Beneficial glycaemic effects of high-amylose barley bread compared to wheat bread in type 2 diabetes

Type 2 diabetes prevalence is increasing, driven by obesity, inactivity and unhealthy diet. Low glycaemic index (GI) and glycaemic load diets can prevent T2D and improve its management and cardiovascular risk. Cereal composition—dietary fibre and starch structure (amylose vs amylopectin)—influences postprandial glucose in both diabetic and non-diabetic individuals. A genetically modified barley line (AmOn) with >99% amylose, achieved by RNAi suppression of starch branching enzymes, shows very low in vitro GI and higher resistant starch, which may slow starch digestion and reduce postprandial glycaemia. Barley generally yields lower GI than wheat and has beneficial effects on gut hormones and appetite, but effects of hulless barley in T2D and of the AmOn high-amylose barley in humans had not been clarified. The study aimed to test whether bread made with AmOn or hulless barley improves 4-hour postprandial glucose responses (iAUC) versus wheat bread in T2D, with concurrent assessment of insulin, glucagon, incretins and lipids. The hypothesis was that AmOn or hulless barley breads would lower postprandial glucose iAUC compared with wheat bread in T2D.

Prior work indicates that higher dietary fibre (notably beta-glucan) and higher amylose content reduce postprandial glycaemia. Barley and oat products have been shown to lower postprandial glucose; resistant starch (RS) can reduce glycaemic responses while affecting incretin hormones. In vitro and some human studies suggest high-amylose starches digest more slowly, increase RS formation (e.g., RS3 upon retrogradation), and lower predicted GI. Mechanistic factors include starch structure, granule architecture, processing, and non-starch components. Barley’s lower GI than wheat and its potential to beneficially modulate gut hormones and insulin sensitivity have been reported, but data specifically in T2D for hulless barley and for the AmOn amylose-only barley were lacking.

Design: Acute, randomized, single-blinded, cross-over intervention with four test meals (breads). Each participant attended four separate visits and consumed one bread per visit in randomized order (randomization via REDCap). Primary outcome: 4-hour postprandial plasma glucose response quantified as incremental area under the curve (iAUC). Secondary outcomes: postprandial insulin, glucagon, triglycerides (TG), free fatty acids (FFA), glucagon-like peptide-1 (GLP-1), and gastric inhibitory polypeptide (GIP), quantified as iAUC or total AUC (tAUC) as appropriate. Participants: Twenty adults with T2D were recruited; 18 completed all meal tests (one excluded due to failed IV access; one withdrew). Inclusion: adults ≥18 years, T2D (IDF criteria), HbA1c 42–78 mmol/mol. Exclusion: type 1 diabetes, insulin-treated T2D, weekly GLP-1 agonist, acarbose, significant cardiovascular/renal/hepatic/endocrine disease, significant psychiatric history, steroid treatment, alcohol/drug abuse, pregnancy/breastfeeding, legal incompetence. Stable treatment for hypertension/hypercholesterolemia allowed if unchanged during study. Ethics approval obtained; trial registered (NCT04646746). Test meals (breads): Four bread types: (1) 100% wheat flour (Manitoba wheat); (2) 50% hulless barley flour + 50% wheat; (3) 75% hulless barley + 25% wheat; (4) 50% genetically modified amylose-only barley flour (AmOn) + 50% wheat. Breads were produced by commercial bakery for wheat/hulless barley and by Plantcarb ApS for AmOn (GM flour not allowed in commercial Danish bakeries). Grains grown in 2020; AmOn based on hulled barley (H. vulgare var. Golden Promise) with RNAi suppression of SBEIIa, SBEIIb, SBEI (>90% reduction) to produce >99% amylose. Hulless barley variety PS3. Portions were calculated to deliver approximately 50 g total carbohydrate (from fibre + starch): 114 g (100% wheat), 119 g (50% hulless), 122 g (75% hulless), 128 g (50% AmOn). Breads were packaged, frozen at −20 °C, and defrosted without heating on study days. Bread composition analyses: Moisture by drying (120 °C, 24 h). Total carbohydrate as fibre + starch. Total starch measured with Megazyme K-TSTA kit variant (DMSO and boiling to solubilize all starch including resistant starch). Total dietary fibre by Megazyme K-TDFR-200A kit (primarily cell wall polysaccharides and some RS). Group differences assessed (see results for values in g/100 g and g/serving). Pre-visit standardization: Standardized evening meal (commercial chili con carne; women: 536 kJ, men: 638 kJ) before each visit; overnight fast from midnight. No smoking overnight/during visits; no alcohol for two days prior; no strenuous exercise day prior. Antihypertensive, cholesterol-lowering, and antidiabetic drugs were paused 24 h before each visit. Washout between visits ≥6 days. Clinical procedures: Arrival 07:00–07:30. IV catheter placed; baseline blood at −10 and 0 min. Test bread consumed within 10 min with 250 mL water. Blood sampling over 240 min at specified times: glucose/insulin/glucagon at −10, 0, 10, 20, 30, 45, 60, 90, 120, 150, 180, 210, 240 min; TG/FFA/GLP-1/GIP at −10, 0, 30, 60, 120, 180, 240 min. Samples centrifuged (3000 g, 10 min, 4 °C), plasma frozen at −20 °C then stored at −80 °C. Assays: Plasma glucose (glucose oxidase, GOD-PAP, Roche). Serum insulin and glucagon (ELISA, Mercodia). Plasma TG and FFA (enzymatic colorimetry; Roche TG kit; Wako FFA kit) run on Cobas c111. GLP-1 and GIP (NL-ELISA, Mercodia). Statistics: Power calculation based on detecting 20% difference in primary outcome at 80% power required n=20 completers (α=0.05, β=0.20). Mixed model ANOVA assessed bread effects vs wheat control; P<0.05 significant. Data presented as mean ± 95% CI (tables) and mean ± SEM (graphs). iAUC (area above fasting) for glucose, insulin, GLP-1, GIP; tAUC (area above zero) for glucagon, TG, FFA. Software: STATA 17 and GraphPad Prism 6.

• Participants: 20 randomized; 18 completed (72% female; mean age 60.5±2.7 y; HbA1c 49.7±1.1 mmol/mol; BMI 30.5±1.1 kg/m²).
• Bread composition per serving (approximate): total carbohydrate similar by g/100 g, but per-serving to target ~50 g differed: 49.3 g (100% wheat), 54.4 g (50% hulless), 51.9 g (75% hulless), 57.1 g (50% AmOn). Fibre per serving highest in 50% AmOn (16.4 g), then 50% hulless (11.2 g), 75% hulless (9.2 g), wheat (5.0 g). Total starch per serving: ~44.3 g (wheat), 45.1 g (50% hulless), 40.7 g (75% hulless), 40.8 g (50% AmOn).
• Fasting values: No differences among breads for glucose, insulin, glucagon, GIP, GLP-1 (P>0.05). Baseline TG and FFA showed some differences by chance (see limitations).
• Primary outcome (4 h glucose iAUC): • 50% AmOn vs comparators: reduced by 34% vs 100% wheat (difference 248 mmol/L·240 min; 95% CI: 175, 321; P<0.001); 27% vs 50% hulless (194; 95% CI: 121, 267; P<0.001); 23% vs 75% hulless (167; 95% CI: 93, 240; P<0.001). • 75% hulless vs 100% wheat: reduced by 11% (81 mmol/L·240 min; 95% CI: 8, 155; P=0.030). • Glucose peak reductions for 50% AmOn vs wheat, 50% hulless, 75% hulless: −1.5 (95% CI: 0.9, 2.0; P<0.001), −1.4 (95% CI: 0.9, 2.0; P<0.001), and −1.0 mmol/L (95% CI: 0.4, 1.5; P=0.001), respectively.
• Insulin iAUC: • 50% AmOn vs 100% wheat: −24% (−7.7 nmol/L·240 min; 95% CI: −14.9, −0.5; P=0.035). • 50% AmOn vs 50% hulless: −35% (−13.3 nmol/L·240 min; 95% CI: −20.5, −6.2; P<0.001). • 75% hulless vs 50% hulless: −22% (−8.5 nmol/L·240 min; 95% CI: −15.7, −1.3; P=0.021). • No significant differences: wheat vs 50% hulless (P=0.121); wheat vs 75% hulless (P=0.422).
• Glucagon tAUC: No differences among breads (P>0.05).
• Incretins: • GIP iAUC: 50% AmOn lower than wheat (−4.6 nmol/L·240 min; 95% CI: −6.2, −3.0; P<0.001), lower than 50% hulless (−3.2; 95% CI: −4.9, −1.6; P<0.001), and lower than 75% hulless (−1.8; 95% CI: −3.4, −0.2; P=0.032). 75% hulless lower than wheat (−2.8; 95% CI: −4.5, −1.2; P=0.001). • GLP-1 iAUC: No differences among breads (P>0.05).
• Lipids: • TG tAUC: 75% hulless higher than wheat (+31.8 mmol/L·240 min; 95% CI: 3.6, 60.0; P=0.028) and higher than 50% hulless (+41.3; 95% CI: 13.1, 69.6; P=0.005). 50% hulless lower than wheat (−31.8; 95% CI: −60.1, −3.6; P=0.028). Interpretation complicated by baseline differences. • FFA tAUC: Suppressed after all breads; suppression greater after 50% hulless than after 50% AmOn (difference −12.4 mmol/L·240 min; 95% CI: −20.0, −4.6; P=0.002). Overall interpretation limited due to baseline differences. Overall, replacing wheat with 50% AmOn or 75% hulless barley improved postprandial glycaemia in T2D.

The study directly tested whether high-amylose (AmOn) or hulless barley breads acutely improve postprandial glycaemic control in T2D. The significant reductions in glucose iAUC with 50% AmOn (23–34% vs other breads) and with 75% hulless barley (11% vs wheat) demonstrate that substituting wheat flour with these barley flours beneficially modulates acute glycaemia without increasing insulin iAUC; indeed, insulin iAUC was lower in key comparisons, suggesting improved glycaemic handling independent of heightened insulin secretion. Lower GIP responses to barley breads, particularly AmOn, align with reduced glycaemic excursions and higher fibre/resistant starch content, whereas GLP-1 was unchanged. Mechanistically, the very high amylose content in AmOn likely slows digestion through reduced gelatinization, increased granule resistance, and greater RS (especially RS3) formation, thereby attenuating glucose release. Higher dietary fibre and possible inclusion of hard-to-degrade amylose in fibre measurements may have contributed. Although TG and FFA responses showed some differences, interpretation is limited by baseline variability. Collectively, findings support the use of high-amylose and hulless barley flours to lower dietary GI and postprandial glycaemia in T2D, with potential cardiometabolic benefits. Practical considerations include barley’s agronomic advantages and the need to address baking quality to enhance consumer acceptability.

Replacing wheat flour with 50% amylose-only high-amylose barley (AmOn) or 75% hulless barley in bread acutely lowers postprandial glucose responses in adults with T2D; insulin iAUC is also reduced in key comparisons, and GIP responses are lower for barley breads without changes in GLP-1. These results indicate a beneficial impact on postprandial glucose regulation. Future research should include longer-term trials integrating such breads into habitual diets to assess sustained effects on glycaemic control (e.g., HbA1c), body weight, lipid metabolism, satiety, and gut microbiota. Work to improve baking qualities of barley-based breads and develop non-GM high-amylose barley varieties could facilitate broader adoption.

• Acute, short-term intervention; no long-term outcomes (e.g., HbA1c, weight, microbiome, satiety) assessed.
• Small sample (n=18 completers), though powered for 20% difference in primary outcome; generalizability limited.
• Baseline (fasting) TG and FFA differed by chance across visits, complicating interpretation of lipid tAUCs.
• Potential misclassification of some hard-to-degrade amylose as fibre with the assay used.
• Portions differed in grams per serving to match ~50 g carbohydrate; absolute intake differences could influence responses.
• Antidiabetic and cardiometabolic medications were paused 24 h pre-visit, which may limit generalizability to usual treated conditions.
• GM AmOn flour is not commercially permitted in Danish bakeries; findings may face translational barriers; sensory/baking quality of barley breads is lower than wheat and was not optimized here.
• No assessment of gut microbiota or subjective appetite/satiety; no heating after thawing could influence starch retrogradation/RS formation but was standardized across breads.