Two mixes of food additives that are very common in ultra-processed foods produce a ‘cocktail effect’ that may increase the risk of type 2 diabetes. The evidence emerges for the first time in a French study (Payen de la Garanderie et al., 2025) published in Plos Medicine, which examined the impact of additive mixes on over 108,000 adults participating in NutriNet-Santé (2009-2023), the prospective cohort study conducted in France to study the relationship between diet and health.
Ultra-processed foods, a reservoir of additives
Ultra-processed foods (UPFs) represent a significant and growing share of the modern Western diet. Their intake has already been associated with an increased risk of chronic non-communicable diseases (NCDs) and premature mortality from all causes.
The French National Agency for Health and Food, Environmental and Occupational Safety, ANSES, highlighted in 2024 among the risk factors of UPFs both poor nutritional profiles and high content of food additives, substances that alter their composition and can lead to the formation of substances that are risky for health (e.g. nitrosamines).
Numerous studies, among others, have associated the intake of individual food additives with various risks for human health, including intestinal dysbiosis, inflammation and metabolic alterations.
Food additives, the cocktail effect
Food additives in the EU are subject to scientific evaluation by EFSA (European Food Safety Authority) and subsequent authorization by the European Commission (Reg. EC/1333/2008). Yet:
- the risk analysis that leads to the determination of acceptable daily intakes (ADI) does not consider the cumulative and synergistic effects of additive mixtures consumed by various population groups in the overall diet;
- an adult can ingest up to 10 kg of food additives per year, according to the study by Chazelas et al. (2021) published in Scientific Reports.
Only one previous study on humans (McCann et al., 2007) has explored the potential health impact of a single mixture of food additives – the so-called ‘Southampton colorants’ – on children’s health. Which recorded symptoms of hyperactivity, after exposure for 6 weeks to a mix of azo dyes and a preservative, still authorized in the EU (although subject to a specific warning on the label):
– carmoisine (E122),
– sunset yellow (E110),
– tartrazine (E102),
– ponceau 4R, cochineal red A (E124),
– allura red AC (E129);
– sodium benzoate (E211).
Study methods
The study by Payen de la Garandie et al. (2025) is the first to evaluate the effect of food additive mixtures to which the general population is actually exposed, in a diet that includes the intake of ultra-processed foods. The research used the following methods.
The NutriNet-Santé cohort
The study in question involved 108,643 French adults, followed for an average of 7.7 years. The participants – average age 42.5 years, 79.2% women – completed five questionnaires on:
- lifestyle and sociodemographic data. Date of birth, sex, level of education, professional occupation, smoking status, number of children;
- health status. Personal and family history, medical treatments;
- eating habits. The data, collected every 6 months, contain logs of food consumed in 24 hours for three non-consecutive days (two weekdays and one weekend), randomly assigned over a 2-week period. The diary includes brand and commercial name of food;
- anthropometric data of participants (height, weight)
- physical activity level. Seven-day assessment using the International Physical Activity Questionnaire, IPAQ.
Information on ethnicity and religion – although of potential interest for research, especially with regard to the kosher diet, where the use of a number of food additives in foods is not permitted – were not recorded as they are considered ‘sensitive data’ by French law.
Dietary data collection
Participants’ food records provided researchers with data to calculate separate daily intakes for food additives, nutrients, and energy.
The NOVA (Food Processing Index) classification was applied to identify ultra-processed foods (UPFs) and calculate their contribution to overall energy intake.
Food additive intake
Examination of individual food records and use of food composition databases (Observatoire de la Qualité de l’Alimentation, Open Food Facts, Global New Products Database, EFSA, and General Standard for Food Additives) allowed us to compile a list of the 269 food additives consumed by study participants.
Only additives consumed by at least 5% of the cohort were included in the mixture modeling.
Type 2 diabetes assessment
The incidence of type 2 diabetes among study participants was assessed using a multi-source approach:
– participant reporting, asking them to report any health events, medical treatments and tests that occurred during the follow-up period;
– information in the national health database, to which the NutriNet-Santé cohort was linked.
Deaths of study participants, and their causes, were recorded via linkage with the French National Mortality Registry (CépiDC).
Study results
The impact of ultra-processed foods on the diet
The examination of participants’ food records showed that ultra-processed foods (classified as NOVA 4) represented on average 33.8% of the daily energy intake.
A total of 75 food additives were consumed by at least 5% of participants and were therefore included in the additive mixture model to compare with the incidence of type 1 diabetes.
Identification of food additive mixtures
Using non-negative matrix factorization (NMF) – useful for reducing a very large data set into representative attributes – the researchers identified 5 main additive mixtures, based on real consumption and composed as follows:
– mixture 1. Related to cakes and biscuits, it contains mainly sodium carbonates, diphosphates, glycerol, ammonium carbonates, potassium carbonates and sorbitol;
– mixture 2. Typical of broth, milk-based desserts, fats and sauces, it is mainly composed of modified starches, pectin, guar gum, carrageenan, polyphosphates, potassium sorbates, curcumin and xanthan gum;
– mixture 3. No uniform group of foods is specifically related to this mix. The additives that are part of it are in fact found in very different foods: from table salt to biscuits and energy drinks. The mix is made up of magnesium carbonates, riboflavin, alpha-tocopherol and ammonium carbonates;
– mixture 4. Like mixture 1, it is linked to salty snacks, cakes and biscuits. It is composed of ammonium carbonates, sodium carbonates, diphosphates, alpha-tocopherol, mono and diglycerides of fatty acids, magnesium carbonates and lecithins;
– mixture 5. It is easy to identify sugary soft drinks or drinks sweetened with non-caloric artificial sweeteners (‘sugar-free’, ‘zero’ drinks) as the target food group of this mix. The additives that compose it are citric acid, sodium citrates, phosphoric acid, sulphite ammonium caramel, acesulfame K, aspartame, sucralose, gum arabic, malic acid, carnauba wax, paprika extract, anthocyanins, guar gum and pectin.
The two food additive mixtures that promote diabetes
Over the 7.7 years of follow-up, 1,131 cases of type 2 diabetes were detected among the study participants.
Regardless of the nutritional quality of the diet (intake of sugars, saturated fats, energy, alcohol, etc.) and after adjustment for a wide range of potential confounders, two food additive mixtures were positively associated with an increased incidence of type 2 diabetes:
– mixture 2, with several emulsifiers (modified starches; pectin; guar gum; carrageenan; polyphosphates; xanthan gum), a preservative (potassium sorbate), and a colorant (curcumin), typically found in a variety of ultra-processed foods including broth, dairy desserts, fats, and sauces;
– mixture 5, composed essentially of food additives found in artificially sweetened beverages and sugar-sweetened beverages. These additives included acidifiers and acidity regulators (citric acid; sodium citrates; phosphoric acid; malic acid), colorants (sulphite ammonium caramel, characteristic of cola drinks; anthocyanins; paprika extract), artificial sweeteners (acesulfame K; aspartame; sucralose) and some emulsifiers (gum arabic; pectin; guar gum).
A clear cocktail effect
The cocktail effect is clear, as the study authors explain:
‘Despite slight attenuations, the associations between mixtures 2 and 5 and the incidence of type 2 diabetes remained similar after adjustment for each characteristic food additive of the mixture using the residual method, suggesting that the associations were not strongly driven by a single additive’.
Previous evidence
Human studies from the NutriNet-Santé cohort
The originality of the research prevents comparison with previous epidemiological literature. In the same NutriNet-Santé cohort, associations between individual food additives and the incidence of type 2 diabetes had already emerged.
‘Consistently, many of the emblematic food additives of mixtures 2 or 5 have been associated with an increased incidence of type 2 diabetes in previous publications on emulsifiers and artificial sweeteners (preservatives and colorants are currently under investigation)’, explain the French researchers.
Evidence from animal models
In the scientific literature, several individual food additives present in mixtures 2 and 5 have also already been associated with an increased incidence of type 2 diabetes. Some examples:
– guar gum (E412) in a mouse model altered the composition of the intestinal microbiota with consequent increase in pro-inflammatory markers and potential metabolic perturbations. The role of the intestinal microbiota in the development of type 2 diabetes concerns the alteration of glucose metabolism pathways;
– carrageenan (E407) impairs glucose metabolism in mice and has inflammatory properties, which may be involved in the etiology of type 2 diabetes;
– artificial sweeteners alter the intestinal microbiota. In particular, in studies on mice, acesulfame-K (E950) and sucralose (E955) cause dysbiosis, which increases glucose intolerance.
Conclusions
Almost 20 years after the surprising Southampton study, the French study on the impact of additive mixtures on such a large audience should alert the institutions responsible for protecting public health.
‘These results suggest that it could be interesting to consider potential interaction/synergistic/antagonistic effects in the safety assessment of food additives and call for a reassessment of the regulations governing their use by the food industry, with the aim of improving consumer protection.
In the meantime, these results support the public health recommendation to limit exposure to ultra-processed foods and related non-essential food additives’, conclude the authors of the study.
Marta Strinati
Cover art copyright © 2025 Dario Dongo (AI-assisted creation)
References
– Payen de la Garanderie M, Hasenbohler A, Dechamp N, Javaux G, Szabo de Edelenyi F, Agaësse C, De Sa A, Bourhis L, Porcher R, Pierre F, Coumoul X, Kesse-Guyot E, Allès B, Fezeu LK, Cosson E, Tatulashvili S, Huybrechts I, Hercberg S, Deschasaux-Tanguy M, Chassaing B, Rytter H, Srour B, Touvier M. Food additive mixtures and type 2 diabetes incidence: Results from the NutriNet-Santé prospective cohort. PLoS Med. 2025 Apr 8;22(4):e1004570. https://doi.org/10.1371/journal.pmed.1004570
– McCann D, Barrett A, Cooper A, Crumpler D, Dalen L, Grimshaw K, Kitchin E, Lok K, Porteous L, Prince E, Sonuga-Barke E, Warner JO, Stevenson J. Food additives and hyperactive behaviour in 3-year-old and 8/9-year-old children in the community: a randomised, double-blinded, placebo-controlled trial. Lancet. 2007 Nov 3;370(9598):1560-7. doi: 10.1016/S0140-6736(07)61306-3. Erratum in: Lancet. 2007 Nov 3;370(9598):1542. PMID: 17825405.
– Debras C, Deschasaux-Tanguy M, Chazelas E, Sellem L, Druesne-Pecollo N, Esseddik Y, Szabo de Edelenyi F, Agaësse C, De Sa A, Lutchia R, Julia C, Kesse-Guyot E, Allès B, Galan P, Hercberg S, Huybrechts I, Cosson E, Tatulashvili S, Srour B, Touvier M. Artificial Sweeteners and Risk of Type 2 Diabetes in the Prospective NutriNet-Santé Cohort. Diabetes Care. 2023 Sep 1;46(9):1681-1690. doi: 10.2337/dc23-0206. PMID: 37490630; PMCID: PMC10465821.
– C Salame, B Srour, E Viennois, I Huybrechts, S Hercberg, M Deschasaux-Tanguy, E Cosson, S Tatulashvili, B Chassaing, M Touvier, Food emulsifiers and risk of type 2 diabetes: results from the prospective NutriNet-Santé cohort, European Journal of Public Health, Volume 33, Issue Supplement_2, October 2023, ckad160.1248, https://doi.org/10.1093/eurpub/ckad160.1248
– Paudel D, Nair DVT, Tian S, Hao F, Goand UK, Joseph G, et al. Dietary fiber guar gum-induced shift in gut microbiota metabolism and intestinal immune activity enhances susceptibility to colonic inflammation. Gut Microbes. 2024;16(1):2341457. pmid:38630030
– Di Vincenzo F, Del Gaudio A, Petito V, Lopetuso LR, Scaldaferri F. Gut microbiota, intestinal permeability, and systemic inflammation: a narrative review. Intern Emerg Med. 2024;19(2):275–93. pmid:37505311
– Bhattacharyya S, O-Sullivan I, Katyal S, Unterman T, Tobacman JK. Exposure to the common food additive carrageenan leads to glucose intolerance, insulin resistance and inhibition of insulin signalling in HepG2 cells and C57BL/6J mice. Diabetologia. 2012;55(1):194–203. pmid:22011715
– Kimilu N, Gładyś-Cieszyńska K, Pieszko M, Mańkowska-Wierzbicka D, Folwarski M. Carrageenan in the diet: friend or foe for inflammatory bowel disease?. Nutrients. 2024;16(11):1780. pmid:38892712
– Abou-Donia MB, El-Masry EM, Abdel-Rahman AA, McLendon RE, Schiffman SS. Splenda alters gut microflora and increases intestinal p-glycoprotein and cytochrome p-450 in male rats. J Toxicol Environ Health A. 2008;71(21):1415–29. pmid:18800291
– Hanawa Y, Higashiyama M, Kurihara C, Tanemoto R, Ito S, Mizoguchi A, et al. Acesulfame potassium induces dysbiosis and intestinal injury with enhanced lymphocyte migration to intestinal mucosa. J Gastroenterol Hepatol. 2021;36(11):3140–8. pmid:34368996
Professional journalist since January 1995, he has worked for newspapers (Il Messaggero, Paese Sera, La Stampa) and periodicals (NumeroUno, Il Salvagente). She is the author of journalistic surveys on food, she has published the book "Reading labels to know what we eat".
