Eosinophilic oesophagitis

What eosinophilic oesophagitis is

The oesophagus is a muscular tube that should do one thing without you noticing it. In eosinophilic oesophagitis a particular white blood cell, the eosinophil, builds up in its lining in response to food proteins. The lining becomes inflamed, and over years the wall beneath it stiffens and scars, so the tube becomes less elastic and narrower.

That sequence explains almost everything about the condition. Early on it causes inflammation and a vague sense that swallowing is not quite right. Later it causes a narrow, rigid oesophagus and food that physically will not pass.

It is an allergic disease, but not the kind most people picture. It is not anaphylaxis, and it is not the sort of allergy that shows up on a skin prick test. It is a slow, chronic inflammation driven by a different arm of the immune system, which is why the usual allergy tests do not find the culprit food. There is a section on that below, because it is the thing people ask about most.

It was barely recognised thirty years ago. Pooled across 40 studies covering more than 288 million people, prevalence rose from about 8 per 100,000 in 1976 to 2001 to about 74 per 100,000 in 2017 to 2022. How much of that is a real increase and how much is better recognition is not settled. Those are global figures. The only Australasian population data are in children, across seven capital cities in Australia and New Zealand, where prevalence ranged from 15 to 54 per 100,000 and incidence rose significantly at every site over ten years.

What it feels like

In adults it is mostly about swallowing:

• food that feels like it stops or slows behind the breastbone, particularly bread and meat
• needing water with every mouthful, or cutting food smaller than other people do
• chest discomfort that is not quite heartburn and does not settle with antacids
• heartburn that has never responded properly to reflux treatment
• episodes where food stops completely and has to be removed at endoscopy

In children it looks different. Vomiting and abdominal pain are commoner, and in small children it shows up as refusing food, eating slowly or failing to gain weight. The reason for the difference is time: in a Dutch cohort of 721 patients, fibrotic changes were present in 76 per cent of adults against 39 per cent of children. Adults have usually had it longer.

When food gets stuck

A food bolus obstruction is when something, usually meat or bread, lodges in the oesophagus and will not go down. You cannot swallow your own saliva, and it is frightening. It needs an emergency department, and the food is removed at gastroscopy.

Here is the part that often gets missed. In a Victorian hospital that adopted a policy of biopsying every food bolus obstruction, eosinophilic oesophagitis was the cause in 28 per cent of 190 presentations. An earlier Victorian series found it in 33 per cent of those who were biopsied. Roughly a quarter to a third.

So if food has been removed from your oesophagus and nobody took biopsies, the question of why it happened is still open. That is worth raising with your GP rather than waiting for it to happen again. In the earlier series, 12 of the 17 people found to have the condition had already had a previous episode.

One Australian detail: these presentations are commoner in spring and summer. In the Victorian series, 39 per cent of spring presentations were attributable to this condition, which fits with pollen driving it.

How it is diagnosed

Three things are needed, and all three matter:

• symptoms of the oesophagus not working properly
• at least 15 eosinophils in a high power field on biopsy
• a proper look at the other things that cause eosinophils to collect there

The biopsies are the diagnosis. There is no blood test for it.

Why biopsies are taken even when it looks normal

At gastroscopy the lining often shows rings, fine lengthwise furrows, white specks, swelling or a narrowing. Pooled across 100 studies and 4,678 patients, furrows were seen in 48 per cent and rings in 44 per cent.

But the lining can look entirely normal in someone who has the condition. Across all the studies in that review that was 17 per cent, falling to 7 per cent when only the prospective studies were counted, so the best estimate is around 1 in 14 rather than 1 in 6. Either way it is not rare. That is why biopsies are taken from anyone whose symptoms suggest it, regardless of how the lining appears. An endoscopy report saying the oesophagus looked normal does not settle the question unless biopsies were taken.

Several are needed, from more than one level, because the inflammation is patchy. A single biopsy picks it up about 55 per cent of the time. Five or more approaches certainty, and the guidelines ask for at least six from two levels. If you have had a gastroscopy with one or two biopsies and the result was negative, that is a less conclusive result than it sounds.

What changed in 2018, and why your older letters may read differently

Until 2018 the pathway involved a trial of acid suppression first. If the eosinophils settled on it, the diagnosis was called something else, and eosinophilic oesophagitis was considered excluded.

An international consensus retired that. Acid suppression reduces the eosinophils in a good proportion of people who genuinely have this condition, so responding to it says nothing about the diagnosis. Acid suppression is now one of the treatments rather than a test.

This matters practically. If you were told years ago that your oesophageal eosinophils were a reflux problem because tablets fixed them, that conclusion was drawn under the old rules and may be worth revisiting.

Why allergy testing does not find your trigger foods

This is the commonest expectation people arrive with, and the answer is firmer than most expect.

The best study on it is Australian. A Melbourne group took consecutive adults who had chosen dietary treatment and ran five different allergy tests on every one of them: skin prick testing, patch testing, blood specific IgE, basophil activation testing and food-specific IgG. Everyone then went through a full elimination diet with foods reintroduced one at a time, each followed by an endoscopy and biopsies, which is the only way to know for certain what a person's triggers are.

None of the five tests predicted the actual trigger foods. The authors concluded that no currently available allergy test identifies food triggers in this condition, and that elimination and rechallenge with biopsies remains the only reliable method.

The numbers back it up from another direction. Pooled across 43 studies and 2,825 patients, diets chosen by allergy testing achieved remission in 39.5 per cent of people, the worst result of any dietary approach tested. Removing a single food, chosen empirically rather than by any test, did better at 46.4 per cent.

Allergy testing is still worth doing if you have immediate reactions to a food, hives or swelling within minutes, because that is a separate condition that can coexist. It just will not tell you what is driving the oesophagus.

Why it should not be left alone

This is the part of the page that matters most, and it rests on three independent cohorts on three continents that all found the same thing.

In 200 Swiss patients the median time from first symptom to diagnosis was six years. Strictures were present in 17 per cent of those diagnosed within two years and in 71 per cent of those who had waited more than twenty.

In 721 Dutch patients the risk of a stricture rose by 9 per cent for every additional year the condition went undiagnosed. Food impactions rose from 24 per cent to 57 per cent across the same range.

In an American series, mean delay was 5 years in those whose oesophagus was still a normal calibre and 14.8 years in those whose oesophagus had narrowed below 10 mm.

The inflammation is treatable. The scarring it leaves is not. That is the whole case for not putting this off, and it is why food that has stuck once is worth investigating rather than working around.

The treatments

A/Prof Kariyawasam teaches treatment here as the three D's: drugs, diet and dilatation. They are not a menu to choose one item from. Drugs are the main treatment and the other two sit alongside.

• Drugs. A swallowed steroid taken together with a proton pump inhibitor. This is the usual starting point for almost everyone, because the steroid clears the inflammation in the large majority of people, a better result than anything else achieves. Using a proton pump inhibitor on its own is uncommon in this practice.
• Diet. Added for some people, particularly those who want to find and remove the trigger rather than suppress the response indefinitely. Used on its own it is also uncommon.
• Dilatation. Stretching the oesophagus, where a narrowing has already formed. It is done alongside the drugs, never instead of them.

A swallowed steroid

A corticosteroid taken so that it coats the oesophagus on the way down, rather than being inhaled or absorbed into the body. It treats the lining directly, which is why the dose that works has little effect elsewhere.

In a randomised trial of 88 adults, the inflammation cleared in 93 per cent on treatment against none on placebo at six weeks, and the stricter target of symptoms and biopsies both clear was met in 58 per cent against none. By twelve weeks 85 per cent had reached remission.

Australia now has a purpose-made product for this, a tablet that dissolves in the mouth rather than being swallowed whole. It is registered here and subsidised for adults, which is a real advantage over the older approach of swallowing an asthma preparation. That older method is still what is used in children, for whom no registered product exists.

The subsidy runs in three stages, and knowing them explains why the follow-up matters so much. To start, a specialist has to confirm at least 15 eosinophils in a high power field on biopsy, which is the same threshold that makes the diagnosis, so anyone with confirmed disease qualifies. That first course is limited to 90 days. Continuing beyond it requires evidence that it worked. Staying on it long term requires documented remission on repeat biopsies, defined as fewer than 5 eosinophils per high power field. This is the practical reason the repeat gastroscopy is not optional, and it is covered in its own section below.

Side effects: thrush in the mouth or oesophagus in about 1 in 20 people on the initial course and 1 in 7 on long-term treatment, easily treated. Concerns about suppressing the adrenal glands come up, and are worth putting in proportion. Biochemical changes have been reported in a minority of people in uncontrolled studies, but none of seven randomised trials found a significant difference against placebo.

The proton pump inhibitor that goes with it

Despite this not being a reflux problem, acid suppression settles the inflammation in roughly half of people on its own. Pooled across 33 studies and 619 patients, the biopsies cleared in 50.5 per cent and symptoms improved in 60.8 per cent. It is simple, cheap and well tolerated, which is why it is given alongside the steroid as a matter of course here rather than being held back.

Used on its own it is a weaker treatment, which is why that is uncommon in this practice. It is also worth knowing that the authors of that review flagged the evidence as of poor quality with publication bias, so half is an approximate figure rather than a precise one. Twice-daily dosing did better than once daily in the pooled figures, though not by enough to be statistically significant. In Australia these medicines are subsidised for reflux, so their use for this condition sits outside the subsidised indication.

Diet

Removing the foods driving the inflammation works, and it is the only treatment that addresses the cause rather than suppressing the response. Pooled across 43 studies, dietary treatment achieved remission in 60.6 per cent.

The field has moved away from removing six foods at once. Current practice starts narrow and steps up only if needed. In a study of 130 patients, removing just milk and wheat achieved remission in 43 per cent, rising to 60 per cent with four foods removed and 79 per cent with six. Among those who responded to the narrow diets, 92 per cent turned out to have only one or two trigger foods. Milk alone was the trigger in 18 per cent of adults and a third of children.

The honest cost is the endoscopies. Finding out which food is responsible means removing foods, confirming on biopsies that the inflammation has settled, then reintroducing foods one at a time with a gastroscopy after each. A full six-food cycle runs to around eight gastroscopies over a year or more. Starting narrow cuts that substantially. This is why many people, told the full picture, choose a tablet.

A dietitian experienced in this makes a large difference, both to whether the diet is actually followed and to whether nutrition holds up. Medicare covers a limited number of allied health visits a year on a GP chronic disease management plan, and that referral comes from your GP rather than from the specialist.

Stretching a narrowing

Where scarring has already narrowed the oesophagus, it is stretched at gastroscopy with a balloon or a dilator. Pooled across 14 studies and 1,607 patients, symptoms improved in about 85 per cent over a median follow-up of a year.

It used to have a reputation for being risky. With modern technique that is not borne out. In a separate review of 2,034 dilatations the perforation rate since 2009 has been 0.03 per cent, none needing surgery and none fatal. Chest discomfort for a day or two is common, in around 1 in 27. Small tears in the lining occur in about a fifth of cases and are the intended effect rather than a complication.

The important limitation: stretching treats the narrowing, not the disease. It does nothing to the inflammation, which will keep producing scar tissue if it is not treated. It is done alongside the drugs, never instead of them.

A/Prof Kariyawasam performs oesophageal dilatation. There is more on the technique on the page about stretching narrowings.

The injection used overseas

A biologic injection is approved in some countries for this condition, and the 2025 American guideline suggests it for people aged 12 and over whose condition has not responded to acid suppression. It is not available for eosinophilic oesophagitis in Australia. It is mentioned here because people read about it overseas, not because it is an option in this country at present.

What shapes the plan

The drugs are the constant. What varies is how much else goes with them, and what usually decides that:

• How much you want to avoid long-term medication. Diet is the option that treats the cause. It asks a great deal more in the first year and then often a great deal less.
• How restrictive the diet turns out to be. Most people have only one or two triggers, and quite often it is just milk. That is a different proposition from cutting out six food groups permanently.
• Appetite for gastroscopies. Diet costs more of them, at least to begin with.
• Whether a narrowing has already formed. If it has, dilatation is likely to be needed as well, alongside the drugs rather than instead of them.

Nothing here is a one-way door. Diet can be added later, or dropped if it turns out to be more trouble than it is worth, and the dose of the steroid is adjusted over time. The plan is expected to change.

The follow-up gastroscopy, and why it is not optional

There are two checks, and they do different jobs. At about three months the response is reviewed clinically: is it working, are you swallowing better, has the food stopped sticking. That review is what allows treatment to continue past the first course. Then, before treatment moves to its long-term phase at around twelve months, the gastroscopy is repeated and biopsies are taken again.

The biopsy is the one people skip, and there are three separate reasons not to.

Feeling better does not mean it has settled. In 269 adults followed prospectively, symptoms identified whether the inflammation had cleared with only about 62 per cent accuracy. People adapt how they eat without realising it, so symptoms understate what is happening in the lining.

The result decides what happens next. Not everyone clears the inflammation on the treatment they start on. Without biopsies there is no way to know which group you are in, and untreated inflammation carries on quietly producing scar tissue.

In Australia it is also what keeps the medicine coming. The subsidy runs in three stages: an initial course of up to 90 days, a continuing phase of up to 36 weeks once a response has been shown, and then long-term treatment. Moving to that last stage requires documented remission on biopsies, meaning fewer than 5 eosinophils per high power field. That is why the twelve month gastroscopy is not optional. It is worth knowing before you start rather than afterwards.

Staying well

Treatment brings this into remission. It does not cure it.

That is clearest in the maintenance trial, where 204 adults already in remission were randomised to continue treatment or to a placebo for up to a year. Around three quarters of those who continued stayed in remission against 4 per cent of those who stopped. Half of those who stopped had relapsed within three months.

So this is managed long term rather than cured, in the way asthma is. The dose is usually the lowest that holds it, and the plan is reviewed rather than set once.

What it does not do

Worth saying plainly, because the word allergic and the word inflammation together make people assume the worst.

• It is not a cancer, and the overall cancer risk is not raised. A Swedish national study of 1,580 people with the condition, matched against 7,533 from the general population and followed for a median of seven years, found no overall increase in cancer, and none in skin, lung, breast or blood cancer. The same study did find oesophageal cancer and Barrett's oesophagus more often in the group with this condition, but on very small numbers: two oesophageal cancers against one. The authors say plainly that the numbers are too small, and the range around them too wide, to draw a conclusion from, and this is not a condition for which cancer surveillance is recommended. It is mentioned here because it is in the paper, not because it changes what is advised.
• It does not spread. Thirty adults followed for an average of seven years had inflammation that stayed in the oesophagus and did not involve the stomach or the bowel.
• It does not cause malnutrition in adults. Those same thirty remained in good health and stable nutritional state throughout. Small children who refuse food are a different and real problem, and are managed differently.
• It is not life-shortening. The thing worth preventing is the scarring, not a threat to life.

ausEE is the Australian patient organisation for eosinophil-associated diseases and is a sensible starting point for anyone who wants to talk to others managing the same thing. The Australasian allergy society publishes a management plan and plain-language guides to the elimination diets used here.

Where eosinophilic oesophagitis is managed in Sydney

A/Prof Viraj Kariyawasam is a consultant gastroenterologist and hepatologist with an advanced endoscopy fellowship. Food that sticks, and eosinophilic oesophagitis, are assessed at consultation in Norwest, Rouse Hill, Pennant Hills and Macquarie Park.

The gastroscopy with biopsies, and dilatation where a narrowing has formed, are performed at Norwest Private Hospital, Macquarie University Hospital and Pennant Hills Day Surgery. A referral from your GP is needed for a consultation.