There has been quite a bit of coverage of a recent Nature paper reporting a link between artificial sweeteners and high blood glucose levels (1) – an important finding if true, as high blood glucose levels are a step towards insulin resistance and type 2 diabetes. The study was carried out mostly in mice and was found to be mediated by the gut microbes. The paper isn’t open access so you’ll need $32 to read it if you don’t have a subscription to Nature.
The authors report that consumption of artificial sweeteners changes the microbes present in the gut of mice (and 4 out of 7 healthy human volunteers), with some microbes becoming more abundant and others less so or disappearing altogether, and that this is correlated with a rise in blood glucose levels. The effect disappears when mice are treated with antibiotics. The effect can also be transferred to animals who haven’t been fed artificial sweeteners by giving those animals a faecal transplant from animals with the altered gut microbiome.
The researchers first tested the effect of giving groups of mice access to drinking water spiked with three different artificial sweeteners currently used by people: saccharin (marketed as ‘Sweet’N Low’ in the USA), sucralose (marketed as ‘Splenda’) and aspartame (marketed as ‘NutraSweet’ and ‘Canderel’). The results of the blood glucose tests in mice are shown in Figure 1b of their paper, reproduced below. Despite not being widely used as an artificial sweetener in processed food and drinks anymore, saccharin was the sweetener the researchers chose to do the rest of their studies with, including feeding it to their 7 healthy volunteers.
So why did the researchers choose saccharin for their studies?
Let’s take a closer look at Figure 1b. It’s a graph that shows the area under curve of the data for the blood glucose tests for the mice. The higher the value, the higher or more prolonged the blood glucose levels. Each individual symbol is the value for an individual animal.
The first three groups of animals are the controls – the first were fed plain water (black circles), the second sucrose (black triangles) and the third glucose (black squares). The sucrose and glucose groups are an attempt to control for the sweet taste of the water containing artificial sweetener. The next three groups (blue) are the animals that have been fed the artificial sweeteners: saccharin (blue circles), sucralose (blue squares) and aspartame (blue triangles). The red and grey groups have been fed different antibiotics to knock down their gut microbes in addition to being either control animals or having had the artificial sweeteners.
What is immediately obvious to me is that there are about 4-5 animals in the saccharin and sucralose groups who have very high area under curve values – they look quite different from the rest of the animals in their groups. I think we can call these outliers. That doesn’t mean the effect isn’t real – just that they might not be quite representative of the rest of their cohort. When looking at the data for the rest of the animals, and for the aspartame group, there is quite a lot of overlap between the control groups and the groups fed the artificial sweeteners. I’d love to have access to this data to reanalyse it because I’m surprised the groups are significantly different from the controls, especially for aspartame. But this does explain why they chose saccharin – it showed the biggest difference when compared to the controls.
So could something else be going on with the saccharin group?
One of the things the researchers did was put the animals in ‘metabolic cages’. This allows the researchers to monitor the food and drink intake of the animals as well as how active they are. In their paper the authors state:
“Metabolic profiling of normal-chow or HFD-fed mice in metabolic cages, including liquids and chow consumption, oxygen consumption, walking distance and energy expenditure, showed similar measures between NAS- and control-drinking mice (Extended Data Fig. 3 and 4.)”
Let’s have a look at Extended Data Fig. 3, shall we?
“Similar measures”? It looks to me like the saccharin group (and glucose group) drank more and ate less than the other groups. The saccharin group also looked like they expended more energy. Hmmmm. I think ‘similar measures’ is stretching the truth a little.
I’d love to know what the reviewers said about this paper. There’s far too much cherry-picking in it for my liking.