Benefits
Blood glucose and insulin after an oral dose
In 5 healthy men, an oral dose of 0.3 g per kg of body weight did not raise serum glucose or insulin, while the same dose of glucose raised both within 30 minutes. About 90% was recovered unchanged in urine within 48 hours, so the authors estimated it supplies less than 0.4 kcal per gram. The study was run at a company research lab.
Glucose tolerance after 5 weeks of daily intake
In a randomized trial of 42 adults with obesity, 36 g of erythritol a day for 5 weeks did not change glucose tolerance, arterial stiffness, abdominal fat, uric acid or liver enzymes compared with xylitol or no sweetener. Digestive tolerance was good apart from a few diarrhea-related symptoms.
Blood erythritol levels and cardiovascular events
In people having heart evaluations in the US (2,149) and Europe (833), those in the top quarter of blood erythritol had about 1.8 and 2.2 times the 3-year risk of heart attack, stroke or death of those in the bottom quarter. This is an association, not proof of cause, and the body also makes erythritol. The same paper found it raised platelet reactivity in lab tests.
Platelet reactivity after a 30 g erythritol drink
In 20 healthy volunteers, a drink with 30 g of erythritol raised blood erythritol more than 1,000-fold and increased platelet responses in every participant, while a 30 g glucose drink did not significantly increase release of the platelet markers measured. In an earlier 8-person test, blood levels stayed raised for more than 2 days. These were lab markers, not clots or heart events.
Regulatory status in the US and EU
In the US, erythritol is used in foods under a GRAS notice that FDA answered with no questions; that is a food safety notice, not a drug approval. EFSA set an acceptable daily intake of 0.5 g per kg of body weight based on diarrhea and found no proof of a causal link with heart disease. The WHO guideline on non-sugar sweeteners does not apply to sugar alcohols.
Digestive tolerance compared with xylitol
In a double-blind study of 70 young adults (64 finished), single doses of 20 and 35 g of erythritol in water caused no significant increase in digestive symptoms compared with 45 g of sugar. At 50 g, erythritol increased nausea and stomach rumbling, while 35 and 50 g of xylitol caused watery stools and more symptoms.
Polyol candies and children's teeth in school trials
In a 3-year double-blind trial in 485 Estonian schoolchildren, about 7.5 g a day of erythritol from candies was followed by fewer teeth with dentin lesions than xylitol or sorbitol candies. A 4-year Finnish school trial in 579 children found no added benefit from erythritol or xylitol lozenges where preventive dental care was already in place.
Mechanism of action
Absorbed in the small intestine and excreted unchanged
EFSA describes erythritol as readily absorbed in a dose-dependent way, with a small fraction converted to erythronate and the rest excreted unchanged in urine. In healthy men, about 90% of an oral dose was recovered in urine within 48 hours. Because so little is broken down, it supplies almost no usable energy and did not raise blood glucose or insulin.
Also made by the body through the pentose phosphate pathway
Human cells can make erythritol from glucose through the pentose phosphate pathway, so blood erythritol reflects both diet and the body's own production. The authors of the large cardiovascular study said levels in their cohorts likely came from both sources. A 30 g serving raises blood levels more than 1,000-fold, far above the usual range.
Platelet reactivity at high blood levels
In laboratory tests, erythritol at levels seen in people increased platelet responses to normal clotting signals, and in mice it sped clot formation. In healthy volunteers, a 30 g drink increased platelet aggregation and the release of platelet activation markers. Whether this leads to clots or heart events in people has not been tested.
Clinical trials
Observational cohort analysis with laboratory platelet tests, mouse clotting experiments and an 8-person feeding test; funded by NIH, the Office of Dietary Supplements and other grants. (Witkowski et al. 2023, Nat Med)
Stable patients having elective heart evaluations: 1,157 in a discovery cohort, 2,149 in a US validation cohort and 833 in a European validation cohort; 8 healthy volunteers in the feeding test.
Top versus bottom quarter of plasma erythritol: adjusted hazard ratio for death, heart attack or stroke over 3 years of 1.80 (95% CI 1.18 to 2.77) in the US cohort and 2.21 (1.20 to 4.07) in the European cohort. Erythritol increased platelet reactivity in lab tests and clot formation in mice. After 30 g of erythritol, blood levels in 8 volunteers stayed raised for more than 2 days. This shows an association, not cause, and blood levels also reflect the body's own production.
Prospective, non-randomized interventional study comparing a 30 g erythritol drink with a 30 g glucose drink, run by the group that reported the cardiovascular association. (Witkowski et al. 2024, Arterioscler Thromb Vasc Biol)
20 healthy volunteers averaging about 30 years of age, 10 per drink group.
Erythritol raised plasma erythritol more than 1,000-fold (median 6,480 micromol/L 30 minutes after the drink versus about 3.8 micromol/L before it) and enhanced stimulus-induced platelet aggregation in every participant, with greater release of the platelet markers serotonin and CXCL4. The comparison drink did not significantly increase release of these markers. Only lab markers were measured, not clots or clinical events.
Single-dose comparison with an equal dose of glucose, carried out at a Nikken Chemicals research lab. (Noda et al. 1994, Eur J Clin Nutr)
5 healthy male volunteers aged 45 to 58 given 0.3 g per kg of body weight of erythritol or glucose.
Erythritol did not raise serum glucose or insulin, while glucose raised both within 30 minutes. Cholesterol, triglycerides, free fatty acids and electrolytes did not change. About 86% of the dose was excreted in urine within 24 hours and 90% within 48 hours.
Randomized pilot comparing erythritol, xylitol and no added sweetener; the authors declared no competing interests. (Bordier et al. 2023, BMJ Nutr Prev Health)
42 adults with obesity randomized to 36 g of erythritol, 24 g of xylitol or nothing daily for 5 weeks.
No significant effect on pulse wave velocity or other vascular measures, abdominal or liver fat, blood lipids, glucose tolerance, uric acid, liver enzymes or creatinine. Digestive tolerance was good apart from a few diarrhea-related symptoms.
Open-label, single-arm pilot with no control group; Cargill is listed as a collaborator on the trial registration (NCT01609088) and three authors worked for Cargill. (Flint et al. 2014, Acta Diabetol)
24 adults with type 2 diabetes who drank 36 g of erythritol a day for 4 weeks, plus a 24 g dose at the first and last visits.
A single dose improved fingertip small-vessel function (0.52 to 0.87 au). After 4 weeks, central pulse pressure fell from 47 to 41 mmHg and carotid-femoral pulse wave velocity did not change significantly (P = 0.06). Without a control group, these before-and-after changes cannot be attributed to erythritol alone.
Randomized, double-blind, placebo-controlled crossover with 7-day washouts; drinks taken at home. (Storey et al. 2007, Eur J Clin Nutr)
70 healthy students aged 18 to 24 (64 completed) given 45 g sucrose or 20, 35 or 50 g of erythritol or xylitol in water.
20 and 35 g of erythritol did not significantly increase digestive symptoms compared with sucrose. 50 g of erythritol increased borborygmi and nausea. 35 and 50 g of xylitol caused watery stools and other symptoms, and xylitol produced more watery stools than erythritol at every dose compared.
Double-blind randomized trial with sorbitol candy as the control, registered as NCT01062633. (Honkala et al. 2014, Caries Res)
485 first- and second-grade children in southeastern Estonia eating candies with about 7.5 g of polyol per school day for 3 years.
At 24 months, and for surfaces at 36 months, the erythritol group had fewer teeth and surfaces with dentin caries than the xylitol or sorbitol groups, and lesions took longer to develop. Caries scores rose less in the erythritol group.
Double-blind, cluster-randomized trial in a low-caries area with natural fluoridation, compared with comprehensive preventive care alone. (Lenkkeri et al. 2012, Int J Paediatr Dent)
579 children aged 10 from 21 Finnish schools (496 analyzed) given about 4.5 g erythritol plus 4.2 g maltitol, or xylitol plus maltitol, daily on school days for 1 or 2 years.
Neither erythritol nor xylitol lozenges reduced dentin caries over 4 years compared with controls. Baseline caries was the strongest predictor of new caries.