Gluten Sensitivity Didn't Survive Its Own Replication. The Monash Team Behind the Famous 2011 Finding Re-Tested 37 Patients Under Blinded Conditions and Found Gluten-Specific Effects in Just 8%

In 2011, a small Australian trial handed the gluten-free industry its founding document. Thirty-four patients with irritable bowel syndrome, none with celiac disease, ate gluten or placebo baked into bread and muffins for up to six weeks, and the gluten group fared markedly worse: 68 percent said their symptoms were no longer under control, against 40 percent on placebo (p = 0.047). Headlines declared proof of non-celiac gluten sensitivity, the market listened, and within a few years one in five Americans told Gallup they were actively seeking gluten-free foods, roughly twenty times the share who actually have celiac disease.

Then the same researcher tried to reproduce her own result, and the story split in two.

Jessica Biesiekierski, the Monash University dietitian behind the 2011 trial, spent two years building a stricter test. Thirty-seven adults with self-reported gluten sensitivity plus IBS, celiac disease carefully ruled out, first spent two weeks on a diet low in FODMAPs, the fermentable carbohydrates her own lab had implicated in gut distress, and every single one improved. Then came the blinded phase: one-week stretches of high-gluten food (16 grams a day, about four slices of bread), low-gluten food (2 grams plus whey protein), or whey protein alone as the control, with neither patients nor researchers knowing who ate what. Twenty-two participants later crossed over for a three-day rechallenge.

The results, published in Gastroenterology in 2013, were a clean null with an asterisk: symptoms improved for every participant during the low-FODMAP run-in, then worsened by roughly the same amount whether the hidden ingredient was gluten or whey. Only 8 percent, about three people, showed effects specific to gluten. Blood and stool markers of inflammation and immune activation did not move on any diet. In the three-day rechallenge, symptoms climbed equally in all groups and the gluten-specific signal failed to reappear. The conclusion was blunt: no evidence of specific or dose-dependent gluten effects in self-reported sensitivity once FODMAPs were controlled.

The design detail is the whole story. The 2011 trial had compared gluten against an inert placebo. The 2013 trial compared gluten against whey protein, and the whey group felt just as bad. Whatever was making people miserable was not specific to gluten. It looked like the gut reacting to being challenged at all, or the mind reacting to the idea of a challenge.

The mind, it turns out, does heavy lifting here: a 2024 double-blind trial found that symptom scores followed what participants expected to eat rather than what they actually ate, with expectancy effects dwarfing ingredient effects (p < 0.001), a result a 2025 Lancet review called landmark evidence for nocebo mechanisms in this field. An earlier Italian challenge found patients felt worst on placebo, with only 4 of 20 correctly identifying the gluten periods. When belief moves the needle more than biochemistry, the condition starts to look less like an intolerance and more like a story the gut tells itself.

Meanwhile an independent lab found the likelier culprit. In 2018, Gry Skodje's team at Oslo University Hospital ran a double-blind crossover in 59 self-reported sensitive patients, pitting gluten (5.7 grams) directly against fructans (2.1 grams, the FODMAP subtype concentrated in wheat) and placebo. Fructans produced the worst overall symptom scores, 38.6 against 33.1 for gluten and 34.3 for placebo, with bloating significantly worse on fructans (p = .004), and the headcount told the same story: twenty-four participants felt worst on fructans, twenty-two on placebo, thirteen on gluten, with gluten statistically indistinguishable from placebo, a result that earned the Rome Foundation's 2019 best-paper prize and made fructans the field's leading suspect. The paper won the Rome Foundation's 2019 best-paper prize. As Monash's Jane Muir put it, cutting wheat removes a large share of fructans, so of course people feel better, yet they can still run into trouble with onions, garlic, and even chickpea crisps. Gluten was standing near the scene; fructans committed the act.

Here is the article's own calculation, with the inputs shown. Gallup finds 21 percent of American adults actively try to include gluten-free foods. Against roughly 260 million adults, that is about 54 million people. Celiac disease affects about 1 percent, or roughly 2.6 million adults. The ratio is twenty to one: for every American who must avoid gluten to protect their small intestine, about twenty avoid it for a sensitivity the best blinded trials cannot locate. The 2013 trial prices the mismatch from the other side, since among people certain gluten is their problem, only 8 percent reacted to gluten specifically, leaving more than nine in ten reacting to something else, whether FODMAPs, expectation, or an IBS gut that never fully settles.

Now the strongest case against this reading, at full strength, because the finding has genuine limits and serious critics: thirty-seven people at one center is a small trial, and every participant had IBS, so nothing here addresses gluten sensitivity without IBS. The challenges ran a week, with a three-day rechallenge; slow, cumulative effects would slip through unseen. The trial measured gut symptoms, not the foggy mind, joint pain, rashes, and headaches that clinicians like Alessio Fasano count as part of the syndrome, a gap Italian researchers flagged directly. Wheat carries other suspects the trial never isolated, notably amylase-trypsin inhibitors, which switch on innate immune cells in laboratory models, and wheat germ agglutinin. The 8 percent matters too: a small subgroup did show gluten-specific effects, and a 2014 trial from the same Monash group found gluten lifted depression scores versus placebo even as gut symptoms held flat. The field's response was discipline, not deletion: the 2015 Salerno criteria kept non-celiac gluten sensitivity as a diagnosis of exclusion, confirmable only by double-blind placebo-controlled challenge, and Biesiekierski herself told NPR the sensitivity probably exists, just rarely, with much still to learn about gluten.

That is the honest shape of the evidence: not that gluten sensitivity is invented, but that the thing millions of people treat with expensive bread is, for the large majority, something else entirely. The 2011 trial was real science, and the 2013 trial was better science by the same hands, which is how the system is supposed to work and almost never how it gets celebrated.