Skip to content
SansgluGluten-Free App · Research sources
Get Sansglu
Source guide

Gluten-free dining, by the numbers.

A reference for writers and diners. Each figure below names what was measured, when it was published, and what it cannot tell us. This is a research summary, not a Sansglu clinical study or medical advice.

One minute, no sound. The full explanation and sources follow below.

A 2026 update: 1.5% seroprevalence, 0.8% biopsy-confirmed

A meta-analysis published online August 18, 2026 included 87 articles and 394,274 participants. Its pooled estimates were 1.5% for positive celiac blood tests (95% CI: 1.3–1.8%) and 0.8% for biopsy-confirmed disease (95% CI: 0.6–1.0%). The literature search ended in April 2024.

Limit: These measure different diagnostic outcomes. Do not add them together or describe 1.5% as biopsy-confirmed disease. Study results varied substantially; the combined participant count is not the denominator for each estimate. We checked the indexed abstract, not the paywalled full text.

Source: Karimzadhagh et al., Gastroenterology (2026) · PubMed record.

Why older pages say 1.4% or 0.7%

A 2018 meta-analysis reported 1.4% seroprevalence among 275,818 participants and 0.7% biopsy-confirmed prevalence among 138,792 participants. It included studies published through March 2016. These remain historical estimates; comparing them with the 2026 review does not establish a population trend.

Source: Singh et al., Clinical Gastroenterology and Hepatology (2018).

About 2 million people in the United States

NIDDK gives an estimate of about 2 million people with celiac disease in the United States. This is the agency's published estimate, not a new 2026 census or a count of diagnosed patients. Celiac disease, wheat allergy, and non-celiac gluten sensitivity are different conditions.

Source: NIDDK: Definition & Facts for Celiac Disease.

32% in one crowd-sourced restaurant-food study

A 2019 U.S. study analyzed 5,624 food tests shared by 804 portable-sensor users over 18 months. Gluten was detected in 32% of foods labeled gluten-free. The 5,624 total includes tests beyond that gluten-free-labeled subset.

Limit: Participants chose what to test and share. This was not a random sample of restaurants. The device could detect gluten below 20 parts per million; a positive result did not establish a regulatory violation or predict symptoms. Do not report this as “32% of restaurants are unsafe,” or as a 2026 contamination rate.

Source: Lerner et al., American Journal of Gastroenterology (2019).

Less than 20 ppm: a U.S. food-labeling criterion

FDA requires less than 20 parts per million of unavoidable gluten as one criterion for a gluten-free claim on foods under its labeling rule. This does not mean zero gluten. FDA says restaurant use of the term should be consistent with the federal definition.

Limit: A menu label does not establish that a kitchen was inspected or that a particular meal was tested. Ask the restaurant about current ingredients and cross-contact practices. This is a U.S. rule; requirements differ by jurisdiction.

Source: FDA: “Gluten-Free” Means What It Says.

Use the number with its limits

Cite the original paper or agency beside the statistic. Keep its publication year, population, and measurement method. Do not turn a pooled estimate into a Sansglu customer count or a promise about an individual restaurant.

Download the source table (CSV). The table is a reading aid, not participant-level data. Credit to this summary is welcome but never required in exchange for a link.

About Sansglu

Sansglu helps people research gluten-free dining options. Restaurant information supports decisions; it does not guarantee safety. Read how we assess restaurant information.

For source corrections or product questions: hello@sansglu.com. For diagnosis or treatment, speak with a qualified healthcare professional.