Cardiovascular Risk and Safety

A widely used sugar substitute under scientific scrutiny — what the research says about erythritol's dental benefits, gut tolerance, and emerging cardiovascular concerns

Erythritol is a naturally occurring sugar alcohol found in small amounts in grapes, pears, and fermented foods like wine and soy sauce — though commercial versions are produced by fermenting glucose, typically from corn starch. About 70% as sweet as sugar, it contributes virtually no calories (0.2 kcal/g versus 4 kcal/g for sucrose) and has a glycemic index of zero, meaning it does not raise blood glucose or trigger an insulin response [6]. It's also gentler on the digestive system than most other sugar alcohols because it is absorbed almost entirely in the small intestine and excreted unchanged in the urine [3]. A 2023 study in Nature Medicine raised concern that higher blood erythritol levels are associated with increased cardiovascular event risk — findings that have prompted both scrutiny and ongoing research [1]. The picture is more nuanced than early headlines suggested, and understanding what the evidence actually says matters for anyone who uses erythritol regularly.

Why Erythritol Is Different from Other Sugar Alcohols

Most sugar alcohols (sorbitol, maltitol, lactulose) are poorly absorbed and reach the colon, where gut bacteria ferment them — causing gas, bloating, and loose stools at moderate doses. Erythritol behaves differently: roughly 90% is absorbed in the small intestine, passes into the bloodstream, and is excreted almost entirely in the urine within 24–48 hours, unchanged [3][6]. This means far less laxative effect and better digestive tolerance. A crossover study in 12 healthy volunteers found that daily doses up to 1 g/kg body weight (approximately 70 g for a 70 kg person) over seven days were well tolerated with no significant gastrointestinal complaints [3].

This absorption profile also explains why erythritol does not contribute meaningfully to energy intake: it is absorbed but not metabolized for fuel. The body excretes it like a water-soluble inert molecule.

Blood Sugar and Insulin: A Genuine Advantage

Erythritol does not stimulate insulin secretion and produces no measurable rise in blood glucose after consumption [6]. This makes it a practical option for people managing blood sugar — people with prediabetes, type 2 diabetes, or those following low-carbohydrate or ketogenic diets. Unlike some sweeteners that may affect gut hormone secretion in complex ways, erythritol's impact on metabolic signaling appears minimal.

In diabetic animal models and short-term human studies, erythritol consumption does not perturb glucose homeostasis or lipid profiles. It is classified as safe for use in diabetes by regulatory agencies in Europe, the US, and Japan [6].

Dental Health: Better Than Xylitol in Some Trials

Erythritol has genuine evidence for caries prevention, and in one well-designed head-to-head trial, it outperformed xylitol. Honkala et al. (2014) conducted a three-year, double-blind randomized controlled trial in 485 Estonian schoolchildren who consumed four erythritol, xylitol, or sorbitol candies three times per school day [4]. At the 24-month follow-up, the erythritol group had significantly fewer new dentin caries than both the xylitol and sorbitol groups — a finding that persisted at the 36-month assessment.

The mechanism differs slightly from xylitol's: erythritol is not metabolized by Streptococcus mutans at all, is not fermented to acids, and may reduce bacterial adhesion to tooth surfaces and decrease total mutans streptococcal counts in plaque over time. At a practical level, erythritol-containing chewing gums, mints, and toothpastes offer meaningful oral health support.

The Cardiovascular Concern: What the 2023 Study Found

The most discussed erythritol research is Witkowski et al. (2023) published in Nature Medicine [1]. The study analyzed metabolomic data from 1,157 patients undergoing cardiac risk assessment at Cleveland Clinic and found that erythritol was the compound most strongly associated with three-year risk of major adverse cardiovascular events (MACE). This was validated in two independent cohorts: a US cohort (n = 2,149) and a European cohort (n = 833). People in the highest quartile of plasma erythritol had approximately double the risk of MACE compared to those in the lowest quartile.

The study went further: it showed that erythritol — at concentrations achieved after consuming a typical erythritol-sweetened beverage or snack bar — enhanced platelet aggregation and accelerated thrombus formation in ex vivo human blood samples. This proposed a potential mechanism linking high erythritol exposure to cardiovascular events.

This is notable research from a credible group using rigorous methodology. It should not be dismissed. However, several important limitations and complications affect interpretation.

Why the Association May Not Mean Dietary Erythritol Is the Problem

A critical complication: most circulating erythritol in human blood is not from food. The body makes its own erythritol endogenously, as a byproduct of glucose metabolism via the pentose phosphate pathway (PPP). Hootman et al. (2017) demonstrated this in a landmark PNAS study using stable isotope tracing [2]. They showed that blood erythritol levels rise with adiposity and elevated HbA1c: young adults who gained central adiposity over 36 months had erythritol concentrations 15-fold higher than stable peers; those with HbA1c above 5.05% had 21-fold higher blood erythritol compared to peers with lower HbA1c.

This means high plasma erythritol may primarily reflect dysregulated glucose metabolism — the pentose phosphate pathway running at elevated capacity due to insulin resistance or high glucose intake — rather than dietary erythritol consumption. In other words, people who are metabolically unhealthy produce more erythritol internally, and those same people have higher cardiovascular risk for many independent reasons. The elevated blood erythritol may be a marker of disease rather than a cause.

Mazi and Stanhope (2023) reviewed the full evidence base and concluded that "it is unlikely that dietary erythritol is mediating these associations" and that high plasma erythritol more likely reflects underlying glycemic dysregulation [6].

Mendelian Randomization: No Causal Evidence Found

Khafagy et al. (2024) applied Mendelian randomization — a technique that uses genetic variants as natural experiments to test causality — to examine whether genetically elevated erythritol leads to cardiometabolic disease [5]. Published in Diabetes, the analysis found no supportive evidence that higher erythritol concentrations cause coronary artery disease, diabetes, or metabolic disease. Interestingly, it found suggestive evidence that genetically higher erythritol may be associated with lower BMI (b = −0.04, p = 1.23 × 10⁻⁵), which runs counter to the concern that erythritol promotes metabolic harm.

This does not fully exonerate dietary erythritol — Mendelian randomization captures genetic variation in endogenous production, not the acute effects of large oral doses — but it does support the interpretation that the observational association between plasma erythritol and cardiovascular events is likely confounded by metabolic disease status.

Practical Guidance

The evidence does not clearly support avoiding erythritol, especially when the alternative is sugar. A zero-calorie, zero-glycemic-index sweetener with dental benefits and excellent gut tolerance remains appealing for most people. The caveats worth noting:

If you have established cardiovascular disease, metabolic syndrome, or insulin resistance: the 2023 data warrant some caution about consuming large amounts of erythritol regularly. The platelet aggregation findings from the Witkowski study are biologically plausible even if causality is unproven. Moderation makes sense.
For healthy people using erythritol occasionally: the evidence does not justify alarm. Daily doses well under 30 g are unlikely to drive meaningful changes in plasma erythritol beyond what the body produces endogenously.
Dental benefit: the caries prevention evidence is genuine. Erythritol in chewing gum or toothpaste is a reasonable oral health strategy regardless of any cardiovascular uncertainty.
As a sugar replacement: for blood sugar management, erythritol remains one of the best-characterized options — genuinely inert with respect to glucose and insulin.

The broader scientific debate is ongoing. Treating erythritol as a free-pass sweetener to be consumed in large quantities is premature. Using it thoughtfully as a sugar substitute in modest quantities appears reasonable given current evidence.

See our xylitol page for comparison with a related sugar alcohol, and our artificial sweeteners page for the broader context of non-nutritive sweeteners.

Evidence Review

The 2023 Nature Medicine Study: Methods and Findings

Witkowski et al. (2023) performed untargeted metabolomics on plasma samples from 1,157 stable patients referred for cardiac risk evaluation (the discovery cohort), followed by targeted validation [1]. Erythritol emerged as the metabolite most strongly associated with incident three-year MACE (composite of non-fatal MI, non-fatal stroke, and all-cause mortality). This association persisted after adjustment for traditional cardiovascular risk factors including age, sex, BMI, smoking, hypertension, diabetes, and kidney function (adjusted hazard ratio in top versus bottom quartile approximately 1.8 to 2.1 across cohorts).

The finding was reproduced in two validation cohorts: a US cohort of 2,149 patients (PREDETERMINE Study, Cleveland Clinic) and a European cohort of 833 patients (University Hospital Charité, Berlin), with consistent hazard ratios in the range of 1.7 to 2.1 for highest versus lowest plasma erythritol quartiles.

The mechanistic experiments showed that erythritol at physiological concentrations (as achieved after ingesting 30 g orally — a typical serving in many keto or sugar-free products) significantly enhanced human platelet aggregation in response to sub-threshold stimuli and accelerated thrombus formation under physiological flow conditions in microfluidic models. These platelet effects persisted for at least two to three days after erythritol ingestion in five healthy volunteers who consumed a 30 g oral dose.

Critical limitations of the Witkowski study: The observational cohorts were patients at elevated cardiovascular risk, not general-population samples — limiting generalizability. Plasma erythritol was measured at a single time point, not longitudinally. Crucially, the study did not measure dietary erythritol intake: it is unclear how much of the plasma erythritol in participants reflected dietary consumption versus endogenous production. The platelet experiments used isolated platelets and microfluidic models, not clinical event data, so the mechanistic connection to actual thrombosis events remains an inference.

Endogenous Production: The Confounding Problem

Hootman et al. (2017) used stable isotope-assisted metabolomics in 264 university students followed over 36 months [2]. Key findings: erythritol is produced endogenously from glucose via the non-oxidative branch of the pentose phosphate pathway; blood erythritol levels predicted prospective adiposity gain (those gaining central adiposity had 15-fold higher erythritol, p < 0.01); erythritol concentrations correlated strongly with HbA1c, with participants above HbA1c 5.05% showing 21-fold higher blood erythritol than those below (p = 3 × 10⁻⁶). This dataset establishes that plasma erythritol is primarily a biomarker of glucose dysregulation and metabolic dysfunction, not dietary intake.

This confounding is substantial. People with cardiometabolic risk factors produce more erythritol endogenously. Those same people have higher cardiovascular event risk. Attributing the excess risk to dietary erythritol based on plasma measurements alone requires demonstrating that plasma erythritol predicts events independently of metabolic status — which the Witkowski study attempted through multivariable adjustment, but residual confounding by unmeasured metabolic variables remains likely.

Mendelian Randomization Analysis

Khafagy et al. (2024) selected genetic instruments associated with circulating erythritol from GWAS data in European-ancestry cohorts and applied two-sample Mendelian randomization to assess causal effects on coronary artery disease (CAD), BMI, waist-hip ratio, HbA1c, fasting glucose, and kidney function [5]. Using three different MR approaches (IVW, MR-Egger, weighted median), none found evidence for a positive causal association between erythritol and CAD (pooled estimate b = −0.033 ± 0.02, p = 0.14, and b = 0.46 ± 0.37, p = 0.23 across instruments). There was a statistically significant inverse association between genetically elevated erythritol and BMI.

The authors concluded that their data do not support a causal role of erythritol in cardiometabolic disease, and that the observational associations more likely reflect reverse causation — metabolic disease elevating endogenous erythritol production rather than erythritol causing disease. Important caveat: MR captures effects of lifelong genetic variation in erythritol production, which may not capture the specific effects of acute, high-dose dietary erythritol consumption on platelet function.

Dental Caries Prevention: 3-Year RCT

Honkala et al. (2014) conducted a rigorously designed cluster-randomized, double-blind controlled trial over three years in 485 Estonian primary schoolchildren [4]. Children were randomized to consume four candies containing erythritol (2.5 g), xylitol (2.5 g), or sorbitol (control, 2.5 g) three times per school day. The primary outcome was development of new dentin caries lesions.

Results: at 36 months, the erythritol group had significantly fewer new dentin caries surfaces than both the xylitol group (p = 0.024) and the sorbitol group (p = 0.0003). The reduction in caries experience (new dfs increment) was approximately 30–35% lower in the erythritol group versus sorbitol controls. A post-trial survival analysis (PMID 27806364) confirmed the caries-preventive effect persisted up to three years after the intervention ended.

Proposed mechanisms include: (1) erythritol is not fermented by S. mutans or other cariogenic bacteria, so it does not lower plaque pH; (2) erythritol may reduce bacterial cell surface adhesion, reducing initial colonization; (3) reductions in total mutans streptococcal counts in saliva were observed over the intervention period. These mechanisms collectively suggest that erythritol has stronger anti-caries properties than xylitol, which is notable given the established evidence base for xylitol.

Gastrointestinal Tolerance

Tetzloff et al. (1996) performed a double-blind crossover study in 12 healthy male volunteers who consumed erythritol or sucrose for 7 days each at escalating doses: 0.3 g/kg on day 1, 0.6 g/kg on day 2, 1.0 g/kg on days 3–7 [3]. At 1 g/kg body weight daily (70 g for a 70 kg person), erythritol was well tolerated with no significant differences in gastrointestinal symptoms versus sucrose.

The tolerance advantage stems from erythritol's small molecular size and high intestinal absorption rate. Unlike sorbitol, which draws water into the intestine via osmotic effect, or lactulose, which is fermented in the colon to gas, erythritol is absorbed before reaching the colon in most of the dose. Unabsorbed erythritol that does reach the colon is also poorly fermented by gut bacteria. Absorption and metabolism dose-ranging studies in humans have confirmed that 80–90% of ingested erythritol is recovered in urine within 24–48 hours, with less than 10% reaching the colon [6]. This profile makes erythritol the best-tolerated sugar alcohol at moderate doses.

Regulatory Status and Overall Safety Profile

Erythritol has GRAS (Generally Recognized As Safe) status in the United States and is approved as a food additive (E 968) in the European Union. A 2023 EFSA re-evaluation found no safety concerns at typical dietary exposure levels. Unlike xylitol and sorbitol, erythritol does not have documented toxicity in dogs or cats, making it a less hazardous household sweetener from a pet-safety standpoint.

The main area of genuine scientific uncertainty remains the cardiovascular signal from the Witkowski study and subsequent work. The weight of mechanistic and genetic evidence currently favors the interpretation that elevated plasma erythritol is a downstream consequence of metabolic dysregulation rather than a causal cardiovascular risk factor. This should be revisited as prospective dietary intervention trials — which would directly test whether increased dietary erythritol raises MACE risk — are conducted. Until such trials exist, the precautionary use of erythritol in moderate quantities as a sugar substitute appears defensible, particularly for individuals without pre-existing cardiovascular or metabolic disease.

References

The artificial sweetener erythritol and cardiovascular event riskMarco Witkowski, Ina Nemet, Hassan Alamri, Jennifer Wilcox, Nilufar Gupta, Nathalie Nimer, Arash Haghikia, Xinyang S Li, Yuping Wu, Elske Lassen, Gary I Ash, Hooman Allayee, W H Wilson Tang, Ulf Landmesser, Stanley L Hazen. Nature Medicine, 2023. PubMed 36849732 →
Erythritol is a pentose-phosphate pathway metabolite and associated with adiposity gain in young adultsKathleen C Hootman, Jean-Pierre Trezzi, Lisa Kraemer, Lindsay S Burwell, Xiaowei Dong, Kristin A Guertin, Molly Reilly, Peter J Mucha, Richard D Beger, Lena Soldatova, Clare F Garland. Proceedings of the National Academy of Sciences USA, 2017. PubMed 28484010 →
Tolerance to subchronic, high-dose ingestion of erythritol in human volunteersWolfgang Tetzloff, François Dauchy, Sined Medimagh, D Carr, Albert Bär. Regulatory Toxicology and Pharmacology, 1996. PubMed 8933645 →
Effect of erythritol and xylitol on dental caries prevention in childrenSisko Honkala, Riina Runnel, Mare Saag, Jana Olak, Rita Nõmmela, Silvia Russak, Pirkko-Liisa Mäkinen, Tero Vahlberg, Gwen Falony, Kauko Mäkinen, Eino Honkala. Caries Research, 2014. PubMed 24852946 →
Erythritol as a Potential Causal Contributor to Cardiometabolic Disease: A Mendelian Randomization StudyRana Khafagy, Andrew D Paterson, Satya Dash. Diabetes, 2024. PubMed 37939167 →
Erythritol: An In-Depth Discussion of Its Potential to Be a Beneficial Dietary ComponentTagreed A Mazi, Kimber L Stanhope. Nutrients, 2023. PubMed 36615861 →