Celiac, Fiber, and Why Gluten-Free Diets Fall Short

Lit Review Friday · Celiac, Fiber, and Why Gluten-Free Diets Fall Short · 2026 · 42 minute read

The Experiment Nobody Controlled

Early in my career I worked a shift with a naturopath who would not have a conversation with a patient who was not willing to take gluten out of their diet. Not would not treat them. Would not even see them as a patient. That was the entry fee.

And here is the uncomfortable part. It mostly worked. People took gluten out and they felt better. He felt justified. And some real patient harm was done along the way, in a catch-all, ableist approach that treated every body as the same body. It is also, I suspect, why gluten-free became a punchline in jokes about California.

The problem is what else left with it. If you eat a standard American diet and I remove the gluten, I have removed most of what you were eating. The bread, the pasta, the crackers, the cereal, the beer, the thing in the vending machine at four o'clock. You feel better. And neither of us has any idea whether the gluten had anything to do with it.

Two variables. Nobody controlled the second one. That sentence explains most of the last forty years, and it explains it in both directions at once, which is the part that usually gets left out.

Both readings come from the same missing control. Neither is a story about one side being smarter than the other.

---

Is Non-Celiac Gluten Sensitivity Real?

Start with the history, because the history is not flattering to anyone.

The phrase appears in The Lancet in 1978, as a case-report letter with a question mark in its title (Ellis and Linaker, 1978). Two years later, a group in Birmingham published eight adult women with abdominal pain and chronic diarrhea that was, in their words, often incapacitating and frequently nocturnal. The patients improved dramatically on a gluten-free diet, relapsed on challenge, did not have celiac disease, and had minor jejunal changes that normalized on the diet (Cooper et al., 1980).

Read that list against the four criteria the field would formally adopt in 2015: symptom response to withdrawal, relapse on challenge, celiac excluded, objective mucosal change. Cooper's group had all of it in 1980, in eight women, and the field left it alone for three decades.

The turn came in February 2011, at a consensus conference in London that produced the modern nomenclature (Sapone et al., 2012). A second meeting in Munich in 2012 chose the name, and the paper records the argument openly: some panelists preferred non-celiac wheat sensitivity, because gluten may not be the operative molecule. That disagreement was never resolved. It was recorded and left standing.

The most useful sentence in that document is the field describing itself:

Then the careful trials arrived, and complicated it again

In 2011, Biesiekierski and colleagues ran a double-blind placebo-controlled trial and found gluten did cause symptoms in people without celiac disease. Thirteen of nineteen on gluten reported inadequately controlled symptoms against six of fifteen on placebo. They found no mechanism, and said so: non-celiac gluten intolerance may exist, but no clues to the mechanism were elucidated.

Two years later the same laboratory reversed itself with a better design. When participants were placed on a low-FODMAP diet first, removing fermentable carbohydrates before the gluten challenge, symptoms worsened equally on gluten and on whey protein. Gluten-specific effects appeared in 8 percent of participants (Biesiekierski et al., 2013).

Then the trial that matters most. Skodje and colleagues gave 59 people who were certain gluten was their problem three kinds of muesli bar in a blinded crossover: gluten at 5.7 grams, fructans at 2.1 grams, or placebo, each for seven days.

The obvious misreading of that result is that the patients invented their symptoms. That is exactly wrong, and it is the most important distinction in this entire article. Their symptoms were real, measurable, and reproducible under blinding. What was wrong was the attribution. Being wrong about the molecule is a completely different sentence from imagining the whole thing.

Zanini and colleagues demonstrated the same thing from another angle. They gave 35 patients blinded flour and asked a simpler question: could they tell which one contained gluten? Twelve identified it correctly. Seventeen were confident the gluten-free flour was the gluten flour, and had significantly worse symptom scores while eating it. Half the group was confidently wrong, and measurably unwell while being wrong.

And the counterweight, which deserves equal billing. Uhde and colleagues found that people reporting wheat sensitivity without celiac disease had elevated soluble CD14, lipopolysaccharide-binding protein, antibody reactivity to bacterial products, and elevated fatty acid-binding protein 2, a marker of intestinal epithelial cell damage. The markers normalized in a subgroup who removed wheat. You cannot produce epithelial damage markers through expectation.

So both are true at once: most people who identify as gluten sensitive do not react to gluten specifically under blinding, and a minority demonstrably do.

What clinicians on both sides actually see

Ask any functional medicine provider or gastroenterologist who works in this space and you will hear the same thing: there are discrete people who clearly react to wheat, and they are not imagining it. The mechanism is unknown. The candidates are several and not mutually exclusive. Fructans, from Skodje's trial. Amylase-trypsin inhibitors, a family of wheat proteins that activate innate immunity through toll-like receptor 4 (Junker et al., 2012). Other components and additives in processed wheat products. And cross-reactivity in a person whose gut lining is already inflamed and letting things through. The field accepted the patients before it found the molecule, which is why the 2015 Salerno criteria exist: a blinded, placebo-controlled crossover challenge at 8 grams of gluten a day, one week on, one week washout, with a 30 percent change in one to three patient-nominated symptoms counting as a positive (Catassi et al., 2015).

The most recent trial comes from the same McMaster group as this week's paper. Twenty-eight adults with irritable bowel syndrome who had improved on a gluten-free diet were challenged blind with wheat, gluten, and a gluten-free sham bar. Symptoms worsened in 39 percent on wheat, 36 percent on gluten, and 29 percent on sham, with no significant difference (Seiler et al., 2025). The authors' conclusion is the right one for both sides: identify the subset who benefit from restriction, and take the stigma off wheat for everyone else.

Zonulin, leaky gut, and the dial we lost the ability to read

The functional medicine version of this story ran through Alessio Fasano's zonulin work, and the early science was beautiful. Gliadin, a component of gluten, binds a receptor called CXCR3 on the gut lining and triggers release of zonulin, a protein that loosens the tight junctions between cells. In mice lacking that receptor, gliadin could not raise permeability (Lammers et al., 2008). Zonulin itself was identified as pre-haptoglobin-2, an uncleaved precursor of a common blood protein (Tripathi et al., 2009). In duodenal biopsy explants, gliadin increased permeability in tissue from everyone, and more so in active celiac and gluten-sensitive tissue than in celiac tissue in remission (Hollon et al., 2015). Fasano's 2011 review called zonulin the biological door to inflammation, autoimmunity and cancer, and that title is where a great deal of the leaky-gut era came from (Fasano, 2011).

What happened next is a specific, fixable kind of failure. The commercial test that most published zonulin studies relied on turned out not to detect pre-haptoglobin-2 at all. In 376 people, kit "zonulin" did not track haptoglobin genotype, which it must if it measures the real protein, and the captured protein looked like properdin, a complement factor (Scheffler et al., 2018). Fasano was a co-author on that paper, which is why this is not an attack story. Independent groups then showed the kits had been built against a sequence unrelated to zonulin, and that their readings do not correlate with functional permeability (Massier et al., 2021). So we have twenty years of published zonulin numbers where nobody can say for certain what was in the tube.

The tide is real. The marker is still standing. Nobody can read the scale.

That is not the same as leaky gut being fake. The barrier is real. The tight junctions are real. Permeability rises in defined stress states, and it is elevated at celiac diagnosis and normal again in most people after a year on the diet. What we lost is the ability to read the dial. And the line I would hand every clinician comes from Camilleri's review for clinicians: no disease is cured by normalizing barrier function alone, and claims in the public domain about repairing the barrier with exclusions or supplements need confirmation before anyone endorses them (Camilleri, 2019).

---

What an IgG Food Panel Actually Measures

IgG food panels, and gluten-free recommendations for nearly everyone. That was my field.

The intuition behind the panel is obvious. You have antibodies to eggs, so you must be reacting to eggs. But that is not what the antibody indicates. The European Academy of Allergy and Clinical Immunology Task Force concluded in 2008 that food-specific IgG4 represents a physiological response of the immune system to food exposure, and that for IgG and IgG4 this may be the normal human response. The American Academy of Allergy, Asthma and Immunology endorsed that position in 2010, and the Canadian society followed in 2012.

Read that carefully. The antibody plausibly marks exposure, and possibly tolerance. It may be pointing the opposite direction from the one it was being read in.

The testing has moved since I trained. IgG was the standard panel then. Some labs now measure IgG4, IgA and IgE alongside IgG, and some report at the level of individual food peptides. That is movement in the right direction, and none of it has yet been validated against a gold-standard challenge the way the older tests were, so I read those panels as a hypothesis generator, not a verdict. And sometimes the old panel did help, when there was one discrete reaction to find. Where it fell apart was the person whose whole gut lining was inflamed and letting things through, who ate the same four foods every day, and whose four foods came back positive. Were those foods driving an immune response, or was the immune system keeping a record of what walked past it?

Now watch the loop close. The panel flags twenty foods, or just the two that are in every processed product. You remove those foods, which is most of a normal diet. You feel better. Feeling better confirms the panel. Nothing anywhere in that sequence ever tests whether the antibody meant anything. The EAACI report predicted the outcome precisely, warning that panel testing produces multiple positives leading to over-restriction of the diet.

Somebody did test it, and it was my own school

In 2018, researchers at the Helfgott Research Institute, the research arm of the National University of Natural Medicine in Portland, ran a randomized controlled trial of exactly this intervention. Participants were randomized two to one. One group received an elimination diet built from their IgG results with health-coach support. The other group received nothing, and were not shown their results.

A naturopathic research institute randomized the intervention, and the antibody levels moved almost as much in the people who changed nothing. That is not critics attacking naturopathic medicine from outside. That is naturopathic researchers running the study properly and publishing what they found, which is the only thing that ever actually settles a question like this.

A separate validity test makes the point from the other direction. Philpott and colleagues studied adults with eosinophilic esophagitis, a condition where the true food trigger can be confirmed by endoscopy and histology after structured elimination and rechallenge. Against that gold standard, they tested five allergy modalities including food-specific IgG. Their conclusion: no allergy test could accurately predict actual food triggers.

Why the food-panel era belongs in a celiac article

It all landed in the same era. Everybody was running food panels. Gluten was the hottest new sensitivity, with dairy right behind it, and dairy is its own can of worms for another day. Here is what makes it a conundrum and not a cautionary tale. A lot of those people really were celiac, or really were gluten sensitive. Then there was a whole other class of people who got swept up, by a panel or by a clinician who just said take it out, and who took it out and felt better anyway.

So if you are a functional medicine clinician coming into this space now, or a patient who never got the history, this is why everybody got put on a gluten-free diet. There were real outcomes behind it, clinicians liked it, and the mechanism had holes in it. All of that is a separate story from the true celiac patients, who got eyes rolled at them for decades, who also benefited from the movement, and whose microbiomes we are only now learning are different in ways we are only starting to measure.

How Common Are Celiac Disease and Gluten Sensitivity?

This is the part to know before deciding what to test for, because the two numbers behave completely differently. One of them is measurable and stable. The other is a moving target, and what moves it is not biology.

The United States data shows the same split from another angle. Between 2009 and 2014 the rate of celiac disease held flat at roughly 0.7 percent, undiagnosed celiac disease halved from 0.6 to 0.3 percent, and the number of people without celiac disease avoiding gluten tripled, from 0.5 percent to 1.7 percent (Choung et al., 2016). Three things at once: the disease stayed put, we got better at finding it, and gluten avoidance ran far ahead of both.

So about 1 in 140 people has celiac disease and most of them do not know it, while a much larger group has removed gluten without ever finding out whether they needed to. Those two facts point at the same fix.

The one thing I would tape to the wall

Do not go gluten-free before you get tested. Gold-standard testing for celiac disease is a sequence. Blood first: tissue transglutaminase IgA antibodies, with a total IgA level alongside, because IgA deficiency makes the antibody test read falsely negative. Then an upper endoscopy with multiple biopsies of the small intestine, taken from more than one site including the duodenal bulb, because the damage is patchy and a single sample can miss it (Rubio-Tapia et al., 2023). If you are already scheduled for an upper endoscopy for any reason, ask for those biopsies. And every step of that sequence requires you to be eating gluten. Take it out first and you are choosing between three things: staying undiagnosed, eating gluten on purpose later until you are sick enough to test, or never knowing.

And knowing matters more than people think, because a celiac diagnosis is not just a diet. Your first-degree relatives get screened. Somebody watches your bone density. The instinct to just take it out and see is generous, and it can cost somebody the ability to ever find out.

---

What They Did

Which brings us to the paper that prompted this episode. Wulczynski and colleagues, from Elena Verdú's and Alberto Caminero's groups at McMaster, with Lawrence David at Duke and Joseph Murray at Mayo, published in Nature Communications in 2026.

The human arm recruited three groups: 16 patients with celiac disease diagnosed within the previous three months, 11 patients treated with a gluten-free diet for at least two years, and 26 healthy controls. They collected duodenal aspirates during endoscopy for 16S rRNA gene sequencing, fecal samples for short-chain fatty acid measurement by gas chromatography mass spectrometry, food frequency questionnaires, and fecal plant DNA sequencing to characterize what people were actually eating, not what they reported eating.

The mouse arm used non-obese diabetic mice carrying the human celiac risk gene HLA-DQ8, sensitized with digested gluten. After establishing that histology normalized by week 10 of a gluten-free diet on its own, they tested two fiber supplements during recovery: inulin at 52 grams per kilogram of diet, and HylonVII, a high-amylose resistant starch, at 136 grams per kilogram. They then ran germ-free experiments to test whether the effects required a microbiota at all, and colonized germ-free mice with a defined ten-member Prevotellaceae consortium.

One methodological point deserves naming up front, because it constrains everything that follows. They could not measure microbial enzymes directly. Duodenal aspirates are low-biomass and heavily contaminated with host DNA, which made shotgun metagenomics unworkable, so enzyme gene abundance was predicted from 16S sequencing using PICRUSt2. The authors title their first figure "Predicted decreases in duodenal microbial saccharolytic enzymes." They also could not measure short-chain fatty acids in the human small intestine at all, for the practical reason that patients arrive for endoscopy fasted and concentrations are low.

---

What They Found

That final line is the one I keep returning to. Butyrate did not recover in the treated group. These were patients on a gluten-free diet for more than two years, doing well by the measures that usually define success, and the metabolite their colonocytes run on was still depressed.

Butyrate is not a minor metabolite. The cells lining the colon burn it as their main fuel. It inhibits histone deacetylases, which changes gene expression in those cells and in immune cells, and through that route it supports the regulatory T cells that keep the immune response to food and microbes in check. It signals through receptors on gut and immune cells, and it tightens the junctions of the barrier (Koh et al., 2016). Those are the reasons a persistent butyrate deficit in a treated patient is an immune finding and not only a digestive one. We spent a whole episode on it: Why Butyrate Never Reaches Your Bloodstream.

There is a second finding tucked in the dietary data that sharpens it. The treated patients had the lowest fiber intake of all three groups, with 86 percent falling below the minimum recommendation, and yet had modestly higher SCFA than the newly diagnosed patients. The authors read this as partial recovery of saccharolytic capacity despite suboptimal substrate. It is also a fairly direct statement that these patients are running a fermentation system on short rations.

This is not an isolated result. A 2026 cross-cohort synthesis pooling more than 900 samples across global datasets, with Verdú as a co-author, found celiac disease is not characterized by large diversity changes, but by subtle and consistent reductions in butyrate producers including Faecalibacterium, Prevotella, Agathobacter and Gemmiger. Those reductions were present before and during disease, and persisted on a gluten-free diet (Prendergast et al., 2026). Two different designs, same direction.

---

Which Fibers Are Affected, and Which Foods?

The paper names two enzymes, which makes this unusually answerable. Alpha-amylase, which breaks down starch. And fructan beta-fructosidase, which the authors describe as an inulin-like fiber degrading enzyme.

Two enzymes, two families of food.

Look at the end of the fructan list. Wheat, rye and barley are significant fructan sources, and in a Western diet wheat is also the dominant source of arabinoxylan, the cereal fiber that Prevotella specializes in degrading.

So the only treatment we have for celiac disease removes a major source of exactly the fiber these patients have the least predicted capacity to process, from a gut that has already lost the organisms that process it.

And the replacements do not fill the gap. Rice flour, corn starch, tapioca and potato starch are rapidly hydrolyzed by human enzymes high in the small intestine. Very little reaches a microbe at all. Pooling 38 studies and 2,114 patients, adults on a gluten-free diet consume about 18.9 grams of fiber daily against a 25 to 30 gram reference, with saturated fat at 13.2 percent of energy and vitamin D insufficient across every age group (Gessaroli et al., 2023).

One complication, because the deficiency is not inevitable. When Italian researchers built a proper gluten-free food composition database and modeled national dietary guideline menus using gluten-free substitutions, the resulting menus met requirements, and were higher in polyunsaturated fat and vitamin E. Their conclusion was that patients can meet requirements by substituting gluten-free cereal products while following general population recommendations. The deficit is a property of how the diet is executed, through reliance on refined processed substitutes, not an inescapable consequence of removing gluten.

---

Can You Just Add the Fiber Back?

This is the obvious question, and the paper went and tested it in mice. The first result is the one worth sitting with.

They gave inulin to germ-free mice, animals with no microbiome at all. Nothing happened. No rise in small intestinal short-chain fatty acids. The fiber passed through an animal with no one there to ferment it.

Then they colonized germ-free mice with a ten-member Prevotellaceae consortium, the family depleted in the patients, and gave inulin. The short-chain fatty acids appeared. Take the workers out and the fiber does nothing. Put the workers back and the fiber works.

In the gluten-sensitized mice, inulin supplementation during the gluten-free diet improved every histological parameter against no added fiber, and accelerated mucosal recovery. In mice, with an intact microbiota. That distinction matters and I will come back to it.

The second fiber is where it gets practical. HylonVII, the high-amylose resistant starch, did not do the same thing. It failed to normalize intraepithelial lymphocytes, and it shifted the community in a specific and unwelcome direction: higher Helicobacter and higher Escherichia and Shigella. Across all groups, Escherichia and Shigella abundance correlated negatively with small intestinal short-chain fatty acids.

So "add fiber" is not one instruction. The specific fiber mattered, and the wrong one was not neutral.

And inulin itself is not a gentle fiber. It is a fructan, it ferments fast, and it is exactly the class these patients are least equipped to handle. The paper's own answer worked in mice with a working microbiota, using the fiber family most likely to make a person with a depleted one feel terrible. In people it has been tried once at any scale, in 34 children on a gluten-free diet, and it moved the chemistry without measuring whether anyone felt better. The details are in the limitations below.

The microbes tune the host, not just the metabolite pool

One more result deserves attention because it changes the shape of the mechanism. Inulin increased expression of the short-chain fatty acid receptors GPR41 and GPR43 at most small intestinal sites in conventional mice. In germ-free mice, it did not.

The microbes are not only supplying the signal. They are changing how loudly the host is listening for it. That echoes work we covered in the episode on tryptophan derivatives, where wiping out butyrate producers with antibiotics reduced signaling through PPAR-gamma in the gut lining, independent of the ligand itself.

The small intestinal SCFA rise was driven by acetic acid, with no differences in propionic or butyric acid. The magnitude of the SCFA increase was roughly ten times higher in feces than in the small intestine, with different dominant species at each site. The small intestine and the colon are different rooms running different chemistry, and most of what we say about butyrate is quietly a colonic statement.

---

Why Do Some People Tolerate Fiber Worse Than Others?

This is the clinical question, and it usually gets asked wrong.

Nobody in this literature is arguing about whether humans need fiber. We need fiber. The argument is about how you deliver it to someone who does not currently have the machinery to use it.

The answer, in my clinical experience and consistent with the trial evidence, is low and slow. And you choose the fiber for tolerance first and fermentability second, which is backwards from how most people approach it. Something that ferments gently and gradually instead of all at once in the first place it meets a bacterium.

Partially hydrolyzed guar gum is the one I reach for, for reasons I lay out in the practical section at the end of this article. The short version: it is a soluble, low-viscosity, non-gelling fiber that ferments slowly, and fermentation still yields acetate, propionate and butyrate. You get the metabolites without the fireworks.

This is the same shape as a finding we covered in an earlier episode on the fiber paradox: in the Stanford work on high-fiber and high-fermented-food diets, people with diverse microbiomes saw inflammatory markers fall, while people with low diversity saw them rise on the same intervention. You are not tiptoeing around a sensitive patient. You are rebuilding a workforce before you increase the shipment.

---

The Honest Limitations

The cohort is small and the authors say so. Their words: our cohort was small and exploratory. The duodenal aspirate arm had 17 controls, 7 active celiac patients and 7 treated patients. The food frequency questionnaire arm had 5, 6 and 7. Do not lean on the dietary intake comparisons.

The enzymes were predicted, not measured. PICRUSt2 infers gene content from 16S taxonomy. It is a reasonable tool and it is not a measurement. Every statement in this article about alpha-amylase and fructan beta-fructosidase inherits that caveat.

Human small intestinal short-chain fatty acids were never measured. The human SCFA data is fecal, which reports on the colon. The small intestinal SCFA data is entirely from mice.

The mucosal healing result is a mouse result. Sensitized HLA-DQ8 mice, six weeks, six animals per group. Neither fiber normalized anti-gliadin or anti-transglutaminase antibodies at six weeks, although all mice became seronegative by twelve.

And the one that matters most for anyone tempted to act on this. Prebiotic fiber has been tested in celiac disease in humans exactly once at any meaningful scale: a randomized trial of oligofructose-enriched inulin at 10 grams daily for 12 weeks in 34 children. Bifidobacteria increased, total short-chain fatty acids rose about 31 percent including butyrate, and serum hepcidin fell 61 percent. But there was no change in ferritin, hemoglobin or C-reactive protein, and no clinical outcome endpoint was measured at all. Nobody recorded whether a single child felt better, healed faster, or absorbed more iron.

The intervention moved the number that was broken. That is the right target. It is not evidence of benefit, and the authors say as much: these results warrant further investigation in a larger cohort. There is no adult trial. There is no trial powered on symptoms, histology, or quality of life. There is no trial in non-responsive celiac disease.

---

Frequently Asked Questions

Should I go gluten-free before getting tested for celiac disease?

No. Both blood tests and biopsy require you to be eating gluten to be interpretable, per the American College of Gastroenterology's 2023 guidelines. If you remove gluten first, your remaining options are staying undiagnosed, or deliberately eating gluten later until you are unwell enough to test. A diagnosis carries consequences a diet alone does not, including screening for first-degree relatives and monitoring of bone health.

Can people with celiac disease eat oats?

Most can, if the oats are certified gluten-free, meaning grown and processed away from wheat, barley and rye. Pooled trial data found no effect of pure oats on symptoms, biopsy findings or antibodies over a year, and a longer review found oats maintained remission while improving iron, fiber and zinc intake. A small subset of people react to oat protein itself, so oats are added with monitoring, not assumed.

Does a low-FODMAP diet help celiac symptoms that continue on a gluten-free diet?

In a randomized trial of adults in full remission with ongoing symptoms, a moderately low-FODMAP version of the gluten-free diet reduced pain, bloating and diarrhea within a week. It works by removing fermentable fibers such as fructans, so it is designed as a short restriction phase followed by reintroduction, and the five-year follow-up found weight loss was more common in people who stayed restricted.

Why do some people feel better off wheat but test negative for celiac disease?

Wheat contains several things besides gluten. In blinded crossover testing, fructans, a fermentable carbohydrate, produced worse symptoms than gluten did in people who believed gluten was their trigger. Removing wheat removes both at once, which is why symptom improvement alone cannot tell you which component mattered.

Are IgG food sensitivity panels accurate?

The major allergy societies do not recommend them for diagnosing food intolerance. Food-specific IgG appears to reflect normal exposure to a food, not a reaction to it. In a randomized trial, antibody levels fell nearly as much in participants who changed nothing as in those following an elimination diet based on their results.

Why do some people react badly when they add fiber?

Fiber is not the active ingredient. The microbes that ferment it are. When the community that processes a given fiber is thin, more of that fiber travels further before anything acts on it, which tends to produce gas and distension instead of benefit. Baseline microbial diversity predicts who tolerates a fiber intervention and who does not.

Does more strictness with a gluten-free diet lead to better outcomes?

Adherence matters, but extreme vigilance beyond it does not appear to. Patients who were the most restrictive reported the worst quality of life while scoring highest on food-label knowledge, and showed no difference in intestinal biopsy findings, nutrient deficiencies, or bone disease compared to less restrictive patients.

---

The Bottom Line

Wolf and colleagues looked at people with celiac disease who were the most vigilant about the diet: the most careful, the most concerned about cross-contamination. Those people reported the worst quality of life, and they were not the least informed. They scored higher on food-label knowledge than everyone else. And when a Vanderbilt group compared patients whose eating met criteria for avoidant/restrictive food intake disorder against those who did not (Bennett et al., 2022), there was no difference in biopsy activity, no difference in adherence, no difference in micronutrient deficiencies, and no difference in bone disease.

Whatever the extra restriction was buying, it was not showing up in the intestine.

My own profession got this wrong. We told people that more restriction was more virtuous. The data says the extra restriction did not make them healthier. It made their lives smaller. That is not an argument for taking gluten less seriously in people who need to avoid it. The gluten is real. The damage is real. The diet is the only treatment there is, and it works.

The precursor is necessary and it is not sufficient. You can eat all the fiber you want, and what actually decides what gets built is the environment those microbes are working in, and which microbes are there to do the work. The food is the easy part. The hard part is the ecosystem that decides what gets made out of it.

---

How Do You Add Fiber Back If You Have Celiac Disease or a Fructan Sensitivity?

This is the section I wish someone had handed me twenty years ago. It is how I think about the problem in practice, built on the trials above and on the ones below. It is general education, not individual medical advice, and the gluten-free diet stays exactly as strict as your diagnosis requires while any of this happens alongside it.

First, the two things with actual outcomes in celiac patients

Certified gluten-free oats. In a randomized, double-blind study of children newly diagnosed with celiac disease, 34 children added oats to their gluten-free diet and 37 followed a standard gluten-free diet for a year. In the standard group, fecal short-chain fatty acids fell significantly over the year. In the oats group they stayed high, with more acetate, more butyrate and more total short-chain fatty acids at twelve months than the standard group (Tjellström et al., 2014). The authors were careful to add that some children receiving oats may develop gut mucosal inflammation, which is the reason the word certified matters and the reason oats are introduced with monitoring, not on a whim.

Pull back to the whole literature and the picture holds. A meta-analysis of 28 studies, including six randomized trials with 661 patients, found no evidence that adding pure, uncontaminated oats to a gluten-free diet affected symptoms, histology, intraepithelial lymphocyte counts or serology over twelve months, with the authors rating the overall quality of evidence as low (Pinto-Sánchez et al., 2017). And a 2026 systematic review of prebiotic interventions in celiac disease, twelve studies and 1,066 participants, found that oats at 50 to 70 grams a day for up to 24 months maintained histologic remission and negative serology while improving dietary iron, fiber, thiamin and zinc intake, with no serious adverse events reported in any included study (Shahzil et al., 2026). The practical risk with oats is contamination with wheat, barley or rye during growing and processing, not the oat itself, so the label is the whole game.

A short, structured low-FODMAP phase for symptoms that persist on a strict diet. Roughly a third of people with celiac disease keep having gut symptoms after the diet has put the intestine into remission by every measure. In Oslo, van Megen and colleagues took 70 adults with biopsy-proven celiac disease, confirmed serologic and mucosal remission and at least a year of strict adherence, and randomized them to four weeks of a moderately low-FODMAP version of their gluten-free diet, about 8 grams of FODMAPs a day, or their usual diet. Symptom scores separated within one week and the gap persisted through week four, with pain, bloating, diarrhea and early fullness all improving. Constipation did not (van Megen et al., 2022).

Which is the trap from earlier in this article, arriving as a treatment. It works by taking out the fructans. Relief on one side, less fiber on the other. That is why the diet is designed in three phases: four to eight weeks of restriction, then deliberate reintroduction, then a personalized diet that puts back every fermentable food you tolerate. When the same group followed up five years later, about half of the original intervention group were still following a personalized, partly low-FODMAP diet, fructans were the most commonly identified trigger, and 44 percent of the intervention group reported weight loss against 23 percent of controls, which the authors read as a sign that the combined diet is too restrictive without dietitian follow-up (van Megen et al., 2026). The restriction phase is a diagnostic tool. It is not a place to live.

One more with an outcome, since people will ask. In a randomized, double-blind, placebo-controlled trial of 109 adults with celiac disease and persistent irritable-bowel-type symptoms despite a strict diet, six weeks of a multispecies probiotic mixture reduced symptom severity scores by about 16 percent while the placebo group worsened by about 8 percent, and bifidobacteria increased in the treated group with no adverse events (Francavilla et al., 2019). It is one trial, of one specific mixture, and it is the strongest intervention trial in this population that I could find.

Then the gentler fibers, and why partially hydrolyzed guar gum is the one I reach for

Partially hydrolyzed guar gum, PHGG, is a soluble fiber made by breaking guar gum into shorter chains so it no longer gels. It dissolves clear in water, it ferments slowly and further down the gut, and the trials behind it are in exactly the population this article is about: people with sensitive guts who need fiber and cannot tolerate the usual way of getting it.

Read those honestly. Two of the three are open-label, the placebo-controlled one moved only bloating, and none of them enrolled celiac patients. I searched for a PHGG trial in celiac disease and there is not one. What the trials do establish is the thing I care about for this population: it is tolerated by people who do not tolerate bran, it is taken at 5 to 6 grams a day, it does not need to be stopped, and it still feeds the fermentation that makes acetate, propionate and butyrate. The people who need the metabolites most are the ones who cannot get through the fireworks. This is a fiber that skips the fireworks.

How I would actually do it

Slow water, many plants, and a bank that took years to build.

And the sentence to keep from the whole episode: the people who need fiber most tolerate it worst, because the workforce that ferments it is thin. Low and slow is not timidity. It is how you rebuild a workforce before you increase the shipment.

---

References

1. Wulczynski M, Constante M, Galipeau HJ, et al. Small intestinal microbial fiber metabolism dysfunction in celiac disease. Nature Communications. 2026;17:2698. https://doi.org/10.1038/s41467-026-70644-4 · FREE FULL TEXT
2. Skodje GI, Sarna VK, Minelle IH, et al. Fructan, rather than gluten, induces symptoms in patients with self-reported non-celiac gluten sensitivity. Gastroenterology. 2018;154(3):529-539. https://doi.org/10.1053/j.gastro.2017.10.040
3. Molina-Infante J, Carroccio A. Suspected nonceliac gluten sensitivity confirmed in few patients after gluten challenge in double-blind, placebo-controlled trials. Clinical Gastroenterology and Hepatology. 2017;15(3):339-348. https://doi.org/10.1016/j.cgh.2016.08.007
4. Biesiekierski JR, Peters SL, Newnham ED, et al. No effects of gluten in patients with self-reported non-celiac gluten sensitivity after dietary reduction of fermentable, poorly absorbed, short-chain carbohydrates. Gastroenterology. 2013;145(2):320-328. https://doi.org/10.1053/j.gastro.2013.04.051
5. Zanini B, Baschè R, Ferraresi A, et al. Randomised clinical study: gluten challenge induces symptom recurrence in only a minority of patients who meet clinical criteria for non-coeliac gluten sensitivity. Alimentary Pharmacology and Therapeutics. 2015;42(8):968-976. https://doi.org/10.1111/apt.13372
6. Uhde M, Ajamian M, Caio G, et al. Intestinal cell damage and systemic immune activation in individuals reporting sensitivity to wheat in the absence of coeliac disease. Gut. 2016;65(12):1930-1937. https://doi.org/10.1136/gutjnl-2016-311964 · FREE FULL TEXT
7. Cooper BT, Holmes GK, Ferguson R, et al. Gluten-sensitive diarrhea without evidence of celiac disease. Gastroenterology. 1980;79(5 Pt 1):801-806.
8. Ellis A, Linaker BD. Non-coeliac gluten sensitivity? The Lancet. 1978;1(8078):1358-1359. https://doi.org/10.1016/s0140-6736(78)92427-3
9. Sapone A, Bai JC, Ciacci C, et al. Spectrum of gluten-related disorders: consensus on new nomenclature and classification. BMC Medicine. 2012;10:13. https://doi.org/10.1186/1741-7015-10-13 · FREE FULL TEXT
10. Catassi C, Bai JC, Bonaz B, et al. Non-celiac gluten sensitivity: the new frontier of gluten related disorders. Nutrients. 2013;5(10):3839-3853. https://doi.org/10.3390/nu5103839 · FREE FULL TEXT
11. Neuendorf R, Corn J, Hanes D, Bradley R. Impact of food immunoglobulin G-based elimination diet on subsequent food immunoglobulin G and quality of life in overweight/obese adults. Journal of Alternative and Complementary Medicine. 2018. https://doi.org/10.1089/acm.2018.0310
12. Stapel SO, Asero R, Ballmer-Weber BK, et al. Testing for IgG4 against foods is not recommended as a diagnostic tool: EAACI Task Force Report. Allergy. 2008;63(7):793-796. https://doi.org/10.1111/j.1398-9995.2008.01705.x
13. Philpott H, Nandurkar S, Royce SG, et al. Allergy tests do not predict food triggers in adult patients with eosinophilic oesophagitis. Alimentary Pharmacology and Therapeutics. 2016;44(3):223-233. https://doi.org/10.1111/apt.13676
14. Gessaroli M, Frazzoni L, Sikandar U, et al. Nutrient intakes in adult and pediatric coeliac disease patients on gluten-free diet: a systematic review and meta-analysis. European Journal of Clinical Nutrition. 2023;77(8):784-793. https://doi.org/10.1038/s41430-023-01280-0
15. Prendergast MB, et al. Cross-cohort analysis of the gut microbiome in coeliac disease. Communications Medicine. 2026;6(1). https://doi.org/10.1038/s43856-026-01627-1 · FREE FULL TEXT
16. Rubio-Tapia A, Hill ID, Semrad C, et al. American College of Gastroenterology guidelines update: diagnosis and management of celiac disease. American Journal of Gastroenterology. 2023;118(1):59-76. https://doi.org/10.14309/ajg.0000000000002075
17. Niv E, Halak A, Tiommny E, et al. Randomized clinical study: partially hydrolyzed guar gum (PHGG) versus placebo in the treatment of patients with irritable bowel syndrome. Nutrition and Metabolism. 2016;13:10. https://doi.org/10.1186/s12986-016-0070-5 · FREE FULL TEXT
18. Zhou J, Ho V. Role of baseline gut microbiota on response to fiber intervention in individuals with irritable bowel syndrome. Nutrients. 2023;15(22):4786. https://doi.org/10.3390/nu15224786 · FREE FULL TEXT
19. Drabińska N, Jarocka-Cyrta E, Markiewicz LH, Krupa-Kozak U. The effect of oligofructose-enriched inulin on faecal bacterial counts and microbiota-associated characteristics in celiac disease children following a gluten-free diet: results of a randomized, placebo-controlled trial. Nutrients. 2018;10(2):201. https://doi.org/10.3390/nu10020201 · FREE FULL TEXT
20. Feruś K, Drabińska N, Krupa-Kozak U, Jarocka-Cyrta E. A randomized, placebo-controlled, pilot clinical trial to evaluate the effect of supplementation with prebiotic Synergy 1 on iron homeostasis in children and adolescents with celiac disease treated with a gluten-free diet. Nutrients. 2018;10(11):1818. https://doi.org/10.3390/nu10111818 · FREE FULL TEXT
21. Wolf RL, Lebwohl B, Lee AR, et al. Hypervigilance to a gluten-free diet and decreased quality of life in teenagers and adults with celiac disease. Digestive Diseases and Sciences. 2018;63(6):1438-1448. https://doi.org/10.1007/s10620-018-4936-4
22. Bennett A, Bery A, Esposito P, Zickgraf H, Adams DW. Avoidant/restrictive food intake disorder characteristics and prevalence in adult celiac disease patients. Gastro Hep Advances. 2022;1(3):321-327. https://doi.org/10.1016/j.gastha.2022.01.002 · FREE FULL TEXT
23. Wastyk HC, Fragiadakis GK, Perelman D, et al. Gut-microbiota-targeted diets modulate human immune status. Cell. 2021;184(16):4137-4153. https://doi.org/10.1016/j.cell.2021.06.019 · FREE FULL TEXT
24. Tjellström B, Stenhammar L, Sundqvist T, et al. The effects of oats on the function of gut microflora in children with coeliac disease. Alimentary Pharmacology and Therapeutics. 2014;39(10):1156-1160. https://doi.org/10.1111/apt.12707
25. Pinto-Sánchez MI, Causada-Calo N, Bercik P, et al. Safety of adding oats to a gluten-free diet for patients with celiac disease: systematic review and meta-analysis of clinical and observational studies. Gastroenterology. 2017;153(2):395-409. https://doi.org/10.1053/j.gastro.2017.04.009
26. Shahzil M, Irfan MZ, Kazmi SK, Muhammad A, Clarke K. Prebiotic supplementation in celiac disease: a systematic review of clinical efficacy, safety, and nutritional outcomes. Journal of Gastrointestinal and Liver Diseases. 2026;35(2):237-249. https://doi.org/10.15403/jgld-6655
27. van Megen F, Skodje GI, Lergenmuller S, et al. A low FODMAP diet reduces symptoms in treated celiac patients with ongoing symptoms: a randomized controlled trial. Clinical Gastroenterology and Hepatology. 2022;20(10):2258-2266. https://doi.org/10.1016/j.cgh.2022.01.011
28. van Megen F, Veierød MB, Lundin KEA, Henriksen C. The low FODMAP diet in celiac disease: 5-year follow-up of a randomized controlled trial. Gastro Hep Advances. 2026;5(8):101009. https://doi.org/10.1016/j.gastha.2026.101009 · FREE FULL TEXT
29. Francavilla R, Piccolo M, Francavilla A, et al. Clinical and microbiological effect of a multispecies probiotic supplementation in celiac patients with persistent IBS-type symptoms: a randomized, double-blind, placebo-controlled, multicenter trial. Journal of Clinical Gastroenterology. 2019;53(3):e117-e125. https://doi.org/10.1097/MCG.0000000000001023 · FREE FULL TEXT
30. Parisi GC, Zilli M, Miani MP, et al. High-fiber diet supplementation in patients with irritable bowel syndrome (IBS): a multicenter, randomized, open trial comparison between wheat bran diet and partially hydrolyzed guar gum (PHGG). Digestive Diseases and Sciences. 2002;47(8):1697-1704. https://doi.org/10.1023/a:1016419906546
31. Parisi G, Bottona E, Carrara M, et al. Treatment effects of partially hydrolyzed guar gum on symptoms and quality of life of patients with irritable bowel syndrome: a multicenter randomized open trial. Digestive Diseases and Sciences. 2005;50(6):1107-1112. https://doi.org/10.1007/s10620-005-2713-7
32. Catassi C, Elli L, Bonaz B, et al. Diagnosis of non-celiac gluten sensitivity (NCGS): the Salerno experts' criteria. Nutrients. 2015;7(6):4966-4977. https://doi.org/10.3390/nu7064966 · FREE FULL TEXT
33. Junker Y, Zeissig S, Kim SJ, et al. Wheat amylase trypsin inhibitors drive intestinal inflammation via activation of toll-like receptor 4. Journal of Experimental Medicine. 2012;209(13):2395-2408. https://doi.org/10.1084/jem.20102660
34. Seiler CL, Rueda GH, Miranda PM, et al. Effect of gluten and wheat on symptoms and behaviours in adults with irritable bowel syndrome: a single-centre, randomised, double-blind, sham-controlled crossover trial. The Lancet Gastroenterology and Hepatology. 2025;10(9):794-805. https://doi.org/10.1016/S2468-1253(25)00090-1
35. Lammers KM, Lu R, Brownley J, et al. Gliadin induces an increase in intestinal permeability and zonulin release by binding to the chemokine receptor CXCR3. Gastroenterology. 2008;135(1):194-204. https://doi.org/10.1053/j.gastro.2008.03.023 · FREE FULL TEXT
36. Tripathi A, Lammers KM, Goldblum S, et al. Identification of human zonulin, a physiological modulator of tight junctions, as prehaptoglobin-2. Proceedings of the National Academy of Sciences. 2009;106(39):16799-16804. https://doi.org/10.1073/pnas.0906773106 · FREE FULL TEXT
37. Hollon J, Puppa EL, Greenwald B, Goldberg E, Guerrerio A, Fasano A. Effect of gliadin on permeability of intestinal biopsy explants from celiac disease patients and patients with non-celiac gluten sensitivity. Nutrients. 2015;7(3):1565-1576. https://doi.org/10.3390/nu7031565 · FREE FULL TEXT
38. Fasano A. Zonulin and its regulation of intestinal barrier function: the biological door to inflammation, autoimmunity, and cancer. Physiological Reviews. 2011;91(1):151-175. https://doi.org/10.1152/physrev.00003.2008
39. Scheffler L, Crane A, Heyne H, et al. Widely used commercial ELISA does not detect precursor of haptoglobin2, but recognizes properdin as a potential second member of the zonulin family. Frontiers in Endocrinology. 2018;9:22. https://doi.org/10.3389/fendo.2018.00022 · FREE FULL TEXT
40. Massier L, Chakaroun R, Kovacs P, Heiker JT. Blurring the picture in leaky gut research: how shortcomings of zonulin as a biomarker mislead the field of intestinal permeability. Gut. 2021;70(9):1801-1802. https://doi.org/10.1136/gutjnl-2020-323026 · FREE FULL TEXT
41. Camilleri M. Leaky gut: mechanisms, measurement and clinical implications in humans. Gut. 2019;68(8):1516-1526. https://doi.org/10.1136/gutjnl-2019-318427 · FREE FULL TEXT
42. Koh A, De Vadder F, Kovatcheva-Datchary P, Bäckhed F. From dietary fiber to host physiology: short-chain fatty acids as key bacterial metabolites. Cell. 2016;165(6):1332-1345. https://doi.org/10.1016/j.cell.2016.05.041
43. Rubio-Tapia A, Ludvigsson JF, Brantner TL, Murray JA, Everhart JE. The prevalence of celiac disease in the United States. American Journal of Gastroenterology. 2012;107(10):1538-1544. https://doi.org/10.1038/ajg.2012.219
44. Singh P, Arora A, Strand TA, et al. Global prevalence of celiac disease: systematic review and meta-analysis. Clinical Gastroenterology and Hepatology. 2018;16(6):823-836. https://doi.org/10.1016/j.cgh.2017.06.037
45. Choung RS, Unalp-Arida A, Ruhl CE, Brantner TL, Everhart JE, Murray JA. Less hidden celiac disease but increased gluten avoidance without a diagnosis in the United States: findings from the National Health and Nutrition Examination Surveys from 2009 to 2014. Mayo Clinic Proceedings. 2016;92(1):30-38. https://doi.org/10.1016/j.mayocp.2016.10.012 · FREE FULL TEXT
46. Croall ID, Trott N, Rej A, et al. A population survey of dietary attitudes towards gluten. Nutrients. 2019;11(6):1276. https://doi.org/10.3390/nu11061276 · FREE FULL TEXT

This post accompanies the Lit Review Friday episode of Learn Something with Thaena.