Insoluble cereal fibres have been shown in large prospective cohort studies to be highly effective in preventing type 2 diabetes, but there is a lack of interventional data. Our 2 year randomised double-blind prospective intervention study compared the effect of an insoluble oat fibre extract with that of placebo on glucose metabolism and incidence of diabetes. METHODS: A total of 180 participants with impaired glucose tolerance underwent a modified version of the 1 year lifestyle training programme PREvention of DIAbetes Self-management (PREDIAS) and were randomised to receive a fibre supplement (n = 89; 7.5 g of insoluble fibre per serving) or placebo (n = 91; 0.8 g of insoluble fibre per serving) twice daily for 2 years. Eligible participants were men and women, were at least 18 years old and did not report corticosteroid or other intensive anti-inflammatory treatment, fibre intolerance or any of the following disorders: overt diabetes, chronic or malignant disease, or severe cardiopulmonary, endocrine, psychiatric, gastrointestinal, autoimmune or eating disorder. Participants were recruited at two clinical wards in Berlin and Nuthetal. The allocation was blinded to participants and study caregivers (physicians, dietitians, study nurses). Randomisation was conducted by non-clinical staff, providing neutrally numbered supplement tins. Both supplements were similar in their visual, olfactory and gustatory appearance. Intention-to-treat analysis was applied to all individuals. RESULTS: After 1 year, 2 h OGTT levels decreased significantly in both groups but without a significant difference between the groups (fibre -0.78 ± 1.88 mmol/l [p ≤ 0.001] vs placebo -0.46 ± 1.80 mmol/l [p = 0.020]; total difference 0.32 ± 0.29 mmol/l; not significant). The 2 year incidence of diabetes was 9/89 (fibre group) compared with 16/91 (placebo group; difference not significant). As secondary outcomes, the change in HbA1c level was significantly different between the two groups (-0.2 ± 4.6 mmol/mol [-0.0 ± 0.0%; not significant] vs +1.2 ± 5.2 mmol/mol [+0.1 ± 0.0%; not significant]; total difference 1.4 ± 0.7 mmol/mol [0.1 + 0.0%]); p = 0.018); insulin sensitivity and hepatic insulin clearance increased in both groups. After 2 years, improved insulin sensitivity was still present in both groups, although the effect size had diminished. Separate analysis of the sexes revealed a significantly greater reduction in 2 h glucose levels for women in the fibre group (-0.88 ± 1.59 mmol/l [p ≤ 0.001] vs -0.22 ± 1.52 mmol/l [p = 0.311]; total difference 0.67 ± 0.31 mmol/l; p = 0.015). Levels of fasting glucose, adipokines and inflammatory markers remained unchanged in the two groups. Significantly increased fibre intake was restricted to the fibre group, despite dietary counselling for both groups. No severe side effects occurred. CONCLUSIONS/INTERPRETATION: We cannot currently provide strong evidence for a beneficial effect of insoluble cereal fibre on glycaemic metabolism, although further studies may support minor effects of fibre supplementation in reducing glucose levels, insulin resistance and the incidence of type 2 diabetes. TRIAL REGISTRATION: clinicaltrials.gov NCT01681173 Funding: German Diabetes Foundation (grant no. 232/11/08).