The two cohorts do not agree on the size of the effect, and that disagreement is the most informative part of the analysis. The Canadian sample yielded a 57 percent higher risk of a major cardiovascular event over six years of follow-up [4]. EPIC-Norfolk yielded 18 percent over 30 years, as the report words it [5]. The larger relative excess, about 3.2 times the smaller [17], came from the follow-up that was one fifth as long [18]. Both can be true, but only if something about the two populations or the two measurements differs enough to account for the gap, and a single baseline blood draw plainly carries less information about year 28 than about year 5.
Be pedantic about the exposure, because that is where the practical question lives. Participants were sorted by circulating xylitol, not by what they reported eating or chewing [15]. Xylitol occurs naturally, including in the body, while commercial products carry it at roughly a thousand times those concentrations [8]. A high blood value can therefore be a record of intake or a record of metabolism, and this design does not separate the two. The thing it does not tell you is whether dropping xylitol-sweetened gum, drinks or toothpaste [9] out of someone's week moves that person out of the top tier.
There is prior reason not to wave the signal away. Earlier work suggested xylitol can help blood platelets form clots, and had already tied high blood xylitol to cardiovascular events among people who were already at elevated risk [10]. The Berlin group's answer to that literature was population and duration, a larger and more general sample followed for longer [14], presented in Munich at the European Society of Cardiology congress [3]. It is also why the graded pattern matters more than any single ratio. Risk that climbs from low to middle to high exposure tiers [6] and holds across age, sex, body mass index, hypertension and diabetes [7] is harder to manufacture by accident than one extreme-versus-extreme comparison.
What the reported account does not describe is adjustment for kidney function or smoking [19], and it gives relative risks without absolute event rates or event counts [16]. A 57 percent increase on an unstated baseline is not yet a number anyone can carry into a consultation.
Dan Atar of the ESC's communication committee set the ceiling himself: causality is difficult to infer from observational work, and further studies are mandated [11]. My own reading, conditional on the full publication showing the exposure cutoffs and a kidney-function adjustment: xylitol has moved out of the settled-safe drawer and into the measure-it drawer. That is less than a headline wants and more than is convenient for anyone specifying a sweetener for a 2027 product line. Marco Witkowski's framing of the problem, that these compounds are consumed by people who think they are healthier than sugar and are generally regarded as safe by regulatory agencies [12], is where the pressure now sits.
