Sep 18 2014

Artificial Sweeteners and Diabetes

A new study published in Nature is getting a lot of press, and it seems making a lot of people worried. The Nature News article discussing the study has the headline: Sugar substitutes linked to obesity. I think this headline is misleading. Here’s a breakdown of what the study does and does not tell us.

The study’s title is more descriptive, as one might expect: Artificial sweeteners induce glucose intolerance by altering the gut microbiota. The authors (A team led by Eran Elinav of the Weizmann Institute of Science in Rehovot, Israel) studied three noncaloric artificial sweeteners (NAS), saccharin, sucralose, and aspartame. They fed mice that either had a microbiota (bacteria colonizing their gastrointestinal system) or those that were germ free either NAS or control food without NAS. After 11 weeks the mice fed NAS showed signs of glucose intolerance – their blood sugar rose more when challenged with a dose of sugar.

They also found that mice treated with antibiotics did not have this response. Further, they performed fecal transplants from NAS treated mice to germ-free mice and found that the glucose intolerance transferred with the bacteria.  They also cultured bacteria with NAS and transplanted that into mice, who then became glucose intolerant.

All of this strongly suggests that consuming NAS alters the gut bacteria, which in turn has an effect of metabolism of the host, leading to glucose intolerance. Glucose intolerance is a risk factor for type II diabetes, which essentially is severe glucose intolerance. The study did not have anything directly to do with obesity.

To see if their results would apply to humans, the researchers fed 7 human subjects NAS and found that 4 of them developed glucose intolerance, just like the mice.

This is all a nice bit of research. It certainly raises some interesting possibilities that deserve follow up study. I don’t, however, think these results are sufficient to recommend ditching NAS, and definitely this research does not suggest that drinking sugary drinks is more healthful than drinks sweetened with NAS.

The scientific community is already starting to pick over the results of this study, and dampen public reaction by putting it into perspective. First, the majority of this work was done in mice, who have a different glucose metabolism, diet, and tolerance than humans. The small study with 7 human subjects is very preliminary, and far from sufficient to conclude that the mice data will be applicable to people.

The Science Magazine article points out that the study was published in a basic science journal, and that a clinical science journal would probably have been much more critical of their clinical speculations.

Another potentially serious criticism is that the researchers combined saccharin, sucralose, and aspartame data. It seems highly unlikely that three very different molecules would all have the same effect on gut microbiota. It’s possible that what the researchers are seeing is isolated to saccharin alone, which the research focused on. Earlier trials used aspartame, which had a smaller effect so the researchers switched to saccharin. The Science Magazine article reports:

“The authors are confounding their conclusions by addressing all these noncaloric artificial sweeteners together,” says Brian Ratcliffe, a nutrition researcher at Robert Gordon University in Aberdeen, U.K. That’s why the title of the paper, “Artificial sweeteners induce glucose intolerance by altering the gut microbiota,” is misleading, he says. “I cannot believe the journal allowed that title.” Still, he says, the data “certainly does suggest that there is something more that needs to be explored about saccharin.”

If this effect is unique to saccharin, that would also explain the disconnect with other data focusing on the consumption of diet soft drinks, which use aspartame and sucralose. A large European epidemiological trial published last year and involving cohorts with >10,000 subjects found an association between drinking sugary drinks and Type II diabetes. It also found an association with drinking NAS containing drinks, but this association vanished when controlled for energy intake and BMI. In other words, people drink diet soda because they are overweight, not the other way around.

A 2011 study of >40,000 men found:

Sugar-sweetened beverage consumption is associated with a significantly elevated risk of type 2 diabetes, whereas the association between artificially sweetened beverages and type 2 diabetes was largely explained by health status, pre-enrollment weight change, dieting, and body mass index.

The two large epidemiological studies are likely to be more reliable than the 7-subject arm of the recent Nature study.


The current study is rigorous and interesting, as far as it goes. It suggests that saccharin has the potential to alter the gut microbiota of mice and has some relationship to glucose metabolism. It would be interesting to work out the various mechanisms involved with future research.

The bottom line that is frequently being communicated to the public about this study, however, is not supported by this data, is misleading, and is likely to lead to poor health decisions.

It is unclear if these results apply to sucralose or aspartame (and therefore diet soft drinks), and it is further unclear if they apply to humans.

Meanwhile, other studies, some of which are massive epidemiological studies, show a clear connection between consuming sugar-sweetened drinks and type II diabetes, and no connection to drinking NAS sweetened drinks. This study should not motivate anyone to abandon their diet drinks for sugar-sweetened drinks, but the way the study is being reported may do just that.

15 responses so far

15 thoughts on “Artificial Sweeteners and Diabetes”

  1. MikeB says:

    Thanks for this. I heard this study discussed on NPR last night, and I had my own inarticulate musings, as a lay person and writing instructor interested in how science is reported in the media.

    The more I experience, the more I’m tempted to promulgate Mike’s Axiom: If it’s “science” reported in mainstream news sources, it’s probably wrong (or at least misleading).

  2. MikeB says:

    I forgot to mention that one of the “scientists” interviewed by NPR admitted that after reading the study he avoided artificially-sweetened drinks.

  3. “All of this strongly suggests that consuming NAS alters the but bacteria”

    I suspect that this is a typo: Likely the “b” in “but” should be a “g”, but I suppose it’s possible that but is missing an extra “t” at the end. 🙂

  4. ScubaSharky says:

    Is it just me, or does the choice to include only artificial non-nutritive sweeteners expose a potential bias of the researchers? If we’re testing the effects of a wide range of chemically dissimilar sweeteners, why not throw steviol glycoside in for good measure? It just seems suspect to me that they’ve targeted three NASs which are the frequent subjects of hysterical internet memes.

  5. RC says:

    ScubaSharky, I was thinking the same thing.

  6. Debbie_M says:

    I read the paper in question. Their data is weird. I looked at their results that supposedly show glucose intolerance in the mice that were fed NAS. I don’t see how they could define the results as showing glucose intolerance. Furthermore, all the mice showed an increase in glucose tolerance once their gut biota was eradicated, not just the mice who were fed NAS. The difference in the AUC measurements was greatest between the mice who had gut bacteria and the ones that had none. This difference was much greater than the (supposed) difference between the NAS and control mice. The authors didn’t discuss this at all. They said that glucose tolerance decreased when NAS treated bacteria were introduced into sterile mice – but they didn’t do a control where they treated the mice with non-NAS treated bacteria.

    They didn’t say how the glucose tolerance test was administered, but they showed a graph of both the blood glucose levels and AUC (area under the curve of glucose levels). I thought it was really strange that the starting glucose levels of the mice were quite varied. Looking at the test results, I could see some differences in the peak glucose levels, but at the end of the test, all the glucose measurements had become the same. I’m a chemist, not a biologist or physician, but don’t glucose intolerant individuals show a higher baseline glucose level? If so, then why would all of the glucose levels go to the same level at the end of the test?
    I apologize for not being more specific. I only got to skim the paper yesterday and I don’t have access to it here at home.

  7. Debbie_M says:

    Corrections: the authors did have a control when they introduced bacteria into sterile mice and it does look like the glucose levels didn’t all return to the same levels in the glucose tolerance tests.

  8. rezistnzisfutl says:

    Just heard a radio interview on NPR on Science Friday with the study’s primary author. Ira asked many questions about the study. Unfortunately, no other scientists participated in the piece. The author went on to make several speculative statements about the role of gut bacteria and their connection with the brain and overall health, going as far as saying how artificial sweeteners may be responsible for a myriad maladies, including diabetes, heart disease, and even cancer. At least he acknowledged the small sample size of human participants, week long duration, and the fact that saccharin was the only substance they really studied and that there are many other forms of artificial sweeteners that should be studied.

    It would be so nice if science reporting included interviews with other scientists.

  9. Argh, really annoying that they mixed Aspartame in with this. I’m actually someone who could actually be at risk from this due to medical history (an IBD in which sparse diversity of gut bacteria is implicated, plus both overweight and with a history of corticosteroids leading to an increased risk of Diabetes Type II anyway) and there’s only a single product I consume regularly that contains artificial sweetners (a no-added-sugar “High Juice” cordial – albeit in considerably more dilute concentrations than recommended); fine if only Saccharin is implicated (as it doesn’t contain any) but not so fine if Aspartame is implicated (as it does.) I guess, for me at least, this is going to have to go into the “interesting, but not really adequately specifically proven for me to take action yet” pile.

  10. superdave says:

    my take is this. there is basically no risk to switching to water from diet soda so if there is even a slight risk to the diet soda, you might as well at least try to cut back while the science settles. If 5 years from now this study is overturned, you have not lost much and saved money on drinks.

  11. BillyJoe7 says:

    The trouble with that policy is that if you apply this to everything you eat or drink you’re not going to be eating or drinking much of anything. Going with the evidence as it stands at present and balancing the risks and benefits works better as a policy

  12. RC says:

    @Superdave – “there is basically no risk to switching to water from diet soda”

    This is basically Pascal’s wager – there’s nothing to lose by believing in God, and everything to gain, and nothing to gain by not believing, and everything to lose.

    But there is something to lose, and there is risk – there’s the risk that I spent 5 years being influenced by a poor study and was deprived of something I love – diet coke.

  13. blynch99 says:

    Nice post. Just a comment about the statement “It seems highly unlikely that three very different molecules would all have the same effect on gut microbiota”. In fact there are hypothesis and experimental evidence that could account for this. There is a growing recognition that sweet receptors, canonically found in the mouth, are present in the gut, on pancreatic beta cells, and in other sites: they are activated by all of the artificial sweeteners almost by definition. A couple of article of interest below.

  14. DLC says:

    While it’s an interesting study, it has nothing to do with obesity in humans. It seems to me to be much on a par with “it kills cancer cells in a petrie dish” studies — interesting but little in the way of real-world application.

  15. ebohlman says:

    RC: Agreed. In the absence of established risk, “because I enjoy it” is a perfectly good reason to do something (I’ve been known to joke that the difference between a conservative and a liberal is that the conservative thinks that people should screw only for procreation and the liberal thinks that people should eat only for nutrition).

    DLC: When any research results regarding nutrition make it into the popular media, they inevitably wind up being framed in terms of weight.

Leave a Reply