September 7, 2022

expert reaction to study suggesting possible link between artificial sweeteners and heart disease

A study published in The BMJ looks at artificial sweeteners and risk of cardiovascular diseases.

Prof Nita Forouhi, MRC Epidemiology Unit, University of Cambridge, said:

“This study adds importantly to a growing body of evidence on potential cardiovascular health harms that were previously based largely on links between beverages that are artificially sweetened, this time extended to artificial sweeteners in the total diet, in foods, drinks and table-top sweeteners.

“The authors have acknowledged that this research is not interventional but observational, and made demonstrable attempts to reduce the well-known limitations of observational research through their analytical approaches. For instance, they accounted for a number of risk factors that were distributed unequally between low and high consumers of artificial sweeteners and performed a number of sensitivity analyses to test the robustness of their findings. Since artificial sweeteners may be related with weight status, the researchers accounted for baseline body mass index, weight loss diet and weight change during follow up, which did not change the results substantially. The statistical models did not account for the observation that the lowest intake of sugary drinks was in non-consumers of artificial sweeteners while the highest sugary drinks intake was in higher consumers of artificial sweeteners. However, the utility of artificial sweeteners as substitutes for added sugars was tested and found not to have any risk reduction for cardiovascular disease. Some research limitations remained, such as this research included mostly women (80% of participants), so the findings in men remain under-studied, the interpretation of the hazard ratios (relative risks) was complex because the results were modelled per log10 of artificial sweetener intake in mg/day+1, a dose response relationship was not clear, the magnitude of associations was modest, confounding cannot be fully excluded and causality is not established.

“This current research is not definitive for public health policy making and further evidence is needed to confirm a possible link between artificial sweeteners and cardiovascular disease. However, we should not ignore these findings. While further improved research should be conducted, in the meantime the key take-away message is that consumption of artificial sweeteners is unlikely to have health benefits for cardiovascular disease.”

Helena Gibson-Moore, Nutrition Scientist, British Nutrition Foundation, said:

“There can be conflicting messages around the safety and health effects of artificial (or low calorie) sweeteners, which can cause confusion among consumers. This is often because messages are based on evidence from individual studies, rather than the totality of the evidence. Health authorities that have looked at the safety of additives such as low calorie sweeteners review the totality of evidence at the time, and provide a robust, evidence-based evaluation. Low calorie sweeteners have been critically evaluated by international authorities such as the European Food Safety Authority (EFSA) and United States Food and Drug Administration (FDA) and have been approved as safe for use in a range of food and drink products. 

“This study is an observational study, and although well designed with a good-sized population group, there are some limitations to the study which the authors highlight.  The study can only show an association and can’t establish true cause and effect, nor can the researchers rule out the possibility that other unknown (confounding) factors might have affected their results. The study population was French and mainly female (80%) so caution is needed to generalise these results to other countries and population groups. Higher consumers compared with non-consumers, tended to be younger, have a higher body mass index, were more likely to smoke, be less physically active, and to follow a weight loss diet. They also had lower total energy intake, and lower alcohol, saturated and polyunsaturated fats, fibre, carbohydrate, fruit and vegetable intakes, and higher intakes of sodium, red and processed meat, dairy products, and beverages with no added sugar. Although the researchers took account of these differences in their analyses this doesn’t guarantee that the confounding factors have been entirely accounted for. Cardiovascular diseases are complex multifactorial diseases that typically develop over a long period of time with dietary risk factors such as low fibre and fruit and vegetable intakes, higher sodium intakes, and other factors such as high body mass index and lower physical activity levels, which are also observed in the higher consumers of low calorie sweeteners, so it may be that the higher consumers had an overall poorer diet and lifestyle. The results of the study are interesting, however they are unlikely to indicate a need to change average consumption habits of low calorie sweeteners with regards to cardiovascular risk. 

“To help reduce the risk of cardiovascular disease it’s recommended we consume a healthy, balanced diet with plenty of fibre rich foods such a wholegrains, fruit and vegetables and pulses, cut down on saturated fat and replace with unsaturated fats, have 2 portions of fish a week (one oily type such as mackerel, salmon or sardines), watch our salt intake (less than 6g a day), as well as be a healthy weight and lead a healthier lifestyle.  

“It’s important that the research on the long-term effects of low calorie sweeteners on disease risk such as cardiovascular diseases continues to be reviewed. The World Health Organization is currently drafting guidelines on the use of low calorie sweeteners, and a public consultation has recently been conducted, but there is no indication yet of when the guidelines may be published.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“This is generally a good, careful study. It uses appropriate statistical methods and reports its findings appropriately. Well over 100,000 French people, who were already part of a wider set of studies, were followed up for nine years on average. Careful records of what they ate and drank over a period of 24 hours were made on repeated occasions, many other factors about their lives were recorded, and the researchers also recorded when any of them had a new diagnosis of a cardiovascular disease event (heart attack, angina, stroke, and similar events).

“The recording of diet, and hence the measurement of each participant’s consumption of artificial sweeteners, seems to me to be much more careful and detailed than one sometimes sees in this kind of study. For example, one sees studies where consumption is measured only once at the beginning of the research project, so there’s no chance of picking up changes in diet that might occur during several years of follow-up. In this new study, diet was recorded repeatedly at regular time intervals. All sources of artificial sweeteners in the diet were recorded, unlike some previous studies that looked only at consumption of artificially sweetened drinks. Also, the statistical analysis is detailed and appropriate. But, inevitably, care needs to be taken with the interpretation of the findings, particularly in terms of what might be causing what.

“This is an observational study. The participants were not told what to eat and drink; the researchers simply observed and recorded what they ate and drank. So, inevitably, there are differences between people who consumed different amounts of sweeteners, apart from the differences in their consumption of artificial sweeteners. Many such differences are shown in Table 1 of the research paper.

“Every one of the factors shown there has a statistically significant difference between the three groups of participants (those who consumed no artificial sweeteners, those who did consume them, but less than the average (median) amount, and those who consumed more than average). But that’s partly because, with so many participants, even small and possibly unimportant differences will show up as statistically significant, because they are too big to be put down to the workings of chance alone. But some of the differences are reasonably substantial –  for instance those who consumed a lot of sweeteners were more likely than non-consumers to have high blood pressure, less likely to smoke, a bit less likely to have high physical activity levels, were more likely to have gone on a weight loss diet, and on average the high sweetener consumers ate quite a lot less whole-grain food than the non-consumers. Interestingly, the participants who consumed higher levels of artificial sweeteners drank, on average, about a fifth more of sugary drinks daily than did those who consumed no sweeteners. Since more than half of the overall consumption of sweeteners in the study came from soft drinks with no added sugar, it looks as if many of those who consumed a lot of artificial sweeteners just liked sweet drinks in general, whether sweetened by sugar or by something else. Some of these other factors have been implicated as risk factors for cardiovascular disease.

“So the problem is that any differences in the rate of cardiovascular disease between those who don’t consume sweeteners and those who consume a lot could be caused, in whole or in part, by differences in one or more of these other factors, and not by the sweetener consumption. The researchers were perfectly aware of that possibility and made appropriate statistical adjustments to allow for differences in many other factors, either in their main results or in other sensitivity analyses that they carried out to check their overall findings. But the snag in any observational study is that one can never be sure that all the possible factors have been taken account of in this way, and anyway no adjustment can be made for a factor on which the researchers have no data. That’s why both the press release and the research paper point out that the study can’t entirely establish that the differences in cardiovascular disease rates are caused by differences in consumption of sweeteners. It’s correlation, not necessarily causation.

“That’s why the researchers call for further large-scale research to confirm their findings, and to look at possible biological mechanisms that might explain how sweetener consumption could lead to heart disease or strokes. And I’m certainly not saying that consumption of artificial sweeteners can have no cause-and-affect relationship with cardiovascular disease, only that this study on its own can’t show that the association must be one of cause and effect.

“That said, in my view the researchers did a pretty good job, as far as they could within a study like this, to try to investigate cause and effect. As well as the statistical adjustments and the sensitivity analyses, they made a good attempt to deal with another bugbear of cause and effect, that can arise in this kind of study. People sometimes change their diet because of a health condition. For instance, people may decide to avoid sugar because they believe it could make a pre-existing health condition worse, so it’s possible that some aspect of people’s health could cause differences in the amount of artificial sweetener they consume, if people substitute artificial sweetener for sugar on health grounds. That’s an example of what’s called reverse causation – the main interest is in whether consuming more sweeteners may lead to ill health, but there’s a possibility that ill health could lead to the consumption of more sweeteners. The researchers made a statistical adjustment (in a sensitivity analysis) for whether participants had been on a weight loss diet. They also excluded from their analysis anyone who already had cardiovascular disease or diabetes before the study began, and indeed they also excluded people who had cardiovascular disease diagnosed during the first two years of follow-up, in case they might have changed their consumption because of some health issue that was leading up to CVD but had not been diagnosed as CVD at the start of the study. All of this can’t entirely rule out the possibility of reverse causation, but it goes quite a long way in my view.

“It’s also got to be borne in mind that some other previous studies have, in various ways, found associations between sweetener consumption and health. Issues of determining cause and effect do remain, because those other studies were generally observational too, and of course there’s also evidence of effects of sugar consumption on health. We’re never going to be able to run randomised clinical trials, where people are randomly allocated to different diets which they must then follow for many years – it’s just not practicable. So evidence on health effects of sweeteners will have to be put together from observational studies and other types of research. This new study is an important contribution to the developing picture.

“Another potential limitation is that the data are based on findings in a specific group of French people, who volunteered for the NutriNet-Santé study which was set up to investigate associations between nutrition and health. As often happens with such studies, the pool of participants doesn’t look all that similar to French society as a whole – for instance, of those in this particular new study of sweeteners, four in five participants are female, and the levels of education and professional status are higher than average. And, again unsurprisingly because they are interested in health and nutrition in order to have volunteered, their health behaviours (including diet) tend to be different from average. That shows up, for instance, in that nearly two in every three participants in this sweetener study did not consume any artificial sweeteners at all. So we can’t be sure that everything works in the same way in people from this cohort as in French, or UK society as a whole. Maybe it does work in the same way, or maybe there are small or even large differences.”

Prof Naveed Sattar, Professor of Metabolic Medicine, University of Glasgow, said:

“Whilst this topic is clearly worthy of study, the present observational study cannot answer the question posed.  This is because there are clearly major differences in many characteristics of people who consume artificial sweeteners compared to those who do not consume any.  This then means that however one tries to adjust the findings for such differences, there remains a very strong chance of what is known as ‘residual confounding’ – i.e. key differences that really account for findings may be factors not measured or poorly measured but not included in analyses or fully captured.  Only randomised trials can get to the truth. Also, the writing in this paper is far too strongly suggestive of a causal link between artificial sweeteners and cardiovascular when this type of evidence is weak and potentially flawed – something the editors should have picked up on.  One cannot say “increased” risk when it is observational data.    There are multiple small randomised trials assessing the impact of artificial sweeteners on risk factors such as weight and glucose and even these are contradictory, with some suggesting harm, others no effect, and some even suggesting benefits.  This means we can draw no real conclusion from the available evidence.  To do this important area justice requires longer term and larger randomised trials and food / governmental agencies should fund these to get closer to the truth.  Observational studies are simply too confounded to make any meaningful conclusions on this question and public health guidance requires far higher quality evidence from trials.”

Dr Duane Mellor, Registered Dietitian and Senior Teaching Fellow, Aston Medical School, Aston University, said:

“This study reports data from a large number of people (over 100,000) who were followed up for an average of 9 years. Unlike previous publications of this Nutri-Net Sante cohort, it does not explain how the 103,388 participants were selected from the overall population sample of 128, 343, which means for an unexplained reason around a fifth of participants were excluded. 

“It is also a shame that the data in a paper looking at the effect of the same artificial sweeteners in the exact same population and associated risk of cancer published in March this year was not combined with this analysis to give an overall assessment of risk associated with artificial sweetener consumption on diet related chronic diseases. Some commentators have criticised other researchers when they have separated of data from the same population into different papers as being like salami slicing, this may not be the case here, but having said that, it is far more use to public health policy makers and the public if data is presented in fewer papers than split across many.

“As the methods are identical to the paper published by this group of researchers in March the limitations are very similar, in that it does not acknowledge that the dietary pattern of those who consume more artificial sweeteners appeared to contain more process foods than those who consumed none. Also higher consumers of artificial sweeteners consumed more sugary drinks and soft drinks in general, with sugary drinks having previously been linked with an increased risk of cardiovascular disease.

“The wording of the press release needs to be considered carefully as the definition of coronary heart disease included not only heart attacks but also acute coronary syndrome, angioplasty (procedures to widen the arteries supplying the heart)  and angina and cerebral vascular disease included both strokes and mini-strokes (known as transient ischaemic attacks). These are all important forms of cardiovascular disease, but these are not all the same as heart attacks and strokes. 

“It is interesting that the effect did not seem to be consistent across all types of sweeteners, with the combined cardiovascular disease risk only being significantly associated with total sweetener intake, however for those related to the range of heart diseases it was only significant for acesulfame – K and sucralose but not for aspartame, with cerebrovascular disease the association was significant for overall aspartame intake and aspartame only. The variation in response to different sweeteners, could be a reflection of the relatively small number of cases of cardiovascular disease, or that the association may be due to other factors which could not be controlled or were not controlled for example sugary drink intake. It is also worth noting that the dietary recall method, although repeated was used to estimate artificial sweetener intake, typically sweeteners are blended, the most common blend being aspartame and acesulfame-K this makes the finding that there is an association between one and not the other somewhat surprising. This could be something that is different to how sweeteners are produced in France, or the product of factors that could not be controlled for.

“As with the very similar study published by the same French group, this study, again reminds us, that when we try to improve our diet, and maybe lose weight, the easier option to switch to reduced sugar and artificially sweetened foods (which can be low fat too), may not be the best way. Instead, it is probably much better to take a bit of time to plan a sustainable meal plan based on minimally processed foods and rich in vegetables, fruit, beans, peas lentils, wholegrain and nuts or seeds (or a Mediterranean type diet) is a far better way to maintain and improve health including helping to reduce our risk of cardiovascular disease.”

‘Artificial sweeteners and risk of cardiovascular diseases: results from the prospective NutriNet-Santé cohort’ by Charlotte Debras et al. was published in The BMJ at 23:30 UK time on Wednesday 7 September 2022.

DOI: 10.1136/bmj-2022-071204

Declared interests

Dr Duane Mellor: “I have previously worked with the International Sweetener Association.”

Prof Naveed Sattar: “I have consulted for many companies that make diabetes and cardiovascular drugs.  I have also been involved in multiple trials of lifestyle for prevention and remission of diabetes.  I do not have any interaction with artificial sweetener studies.”

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  My quote above is in my capacity as an independent professional statistician.”

Prof Nita Forouhi: “None.”

Helena Gibson-Moore: “Funding to support the British Nutrition Foundation’s charitable aims and objectives comes from a range of sources including membership, donations and project grants from food producers and manufacturers, retailers and food service companies, contracts with government departments; conferences, publications and training; overseas projects; funding from grant providing bodies, trusts and other charities. Further information about the British Nutrition Foundation’s activities and funding can be found at http://www.nutrition.org.uk/aboutbnf/”