Experts point out the importance of study design in research about low calorie sweeteners and obesity

A commentary on the study by Azad et al.

Highlights:

  • In a recent publication by Azad et al., claiming that low calorie sweeteners may be linked to increased body weight and cardiometabolic risk, the data from clinical trials do not agree with outcomes from observational studies.
  • Study design is a critical aspect in nutrition research. While observational studies are useful in examining possible links or relationships, these associations should be further examined in randomised clinical trials, which is the gold standard in research.
  • Taken together, the available evidence supports the intended benefit of low calorie sweeteners as being helpful in reducing overall energy intake, when used to replace sugar.

From time to time you may be crossing stories online casting doubts on low calorie sweeteners’ effect on body weight without necessarily being able to critically judge these assertions, especially if you don’t have a nutrition or medical background. Similarly, when a new study by Azad et al.1 supporting that routine intake of non-nutritive sweeteners may be associated with increased BMI and cardiometabolic risk based on observational data, led to sensational headlines worldwide in mid-July, it was probably difficult for most people to read behind the headlines.

Indeed, media titles are often misleading and the articles frequently miss to clarify that epidemiological studies, due to their observational nature, cannot provide evidence of causation, but only to examine the association between the factors they are meant to study. Therefore, in this article we present comments from different experts who carefully reviewed the study by Azad et al. aiming to provide a more critical view and interpretation of the findings.

The conclusion by Azad et al. seems unjustified, experts support

In a response published online2 in the Canadian Medical Association Journal by Dr John Sievenpiper and his colleagues at the University of Toronto in Canada, the authors point out that “the conclusion by Azad et al. that the evidence does not support the intended benefits of low calorie sweeteners seems unjustified” and express their concerns that important methodological considerations were overlooked.

As the intended benefit of use of low calorie sweeteners is to help reduce caloric intake by replacing sugar with low/ no calories while maintaining the desired sweet taste, one would expect that the study would aim to examine the effect of this caloric displacement on body weight. However, based on the selection criteria of the meta-analysis of randomised controlled trials (RCTs), the study by Azad et al. included trials which used as a comparator water, or placebo, therefore failing to account for the nature of the comparator. On this basis, the study cannot provide evidence about the intended benefit of low calorie sweeteners.

Furthermore, Dr Sievenpiper and his colleagues emphasize the limitations of observational studies, which, by design, cannot prove ‘cause and effect’. Despite the claims of Azad et al., the authors of this response note that the prospective design of cohort studies does not limit residual confounding and cannot rule out the possibility of reverse causality, meaning that people with higher BMI may be choosing low calorie sweeteners more often in their effort to reduce sugar and caloric intake and manage their body weight. Therefore, they conclude that Azad et al. accorded too much weight to prospective cohort studies in their conclusions and that ‘These uncertainties suggest that the current estimates from prospective cohort studies are not trustworthy and new research is likely to have an important impact on their direction, magnitude, and precision.’

Study limitations should be taken into consideration in interpreting the findings

As an epidemiologist, Professor Carlo La Vecchia, University of Milan, Italy, highlights that ‘the data from observational (cohort) studies are subject to indication bias or reverse causation, and – not surprisingly – show opposite effects on BMI’ [compared to RCTs]. Furthermore, he points out the few data used in the meta-analysis of RCTs to detect the differences in mean BMI (n=120) and in body weight (n=345) between low calorie sweeteners’ consumers and control groups.

Professor Peter Rogers, University of Bristol, UK, who has recently conducted a thorough systematic review and meta-analysis3 on this topic, noted that: ‘By using a 6-month duration criterion, the authors [in Azad et al study] have excluded several relevant intervention studies. Oddly, their intervention studies are a mixture of those comparing LCS with water and those comparing LCS with sugar. Studies with children are also excluded, so the very clear result from de Ruyter et al (2012)4 – the best powered study to date – is not discussed.’ Contrary to the conclusions of Azad et al study, the extensive systematic review of the literature and meta-analysis led by Prof Rogers using data from 56 short-term trials reporting 218 comparisons, 13 sustained intervention trials and 10 prospective cohort studies, found that, overall, the balance of evidence indicates that use of low calorie sweeteners (LCS) in place of sugar, in children and adults, leads to reduced energy intake and body weight.

Similarly, the Registered Nutritionist Sigrid Gibson, also author of several scientific publications on the topic of low calorie sweeteners, agrees and notes that “the comparison selected by Azad in the 7 RCTs was LCS versus water or placebo, not LCS versus sugar or sugar-sweetened beverages. Therefore the paper’s conclusion, that “Evidence from RCTs does not clearly support the intended benefits of non-nutritive sweeteners for weight management” is misleading because it did not address their intended use, which is to replace sugar. That LCS have similar impact on bodyweight to water or placebo is exactly as expected given their negligible calorie content. The meta-analysis of the trial data shows the overall effect of LCS was a very small, nonsignificant decrease in weight or BMI. The papers conclusions seems to give more weight to the cohort data, which are observational and prone to confounding.

Despite media headlines, low calorie sweeteners should be viewed exactly for what they are and for what the science supports about their use; they are sweet-tasting ingredients with no, or virtually no, calories added in foods and drinks in tiny amounts in place of sugar. While low calorie sweeteners are not a magic bullet for weight loss, experts conclude that, taken together, the scientific evidence supports the intended use of low calorie sweeteners, in place of sugar, in order to help reduce overall energy intake and hence help in weight control.

In times when obesity rates are reaching epidemic proportions globally and all available dietary strategies for calorie reduction and weight loss are important, people deserve clear and evidence based information in media about low calorie sweeteners’ impact on body weight.

To read the ISA statement in response to the study by Azad et al., please click here.

  1. Azad M., Abou-Setta AM., Chauhan BF., et al. Nonnutritive sweeteners and cardiometabolic health: a systematic review and meta-analysis of randomised controlled trials and prospective cohort studies. Canadian Medical Association Journal, July 2017; 189: E929-39
  2. Sievenpiper JL, Khan TA, Ha V, Viguiliouk E, Auyeung R. The importance of study design in the assessment of non-nutritive sweeteners and cardiometabolic health. A letter in response to Azad et al study in CMAJ. Available online: http://www.cmaj.ca/content/189/28/E929/reply#cmaj_el_733381
  3. Rogers PJ, Hogenkamp PS, de Graaf C, et al. Does low-energy sweetener consumption affect energy intake and body weight? A systematic review, including meta-analyses, of the evidence from human and animal studies. Int J Obes (Lond) 2016; 40: 381-94.
  4. de Ruyter JC, Olthof MR, Seidell JC, Katan MB. A trial of sugar-free or sugar-sweetened beverages and body weight in children. N Engl J Med 2012; 367: 1397–1406