Metabolic & Cardiometabolic
starch-polyphenol complex decreases starch digestibility
In plain terms: Do plant compounds slow how fast starch digests?
Strong support Metabolic & Cardiometabolic 🐭 Non-human evidence
RefutedContestedStrong support
consensus score 0.96
Yes, but only shown in the lab and rodents so far, with hurdles like bitterness and baking breakdown.
📅 Last reviewed: 2026-07-15 ⓘ
Evidence ladder
How far up the ladder this claim has climbed. A high consensus on a low rung means "consistent so far," not "proven in people."
Top evidence so far: Animal studies (Animal)
MechanismIn-vitroAnimalObservationalRCTMeta-analysis
How the studies fall
21 support 0 contradict 0 tested null 1 mixed · 22 sources, 20 independent groups
What the evidence shows
Starch-polyphenol V-type complexes slow amylolysis by a dual mechanism (physical barrier + direct amylase/glucosidase inhibition), lowering glucose and even raising GLP-1/PYY in rodents — partly achievable by adding polyphenol-rich flours to dough. Limits: astringency at effective dose, polyphenol bake-instability, no human glycemic data.
The evidence (27)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Feliziani G, Tagliasco M, Suo X, Pellegrini N, Vittadini E, Bordoni L, Gabbianelli R 2026 · Food Science & Nutrition | in-vitro | contradicts | low | Pasta and bread were made from pigmented wholemeal wheat (Grano Mischio) and from non-pigmented Senatore Cappelli semolina, then digested in a test tube. The pigmented pasta digested faster, not slower: rapidly digestible starch was 70.3% of total starch versus 50.2% in the control pasta, and slowly digestible starch 6.3% versus 14.8%. The two breads did not differ on any starch fraction. The authors attribute the pasta result to polyphenols weakening the gluten network so the strands break up faster. No starch-polyphenol complex was made or measured. |
| Xu L, Wang Z, Feng M, Feng Y, Zhang S, Lu Y, Teng C, Ma K, Hu X, Li Y, Ding B 2025 · Food Chem | in-vitro | supports | moderate | Wheat starch was complexed with polyphenols from red, white and black peanut skins. The estimated glycemic index fell after complexation, to 79.9, 77.4 and 76.9 respectively - values still in the high range, and the abstract does not give the figure for unmodified wheat starch, so the size of the drop cannot be read from it. Short-range order and total crystallinity both rose (3.91% black, 3.45% red, 3.44% white). Polyphenols released during digestion suppressed alpha-amylase, and regression analysis identified that enzyme inhibition, not the ordered structure, as the main driver. |
| Fang S, Lan W, Cheng W, Wang C, Li R, Chen Y, Liu Y, Du X, Wang N 2026 · Int J Biol Macromol | in-vitro | supports | moderate | Ginkgo starch was debranched with pullulanase for 0, 3, 6, 12 or 24 hours and complexed with ginkgo biloba extract. Longer debranching raised the share of long chains (DP 25-36), which tracked the complexation index (r = 0.95, p < 0.05). Complexation peaked at 63.94% in the 12-hour sample, which also showed the highest crystallinity (19.63%) and molecular order, and the highest resistant starch, 52.31%. Every sample carried the extract, so the comparison reported is between debranching times rather than against starch with no polyphenol. |
| Aalim H, Khalifa I, Shishir MRI, Sun J, Zhou C, Zou X 2026 · Food Chemistry: X | in-vitro | mixed | moderate | Rice starch was gelatinized with 0-15% rice protein, then bound with the anthocyanin cyanidin-3-O-glucoside, with an anthocyanin-free control at each protein level. The results run both ways. In starch with no protein, the anthocyanin significantly raised rapidly digestible starch and significantly lowered resistant starch - the opposite of slowing digestion. Resistant starch rose across the protein series instead. The abstract says slowly digestible starch fell, but the results section reports a slight rise (52.6-59.4%). Test-tube digestion in triplicate; no food, animal or person tested. |
| Jiapeng Tian, Xuan Zhang, Wendi Zhang, Kexue Zhu, Xiaoai Chen, Yutong Zhang, Zuohua Xie, Lixiang Zhou, Yanru Zhou, Yanjun Zhang, Juxiu Li 2026 · Foods. 2026 Feb 8;15(4):612 | in-vitro | supports | moderate | In-vitro study of breadfruit starch complexed with three phenolic acids. Against native starch (resistant starch 47.19%), the phenolic-acid complexes raised resistant starch to 59.30% (gallic), 57.29% (DOPAC) and 70.14% (caffeic) and cut the hydrolysis index from 80.61 to 63.24, 65.17 and 52.60. Adding lauric acid alongside a phenolic acid was antagonistic: every ternary complex held less resistant starch than its phenolic-only counterpart, and the gallic-acid ternary (49.12%) did not differ significantly from native starch. Three replicates, one starch source; no human glycemic data. |
| Ma S, Wang Z, Fan H, He H 2026 · Foods | in-vitro | supports | moderate | Rice starch was extruded with propyl gallate, a phenolic food antioxidant, at 0, 25, 50 and 100 mg/g. Complexes formed only when PG was present: complex index 88.28% at 50 mg/g, falling to 66.06% at 100 mg/g, with V-type crystalline features. In a simulated digestion, resistant starch was 1.59% in native rice starch, 8.09% after extrusion with no PG, and 25.78% at 100 mg/g PG; rapidly digestible starch fell from 94.89% to 86.52% to 65.29%, and the digestion rate constant from 0.18 to 0.11 per minute. No animal or human testing, and no other polyphenol was tested. |
| Fan H, Sun L, Yao X, Wen Y, Xiang X, Li H, Wang J, Sun B 2025 · Carbohydr Polym | in-vitro | supports | moderate | High-amylose corn starch was complexed with nine polyphenols; four - caffeic acid, dihydroxybenzoic acid, soy isoflavone and naringenin - were compared through in-vitro digestion. The complexes released polyphenols gradually during digestion, which inhibited alpha-amylase: release ranked dihydroxybenzoic acid < soy isoflavone < caffeic acid = naringenin, with inhibition in the opposite order. Digested complexes showed more super-long amylopectin chains and greater short-range order, which the authors link to the enzymatic resistance. The abstract gives no numerical digestibility values. |
| Lv G, Li D, Hou C, Wang M, Zhang X, Zhang W, Bao Y, Zhang Y, Zhang Y 2026 · Food Chem X | in-vitro | supports | moderate | In-vitro study of cassava starch complexed with the phenolic acid sinapic acid, alone and together with the fibre hemicellulose. Against plain cassava starch (hydrolysis index 85.97%, estimated glycemic index 86.91%), the starch-sinapic acid complex cut hydrolysis to 58.27% and estimated GI to 71.70%, and left 47.71% resistant starch. The three-way starch-hemicellulose-sinapic acid complex was lowest of all (hydrolysis 30.44%, estimated GI 56.42%). Test-tube digestion of purified ingredients; the paper does not test any food, animal or person. |
| Li K, Yao F, Du J, Deng X, Li C 2018 · J Agric Food Chem | animal | supports | moderate | Persimmon tannin was tested on starch digestion in test tubes, in intestinal cells and in rats. Oral persimmon tannin-starch complexes lowered blood glucose and the glucose area under the curve dose-dependently, compared with starch alone. The tannin bound starch directly and also inhibited both starch-digesting enzymes, alpha-amylase (IC50 0.35 mg/mL) and alpha-glucosidase (IC50 0.24 mg/mL); 20 ug/mL cut glucose uptake and transport in Caco-2 intestinal cells. Read from the abstract, which gives no rat numbers, doses or variability. |
| Zheng Y, Chen T, Li K, Wang L, Liu X, Zhao L, Hu Z, Wang K 2025 · Carbohydr Polym | animal | supports | moderate | Gallic acid and tannic acid were bound to high-amylose maize starch, forming V-type crystalline complexes. In laboratory digestion the complexes broke down more slowly and left less digestible starch than uncomplexed starch, which the authors attribute to higher crystallinity plus inhibition of alpha-amylase; the tannic-acid complex, carrying more polyphenol, slowed digestion more than the gallic-acid one. Mice fed the complexes had lower blood glucose, higher GLP-1 and PYY, and greater satiety. The paper reports no digestion or glucose figures and no group sizes. |
| Bharathvaj VR, Sunil CK, Rawson A, Chandrasekar V, Baskaran N, Venkatachalapathy N 2026 · Food Chemistry | in-vitro | supports | moderate | Kodo millet starch was annealed, heat-moisture treated or ultrasonicated before being complexed with quercetin. Ultrasound gave the highest complexation index (18.54%), while heat-moisture gave the highest resistant starch (38.7%, against 24.57% for the ultrasound complex). Complexed starches also had lower peak viscosity (24.08%), more short-range crystallinity (4.86%) and better freeze-thaw stability (up to 83.44%). Digestion was tested in a laboratory assay, and no resistant-starch figure for plain uncomplexed starch is reported, so the size of the gain is not quantified. |
| Wang R, Dong L, Shi P, Liu J, Liu S, Chang X 2026 · Food Chem | animal | supports | moderate | Whole chestnut powder was treated with high hydrostatic pressure (200-600 MPa) plus catechin (3-9%). Catechin binding to the starch rose with pressure and peaked at 600 MPa, and higher catechin levels significantly raised resistant starch content, solubility and freeze-thaw stability (p < 0.05). Digestibility fell in a laboratory digestion model, and in mice the 600 MPa / 6% complex — not the treated powder generally — gave the lowest 2-hour postprandial blood glucose. The published abstract gives no figure for how much the glucose fell. |
| Li Yongxing, Ji Haiyu, Yang Shuang, Cao Wenhao, Zou Jiarong, Zhu Zhangliang, Zhang Jinfang, Li Mei 2026 · Food Res Int | in-vitro | supports | high | Young apple polyphenols were autoclaved with corn (A-type), potato (B-type) and pea (C-type) starch, reorganising all three into A+V or B+V crystalline complexes. Resistant starch in the resulting complexes reached 52.02% (corn), 56.55% (potato) and 57.19% (pea), and potato showed the largest drop in estimated glycemic index (14.83 points). Corn complexes carried the most polyphenol (785.61 mg/g), pea the highest encapsulation efficiency (97.40%). Digestibility was measured in a test tube; the glycemic index is an estimate from that assay, not a measurement in people. |
| Deng X, Zhang Q, Wang Hin NG, Ma Y, Cheung KC, Ho TC, Li C 2026 · Int J Biol Macromol | in-vitro | supports | moderate | Corn and potato starch were complexed with four polyphenols, but only a small share bound (complexing index 2.8-8.3%). Corn starch formed V-type inclusion complexes that physically blocked digestion; potato starch formed non-inclusion complexes that released polyphenols during digestion and lowered enzyme activity. Both created a slowly digestible fraction on amylopectin chains (DP 15-18). Across a Caco-2 cell monolayer the complexes cut glucose absorption by 12-29%, with SGLT1 and GLUT2 expression reduced. No number is given for the digestibility drop, and no animals or people were tested. |
| Jiapeng Tian, Xuan Zhang, Wendi Zhang, Kexue Zhu, Xiaoai Chen, Yutong Zhang, Zuohua Xie, Lixiang Zhou, Yanru Zhou, Yanjun Zhang, Juxiu Li 2025 · Preprints.org | in-vitro | supports | low | Preprint (breadfruit starch + lauric acid + phenolic acids): caffeic-acid binary complex gave highest RS, lowest hydrolysis; ternary complexes' antidigestibility not superior to binary but above native starch. |
| Hu Z, Xu Y, Xiong Y, Huang G 2025 · Foods | in-vitro | supports | low | A narrative review of laboratory work. Polyphenols slot into starch helices (V-type inclusion) and block alpha-amylase: quercetin raised resistant starch in high-temperature maize starch from 8.3% to 32.5% and slowed digestion by more than 40%; a potato starch complex rose from 5.1% to 18.7% resistant starch with a 35% slower enzyme rate; annealing plus acid hydrolysis reached 51%. The authors say human bioavailability is still unvalidated and the products remain at laboratory or small-scale stage. |
| Yuan Y, Chai Z, Zheng Y, Ren Y, Ye X, Kong X, Tian J 2024 · J Food Sci | animal | supports | moderate | Hawthorn polyphenol extract was mixed into corn starch at 1-4% by weight. The polyphenol drove small V-type crystal formation, lowered paste viscosity and raised the gelatinization temperature, and increased the slowly digestible and resistant starch fractions to 36.32% and 33.32%; the abstract does not say which polyphenol level produced those two figures, nor what the plain corn starch measured. Mice given the hawthorn-polyphenol starch had a significantly lower 30-minute blood glucose peak than mice given raw corn starch, 11.77 versus 14.30 mmol/L. The number of mice is not reported. |
| Li L, Xu Y, Huang Y, Wu Y, Li C 2026 · Foods | animal | supports | moderate | Chestnut starch-resveratrol complexes were extruded into fresh noodles. In simulated digestion these noodles were digested least of all samples at 180 minutes (45.92% hydrolysed, against 70.58% for commercial wheat noodles and 82.96% for white bread), giving a predicted glycemic index of 53.35 versus 70.65 - predicted from the hydrolysis curve, not measured in people. A separate mouse experiment gavaged gelatinised starch pastes rather than the noodles: 40 mice, 8 per group, peak blood-glucose rise 4.31 mmol/L on the resveratrol-complexed chestnut starch against 7.86 mmol/L on wheat starch. |
| Lv W, Zhang L, Zhang M, Liu X, Zhou Y, Wang Y 2025 · Food Chem | in-vitro | supports | moderate | In a test-tube digestion model, wheat starch gels resisted breakdown when nanoparticles were added: starch hydrolysis fell to 57.7% with resveratrol alone, 62.8% with plain gliadin nanoparticles, and 41.0% with resveratrol-loaded gliadin nanoparticles. Alpha-amylase activity fell to 51.5%, 68.4% and 32.9% respectively, and fluorescence, docking and confocal imaging showed resveratrol binding the enzyme (7 hydrogen bonds, 49 hydrophobic interactions). The glycemic index of about 45 was calculated from the digestion curve, not measured in people. |
| Meng R, Feng R, Wu SJ, Ji QY, Tao H, Zhang B 2025 · Food Chem | in-vitro | supports | moderate | Rice starch was complexed in the lab with each of several polyphenols plus whey protein. Of the complexes tested, the quercetin one digested least (55.7% digestibility) and held the most resistant starch (46.0%), which the authors put down to quercetin inhibiting amylases most strongly; quercetin also showed the highest enzyme-inhibitory activity in the presence of whey protein, and flavonoid polyphenols bound the protein more tightly than others. Every complex with a digestibility figure also contained whey protein, and no digestibility value for plain rice starch is reported. |
| Xie W, Wang W, Feng X, Zheng R, Chen L, Ding N, Chen Q, Lin S 2026 · Foods | in-vitro | mixed | low | Quercetin was bound to oat beta-glucan, a soluble fibre rather than a starch, forming a stable complex held together by hydrogen bonds. Stirred into corn starch pastes the complex made them thicker, and in a test tube it inhibited the starch-digesting enzyme alpha-amylase more strongly than either ingredient alone: half-maximal inhibition at 2.629 mg/mL, against 3.32 for quercetin and 5.297 for beta-glucan. The authors say the study did not measure starch digestion or glucose release, so it shows no direct effect on starch digestibility. |
| Zhang S, Li W, Liu K, Huang Z, He X, Li H, Sun J 2026 · Foods | in-vitro | mixed | low | Cookies were baked three ways: a plain butter control, algae powder only, and algae powder plus a corn-oil emulsion gel carrying 0.1% grape-seed polyphenol. Rapidly digestible starch came out at 59.6%, 50.2% and 54.0%; resistant starch at 22.2%, 28.7% and 25.8%. Adding the polyphenol-bearing gel to the algae cookie significantly raised rapidly digestible starch (p<0.05) - the authors say the gel partly counteracted the algae's slowing effect. No gel-only cookie was baked, and no starch-polyphenol complex was made or measured. |
| Kusumawardani S, Luangsakul N 2026 · Ultrasonics Sonochemistry | in-vitro | supports | moderate | Corn (A-type), potato (B-type) and mung bean (C-type) starch were complexed with red rice bran polyphenols by gelatinization plus ultrasound, in triplicate. At the best setting, 60% amplitude, rapidly digestible starch fell (corn 36.5% to 15.2%) and resistant starch rose (43.9% to 58.9%), taking the estimated glycemic index from 75-77 in the untreated starches down to 65.9-67.9; the other settings gave 67.3-72.8. Potato starch bound the most polyphenol (8.04 mg GAE/g). X-ray diffraction showed no new peaks, so the binding was not V-type. The only comparator was untreated starch. |
| Zhu Z, Ma X, Pan X, Du X, Chen X 2025 · Food Chem X | in-vitro | supports | moderate | Lab study of pre-formed V-type wheat starch complexed with green-tea polyphenols. Digestion was measured on the complex powder itself, not on noodles: resistant starch rose from 31.77% in the polyphenol-free control starch to 43.3% at 5% polyphenols and 46.79% at 10% (the 47.28% relative rise), while rapidly digestible starch fell from 46.59% to 37.83%; 7.5% and 10% were no better than 5%. The steamed cold noodles made with the complexes were tested only for colour and texture - 10% polyphenols cut hardness 51.46% - never for digestion or blood glucose. |
| Meng R, Feng R, Wu SJ, Ji QY, Tao H, Zhang B 2025 · Food Chem | in-vitro | supports | moderate | Rice starch was complexed in the lab with quercetin and whey protein, then digested at different pH values. In acid (pH 3) the complex became more ordered, more thermally stable, less soluble and less swollen, digested more slowly and left 49.25% resistant starch. In alkali (pH 9) the quercetin degraded, the ordered structure broke down, starch was hydrolysed faster and resistant starch fell. The contrast tested is between pH conditions of the same starch-quercetin-protein complex, not between starch with and without quercetin. |
| Li C, Zhu W, Luo F 2026 · Ultrasonics Sonochemistry | in-vitro | supports | moderate | Chestnut starch complexed with cyanidin-3-O-glucoside digested more slowly in a test tube than plain chestnut starch: 18.30% hydrolysed at 20 min and 77.33% at 180 min versus 25.10% and 85.53%, resistant starch 43.93%, modelled end-point of digestion 86.17% versus 93.20%. Ultrasound pretreatment before complexing pushed resistant starch to 49.27% and 180-min hydrolysis down to 68.33%. C3G also inhibited alpha-amylase directly, 8.33-23.81% at 0.05-0.2 mg/mL. Simulation showed it detaching from amylose after 80 ns, so binding is transient. No animal or human testing. |
| Yu M, Qu C, Li D, Jiang Z, Liu J, Yang F, Liu C, Yue W, Wu Q 2024 · Int J Biol Macromol | in-vitro | supports | moderate | Five phenolics abundant in Euryales semen (fox nut) - naringenin, isoquercitrin, gallic acid, epicatechin and quercetin - were complexed with Euryales semen starch. The complexes markedly lowered the starch's in-vitro digestibility, swelling power and gelatinisation enthalpy and raised its solubility; infrared and X-ray analysis showed the phenolics bound the starch non-covalently. Separately, all five inhibited alpha-amylase by mixed competitive inhibition, ranked quercetin > epicatechin > gallic acid > isoquercitrin > naringenin. The paper reports no digestibility or inhibition figures. |
Disagree, or know a study we missed?
We grade by evidence, not opinions. The way to weigh in is to point us to a study we haven't cited (check the evidence table above first), or to flag a problem with one we have. Every submission is reviewed; if it holds up, the grade updates and shows in Science Changes Its Mind.
Opens a short form. You'll sign in with Google so submissions are tied to a real account — we don't display your identity, and we only accept a link we can verify (PubMed, DOI, ClinicalTrials.gov).
Educational only, not medical advice. Grades and scores reflect published evidence weighted by study design and quality; see the methodology.