One standardseparate promising findings from demonstrated outcomes
START HERE
Five questions, connected—not a five-part cure
Our objective is durable gluten-specific immune control together with preserved or restored intestinal function. The organization below is our research framework, not a proven sequence of treatments. Different questions may require different models, datasets and scientific teams.
These are original summaries of selected published findings. Publication dates identify the evidence being discussed, not the present regulatory status of a product. An experimental effect is not a treatment recommendation. Authors and institutions cited below are sources, not represented as partners or endorsers.
lnc13: a possible regulator of the response threshold
Yang-Fischer and colleagues · Nature Immunology · 2026
Published finding. In HLA-DQ8 mice, loss of the regulatory RNA lnc13 increased gluten-associated inflammatory and cytotoxic responses. The work links lnc13 to restraint of immune responsiveness, including IL-15-associated activity.
Important limit. A genetic loss-of-function experiment does not demonstrate that restoring this regulator safely reverses established human disease. The paper discusses the differences between its experimental model and complex human celiac disease.
Why it belongs here. It supplies a mechanistic basis for asking about the transition from susceptibility to injury. Our question is whether a reversible regulatory state explains that transition and can be tested without broadly suppressing immunity. Source and access scope.
Wang and colleagues · Nature · Online 2025; issue 2026
Published finding. Human intestinal M cells in organoid–T-cell experiments processed and presented gluten antigen through MHC-II. The authors also report important differences between human and mouse M-cell biology.
Important limit. Demonstrating this ability does not show that the route is indispensable in an intact intestine. Other antigen-presenting populations may contribute or compensate.
Why it belongs here. The next question is the additional benefit of interrupting this route when alternatives remain present—not merely repeating the demonstration that M cells can activate T cells. Source and access scope.
TEV-408: reported protection from intestinal injury
Teva · Phase 2a announcement · September 2, 2026
Reported finding. Teva announced reduced gluten-induced intestinal damage versus placebo in its study of an investigational anti-IL-15 antibody.
Important limit. The evidence summarized here is a company topline announcement, not our independent patient-level analysis. It does not establish lasting treatment-free tolerance, and this library does not equate an early safety statement with comprehensive safety proof.
Why it belongs here. Tissue protection matters even when it is not a cure. Our question asks what remains after an intervention has genuinely ceased acting: persistent immune change, continued protection by another mechanism, or return of the harmful response. Original sponsor report.
TAK-101: an immune signal that needs careful interpretation
Kelly and colleagues · Gastroenterology · 2021
Published finding. The Phase 2a study reported an 88% reduction in the change in a gluten-specific interferon-γ response versus placebo. Thirty-three randomized participants completed the gluten challenge.
Important limit. The reported between-group change in villus-height-to-crypt-depth ratio did not reach conventional statistical significance (p=0.08), and intraepithelial lymphocyte numbers were not different. The authors interpreted the immune findings as evidence suggesting antigen-specific tolerance; that is not the same as demonstrated lifelong, treatment-free protection.
Why it belongs here. A positive immune endpoint should be retained without upgrading every tissue endpoint to a proven benefit. The later program's outcomes must be evaluated separately rather than inferred from this early report. Published trial.
Nexvax2 / RESET CeD: keep the negative evidence visible
Tye-Din and colleagues · The Lancet Gastroenterology & Hepatology · 2023
Published finding. The randomized Phase 2 study was discontinued after its planned interim analysis. Nexvax2 did not reduce the acute gluten-induced gastrointestinal symptoms assessed by the study.
Important limit. This negative result applies to the product, population, regimen and endpoints studied. It does not prove that every antigen-specific approach is ineffective. Nor is it a head-to-head comparison with TAK-101, which used a different primary endpoint.
Why it belongs here. A credible tolerance program must explain which response it aims to change and show a meaningful downstream outcome. The failure is a design lesson, not an inconvenient result to omit. Terminated-trial publication.
Gluten-reactive cell lineages can persist for decades
Risnes and colleagues · Journal of Clinical Investigation · 2018
Published finding. Researchers followed gluten-specific CD4 T-cell repertoires and found persistent clonotypes, with recall responses largely drawing on established lineages. A historical-sample subset supported persistence over decades.
Important limit. A persistent lineage is not proof that an individual cell lived for that entire period. This study does not require a future therapy to eliminate every gluten-recognizing cell.
Why it belongs here. Our interpretation is that a brief reduction in tissue injury cannot, by itself, establish control over the body's longer-lived immune memory. Functional reprogramming, selective depletion and tissue protection are different possible mechanisms that must not be conflated. Primary study.
Published finding. Biopsy-derived air–liquid interface organoids retained intestinal epithelium, supporting tissue and several immune populations. The study reproduced gluten-dependent injury and identified IL-7 activity in this experimental setting.
Important limit. These are laboratory findings, not demonstrated IL-7 treatment efficacy in patients. Long-lived epithelial growth is also not proof that the relevant immune response remains functional for an equally long period.
Why it belongs here. This is a concrete model to evaluate for injury and protection questions. Our proposed persistence comparison first needs a time-matched untreated model that still responds meaningfully; otherwise, a fading culture could resemble successful treatment. Publication · Model and dataset guide.
Wulczynski and colleagues · Nature Communications · 2026
Published finding. Human celiac samples showed altered small-intestinal microbial fiber metabolism. In sensitized mice, a dietary intervention during a gluten-free recovery period supported microbial metabolism and faster mucosal healing.
Important limit. Human observations and mouse intervention results are different evidence types. This does not establish a microbiome cure or treatment-free tolerance in people. A post-exposure recovery intervention is already part of the published design; we do not claim to have originated that idea.
Why it belongs here. Our refined question is whether an added benefit reflects tissue repair, reduced continuing immune injury during recovery, or both—and whether it persists. This is not a recommendation to take fiber supplements. Primary study.
LEARNING FROM WHAT DID NOT WORK
Negative findings and missing results are different
A published negative result
RESET CeD provides an answer to its studied symptom question: the intervention did not improve that endpoint. Retaining that result helps prevent a new proposal from ignoring a test that has already been performed. It does not justify dismissing an entire biological strategy. Read the primary evidence.
An unresolved result
An unlocated result is not evidence of benefit or failure. For a completed, delayed or discontinued program, our review method is to reconcile the registry record, publication, sponsor statement and dates; record what is actually missing; and keep the outcome unresolved until evidence supports a conclusion. This is a method, not an allegation of concealment.
Historical success is not current-program status. The early TAK-101 paper cannot answer what a later trial found. Its separate study record NCT04530123 is provided for checking the later trial. This page does not certify its current recruitment or results-posting status. Different study identifiers, populations and endpoints must stay distinct.
Our comparison rule. A statistically significant change within one group and a nonsignificant change within another do not themselves prove that the groups differ. Read the actual between-group comparison, endpoint and follow-up period. This is why the positive immune finding and uncertain tissue comparison in the 2021 TAK-101 report are both retained.
RESOURCES FOR RESEARCHERS
Choose a model that can answer the question
The resources below belong to the cited investigators and repositories. Links do not establish our access to biological material, permission to redistribute it, a laboratory partnership or a completed replication. A model's scope is as important as its sophistication.
Model suitability: published capabilities and our interpretation of the limits
Resource
Useful starting question
Limit that must remain visible
Human M-cell organoids
Antigen processing and presentation to T cells.
Presentation capability does not establish necessity in a system with other presenting cells. Wang et al.
Immune–epithelial ALI organoids
Gluten-peptide-triggered injury and downstream immune interactions.
Wang's discussion notes that the Santos model lacks M cells and uses modified gliadin peptide; it is not automatically an M-cell necessity test. Discussion · Model paper
Longitudinal human immune samples
Persistence and recall of antigen-reactive lineages.
Repertoire persistence is not an intervention trial or proof of tissue protection. Risnes et al.
Dataset spotlight: GSE200075
The Santos paper links this public single-cell dataset. Our inspection of its deposited sample descriptions identified 30 library records associated with six donor labels, including gene-expression and immune-receptor measurements from shared material—not 30 independent patients. Four donors have potential exposure/control contrasts, with control wording and cell-selection differences that require care. This is our metadata assessment, not a gene-expression reanalysis or criticism of every experiment in the paper. Open the deposited dataset.
The deposited organoid comparison is two days after stimulation, not a post-withdrawal recovery series. Cells captured by different selection methods cannot simply be treated as equivalent samples of the whole intestine. We have not calculated a pathway ranking, drug response or cure prediction from its expression matrices.
Additional deposited resources
Wang and colleagues identify GSE275771 and GSE275772. These are data pointers, not claims that we have analyzed every file. Before reuse, check the exact sample mapping, quality, associated publication, applicable terms and fit to the proposed question.
PROMISE VERSUS PROOF
What would make a stronger result?
This is our framework for interpreting the five questions—not a validated scoring system, clinical definition or prescribed experimental protocol. No single item below establishes a cure on its own.
Relevant immune function
Measure the gluten-specific response, not just a general reduction in inflammation. A quiet model with missing or failing immune cells can be misleading.
Independent tissue outcomes
Connect the proposed immune change to intestinal structure or function. A better blood measurement is not automatically restored tissue.
Persistence after withdrawal
Establish that the intervention has genuinely ceased acting and comparison cultures remain responsive. Residual drug activity is not an off-treatment result.
Selective rather than broad effects
Distinguish functional reprogramming, selective loss of pathogenic lineages and nonspecific immune suppression. Do not impose elimination of every gluten-reactive cell as an invented universal requirement.
Independent replication
Count donors and biological units correctly. Many cells or repeat measurements from the same source do not create a large independent patient study.
A result that can be challenged
State what would weaken the hypothesis, preserve contrary evidence and use a simpler explanation when it fits just as well. Report uncertainty instead of hiding it behind a single score.
These interpretation questions draw on the limitations of the studies above, especially the distinct endpoints in TAK-101, the organoid model's scope and the human immune-memory evidence. They remain open to scientific criticism.
A short guide to the terminology
Organoid
A laboratory-grown tissue model. What it represents depends on which cell types and functions it retains.
Antigen presentation
Displaying a protein fragment to immune cells. Showing that one cell type can do this does not show that it is the only route.
Immune tolerance
In this research agenda, selective control of the harmful gluten response—not simply reducing every immune response.
Clonotype
A lineage distinguished by its immune-receptor sequence. It is not interchangeable with one individual cell.
Washout / withdrawal
Removing an intervention and establishing that meaningful residual activity does not explain an apparent lasting effect.
Falsifier
An observation that would make us reject or narrow a proposed explanation. Each of the five questions includes one.
MAKE THE NEXT RESPONSE USEFUL
Criticism is a contribution
A short, specific response is more useful than a general endorsement. Identify the question number, the exact claim you disagree with, a relevant paper or dataset, and the comparison that would resolve the uncertainty. A reference showing that a question is already answered is valuable too.
Researchers who received an invitation can reply to the original email; a full review or meeting is not required. A model or dataset suggestion does not create an obligation to provide materials or perform work. Please do not send patient-identifiable records, confidential proposals or unpublished third-party material.
We do not list people who receive an email as collaborators. A reply is not permission to publish private correspondence or claim institutional endorsement. Formal studies, material transfers, funding and data-use agreements would require separate decisions by the actual responsible parties.
Public: the five-question discussion brief and this supporting evidence library. Analysis preparation: selected source-number checks and a deposited sample-design review. Not established by these outputs: an independently replicated biological result, a validated therapy, a laboratory agreement to test the proposals or demonstrated patient benefit.
The source-number checks were descriptive arithmetic, not an independent clinical trial. Metadata inspection establishes what a dataset appears to contain; it does not substitute for analyzing its expression data. Research questions and interpretations here are AI-GENERATED / UNTESTED / NOT NOVELTY-CLEARED.
This public education and scientific-discussion effort is separate from any grant application. A funding deadline does not define when the research mission ends.
Version record
— Library version 1.0: eight study/report guides, negative-evidence discussion, model/data directory and reading criteria added around the existing five-question brief. This date records editorial work, not new laboratory findings.
FOLLOW THE EVIDENCE
Sources and review scope
This is a curated starting library, not an exhaustive or systematic review, medical guideline, full trial-pipeline tracker or complete raw-data audit. Source coverage includes accessible primary abstracts, article text, selected legends and the specifically described source-data/metadata checks. Source links were reviewed September 23, 2026; not every linked full text or supplemental file was obtained.
Published results belong to the cited investigators. The connecting research questions, model-suitability judgments and reading framework are our AI-assisted interpretation. We do not reproduce third-party figures, full papers, original datasets, private correspondence or protected implementation details. Access to a public record does not certify every possible reuse right.
Accessible article, methods and published Figure 3. Its linked source table was not obtained in our review; no new statistical reanalysis is claimed.
Dataset and registry links appear beside their relevant discussions. Publication and data availability are not evidence of affiliation, funding, treatment access or an invitation for self-experimentation.
Research updates
Historical research: DONQ52 studied as a selective blocker of gluten-related immunity
Research update — AI-assisted summary of research by the cited authors.
In a study published in December 2023, Okura and colleagues developed DONQ52, an antibody designed to recognize several gluten-peptide complexes presented by HLA-DQ2.5. The authors reported that it blocked gluten-directed immunity in genetically modified mice carrying HLA-DQ2.5, while systemic immunity was unaffected in the experiments described. They presented these preclinical findings as a reason for clinical testing. This paper does not establish that DONQ52 improves symptoms or intestinal damage in people with celiac disease.
Limits of the evidence
This is historical preclinical evidence, not a new clinical breakthrough. The mouse findings do not establish human benefit or safety. Chugai funded the study, and the authors disclosed company employment and patent interests. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Historical research: how gluten-reactive T cells recognize a gliadin fragment
Research update — AI-assisted summary of research by the cited authors.
A 2015 study examined gluten-reactive T cells from intestinal tissue of people with HLA-DQ8-associated celiac disease. Researchers identified two T-cell populations that differed in receptor gene usage but recognized the same gliadin-derived peptide. Structural comparisons showed a conserved contact between an arginine in the T-cell receptors and a serine in the presented peptide. Replacing that serine with alanine prevented proliferation of the studied T-cell clones in response to the peptide-HLA complex.
Limits of the evidence
These are laboratory findings about antigen recognition in a particular HLA-DQ8-associated setting. They do not establish a treatment benefit, demonstrate prevention of celiac disease, or show that the same result applies to every form of celiac disease. This is a summary of a historical study, not a new clinical result. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Historical research: enzyme-targeting B cells in celiac-related skin disease
Research update — AI-assisted summary of research by the cited authors.
A 2025 laboratory study examined dermatitis herpetiformis, a gluten-sensitive skin disorder associated with celiac disease. Researchers reported that B cells recognizing transglutaminase 3 could take up complexes containing the enzyme and gluten, then activate gluten-reactive T cells. Certain antibodies also increased the enzyme's activity; when used as B-cell receptors, they increased gluten presentation to T cells. The findings describe an immune mechanism, rather than a treatment result.
Limits of the evidence
This summary covers the published abstract and bibliographic record. Laboratory and structural findings do not establish treatment benefit in people. The study concerns transglutaminase 3 and dermatitis herpetiformis; it does not answer whether transglutaminase 2 is necessary in intestinal celiac disease. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Historical research: reading food labels with celiac disease
Research update — AI-assisted summary of research by the cited authors.
In a 2020 cohort study, Gutowski and colleagues followed adults recently diagnosed with celiac disease and tested whether they could identify gluten-free foods from product labels. The authors reported label-reading errors across follow-up and found the fewest errors for products with explicit gluten-free claims. Quiz scores were not correlated with the study's blood antibody measure or questionnaire-based diet-adherence scores.
Limits of the evidence
This observational study tested label-reading performance. It did not establish that labeling errors caused persistent intestinal damage. The authors called for further research on that link. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Historical research: blood microRNAs before celiac diagnosis
Research update — AI-assisted summary of research by the cited authors.
In a 2021 study, Tan and colleagues examined small RNA molecules in blood samples from children followed in the PreventCD study and compared findings with additional cohorts. They found circulating microRNA patterns associated with celiac disease, including differences detectable before the first positive anti-transglutaminase antibody result. Some candidate markers also differed in intestinal biopsy samples. The authors proposed these patterns as potential early biomarkers.
Limits of the evidence
These are candidate associations, not a validated screening or diagnostic test. The study does not establish that changing a microRNA prevents celiac disease, and it does not support changes to anyone's diet or care. This historical report does not establish current clinical availability. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Historical research: an enzyme inhibitor tested for celiac intestinal injury
Research update — AI-assisted summary of research by the cited authors.
In a 2021 randomized trial, Schuppan and colleagues compared the investigational enzyme inhibitor ZED1227 with placebo in adults with well-controlled celiac disease during a supervised gluten challenge. They reported less damage to the small-intestinal lining with the drug, measured in biopsy samples. The study also recorded adverse events, including gastrointestinal symptoms and rash. Funding came from Dr. Falk Pharma.
Limits of the evidence
This was a preliminary, short-term trial, not proof of a cure or lasting gluten tolerance. Its biopsy outcome does not establish long-term benefit or safety. This historical summary is based on the published abstract and does not establish current clinical availability. It provides no instructions for medication use or gluten exposure. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Research archive: tissue-resident T cells and recurring intestinal inflammation (2021)
Research update — AI-assisted summary of research by the cited authors.
A 2021 review by Barreto de Albuquerque and colleagues examines memory T cells that remain in tissues, with a focus on the intestine. These cells help protect against infections, but the authors also discuss their involvement in recurring inflammatory disorders, including celiac disease and inflammatory bowel disease. The review considers tissue-resident T cells as a possible research target and discusses potential ways to interfere with inflammation involving these cells.
Limits of the evidence
This is a historical review of existing research, not a new experiment or clinical trial. Its abstract does not establish that targeting these cells safely prevents or treats celiac disease. Findings across different inflammatory disorders cannot automatically be applied to celiac disease. This summary covers the published abstract and bibliographic metadata, not an independent assessment of every cited study or current treatment advice. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Celiac research: distinct metabolic patterns found in seronegative relatives
Research update — AI-assisted summary of research by the cited authors.
A study by Upadhyay and colleagues compared metabolic profiles in serologically negative first-degree relatives of people with celiac disease, people with celiac disease, and control groups. Researchers examined intestinal biopsies, blood plasma and urine using proton NMR spectroscopy. The relatives had metabolic profiles that differed from both the celiac and control groups. The authors interpret differences in several intestinal metabolites as possible signs of an adaptive response that helps maintain intestinal integrity despite genetic predisposition.
Limits of the evidence
This study compared biological profiles; it did not test a preventive treatment. The proposed protective mechanisms are interpretations of the observed associations, not proof that changing a metabolite prevents celiac disease. The abstract does not establish whether these patterns predict future disease or apply to all relatives. This summary covers the published abstract and bibliographic metadata, not a full-report assessment or treatment advice. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Celiac research: yeast strain activates a microbial signaling pathway in preclinical experiments
Research update — AI-assisted summary of research by the cited authors.
Kan and colleagues studied how a strain of Saccharomyces boulardii interacts with intestinal microbes in celiac disease research. In gluten-sensitized mice expressing HLA-DQ8, the yeast activated the aryl hydrocarbon receptor pathway and improved gluten-related immune pathology. The researchers also found that it enhanced the capacity of celiac patients' microbiota to activate this pathway when certain indole-producing microbes were present.
Limits of the evidence
These are preclinical findings. They do not establish that taking this yeast treats or prevents celiac disease in people. The authors call for clinical testing. This summary covers the published abstract and bibliographic metadata; it is not a full-report assessment or treatment recommendation. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Research archive: anti-IL-15 trial did not meet its main intestinal-injury endpoint
Research update — AI-assisted summary of research by the cited authors.
Lähdeaho and colleagues tested AMG 714, an antibody targeting interleukin 15, in adults with celiac disease during a controlled research gluten challenge. For both study comparisons, the primary measure of intestinal structure did not differ significantly from placebo. Exploratory findings involving immune cells and symptoms did not change that primary result.
Limits of the evidence
This experimental study does not demonstrate that people can safely consume gluten or replace a gluten-free diet. Secondary findings were exploratory. This summary covers the published abstract and bibliographic metadata, not a full-report assessment or personal treatment advice. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
Mouse study links IgA deficiency, gut microbes and gluten sensitivity
Research update — AI-assisted summary of research by the cited authors.
In IgA-deficient mice, researchers observed gluten-associated changes in intestinal structure and immune responses. The paper also reports experiments transferring gut microbes into germ-free mice, including colonization with a Streptococcus species. Those experiments reproduced features of gluten-sensitive intestinal disease, supporting a microbiota-dependent pathway in the studied mouse models.
Limits of the evidence
These are preclinical findings in selected mouse models, not evidence that the same pathway explains human celiac disease or that a treatment is effective. This summary covers the published abstract and bibliographic metadata, not a full-report review. It does not provide dietary, supplement, testing or treatment advice. Based on published abstracts; details beyond those abstracts have not been checked.
This is research information, not personal medical advice.
PreventCD design paper outlines two approaches to studying celiac risk
Research update — AI-assisted summary of research by the cited authors.
A historical study-design paper described the PreventCD project, which was intended to study how infant nutrition and genetic, immune and environmental factors relate to celiac disease. It outlined a prospective, double-blind randomized dietary-intervention study involving infants from high-risk families, alongside population-based screening in Swedish children from birth cohorts with different infant-feeding practices. The abstract describes the research plans rather than reporting prevention results.
Limits of the evidence
This is a study-design article, not a completed outcomes report. Planned comparisons and the investigators' proposed hypothesis do not establish that a feeding practice prevents celiac disease. This summary covers the published abstract and bibliographic metadata only, without reviewing the full methods or later trial results. It provides no infant-feeding instructions or clinical recommendations. Based on published abstracts; details beyond those abstracts have not been checked.
Bibliographic correction (November 3, 2026): the volume, pages and issue month were corrected from the linked PubMed record. The research summary is unchanged.
This is research information, not personal medical advice.
What a Celiac hookworm trial found about gluten tolerance
Research update — AI-assisted summary of research by the cited authors.
A randomized, placebo-controlled Celiac study tested whether experimental hookworm infection changed how participants responded to gluten. The main outcome was whether participants successfully completed the 42-week challenge under study criteria that included an antibody test and an intestinal biopsy. The trial found no statistically significant difference between groups on that outcome. Measures of intestinal lining damage worsened across the groups. Participants in the active groups had fewer recorded gluten-related adverse events. Before the gluten challenge began, two people in an active group withdrew with temporary gut inflammation the researchers suspected was related to hookworm infection.
Limits of the evidence
The trial did not establish that hookworm infection protected the intestinal lining or that the groups had equivalent outcomes. The researchers concluded that hookworm infection did not restore tolerance to sustained gluten consumption. In a separate exploratory analysis, the researchers regrouped participants by whether infection was detected. An interim visit during intermittent gluten exposure suggested better celiac-related quality-of-life scores among participants with detected infection. By the end of the trial, the change in scores was similar in the two regrouped groups. The regrouped comparison did not retain the original randomized groups.
This is research information, not personal medical advice.
An early Celiac hookworm trial found intestinal injury in both groups
Research update — AI-assisted summary of research by the cited authors.
A small blinded study compared experimental hookworm infection with placebo in adults with Celiac disease. After a short wheat challenge, biopsies showed more damage to the lining of the small intestine in both groups. The researchers did not find a statistically significant difference between groups on the biopsy grading scale.
Technical details. The randomized, double-blind, placebo-controlled trial specified duodenal Marsh histology scores and a blood-cell immune assay as primary endpoints. Figure 6A shows a change over time, with no significant group effect or interaction on the histology measure. The separate villus-height-to-crypt-depth measure also worsened in both groups.
Limits of the evidence
A short challenge in a small trial cannot establish lasting tolerance or rule out every possible effect. This article summarizes published research for educational purposes.
This is research information, not personal medical advice.
Feeling better and intestinal healing differed in one Celiac cohort
Research update — AI-assisted summary of research by the cited authors.
Researchers asked whether clinical improvement matched healing of the small-intestine lining in adults with Celiac disease. In one Mayo Clinic cohort, people often improved clinically, but that improvement did not reliably indicate that the lining had healed on a follow-up biopsy.
Technical details. The investigators reviewed records retrospectively. Of 381 adults considered, 241 had both the diagnostic and follow-up biopsies available for review. In that group, time-to-event estimates of confirmed healing were 34% by two years after diagnosis (95% confidence interval, 27% to 40%) and 66% by five years (58% to 74%). The timing of follow-up biopsies varied.
Limits of the evidence
The analysis selected people with paired biopsies from one center. The estimates describe that cohort and may not apply to everyone with Celiac disease. The study does not establish when any individual should be tested. This article summarizes published research for educational purposes.
This is research information, not personal medical advice.
Childhood enterovirus and celiac disease: what the cohort studies found
Research update — AI-assisted summary of research by the cited authors.
Several studies followed children from birth and looked for signs of common infections before celiac-related antibodies appeared. Some found a link with enterovirus. A smaller study using a shorter sampling window did not detect the virus in stool or nasal samples from either group. The studies used different tests and measured different outcomes. They cannot tell us whether an infection caused celiac disease or whether preventing it would change a child’s risk.
Norwegian birth cohort. Kahrs and colleagues used serial stool samples in a nested case-control study. Enterovirus appeared more often in samples taken before celiac-related antibodies developed in children who later had celiac disease than in matched controls. The adjusted odds ratio was 1.49 (95% confidence interval 1.07–2.06). This is a comparison of detection in the sampled groups, not an estimate of an individual child’s disease risk.
TEDDY. Lindfors and colleagues examined stool virome data in 83 matched pairs of children selected for genetic susceptibility. They reported an association involving enterovirus and gluten intake with celiac autoimmunity, defined by persistent celiac-related antibodies. Of the 83 children with that antibody outcome, 28 developed celiac disease during follow-up. The antibody endpoint and a disease diagnosis are different.
Finnish DIPP cohort. Oikarinen and colleagues compared 41 children with biopsy-confirmed celiac disease and 53 matched controls. They inferred infection from changes in antibodies to enterovirus in serial blood samples. Before celiac-related antibodies appeared, the reported odds ratio for enterovirus infection was 6.3 (95% confidence interval 1.8–22.3). The wide interval shows substantial uncertainty. The studies' odds ratios describe different measurements and should not be pooled from these figures.
DIABIMMUNE. Simre and colleagues compared 29 children who developed celiac disease with 29 matched controls. Their early-infancy stool and nasal samples yielded no enterovirus detections in either group. Serum enterovirus antibody measures did not differ significantly. The study was small, and early-infancy sampling covered a narrower window than the Norwegian study, which collected stool samples through age three.
Limits of the evidence
All four analyses are observational. The timing is informative, but the tests capture different aspects of infection, and the cohorts and endpoints differ. These findings do not establish a prevention strategy or support changing a child’s diet or care.
This is research information, not personal medical advice.
What one hospital’s celiac biopsy records can tell us about later diagnosis
Research update — AI-assisted summary of research by the cited authors.
Researchers reviewed records from a children’s hospital to see whether some children with biopsy-confirmed celiac disease had an earlier biopsy that did not show the disease. They found that pattern in 11 children. The 11 cases are part of the group with a confirmed diagnosis. Because the hospital did not systematically follow every child whose first biopsy was negative, these records cannot give a reliable progression rate for that group.
In a retrospective review at St. Louis Children’s Hospital, 316 of 327 children with biopsy-confirmed celiac disease were diagnosed on their first biopsy. The other 11 had an initially negative biopsy and a later positive biopsy. The denominator is the number with a confirmed diagnosis in the study, not all children with an initially negative biopsy.
Limits of the evidence
The hospital did not systematically follow all children with an initially negative biopsy. Testing methods and available laboratory results varied. The study describes selected records and does not provide a personal risk estimate, diagnostic threshold, or follow-up schedule.