Published March 9, 2023. Revised on March 13, 2023.
In a recent publication in Nature Medicine, researchers linked the sugar-substitute erythritol to an increased risk of heart attacks and strokes. This paper received a lot of coverage from traditional media outlets as well as on social media because erythritol has increasingly been utilized as a non-caloric sweetener, particularly in the low-carb- and keto-diet communities.
In this story, I summarize the study’s main findings and outline my main concerns to give you a sense of whether it may indeed be a good idea to limit your consumption of erythritol.
Summary of the study
The primary objective of the investigation described in this paper was to identify novel biomarkers of cardiovascular disease risk. To this end, the researchers measured the blood concentrations of a large number of metabolites in a cohort of people at high risk of a cardiovascular event.
They found that the risk of developing a heart attack or a stroke was substantially increased in participants with elevated concentrations of the sugar alcohol erythritol in their blood.
An association such as this could be a chance finding, because the original investigation was not focused on erythritol. The authors, therefore, sought to confirm this association between blood erythritol and cardiovascular events in an additional cohort in the United States and another one in Europe. Reassuringly, the association could also be found in both of these cohorts.
Because an association does not necessarily reflect a cause-and-effect relationship, the investigators also conducted several mechanistic experiments. They found that elevated concentrations of erythritol increase platelet aggregation and blood clotting in test tubes, and also accelerated the rate of thrombus formation in vivo in mice. Because an increased tendency of blood platelets to form clots could plausibly increase the risk of heart attacks and strokes, the authors suggested that increased coagulation and thrombosis may be mechanistic links between elevated blood levels of erythritol and an increased risk of cardiovascular events.
Now to the part of the investigation that garnered the most interest from the public: erythritol is a component of the human diet. It occurs naturally in some foods, albeit in very small amounts, and is added as a non-caloric sweetener in many processed foods. This raised the possibility that blood erythritol levels may be increased mostly in people who consume a lot of erythritol. And this, in turn, would mean that consuming foods sweetened with erythritol may increase the risk of heart attacks and strokes.
To assess the degree to which dietary erythritol affects blood erythritol levels, the investigators conducted a small human feeding experiment. They showed that a single dose of 30g of erythritol (a typical daily dose of someone who consumes a lot of erythritol) raises blood erythritol concentrations by up to 1,000-fold in healthy people, who usually have very low baseline erythritol concentrations (3–5 uM).
Taken together, the study suggested that regular consumption of erythritol substantially raises blood erythritol levels, that such elevated blood erythritol levels are associated with an increased risk of cardiovascular events such as a heart attack or a stroke, and that high concentrations of erythritol also increase blood clotting and thrombosis in test tubes and in mice (see figure below).
This may sound like a compelling story, but I am concerned about several limitations that substantially reduce my confidence that the observed associations between erythritol and cardiovascular events reflect cause-and-effect relationships. Let me guide you through these concerns.
Concern #1: Potential Lack of Generalizability
The cohorts studied in this investigation all consisted of elderly individuals in their mid-60s to mid-70s who suffered from numerous pre-existing conditions that dramatically increased their risk of cardiovascular events. Specifically, many participants were overweight or obese, about half had previously had a heart attack, about three-quarters suffered from hypertension, and many had type 2 diabetes.
Thus, it’s critical to be clear that the results from this study cannot necessarily be generalized to a younger or healthier population.
Concern #2: Blood Erythritol May Not Be a Biomarker of Dietary Erythritol Intake
While the study showed that a dose of dietary erythritol substantially increases blood erythritol levels, it is also well known that erythritol can be synthesized endogenously from glucose, in the pentose phosphate pathway. It is also possible that the rates of absorption or excretion (through the urine) may differ between people.
Therefore, it has thus far remained unclear whether blood erythritol can serve as a biomarker of dietary erythritol intake. It must be emphasized that research conducted using data on blood erythritol may provide little or no information about the health effects of dietary erythritol.
Concern #3: Potential Confounding by Pre-Existing Conditions
Many previous studies have shown that blood erythritol concentrations are elevated in individuals with obesity, type 2 diabetes, or chronic kidney disease, and are also associated with the later development of some of these conditions. This raises the possibility that different conditions increase the endogenous synthesis of erythritol through the pentose phosphate pathway, and/or affect the rate of excretion through the urine.
In fact, I mentioned above that healthy people have low blood erythritol levels of about 3–5 uM, unless they consume dietary erythritol. In the present study, blood erythritol concentrations seemed higher than this figure of 3–5 uM in all 4,000 participants studied across all three cohorts. In most study participants, blood erythritol levels were 10–1,000-fold higher than the 3–5 uM we would expect in healthy people who have not recently consumed erythritol. This may mean that all 4,000 participants were regular consumers of erythritol, or that — much more likely — blood erythritol levels were elevated in these participants mostly because they suffered from pre-existing medical conditions such as obesity, type 2 diabetes, or chronic kidney disease that are known to cause increased endogenous erythritol synthesis.
Because obesity, kidney disease, and type 2 diabetes are all major risk factors for cardiovascular disease, this raises the possibility that the relationship between blood erythritol and cardiovascular events may be confounded.
Confounding means ‘confusion of effects’. In this context, it means that the observed association between blood erythritol and cardiovascular events may not be reflective of a cause-and-effect relationship. Instead, elevated blood erythritol may just be an innocent bystander in what may truly be going on here: that those who are more obese and sicker are at higher risk of cardiovascular events.
The authors adequately adjusted their analyses for some of these potential confounding factors. However, they did not adjust for all conditions associated with elevated blood erythritol (i.e., chronic kidney disease), and some conditions such as type 2 diabetes are underdiagnosed. Therefore, the potential for what we call residual confounding remains. Residual confounding means that even though the authors adjusted for some potential confounding variables, the possibility for confounding was probably not entirely eliminated.
Given that many of the participants in all three cohorts already had pre-existing cardiovascular disease, one could even argue that the observed association between blood erythritol levels and cardiovascular events could be explained by reverse causation. Reverse causation would mean that the observed association may be reflective not of blood erythritol contributing to the development of a cardiovascular event, but of (pre-clinical) cardiovascular disease causing increased blood erythritol levels.
Concern #4: Potential Confounding by Dietary Factors
We similarly need to be aware that people who consume the most erythritol almost certainly follow a diet that differs in many other ways from the diet of people who consume no or little erythritol. Because erythritol is very commonly used to sweeten products that cater to the low-carb- or keto-diet communities, it is likely, for example, that high erythritol intake is associated with higher fat and saturated fatty acid intake, but lower consumption of foods rich in carbohydrates, leading to lower intake of fiber. Because erythritol is often added to industrial foods, such as energy drinks or snack bars, ultra-processed food intake may also be higher in those consuming the most erythritol.
And in turn, it is certainly possible that any of these associated dietary factors could affect the risk of a cardiovascular event. This could confound the relationship between blood erythritol, heart disease, and strokes. For example, those with the highest plasma erythritol concentrations may have a higher risk for cardiovascular events not because of their elevated blood erythritol but because of their diet rich in ultra-processed foods and low in fiber. The authors could not statistically adjust for this possibility in this study because data on dietary intakes were not available.
Concern #5: Mechanistic Studies of Erythritol Effects on Blood Clotting and Thrombus Formation May not be Relevant for Humans
Given the substantial potential for confounding and even reverse causation mentioned above, the merits of the study depend heavily on the mechanistic studies in which the authors showed that erythritol increases coagulation and thrombosis. Given that these were conducted in test tubes, with one experiment in vivo in mice, questions remain about whether these data are relevant for humans with elevated levels of erythritol in their blood.
This is even more so because the clinicaltrials.gov entry of the human feeding experiment suggested that the authors aimed to test the effect of a 30g dose of erythritol on blood erythritol levels and platelet aggregation in 40 healthy participants.
The fact that they shared only data on changes in blood erythritol levels from 8 participants, and no data on changes in platelet aggregation raises a major red flag: why were in vivo human data on the acute effects of an oral dose of erythritol on platelet aggregation not reported in the paper?
Given that at least eight participants had been studied suggests that the effects on platelet aggregation could have easily been measured. The authors decided not to share these data, explaining in the paper that these studies are “ongoing and not part of this manuscript”. We might want to consider that while these data could substantially strengthen the story this manuscript aims to convey, they could also derail the story entirely if erythritol was not found to increase platelet aggregation in humans.
I have seen many people on social media argue that they consider the paper conclusive that dietary erythritol increases the risk of cardiovascular events, and that they will no longer consume erythritol. I do not share this interpretation. The study has substantial potential for confounding by both pre-existing medical conditions and other dietary factors, and the story that dietary erythritol increases the risk of cardiovascular events depends heavily on the suggested effect of erythritol on platelet aggregation. These data were generated in vitro and in mice, and potentially available human data on platelet aggregation were not included in the manuscript. Thus, questions remain as to whether erythritol indeed impacts platelet aggregation and thrombosis. I am skeptical.
Taken together, even though the paper was published in a highly prestigious medical journal, it suffers from numerous limitations. It is very well possible that blood erythritol is just an innocent bystander in people at extremely high risk of cardiovascular disease, and that in vivo in humans, erythritol will prove to be benign and not affect coagulation and blood clotting.
However, the publication does raise the possibility that dietary erythritol could increase the risk of cardiovascular events, potentially by increasing platelet aggregation and thrombus formation. The paper leaves both possibilities open.
However, I want you to take away one thing from this story: the science is not nearly as conclusively settled as some have made it out to be. At this point, it is anyone’s guess whether dietary erythritol indeed increases cardiovascular event risks, and we will need to wait for better evidence to know for sure.
- Witkowski M et al.; The artificial sweetener erythritol and cardiovascular even risk. Nature Medicine 2023; published online on February 27, 2023.
- Hootman KC et al.; Erythritol is a pentose-phosphate pathway metabolite and associated with adiposity gain in young adults. Proceedings of the National Academies of Sciences USA 2017; 114: E4233-40.
- Haukka JK et al.; Metabolomic profile predicts development of microalbuminuria in individuals with type 1 diabetes. Scientific Reports 2018; 8: 13853.
- Tang W. Consumption of oral artificial sweeteners on platelet aggregation and polyol excretion (COSETTE). Registration of clinical study on https://www.clinicaltrials.gov. Accessed on March 9, 2023.
- Mazi TA and Stanhope KL. Erythritol: an in-depth discussion of its potential to be a beneficial dietary component. Nutrients 2023; 15: 2023.