2.11 open urgent direct access ultrasound routes from an upper abdominal mass to gallbladder cancer and to liver cancer respectively. [Note 22] That is to say, a person with a palpable upper abdominal mass has not one route in NG12 but three — and the last two open an ultrasound, not a gastroscopy. What that means for the reader in practice: feeling something in the upper abdomen is not a situation that "a gastroscopy will settle"; which organ is investigated, and with what, is decided by the doctor on what they feel.

The Department of Health's Indigestion list runs to six items in full, and the Chinese page's original text is reproduced in the notes. [Note 23] The list corresponds item by item between the Chinese and English pages, with one exception, the first item: the English reads 「Persistent dyspepsia or poor appetite for an unusual long period」 while the Chinese reads only 「持續性食慾不振」, persistent loss of appetite, without the persistent indigestion. The English version's first item is wider than the Chinese one.

So what is the action? Here it has to be said plainly that the three sources write the urgency differently. NICE NG12 speaks of a "suspected cancer pathway referral" and a "non-urgent direct access" tier, which are terms of art in the UK National Health Service with no equivalent mechanism in Hong Kong. The Department of Health version's instruction to act is to seek medical attention early if in doubt. NICE CG184's instruction for dyspepsia with severe acute gastrointestinal bleeding is the most definite of them: same-day specialist referral. [Note 24] And the Hospital Authority's gastroscopy leaflet states for itself that where a serious incident occurs after the examination (bleeding in the gut or severe abdominal pain, for example), the nearest accident and emergency department should be attended. [Note 25]

⚠️ There is one further instruction to act, on the same Department of Health Indigestion page but not inside the table of signs above — it is in the "using medicines properly" box, and it is precisely the other half of the fifth item of the table. The fifth item of the table above is that symptoms of indigestion in someone taking aspirin or painkillers long term are signs not to be ignored; the other box on the same page is what to do about it: aspirin, painkillers and non-steroidal anti-inflammatory drugs can all cause peptic ulcer, do not take them on your own initiative, and joint pain should be prescribed for by a doctor first. [Note 26]

Only the two boxes together are complete: one says "watch out if this is your situation", the other says "so do not buy them for yourself, and see a doctor about the joint pain first". Reading only the six items of the sign table misses that sentence — and in Hong Kong a great many people buy painkillers for themselves long term. That sentence, NICE CG184 1.3.2 below, and the odds ratio of 2.3 (95% confidence interval 1.5–3.6) Wong 2003 found in Hong Kong for the use of non-steroidal anti-inflammatory painkillers are three sources on the same thing.

Who this section is most use to: anyone ticking their way down an online "red flag list", finding they match none of it, and deciding not to see a doctor. The three lists differ in threshold and in items, and two of them say in terms that they are only examples. Not being on the list does not mean it does not matter.

---

#Should you have a gastroscopy? Hong Kong and the UK and US do not give the same answer

There is a real and current divergence here: the Hospital Authority's own gastroscopy leaflet says that patients with acid reflux 「均應接受胃鏡檢查」 — should all undergo gastroscopy — while both the NICE and the ACG guidelines say most such people should not have a gastroscopy first. Both are primary documents and both are still in force. And the NICE side is not a flat "do not do it" either — CG184 1.6.11 states in terms that it may be considered where the person has GORD.

What the ACG recommendation says is: for patients with the typical GERD symptoms of heartburn and regurgitation and without alarm symptoms, a trial of once-daily PPI treatment for 8 weeks is recommended first. And the recommendation immediately after it covers the opposite box: for patients with alarm symptoms (three listed in its brackets — weight loss and gastrointestinal bleeding), endoscopy is recommended as the first evaluation. But another place in the body of the guideline lists five, and adds a time limit: dysphagia, weight loss, bleeding, vomiting and anaemia, with such patients to be endoscoped as soon as feasible. [Note 27]

The two places differ by two items: vomiting and anaemia. The three in the brackets are not ACG's complete definition of "alarm symptoms"; both places are written openly ("other alarm symptoms", "such as") and the numbers themselves do not agree. So using ACG's "no alarm symptoms means medication first" to place yourself is not safe — particularly for anyone whose blood test has found anaemia, or who is vomiting persistently: they do not find themselves in the three-item version and do find themselves in the five-item one.

⚠️ And in ACG's own list of recommendations for diagnosing GERD, a third one sits between those two — the only "strong" one, and the one that covers this article's largest group of readers: people for whom the drug did not work well enough, or whose symptoms came back after stopping. What the original means is: for patients with persistent symptoms in whom GERD is suspected, diagnostic endoscopy is recommended, preferably two to four weeks after stopping the PPI. [Note 28]

That one deserves separate treatment, for three reasons.

First, it is a "strong" recommendation. Few in the whole ACG guideline are rated strong, and elsewhere in this article (LA grade C or D needing long-term medication, for instance) "strong" has been treated as a watershed. This one is a "strong" recommendation that a person with no alarm symptoms can fall into — reading down the other two branches, someone with no dysphagia, no weight loss and no bleeding easily assumes they are permanently in the "just take the drug" box. That is not what ACG writes.

Second, its population is the very Hong Kong situation this article has itself pointed out. The Asia-Pacific consensus section above noted that PPI-refractory reflux disease is considered fairly common in Asia; and the section on stopping below deals with symptoms recurring after the drug is stopped. Neither of those situations is "try a bit longer" in ACG; both are an indication for endoscopy.

Third, the half-sentence "preferably two to four weeks after stopping the PPI" is very practical. It means stopping the drug before the examination, because being on a PPI heals the oesophagitis and leaves nothing for the endoscope to see. So if a doctor arranges a gastroscopy and tells you to stop the drug for a few weeks, that is not an oversight; it is what the guideline says. That half-sentence also explains why you should not arrange to be scoped while still taking the drug.

What has to be read carefully: "strong" refers to the strength of the recommendation and the evidence rating is "low"; and this is an indication written for a doctor, not a self-service booking voucher. Whether it can be had in Hong Kong is still a question of referral and waiting. But someone who is still not well enough after eight weeks of a PPI, or who relapses as soon as they stop, at least knows they are not out of investigations.

Two further recommendations in the same list, also rated "strong", follow immediately: for GERD suspected but unproven at endoscopy, reflux monitoring off medication is recommended to establish the diagnosis; and where endoscopy has already shown Los Angeles grade C or D oesophagitis, or long-segment Barrett's oesophagus, off-medication reflux monitoring to prove GERD is not recommended. [Note 29] The former is precisely the next step for the group whose gastroscopy showed nothing at all — the non-erosive reflux disease described above; a normal gastroscopy is not the end of a road.

(One internal divergence in the source to be marked: those two are worded slightly differently in ACG's table one and in the body — the table reads 「We suggest against」 where the body reads 「We recommend against」, and the table reads 「LA grade C or D」 where the body reads 「Los Angeles grade C or D」; the ratings are "strong/low" in both places. The notes reproduce the body version.)

How this divergence should be read. Four limitations have to be carried along with it:

1. The Hospital Authority document is an examination information leaflet, not a referral guideline — it describes what a gastroscopy is for, and does not mean the Authority operates direct access gastroscopy for reflux patients.

2. Hong Kong's position on asymptomatic average-risk people is expressly the opposite — but the same passage also sets down a second route, and the two have to be given together. That position is taken by the government's Cancer Expert Working Group on Cancer Prevention and Screening and appears on the Cancer Online Resource Hub's Stomach cancer page, which writes it in two boxes. ⚠️ The Chinese and English versions of that page contradict each other in the heading of the second box: the English reads 「For symptomatic persons at increased risk」 and the Chinese reads 「較高風險的無症狀人士」 — one says symptomatic and the other asymptomatic, on the same government page and in the same box. This article cannot determine which version is the working group's intention, so it reports the divergence and does not take a side; the originals of both boxes in both languages are reproduced in the notes. [Note 30]

   The body text of the two boxes agrees between the Chinese and English versions, but the two boxes do not state the same kind of condition: the first box's body text is two conditions together — average risk, plus being asymptomatic (「無症狀」 appears twice in the Chinese body text, and asymptomatic twice in the English); the second box's body text states risk only, and the whole sentence never mentions symptoms. So what separates the two boxes is the level of risk, and on that the Chinese and English agree; but it cannot be said that neither box's body text has anything to do with symptoms — the first box's body text expressly writes "asymptomatic" into its conditions.

   That distinction is very practical for the reader. The first box's sentence about screening not being recommended is, on its own words, about people at average risk and without symptoms; a person with symptoms is outside the scope of that sentence to begin with. And the page does not set down a route for symptomatic people in the "screening" section — it deals with them in the section immediately following, 「如何檢查和診斷胃癌？」, on investigating and diagnosing stomach cancer, whose first sentence is 「出現上述徵狀的人士，應向醫生求診，儘快接受檢查。」 — anyone with the signs above should see a doctor and be investigated as soon as possible. Which is to say that in the page's own structure, symptomatic people go to the "diagnosis" section, not the "screening" section.

   Why both boxes must nevertheless be printed: print only the first, and a Hong Kong reader with a family history of stomach cancer, or whose report records a precancerous change, will take the official position to be "no need to investigate". The second box's body text in both languages states in terms that these people may consider discussing with a doctor whether and how to be screened — the wording is "may consider … seeking a doctor's advice", not "should be screened". One more thing has to be given with it, the first box's second sentence: not only stomach cancer screening, but Helicobacter pylori screening is stated in the same box not to be recommended in asymptomatic people in the general population. That sentence is particularly easy to skip in Hong Kong, because Helicobacter pylori testing is very common in commercial body-check packages.

3. Whether it can actually be had depends on referral and waiting, and that leaflet does not address it.

4. Section 1.3 of NICE CG184 does not tell a doctor to endoscope for reflux, but it is not a blank section. It has four recommendations: 1.3.1, that severe acute gastrointestinal bleeding calls for same-day specialist referral; 1.3.2, that the doctor should review whether medication is itself causing the dyspepsia (the examples given include calcium channel blockers, bisphosphonates and non-steroidal anti-inflammatory painkillers), and its second half is an instruction to act: for a person who needs referral, NICE tells the doctor to suspend the non-steroidal anti-inflammatory painkiller; 1.3.3, that the doctor should think of cardiac or biliary disease as a differential diagnosis; and 1.3.4, which is written for people who have already had a gastroscopy. [Note 31]

   That sentence in 1.3.2 runs in exactly the same direction as the Department of Health Indigestion box above, and is more specific. But note that it is written for the doctor: it is about the doctor suspending the drug at the point of referral, not an instruction to the reader to stop a drug on their own — many people are taking a painkiller for something else, and stopping it by yourself carries its own risks.

   Recommendation 1.3.3 is in the reflux guideline itself, not in the cardiac guideline. Which is to say: even the UK guideline devoted to dyspepsia and reflux tells the doctor, in its section on when to endoscope, to remember the heart.

   And 1.3.4 is the only one of the four written directly for "people who have had a gastroscopy". Many people ask "I had one a few years ago, do I need another?" — NICE's answer is neither "of course you do" nor "once done, never again", but to carry on managing on the basis of the previous result, on condition that there are no new alarm signs. That condition is the point: the word "new" in "no new alarm signs" means something that has appeared since the last examination. So the red flag section above is not a one-off — if dysphagia, haematemesis, weight loss, anaemia or an upper abdominal mass has appeared since the last examination, 1.3.4 does not apply. And after 1.3.4 NICE itself prints a signpost to NG12 for the definition of "alarm signs" — CG184 does not define "alarm signs" for itself; it hands the whole question to NG12. So the join between the two sections of this article is not one this article has forced; it is printed by NICE.

The Chinese and English versions do not line up with each other either, but not all in one direction. Two different lists have to be dealt with separately here. First, the indications sentence. After its string of indications the Chinese version writes 「等情況」, and the corresponding English sentence ends with no open marker at all. On that sentence alone, the Chinese marks the list as open and the English reads as closed. Second, the complications list — and this one runs the opposite way. The English likewise states that the list 「is not exhaustive」, and adds a whole sentence the Chinese does not have: 「Other unforeseen complications may occasionally occur.」 That is to say, on the complications list both language versions mark it as open, and the English marks it more widely than the Chinese. [Note 32]

⚠️ So no general rule of the form "the Chinese is open and the English closed" can be applied to this leaflet. The indications sentence runs one way and the complications list the other; inferring from the open marker of one list to the other does not hold.

#What to prepare before a gastroscopy

The leaflet's "preparation before the examination" box runs to six sentences in all, every one of them a direct instruction to the patient, so the notes reproduce all six without selecting. [Note 33] Three of the six bear directly on other sections of this article.

One, the sentence about 「特別是某些影響凝血的藥物」 — a Chinese-English divergence, and this time the English is wider. The English reads 「especially antiplatelet and anticoagulation drugs」, while the Chinese says only "certain drugs affecting clotting". The English names antiplatelet and anticoagulant drugs directly; the Chinese names neither. For a reader taking clopidogrel or aspirin, the English version makes it clear this is about them and the Chinese version may not. This leaflet tells that group in terms to declare those drugs before a gastroscopy.

Two, the sentence about driving is a sentence in two halves, and the second half is about alcohol and smoking. The Chinese original separates the halves with a semicolon: the first, that outpatients should not drive themselves to the examination; the second, that they should also avoid drinking heavily, smoking or taking an inappropriate dose of a sedative before the examination. Which is to say that the Hong Kong government, in this leaflet, tells patients in terms to avoid heavy drinking and smoking before a gastroscopy. That sentence has to be read whole, past the semicolon: read only the first half and it looks like a traffic instruction. As to the Chinese-English difference in the first half: the Chinese "should not drive themselves here" is about the journey to the examination, applies to all outpatients, and has nothing to do with whether a sedative is given; the English is softer. The "no driving on the day if you have had a sedative" below is a restriction after the examination — two different things, and both count.

Three, the second sentence tells people with other diseases to declare them and 「聽從其指示服藥」 — to take their medicines as instructed. Diabetes, high blood pressure, valvular heart disease, pregnancy — the leaflet names each. It is not "do not take your medicines on the day of the gastroscopy"; it is "tell the staff, and then take your medicines as they instruct", which is not the same thing.

(The remaining three sentences — six hours nil by mouth, that an elderly or infirm person should be accompanied by a family member, and that dentures, glasses and metal ornaments must be removed — are practical instructions, equally part of the whole passage, and are reproduced with the rest in the notes.)

#The risks of the gastroscopy itself, with Hong Kong's own figures

The Hospital Authority leaflet states: the more serious complications include perforation of the gut (less than one ten-thousandth), bleeding (less than three ten-thousandths) and death (less than one ten-thousandth), with cardiopulmonary complications and infection also possible but with no figure given. The last sentence belongs to the same paragraph as those figures — the leaflet does not separate them: if a complication occurs the patient may need surgery to repair it, and in a severe case it can lead to the patient's death. The figures are about probability and the last sentence is about consequence. [Note 34]

⚠️ And the Chinese and English versions of that sentence do not line up either, this time with the Chinese the wider. The corresponding English sentence has the two qualifiers 「major」 and 「rarely」 and the Chinese has neither. Someone reading the Chinese version comes away with a graver impression than someone reading the English. That direction is the opposite of the indications sentence above and the opposite again of the complications note — so there is no single answer to "which language version is wider" for this leaflet; it has to be taken sentence by sentence.

The leaflet also notes that at least 6 hours of fasting is required before the examination, and that minor discomfort such as nausea, bloating and slight sore throat is common and usually settles within a day; the Chinese version adds that the throat anaesthetic causes 「一個多小時的吞嚥困難」 — difficulty swallowing for an hour or so — mostly resolving within 24 hours.

⚠️ The "after the examination" box runs to two sentences, and it is the first that bears on safety: the patient must fast for a further hour after the examination, until the throat anaesthetic or sedative has worn off, to avoid choking. [Note 35] That is the sentence in the whole leaflet most easily broken in practice. You have been fasting all morning, someone has come to collect you, the family has food waiting — but while the throat anaesthetic has not worn off the swallowing reflex has not fully returned, and the risk the leaflet names is choking. Note that the Chinese and English are written differently: the Chinese says an hour outright, and the English says until the anaesthetic wears off, which 「約持續一小時」 — lasts about an hour. Both point at the same thing: it is not a matter of watching the clock but of whether the anaesthetic has worn off.

The second sentence is the same-day restriction after a sedative, which is particularly practical with a private package: after a sedative, for the whole day you must not operate heavy machinery, sign legal documents or drive. The three private hospital quotations below are all intravenous sedation packages — that is, no driving and no signing documents on the day. Allow for that when arranging transport and leave.

#What it costs

On the public side (Hospital Authority charges, effective 1 January 2026): for eligible persons, a day procedure and treatment at a day healthcare facility costs 250 dollars per attendance; for non-eligible persons the same item is 7,400 dollars per attendance. A specialist clinic is 250 dollars per attendance and 20 dollars per drug item; a family medicine clinic is 150 dollars per attendance and 5 dollars per drug item.

On the private side, the point is not which is cheapest but whether those prices include the same things:

• Union Hospital's day endoscopy price list (document effective 1 July 2025) gives OGD (intravenous sedation) at 4,200 dollars, but the same document states that these are minimum charges, excluding the consultation fees before and after, the doctor's fee, the anaesthetist's fee and pathology charges.
• CUHK Medical Centre's endoscopy packages (charges effective 1 November 2025): gastroscopy (diagnostic) (sedation) at 10,550 dollars and 13,350 dollars — the two figures being the day procedure and inpatient class one respectively, the package including the doctor's fee.
• Gleneagles Hospital's all-inclusive package price list (effective 14 April 2025): gastroscopy (intravenous sedation), excluding polypectomy and biopsy, day case 11,100 dollars (average risk) and 14,430 dollars (moderate risk).

Do the arithmetic once: where the difference is. Union's 4,200 dollars and CUHK Medical Centre's 10,550 dollars differ by 6,350 dollars. Those 6,350 dollars are not "pay more, get better"; they are a difference in what is inside the package. The former is a minimum charge before the doctor's fee, the anaesthetist's fee and pathology; the latter is a package price with the doctor's fee inside it. To compare them you have to add back to the same baseline: on top of Union's 4,200 dollars come the consultation fees, the doctor's fee, the anaesthetist's fee and the pathology charges, and the amounts of those are not listed in that price list, so the final total cannot be worked out from that document at all.

Who this section is most use to: anyone holding a quotation and wondering whether it is reasonable. The first thing to ask is not the number but whether the price includes the doctor's fee, the anaesthetist's fee and the laboratory charges.

---

#How safe are PPIs? The evidence on both sides has to be read together

Observational studies have found a great many associations; a randomised controlled trial in 17,598 people found no difference in anything except intestinal infection. Both are true, and the difference is in the study design — and ACG states for itself that even a randomised controlled trial of that size cannot exclude the possibility of what its own words call 「a modest risk」.

#The first side: the associations found in observational studies

Chronic kidney disease — Lazarus et al 2016 (JAMA Intern Med, PMID 26752337): among 10,482 participants in the ARIC study (mean age 63.0, 43.9% male), the adjusted hazard ratio for incident chronic kidney disease in PPI users against non-users was 1.50 (95% confidence interval 1.14–1.96); and in a replication cohort of 248,751 people in the Geisinger health system, a time-varying new-user design gave an adjusted hazard ratio of 1.24 (95% confidence interval 1.20–1.28). The hazard ratio for twice-daily dosing (1.46, 95% CI 1.28–1.67) was higher than for once daily (1.15, 95% CI 1.09–1.21). The study records its own confounders: compared with non-users, PPI users were more likely to be white, obese and taking antihypertensive drugs.

Dementia — Gomm et al 2016 (JAMA Neurol 2016;73(4):410–6, PMID 26882076): among 73,679 participants aged 75 or over without dementia at baseline, the 2,950 taking a PPI regularly had a hazard ratio for incident dementia of 1.44 against the 70,729 who did not (95% confidence interval 1.36–1.52, P<.001). [Note 36]

That study's conclusion paragraph has to be read whole rather than by its first sentence alone — the authors themselves say that a randomised prospective trial is needed to test causality. Three years later a randomised trial in 17,598 people was in fact done, and it is the one below. So this is not "one study knocked down by another" but a question an observational study raised for itself, later answered by a randomised trial. Quoting only the first sentence and not the last makes the study look more certain than it claims to be. [Note 36]

#The second side: the randomised controlled trial

The COMPASS trial (Moayyedi et al 2019, Gastroenterology, PMID 31152740, ClinicalTrials.gov NCT01776424) was a 3 × 2 partial factorial double-blind trial in which 17,598 participants were randomised to pantoprazole or placebo, with a median follow-up of three years. [Note 37] The trial also recorded roughly twice as much Clostridioides difficile infection in the PPI arm, but there were only 13 events in the whole trial and the difference was not statistically significant.

#The two sides together: ACG's table 5

Several further rows in the table carry no RCT figure at all: cardiovascular events in patients taking clopidogrel, kidney disease (including acute interstitial nephritis), small intestinal bacterial overgrowth, and spontaneous bacterial peritonitis in patients with cirrhosis.

⚠️ And there is a footnote under that table, the most practical sentence in the whole of it, easily skipped along with the table itself: the FDA advises avoiding the use of clopidogrel together with omeprazole. The mechanism column of the same row in the table gives the reason: PPIs are metabolised by the same enzyme (CYP2C19) needed to activate clopidogrel, so using them together may weaken clopidogrel's antiplatelet effect. But ACG's own recommendation immediately afterwards is the other half, and the two have to be given together, or it becomes an instruction to stop a drug by yourself: for GERD patients on clopidogrel who have LA grade C or D oesophagitis, or whose symptoms cannot be controlled on other drugs, the established benefit of the PPI outweighs a cardiovascular risk the original calls 「highly questionable」. [Note 38]

In one sentence: what the FDA addresses is the specific combination of clopidogrel with omeprazole; what ACG addresses is the group who already need a PPI. The two do not conflict — one is about which to choose and the other about whether to use one at all. For someone who has had angioplasty and a stent and is also on a stomach drug, the two together make a question to take to the next appointment: "which PPI am I on?" Incidentally, omeprazole does not appear at all in version 22.1 of the Hospital Authority Drug Formulary, so someone collecting drugs in the public system will mostly not run into the combination the FDA names; but anyone buying privately or over the counter should ask.

The most important sentence about that table: only one row in the whole of it has a 95% confidence interval that excludes 1.0 — enteric infection, odds ratio 1.33 (1.01–1.75). Put differently, within the duration and size of this randomised trial, no statistically significant increase was found in any other putative harm, dementia, fracture and kidney disease included. And for gastric cancer, B12 deficiency and hypomagnesaemia the table carries no RCT figure at all — so "already shown to be fine" cannot be said of those three.

ACG's own reading of COMPASS, including the limits it places on how reassuring it is and the passage recommending how a doctor should put it to a patient, is reproduced in the notes. [Note 39]

Having dealt with the putative serious harms, the genuinely common side effects have to be given too — ACG lists them at the head of the same section, and this is the sentence describing what most people actually meet when taking the drug: the PPI side effects identified in clinical trials include headache, diarrhoea and constipation, and ACG states that side effects of that kind may be related to the particular preparation, so that switching to another is a reasonable option. [Note 40] Whether to switch, and to what, is for the doctor.

#Should magnesium and B12 be checked regularly?

This is one of the specific questions readers most often ask, and the two things have to be answered separately: on serum magnesium there is no consensus at all in the current literature, while on vitamin B12 it is not a matter of no consensus — ACG has written on it in terms.

The United States Food and Drug Administration (FDA)'s drug safety communication of March 2011 opens by stating that prescription PPIs taken long term (longer than a year in most cases) may cause hypomagnesaemia; and that in the cases it reviewed, about a quarter were not manageable by magnesium supplementation alone and required the PPI to be stopped as well. [Note 41] That "about a quarter" matters, because it changes the character of the whole thing: low magnesium is not a small problem that a magnesium tablet settles.

⚠️ But the "longer than a year" span is given a lower bound by the same communication in its data summary section, and that lower bound is where most Hong Kong readers actually are: hypomagnesaemia has been reported in adult patients after at least three months of PPI treatment, though most cases occurred after a year of treatment. [Note 41] "A year" is not a threshold — that sentence says which side most cases fall on, and the same sentence also states that reports start from three months. So "I have only been on it six months" is not a reason to exclude low magnesium for yourself.

Only then comes the monitoring advice: healthcare professionals should consider checking serum magnesium before starting long-term treatment and periodically during it; and for patients taking digoxin (a heart drug) the FDA adds the words "especially important", on the ground that low magnesium makes digoxin's serious side effects more likely. [Note 41] Anyone on digoxin as well as a PPI is in a box the FDA names in terms — which is something to raise with a doctor, not a reason to add tests or stop drugs by yourself.

ACG 2022 records a three-way divergence: the FDA says consider checking, AGA says do not check routinely, and ACG says the data are insufficient to say. [Note 42] Note that all three of those are about serum magnesium. Not one of those three passages mentions vitamin B12 anywhere. But B12 is not unaddressed — another sentence in the same ACG guideline deals with it directly, and its direction is "do not check": for people on a PPI without other risk factors for B12 deficiency, ACG writes that increased vitamin B12 intake is not recommended and neither is routine monitoring of serum B12. [Note 43]

⚠️ That condition has to be read with it: "without other risk factors for B12 deficiency". ACG's sentence speaks only of the group with no other risk factors; what someone with other risk factors for B12 deficiency should do is not stated in that sentence. So any statement that "on a PPI for more than a year you should have magnesium and B12 checked regularly" turns a question without consensus into a consensus.

⚠️ On symptoms, the FDA's list of symptoms is immediately followed by a limitation, and that limitation is the crux of the whole "should it be checked" question: 「patients do not always have these symptoms」. [Note 44] That directly overturns a very common line of reasoning: "I have no cramps and no palpitations, so I do not need my magnesium checked". On the FDA's own account, having no symptoms does not mean the serum magnesium is normal. That does not conflict with the three-way divergence above: what the three are arguing about is whether to check routinely, and none of them says that no symptoms means nothing is wrong. The same FDA communication also has an instruction written directly to patients: do not stop a prescription PPI on your own without discussing it with a healthcare professional. [Note 44]

The FDA also distinguishes prescription from over-the-counter: over-the-counter PPIs are low dose and designed for courses of 14 days at a time, at most 3 courses a year, and when used as labelled the FDA considers the risk of hypomagnesaemia to be very small. ⚠️ That "at most 3 courses a year" has a further condition immediately after it, easily missed — and missing it turns the sentence into "three courses a year, arranged however you like": the over-the-counter PPI is labelled for a 14-day course, and that course may only be repeated once every 4 months. [Note 45]

Which is to say that "over-the-counter PPIs are very safe" is tied tightly to the pattern of use the FDA describes — 14 days at a time, repeated at most once every 4 months, at most 3 courses a year. Buying one at a Hong Kong pharmacy and then taking it for months on end is not covered by that reassurance — the FDA tells healthcare professionals in terms to be alert to the risk of low magnesium in exactly that situation. And one more thing that is very practical for anyone who buys over the counter: precisely because the risk when used as labelled is very small, the FDA decided not to add the risk of hypomagnesaemia to the Drug Facts label of over-the-counter PPIs. [Note 45] That is, the box you pick up in the pharmacy will not carry that warning on it — and the absence of a warning on the box does not mean the risk of long-term use is absent.

#Which PPI do you actually get in Hong Kong?

In the Hospital Authority Drug Formulary (version 22.1, effective 25 July 2026):

• PANTOPRAZOLE and RABEPRAZOLE are listed as general drugs, which on the formulary's own definition means they are provided at standard charges in public hospitals and clinics — that is, the patient pays 20 dollars per drug item at a specialist clinic and 5 dollars per drug item at a family medicine clinic, not the cost of the drug.
• DEXLANSOPRAZOLE, ESOMEPRAZOLE, LANSOPRAZOLE and VONOPRAZAN are in the special drugs column. The definition of that column has three sentences and all three have to be read: special drugs are for use under specific clinical conditions with specific authorisation by a specialist; when prescribed under the specified clinical conditions they are provided at standard charges in public hospitals and clinics; and patients who choose a special drug without meeting the specified clinical conditions pay for it themselves. That is: a special drug does not necessarily mean paying for it yourself — where the conditions are met and a specialist prescribes it, it is the same 20 dollars or 5 dollars per drug item. Quoting only the last sentence frightens off the people who do qualify.
• PANTOPRAZOLE appears in both the general and the special columns. The formulary page lists no route or indication restriction, so why it is listed twice cannot be seen from this source, and this article does not speculate.
• OMEPRAZOLE does not appear anywhere in the formulary at all. It is the one the public knows best, and the one the FDA names above for its interaction with clopidogrel, but it does not exist on version 22.1 of the Hospital Authority Drug Formulary.
• The same formulary has two more boxes beside the PPIs, holding things readers will actually be given. The general drugs column of "ANTISECRETORY DRUGS AND MUCOSAL PROTECTANTS" contains, besides PANTOPRAZOLE and RABEPRAZOLE, FAMOTIDINE (an H2 receptor antagonist), BISMUTH SUBSALICYLATE, MISOPROSTOL and SUCRALFATE. And another box is headed "DYSPEPSIA AND GASTRO-OESOPHAGEAL REFLUX DISEASE", whose general drugs column is entirely antacids and alginates: ALUMINIUM/MAGNESIUM HYDROXIDE AND SIMETHICONE, ALUMINIUM HYDROXIDE, CHARCOAL ACTIVATED, DIMETHYLPOLYSILOXANE, GAVISCON (OR EQUIV), MAGNESIUM TRISILICATE and MYLANTA (OR EQUIV). That box matters, because NICE CG184 1.5.2 below tells long-term users to go back to self-management with an antacid or alginate — and in Hong Kong those things are in the general drugs column of the same formulary, not something to be paid for.

There is one more point about FAMOTIDINE where the two sides are written in opposite directions, and it is worth knowing: NICE CG184 has two recommendations about it, worded identically — 1.4.6 (uninvestigated dyspepsia) and 1.6.5 (GORD) both state that where the response to a PPI is inadequate, H2 receptor antagonist therapy should be offered, the verb being "Offer", which in NICE's usage is one of the harder tiers. But the American side writes the same thing the other way: ACG has two "strong" recommendations stating that for healing and maintaining healing of erosive oesophagitis, a PPI is recommended over an H2RA. The two do not conflict — NICE is speaking about adding or switching after an inadequate PPI response, and ACG about which to choose at the start; which is used is for the doctor.

Do the arithmetic once: a year of PPI in the public system, consultation plus drug. On the Hospital Authority charges effective 1 January 2026: a family medicine clinic is 150 dollars per attendance and 5 dollars per drug item. Assume 4 follow-ups a year with one drug item dispensed each time: 4 × (150 + 5) = 620 dollars. The same set of follow-ups at a specialist clinic (250 dollars per attendance, 20 dollars per drug item): 4 × (250 + 20) = 1,080 dollars. Both are far below the ten thousand dollar annual cap on public healthcare charges for the same year, so for most people who are simply on a drug and attending follow-ups the cap will never come into play; what actually makes the total jump is investigations and procedures, not drugs.

Who this section is most use to: anyone who has read a headline saying PPIs increase the risk of dementia and stopped the drug themselves. Table 5 shows dementia in the RCT at an odds ratio of 1.20 (0.81–1.78), a confidence interval including 1.0; but the same table shows no RCT figure at all for gastric cancer, B12 or low magnesium. Being informed means knowing both directions.

---

#After some years on it, can it be stopped?

AGA's de-prescribing advice of 2022 is a numbered document devoted to reducing and stopping, and all ten of its best practice statements are in the table below. Its statement 9 says plainly that tapering the dose and stopping outright may both be considered. But it is not the only document saying who should not stop — NICE CG184 has two recommendations on that too.

Those ten are all ten of the statements that document prints, unabridged. But note that statement 4 uses "such as" — the list of complicated reflux disease histories is itself an example and not exhaustive. So the inference "I am not in the situations listed at 4, 5 and 7, so I can stop" does not hold: the list is open, and statement 6 requires the bleeding risk to be assessed professionally. That document's guiding principle is reproduced in the notes as well. [Note 46]

ACG has two recommendations of its own about stopping, running in the same direction as AGA, and one of them offers a third option — "on demand". [Note 46] ⚠️ Note the second half of the second one: the short version in ACG's summary table stops at "an attempt should be made to stop the PPI", while the body version has a further option, reading 「or to switch to on-demand therapy」 — so it is not a choice between taking and stopping, with a route in between called taking it on demand. ACG has another recommendation saying so directly: for people with non-erosive reflux disease, on-demand or intermittent PPI is recommended to control heartburn. As noted above, Asia is predominantly NERD — so this option applies to a wider group of Hong Kong readers than the wording on its face suggests.

NICE CG184 also has two recommendations on who should stay on it long term, in the same direction as AGA's statements 4, 5 and 7: people who have had an oesophageal stricture dilated should stay on long-term full-dose PPI (1.6.6); and people with severe oesophagitis should be offered long-term full-dose maintenance treatment (1.6.9). [Note 47]

#Is the "rebound" real or not? All three positions have to be set out

That question is the most practical worry at the point of stopping, and the three sources do not agree:

1. The original trial (Reimer et al 2009, Gastroenterology, PMID 19362552): 120 healthy volunteers were randomised to 12 weeks of placebo, or to 8 weeks of esomeprazole 40 mg followed by 4 weeks of placebo. During weeks 9 to 12, 44% of the PPI arm (26/59) reported at least one relevant acid-related symptom against 15% of the placebo arm (9/59), P<.001. The denominator is the point: these were healthy volunteers, not GERD patients. And that trial's own concluding sentence follows immediately on the figure: that the study reveals an unrecognised aspect of PPI withdrawal and supports the hypothesis that rebound acid hypersecretion is clinically significant. The authors themselves consider it clinically significant — so ACG's contrary position below is a genuine divergence.
2. AGA statement 8: patients should be told that transient rebound symptoms may occur.
3. ACG 2022: although rebound acid hypersecretion has been demonstrated in healthy controls, strong evidence that symptoms increase after a PPI is stopped abruptly is still lacking.

In one sentence: the physiological rebound can be demonstrated in healthy people; whether patients' symptoms actually worsen after stopping is, for ACG, insufficiently evidenced, and for AGA worth warning about in advance. Neither position supports "you must always come off slowly" — AGA's statement 9 says in terms that either way may be considered.

#Use the lowest effective dose: the UK and the US agree

ACG 2022 states: for GERD patients who need maintenance treatment, the lowest dose that effectively controls symptoms should be used; and another recommendation is that the PPI should be taken 30 to 60 minutes before a meal. [Note 48]

The NICE CG184 side is a ladder — in fact two parallel ladders, which have to be looked at separately. [Note 49]

The "dyspepsia" ladder: 1.5.1 — offer people who need long-term management of dyspepsia an annual review of treatment, and encourage them to step down or stop; 1.5.2 — tell patients that where appropriate they can self-manage with "on-demand" use, with antacids and alginates a step further down; and 1.2.5 sets down the order of the whole stepping-down ladder.

The GORD ladder (section 1.6, "Interventions for GORD"): a full-dose course first (4 weeks or 8 weeks, two lengths and not a range, 1.6.2) → on relapse, step down to the lowest effective dose → then teach the patient to take it only as needed → and where the PPI response is inadequate, the branch of adding an H2 receptor antagonist. Note that all four use "Offer" or "Discuss" and not "Consider" — a harder tier in NICE's usage.

The same section has three recommendations on severe oesophagitis, and all three have to be read together — read only the middle one (1.6.9, long-term maintenance) and both the first and the last are missed: 1.6.7 is about healing severe oesophagitis with a full-dose PPI for 8 weeks first; 1.6.9 is about long-term maintenance after healing; and 1.6.10 is about maintenance treatment itself failing — the original tells the doctor to conduct a clinical review and consider switching to another full-dose or high-dose PPI, and/or seeking specialist advice. Only all three together are complete: heal → maintain → review after failure. For someone on a maintenance dose whose symptoms have come back, 1.6.10 is their box, and the instruction in that box is not "increase the dose" but a fresh clinical review. (1.6.8 is about switching after initial healing fails, equally at the doctor's level.)

So what the UK guideline sets out is not a choice between taking and stopping but four tiers: full dose → lowest effective dose → on demand → antacid or alginate. As the formulary section above noted, that last tier sits in the general drugs column of the Hospital Authority Drug Formulary. Whether and when to step down is still for the doctor to decide on your situation; all that is set out here is how many tiers the ladder has.

But LA grade C or D oesophagitis is the exception, and ACG's recommendation for that group is indefinite maintenance PPI or antireflux surgery (see the mechanism section above; that one is a "strong" recommendation).

Who this section is most use to: anyone who has been on a PPI for many years, has not been scoped since, and wants to cut down on their own. The crucial thing here is that whether you fall into the groups at AGA's statements 4, 5 and 7 turns on your grade of oesophagitis and your bleeding risk — and those two things have answers only in an examination report and a doctor's assessment, not in how the symptoms feel.

---

#Does lifestyle really help? What has a source and what does not

ACG has five lifestyle recommendations; but the specific figure that circulates — how many centimetres to raise the head of the bed — is given by neither guideline. As for which side to sleep on, ACG's recommendations do not contain that item, but its body text does discuss it, and the two things have to be kept apart.

The five lifestyle recommendations in ACG 2022's "GERD Medical Management" section are reproduced in full in the notes. [Note 50] Only one of the five, weight loss, is a "strong" recommendation; the other four are all "conditional, low evidence". That distinction is not a detail: ACG states in its own body text that the data supporting recommendations of this kind are 「limited and variable」, mostly small uncontrolled studies.

NICE CG184's opening in this box, 1.2.1, is a general statement: offer simple lifestyle advice, including advice on healthy eating, weight reduction and smoking cessation. The original introduces it with "including", so it is an example; and weight reduction and smoking cessation map back onto weight loss (the first) and smoking cessation (the third) among ACG's five. The recommendation immediately after it, 1.2.2, is the one that lists specific precipitants, and is written more guardedly — advise patients to avoid known precipitants they associate with their dyspepsia, such as strong coffee, alcohol, chocolate and fatty foods; and NICE's words are "may help some people", not a general rule. [Note 51]

⚠️ That box does not stop at 1.2.2. Recommendation 1.2.4 is not about food but about an entirely different kind of management: psychological therapies (cognitive behavioural therapy and psychodynamic therapy, for instance) may reduce dyspeptic symptoms in the short term in individual people. As the Hong Kong data section above noted, Mak 2019's local survey of 2,011 people concluded for itself that symptom frequency has a dose-response relationship with anxiety and depression. That NICE recommendation maps onto the local finding — that is, "managing reflux" is not only a matter of food and drugs. But it has to be read accurately: the three qualifiers "may", "short term" and "individual people" cannot be dropped, and this is not a general claim of efficacy. [Note 51]

Two things in ACG's body text are not in its recommendations. First, ACG names alcohol as NICE does (the original reads 「tobacco and alcohol cessation」) — it is not only NICE that says so. Second, staying upright during and after meals (the original's 「staying upright during and after meals」) appears nowhere in the five recommendations, but it is the same thing as the Department of Health Indigestion page's item about not lying down immediately after eating below. That sentence is introduced with "include" and "such as", so it is an example. [Note 52]

Neither guideline says how high to raise it. So the "15 to 20 centimetres" commonly quoted has no source in either of these guidelines — not that it is wrong, but that neither guideline has ever given that figure. ⚠️ But "not saying how high" is not the same as "not much evidence". Raising the head of the bed is rated "conditional, low evidence" in the recommendations, but the body text of the same guideline states that the studies supporting it include 「including several randomized controlled trials」. A low rating does not mean no trials — ACG simply chose to write it with a low rating. As for the height, that same body passage gives no dimension either, saying only to raise the head of the bed or sleep on a wedge. [Note 52]

As for "the left side is better than the right" — that has to be taken in two layers. It is not among ACG's five recommendations, so it is not a GRADE-rated recommendation. But it cannot simply be treated as description either: in the sentence immediately following, ACG writes an action — patients may be advised to avoid sleeping on the right side. [Note 52] That is: ACG did not promote it to a numbered recommendation, but it did say in the body that patients may be advised to do it. Both things have to be said: it carries less weight than the five recommendations, but it is not that nothing was said about what to do.

On the Hong Kong government side. The passage on the Department of Health Elderly Health Service's Indigestion page is headed with tips on preventing and relieving indigestion, and has four boxes below it, of which dietary adjustment is only the second. [Note 53]

The second box, on dietary adjustment, lists eleven items, five of which coincide with the two guidelines above: avoid high-fat foods such as fried, deep-fried and greasy food; avoid foods and drinks containing caffeine, such as chocolate, coffee and tea; avoid wearing clothes that are too tight; eat only to seven tenths of fullness; and do not lie down immediately after eating, to prevent food refluxing into the oesophagus. High-fat food and coffee and chocolate are each named in NICE CG184 1.2.2; "eat only to seven tenths of fullness" runs in the same direction as ACG's "do not eat within 2 to 3 hours of bedtime" but is not the same rule (one is about quantity, the other about timing) and the two are not to be treated as one. And "do not lie down immediately after eating" maps onto the "staying upright during and after meals" of ACG's body text — that item is not in ACG's five recommendations, only in the body, so it takes this Department of Health page and ACG's body text together to see that it has a source. As for "avoid wearing clothes that are too tight", neither the ACG nor the NICE guideline cited here has that item — it appears only on this Department of Health page. (Not all eleven items in that box concern reflux, so only the ones that coincide are taken here.)

⚠️ Reading only the second box misses the first and the third — and several items in those two are exactly the things this article's other sections consider to carry most weight. The first box, on healthy living, has four lines: keep regular hours and take suitable exercise; do not smoke, or stop smoking; avoid alcohol. Those two deserve separate mention, because they directly overturn an impression that is easy to form — reading straight through, a reader easily assumes that stopping smoking and drinking is something only the foreign guidelines say, and that the Hong Kong government side talks only about food. It is not. Stopping smoking and avoiding alcohol are printed on the same Department of Health page, in the first of the same four boxes of tips.

What they map onto: smoking cessation is at once the third of ACG's five lifestyle recommendations, NICE CG184 1.2.1 (which names smoking cessation in the general advice) and 1.2.2 (the first item in its list of precipitants); and alcohol is NICE CG184 1.2.2 and ACG's body text. That is to say, tobacco and alcohol are the only two lifestyle items in this article's four sources that all four of them mention — and the Hong Kong government side not only mentions them but puts them at the very front of its four boxes of tips. Where the same page explains why indigestion happens, tobacco and alcohol appear again, this time as causes: poor lifestyle habits including bingeing, irregular meals, smoking and heavy drinking. [Note 53]

The third box, on managing emotions and stress, should not be skipped either, because it maps onto the local research above: understanding the process of ageing and one's own bodily function can reduce unnecessary worry; and practising relaxation of body and mind helps ease tension. [Note 53] That, NICE CG184 1.2.4 on psychological therapy, and Mak 2019's local finding are three sources running in the same direction.

Who this section is most use to: anyone who wants to start with lifestyle and does not want to be led astray by what is online. Five items are supported by an ACG recommendation, of which only weight loss is rated "strong"; tobacco and alcohol are the two that all four sources mention; and how many centimetres to raise the head of the bed is a figure with no source.

---

#Barrett's oesophagus and oesophageal cancer: how large is the risk really?

The annual risk of progression printed in ACG's Barrett's oesophagus guideline (2022) is given by grade of dysplasia: 0.2%–0.05% a year for non-dysplastic Barrett's oesophagus (reproduced as printed, including the reversed range as the original prints it), and about 0.7% a year for low-grade dysplasia. [Note 54] Put differently, under the same name "Barrett's oesophagus", the annual risk without dysplasia and with low-grade dysplasia can differ more than tenfold. Any single figure that does not state the grade is incomplete.

The presence or absence of intestinal metaplasia also changes the risk — but that point is contested within ACG's own guideline and cannot be treated as settled. The same guideline cites a population study of 8,522 Barrett's oesophagus patients from the Northern Ireland cancer registry: the risk of oesophageal adenocarcinoma in those with intestinal metaplasia at first endoscopy against those without was 0.38% against 0.07% a year (hazard ratio 3.54, 95% confidence interval 2.09–6.00). ⚠️ But immediately afterwards in the same passage ACG sets down studies running the other way — a single-centre study of 688 people (0.37% against 0.30% a year) and a multicentre study of 1,751 people (hazard ratio 1.36, 95% confidence interval 0.63–2.96) found no difference. [Note 54] That is to say, whether intestinal metaplasia is present is a question still in dispute within that ACG guideline, not a settled tool for stratifying risk.

The difference between the sexes is large. ACG records that the risk of progression from Barrett's oesophagus to oesophageal adenocarcinoma is markedly higher in men than in women (adjusted odds ratio 2.2, 95% confidence interval 1.8–2.5). A modelling study estimated the incidence of oesophageal adenocarcinoma at 3.9 per 100,000 person-years in women aged 60 with weekly GERD symptoms, against 61 per 100,000 person-years in men; and on that basis ACG takes the view that the yield of screening women for Barrett's oesophagus may be very low. [Note 54]

⚠️ But that ACG passage does not stop there — the sentence immediately following is written for women at high risk, and it runs the other way from "the yield may be very low": for women with multiple risk factors, the benefits and harms of screening may be discussed with the patient. [Note 54] Who that sentence matters to: a woman reader who also has several risk factors — obesity, a smoking history, a first-degree relative with Barrett's oesophagus or oesophageal adenocarcinoma — should not close the question off on the strength of the "low yield in women" sentence above. ACG states in terms that screening may be appropriate in that situation, and states that the way to do it is to discuss the benefits and harms with a doctor, not to be scoped automatically or not scoped automatically.

Who is actually recommended one screening endoscopy? ACG's recommendation 5: for patients with chronic GERD symptoms and three or more risk factors (including male sex, age over 40, white race, smoking, obesity, and a first-degree relative with Barrett's oesophagus or oesophageal adenocarcinoma), a single screening endoscopy is recommended. [Note 55] Three things to note: first, it is "3 or more", not any one of them; second, the original introduces the risk factors with "including", so they are examples; and third, it is 「a single screening endoscopy」 — once, not repeated at intervals. ACG also states honestly the fundamental limitation of the whole thing: there is no randomised controlled trial evidence that screening reduces mortality. [Note 55]

The definition itself is still disputed. The Asia-Pacific consensus of 2016 changed the definition of Barrett's oesophagus to require a minimum length of 1 cm and no longer to require intestinal metaplasia; ACG's definition does require intestinal metaplasia, and acknowledges the divergence in its own guideline. The practical consequence: the same Hong Kong patient may be told they do not have Barrett's oesophagus on ACG's definition and that they do on the Asia-Pacific one.

NICE CG184's position on surveillance has to be given with it — and the recommendation immediately before it, 1.6.11, is the one that decides whether to scope at all, so the two have to be read together: do not routinely offer endoscopy to diagnose Barrett's oesophagus, but consider it if the person has GORD, discussing with them their preferences and their individual risk factors (long symptom duration, increasing symptom frequency, previous oesophagitis, previous hiatus hernia, oesophageal stricture or oesophageal ulcers, or male sex, for example). [Note 56] The half-sentence "but consider it if the person has GORD" is precisely the exception to "do not routinely offer", and the group that exception covers is this article's readers. The list of risk factors in the brackets is introduced with "for example" and so is illustrative; and it is not the same as ACG's — ACG requires "3 or more" and NICE sets no number, only telling the doctor to discuss. The two guidelines use different criteria on the same question, so the inference "I do not have 3, so I do not need it" does not hold on the NICE side.

And only then comes the surveillance recommendation. It is not in section 1.6 but in a separate section 1.12 of CG184, which has one recommendation in all, numbered 1.12.1. [Note 56] ⚠️ That one has to be read whole, and not by its last sentence. The sentence about the harms of surveillance possibly outweighing the benefits is not the recommendation itself but a caution at the tail of it; the main clause of NICE's recommendation is "Consider surveillance". Someone who has just received a report, seen "Barrett's oesophagus" and been frightened will, reading only the last sentence, take NICE's position to be that they should not be under surveillance — NICE's position is the opposite: consider surveillance, and then set out the harms plainly.

Who this section is most use to: anyone whose report says "Barrett's oesophagus" and who is frightened by it. What to ask the doctor is: is there dysplasia, of what grade, what length, and when is the next surveillance.

---

#What to do next

Every item below is the stated position of a source already cited above.

1. With chest pain, deal with the heart first and not the stomach. ACG states that reflux chest pain and cardiac chest pain cannot be told apart, and its own order is that GERD is investigated after cardiac disease has been excluded. If you are in pain now, NICE CG95 1.2.1.7 says emergency admission; if the pain has resolved but there are signs of a complication, 1.2.1.9 says referral for hospital assessment; and if the previous episode was already suspected to be cardiac and the pain is back, 1.2.1.11 says emergency referral. In Hong Kong the corresponding route is accident and emergency, at 400 dollars per attendance, with triage categories I and II exempt.
2. Do not draw conclusions from "the stomach drug helped". The systematic review ACG cites shows that the PPI response does not reliably predict the diagnosis. A treatment trial is not a diagnostic test.
3. Leave the red flag symptoms to a doctor rather than ticking a list yourself. The three lists differ in threshold and in items, and dysphagia and haematemesis have to be dealt with at any age under NG12. Not being on the list does not mean it does not matter.
4. If you take painkillers long term, take that to your next appointment. The Department of Health writes that aspirin, painkillers and non-steroidal anti-inflammatory drugs are 「切勿自行服用」 — not to be taken on your own initiative; NICE CG184 1.3.2 tells the doctor to review medication as a cause and to suspend the non-steroidal anti-inflammatory painkiller at referral. That is an instruction to the doctor, not an instruction to you to stop a drug.
5. On whether to have a gastroscopy, take three things to ask about: whether there are alarm symptoms (ACG lists three in one place and five in another), whether eight weeks of a PPI was good enough, and when the last gastroscopy was and whether any alarm sign has appeared since (NICE CG184 1.3.4). If a doctor arranges a gastroscopy and tells you to stop the drug for a few weeks first, that is what the guideline says.
6. Before cutting down or stopping, ask whether you are in one of the "should not stop" groups. AGA's statements 4, 5 and 7 and NICE CG184 1.6.6 and 1.6.9 set those groups out; and the judgment rests on the grade of oesophagitis and the bleeding risk, not on how the symptoms feel. And stopping is not a choice between two states — ACG also writes down taking it on demand, and NICE also writes down the further step down to antacids and alginates.
7. If you take clopidogrel, find out which PPI you are on. What the FDA names is the combination of clopidogrel with omeprazole; and omeprazole does not appear in version 22.1 of the Hospital Authority Drug Formulary. Do not stop the drug yourself — that is the FDA's own instruction to patients.
8. On lifestyle, start with the two things all four sources mention: stop smoking and stop drinking. Weight loss is the only one of ACG's five rated "strong". Raising the head of the bed has a source; how many centimetres does not.

---

#Frequently asked questions

The chest pain went away after a stomach drug — does that prove it was acid reflux?

No. The systematic review ACG cites found that when endoscopy and pH monitoring are normal there is no significant response to a PPI over placebo, and that chest pain and heartburn 「did not reliably predict a PPI response」. A treatment trial is not a diagnostic test. The order is for a doctor to exclude cardiac disease first.

Does 1 adult in every 5 in Hong Kong have acid reflux?

No Hong Kong data supports that. Across three Hong Kong telephone surveys, the proportion with weekly symptoms was 2.5% (2002, 2,209 people) and 4.2% (published 2019, 2,011 people); the 2011 survey (2,074 people) used the Montreal definition and arrived at 3.8%, which is not a weekly proportion. The largest figure is Wong 2003's annual prevalence of 29.8%, which counts anyone who has had it once within a year — not the same thing as regular symptoms.

New symptoms after the age of 40 — should there be a gastroscopy?

Among the referral criteria that can be cited here, not one uses 40 to trigger a gastroscopy for reflux. To be accurate: section 1.2 of NG12 does have one recommendation that uses 40 — 1.2.4, aged 40 or over with jaundice, referral on a suspected cancer pathway for pancreatic cancer (and 1.2.5 in the same section uses 60, with weight loss, for a CT scan). But that one is about jaundice, the pancreas and CT, not reflux and gastroscopy. On the gastroscopy side, NICE NG12's threshold is 55 and weight loss has to be present as well; dysphagia on its own is referred at any age, and haematemesis and a palpable upper abdominal mass also have to be considered at any age (NG12 1.2.2 and 1.2.6). On the other side, the Hospital Authority gastroscopy leaflet states that patients with acid reflux 「均應接受胃鏡檢查」. Both are documents in force, and whether to scope is decided by a doctor on your situation.

A blood test found anaemia, or I vomit often — do those count as alarm symptoms?

The brackets in ACG's recommendation list only three (weight loss, gastrointestinal bleeding), but the body of the same guideline lists five: dysphagia, weight loss, bleeding, vomiting and anaemia, stating that such patients should be endoscoped as soon as feasible. Both places are written openly. And the NICE NG12 side is not silent either: within 1.2.3, aged 55 or over with "upper abdominal pain with a low haemoglobin" stands as a box of its own, and "nausea or vomiting" with any one of weight loss, reflux, dyspepsia or upper abdominal pain stands as another, both of them non-urgent direct access endoscopy. So the answer is: anaemia and vomiting both count in ACG's body listing, and each has a route in NG12 as well, only NG12 adds the age of 55 and a pairing condition. Whether and when to scope is for a doctor.

I take a blood thinner or an antiplatelet drug — will a stomach drug interact?

The footnote to ACG's table 5 states that the FDA advises avoiding the use of clopidogrel together with omeprazole, the mechanism being that both use the enzyme CYP2C19. But the same guideline also states that for patients who have LA grade C or D oesophagitis, or whose symptoms cannot be controlled on other drugs, the established benefit of the PPI outweighs a cardiovascular risk the original calls 「highly questionable」. Do not stop the drug yourself — the patient instruction in the same FDA communication is 「Do not stop taking your prescription PPI drug without talking to your healthcare professional.」 Incidentally, omeprazole does not appear in version 22.1 of the Hospital Authority Drug Formulary.

On a PPI for more than a year — should magnesium and B12 be checked regularly?

The two have to be answered separately. Serum magnesium: there is no consensus at present — the FDA advises considering a check before starting and periodically thereafter; AGA takes the view that long-term users should not be routinely screened or monitored; and ACG takes the view that the data are insufficient to make a recommendation. Vitamin B12: none of the three has ever recommended checking it, and ACG writes in terms that it need not be — for people on a PPI without other risk factors for B12 deficiency, ACG writes that increased vitamin B12 intake is not recommended and neither is routine monitoring of serum B12; what someone with other risk factors for B12 deficiency should do is not stated by ACG. The FDA communication never mentions B12 at all. But none of the three has ever said that no symptoms means nothing is wrong — the same FDA communication states 「patients do not always have these symptoms」, and in the cases it reviewed about a quarter were not manageable by magnesium supplementation alone and required the PPI to be stopped as well. The "one year" is the FDA's own wording, but it has to be read together with the sentence in its data summary: hypomagnesaemia has been reported in adult patients after at least three months of a PPI, though most cases occurred after a year of treatment — a year is not a threshold; three months is the lower bound of what has been reported — so "I have not been on it a year" is not a reason to rule it out for yourself. And people on digoxin are a box the FDA names specifically in that passage: for patients taking digoxin (a heart drug), this is especially important.

Does stopping a PPI have to be done by tapering slowly?

Statement 9 of AGA's 2022 de-prescribing advice states that tapering the dose and stopping outright may both be considered. ACG also writes down a third option: switching to on-demand use (take it when there are symptoms, stop when they settle), and NICE CG184 1.2.5 and 1.5.2 go a step further down, to self-management with antacids and alginates. But statements 4, 5 and 7 of the same AGA document set out several groups who generally should not stop: those with a history of severe erosive oesophagitis, oesophageal ulcer or peptic stricture; those with known Barrett's oesophagus, eosinophilic oesophagitis or idiopathic pulmonary fibrosis; and those at high risk of upper gastrointestinal bleeding. NICE CG184 has two more in the same direction: people who have had an oesophageal stricture dilated should stay on long-term full-dose PPI (1.6.6), and people with severe oesophagitis should be offered long-term full-dose maintenance treatment (1.6.9). The list at statement 4 uses "such as" in the original and is illustrative. Whether to stop is for a doctor.

Do public hospitals have PPIs, and what does a patient pay?

Version 22.1 of the Hospital Authority Drug Formulary (effective 25 July 2026) lists PANTOPRAZOLE and RABEPRAZOLE as general drugs, provided at standard charges — 20 dollars per drug item at a specialist clinic and 5 dollars per drug item at a family medicine clinic. OMEPRAZOLE does not appear anywhere in the formulary.

#Notes: the original texts

[Note 1] The objective definition of acid reflux: the American College of Gastroenterology (ACG) guideline of 2022, and the UK NICE clinical guideline CG184.

「GERD is objectively defined by the presence of characteristic mucosal injury seen at endoscopy and/or abnormal esophageal acid exposure demonstrated on a reflux monitoring study.」

「In this guideline, GORD refers to endoscopically determined oesophagitis or endoscopy-negative reflux disease.」

[Note 2] ACG 2022 on the typical symptoms, with its last sentence about the approach differing according to which symptom predominates.

「Typical symptoms of GERD include heartburn and regurgitation. Heartburn is the most common GERD symptom, and is described as substernal burning sensation rising from the epigastrium up toward the neck. Regurgitation is the effortless return of gastric contents upward toward the mouth, often accompanied by an acid or bitter taste. While both heartburn and regurgitation are major symptoms of GERD, the genesis of these symptoms are not the same, and the diagnostic and management approaches vary depending on which symptom predominates.」

[Note 3] ACG 2022's two lists of extra-oesophageal manifestations — the items on the two lists are not the same, and both are introduced with 「such as」.

「Extraesophageal manifestations of GERD can include laryngeal and pulmonary symptoms such as hoarseness, throat clearing, and chronic cough, and conditions such as laryngitis, pharyngitis, and pulmonary fibrosis. It also has been proposed that GERD might exacerbate asthma. These extraesophageal manifestations are challenging for patients and physicians because, while they may result from GERD, they may also be due to a host of other causes. Even in patients with established GERD, it can be difficult to establish that GERD is the cause of these extraesophageal problems.」

「Atypical and extraesophageal symptoms and conditions such as chronic cough, dysphonia, asthma, sinusitis, laryngitis, and dental erosions have been associated with GERD. However, these symptoms and conditions have poor sensitivity and specificity for the diagnosis of GERD. Diagnoses of GERD by extraesophageal symptoms alone or by their response to PPIs are unreliable due to poor sensitivity and specificity for GERD and not recommended」

[Note 4] ACG 2022's two "do not diagnose this way" recommendations: do not use upper endoscopy to establish an extra-oesophageal diagnosis; and do not diagnose laryngopharyngeal reflux on laryngoscopic findings alone.

「We suggest that upper endoscopy should not be used as the method to establish a diagnosis of GERD-related asthma, chronic cough, or LPR. Low Conditional」

「We suggest against a diagnosis of LPR based on laryngoscopy findings alone and recommend additional testing should be considered. Low Conditional」

[Note 5] The first three positive recommendations of the same ACG 2022 section: evaluate the non-GERD causes first; reflux testing first where there are no typical symptoms; and where both are present, a trial of twice-daily PPI for 8 to 12 weeks may be tried first.

「We recommend evaluation for non-GERD causes in patients with possible extraesophageal manifestations before ascribing symptoms to GERD. Moderate Strong」

「We recommend that patients who have extraesophageal manifestations of GERD without typical GERD symptoms (e.g., heartburn and regurgitation) undergo reflux testing for evaluation before PPI therapy. Moderate Strong」

「For patients who have both extraesophageal and typical GERD symptoms, we suggest considering a trial of twice-daily PPI therapy for 8–12 wk before additional testing. Low Conditional」

[Note 6] ACG 2022 on the pathophysiology of GERD.

「The pathophysiology of GERD includes a poorly functioning esophagogastric junction; the antireflux barrier composed of the lower esophageal sphincter and crural diaphragm, coupled with impaired esophageal clearance and alterations in esophageal mucosal integrity. Reflux esophagitis develops when refluxed gastric juice triggers the release of cytokines and chemokines that attract inflammatory cells, and that also might contribute to symptoms. Other contributors to GERD symptoms may include decreased salivary production, delayed gastric emptying, and esophageal hypersensitivity. As such, GERD can no longer be approached as a single disease, but one with multiple phenotypic presentations and different diagnostic considerations.」

[Note 7] ACG 2022: for Los Angeles grade C or D oesophagitis, indefinite maintenance PPI treatment or antireflux surgery is recommended (a strong recommendation).

「We recommend maintenance PPI therapy indefinitely or antireflux surgery for patients with LA grade C or D esophagitis. Moderate Strong」

[Note 8] The Seoul consensus of 2020: in the setting of typical symptoms, a PPI test can serve as a diagnostic tool.

「In GERD with typical symptoms, a proton pump inhibitor test can be recommended as a sensitive, cost-effective, and practical test for GERD diagnosis. Based on a meta-analysis of 19 estimated acid-exposure time values in Asians, the reference range upper limit for esophageal acid exposure time was 3.2% (95% confidence interval, 2.7-3.9%) in the Asian countries.」

[Note 9] The three sets of thresholds for oesophageal acid exposure time: the 4.0% of the Seoul consensus body text, the ≥ 4% of its statement 13, the 6% of the Lyon consensus and the band between, and the Porto consensus's range of unproven diagnosis.

「Therefore, we propose that more than 4.0% of AET be judged as abnormally high in Asians. However, while this is a reasonable threshold for determining the abnormal range, it may not be the threshold for inducing GERD symptoms.」

「Statement 13: A value of total esophageal acid exposure time of ≥ 4% is defined as an abnormal finding in Asian adults. Level of evidence: moderate Strength of recommendation: weak」

「Lyon consensus proposes that AET < 4% is considered normal (physiological), and > 6% is deemed abnormal and presents the concept of inconclusive with intermediate values between these limits.」

「The Porto consensus suggests that esophageal acid exposure is considered pathological if AET > 6% on pH testing as new concepts of areas of uncertain diagnosis appear (LA classification grade A and B and AET between 4-6%).」

[Note 10] The concluding sentences of the three Hong Kong telephone surveys: Tan 2016's rising trend, Wong 2003's comparison, and the direction recorded by the Asia-Pacific consensus.

「The prevalence of weekly GERD had increased by 1.3% between 2002 and 2011, which represents an at least 50% relative increase (P<0.0005).」

「The prevalence of gastro-oesophageal reflux disease was lower than that in Western populations, but carried a significant socio-economic burden in the studied Chinese population.」

「A rise in the prevalence rates of gastro-oesophageal reflux disease in Asia was noted, with the majority being non-erosive reflux disease. Overweight and obesity contributed to the rise. Proton pump inhibitor-refractory reflux disease was recognised to be common.」

[Note 11] Four passages of ACG 2022 on chest pain: 「indistinguishable」 from cardiac pain, the two wordings in the recommendation and in the body, and the systematic review finding no PPI effect in chest pain.

「Chest pain, indistinguishable from cardiac pain, may present in conjunction with heartburn and regurgitation, or as the only GERD symptom. The symptoms of GERD are nonspecific and may overlap or be confused with those of other disorders such as rumination, achalasia, eosinophilic esophagitis, reflux hypersensitivity, functional disease, cardiac or pulmonary disease, and paraesophageal hernia.」

「In patients with chest pain who have had adequate evaluation to exclude heart disease, objective testing for GERD (endoscopy and/or reflux monitoring) is recommended. Low Conditional」

「In patients who have chest pain without heartburn and who have had adequate evaluation to exclude heart disease, objective testing for GERD (endoscopy and/or reflux monitoring) is recommended. (Conditional recommendation, low level of evidence)」

「There was no significant response to PPIs compared with placebo when endoscopy and pH monitoring were normal, and the symptoms of chest pain and heartburn did not reliably predict a PPI response」

[Note 12] NICE CG95 recommendation 1.2.1.3: the four symptoms that may indicate an acute coronary syndrome, reproduced in full.

「Initially assess people for any of the following symptoms, which may indicate an ACS: pain in the chest and/or other areas (for example, the arms, back or jaw) lasting longer than 15 minutes chest pain associated with nausea and vomiting, marked sweating, breathlessness, or particularly a combination of these chest pain associated with haemodynamic instability new onset chest pain, or abrupt deterioration in previously stable angina, with recurrent chest pain occurring frequently and with little or no exertion, and with episodes often lasting longer than 15 minutes. [2010]」

[Note 13] NICE CG95 recommendation 1.2.1.7: emergency admission, including the box for an electrocardiogram that is abnormal or not available.

「Refer people to hospital as an emergency if an ACS is suspected (see recommendation 1.2.1.3) and: they currently have chest pain or they are currently pain free, but had chest pain in the last 12 hours, and a resting 12-lead ECG is abnormal or not available. [2010]」

(In substance, and as an abridged rendering rather than the original: where an acute coronary syndrome is suspected and the patient currently has chest pain, or is currently pain free but had chest pain within the past 12 hours and the resting twelve-lead electrocardiogram is abnormal or cannot be obtained, refer to hospital as an emergency.)

[Note 14] NICE CG95 recommendations 1.2.1.8, 1.2.1.9 and 1.2.1.10: the three boxes for after the pain has passed, with the three sub-items of 1.2.1.10.

「1.2.1.8 If an ACS is suspected (see recommendation 1.2.1.3) and there are no reasons for emergency referral, refer people for urgent same-day assessment if: they had chest pain in the last 12 hours, but are now pain free with a normal resting 12-lead ECG or the last episode of pain was 12 to 72 hours ago. [2010]」

「1.2.1.9 Refer people for assessment in hospital if an ACS is suspected (see recommendation 1.2.1.3) and: the pain has resolved and there are signs of complications such as pulmonary oedema. Use clinical judgement to decide whether referral should be as an emergency or urgent same-day assessment. [2010]」

「1.2.1.10 If a recent ACS is suspected in people whose last episode of chest pain was more than 72 hours ago and who have no complications such as pulmonary oedema: carry out a detailed clinical assessment (see recommendations 1.2.4.2 and 1.2.4.3) confirm the diagnosis by resting 12-lead ECG and blood troponin level take into account the length of time since the suspected ACS when interpreting the troponin level. Use clinical judgement to decide whether referral is necessary and how urgent this should be. [2010]」

[Note 15] NICE CG95 recommendations 1.2.1.11 and 1.2.1.12: recurrent pain after a recent acute coronary syndrome calls for emergency referral; and transfer to hospital must not be delayed to do an electrocardiogram.

「1.2.1.11 Refer people to hospital as an emergency if they have a recent (confirmed or suspected) ACS and develop further chest pain. [2010]」

「1.2.1.12 When an ACS is suspected, start management immediately in the order appropriate to the circumstances (see the section on immediate management of a suspected acute coronary syndrome) and take a resting 12-lead ECG (see the section on resting 12-lead ECG). Take the ECG as soon as possible, but do not delay transfer to hospital. [2010]」

[Note 16] Four recommendations from section 1.2.2 of NICE CG95: 1.2.2.5 (a normal electrocardiogram cannot exclude), 1.2.2.4, 1.2.2.6 (the results may not be conclusive) and 1.2.2.8 (where the cardiac possibility is reduced, think of pulmonary embolism, aortic dissection and pneumonia).

「1.2.2.5 Do not exclude an ACS when people have a normal resting 12-lead ECG. [2010]」

「1.2.2.4 Even in the absence of ST-segment changes, have an increased suspicion of an ACS if there are other changes in the resting 12-lead ECG, specifically Q waves and T wave changes. Consider following the NICE guideline on acute coronary syndromes if these conditions are likely. Continue to monitor (see recommendation 1.2.3.4). [2010]」

「1.2.2.6 If a diagnosis of ACS is in doubt, consider: taking serial resting 12-lead ECGs, reviewing previous resting 12-lead ECGs, recording additional ECG leads. Use clinical judgement to decide how often this should be done. Note that the results may not be conclusive. [2010]」

「1.2.2.8 If clinical assessment (as described in recommendation 1.2.1.10) and a resting 12-lead ECG make a diagnosis of ACS less likely, consider other acute conditions. First consider those that are life-threatening such as pulmonary embolism, aortic dissection or pneumonia. Continue to monitor (see recommendation 1.2.3.4). [2010]」

[Note 17] The three ways of telling them apart that NICE CG95 expressly forbids: the response to glyceryl trinitrate, by gender, and by ethnicity.

「Do not use people's response to glyceryl trinitrate (GTN) to make a diagnosis. [2010]」

「Do not assess symptoms of an ACS differently in men and women. Not all people with an ACS present with central chest pain as the predominant feature. [2010]」

「Do not assess symptoms of an ACS differently in ethnic groups. There are no major differences in symptoms of an ACS among different ethnic groups. [2010]」

[Note 18] The originals from two Hong Kong government websites on the signs of coronary heart disease: the Department of Health Elderly Health Service's Coronary heart disease page (with its sentence about attending accident and emergency and its statement that about 10% have no signs), and the Centre for Health Protection's Heart disease page.

"Some people may develop the following signs / Angina / Usually occurs only when the load on the heart increases, as with exercise, emotional stress or walking quickly. / Signs of myocardial infarction / Severe pain in the front of the chest may extend to the neck, arms and jaw; profuse sweating and breathlessness; a feeling of oppression and nausea; dizziness, and even loss of consciousness. If any of the above signs appear, see a doctor at once or attend accident and emergency." (our translation from the Chinese original)

Chinese original:

「有些人可能出現下列徵狀 心絞痛 通常在心臟負荷加重時才出現；如運動、情緒受壓抑和快速行走等。 心肌栓塞病徵 胸前劇痛可能推展到頸、手臂、下顎；大量出汗、氣促；翳悶、作嘔；暈眩，甚至不省人事。如有以上徵象應該立即請醫生診斷，或到急症室求診。」

"About 10% of patients have no signs at all and few abnormal findings on physical examination, and trouble usually appears only once it has become very severe — that is, sudden death with no warning. People therefore often overlook the possibility of this kind of heart disease." (our translation from the Chinese original)

Chinese original:

「大概百分之十的患者沒有任何徵狀，在體格檢查時也很少有異常發現，通常在發展到很嚴重時才會出現問題 - 即毫無訊號下突然死亡。人們因此常常忽略此種心臟病的可能性。」

"Patients with coronary heart disease often develop a crushing angina after vigorous exercise, the pain spreading to the arms, shoulders, neck and jaw, but improving after rest. During a heart attack the angina becomes more severe and lasts longer. Other signs that may appear include disturbance of heart rhythm, dizziness, sweating, nausea and weakness of the limbs, and patients in heart failure will also have breathlessness and swelling of both feet." (our translation from the Chinese original)

Chinese original:

「患有冠心病的病人往往會在劇烈運動後產生壓迫性的心絞痛，痛楚可擴散至手臂、肩膀、頸部和下顎，但休息後便會有所好轉。病人在心臟病發時，心絞痛的程度會加劇，時間也會延長。其他可能出現的病徵包括心律紊亂、暈眩、出汗、噁心和四肢無力，心臟衰竭的病人更會氣促和雙腳水腫。」 (The spaces between the characters in the Chinese original are that page's own typesetting, and are reproduced as found.)

[Note 19] Two sentences from the results paragraph of Wong 2003's abstract: the association with non-cardiac chest pain and the 「but not with」 string, and immediately after it the anxiety and depression scores and the days of sick leave.

「Gastro-oesophageal reflux disease symptoms were associated with non-cardiac chest pain [odds ratio (OR), 2.3; 95% confidence interval (95% CI), 1.7-3.1], dyspepsia (OR, 1.9; 95% CI, 1.4-2.5), globus (OR, 1.8; 95% CI, 1.2-2.7), acid feeling in the stomach (OR, 5.8; 95% CI, 4.5-7.5) and the use of non-steroidal anti-inflammatory drugs (OR, 2.3; 95% CI, 1.5-3.6), but not with dysphagia, bronchitis, asthma, hoarseness and pneumonia.」

「Patients with gastro-oesophageal reflux disease had a significantly higher anxiety and depression score and required more days off work when compared with subjects without.」

[Note 20] NICE NG12 recommendations 1.2.1, 1.2.2 and 1.2.6: the threshold of 55 with dysphagia at any age, haematemesis at any age, and the upper abdominal mass.

「1.2.1 Refer people using a suspected cancer pathway referral for oesophageal cancer if they: have dysphagia, or are aged 55 and over, with weight loss, and they have any of the following: upper abdominal pain reflux dyspepsia. [2015, amended 2025]」

「1.2.2 Consider non-urgent, direct access upper gastrointestinal endoscopy to assess for oesophageal cancer in people with haematemesis. [2015]」

「1.2.6 Consider a suspected cancer pathway referral for people with an upper abdominal mass consistent with stomach cancer. [2015]」 (Recommendations 1.2.7, 1.2.8 and 1.2.9 of the gastric cancer section repeat the criteria of 1.2.1, 1.2.2 and 1.2.3 respectively, with the same threshold of 55, the same haematemesis rule and the same set of non-urgent direct access conditions.)

[Note 21] The whole of NICE NG12 recommendation 1.2.3 with its four boxes — the whole of it governed by 「aged 55 or over」, the raised platelet count box pairing with six items and the nausea or vomiting box with four.

「1.2.3 Consider non-urgent, direct access upper gastrointestinal endoscopy to assess for oesophageal cancer in people aged 55 or over with: treatment-resistant dyspepsia, or upper abdominal pain with low haemoglobin levels, or raised platelet count with any of the following: nausea, vomiting, weight loss, reflux, dyspepsia, upper abdominal pain, or nausea or vomiting with any of the following: weight loss, reflux, dyspepsia, upper abdominal pain. [2015]」

[Note 22] NICE NG12 recommendations 1.2.10 and 1.2.11: the urgent direct access ultrasound routes from an upper abdominal mass to gallbladder cancer and to liver cancer.

「1.2.10 Consider an urgent, direct access ultrasound scan to assess for gall bladder cancer in people with an upper abdominal mass consistent with an enlarged gall bladder. [2015]」

「1.2.11 Consider an urgent, direct access ultrasound scan to assess for liver cancer in people with an upper abdominal mass consistent with an enlarged liver. [2015]」

[Note 23] The six signs not to be ignored on the Department of Health Indigestion page, reproduced from the Chinese page as found.

"Indigestion may be only a temporary minor problem, but these signs may also indicate a peptic ulcer, commonly called a stomach complaint, which if ignored can in severe cases lead to perforation or gastric bleeding. Early gastric cancer may also produce similar signs, so if in doubt seek medical attention early. The following are some signs not to be ignored: persistent loss of appetite; unexplained weight loss; passing tarry black stools like sesame paste; sudden or persistent severe abdominal pain; symptoms of indigestion in someone taking aspirin or painkillers long term; a sudden worsening of symptoms in someone already known to have a peptic ulcer." (our translation from the Chinese original)

Chinese original:

「消化不良可能只是暫時的小問題，但這些病徵亦可能顯示已有消化性潰瘍，即俗稱胃病，假若不加理會，嚴重的可導致穿孔或胃出血。此外，早期胃癌亦可能有類此的病徵，因此，如有懷疑就應及早延醫診治。以下是一些不容忽視的病徵： 持續性食慾不振； 無故消瘦； 排泄出墨黑色像芝麻糊的糊狀糞便； 突發或持續的腹部劇痛； 長期服用亞士匹靈或止痛藥人士，出現消化不良徵狀； 已患有消化性潰瘍的人士，徵狀突然惡化。」

[Note 24] NICE CG184 recommendation 1.3.1: same-day referral for dyspepsia with severe acute gastrointestinal bleeding.

「1.3.1 For people presenting with dyspepsia together with significant acute gastrointestinal bleeding, refer them immediately (on the same day) to a specialist. Also see the NICE guideline on acute upper gastrointestinal bleeding in over 16s: management. [2004]」

[Note 25] The Hospital Authority gastroscopy leaflet's instruction on a serious incident after the examination.

"If a patient has minor discomfort after the examination, or has questions about the results or about medication, they should telephone the Endoscopy Centre during office hours; but if a serious incident occurs, such as bleeding in the gut or severe abdominal pain, they should attend the nearest accident and emergency department." (our translation from the Chinese original)

Chinese original:

「在檢查後如病人出現輕微不適，或對檢查結果、服藥有疑問者，應於辦公時間內致電“內窺鏡中心”查詢；但如出現嚴重事故，如腸胃出血、劇烈腹痛等，則應到就近急症室求診。」

[Note 26] The "using medicines properly" box on the Department of Health Indigestion page: aspirin, painkillers and non-steroidal anti-inflammatory drugs 「切勿自行服用」, not to be taken on your own initiative.

"Certain drugs such as aspirin, painkillers and non-steroidal anti-inflammatory drugs can all cause peptic ulcer, and must not be taken on your own initiative. If you have a problem with joint pain, it must be prescribed by a doctor before you take it." (our translation from the Chinese original)

Chinese original:

「一些藥物如亞士匹靈、止痛藥或非類固醇消炎藥等，均能引致消化性潰瘍，切勿自行服用。如有關節痛症的問題，要先由醫生處方才可服食。」

[Note 27] ACG 2022's three recommendations on endoscopy, and the five alarm symptoms with their time limit listed elsewhere in the body of the guideline.

「For patients with classic GERD symptoms of heartburn and regurgitation who have no alarm symptoms, we recommend an 8-wk trial of empiric PPIs once daily before a meal. Moderate Strong」

「We recommend endoscopy as the first test for evaluation of patients presenting with dysphagia or other alarm symptoms (weight loss and GI bleeding) and for patients with multiple risk factors for Barrett's esophagus. Low Strong」

「For patients with GERD symptoms who also have alarm symptoms such as dysphagia, weight loss, bleeding, vomiting, and/or anemia, endoscopy should be performed as soon as feasible.」

[Note 28] ACG 2022's only strong recommendation written for "the drug was not good enough, or the symptoms came back after stopping" — preferably scoped two to four weeks after the PPI is stopped.

「We recommend diagnostic endoscopy, ideally after PPIs are stopped for 2 to 4 weeks, in patients whose classic GERD symptoms do not respond adequately to an 8-week empiric trial of PPIs, or whose symptoms return when PPIs are discontinued. (Strong recommendation, low level of evidence)」

[Note 29] ACG 2022's two strong recommendations on reflux monitoring: the next step for those whose endoscopy was negative, and that it need not be done where grade C or D or long-segment Barrett's oesophagus has already been seen.

「In patients for whom the diagnosis of GERD is suspected but not clear, and endoscopy shows no objective evidence of GERD, we recommend reflux monitoring be performed off therapy to establish the diagnosis. (Strong recommendation, low level of evidence)」

「We recommend against performing reflux monitoring off therapy solely as a diagnostic test for GERD in patients known to have endoscopic evidence of Los Angeles grade C or D reflux esophagitis, or in patients with long-segment Barrett's esophagus. (Strong recommendation, low level of evidence)」

[Note 30] The two boxes of the screening section of the Cancer Online Resource Hub's Stomach cancer page, reproduced in both languages — the headings of the second box contradicting each other between the two versions.

The two boxes of the English version:

「For asymptomatic population at average risk」 「Screening for stomach cancer (by upper gastrointestinal series, upper endoscopy or biomarkers, such as H. pylori serology) is not recommended in asymptomatic persons at average risk. Screening for H. pylori infection (by urea breath test, serology or stool antigen test) among asymptomatic persons in the general population is also not recommended.」

「For symptomatic persons at increased risk」 「Persons at increased risk, e.g. with precancerous lesion of stomach cancer or family history of stomach cancer, may consider seeking advice from doctors regarding the need for and approach of screening.」

The two boxes of the Chinese version (that page's own Chinese, not a translation made here; the spaces around the brackets are the page's own typesetting, reproduced as found):

「一般風險的無症狀人士」 「不建議一般風險的無症狀人士進行胃癌篩查 (包括上消化道攝影檢查、胃內視鏡檢查或 生物標記測試，例如幽門螺旋菌血清測試)。一般市民如無症狀，亦不建議接受幽門螺旋菌感染測試(包括尿素呼氣測試、血清測試或大便抗原測試)。」

「較高風險的無症狀人士」 「較高風險人士(如出現胃癌癌前病變或有胃癌家族史者)可考慮就是否需要接受篩查及篩查方式諮詢醫生意見。」

[Note 31] The remaining three recommendations of section 1.3 of NICE CG184: 1.3.2 (medication as a cause, and suspending the non-steroidal anti-inflammatory painkiller at referral), 1.3.3 (think of the heart or the biliary tract) and 1.3.4 (manage on the previous result where there are no new alarm signs after a gastroscopy), together with NICE's own signpost to NG12.

「1.3.2 Review medications for possible causes of dyspepsia (for example, calcium antagonists, nitrates, theophyllines, bisphosphonates, corticosteroids and non-steroidal anti-inflammatory drugs [NSAIDs]). In people needing referral, suspend NSAID use. [2004]」

「1.3.3 Think about the possibility of cardiac or biliary disease as part of the differential diagnosis. [2004, amended 2014]」

「1.3.4 If people have had a previous endoscopy and do not have any new alarm signs, consider continuing management according to previous endoscopic findings. [2004]」

「For more information about alarm signs and when to refer people to specialists when they present with symptoms that could be caused by cancer, see the NICE guideline on suspected cancer: recognition and referral.」

[Note 32] The note on the Hospital Authority gastroscopy leaflet, Chinese against English — the English version having one sentence more, 「Other unforeseen complications may occasionally occur.」

"This leaflet gives only basic information about the examination, and the risks or complications that may occur cannot be exhaustively listed. The degree of risk also differs for certain kinds of patient. If you have questions, please contact your doctor." (our translation from the Chinese original)

Chinese original:

「本單張只提供有關檢查的基本資料，可能發生的風險或併發症不能盡錄。某類病人的風險程度亦為不同。如有查詢，請聯絡你的醫生。」

「This is general information only and the list of complications is not exhaustive. Other unforeseen complications may occasionally occur. In special patient groups, the actual risk may be different. For further information, please contact your doctor.」

[Note 33] All six sentences of the "preparation before the examination" box of the Hospital Authority gastroscopy leaflet, reproduced whole in both languages.

"The patient must fast for at least six hours before the endoscopy. A patient with another disease such as diabetes, high blood pressure, valvular heart disease or pregnancy must tell the healthcare staff and take their medicines as instructed. The patient should also give details of the medicines they are currently taking, especially certain drugs affecting clotting, and any information about allergic reactions. Outpatients should not drive themselves here; they should also avoid drinking heavily, smoking or taking an inappropriate dose of a sedative before the examination. Elderly or infirm patients should be accompanied to the examination by a family member. Dentures, glasses and metal ornaments must be removed before the examination." (our translation from the Chinese original)

Chinese original:

「病人最少須於內鏡檢查前六小時禁飲食。病人如有其他疾病如糖尿病、高血壓、心臟瓣膜性疾病或懷孕等，須告訴醫護人員並聽從其指示服藥。病人亦應提供現所服用藥物的詳情，特別是某些影響凝血的藥物及任何過敏反應資料。門診病人不宜親自駕駛前來；亦應避免在檢查前酗酒、抽煙或服用不當份量的鎮靜藥物。年老、行動不便的病人宜由家人陪伴前來檢查。假牙、眼鏡及金屬飾物須在檢查前脫下。」

「Patients need to be fasted for at least 6 hours before the procedure. Patients should inform the medical staff of any major medical problems including diabetes, hypertension, valvular heart disease and pregnancy, and continue their medications as instructed. Patients should also provide information concerning the current medications used especially antiplatelet and anticoagulation drugs and any allergic history. Patients should avoid driving to attend the outpatient procedure and also avoid heavy drinking, smoking or use of sedative before the procedure. Elderly patients and those with difficulty in walking should be accompanied by family member. Dentures, spectacles and metallic objects should be removed before the procedure.」

[Note 34] The complication rates in the Hospital Authority gastroscopy leaflet, with the consequence sentence that follows in the same paragraph.

"The more serious complications include perforation of the gut (less than one ten-thousandth), bleeding (less than three ten-thousandths) and death (less than one ten-thousandth), and cardiopulmonary complications and infection may also occur. The chance of each complication varies with the patient's own circumstances and with the examination and treatment methods, and patients should ask their attending doctor for the details. If a complication occurs the patient may need surgery to repair it, and in a severe case it can lead to the patient's death." (our translation from the Chinese original)

Chinese original:

「較嚴重之併發症包括腸道穿孔( 小於萬份之一) 、出血(小於萬份之三)、死亡(小於萬份之一) ，心肺併發症及感染等均可出現。各併發症發生之機會隨著病人不同之情況、檢驗及治療方法而有所差異，病人應主動詢問主診醫生以明瞭詳情。如出現併發症，病人或需接受外科手術補救，嚴重者可導致病人死亡。」

[Note 35] The two sentences of the "after the examination" box of the Hospital Authority gastroscopy leaflet.

"The patient must fast for a further hour after the examination, until the throat anaesthetic or the sedative has worn off, before eating, to prevent choking. If the patient has been given a sedative injection, they must not operate heavy machinery, sign legal documents or drive for the whole day, to prevent an accident." (our translation from the Chinese original)

Chinese original:

「病人檢查後尚須禁食一小時，以待喉部麻醉藥或鎮靜劑藥力減退後才可進食，以防哽嚥。如病人接受鎮靜劑注射，則整天不可操作重型機器、簽署法律文件或駕駛，以防意外發生。」

[Note 36] The design and conclusion paragraphs of Gomm et al 2016 (JAMA Neurol) in full, with the last sentence that randomised prospective trials are needed to examine causality.

「We conducted a prospective cohort study using observational data from 2004 to 2011, derived from the largest German statutory health insurer, Allgemeine Ortskrankenkassen (AOK).」

「CONCLUSIONS AND RELEVANCE: The avoidance of PPI medication may prevent the development of dementia. This finding is supported by recent pharmacoepidemiological analyses on primary data and is in line with mouse models in which the use of PPIs increased the levels of β-amyloid in the brains of mice. Randomized, prospective clinical trials are needed to examine this connection in more detail.」

[Note 37] The COMPASS trial (Moayyedi et al 2019).

「We performed a 3 × 2 partial factorial double-blind trial of 17,598 participants with stable cardiovascular disease and peripheral artery disease randomly assigned to groups given pantoprazole (40 mg daily, n = 8791) or placebo (n = 8807). … Patients were followed up for a median of 3.01 years, with 53,152 patient-years of follow-up.」

「There was no statistically significant difference between the pantoprazole and placebo groups in safety events except for enteric infections (1.4% vs 1.0% in the placebo group; odds ratio, 1.33; 95% confidence interval, 1.01-1.75).」

(In substance: this 3 × 2 partial factorial double-blind trial enrolled 17,598 patients with stable cardiovascular disease and peripheral artery disease, randomised to pantoprazole 40 mg once daily (8,791 people) or placebo (8,807 people), with a median follow-up of 3.01 years and 53,152 patient-years. Apart from enteric infections (1.4% in the pantoprazole arm against 1.0% in the placebo arm, odds ratio 1.33, 95% confidence interval 1.01–1.75), there was no statistically significant difference between the two arms in safety events.)

[Note 38] Footnote c to ACG's table 5 (the FDA advising against using clopidogrel together with omeprazole), and ACG's own other half of the advice immediately following.

「c The US Food and Drug Administration recommends avoiding the concomitant use of clopidogrel and omeprazole.」

「For patients with GERD on clopidogrel who have LA grade C or D esophagitis or whose GERD symptoms are not adequately controlled with alternative medical therapies, the highest quality data available suggest that the established benefits of PPI treatment outweigh their proposed but highly questionable cardiovascular risks.」

[Note 39] ACG's reading of COMPASS, with the limits it places on how reassuring it is; and the passage recommending how a doctor should put it to a patient.

「Nevertheless, this study cannot exclude the possibility that PPIs confer a modest risk for any of these adverse events (i.e., the upper limit of the 95% confidence intervals all are >1), and even a modest risk for such serious events is cause for concern. As the authors themselves acknowledge, the possibility that PPIs confer a modest risk for these putative adverse events can never be excluded no matter how large the study sample size」

「PPIs are the most effective medical treatment for GERD. Some medical studies have identified an association between the long-term use of PPIs and the development of numerous adverse conditions including intestinal infections, pneumonia, stomach cancer, osteoporosis-related bone fractures, chronic kidney disease, deficiencies of certain vitamins and minerals, heart attacks, strokes, dementia, and early death. Those studies have flaws, are not considered definitive, and do not establish a cause-and-effect relationship between PPIs and the adverse conditions. High-quality studies have found that PPIs do not significantly increase the risk of any of these conditions except intestinal infections. Nevertheless, we cannot exclude the possibility that PPIs might confer a small increase in the risk of developing these adverse conditions. For the treatment of GERD, gastroenterologists generally agree that the well-established benefits of PPIs far outweigh their theoretical risks.」

[Note 40] ACG on the common side effects of PPIs — headache, diarrhoea, constipation and the like — and its suggestion that they may be related to the particular preparation.

「Side effects of PPIs that have been identified in clinical trials and listed on FDA labels as the 'most common adverse reactions' include headache, abdominal pain, nausea, vomiting, diarrhea, constipation, and flatulence. These relatively minor side effects occur infrequently and abate when the medications are stopped. Limited data also suggest that these side effects sometimes can be PPI preparation-specific and, for patients who experience them, a trial of switching from one PPI to another is a reasonable management strategy」

[Note 41] The United States Food and Drug Administration's drug safety communication of March 2011 on hypomagnesaemia: its opening sentence, the three-month lower bound in its data summary section, and the monitoring advice (with 「especially important」 for digoxin).

「The U.S. Food and Drug Administration (FDA) is informing the public that prescription proton pump inhibitor (PPI) drugs may cause low serum magnesium levels (hypomagnesemia) if taken for prolonged periods of time (in most cases, longer than one year). In approximately one-quarter of the cases reviewed, magnesium supplementation alone did not improve low serum magnesium levels and the PPI had to be discontinued.」

「Hypomagnesemia has been reported in adult patients taking PPIs for at least three months, but most cases occurred after a year of treatment. Approximately one-quarter of these cases required discontinuation of PPI treatment in addition to magnesium supplementation. Some cases cited both positive dechallenge as well as positive rechallenge (i.e., resolution of hypomagnesemia with PPI cessation and recurrent hypomagnesemia with PPI resumption). After discontinuing the PPI, the median time required for the magnesium to normalize was one week. After restarting the PPI, the median time to develop hypomagnesemia again was two weeks.」

「Healthcare professionals should consider obtaining serum magnesium levels prior to initiation of prescription PPI treatment in patients expected to be on these drugs for long periods of time, as well as patients who take PPIs with medications such as digoxin, diuretics or drugs that may cause hypomagnesemia. For patients taking digoxin, a heart medicine, this is especially important because low magnesium can increase the likelihood of serious side effects. Healthcare professionals should consider obtaining magnesium levels periodically in these patients.」

[Note 42] The three-way divergence recorded by ACG 2022: the FDA, AGA and ACG accounts of whether serum magnesium should be checked.

「A recent AGA Best Practice Recommendation concluded that long-term PPI users should not routinely screen or monitor serum magnesium levels […], whereas the FDA suggests that health care providers should consider monitoring magnesium levels prior to initiation of PPI treatment and then periodically […]. We feel that presently there is insufficient data to make a meaningful recommendation」

[Note 43] ACG 2022's express statement on vitamin B12 — for patients without other risk factors for B12 deficiency, routine monitoring is not recommended.

「For GERD patients on PPIs who have no other risk factors for vitamin B12 deficiency, we do not recommend that they raise their intake of vitamin B12, or that they have routine monitoring of serum B12 levels.」

[Note 44] The symptom list in the same FDA communication with its open marker 「patients do not always have these symptoms」, and the instruction written to patients.

「Low serum magnesium levels can result in serious adverse events including muscle spasm (tetany), irregular heartbeat (arrhythmias), and convulsions (seizures); however, patients do not always have these symptoms.」

「Seek immediate care if you (or your child) experience an abnormal heart rate or rhythm, or symptoms such as a racing heartbeat, palpitations, muscle spasm, tremor or convulsions while taking a PPI drug.」

「Do not stop taking your prescription PPI drug without talking to your healthcare professional.」

[Note 45] The FDA on over-the-counter PPIs: labelled for a 14-day course, repeatable only every 4 months, at most 3 courses a year, and its warning about use beyond that.

「OTC PPIs are labeled for 14 days of use, and this treatment course may be repeated every 4 months, up to 3 times per year.」

「FDA acknowledges that consumers, either on their own, or based on a healthcare professional's recommendation, may take these products for periods of time that exceed the directions on the OTC label. This is considered an off-label (unapproved) use, based on the directions of use for OTC PPIs. Healthcare professionals should be aware of the risk of hypomagnesemia if they are recommending use of OTC PPIs for longer periods of time than in the OTC PPI label.」

[Note 46] The guiding principle of AGA's 2022 de-prescribing document, and ACG's own two recommendations on stopping (including the third option of switching to on-demand therapy).

「Our guiding principle was that, although PPIs are generally safe, patients should not use any medication when there is not a reasonable expectation of benefit based on scientific evidence or prior treatment response.」

「We recommend attempting to discontinue the PPIs in patients whose classic GERD symptoms respond to an 8-wk empiric trial of PPIs. Low Conditional」

「For GERD patients who do not have erosive esophagitis or Barrett's esophagus, and whose symptoms have resolved with PPI therapy, an attempt should be made to discontinue PPIs or to switch to on-demand therapy in which PPIs are taken only when symptoms occur and discontinued when they are relieved. (Conditional recommendation, low level of evidence)」

[Note 47] NICE CG184 recommendations 1.6.6 and 1.6.9: long-term full-dose maintenance for those who have had an oesophageal stricture dilated and for those with severe oesophagitis.

「1.6.6 People who have had dilatation of an oesophageal stricture should remain on long-term full-dose PPI therapy (see table 1 in appendix A). [2004]」

「1.6.9 Offer a full-dose PPI (see table 2 in appendix A) long-term as maintenance treatment for people with severe oesophagitis, taking into account the person's preference and clinical circumstances (for example, tolerability of the PPI, underlying health conditions and possible interactions with other drugs), and the acquisition cost of the PPI. [new 2014]」

[Note 48] ACG 2022: the lowest effective dose, and taking the drug 30 to 60 minutes before a meal.

「For patients with GERD who require maintenance therapy with PPIs, the PPIs should be administered in the lowest dose that effectively controls GERD symptoms and maintains healing of reflux esophagitis. Low Conditional」

「We recommend PPI administration 30–60 min before a meal rather than at bedtime for GERD symptom control. Moderate Strong」

[Note 49] The four recommendations of NICE CG184's stepping-down ladder: 1.5.1, 1.5.2, 1.2.5, and the GORD version at 1.6.2.

「1.5.1 Offer people who need long-term management of dyspepsia symptoms an annual review of their condition, and encourage them to try stepping down or stopping treatment (unless there is an underlying condition or comedication that needs continuing treatment). [2004, amended 2014]」

「1.5.2 Advise people that it may be appropriate for them to return to self-treatment with antacid and/or alginate therapy (either prescribed or purchased over-the-counter and taken as needed). [2004, amended 2014]」

「1.2.5 Encourage people who need long-term management of dyspepsia symptoms to reduce their use of prescribed medication stepwise: by using the effective lowest dose, by trying 'as-needed' use when appropriate, and by returning to self-treatment with antacid and/or alginate therapy (unless there is an underlying condition or comedication that needs continuing treatment). [2004, amended 2014]」

「1.6.2 Offer people with GORD a full-dose PPI (see table 1 in appendix A) for 4 or 8 weeks. [2004]」 「1.6.3 If symptoms recur after initial treatment, offer a PPI at the lowest dose possible to control symptoms. [2004, amended 2014]」 「1.6.4 Discuss with people how they can manage their own symptoms by using the treatment when they need it. [2004]」 「1.6.5 Offer H2RA therapy if there is an inadequate response to a PPI. [2004, amended 2014]」

(In substance: 1.6.2 offer people with GORD a full-dose PPI for 4 weeks or 8 weeks. 1.6.3 if symptoms recur after initial treatment, offer a PPI at the lowest dose that controls symptoms. 1.6.4 discuss with people how they can manage their own symptoms by taking the treatment when they need it. 1.6.5 offer H2 receptor antagonist therapy where the response to a PPI is inadequate.)

[Note 50] ACG 2022's five lifestyle recommendations, reproduced in full — only weight loss is rated strong.

「We recommend weight loss in overweight and obese patients for improvement of GERD symptoms. (Strong recommendation, moderate level of evidence) We suggest avoiding meals within 2-3 hours of bedtime. (Conditional recommendation, low level of evidence) We suggest avoidance of tobacco products/smoking in patients with GERD symptoms. (Conditional recommendation, low level of evidence) We suggest avoidance of "trigger foods" for GERD symptom control. (Conditional recommendation, low level of evidence) We suggest elevating head of bed for nighttime GERD symptoms. (Conditional recommendation, low level of evidence)」

(In substance, in order: weight loss in overweight and obese patients to improve GERD symptoms (strong, moderate); avoiding meals within 2 to 3 hours of bedtime (conditional, low); avoiding tobacco products and smoking in patients with GERD symptoms (conditional, low); avoiding trigger foods to control GERD symptoms (conditional, low); and elevating the head of the bed for night-time GERD symptoms (conditional, low).)

[Note 51] NICE CG184 recommendations 1.2.1, 1.2.2 and 1.2.4: the general lifestyle advice, the list of precipitants, and the three qualifiers in the psychological therapy recommendation.

「1.2.1 Offer simple lifestyle advice, including advice on healthy eating, weight reduction and smoking cessation. [2004]」

「1.2.2 Advise people to avoid known precipitants they associate with their dyspepsia where possible. These include smoking, alcohol, coffee, chocolate, fatty foods and being overweight. Raising the head of the bed and having a main meal well before going to bed may help some people. [2004]」

「1.2.4 Recognise that psychological therapies, such as cognitive behavioural therapy and psychotherapy, may reduce dyspeptic symptoms in the short term in individual people. [2004, amended 2014]」

[Note 52] Two passages on lifestyle from the body of ACG 2022 — including 「staying upright during and after meals」, the 「including several randomized controlled trials」 for raising the head of the bed, and the two sentences on sleeping position.

「Common recommendations include weight loss for overweight patients, elevating the head of the bed, tobacco and alcohol cessation, avoidance of late night meals and bedtime snacks, staying upright during and after meals, and cessation of foods that potentially aggravate reflux symptoms such as coffee, chocolate, carbonated beverages, spicy foods, acidic foods such as citrus and tomatoes, and foods with high fat content」

「However, multiple studies, including several randomized controlled trials, have demonstrated improvement in nocturnal GERD symptoms and nocturnal esophageal acid exposure with head of bed elevation or sleeping on a wedge. Also, compared to lying left-side down, lying right-side down increases nocturnal reflux and reflux after meals, presumably because right-sided recumbency places the EGJ in a dependent position relative to the pool of gastric contents that favors reflux. Thus, patients might be advised to avoid sleeping right side down.」

[Note 53] The four boxes of tips and the causes paragraph on the Department of Health Indigestion page: the dietary adjustment items, the stopping smoking and avoiding alcohol in the first box on healthy living, the third box on managing emotions and stress, and the 「吸煙和酗酒」 of the causes paragraph.

"Do not lie down immediately after eating, to prevent food refluxing back into the oesophagus." (our translation from the Chinese original)

Chinese original:

「進食後不宜立刻躺下，以防止食物倒流回食道。」

"Avoid high-fat foods such as fried, deep-fried and greasy food." / "Avoid foods and drinks containing caffeine, such as chocolate, coffee and tea." / "Avoid wearing clothes that are too tight." / "Eat only to seven tenths of fullness." (our translation from the Chinese original)

Chinese original:

「避免煎、炸、肥膩等高脂肪食物。」 「避免含咖啡因之食物和飲品，如朱古力、咖啡、茶等。」 「避免穿著太緊之衣服。」 「只吃至七分飽。」

"1. Healthy living / Keep regular hours and take suitable exercise. / Do not smoke, or stop smoking. / Avoid alcohol." (our translation from the Chinese original)

Chinese original:

「1. 健康生活 作息定時，並有適量運動。 不吸煙或戒煙。 避免飲酒。」

"Poor lifestyle habits: bingeing, irregular meals, smoking and heavy drinking." (our translation from the Chinese original)

Chinese original:

「不良生活習慣 : 暴飲暴食、飲食不定時、吸煙和酗酒。」

"3. Managing emotions and stress / Understanding the process of ageing and one's own bodily function can reduce unnecessary worry. / Practising relaxation of body and mind helps ease tension. / Arranging work or housework properly can reduce the mental stress that comes from time pressure. / Appreciating yourself, other people and the things around you more will make you happier." (our translation from the Chinese original)

Chinese original:

「3. 情緒與壓力的處理 認識老化過程及身體機能，可減少不必要的憂慮。 實踐身心鬆弛法，有助紓緩緊張情緒。 妥善安排工作或家務，可減低時間緊迫所帶來的精神壓力。 多欣賞自己、別人和週遭事物，令自己更開懷。」

[Note 54] ACG's Barrett's oesophagus guideline 2022: the annual risk of progression, the intestinal metaplasia comparison and the contrary evidence immediately following it, the difference between the sexes and the modelling study, and the sentence written immediately afterwards for women at high risk.

「The annual incidence of cancer progression in BE is estimated at 0.2%–0.05% per year for NDBE and approximately 0.7% per year for LGD」

「On the other hand, other studies suggest no difference in cancer risk of columnar epithelium with or without IM. A single-center UK study of 688 patients with a median follow-up of 12 years found no difference in cancer risk for those with a columnar-lined esophagus with or without IM: 0.37% vs 0.30%/year. Similarly, a multicenter UK study of 1,751 patients found a similar cancer risk in patients with and without IM (HR 1.36; 95% CI 0.63–2.96).」

「In a modeling study, the incidence of EAC in women with weekly symptoms of GERD at age 60 years was markedly lower (3.9/100,000 person-years) compared with men (61/100,000 person-years). Hence, BE screening in women is likely low yield in terms of reducing EAC incidence. However, screening women with multiple risk factors for BE and EAC may be appropriate following discussion with the patient on the pros and cons of such an approach.」

[Note 55] Recommendation 5 of the same guideline (3 or more risk factors, a single screening endoscopy), and the fundamental limitation it sets down for itself.

「We suggest a single screening endoscopy for patients with chronic GERD symptoms and 3 or more additional risk factors for BE, including male sex, age >50 years, White race, tobacco smoking, obesity, and family history of BE or EAC in a first-degree relative (strength of recommendation: conditional; quality of evidence: very low)」

「Unfortunately, there is no randomized controlled trial evidence demonstrating reduced EAC mortality with BE screening.」

[Note 56] NICE CG184 recommendation 1.6.11 and the only recommendation of section 1.12, 1.12.1: do not routinely endoscope to diagnose Barrett's oesophagus (but consider it where the person has GORD), and the main clause and closing sentence of the surveillance recommendation.

「1.6.11 Do not routinely offer endoscopy to diagnose Barrett's oesophagus, but consider it if the person has GORD. Discuss the person's preferences and their individual risk factors (for example, long duration of symptoms, increased frequency of symptoms, previous oesophagitis, previous hiatus hernia, oesophageal stricture or oesophageal ulcers, or male gender). [new 2014]」

「1.12.1 Consider surveillance to check progression to cancer for people who have a diagnosis of Barrett's oesophagus (confirmed by endoscopy and histopathology), taking into account: the presence of dysplasia (also see the NICE guideline on Barrett's oesophagus and stage 1 oesophageal adenocarcinoma) the person's individual preference the person's risk factors (for example, male gender, older age and the length of the Barrett's oesophagus segment). Emphasise that the harms of endoscopic surveillance may outweigh the benefits in people who are at low risk of progression to cancer (for example, people with stable non-dysplastic Barrett's oesophagus). [new 2014]」

#What this article does not state

The following have no support in the sources cited here, and so are not written:

• "1 adult in every 5 has regular symptoms" — no Hong Kong source supports 20%. The weekly proportions in the three Hong Kong surveys are 2.5%–4.2%.
• "Symptoms appearing for the first time after 40 call for a gastroscopy" — the endoscopy route NG12 sets out for reflux has an age threshold of 55, and reflux on its own never triggers a referral. To be accurate: section 1.2 of NG12 does itself use three age thresholds, 40, 55 and 60 — 1.2.4 is 40 with jaundice for referral to assess pancreatic cancer, and 1.2.5 is 60 with weight loss for a CT scan — but neither of those is the gastroscopy route for reflux. What this article does not state is "40 plus reflux symptoms calls for a gastroscopy", not "NG12 never uses 40".
• "The public gastroscopy wait is 8–60 weeks" — this article cites no waiting time in weeks for gastroscopy, because no published Hospital Authority gastroscopy waiting figure could be found, not because the Authority publishes only by specialty. To be accurate: the Authority's specialist outpatient waiting times are published by specialty (medicine, for instance); but the Authority also publishes waiting times by procedure, and the "booked surgery" column under its service guide lists the waiting times for cataract surgery and for total joint replacement separately. So the statement that the Authority does not publish by procedure does not hold. All this article can state is that no waiting time in weeks for gastroscopy (or endoscopy) could be found on the pages searched; that is "unable to confirm whether it is published", not "confirmed not published".
• "Raise the head of the bed by 15–20 centimetres when sleeping" — both ACG and NICE mention raising the head of the bed, and neither gives a height; ACG's body text says only 「head of bed elevation or sleeping on a wedge」. What this article does not state is how many centimetres, not that the thing has no evidence — the same ACG body sentence states that the studies supporting it are 「including several randomized controlled trials」.
• "Sleeping on the left side rather than the right reduces night-time reflux" (as a graded guideline recommendation) — sleeping position is not among ACG's five lifestyle recommendations, so it is not a numbered recommendation; but the ACG body text has two sentences stating that sleeping right side down increases night-time and postprandial reflux, immediately followed by 「Thus, patients might be advised to avoid sleeping right side down.」, both of them quoted in full in the lifestyle section above. What this article does not state is that this is a GRADE-rated recommendation, not that ACG never said patients might be advised so.
• "Barrett's oesophagus progresses to oesophageal adenocarcinoma at 0.1%–0.5% a year" — the figures ACG prints are 0.2%–0.05% a year for non-dysplastic and about 0.7% a year for low-grade dysplasia, and the grade of dysplasia changes the answer by more than an order of magnitude.
• "Long-term users are advised to have magnesium and B12 checked regularly" — that is a contested position and not a consensus: on serum magnesium the FDA, AGA and ACG say three different things; on vitamin B12 none of the three has ever recommended checking, and ACG states expressly that for patients without other risk factors for B12 deficiency routine monitoring of serum B12 is not recommended (what those with other risk factors should do is not addressed by ACG).
• "PPIs do not directly cause dementia or gastric cancer in the elderly" — that is stated too absolutely. ACG's own formulation is that the randomised trial 「cannot exclude the possibility that PPIs confer a modest risk」; and gastric cancer has no RCT estimate at all in ACG's table 5, which prints "NA".
• A symptom list for telling acid reflux from heart disease — no source that can be cited provides one. ACG says the two are 「indistinguishable」, and NICE CG95 expressly forbids telling them apart by the response to glyceryl trinitrate, by gender or by ethnicity.
• The range "HK$8,000–15,000" for a private gastroscopy in Hong Kong — the hospitals' prices are on different bases (minimum charge against all-inclusive package) and cannot be stretched into one range. This article lists each hospital's price and what it covers instead.
• A Hong Kong local clinical guideline for GERD — none is cited, because none could be found after searching (see the next item). Every clinical rule in this article comes from the UK NICE, the American ACG and AGA, and the Asia-Pacific and Seoul consensuses; the Hong Kong primary documents are only the Hospital Authority gastroscopy leaflet (examination information, not a referral guideline), the Department of Health Elderly Health Service's coronary heart disease and indigestion pages, the Cancer Online Resource Hub's Stomach cancer page, and the Hospital Authority Drug Formulary and fee schedule.
• A Hong Kong regulator's assessment of PPI safety — none is cited; every regulatory position in this article comes from the United States FDA.
• A GERD clinical guideline or position statement issued by the Hong Kong Society of Gastroenterology — after searching that society's website hksge.org (9 August 2026, HTTP 200) it is confirmed: there is none. The Position Statement column under its "Professional Resources" holds only two documents, one on Helicobacter pylori and one on the management of anaemia in patients with gastrointestinal bleeding, neither of them about reflux; and the website has no "Guidelines" column. But the society does have public education material on reflux: a page titled "上班族辛酸事 – GI/GERD Public Education 2014", carrying a video, doctor interviews, a PDF handbook on preventing and treating acid reflux, and a GERD-R self-test questionnaire. This article does not cite that material, because it cites only clinical guidelines and government primary documents, and that material is a public education resource carrying no date of publication or review. (The society's domain is hksge.org; hksgh.org is not its domain and does not exist.)

#Sources

• Department of Health Elderly Health Service, Coronary heart disease (Chinese version): https://www.elderly.gov.hk/tc_chi/health_information/hypertension_heart_disease/chd.html (the page carrying no revision date; retrieved 9 August 2026) — the signs of angina and of myocardial infarction, 「現時心臟病是僅次於癌症的香港第二號殺手。」 (inconsistent with the Centre for Health Protection's "ranks third" on its page of 31 March 2026; this article uses the latter), the open lead-in 「有些人可能出現下列徵狀」, the instruction 「到急症室求診」, and that about 10% of patients with coronary heart disease have no signs. ⚠ The old path /tc_chi/health_information/cardiovascular_diseases/chd.html returned HTTP 404 on 3 August 2026, this page having moved to the hypertension_heart_disease directory
• Elderly Health Service, Department of Health, Coronary heart disease (English version): https://www.elderly.gov.hk/english/health_information/hypertension_heart_disease/chd.html (retrieved 3 August 2026) — the Chinese and English symptom lists not corresponding item by item
• Centre for Health Protection, Heart disease: https://www.chp.gov.hk/tc/healthtopics/content/25/57.html · English version https://www.chp.gov.hk/en/healthtopics/content/25/57.html (page dated 31 March 2026; retrieved 3 August 2026) — the signs of coronary heart disease (including disturbance of heart rhythm, weakness of the limbs and swelling of both feet, and radiation to 「手臂、肩膀、頸部和下顎」), the 2024 heart disease death figures, and the coronary heart disease prevalence in the 2020–22 population health survey; that page contains no instruction to seek care
• Department of Health Elderly Health Service, Indigestion (Chinese version): https://www.elderly.gov.hk/tc_chi/health_information/digestive_problems/dyspepsia.html · English version https://www.elderly.gov.hk/english/health_information/digestive_problems/dyspepsia.html (the page carrying no date; retrieved 7 August 2026) — the six signs not to be ignored, all four boxes of the tips on preventing and relieving indigestion having been gone through: the 「不吸煙或戒煙」 and 「避免飲酒」 of the first box on healthy living, and the four items of the third box on managing emotions and stress, the causes paragraph's 「不良生活習慣 : 暴飲暴食、飲食不定時、吸煙和酗酒」, and, of the eleven items in the dietary adjustment box, high-fat food, caffeine (chocolate, coffee, tea), clothes that are too tight, eating only to seven tenths of fullness, and not lying down immediately after eating; and on the same page the "using medicines properly" box with its 「切勿自行服用」 for aspirin, painkillers and non-steroidal anti-inflammatory drugs and its requirement that joint pain be prescribed for by a doctor first
• Hospital Authority, Oesophago-gastro-duodenoscopy PILIC0049C version 3.0: https://www.ekg.org.hk/pilic/public/IM_PILIC/IM_OGD_0049_chi.pdf · English version PILIC0049E: https://www.ekg.org.hk/pilic/public/IM_PILIC/IM_OGD_0049_eng.pdf (effective 30 November 2023; last reviewed 15 October 2025; retrieved 3 August 2026) — acid reflux listed as an indication for gastroscopy, the complication rates and 「或需接受外科手術補救，嚴重者可導致病人死亡」, all six sentences of the whole preparation passage in both languages (six hours nil by mouth; other diseases to be declared with medicines taken as instructed; details of current medicines, the Chinese saying 「某些影響凝血的藥物」 where the English names antiplatelet and anticoagulation; the whole sentence 「門診病人不宜親自駕駛前來；亦應避免在檢查前酗酒、抽煙或服用不當份量的鎮靜藥物」; that elderly and infirm patients should be accompanied by a family member; and that dentures, glasses and metal ornaments must be removed), both sentences of the whole "after the examination" passage (including 「病人檢查後尚須禁食一小時…以防哽嚥」), the whole of the Chinese and English remarks (the English also carrying 「the list of complications is not exhaustive」 and the sentence 「Other unforeseen complications may occasionally occur.」), the same-day restriction after a sedative, and the sentence-by-sentence Chinese-English divergences (the indications sentence, the complications sentence and the remarks running in three different directions)
• Hospital Authority Drug Formulary version 22.1 (effective 25 July 2026): https://www.ha.org.hk/hadf/en/Others/Search-Result.html · formulary categories and charges: https://www.ha.org.hk/hadf/en/Drug-Formulary/Drug-Formulary-Categories-And-Charges.html (retrieved 3 August 2026) — the classification of the PPIs, all three sentences of the definitions of general and special drugs, the antisecretory drugs including FAMOTIDINE, and the antacids and alginates in the "DYSPEPSIA AND GASTRO-OESOPHAGEAL REFLUX DISEASE" column; the formulary is published in English only
• Hospital Authority public healthcare charges (effective 1 January 2026, Chinese version): https://www.ha.org.hk/visitor/ha_view_content.asp?Content_ID=10045&Lang=CHIB5&Dimension=100&Ver=HTML (retrieved 1 August 2026) — accident and emergency at 400 dollars with categories I and II exempt, the specialist and family medicine clinic consultation and drug charges, and the day procedure charges
• Hospital Authority "annual charge cap": https://www.ha.org.hk/visitor/ha_view_content.asp?Content_ID=281820&Lang=CHIB5&Dimension=100&Ver=HTML (effective 1 January 2026; retrieved 1 August 2026) — the annual cap of ten thousand dollars, with no means test
• Union Hospital day endoscopy price list: https://www.union.org/gicentre/files/Day_Case_Endoscopy_Plan_EN.pdf (document effective 1 July 2025; retrieved 1 August 2026) — OGD at 4,200 dollars as a minimum charge, excluding the consultation fee, the doctor's fee, the anaesthetist's fee and pathology charges
• CUHK Medical Centre endoscopy package charges: https://www.cuhkmc.hk/f/page/1015/24848/(CMP)%20Endo%20Package_20251101_2.pdf (effective 1 November 2025; retrieved 1 August 2026) — gastroscopy (diagnostic) (sedation) at $10,550 and $13,350, the package including the doctor's fee
• Gleneagles Hospital all-inclusive package price list (END02): https://gleneagles.hk/media/GHK-All-Inclusive-Package-Price-List-END02_250414.pdf (effective 14 April 2025; retrieved 1 August 2026) — gastroscopy (intravenous sedation), day case, at 11,100 and 14,430 dollars
• NICE clinical guideline CG95, Recent-onset chest pain of suspected cardiac origin: https://www.nice.org.uk/guidance/cg95/chapter/Recommendations (published 24 March 2010; updated 30 November 2016; retrieved 3 August 2026) — recommendation 1.2.1.3, the four symptoms that may indicate an acute coronary syndrome; 1.2.1.7 emergency referral; 1.2.1.8 same-day assessment (including the 12–72 hour box); 1.2.1.9 that pain having resolved with signs of a complication still calls for referral to hospital; 1.2.1.10 beyond 72 hours (with its three sub-items: a detailed clinical assessment, confirmation by electrocardiogram and troponin, and interpreting the troponin by the time since onset); 1.2.1.11 that recurrent chest pain after a recent acute coronary syndrome calls for emergency referral; 1.2.1.12 that management starts at once but transfer must not be delayed for an electrocardiogram; 1.2.1.13 other causes; 1.2.2.5 「Do not exclude an ACS when people have a normal resting 12-lead ECG.」; the three prohibitions at 1.2.1.4–1.2.1.6; the whole of section 1.2.2, eight recommendations (with 1.2.2.4 on changes other than ST-segment change, 1.2.2.6 that the results may still not be conclusive, and 1.2.2.8 that pulmonary embolism, aortic dissection and pneumonia come first once an acute coronary syndrome is made less likely). The scope of the sections: 1.2.1 has thirteen recommendations and 1.2.2 has eight. A statement of quoting practice: the NICE originals quoted here always keep their internal cross-references (「(see recommendation …)」, 「(see the section on …)」) and are not cut without marking; and where the NICE page uses a non-breaking hyphen (U+2011), as in 「12‑lead」, this article writes an ordinary hyphen "-", which is glyph normalisation and involves no addition or removal of words
• NICE clinical guideline CG184, Gastro-oesophageal reflux disease and dyspepsia in adults: https://www.nice.org.uk/guidance/cg184/chapter/Recommendations (published 3 September 2014; updated 18 October 2019; retrieved 3 August 2026) — the definition of GORD; the whole of section 1.3 (including 1.3.1 same-day referral, 1.3.2 medication as a cause and 1.3.3 the cardiac and biliary differential diagnosis); 1.2.1 the general lifestyle advice; 1.2.2 the precipitants and 「may help some people」; 1.2.4 psychological therapy; 1.2.5 stepping down; 1.4.3 and 1.6.2 empirical PPI; 1.5.1 the annual review; 1.5.2 antacids and alginates; the whole of section 1.6, eleven recommendations (including 1.6.2's 「4 or 8 weeks」, 1.6.3 the lowest dose after relapse, 1.6.4 self-management as needed, 1.6.5 H2RA, 1.6.6 and 1.6.9 those who should stay on long-term treatment, 1.6.7 healing severe oesophagitis in 8 weeks, 1.6.10 the review after maintenance treatment fails, and 1.6.11 diagnostic endoscopy for Barrett's); and the whole of 1.12.1 in section 1.12, the Barrett's surveillance recommendation (whose main clause is 「Consider surveillance」, the sentence about harms being a closing caution)
• NICE guideline NG12, Suspected cancer: recognition and referral: https://www.nice.org.uk/guidance/ng12/chapter/Recommendations-organised-by-site-of-cancer (the oesophageal and gastric cancer recommendations marked [2015, amended 2025]; retrieved 3 August 2026) — the whole of section 1.2 on upper gastrointestinal cancers, all eleven recommendations (1.2.1 to 1.2.11) gone through: the threshold of 55 and dysphagia at any age at 1.2.1 and 1.2.7; haematemesis at any age at 1.2.2 and 1.2.8; the whole of the non-urgent direct access route at 1.2.3 and 1.2.9 with its four boxes (including the six-item pairing list for a raised platelet count and the four-item pairing list for nausea or vomiting); 1.2.6 the upper abdominal mass; 1.2.4's threshold of 40 with jaundice for pancreatic cancer and 1.2.5's route of 60 with weight loss to CT; and the urgent direct access ultrasound routes from an upper abdominal mass to gallbladder cancer at 1.2.10 and to liver cancer at 1.2.11
• Katz PO, Dunbar KB, Schnoll-Sussman FH, Greer KB, Yadlapati R, Spechler SJ. ACG Clinical Guideline for the Diagnosis and Management of Gastroesophageal Reflux Disease. Am J Gastroenterol 2022;117(1):27–56 (DOI 10.14309/ajg.0000000000001538; PMID 34807007; https://pmc.ncbi.nlm.nih.gov/articles/PMC8754510/ ; retrieved 3 August 2026) — the definition and mechanism of GERD; the two lists of typical and extra-oesophageal symptoms; chest pain 「indistinguishable」; the two wordings of the chest pain recommendation in the table and in the body; the unreliability of the PPI trial; endoscopy not to be used to diagnose extra-oesophageal GERD or LPR; the 8-week empirical PPI; the three-item and five-item versions of the alarm symptoms and 「as soon as feasible」; the five lifestyle recommendations; the body sentence 「Common recommendations include…」 (including alcohol cessation and 「staying upright during and after meals」); the three consecutive sentences on raising the head of the bed, left and right side sleeping and 「patients might be advised to avoid sleeping right side down」; the whole of table 5 and its footnote c (clopidogrel and omeprazole); the most common adverse reactions on the FDA label; the benefit and harm advice for patients on clopidogrel; the recommendations on stopping and on on-demand use; the "strong" recommendation of diagnostic endoscopy for symptoms returning after stopping or an inadequate response at 8 weeks (including 「ideally after PPIs are stopped for 2 to 4 weeks」); the body text of the remaining two "strong" recommendations in the "diagnosis of GERD" list (reflux monitoring off medication where endoscopy gives no objective evidence; and that those with known LA grade C or D oesophagitis or long-segment Barrett's oesophagus should not stop medication for reflux monitoring merely to make the diagnosis), with the wording difference between table one and the body (the table reading 「We suggest against」 and 「LA grade C or D」, the body reading 「We recommend against」 and 「Los Angeles grade C or D」, both rated strong/low); the whole of the extra-oesophageal symptoms section, six recommendations (including the two "strong" ones on evaluating non-GERD causes first and reflux testing first where there are no typical symptoms, and the one on twice-daily PPI for 8–12 weeks); the two "strong" recommendations that a PPI rather than an H2RA is recommended both for healing and for maintaining healing of erosive oesophagitis; the three-way divergence on magnesium monitoring; the sentence declining to recommend routine monitoring of vitamin B12 (together with bone mineral density and serum creatinine), which appears both in the "Key concept statements" and in the body with slightly different wording, this article quoting the body version; and the insufficiency of the rebound evidence
• Shaheen NJ, Falk GW, Iyer PG, Souza RF, Yadlapati RH, Sauer BG, Wani S. Diagnosis and Management of Barrett's Esophagus: An Updated ACG Guideline. Am J Gastroenterol 2022;117(4):559–587 (DOI 10.14309/ajg.0000000000001680; PMID 35354777; https://pmc.ncbi.nlm.nih.gov/articles/PMC10259184/ ; retrieved 2 August 2026) — the annual risk of progression; the intestinal metaplasia comparison and the contrary evidence immediately following it (a single-centre study of 688 people, 0.37% against 0.30% a year; a multicentre study of 1,751 people, HR 1.36, 95% CI 0.63–2.96); the difference between the sexes and the sentence immediately following, 「However, screening women with multiple risk factors…may be appropriate」; the adjusted odds ratio for progression between the sexes of 2.2 (1.8–2.5); the single screening endoscopy recommendation (recommendation 5); and the absence of RCT mortality evidence
• Targownik LE, Fisher DA, Saini SD. AGA Clinical Practice Update on De-Prescribing of Proton Pump Inhibitors: Expert Review. Gastroenterology 2022;162(4):1334–1342 (DOI 10.1053/j.gastro.2021.12.247; PMID 35183361; retrieved 2 August 2026) — all ten best practice statements in full
• Moayyedi P, Eikelboom JW, Bosch J, et al.; COMPASS Investigators. Safety of Proton Pump Inhibitors Based on a Large, Multi-Year, Randomized Trial of Patients Receiving Rivaroxaban or Aspirin. Gastroenterology 2019;157(3):682–691.e2 (DOI 10.1053/j.gastro.2019.05.056; PMID 31152740; NCT01776424; retrieved 2 August 2026) — the randomised trial in 17,598 people, with no significant difference outside enteric infection
• Lazarus B, Chen Y, Wilson FP, Sang Y, Chang AR, Coresh J, Grams ME. Proton Pump Inhibitor Use and the Risk of Chronic Kidney Disease. JAMA Intern Med 2016 (PMID 26752337; retrieved 2 August 2026) — the chronic kidney disease association and its confounders
• Gomm W, von Holt K, Thomé F, Broich K, Maier W, Fink A, Doblhammer G, Haenisch B. Association of Proton Pump Inhibitors With Risk of Dementia: A Pharmacoepidemiological Claims Data Analysis. JAMA Neurol 2016;73(4):410–6 (DOI 10.1001/jamaneurol.2015.4791; PMID 26882076; retrieved 3 August 2026, the abstract obtained in full through NCBI efetch) — the dementia association, the study design (observational data from AOK, Germany's largest statutory health insurer, 2004–2011), and the whole conclusion paragraph (including its last sentence 「Randomized, prospective clinical trials are needed to examine this connection in more detail.」)
• Reimer C, Søndergaard B, Hilsted L, Bytzer P. Proton-pump inhibitor therapy induces acid-related symptoms in healthy volunteers after withdrawal of therapy. Gastroenterology 2009;137(1):80–7 (DOI 10.1053/j.gastro.2009.03.058; PMID 19362552; retrieved 3 August 2026) — the original rebound acid hypersecretion trial (in healthy volunteers) and its concluding sentence 「supports the hypothesis that RAHS has clinical implications」
• U.S. Food and Drug Administration, FDA Drug Safety Communication: Low magnesium levels can be associated with long-term use of Proton Pump Inhibitor drugs (PPIs) (document dated 2 March 2011): https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-low-magnesium-levels-can-be-associated-long-term-use-proton-pump (that URL is no longer available; retrieved 7 August 2026 from a site archive of 21 January 2026) — 「longer than one year」, the 「at least three months」 lower bound in the data summary and the one-week and two-week median times, that about a quarter of cases required the drug to be stopped, 「patients do not always have these symptoms」, the magnesium monitoring advice and the 「especially important」 sentence about digoxin, the instruction to patients, the difference between prescription and over-the-counter PPIs, the over-the-counter course 「every 4 months, up to 3 times per year」, the warning about use beyond the label, and that the risk of hypomagnesaemia will not be added to the Drug Facts box
• Wong WM, Lai KC, Lam KF, et al. Prevalence, clinical spectrum and health care utilization of gastro-oesophageal reflux disease in a Chinese population: a population-based study. Aliment Pharmacol Ther 2003;18(6):595–604 (DOI 10.1046/j.1365-2036.2003.01737.x; PMID 12969086; retrieved 3 August 2026) — the weekly, monthly and annual prevalence in Hong Kong in 2002, the odds ratio for non-cardiac chest pain, the absence of association with dysphagia, bronchitis, asthma, hoarseness and pneumonia, the sentence immediately following in the results paragraph of the abstract, 「Patients with gastro-oesophageal reflux disease had a significantly higher anxiety and depression score and required more days off work when compared with subjects without.」, and the concluding sentence 「The prevalence of gastro-oesophageal reflux disease was lower than that in Western populations, but carried a significant socio-economic burden in the studied Chinese population.」
• Tan VP, Wong BC, Wong WM, et al. Gastroesophageal Reflux Disease: Cross-Sectional Study Demonstrating Rising Prevalence in a Chinese Population. J Clin Gastroenterol 2016;50(1):e1–7 (DOI 10.1097/MCG.0000000000000304; PMID 25751371; retrieved 2 August 2026) — 3.8% on the Montreal definition in 2011, and the rise of 1.3 percentage points over ten years
• Mak ADP, Wu JCY, Chan Y, Tse YK, Lee S. Associations between Gastro-oesophageal Reflux Disease, Generalised Anxiety Disorder, Major Depressive Episodes, and Healthcare Utilisation: a Community-based Study. East Asian Arch Psychiatry 2019;29(2):41–47 (PMID 31237245; retrieved 2 August 2026) — 4.2% weekly, 13.9% monthly, and the dose-response relationship and conclusion that those with monthly symptoms were at higher risk of GAD (p = 0.01) and MDE (p < 0.001)
• Fock KM, Talley N, Goh KL, et al. (including Wu J of the Chinese University of Hong Kong) Asia-Pacific consensus on the management of gastro-oesophageal reflux disease: an update focusing on refractory reflux disease and Barrett's oesophagus. Gut 2016;65(9):1402–15 (DOI 10.1136/gutjnl-2016-311715; PMID 27261337; retrieved 3 August 2026) — the rising prevalence in Asia, the predominance of non-erosive disease, the commonness of PPI-refractory reflux disease, and the divergence over the definition of Barrett's
• Jung HK, Tae CH, Song KH, et al. (including Wu JCY of the Chinese University of Hong Kong) 2020 Seoul Consensus on the Diagnosis and Management of Gastroesophageal Reflux Disease. J Neurogastroenterol Motil 2021;27(4):453–481 (DOI 10.5056/jnm21077; PMID 34642267; retrieved 3 August 2026) — the position on the PPI test in the setting of typical symptoms; the Asian upper reference limit for oesophageal acid exposure time of 3.2% (95% CI 2.7–3.9%) and the 4.0% Asian abnormal threshold proposed immediately afterwards with its limiting sentence 「may not be the threshold for inducing GERD symptoms」, and statement 13 (≥ 4%, moderate evidence, weak recommendation) and that consensus's own internal inconsistency about the boundary (the statement reading 「≥ 4%」 and the body 「more than 4.0%」, the difference being the point 4.0% itself, this article setting both side by side); and the three boxes of the Lyon consensus it quotes, < 4%, > 6% and inconclusive, and the 4–6% range of unproven diagnosis of the Porto consensus it quotes
• Hong Kong Society of Gastroenterology: https://www.hksge.org/ (retrieved 9 August 2026, HTTP 200) — About us, Professional Resources (e-Newsletter, Position Statement, Links), Public Education and Events & News all gone through. The Position Statement column holds only two documents (Helicobacter pylori; the management of anaemia in patients with gastrointestinal bleeding), and no GERD guideline or position statement, the website having no Guidelines column either; the only reflux material is the public education page "上班族辛酸事 – GI/GERD Public Education 2014" (https://www.hksge.org/gi-gerd-public-education-2014/ , with a handbook PDF and a GERD-R self-test questionnaire), which this article does not cite. hksgh.org is not the society's domain and does not resolve in DNS
• Cancer Online Resource Hub, Stomach cancer (carrying the position of the government's Cancer Expert Working Group on Cancer Prevention and Screening): https://www.cancer.gov.hk/en/hong_kong_cancer/common_cancers_in_hong_kong/stomach_cancer.html · Chinese version https://www.cancer.gov.hk/tc/hong_kong_cancer/common_cancers_in_hong_kong/stomach_cancer.html (page last revised 28 April 2026; retrieved 9 August 2026) — the whole of both boxes of the screening section, reproduced in both languages: that screening for stomach cancer is not recommended in asymptomatic people at average risk, and neither is screening for Helicobacter pylori; that people at increased risk (a precancerous lesion or a family history) may consider seeking a doctor's advice; and the first sentence of the section on investigating and diagnosing stomach cancer (the Chinese 「出現上述徵狀的人士，應向醫生求診，儘快接受檢查。」 and the English 「People with aforementioned symptoms should consult a doctor as soon as possible and undergo tests.」). ⚠ The headings of the second box on that page contradict each other between the languages: the English 「For symptomatic persons at increased risk」 and the Chinese 「較高風險的無症狀人士」. The body text of both boxes agrees between the languages, but the two boxes do not state the same condition: the first box's body text expressly makes being asymptomatic a condition of application (「無症狀」 twice in the Chinese, asymptomatic twice in the English), while the second box's body text states risk only, with no word about symptoms in either language. So what separates the boxes is the level of risk, but it cannot be said on that basis that neither box's body text mentions symptoms. This article reproduces both headings in both languages and reports the divergence without choosing between them; and it does not use the heading in either language alone to support an inference that the boxes are divided by symptoms. That page contains no word for "coordinated"

#Further reading

• On this platform: Reading laboratory and imaging reports
• On this platform: General practitioner vs specialist
• On this platform: Public waiting times too long? Private and PPP alternatives
• Centre for Health Protection — information related to the upper gastrointestinal tract: chp.gov.hk
• On this platform: Weight-loss drugs