Fiber and future diabetes: the weight of an association
A major European analysis shows why body-size adjustment belongs in the headline.

Illustrative food image, not a photograph of the study intervention.
Follow people before the diagnosis
The EPIC-InterAct analysis asked whether reported fiber intake was associated with later type 2 diabetes. It did not assign participants to high- and low-fiber diets. Investigators identified 11,559 diabetes cases during a median follow-up of 10.8 years and used a case-cohort design across eight European countries. They also combined the findings with other prospective studies. This is a different question from whether a dietary intervention changes HbA1c in people already living with diabetes. Keeping those questions separate prevents a prevention association from being advertised as a treatment result.
InterAct Consortium. Diabetologia, 2015. PMID 26021487 · DOI
The adjustment that changes the headline
In EPIC-InterAct, people in the highest category of total fiber intake had a lower estimated diabetes risk than those in the lowest category after adjustment for lifestyle and dietary factors. The hazard ratio was 0.82, with a confidence interval of 0.69 to 0.97. After additional adjustment for body mass index, the associations weakened and were no longer statistically significant. That second analysis belongs beside the first, not in a footnote. The authors suggested that body weight might partly explain the association. The analysis does not settle whether weight is a pathway, a source of confounding, or some combination.
InterAct Consortium. Diabetologia, 2015. PMID 26021487 · DOI
Sources of fiber are not interchangeable results
Across the meta-analysis of 19 cohorts, higher total fiber and cereal fiber intakes were inversely associated with incident diabetes. The pooled estimates for fruit fiber and vegetable fiber were less conclusive, with confidence intervals that included no association. This is not evidence that fruit or vegetables cause harm, nor does it establish that cereal fiber alone accounts for the pattern. Different food-source associations are different statistical questions. A summary about total fiber should not silently become a claim that every isolated fiber product has demonstrated the same outcome. No supplement comparison was assigned in this cohort analysis.
InterAct Consortium. Diabetologia, 2015. PMID 26021487 · DOI
Use the association without turning it into insurance
When reading a fiber headline, ask three questions: Was diet assigned or reported? Were the results stable after body-size adjustment? Did the outcome concern new diagnoses or a laboratory marker? Here the answers make the finding interesting but not a guarantee. Observational comparisons cannot exclude all alternative explanations simply because many variables were included in a model. For everyday meal planning, you can use the paper as a reason to examine your pattern rather than calculate a personal prevention percentage. It does not establish a prescribed amount for you, test a branded food, or certify a recipe as protection against diabetes.
InterAct Consortium. Diabetologia, 2015. PMID 26021487 · DOI