- "Stevia" is not one substance. Whole-leaf/green powder is not FDA-GRAS and is subject to import detention, while refined white steviol glycoside extract (≥95% purity) is GRAS and dominates the market (FDA Import Alert 45-06).
- The best independent synthesis — a 9-RCT, 756-participant Oxford meta-analysis — found stevioside produces modest but statistically significant reductions in fasting glucose and diastolic blood pressure, while Rebaudioside A shows no effect on either (Onakpoya & Heneghan 2015).
- A 12-week human RCT found stevia consumption did not significantly alter gut microbiome composition (PubMed 38257188), though a related industry-funded gut study is flagged as conflicted.
- The Acceptable Daily Intake, set jointly by EFSA and JECFA, is 4 mg/kg body weight/day of steviol equivalents — a limit regulators say could be exceeded by high-consuming children (EFSA).
- Native to Paraguay and Brazil, global stevia supply is now overwhelmingly grown and processed in China, with Reb M/Reb D increasingly made by fermentation rather than leaf extraction (Dialogue Earth; FDA GRAS Notice 882).
- Overall evidence grade: Moderate for purified steviol glycosides (≥95%); Insufficient for whole-leaf/crude stevia.
Table of contents
- Evidence summary
- What stevia is
- All forms and grades
- How it works
- Benefits by claim
- What works and what does not
- Risks and all side effects
- All interactions
- Who should avoid stevia
- Dosage and how to take
- Animal and in-vitro evidence excluded
- Independent funding and conflict notes
- Frequently asked questions
- Sources and funding notes
| Claim | Evidence | Source | Funding/conflict | Strength |
|---|---|---|---|---|
| Stevioside lowers fasting blood glucose | 9-RCT meta-analysis, 756 participants; small but statistically significant reduction | Onakpoya & Heneghan 2015 | Independent (Oxford academic, no industry funding disclosed) | Moderate |
| Stevioside lowers diastolic blood pressure | Same meta-analysis; systolic effect not significant (MD −2.98 mmHg, 95% CI −6.23 to 0.27) | Onakpoya & Heneghan 2015 | Independent | Moderate |
| Reb A affects blood pressure or cardiometabolic risk | No significant effect found | Onakpoya & Heneghan 2015 | Independent | Weak (null finding) |
| Stevia vs. sucralose in type 2 diabetes (8-week RCT, n=34) | No significant difference in fasting glucose, HbA1c, insulin, or lipids | PMC7103435 | Not disclosed as industry-funded in source | Moderate |
| Stevia alters gut microbiome composition over 12 weeks | No significant alteration found | PubMed 38257188 | Independent academic RCT | Moderate |
| Whole-leaf/crude stevia is safe for regular human consumption | No FDA/EFSA-grade toxicological dossier exists | FDA Import Alert 45-06 | Regulatory determination | Insufficient |
| ADI of 4 mg/kg bw/day steviol equivalents is protective for all age groups | EFSA flags possible exceedance in high-consuming children | EFSA | Independent regulator | Moderate |
What stevia is
Stevia is a sweetener derived from the leaves of Stevia rebaudiana, a shrub native to Paraguay and Brazil with a long history of traditional use in the region (Wikipedia: Stevia cultivation in Paraguay; Public Eye report on Paraguay stevia). The sweetness comes from a family of plant compounds called steviol glycosides, which are roughly 200–300 times sweeter than table sugar (FDA Import Alert 45-06; EFSA). Critically, "stevia" as sold today spans two very different regulatory categories: unrefined whole-leaf material, which retains the plant's full phytochemical matrix and has never cleared FDA safety review, and highly purified extracts (≥95% steviol glycosides), which are the form actually approved as GRAS and used in almost all commercial food and beverage products. Despite common marketing language treating stevia as a single "natural" ingredient, the evidence base, safety clearance, and legal status differ sharply between these forms.
All forms and grades
Stevia reaches consumers in several distinct product forms with very different regulatory and evidentiary status. The table below summarizes the main forms found on the market.
| Form | Composition | Purity/grade | Regulatory status | Sweetness vs. sucrose |
|---|---|---|---|---|
| Whole-leaf / green stevia powder | Dried, ground leaves; full phytochemical matrix (steviol glycosides plus flavonoids, chlorophyll, other plant compounds) | Not purified; variable glycoside content | Not FDA-GRAS, not an approved food additive in the US; sold as a dietary supplement; approved as a food ingredient in Japan, Brazil, Australia, and New Zealand | Variable, generally lower/less consistent than refined extract |
| Refined white steviol glycoside extract | Purified stevioside, Reb A, Reb D, Reb M, or enzyme-modified glycosides | ≥95% steviol glycosides on a dried-weight basis (FDA's regulatory line) | GRAS via multiple FDA GRAS notices | ~200–300x |
| Liquid stevia drops/extract | Concentrated steviol glycoside solution, typically standardized to Reb A or Reb M | ≥95% glycoside basis typical | GRAS (as purified extract) | ~200–300x concentrate |
| Erythritol/maltodextrin stevia blends | Refined steviol glycosides combined with a bulking agent to mimic sugar's volume/mouthfeel (e.g., Truvia-style products) | Glycoside component ≥95% purity; bulking agent is majority of blend by weight | GRAS (glycoside component) | Diluted; formulated to approximate 1:1 sugar substitution by volume |
| Tabletop packets | Purified glycosides blended with maltodextrin or dextrose as carrier | ≥95% glycoside basis in active ingredient | GRAS | Formulated per-packet equivalence (usually ~1 packet = 2 tsp sugar) |
Glycosides: stevioside vs. Reb A vs. Reb M vs. Reb D
Stevia leaves contain a family of steviol glycosides that differ in sweetness intensity, bitterness, and natural abundance:
- Stevioside and Rebaudioside A (Reb A) are the most abundant glycosides in conventional stevia extract and the most studied in human trials (FDA Import Alert 45-06; EFSA).
- Rebaudioside M (Reb M) and Rebaudioside D (Reb D) are minor components of the native leaf, present in far smaller natural quantities, but valued for a cleaner taste with less licorice/bitter aftertaste. They are now produced at commercial scale via fermentation using engineered yeast (Yarrowia lipolytica), which the FDA has confirmed is chemically identical to leaf-extracted Reb M (FDA GRAS Notice 882).
- The FDA's regulatory line is purity-based, not glycoside-specific. Highly purified steviol glycosides (≥95% on a dried-weight basis) — whether stevioside, Reb A, Reb D, Reb M, or enzyme-modified glycosides — are the substances that have cleared GRAS review via multiple notices (GRAS Notice 702, GRAS Notice 790, GRAS Notice 838). Whole leaf and "crude" (non-95%-pure) extracts remain not GRAS and are subject to FDA import detention (FDA Import Alert 45-06).
How it works
In humans, steviol glycosides are not absorbed intact in the small intestine; they pass to the colon where gut bacteria hydrolyze them to steviol, which is absorbed, metabolized in the liver, and excreted primarily as steviol glucuronide in urine. The proposed mechanism behind stevioside's modest glucose- and blood-pressure-lowering signal involves stimulation of insulin secretion and vasodilatory effects, though the exact human pathway is not fully characterized by RCT-level mechanistic data. One IN-VITRO (non-human) evidence data point offers plausible mechanistic color: experiments using human intestinal enteroendocrine cell lines (STC-1 and HuTu-80) found that Rebaudioside A stimulates GLP-1 — a gut hormone involved in insulin secretion and satiety — release in a concentration-dependent manner (Food & Function, RSC). This finding is limited: the same study also incorporated mouse organoid and pig intestinal segment data, which is excluded here, and even the retained human cell-line portion is a reductionist single-cell-type model that cannot capture whole-body hormonal, neural, or systemic feedback present in actual human physiology. It should be read only as a plausible mechanism, not as proof of a clinical effect — particularly since the same meta-analysis that established stevioside's glucose effect found no comparable effect from Reb A on blood pressure or cardiovascular risk factors (Onakpoya & Heneghan 2015).
Benefits by claim
Blood glucose and insulin
The best available synthesis is a systematic review and meta-analysis of randomized clinical trials by Onakpoya and Heneghan (Oxford University, European Journal of Preventive Cardiology, 2015), pooling 9 RCTs with 756 participants. It found stevioside produced small but statistically significant reductions in fasting blood glucose and diastolic blood pressure, while Rebaudioside A showed no significant effect on blood pressure or cardiovascular risk factors; heterogeneity across trials was substantial, limiting confidence (Onakpoya & Heneghan 2015). This is the single most important finding in the stevia literature: the glycoside matters. Stevioside-dominant products show a real, if modest, glucose and blood-pressure signal; Reb A-dominant products (the most common commercial glycoside) do not.
A separate double-blind RCT in 34 people with type 2 diabetes compared stevia-sweetened tea to sucralose over 8 weeks and found no significant differences in fasting blood sugar, HbA1c, insulin, or lipids between groups — i.e., stevia was glycemically neutral, neither harmful nor beneficial versus sucralose (PMC7103435). A related meta-analysis, "Effect of Steviol Glycosides on Human Health with Emphasis on Type 2 Diabetic Biomarkers," similarly found modest, inconsistent effects on glycemic biomarkers (PMC6770957). Dose used in the Oxford-pooled trials generally ranged in the range studied by the original stevioside RCTs, several of which used doses in the range of ~250–1500 mg/day stevioside; readers should note the pooled analysis itself flags substantial heterogeneity, so precise dose-response is not firmly established.
Blood pressure
The Oxford meta-analysis found a significant reduction in diastolic blood pressure with stevioside, though the systolic effect was not statistically significant (mean difference −2.98 mmHg, 95% CI −6.23 to 0.27), and Reb A specifically showed no blood-pressure effect (Onakpoya & Heneghan 2015). The authors are affiliated with Oxford's Nuffield Department of Primary Care Health Sciences (Centre for Evidence-Based Medicine); no stevia-industry funding is disclosed. Independence: Independent. Credibility: Strong (transparent Cochrane-style methodology, academic affiliation, no product funding found).
Gut microbiome
A 12-week human RCT titled "Consumption of the Non-Nutritive Sweetener Stevia for 12 Weeks" found no significant alteration of gut microbiota composition in humans (PubMed 38257188; reported by News-Medical). A companion human comparison study, "Comparison of a Daily Steviol Glycoside Beverage" versus other options, reached similar conclusions of minimal microbiome disruption (PubMed 38408729). A widely cited "impact on gut microbiome" study by Cargill-funded researchers compared human and primate (Cebus apella) data (ACS) — the primate/animal-comparison portion is excluded from this review entirely; only the human-subject data from that line of research is retained, and the Cargill funding source is flagged as a conflicted source requiring independent replication. Overall gut-microbiome evidence grade: Moderate — consistent human data showing minimal disruption, but one supporting study line is funded partly by an ingredient manufacturer.
Weight management
A 12-week open-label RCT ("Daily Consumption of Stevia Drops") in healthy adults examined glycemia, body weight, and energy intake (ScienceDirect / PMC7257954). A randomized crossover trial from the EU Horizon 2020-funded SWEET consortium — a mixed public-private group including food-industry partners — compared neotame, Stevia Reb M, and sucrose-sweetened biscuits in overweight/obese adults over two weeks, examining postprandial appetite and endocrine response (The Lancet eBioMedicine). Because SWEET includes food and beverage industry partners, its findings are treated here as probably independent but flagged for partial industry funding, not fully independent evidence.
Taste and adverse effects in trials
Human RCTs consistently report a licorice-like aftertaste and bitterness as the most common complaint, more pronounced with stevioside and Reb A than with Reb M. The Oxford meta-analysis recorded adverse events including abdominal fullness, epigastric pain, and dizziness in a minority of participants across trials (Onakpoya & Heneghan 2015).
What works and what does not
| Claim | Verdict | Evidence basis |
|---|---|---|
| Stevioside modestly lowers fasting glucose | Supported, modest effect | 9-RCT/756-participant meta-analysis (Onakpoya & Heneghan 2015) |
| Stevioside modestly lowers diastolic blood pressure | Supported, modest effect; systolic not significant | Onakpoya & Heneghan 2015 |
| Reb A lowers blood pressure or improves cardiometabolic risk | Not supported | No significant effect found (Onakpoya & Heneghan 2015) |
| Stevia is glycemically superior to sucralose in type 2 diabetes | Not supported — no significant difference found | PMC7103435 |
| Stevia disrupts gut microbiome balance | Not supported over 12 weeks | PubMed 38257188 |
| Whole-leaf/green stevia powder is as safe as purified extract | Not established — lacks regulatory-grade human safety data | FDA Import Alert 45-06 |
| Fermentation-derived Reb M is chemically different from leaf-extracted Reb M | Not supported — FDA confirms chemical identity | FDA GRAS Notice 882 |
Risks and all side effects
| Side effect | Frequency/context | Source |
|---|---|---|
| Bitter/licorice-like aftertaste | Common, especially with stevioside and Reb A; less pronounced with Reb M | Onakpoya & Heneghan 2015 |
| Mild GI discomfort (abdominal fullness, epigastric pain, bloating) | Reported in a minority of RCT participants at higher doses | Onakpoya & Heneghan 2015 |
| Dizziness | Reported in a minority of RCT participants | Onakpoya & Heneghan 2015 |
| Excess intake beyond ADI, particularly in children | Regulatory concern, not an observed adverse-event rate | EFSA |
| Unknown long-term safety of whole-leaf/crude extract | Data gap — not GRAS, no equivalent human toxicology dossier | FDA Import Alert 45-06 |
No serious or life-threatening adverse effects have been documented in human RCTs of purified steviol glycosides at studied doses. The main established human safety concerns concentrate on the whole-leaf/crude form, which lacks the toxicological data package the FDA and EFSA require, and on the theoretical risk of exceeding the ADI.
All interactions
| Drug/substance class | Mechanism of concern | Severity/guidance | Evidence status |
|---|---|---|---|
| Antihypertensive medications | Stevioside's diastolic blood-pressure-lowering effect could theoretically be additive | Use with awareness; monitor blood pressure | Theoretical, extrapolated from RCT effect size; no dedicated interaction trial identified |
| Antidiabetic medications (insulin, metformin, sulfonylureas) | Stevioside's modest fasting-glucose-lowering effect could theoretically compound glucose-lowering drug effects | Use with awareness; monitor glucose | Theoretical, extrapolated from RCT effect size; no dedicated interaction trial identified |
| Anticoagulants/antiplatelets, antidepressants, sedatives, thyroid medication, statins, PPIs, oral contraceptives, antibiotics, antiepileptics, immunosuppressants | No documented mechanism identified in the reviewed literature | No specific guidance available | Data gap |
Who should avoid stevia
- Anyone relying on whole-leaf or crude ("green") stevia powder as a daily sweetener, given the absence of FDA/EFSA-grade human safety data for that form (FDA Import Alert 45-06).
- People on blood pressure-lowering or glucose-lowering medications using stevioside-dominant products in large amounts, until dedicated interaction data exists — monitor for additive effects rather than assuming none.
- Children and other high-consumers of multiple stevia-sweetened products, who EFSA notes could plausibly exceed the 4 mg/kg bw/day ADI at proposed maximum use levels (EFSA).
- Pregnant or breastfeeding people considering whole-leaf/crude stevia specifically — this form lacks dedicated human pregnancy-safety data and should be avoided under the precautionary principle applied to any non-GRAS ingredient; purified, GRAS steviol glycosides within ADI limits have not been flagged as a pregnancy-specific concern by FDA or EFSA (PMC4229159).
Dosage and how to take
| Parameter | Value | Source |
|---|---|---|
| Acceptable Daily Intake (ADI) | 4 mg/kg body weight/day, expressed as steviol equivalents | EFSA; JECFA database |
| Regulatory basis | Harmonized between EFSA and JECFA; applies to purified (≥95%) steviol glycosides | EFSA |
| Population at risk of exceeding ADI | High consumers, particularly children, at proposed maximum use levels | EFSA |
| Doses studied for glucose/blood-pressure effect (stevioside) | Varied across the 9 pooled RCTs; substantial heterogeneity noted by authors | Onakpoya & Heneghan 2015 |
| Whole-leaf/crude form dosing guidance | None established — not GRAS, no regulatory dosing framework | FDA Import Alert 45-06 |
Animal and in-vitro evidence excluded
This review relies on independent human-trial evidence only. The following animal studies surfaced during research and are explicitly excluded from all safety and efficacy conclusions:
- Excluded — animal study: A rodent toxicology study claiming stevia and sucralose cause liver enzyme elevation, hemoglobin reduction, or "hidden hazardous effects" is a mouse study and cannot be used to support any human harm claim (PMC7584803).
- Excluded — animal/primate comparison: The primate (Cebus apella) portion of a Cargill-funded gut-microbiome study is excluded; only the human-subject data from that research line is retained above, and even that is flagged as industry-funded (ACS).
One piece of IN-VITRO (non-human) evidence is used in this article, in the "How it works" section above: human enteroendocrine cell-line data on Reb A and GLP-1 release (Food & Function, RSC), explicitly flagged there with its limitations. No other in-vitro or animal data is used to support any claim in this article.
Independent funding and conflict notes
| Source | Funding/affiliation | Independence rating |
|---|---|---|
| Onakpoya & Heneghan 2015 | Oxford Nuffield Department of Primary Care Health Sciences (Centre for Evidence-Based Medicine); no stevia-industry funding disclosed | Independent |
| PubMed 38257188 (12-week gut microbiome RCT) | Academic human RCT | Independent |
| ACS gut microbiome study (human + primate) | Cargill-funded | Conflicted — human-only portion retained, flagged for independent replication |
| SWEET consortium crossover trial | EU Horizon 2020 public funding plus food-industry consortium partners | Probably independent, flagged for partial industry funding |
| EFSA steviol glycoside opinion | EU independent food safety regulator | Independent regulator |
| JECFA database | WHO/FAO joint expert committee | Independent regulator |
| FDA Import Alert 45-06 | US federal regulator | Independent regulator |
Frequently asked questions
Is stevia the same thing regardless of which product I buy?
No. Whole-leaf/green stevia powder and refined white steviol glycoside extract (≥95% purity) are regulated completely differently in the US — only the purified extract is GRAS, while whole-leaf material is not approved as a food additive and is subject to FDA import detention (FDA Import Alert 45-06).
Does stevia lower blood sugar?
Stevioside-dominant stevia shows a small but statistically significant reduction in fasting blood glucose across a 9-RCT meta-analysis, while Reb A-dominant stevia — the most common commercial form — shows no such effect (Onakpoya & Heneghan 2015). In a head-to-head RCT against sucralose in people with type 2 diabetes, stevia showed no significant glycemic advantage (PMC7103435).
Does stevia harm gut bacteria?
A 12-week human RCT found no significant alteration of gut microbiota composition from stevia consumption (PubMed 38257188). This is reassuring but one supporting study in this research area was industry-funded and is flagged as conflicted rather than independent.
What is the difference between stevioside, Reb A, Reb M, and Reb D?
These are all steviol glycosides found in the stevia leaf, differing mainly in sweetness intensity, bitterness, and natural abundance. Stevioside and Reb A are the most abundant and most-studied for glucose/blood pressure effects; Reb M and Reb D are minor leaf components valued for cleaner taste and are increasingly produced by fermentation rather than extraction (FDA GRAS Notice 882). The FDA's GRAS approval line is based on purity (≥95%), not on which specific glycoside is used (FDA Import Alert 45-06).
How much stevia is safe to consume daily?
EFSA and JECFA have harmonized on an ADI of 4 mg/kg body weight/day, expressed as steviol equivalents. EFSA has specifically noted this limit could be exceeded by high consumers, particularly children, at proposed maximum use levels (EFSA).
Where does commercial stevia actually come from?
Stevia rebaudiana is native to Paraguay and Brazil, but the overwhelming majority of commercial stevia leaf and extract today is cultivated and processed in China, which dominates global supply (Dialogue Earth). Paraguay's traditional growers have seen commercialization increasingly captured by large multinational buyers rather than local producers (Public Eye report on Paraguay stevia).
Sources and funding notes
- Onakpoya & Heneghan 2015, meta-analysis of 9 RCTs on steviol glycosides, European Journal of Preventive Cardiology — independent, Oxford academic.
- PMC7103435, RCT comparing stevia and sucralose in type 2 diabetes.
- PMC6770957, meta-analysis on steviol glycosides and type 2 diabetes biomarkers.
- PubMed 38257188, 12-week human RCT on stevia and gut microbiota.
- News-Medical coverage of the 12-week gut microbiota RCT.
- PubMed 38408729, companion steviol glycoside beverage comparison study.
- ACS, Cargill-funded human/primate gut microbiome study (primate portion excluded).
- ScienceDirect, 12-week RCT on daily stevia drop consumption / PMC7257954.
- The Lancet eBioMedicine, SWEET consortium crossover RCT — mixed public/industry funding.
- FDA Import Alert 45-06, stevia leaf and crude extract.
- FDA GRAS Notice 702, purified steviol glycosides.
- FDA GRAS Notice 790.
- FDA GRAS Notice 838.
- FDA GRAS Notice 882, fermentation-derived Reb M.
- EFSA, steviol glycosides safety opinion and ADI — independent regulator.
- JECFA (WHO/FAO) database entry for steviol glycosides — independent regulator.
- PMC4229159, review of sugar substitutes in pregnancy.
- Food & Function (RSC), human enteroendocrine cell-line study on Reb A and GLP-1 — IN-VITRO (non-human) evidence, flagged.
- PMC7584803, rodent toxicology study — excluded animal study.
- Wikipedia: Stevia cultivation in Paraguay.
- Public Eye report on Paraguay stevia and biopiracy.
- Dialogue Earth, commercialization of Paraguay's stevia and China's dominant supply role.
Last reviewed: July 4, 2026.