Supplements · Sweeteners · Diets · Metabolic & Cardiometabolic
allulose increases GLP-1
In plain terms: Does the rare sugar allulose boost the appetite-suppressing gut hormone GLP-1?
Strong support Supplements 💰 Industry COI noted
Part of: 🧪 allulose
RefutedContestedStrong support
consensus score 1.00
Yes, but mostly shown in mice; a clean human trial didn't see a reliable effect yet.
📅 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: Human trials (RCT / n-of-1)
MechanismIn-vitroAnimalObservationalRCTMeta-analysis
How the studies fall
9 support 0 contradict 0 tested null 0 mixed · 9 sources, 9 independent groups
What the evidence shows
Allulose (D-psicose) acutely increases GLP-1 secretion. A high-quality human RCT (Wölnerhanssen 2022, J Nutr, PMID 35135006) found an oral allulose bolus induced a significant plasma GLP-1 release vs control (P_Holm<0.0001, large effect d_z>1), corroborated by consistent rodent data and enteroendocrine-cell mechanism. The acute incretin effect is real and established.
The evidence (14)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Iba K, Kyo M, Ishihara H, Nagao A, Kawabe M, Ohbayashi K, Yada T, Iwasaki Y 2026 · Int J Mol Sci | animal | supports | moderate | OFF-SCOPE, zero weight - the paper never measured GLP-1. [FT-verified, PMC JATS full text PMC12940309 (111,059 bytes); the source note said provenance: abstract and tags abstract-only, and the full text has been cached since 2026-07-17.] Iba 2026, Int J Mol Sci 27(4):1651, doi 10.3390/ijms27041651, Iwasaki/Yada lab, Kyoto Prefectural University. THE CLAIM'S OBJECT IS GLP-1 THE HORMONE (claim scope: 'acute GLP-1 secretion'); across all three experiments the only immunoassays run were 17beta-estradiol (an ovariectomy check) and insulin - 'Plasma insulin concentrations were determined using an insulin ELISA kit' - and no plasma, portal or intestinal GLP-1 concentration appears in any Result, Figure legend or Method. GLP-1 enters the paper only as a RECEPTOR NECESSITY test: 'Finally, we investigated the involvement of the GLP-1 receptor in these effects using GLP-1 receptor knockout mice.' The design actively steers AWAY from the incretin axis: 'In this study, IPGTT was used instead of an oral glucose tolerance test to minimize the influence of gastrointestinal digestion, absorption, and gut-derived incretin effects on blood glucose responses.' A Glp1r-KO abolition cannot discriminate a rise in the LIGAND from potentiated RECEPTOR signalling - sibling ev-s41327798 tested exactly the latter (allulose stabilising GLP-1R) and was off-scoped on this claim, as was ev-s35561584 (GLP-1 applied as a reagent); this is the third row of that shape. THE PRIOR EXTRACT QUOTED THE TITLE, NOT A RESULT: 'D-allulose-induced GLP-1 release ameliorated sucrose-induced obesity and glucose intolerance in ovariectomized mice' is a paraphrase of the article title ('GLP-1 Release by Rare Sugar D-Allulose Ameliorates Sucrose-Induced Obesity and Glucose Intolerance in Ovariectomized Mice'); the release half of that title is imported from this lab's earlier work, and the paper's own restatement is citation-carried, not measured - 'our results show that D-allulose, which stimulates endogenous GLP-1 secretion and activates GLP-1 receptor signaling, effectively improves glucose metabolic abnormalities in OVX mice.' WHAT WAS MEASURED, WITH EACH EXPERIMENT'S OWN DENOMINATOR: Exp 1 - OVX n=10 vs sham n=10, water only, IPGTT Day 43, cull Day 63. Exp 2 - OVX n=9 vs sham n=10, two-bottle 25% sucrose from Day 33, IPGTT Day 71, cull Day 82. Exp 3 (THE ALLULOSE EXPERIMENT, n=7 PER GROUP by the Figure 4-6 legends) - 25% (w/w) sucrose-enriched diet from Day -5, D-allulose 3 g/10 mL/kg or water by once-daily oral gavage from Day 5 for 2 weeks, IPGTT Day 15, cull Day 20, run separately in wild-type and Glp1r-KO mice with 'statistical analyses performed separately within wild-type and GLP-1R KO mice'. Endpoints: body weight, food intake, intraperitoneal glucose AUC, fasting glucose, plasma insulin, HOMA-IR, visceral (mesenteric + perirenal + periuterine) fat mass, uterine weight. Those findings are logged as candidates, including one that cuts against the marketing story: 'In contrast, D-allulose did not markedly affect body weight gain in OVX mice' (n=7) - un-appraised evidence bearing on claim-allulose-reduces-body-fat, whose SPO object is body weight. Grade (animal) and quality (moderate) left unchanged as moot at zero weight, and stance left untouched, matching how siblings ev-s35561584, ev-s41327798, ev-s32193016 and ev-s41713874 were off-scoped on this claim. SPONSOR - confirms from the paper's own text the check reread #697 made against the cache: 'This study was supported by a grant from Matsutani Chemical Industry Co., Ltd. to T.Y. and Y.I.'; 'D-allulose was provided by Matsutani Chemical Industry Co. Ltd. (Itami, Japan), with purities exceeding 98%'; and the COI adds 'T.Y. and Y.I. have received grant support from Matsutani Chemical Industry Co. Ltd.' with the mitigation that the company 'only provided D-allulose but was not involved in the conduction of the current study'. INDEPENDENCE FLAG (not acted on, rubric axis 7): the author block Iba K, Kyo M, Ishihara H, Nagao A, Kawabe M, Ohbayashi K, Yada T, Iwasaki Y is the same Kyoto Prefectural laboratory as s40218979 and shares Iwasaki Y and Yada T with s29317623 - group_id 'iba-ovx' is the FIFTH label for one collaboration network on this claim (followup 176). |
| Mizuma S, Hayakawa M, Hira T 2025 · Endocrinology | animal | supports | moderate | [Abstract-only; NCBI efetch XML, no PMC full text - BioC pmcoa returns no result and the PubMed record carries no PMCID, so every n, effect size and statistic below is CANNOT-TELL rather than absent.] Mizuma 2025, Endocrinology, male rats (MeSH: Rats, Sprague-Dawley). ON THE CLAIM'S OBJECT the paper is direct and confirmatory: 'Our results demonstrate that allulose promotes GLP-1 secretion, at least in part, via intestinal distension as a novel GLP-1 secretory mechanism.' Dose relation is stated as amount, not concentration - 'The secretion of GLP-1 increased as the absolute amount of allulose increased, independent of the concentration' - and 'The extent of the GLP-1 secretory response was positively correlated with the intestinal content volume and diameter after allulose administration.' WHAT THE MECHANISM MEANS FOR THE CLAIM (off-object, logged as candidates, not scored here): the release is attributed to a NON-SPECIFIC physical stimulus, not to allulose chemistry - 'Polyethylene glycol 1000 and D-mannitol, which are water-soluble and poorly absorbable, stimulated GLP-1 secretion' and 'Furthermore, the intraileal administration of air expanded the intestine-induced secretion of GLP-1', i.e. GLP-1 rose with no allulose present at all. Real specificity controls are reported and they failed to stimulate - 'However, cellulose (insoluble), and tetra ethylene glycol (water-soluble and absorbable) did not' - which is a genuine internal-validity strength and is why quality stays moderate despite no recoverable numbers. NO plasma compartment, timepoint, group size, effect size, p-value, randomisation or blinding is recoverable from the abstract. NOT an independent replication by the paper's own account: 'We previously observed intestinal distension after the oral administration of allulose, accompanied by increased GLP-1 secretion in rats' - it extends this lab's prior result. SPONSOR (answers the unchecked slot in followup 175): sole listed funder in the PubMed GrantList is 'Matsutani Chemical Industry Co. Ltd.', the allulose manufacturer - so this is a SIXTH supplier-linked row among the claim's nine animal rows, and a fifth Matsutani-linked one. Supplier of the test allulose is not stated in the abstract (cannot-tell). INDEPENDENCE FLAG (not acted on): third/senior author Hira T, Laboratory of Nutritional Biochemistry, Hokkaido University - the same corresponding author and lab as s29402406 (Hayakawa 2018) and a co-author of s29317623, both already named in followup 171 as one Hokkaido/Matsutani lab counted as two voices. group_id 'mizuma' is a THIRD label for that same lab; merge belongs in the allulose pass. |
| Wolnerhanssen BK, et al. 2022 · J Nutr | RCT | supports | high | human RCT n=18 DB 25g intragastric D-allulose raised GLP-1/CCK/PYY vs water P<.0001 dz>1; independent |
| Iwasaki Y, Sendo M, Dezaki K, Hira T, Hara H, Drucker DJ, Tokuda M, Yada T, et al 2018 · Nat Commun | animal | supports | high | [FT-verified] Mouse: p.o. D-allulose 1 g/kg raised portal-vein ACTIVE GLP-1 at 0.5 h, plateau 1-2 h, baseline by 3 h (n=5-9/group); dose-dependent at 1 and 3 g/kg but NOT 0.3 (n=5-15/dose); GIP, CCK and PYY unchanged while oral glucose raised GIP not GLP-1; rise ~2-fold larger in HFD-fed mice (n=5-6). Oral route only - i.p. allulose raised no GLP-1. Release is UPSTREAM of the vagus: intact in Glp1r KO mice, and 20 mM allulose did not activate nodose neurons; vagal afferents carry the downstream feeding/glucose effects, not the release. ANIMAL only, no human data. Allulose supplied by Matsutani Chemical; 2 authors hold Matsutani grant support. |
| Noh 2026 · Nutrients | RCT | supports | low | [FT-verified] human RCT n=10 allulose+meal raised GLP-1 tAUC +12% p=.042 lower glucose/insulin; Samyang COI; not null |
| Deemer SE, Plaisance EP, Martins C 2020 · Nutrition Research | mechanism | mixed | low | OFF-SCOPE, zero weight. The paper's own statement of subject and outcome: 'The aim of this review is to describe the status of knowledge regarding the impact of ketosis, induced by diet or exogenous ketones (ketone esters), on appetite and the potential mechanisms involved.' Its findings are about ghrelin, hunger and satiety — 'In humans, acute consumption of a ketone ester drink reduced feelings of hunger and increased satiety compared to a dextrose drink.' The previous extract asserted the review 'notes non-metabolizable sweeteners and ketones modulate gut satiety hormones' and read that as context for allulose acting via a gut sweet-taste/GLP-1 pathway; neither allulose nor GLP-1 nor any sweetener occurs in the abstract on record, and the full text is paywalled (Europe PMC isOpenAccess:N), so that bridge is unsupported by any text this vault has read. It is also a narrative review, which is zero weight regardless of subject. |
| Hayakawa M, Hira T, Nakamura M, Iida T, Kishimoto Y, Hara H 2018 · Biochem Biophys Res Commun | animal | supports | moderate | [FT-verified from the Hokkaido HUSCAP open-access accepted manuscript; source note previously said provenance: abstract] Rat in vivo, male SD, 16 h fasted. Oral D-allulose 0.5, 1.0 and 2.0 g/kg raised plasma total GLP-1 for more than 2 h dose-dependently (dose-response n=8/group; 2.0 g/kg vs water n=12/group), and induced clearly higher GLP-1 than fructose (n=7 vs 6) or resistant maltodextrin, while maltodextrin did not differ from water. In the PORTAL vein after ORAL 2.0 g/kg, total GLP-1 was higher at 60 and 150 min and active GLP-1 at 60 min, while total GIP did not change; that GIP null has its own denominator of n=4 per time point vs n=8 untreated, and comes from this oral portal experiment, NOT from the luminal experiments. Oral glucose did the reverse (raised GIP, no GLP-1 rise). Site of action: intraperitoneal allulose produced no GLP-1 rise (n=5-6), whereas duodenal (n=6-8) and ligated-ileal-loop (n=8) allulose raised portal GLP-1, so the effect is luminal. Mechanism: duodenal xanthohumol (glucose/fructose transport inhibitor) left GLP-1 only slightly increased (abstract says abolished; n=6-8), while the SGLT1 inhibitor phloridzin and the sweet-taste antagonist lactisole did not reduce the response; allulose produced no GLP-1 release in GLUTag cells (reported as data not shown). ANIMAL only, no human data, no randomisation or blinding described. Funded by Matsutani Chemical Industry, which also supplied all test sugars; 3 of 6 authors (Nakamura, Iida, Kishimoto) are Matsutani employees. INDEPENDENCE FLAG (not acted on): corresponding author Hira T and senior author Hara H also co-author Iwasaki 2018 (s29317623), stored under a different group_id. |
| Masuda Y, Ohbayashi K, Iba K, Kitano R, Kimura T, Yamada T, Hira T, Yada T, Iwasaki Y 2025 · Nutrients | animal | supports | moderate | [FT-verified, PMC JATS full text PMC11990814; source note said provenance: abstract and carried no venue and no DOI] Masuda 2025, Nutrients, Iwasaki/Yada lab, Kyoto Prefectural University. THIS ROW'S OWN ENDPOINT AND DENOMINATOR: overnight-fasted male C57BL/6J mice, PORTAL-vein plasma ACTIVE GLP-1 (Millipore ELISA) at a SINGLE 1 h timepoint after peroral gavage, n = 5-15 per group (the Figure 2 legend gives only that range, not per-group n). Allulose's own numbers: 1 g/kg 35.8 +/- 5.52 pM vs saline 9.22 +/- 1.02 pM, and 3 g/kg 92.1 +/- 10.7 pM vs saline 15.2 +/- 1.75 pM (one-way ANOVA, Tukey, p < 0.05) - a genuine two-dose dose-response, with both aldohexoses flat at both doses as an internal negative control and the saline vehicle validated against water. WHAT THE PRIOR EXTRACT LEFT OUT. (a) It omitted D-FRUCTOSE, an ordinary caloric sugar, which raised GLP-1 just as much: 'There were no significant differences in the ability to stimulate GLP-1 secretion among the four ketohexoses at each administered dose' (fructose 37.6 and 66.72 pM vs allulose 35.8 and 92.1). On this paper's own data the effect is a ketohexose CLASS property, not something specific to allulose. (b) 'GLP-1-receptor-antagonist-reversible' was attached to the GLP-1 rise; a receptor antagonist cannot block secretion. Ex(9-39) was tested against FOOD INTAKE only, and even there the paper says it 'significantly attenuated, or tended to attenuate' the effect, with per-sugar resolution figure-level only. (c) Food intake is off this claim's object and the suppression 'ceased by 24 h'; it and the antagonist result are logged as candidates. ASSAY CAVEAT THE AUTHORS STATE ABOUT THEIR OWN MEASUREMENT: oral glucose raised no GLP-1, against established human data, and they attribute it to detection failure - 'In mice, GLP-1 undergoes pronounced additional endoproteolytic degradation by neprilysin (neutral endopeptidase 24.11, NEP), which may hinder its detection using ELISA kits'. That argues false negatives, not a false-positive allulose result, so the finding stands. Single timepoint, no time course, one hormone, males only by design, no randomisation or blinding described (cannot-tell). QUALITY HELD AT MODERATE, deliberately: level with sibling s29402406, which also pairs a dose-response with heavy Matsutani involvement, above the low rows s35956407, s36937886 and s38931176, which have neither dose-response nor time course, and well below s29317623, which added a time course, a negative dose, route controls and Glp1r-KO mice. NOT AN INDEPENDENT TEST OF THIS CLAIM: allulose is the paper's carried-forward positive control - 'We previously reported that the rare sugar D-allulose promotes GLP-1 release, thereby preventing hyperphagic obesity' (that reference IS sibling s29317623), and the 1 g/kg dose was chosen from it. SPONSOR (closes an unchecked slot in followup 175): 'T.K. and T.Y. (Takako Yamada) were employed by Matsutani Chemical Industry Co., Ltd.', the company funded the study, supplied all four rare sugars at over 98% purity, and 'The funder provided rare sugars and was partly involved in designing and performing the experiments' - the most explicit sponsor entanglement of any row on this claim, and a SEVENTH supplier-linked animal row. Checking the last unchecked row, s41751787, against its cached full text shows it is Matsutani-funded and Matsutani-supplied too, so 8 of the claim's 9 animal rows are supplier-linked and only s36937886 is known manufacturer-independent. INDEPENDENCE FLAG (not acted on): group_id '40218979' is a FOURTH label for one network - Iba K, Ohbayashi K, Yada T and Iwasaki Y are also the author block of s41751787 (group_id iba-ovx, same Kyoto Prefectural laboratory), Iwasaki/Yada/Hira are s29317623 (iwasaki-yada), and co-author Hira T of Hokkaido links on to s29402406 (hayakawa) and s39821080 (mizuma). Five group_ids, one collaborating network, on a claim whose animal weight they dominate. |
| Busch 2026 · JMIR Research Protocols | observational | tested-null | low | OFF-SCOPE, zero weight - a PROTOCOL, not a result. [FT-verified from PMC JATS full text PMC12963978; the source note's 'abstract-only' tag is wrong and the full text has been cached since 2026-07-17.] Busch 2026, JMIR Research Protocols 15:e81857, doi 10.2196/81857 - the LisA trial protocol (Max Rubner-Institut, Karlsruhe; DRKS00028521; IRRID DERR1-10.2196/81857). THE PAPER REPORTS NO GLP-1 DATA. Its planned primary outcome is exactly this claim's object - 'The primary outcome is the postprandial profile of GLP-1' - but nothing had been analysed when it went to press: 'The statistical analysis for the primary outcome and key secondary clinical outcomes is scheduled to commence immediately following final data lock in January 2026', 'Data processing and statistical analysis for the primary outcome are expected to be completed by June 2026', and 'We anticipate the main results manuscript, focusing on the primary and key secondary clinical outcomes, to be completed and submitted for publication by June 2026.' The direction is a HYPOTHESIS, stated as one - 'We hypothesize that the consumption of allulose will lead to a measurably higher acute postprandial GLP-1 response compared with the aspartame placebo' - and the article is 'poised to provide confirmatory evidence for the acute GLP-1 effects of allulose', i.e. that evidence does not exist yet. THE PRIOR STANCE WAS FALSE ABOUT THE PAPER, not merely unweighted: 'tested-null' asserts the study tested the effect and found none, when the test had not been run. It is left in the field only because it is arithmetically inert at zero weight, matching how siblings ev-s35561584, ev-s41327798 and ev-s32193016 were off-scoped on this claim with stance untouched; if this row is ever restored to counting weight the stance must be corrected in the same edit. Grade 'observational' is likewise moot and likewise wrong - the source note says RCT and the trial is a randomised double-blind crossover - flagged for adjudication, not silently upgraded. WHAT THE PAPER DOES CONTAIN, with its own denominators: 10 COMPLETERS (4F/6M, mean age 31.2 SD 6.8 y, BMI 25.1 SD 2.6 kg/m2), which is a completer figure, not enrolment - 67 were telephone-screened, 35 invited to screening, 14 invited to the study, and 'Of the 12 participants who started the study, 2 withdrew their participation', leaving 4 in sequence AB and 6 in sequence BA; 'All biological samples and electronic data records were collected as planned by November 16, 2023.' Design: two phases of 7-day baseline plus 4-week intervention separated by a 3-4 week washout, 3 daily beverages of allulose at 50 g/L (volume individualised to 5.1% of daily energy requirement) vs sweetness-matched aspartame at 113.75 mg/L, primary analysis by iAUC with a 2-tailed paired t test. The authors bound their own generalisability: 'While sufficiently powered for the primary outcome, the modest overall sample size (N=10) confines the interpretation of all secondary endpoints to hypothesis-generating findings and significantly limits the generalizability of the results.' DUPLICATE-POPULATION WARNING FOR THE FUTURE: the LisA results paper is expected June 2026 and will report these same 10 participants on this same outcome; it must SUPERSEDE this row, never count beside it - match on DRKS00028521. INDEPENDENCE (flag, not acted on, and for once it points the good way): funded by the German Research Foundation, 'project number 517694181', with no sweetener manufacturer among the funders and no supplier acknowledgement - so when its results land it would be one of very few manufacturer-independent human tests on a claim where 8 of 9 animal rows are supplier-linked (Matsutani, Anderson Advanced Ingredients, Samyang). group_id 'busch-protocol' is correct and shares no authors with the Hokkaido/Kyoto/Matsutani network. |
| Medak KD, Weber AJ, Shamshoum H, McKie GL, Hahn MK, Wright DC 2023 · Frontiers in Pharmacology | animal | supports | low | [FT-verified, PMC JATS full text; source note said provenance: abstract, and the prior extract read this paper as INDIRECT pathway support - it is not, the paper measured the claim head-on.] Medak 2023, Wright lab, University of Guelph. THIS ROW'S OWN DENOMINATOR IS EXPERIMENT 1 ONLY: male C57BL/6J mice, n=6-11/group, allulose 1 g/kg oral gavage, serum at a SINGLE terminal 120-min timepoint. There, 'there were main effects of both allulose and olanzapine to increase total GLP1 (Mean +/- SD: Vehicle-Control = 13.72 +/- 5.77, Olanzapine-Control = 30.26 +/- 10.86, Vehicle-Allulose = 22.63 +/- 5.89, Olanzapine-Allulose = 40.43 +/- 17.09)'. FRACTION-SPECIFIC: TOTAL GLP-1 only - the active-GLP-1 main effect at 120 min was olanzapine's, not allulose's (vehicle 7.86 to vehicle-allulose 8.48); most likely compartment and timing rather than contradiction, since siblings s29317623 and s29402406 sampled PORTAL vein with a time course, and the authors note 'It could be that peak increases in circulating active GLP1 were missed in the current study'. The authors call the rise 'minor' and 'marginal'. WHAT THE PRIOR EXTRACT BORROWED FROM THE WRONG ARM: allulose ALONE dampened nothing - 'The minor increase in serum GLP1 did not impact olanzapine induced increases in blood glucose'; the dampening in the title and abstract belongs to a three-drug cocktail (allulose + sitagliptin + SSTR5 antagonist) given WITH an oral glucose load (n=4-6/group), and is logged as a candidate finding, not as this claim's evidence. Quality stays LOW on its merits: one terminal timepoint, no time course, no dose-response, SDs of the same order as the means, males only by design, no randomisation or blinding described (cannot-tell), and GLP-1 here is the manipulation check rather than the question - the same shape sibling s35956407 is graded low for. INDEPENDENCE FLAG (not acted on, and for once it points the good way): CIHR Project Grant PJT 159538 and NSERC, COI states 'the research was conducted in the absence of any commercial or financial relationships', allulose purchased from Cedarlane - so unlike siblings s29317623, s29402406, s35561584 and s35956407 this row carries no Matsutani funding or Matsutani-supplied allulose. Only Wright-lab source in the vault; group_id medak-olanzapine is correct. |
| Yermek R, Wang L, Kaneko K, Han W, Seino Y, Yabe D, Yada T 2022 · Biochem Biophys Res Commun | animal | supports | moderate | OFF-SCOPE, zero weight — true about the paper, filed against the wrong object. The claim's object is GLP-1 itself; this study never measured GLP-1 in any compartment. Its own statement of what it did: 'This study explored the effect of central injection of D-Allulose on feeding behavior in mice. We also examined direct effects of D-Allulose on the neurons in the hypothalamic arcuate nucleus (ARC) that regulate feeding, including the anorexigenic glucagon-like peptide-1 (GLP-1)-responsive neurons and proopiomelanocortin (POMC) neurons. Single neurons were isolated from ARC and cytosolic Ca2+ concentration ([Ca2+]i) was measured by fura-2 microfluorometry.' The GLP-1 sentence the prior extract rested on is about allulose potentiating GLP-1's ACTION on neurons, not about GLP-1 release: 'D-Allulose potentiated the GLP-1 action to increase [Ca2+]i in ARC neurons including POMC neurons' — GLP-1 here is the applied agonist, so the finding is logically independent of whether allulose raises GLP-1 levels. Findings, with their own denominators named: allulose 5.6/16.7/56 mM 'concentration-dependently increased [Ca2+]i in ARC neurons', not an osmotic artefact ('The [Ca2+]i increases took place similarly when the osmolarity of superfusion solution was kept constant'); '33% of POMC neurons in ARC' responded (denominator = POMC neurons, not all ARC neurons); the 40% co-response figure is over D-ALLULOSE-RESPONSIVE neurons only — 'The majority (40%) of the D-Allulose-responsive neurons also responded to GLP-1 with [Ca2+]i increases' (the word 'majority' contradicts the 40%; recorded as the authors wrote it); and in vivo, 'Intracerebroventricular injection of D-Allulose significantly decreased food intake at 1 and 2 h after injection.' Abstract-only: no neuron counts, no mice per group, no effect sizes, no randomisation or blinding described — cannot-tell, not absent. COI: 'T.Y. received grant support from Matsutani Chemical Industry Co. Matsutani Chemical Industry Co. provided D-Allulose', the same sponsor that funded and supplied the other animal rows on this claim. Grade and quality left unchanged as moot at zero weight. Consistent with ev-s32193016, the sibling already off-scoped on this claim for the same wrong-outcome shape. The ICV feeding, POMC-activation and GLP-1-potentiation results are logged as candidate claims in the re-read record. |
| Suzuki T et al 2022 · Nutrients | animal | supports | low | [FT-verified, PMC JATS full text; source note was provenance: abstract] Suzuki 2022, Nutrients, three animal experiments. GLP-1 BEARS ON THIS CLAIM FROM ONE EXPERIMENT ONLY, with its own denominator: normal male ICR mice, overnight-fasted, PORTAL-vein plasma total GLP-1 at a single 30-min timepoint after oral gavage, n=6 per group vs saline (Figure 8B). There, 'GLP-2 and GLP-1 were significantly induced by oral administration of both fructose and allulose, but glucose administration was not induced by these hormones' (ANOVA + Dunnett vs saline). GLP-1 was NOT measured in the TPN rats - the prior extract's 'monitored GLP-1 ... in TPN rats' is wrong: those peripheral-vein samples (n=4/group) were assayed for GLP-2 and IGF-2 only (Figure 8A), and the 'paralleled GLP-2' finding is a different hormone from this claim's object. Dose is internally inconsistent (Methods 15 mmol/kgBW, Results 30 mmol/kgBW); total, not active, GLP-1; no time course, no dose-response, no randomisation or blinding described; the transcriptome/TPN arms measured no GLP-1 at all. COI, stated the unusual way round: 'The funders excluding Matsutani Chemical Industry Co., Ltd. had no role in the design of the study; in the collection, analysis, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results' - i.e. the allulose manufacturer, which also supplied the allulose under an MTA, is the one funder NOT excluded. Supports the claim as a single-timepoint confirmation of an effect the paper itself set out to confirm from prior reports; quality low, below the FT-verified dose-response siblings s29317623 and s29402406. INDEPENDENCE FLAG (not acted on): Matsutani funded and supplied the allulose for this row and for s29317623, s29402406 and s35561584 - four group_ids, one sponsor. |
| Cayabyab KB, ... Norwitz NG, Arroyo JA, Reynolds PR, Bikman BT 2024 · Nutrients | animal | supports | low | [FT-verified, PMC JATS full text (PMC11207032); source note said provenance: abstract] Cayabyab 2024, Bikman lab, Brigham Young University. A CHRONIC, FASTING-STATE row, not an acute secretion test: 12-week 2x2 trial in healthy 12-week-old Wistar rats randomised to standard or Western diet with either allulose or stevia in the DRINKING WATER (3%, ~1.9 g/kg/day; n=10/group, 5F/5M), with 8-h-fasted tail-vein ACTIVE GLP-1 sampled every four weeks. THIS ROW'S OWN DENOMINATOR IS n=8 PER GROUP, not the 10 randomised: 'GLP-1 was significantly elevated throughout the trial in both groups consuming allulose ((F); n = 8).' The rise was NOT conditional on the obese/diabetic state the abstract frames - 'the higher GLP-1 levels were observed in both allulose groups, but the benefit on weight and metabolic parameters was only observed in the WD+allulose group' - which decouples the GLP-1 effect from the metabolic benefit and, on this claim, strengthens it. The authors claim novelty for a lingering fasted rise: 'we believe our results are the first to report an increased GLP-1 in a state' (a word is missing in the published sentence; Methods make it the 8-h-fasted state). WHAT THE PRIOR EXTRACT GOT WRONG: it led with 'lowered glucose', which is off-object here and is also not what happened - allulose PREVENTED, in the Western-diet arm only ('allulose also blocked the development of hyperinsulinemia ... hyperglycemia'); the rats were not diet-induced-obese/T2D at randomisation but got allulose from day one; and 'associated with enhanced incretin/GLP-1 signaling' softened a direct serum measurement into an inference. QUALITY moderate -> LOW on reliability: active GLP-1 was read semi-quantitatively off a custom rodent antibody array by densitometry ('Signal intensities were compared to positive controls included on each membrane'); NO mean, SEM, p-value or concentration for GLP-1 appears anywhere in the text or legends, so no magnitude is recoverable; single dose, no dose-response; statistics are 'ANOVA, followed by student t tests. Results are representative' across ~20 outcomes with no multiplicity control; blinding not described (cannot-tell). The comparator is stevia with no water arm and the authors flag the confound on this very object - 'while stevia has been shown to stimulate GLP-1 secretion in cell culture, we know of no such evidence in rodents or humans' and 'the lack of a water-only group should be noted' (it would bias toward the null, not toward the claim). Held above the floor by 12-week repeated measurement and significance in two independent allulose arms; below the dose-response siblings s29317623 (high) and s29402406 (moderate), level with s35956407 and s36937886. SPONSOR FLAG (recorded, not acted on): 'This research was funded by a sponsored research agreement awarded to Brigham Young University by Anderson Advanced Ingredients' - the same firm that supplied the allSWEET allulose; 8 of 13 authors advise Nutrishus Brands, and the author block (Norwitz, D'Agostino, Diamond, Koutnik, Johnson) overlaps the vault's lmhr-network group_id, though no lmhr-network source votes on any allulose claim. A FIFTH supplier-funded animal row on this claim under a fifth group_id - a different sponsor from the Matsutani cluster, so no merge. Of the claim's nine animal rows, five are now known supplier-linked and only s36937886 is known manufacturer-independent; three others have not been funding-checked. |
| Lee 2025 · Antioxid Redox Signal | in-vitro | supports | moderate | OFF-SCOPE, zero weight — receptor, not hormone. The claim's object is GLP-1 itself; this paper measures GLP-1R. Its own framing: 'D-allulose is predicted to modulate GLP-1R via mechanisms linked to endoplasmic reticulum stress and reactive oxygen species (ROS) pathways, positively influencing GLP-1R stability and functionality.' Its mechanistic result names the receptor transcript as the substrate: 'D-allulose effectively regulated adipocyte differentiation by inhibiting the NADP+/NADPH-ROS-inositol-requiring enzyme 1a (IRE1a)-regulated IRE1-dependent decay (RIDD) axis, resulting in controlled decay of GLP-1R, a newly identified RIDD target.' Its in-vivo test is a NECESSITY test of the receptor, not a measurement of the ligand: 'However, these effects were not observed in GLP-1R KO mice, suggesting that the antiobesity effects of D-allulose rely on the presence of GLP-1R' — logically independent of whether allulose raises GLP-1 levels, exactly as sibling ev-s35561584 (GLP-1 applied as reagent) was off-scoped for. No plasma, portal or intestinal GLP-1 concentration is reported. The prior extract, 'D-allulose modulated the GLP-1 receptor pathway ... to regulate obesity', is accurate about the paper and states the receptor plainly; it was simply filed against a secretion claim. ABSENCE IS NOT A FINDING: full text is unobtainable (no PMCID; Europe PMC isOpenAccess:N), so whether a GLP-1 assay exists in the Results is cannot-tell — but nothing in the recorded provenance can carry a vote on hormone release. Grade (in-vitro) and quality (moderate) left unchanged as moot at zero weight; the paper is in fact in-vitro PLUS a 12-week in-vivo mouse arm, and the sibling row on this same paper grades it animal. INDEPENDENCE (flag, not acted on): two authors are at '(5)Food Biotech R&D Center, Samyang Corp., Seongnam-si, Republic of Korea.', the allulose manufacturer already flagged for COI on ev-s42280385 — a second supplier-linked row on this claim under a different group_id, alongside the Matsutani cluster catalogued in the sibling extracts. |
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.