The Story Everyone Knows
Sugar is the villain of modern nutrition, and sugar substitutes were supposed to be the escape hatch: all of the sweetness with none of the calories, the blood-sugar spike, or the guilt. Dietary guidelines steer people with obesity and diabetes toward these substitutes, and dentists go further, actively recommending xylitol gum because it starves cavity-causing bacteria. It even sounds wholesome: birch sugar, keto-friendly, nothing to hide.
The pitch had one unexamined assumption. Regulators declared sugar alcohols safe based on classic toxicology: feed rats a lot of it, watch for tumors and organ damage, find none, stamp the approval. Nobody ran the other experiment: what these compounds do to blood clotting in humans over decades. The safety file was complete on paper, and empty where it mattered.
The Study That Spoiled the Sweetener
In June 2024, the European Heart Journal published the paper that opened the file, in which Marco Witkowski, then at the Cleveland Clinic's Lerner Research Institute, and senior author Stanley Hazen ran targeted metabolomics on blood from 1,157 patients undergoing cardiac evaluation, then validated the signal in 2,149 more. Patients in the highest third of blood xylitol were about 50 percent more likely to suffer a heart attack, stroke, or cardiovascular death within three years than those in the lowest third after adjusting for established risk factors.
An association is only a hint, so the team chased the mechanism across three levels: in isolated human blood, xylitol made platelets more sensitive to clotting signals; in mice, raising blood xylitol sped up clot formation and artery blockage; and in ten healthy volunteers who drank a xylitol-sweetened beverage, blood xylitol rocketed 1,000-fold within 30 minutes while platelets grew more reactive, settling to baseline within four to six hours.
This was the sequel, not the premiere. A year earlier the same group had published the erythritol version in Nature Medicine: 4,139 patients, same pattern, with the highest erythritol quartile carrying roughly double the three-year event risk. Then a 2024 follow-up closed the most obvious loophole by testing diet directly: twenty healthy volunteers received either 30 grams of erythritol or 30 grams of glucose, and in every erythritol subject, across every clotting trigger and dose tested, platelets grew more reactive, while glucose did nothing.
The Arithmetic of a Single Serving
Run the numbers from that feeding trial and the scale of the exposure becomes concrete. Baseline blood erythritol sat at 3.75 micromoles per liter; after one 30-gram serving, the median hit 6,480, a jump of more than 1,000-fold. The FDA's technical memo on the earlier paper noted that platelet activation in laboratory preparations began at 4.5 to 45 micromoles per liter, so divide 6,480 by 45 and you get 144. A single serving drove blood levels roughly 140 times above the concentration that switched platelets on in the dish, and the xylitol pilot showed the same thousand-fold spike within half an hour of one sweetened drink, with levels staying elevated for hours.
Then Came the Replication
The fair objection went like this: cardiac patients are not the general population, and blood levels are not diet. In August 2026, Witkowski, now at Charité in Berlin, answered with scale, presenting data from 17,710 adults in two general-population cohorts, the Canadian Longitudinal Study on Aging and EPIC-Norfolk, at the European Society of Cardiology congress. In the Canadian cohort, the highest xylitol quartile faced a 57 percent higher risk of heart attack, stroke, or cardiovascular death over six years than the lowest quartile, after adjusting for standard risk factors; in EPIC-Norfolk, followed up to 30 years, the gap was 18 percent. Event rates rose steadily with blood levels, a dose-response, and the association held across age and sex. One honest flag: this is a conference presentation, and peer-reviewed publication is still pending.
The Strongest Counterargument
The best case against this story is that the molecule is a bystander. The human body manufactures xylitol and erythritol on its own through the pentose phosphate pathway, and metabolic dysfunction cranks up the production line, with obesity linked to 15-fold higher blood erythritol in one analysis. On this reading, high blood xylitol marks a broken metabolism the way smoke marks a fire. The dietary sugar did the damage; the sweetener is just what the lab happens to measure. A published letter re-examining the 2023 paper's supplementary tables found cardiovascular risk factors climbing in lockstep with erythritol quartiles, which is exactly what residual confounding looks like.
The animal literature adds its own awkward fact: two-year rodent feeding trials, with diets up to 20 percent sugar alcohol, reported no cardiovascular abnormalities and no shortened lifespan, while the human pilots were tiny, ten to twenty people, with no placebo arm in the parallel designs, a single post-drink measurement, and clotting tested in a dish rather than in a living artery. A transient post-meal twitch in platelet reactivity may never translate into a heart attack years later. Regulators have not moved. The FDA reviewed the 2023 paper in a detailed technical memo and left erythritol's generally-recognized-as-safe status exactly where it was.
That is the honest stalemate: the association has replicated and the mechanism is plausible, but no randomized trial with hard endpoints exists, and given the decades one would take, none soon will.
What We Didn't Prove
No study here randomized people to years of xylitol versus sugar and counted heart attacks, because that trial does not exist. The cohorts measured blood levels, not diet, so the data cannot separate sweetener intake from the body's own production, and the 2026 replication, the largest dataset of all, has not finished peer review. The feeding pilots were small and brief, and xylitol and erythritol are different molecules, and while both now carry suspicious files, the dose-response replication exists so far only for xylitol. None of this erases xylitol's dental record: the cavity evidence comes from separate trials, and a stick of gum is a far smaller dose than the pilots' sweetened drinks.
The Bottom Line
The safest-seeming sweetener now has the most suspicious cardiovascular file. Two sugar alcohols track with higher clot-related risk in patients and the general population, and feeding people an ordinary serving makes their platelets more reactive within the hour. The counterevidence is real but has not made the signal go away. Until a long-term randomized trial settles it, sugar-free should stop reading as risk-free.
What You Can Do
Read labels for xylitol (also birch sugar, wood sugar, E967) and erythritol in everyday products: gum, mints, keto snacks, protein bars, and sugar-free drinks are the usual carriers. Heavy daily use of either is now an open scientific question, not a settled safe choice, so treat it that way. If you switched to sugar substitutes to protect a diabetic or cardiac condition, do not overhaul your diet on the basis of this article; the absolute risk increases are modest, sugar carries its own well-documented harms, and any change runs through your clinician.
Keep the dental picture separate. Xylitol gum's cavity benefit rests on its own trials, and the doses in gum are small. This is a story about high, repeated dietary exposure, not about brushing your teeth. The least-regret move: less sweetness overall beats swapping one sweetener for another.