About one person in a hundred has celiac disease and must avoid gluten permanently. A much smaller number have wheat allergy. Beyond those two groups the evidence that removing wheat improves anything is weak, and the best-designed challenge studies point at fructans rather than gluten as the cause of most self-reported symptoms. In 2013 this blog published a review of Wheat Belly telling every reader to quit wheat. That advice was wrong, and this article explains why — starting with the piece of information that matters most.
Do not go gluten-free before you are tested
This is the single most useful sentence on this page. Celiac testing only works while gluten is still in the diet. The antibodies that serology detects fall once gluten is removed, and the intestinal damage that biopsy looks for begins to heal, so a person who quits first and asks questions later usually cannot be diagnosed.
Getting an answer afterward means a gluten challenge: deliberately eating gluten daily for weeks before retesting, which is unpleasant for someone who felt better without it. The American College of Gastroenterology guideline describes the sequence — serology first, biopsy in most patients to confirm — and all of it assumes gluten is still being eaten.
The diagnosis is worth having. It determines whether the diet must be absolute or merely convenient, triggers screening of first-degree relatives, and prompts follow-up for the deficiencies and bone loss associated with untreated celiac disease. “I feel better without bread” tells nobody any of that.
Celiac disease, and how it is actually diagnosed
Celiac disease is an immune-mediated reaction to gluten in genetically susceptible people, in which the immune response damages the lining of the small intestine. It is neither an intolerance nor an allergy.
A 2018 systematic review and meta-analysis in Clinical Gastroenterology and Hepatology pooled 96 studies covering 275,818 people tested serologically and 138,792 assessed by biopsy. Pooled global prevalence was 1.4 percent by antibody testing and 0.7 percent by biopsy confirmation, with regional variation from 0.4 percent in South America to 0.8 percent in Europe.
Diagnosis normally runs in two steps. First, blood tests: tissue transglutaminase IgA measured alongside total IgA, because the two to three percent of people who are IgA deficient will otherwise return a false negative and need IgG-based testing. Second, upper endoscopy with multiple duodenal biopsies including the bulb, since damage can be patchy. Guidelines also allow a no-biopsy route for selected children with very high antibody titers confirmed on a second sample. The NIDDK overview covers the clinical picture, which ranges from diarrhea and weight loss to anemia, fatigue, dermatitis herpetiformis or nothing obvious at all.
Wheat allergy is a different condition
Wheat allergy is an IgE-mediated response to wheat proteins, not an autoimmune reaction to gluten. Symptoms appear within minutes to a couple of hours — hives, swelling, wheezing, vomiting, occasionally anaphylaxis — and it is diagnosed with skin prick or specific IgE testing and, where needed, a supervised oral food challenge.
One variant is worth knowing: wheat-dependent exercise-induced anaphylaxis, where reactions occur only when wheat is followed by exertion, usually linked to omega-5 gliadin. Wheat is a major food allergen that must be declared on US labels under FDA allergen rules, which is why “contains wheat” appears on packaging that has nothing to do with celiac disease.
Non-celiac gluten sensitivity, and the fructan problem
A third group reports real symptoms from wheat with celiac disease and wheat allergy both excluded. There is no biomarker and no test, so the label is applied by exclusion. The symptoms are not in dispute. The culprit is.
The turn in this literature came from Biesiekierski and colleagues in Gastroenterology in 2013, who put 37 people with self-reported gluten sensitivity on a reduced-FODMAP diet first, then challenged them blind with high gluten, low gluten or placebo. Symptoms improved on the FODMAP reduction and did not track gluten dose. The same group had reported a gluten effect two years earlier without the FODMAP run-in.
The clearest test came from Skodje and colleagues in Gastroenterology in 2018: 59 adults on a self-imposed gluten-free diet, celiac disease excluded, given muesli bars containing 5.7 g gluten, 2.1 g fructan, or placebo, for seven days each in double-blind crossover with washouts. Fructans produced the highest symptom scores at 38.6 on the GSRS-IBS scale, against 33.1 for gluten, and there was no difference between gluten and placebo. Bloating was worst on fructan.
Fructans are fermentable carbohydrates, and wheat is the largest single source of them in a typical Western diet. Cutting wheat cuts fructans, which is a plausible explanation for why so many people feel better. A 2017 review in Clinical Gastroenterology and Hepatology pooled the double-blind challenge trials and found gluten-specific responses confirmed in only a small minority of patients, with a substantial share reacting to placebo. None of this means the symptoms are imagined. It means the label on the bottle may be wrong.
What Wheat Belly and Grain Brain claimed
The 2013 post on this blog repeated four claims from Wheat Belly. Each can be checked.
That modern wheat was bred to contain far more gluten. A 2013 analysis in the Journal of Agricultural and Food Chemistry surveyed twentieth and twenty-first century US data and found no support for the idea that breeding increased the protein content — and therefore the gluten content — of American wheat. Semi-dwarf wheat was produced by conventional breeding for height, and no genetically modified wheat is grown commercially in the United States.
That two slices of whole wheat bread raise blood sugar more than a candy bar. This compares glycemic index values measured on 50 g of available carbohydrate eaten alone. Whole wheat bread and white bread sit close together, and table sugar sits below both, which is a curiosity of the index rather than a nutritional insight. Glycemic load in a real mixed meal behaves differently.
That gliadin produces morphine-like fragments that make wheat addictive. Gluten exorphins can be generated in laboratory digestion. There is no human evidence that they cross the blood-brain barrier and act centrally at dietary intakes, and the small naloxone studies sometimes cited have not been replicated in a way that supports the claim.
That grains drive obesity and cognitive decline. Neither the cohort data nor any trial supports this. And there is a confound in every personal experiment: quitting wheat usually means quitting pizza, pastries, beer and sandwiches, which changes calories and food quality at once. Our piece on cereal grasses covers a related confusion, since wheat grass is cut before the grain forms and contains no gluten.
What the cohort data show about grains
The evidence running the other way is large and consistent. A 2016 dose-response meta-analysis in the BMJ pooled 45 prospective studies. Per 90 g a day of whole grains — about three servings — the summary relative risks were 0.81 (95% CI 0.75 to 0.87) for coronary heart disease, 0.78 (0.73 to 0.85) for cardiovascular disease, 0.85 (0.80 to 0.91) for total cancer and 0.83 (0.77 to 0.90) for all-cause mortality, with risk continuing to fall up to about 210 to 225 g a day.
Gluten itself was examined directly. A 2017 BMJ cohort analysis followed 64,714 women and 45,303 men for 26 years, recording 6,529 cases of coronary heart disease. Comparing the highest fifth of gluten intake with the lowest, the multivariable hazard ratio was 0.95 (0.88 to 1.02), and 1.00 after adjusting for whole grain intake. The authors’ conclusion was that gluten intake was not associated with coronary heart disease, and that avoiding it may reduce whole grain intake.
These are observational studies with the usual problem: people who eat more whole grains differ in many other ways, and adjustment is imperfect. What can be said is that the large cohorts repeatedly failed to find the harm the popular books predicted, and consistently found the opposite association. Our overview of diet and long-term disease risk sets out how much weight this kind of study carries.
What unnecessary avoidance costs
Going gluten-free without a reason is not free. Enriched wheat flour in the United States carries added iron, thiamin, riboflavin, niacin and folic acid; gluten-free flours generally do not, and gluten-free substitute products are commonly lower in fiber and protein as well.
Rice flour, the usual substitute, brings its own issue. A 2017 analysis in Epidemiology examined 7,471 NHANES participants, 73 of whom reported a gluten-free diet, and found higher blood and urinary arsenic, cadmium and mercury in that group. The sample was small and cross-sectional — a signal rather than a verdict.
Gluten-free versions also cost consistently more than their conventional equivalents, and the diagnostic cost is the one already covered: quitting first forecloses the test. For anyone who does need the diet, the rules that make labels readable are covered in our guide to gluten-free labeling.
Common questions
Can I still be tested if I already stopped eating gluten?
Not accurately. Antibody levels fall and the intestinal lining heals on a gluten-free diet, so both serology and biopsy can read normal. The route back is a supervised gluten challenge — eating gluten daily for a period of weeks before retesting — which is worth discussing with a gastroenterologist rather than improvising.
If I feel better without wheat, does that mean I am gluten sensitive?
It means something changed. In the 2018 crossover trial, fructans produced worse symptoms than gluten, and gluten was indistinguishable from placebo. Cutting wheat also cuts fructans, a great deal of ultra-processed food and often a lot of calories. Celiac testing first, then a structured low-FODMAP trial with a dietitian, separates those possibilities better than guessing.
Are spelt, einkorn or sourdough safe for celiac disease?
No. Spelt, einkorn, emmer and khorasan are all wheat and all contain gluten. Long sourdough fermentation lowers fructan content and partly degrades gluten, but not reliably to a level safe for celiac disease. Anyone with a diagnosis needs food that meets the labeling standard, not artisanal bread that may be gentler.
Is a gluten-free diet healthier for people without celiac disease?
There is no evidence that it is, and some evidence pointing the other way through reduced whole grain and fiber intake. A diet built on vegetables, legumes, nuts and intact grains is well supported. A diet built on gluten-free packaged replacements is not the same thing.
Where this leaves wheat
For the one percent with celiac disease, gluten avoidance is lifelong, strict and non-negotiable — and the diagnosis has to come first. For people with wheat allergy, the same holds for wheat. For everyone else the evidence supports whole grains and offers no reason to remove them, however confidently the books argued otherwise. We regret publishing that advice.
The reliable gap in most American diets is vegetables rather than grains. Greens Plus has been making green superfood blends since 1993.
About the Author
