Fiber-Rich Gluten-Free Foods: No Overall Glucose Benefit in Pilot

Key Takeaways

• Among women with celiac disease in a remotely conducted Norwegian comparison, glucose iAUC did not improve significantly with fiber-rich experimental foods versus benchmark gluten-free foods.
• Glucose at the 120-minute OGTT time point fell more with experimental foods, a finding distinct from the overall iAUC comparison.
• Gastrointestinal symptoms were generally unchanged, while diarrhea scores improved more among controls.
• Nominal differences in targeted gut bacteria did not remain significant after correction for multiple testing.

Fiber content can be difficult to maintain when bread and breakfast cereals are replaced on a gluten-free diet. Fiber-rich alternatives raise separate questions about glucose response and gastrointestinal symptoms: a different nutrient profile might affect blood glucose without necessarily changing how a person feels. For women living with celiac disease, those questions concern foods eaten regularly, not just the composition stated on a package. Investigators tested both responses during a substitution of everyday gluten-free staples.

Investigators conducted a randomized trial of fiber-rich gluten-free foods in women with celiac disease as a remotely managed, double-blind, parallel-group exploratory pilot in Norway. Thirty women with medically confirmed celiac disease completed a benchmark-product run-in, then spent three weeks replacing usual breads and cereals with either experimental gluten-free bread, rolls, and chia muesli containing oats, buckwheat, and quinoa or commercially available gluten-free breads, including oat bread, and cornflakes. Prescribed minimum quantities of study products supplied 24.8 versus 12.3 g/day of fiber, respectively; these were not measures of total daily fiber intake. Primary measures included fasting and post-load glucose, including glucose incremental area under the curve (iAUC) during an oral glucose tolerance test (OGTT). Secondary assessments included the Gastrointestinal Symptom Rating Scale–Irritable Bowel Syndrome (GSRS-IBS), Bristol Stool Chart, targeted stool microbiota, blood metabolic and inflammatory markers, and peptide YY (PYY).

Change in glucose iAUC did not differ significantly between groups after the intervention. Within the experimental group, iAUC decreased (p = 0.049), but that within-group change did not establish an advantage over the benchmark products. Change in glucose at the 120-minute OGTT time point did differ between groups (p = 0.04), with glucose decreasing after experimental foods. That isolated time-point difference did not alter the overall iAUC comparison.

Gastrointestinal symptom ratings were generally stable, although the reduction in GSRS-IBS diarrhea score was greater among controls than among women receiving experimental foods (between-group p = 0.04). Five of 83 statistically analyzed targeted microbial taxa showed nominal between-group abundance-change differences: Anaerostipes hadrus, Bacteroides caccae, Dorea longicatena, Escherichia coli, and Faecalibacterium prausnitzii A2-165. Abundance decreased more with control foods, but none of these differences remained significant after Benjamini–Hochberg false-discovery-rate (FDR) correction (all q > 0.05). Total cholesterol, C-peptide, analyzed inflammatory markers, and PYY showed no significant differences.

Experimental and benchmark foods also differed in energy, protein, fat, and carbohydrate. Total fiber consumed during the intervention was unknown, and soluble and insoluble fiber in the products was not measured, so responses cannot be assigned to fiber or pseudocereals alone. Investigators noted the brief, small women-only exploratory trial, lack of a power calculation, and a diet that was not fully controlled. Neutral packaging may not have prevented recognition of commercial control foods; mean-value imputation and difficulties collecting samples remotely further limited interpretation. Conducted in Norway, the trial does not establish longer-term effects or applicability to all people with celiac disease.

The investigators concluded that pseudocereal-containing gluten-free foods did not improve the between-group overall glucose response or generally increase measured gastrointestinal symptoms during the intervention. Whether the foods improve glycemic regulation or act through gut-microbiota changes remains unresolved, according to the investigators.

Clinician Questions

Do findings on fiber-rich gluten-free foods in celiac disease extend to patients with diabetes?

The Norwegian comparison enrolled women with medically confirmed celiac disease but excluded people with diabetes and other chronic metabolic or intestinal diseases. It cannot establish the glucose response to these foods in people with both celiac disease and diabetes.

How was glucose iAUC measured after gluten-free product substitution in women with celiac disease?

Investigators used a 75-g glucose-equivalent oral load and serial finger-prick glucose readings in women with celiac disease. They subtracted the fasting value from post-load measurements and calculated the incremental area under the resulting curve. Unlike an individual post-load reading, iAUC captures glucose response across the test period.

Did every woman with celiac disease provide glucose and stool samples during the gluten-free product intervention?

Among women with medically confirmed celiac disease assigned fiber-rich pseudocereal-based or benchmark gluten-free products, 29 completed oral glucose tolerance testing, 27 provided stool samples, and all completers supplied dried blood spots and gastrointestinal symptom reports.