Artificial sweeteners are either killing you or not killing you
Did you happen to come across the news reports on the potential cardiovascular harms of a common artificial sweetener?
The headlines all had the same general gist: new research suggests an artificial sweetener has a “link” to heart disease:
Here’s the New York Times: “Study suggests possible link between sugar substitute and heart issues.”
The Washington Post: “Artificial sweetener linked to higher heart attack risk, study says.”
The Wall Street Journal: “Are artificial sweeteners bad for you? New worries about link to strokes, heart attacks.”
The Today Show: “What is erythritol? Sugar substitute linked to heart attacks, stroke”
The word “link” is doing a lot of work in all of those headlines. It’s really hard to know what it means - link could be really bad, or it could be a real nothingburger.
So should you be scared? Or, more importantly, should you be changing your habits in response to a study like this?
Let’s take a look.
Erythritol - the “artificial” sweetener in question, is not artificial. It’s even naturally made by our body
Erythritol is a naturally occurring sugar alcohol that is found in fruits such as melon, watermelon, pears, grapes as well as in fermented foods like cheese, miso, soy sauce, and sake.
It’s also made by our bodies. It’s been described that people with central adiposity (AKA belly fat) make more erythritol and that oxidative stress (the type of thing that antioxidants fight against) increases erythritol production as well.
In other words, the sicker we are, the more erythritol our bodies make.
This is important as we grapple with the implications of the study.
This study is interesting and complicated - it’s actually a bunch of studies all published in one paper
The paper in question was published in Nature Medicine and titled, “The artificial sweetener erythritol and cardiovascular event risk.”
The group that worked on this study previously achieved media notoriety with work on a substance called TMAO, which linked eggs and red meat to heart disease.
I’m not going to get into the TMAO stuff today, so let’s get back to erythritol.
It’s worth thinking about each subset of this study before we put it all together to draw some conclusions.
Part 1: Discovery cohort
They looked at bloodwork done of patients being evaluated for cardiovascular disease and measured levels of a whole bunch of different substances in their blood (“untargeted metabolomics”) and found that erythritol as well as a number of other related molecules seemed to be present in higher quantities in people who developed heart disease.
The quartile that had the highest blood levels of erythritol had a higher risk of cardiovascular events over 3 years, based on unadjusted data, the people with the highest erythritol levels had a 3-fold higher incidence of cardiovascular events
Part 2: Validation cohort
They performed much more targeted evaluation of erythritol levels, both more precisely measuring erythritol and assessing multiple populations of people.
The curves look similar to the discovery cohort, where the top quartile has a higher incidence of heart disease:
In the validation cohort, they also found that people with established cardiovascular disease tended to have higher blood levels of erythritol than people who didn’t have cardiovascular disease.
Part 3: Platelets exposed to erythritol
This is the blood clotting part of the study that got a lot of attention.
A lot of the articles mention that a possible mechanism of why erythritol might cause heart attacks or strokes is through increased blood clotting.
The authors did a number of different tests to on the impact of erythritol on human platelets, human whole blood, and mice, measuring markers of blood clotting, where they found that exposure to high levels of erythritol stimulated platelet aggregation in a test tube and stimulated blood clotting in the carotid arteries of mice.
This is tough to interpret - a lot of things make platelets stickier in response to ADP, and demonstrating that the effect you see on platelets in a test tube or in a mouse translates to an increased risk of blood clots in people doesn’t usually pan out.
This is the complicated part of the study, and if you want to read more details about it, I would highly recommend this article from Kevin Klatt.
There are a lot of steps to go before you can really conclude anything about blood clotting in humans after having food or drink sweetened with erythritol.
One thing that you may note that the authors did not do here: establish that giving erythritol to people increased risks of blood clotting.
Part 4: Giving humans erythritol in the form of an erythritol sweetened beverage
The authors then gave 8 human volunteers a 30g challenge or erythritol in the form of a drink and measured their blood levels afterwards.
They clearly showed that erythritol consumption raises blood levels of erythritol:
This paper is really clever, but doesn’t actually prove anything about anything
If you want to make a persuasive argument that using erythritol as a sweetener increases cardiovascular risk, you need to demonstrate that giving erythritol to people increases cardiovascular risk.
The authors in the study did not do that.
They played a clever game of using the transitive property to insinuate causation.
Step 1: higher levels of erythritol in the blood are seen in patients who have a higher risk of cardiovascular disease
Step 2: erythritol makes platelets more likely to clot in a test tube. It also may make mice have more blood clotting under artificial circumstances
Step 3: giving people erythritol raises blood levels of erythritol
You can almost see how this could be persuasive - you give people erythritol and their blood becomes stickier, so they have more heart attacks and more strokes. And that’s part of why we can observe that people with more cardiovascular disease have higher levels of erythritol in their blood.
But the authors didn’t actually prove that consuming erythritol causes anything clinically relevant to happen.
The data here is just so preliminary, I don’t see how you can conclude anything other than the idea that erythritol may increase heart disease is a hypothesis that needs to be directly tested.
We don’t need more indirect studies like this to “raise questions.”
I can’t stress this strongly enough - just because there’s no proof that erythritol does anything bad doesn’t mean it should be a huge part of your diet
I think that this is a useless study.
It changes nothing about my behavior, my advice to my patients, or my sentiments about erythritol versus sugar, sucralose, aspartame, or stevia.
My take on the sum total of the evidence is that it’s fairly likely that sugar substitutes are better for us than sugar. I would never advise a patient have a regular Coke instead of a Diet Coke because of preliminary data like this.
But I don’t mean that as a suggestion that our diets should be filled with sugar substitutes (whether natural or artificial).
In my rules about nutrition, I wrote that “most of us know what junk food is and also know that we should avoid it. A lot of the questions about eating healthy are folks trying to find loopholes in this concept.”
And so when it comes to this stuff, use common sense and don’t make these major parts of your diet. I get the sense that using sweeteners like erythritol to make keto versions of baked goods is a reason why these diets sometimes fail.
But when it comes to erythritol and cardiovascular risk, I’m not even ready to say that there’s smoke here.
This is incredibly preliminary data that needs to be studied more before we have any clue whether its nonsense or something alarming.
The best advice on how to use this study to change your behavior comes in the form of this tweet: