Supplements · Sweeteners · Metabolic & Cardiometabolic
stevia improves glycemic control in diabetes
In plain terms: Does stevia help control blood sugar if you have diabetes?
Part of: • Stevia
Not in any lasting way that's been proven. There's a real short-term signal (one high-dose meal study blunted the glucose spike) and supportive animal data, but the longer human trials and a Cochrane review find no reliable improvement in HbA1c or fasting glucose. Stevia is fine as a sugar substitute, but it isn't a blood-sugar treatment.
📅 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: All trials, pooled (Meta-analysis)
How the studies fall
What the evidence shows
Beyond simply not raising blood sugar, the stronger claim that stevia actively improves diabetes control is not supported by durable human evidence.
The evidence (11)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Nichol AD, Holle MJ, An R 2018 · Eur J Clin Nutr | meta-analysis | tested-null | moderate | Meta (NNS class incl. stevia): no significant glycemic impact in most comparisons; stevia underrepresented. |
| Kang L, Zhang C, Wang R, Li K, Bai X, Qi N, Qu H, Li G 2025 · Nutrients | animal | supports | low | Preclinical SR (40 animal + 5 cell studies): steviol glycosides improve glucose metabolism; authors state CLINICAL confirmation still needed. |
| Barriocanal LA, Palacios M, Benitez G, Benitez S, Jimenez JT, Jimenez N, Rojas V 2008 · Regul Toxicol Pharmacol | RCT | contradicts | moderate | Randomised double-blind placebo-controlled pilot, 3 months, stevioside 250 mg three times daily vs placebo in three groups (type 1 diabetes, type 2 diabetes, non-diabetic normo/low-BP); group sizes not given in the abstract. 'Post-treatment systolic BP, diastolic BP, glucose and glycated hemoglobin (HbA1c) were not significantly different from baseline measurements'; authors conclude steviol glycosides 'have no pharmacological effect'. Clean null on the directional claim (no point estimates reported, no trend toward benefit described) -> contradicts per CONVENTIONS §5. Comparisons are reported against baseline, not between arms; pilot scale. |
| Anker CCB, Rafiq S, Jeppesen PB 2019 · Nutrients | meta-analysis | tested-null | low | Claim is about glycemic control IN DIABETES, so the diabetic subgroups decide the row. Diabetic subgroup: HbA1c MD -0.30% (95% CI -0.89 to 0.29, p=0.31) and fasting glucose MD -44.01 mg/dL (95% CI -120.41 to 32.38, p=0.26, I2=89%) - both point toward improvement with intervals including zero: underpowered trend, so tested-null per CONVENTIONS §5 second clarification. Whole-population pools are flat (HbA1c 3 studies/5 RCTs, n=127: 0.00%, CI -0.24 to 0.25, p=0.98; fasting glucose 6 studies/8 RCTs, n=438: -2.63 mg/dL, CI -7.77 to 2.51, p=0.32). The authors' remark that 'only two studies concerning type 2 diabetics were included' sits in their triglyceride discussion, not the glucose results. No post-meal glucose outcome was measured. |
| Stamataki NS, Crooks B, Ahmed A, McLaughlin JT 2020 · Nutrients | RCT | contradicts | low | 12-week randomised open-label trial in 28 healthy, normal-weight adults (14 stevia, 14 no-treatment control): daily stevia extract produced no significant change in the glucose or insulin response to an oral glucose tolerance test. Glucose iAUC in the stevia group was 132 mmol/L x min at baseline and 133 at week 12 (control 131 then 159); peak glucose went 7.08 to 6.91 mmol/L on stevia. Every participant was normoglycaemic at entry, so the trial tests whether stevia improves glucose handling in healthy people - it says nothing about managing established diabetes. |
| Gregersen S, Jeppesen PB, Holst JJ, Hermansen K 2004 · Metabolism | RCT | supports | moderate | Acute cross-over trial in 12 people with type 2 diabetes: one test meal with 1 g of stevioside versus the same meal with 1 g of maize starch. Stevioside cut the incremental glucose area under the curve by 18% (P=.013) and raised the insulinogenic index - insulin AUC divided by glucose AUC - by about 40% (P<.001). But the insulin area itself did not change significantly, so the index rose mainly because glucose fell, not because more insulin was released. Glucagon tended to fall. HbA1c and longer-term glucose control were not measured. |
| Chen J, Jeppesen PB, Abudula R, Dyrskog SE, Colombo M, Hermansen K 2006 · APMIS | in-vitro | supports | low | Isolated islets: stevioside enhanced glucose-stimulated insulin secretion without desensitizing beta-cells - mechanism only. |
| Concha Celume F, Pérez-Bravo F, Magne F, Olivares R, Gotteland M 2026 · Frontiers in nutrition | animal | tested-null | low | Mice: parental stevia (0.1mg/mL, 16wk) vs water; F0 OGTT unchanged; F1 male Stevia mildly altered gene expr (Tlr4/Tnf up), normalized by F2; no direct glycemic-control benefit shown |
| Lohner S, Kuellenberg de Gaudry D, Toews I, Ferenci T, Meerpohl JJ 2020 · Cochrane Database Syst Rev | meta-analysis | contradicts | high | Cochrane review (NNS in diabetes): insufficient / low-certainty evidence that NNS incl. stevia improves glycemic outcomes. |
| Abudula R, Jeppesen PB, Rolfsen SE, Xiao J, Hermansen K 2004 · Metabolism | in-vitro | supports | low | Isolated islets: RebA dose-dependently stimulated insulin secretion (glucose/calcium-dependent) - mechanism only. |
| Jeppesen PB, Gregersen S, Rolfsen SE, Jepsen M, Colombo M, Agger A, Xiao J, Kruhøffer M, Orntoft T, Hermansen K 2003 · Metabolism | animal | supports | low | Diabetic rats: stevioside improved glucose tolerance and first-phase insulin - mechanism only. |
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.