One of the largest investigations into the health effects of food additives ever conducted suggests that the real risk may lie not in any single ingredient, but in the combinations of additives that people consume together every day. In a prospective cohort study published in BMC Medicine, researchers analyzed data from 112,050 French adults participating in the NutriNet-Santé cohort and found that several real-world mixtures of food additives — combinations of emulsifiers, thickeners, sweeteners, acids and raising agents commonly found in ultra-processed foods — were associated with a significantly higher incidence of hypertension and cardiovascular disease.
The study, led by Marie Payen de La Garanderie and Mathilde Touvier of the Nutritional Epidemiology Research Team at Université Sorbonne Paris Nord and Université Paris Cité, together with colleagues at INSEM, INRAE, CNAM and partner institutions including the International Agency for Research on Cancer, set out to address a fundamental blind spot in food safety science. Regulators such as the European Food Safety Authority and the U.S. Food and Drug Administration traditionally evaluate food additives one substance at a time, determining acceptable daily intakes for individual compounds. But consumers do not eat isolated additives in controlled doses. They eat packaged breads, sodas, processed meats, desserts and sauces that each contain a cocktail of several additives, and across a day’s diet, dozens of these substances interact in the digestive tract simultaneously. The new research represents one of the most ambitious attempts to capture that real-life exposure in an epidemiological framework.
The evidence base came from the NutriNet-Santé cohort, an ongoing web-based study of French adults launched in 2009 with follow-up extended to 2024. Participants in the analysis had a mean age of 42.8 years, with a standard deviation of 14.6, and 79 percent were women. A distinctive strength of the cohort lies in its dietary assessment method: rather than relying on generic food frequency questionnaires, participants completed up to three detailed 24-hour dietary records every six months. Crucially, these records captured the specific commercial brands of industrial products consumed. That brand-level detail allowed the researchers to reconstruct individual exposure to additives with unusual precision, using comprehensive food composition databases supplemented by ad-hoc laboratory analyses of actual food matrices to quantify the additive content of specific products.
To move beyond studying additives in isolation, the team employed Non-Negative Matrix Factorisation, an unsupervised machine-learning technique that identifies data-driven patterns of additives that co-occur in real diets. Rather than predefining which additives might cluster together, the algorithm extracted five recurring mixture profiles from the actual consumption data of more than one hundred thousand people. This approach ensured that the mixtures analyzed reflected genuine dietary patterns — the combinations a consumer of diet sodas, industrial cakes and processed convenience foods would actually be exposed to — rather than hypothetical pairings constructed by the researchers.
Participants’ health outcomes were then tracked through time. Over a median follow-up of 7.92 years, 5,543 participants developed hypertension, 2,450 developed some form of cardiovascular disease, 1,142 experienced a cerebrovascular accident such as a stroke, and 1,308 developed coronary heart disease. The researchers used multi-adjusted Cox proportional hazards models, treating exposure to the additive mixtures as time-dependent so that changes in diet over the years of follow-up were reflected in the analysis. Results were expressed as hazard ratios comparing participants in the highest tertile of consumption of each mixture with those in the lowest, adjusting for a wide range of potential confounders including overall diet quality, lifestyle factors and anthropometric characteristics.
The findings were striking for four of the five mixtures identified. Mixture 1, characterized mainly by sodium carbonates, diphosphates, glycerol and ammonium carbonates — additives typically found in baked goods and confectionery — was associated with a 14 percent higher incidence of hypertension (hazard ratio 1.14, 95 percent confidence interval 1.07–1.22) and a 12 percent higher incidence of cardiovascular disease overall (1.12, 1.00–1.24). Mixture 2, dominated by modified starches, pectin, guar gum and carrageenan, a suite of thickening and gelling agents ubiquitous in dairy desserts, plant-based drinks and processed sauces, showed associations with hypertension (1.12, 1.05–1.19), cardiovascular disease (1.18, 1.07–1.31) and, notably, the strongest single association in the study: a 26 percent increased incidence of coronary heart disease (1.26, 1.10–1.45).
Mixture 4, combining ammonium and sodium carbonates, diphosphates and alpha-tocopherol, was linked to elevated risks of hypertension (1.12, 1.05–1.19), cardiovascular disease (1.13, 1.02–1.26) and coronary heart disease (1.16, 1.01–1.34). Meanwhile, Mixture 5 — an acidic and sweetening cluster containing citric acid, sodium citrates, phosphoric acid and the artificial sweetener acesulfame-K, a profile strongly reminiscent of soft drinks — was associated with a 19 percent higher incidence of hypertension (1.19, 1.11–1.28), though it did not reach statistical significance for the broader cardiovascular endpoints. Only Mixture 3, composed of magnesium carbonates, riboflavin, alpha-tocopherol and ammonium carbonates, showed no detectable association with any of the studied outcomes.
Beyond the headline hazard ratios, the study’s exploratory analyses of interactions between emblematic additives within the mixtures point to something potentially more consequential: evidence of synergy. When two or more substances are present simultaneously, their combined effect may exceed the sum of their individual effects — a synergistic interaction — or, in some cases, one substance may dampen the effect of another, an antagonistic interaction. The researchers report several such interactions among additives composing the identified mixtures, supporting the hypothesis that the health impact of a food’s additive content cannot be predicted by testing each component separately. This resonates with a growing body of experimental work from toxicology laboratories, including work from co-authors at the Toxalim Research Centre in Food Toxicology in Toulouse, examining how combinations of emulsifiers and other additives affect the gut barrier, the intestinal microbiome and inflammatory pathways — mechanisms plausibly connected to hypertension and atherosclerosis.
The findings arrive amid intensifying scientific scrutiny of ultra-processed foods, which previous studies — many from the same NutriNet-Santé team — have linked to cardiovascular disease, cancer, type 2 diabetes and mortality. Food additives have long been suspected as one of the mediating factors behind those associations, but isolating their contribution has proven methodologically difficult. The present study adds an important layer: even within diets that may share a similar degree of processing, the specific cocktail of additives consumed appears to matter, and some cocktails are more strongly implicated in cardiovascular harm than others. The association with carrageenan and other emulsifier-type substances in Mixture 2, in particular, aligns with experimental evidence that certain emulsifiers can erode the protective mucus layer of the gut and promote low-grade inflammation.
The authors are careful to frame their conclusions appropriately. This is an observational study, and while the statistical models adjusted for numerous confounding factors, residual confounding — for example, by unmeasured lifestyle characteristics associated with both additive-rich diets and cardiovascular risk — cannot be entirely excluded. Hazard ratios in the range of 1.1 to 1.3 describe modest elevations in relative risk, though ones that could translate into substantial absolute public health impact given how widely these additives are consumed across populations. The cohort is also predominantly female and composed of volunteers who may be more health-conscious than the general population, factors that the authors note warrant replication in other, more diverse cohorts.
Nevertheless, the study is registered at ClinicalTrials.gov (NCT03335644), was conducted under Declaration of Helsinki guidelines with ethical approval from the French Institute for Health and Medical Research review board and the French data protection authority, and its funders — including a European Research Council Horizon 2020 grant dedicated specifically to studying food additives — played no role in the design, analysis or reporting. The work was published open access, and the authors declare no competing interests.
The implications for food policy could be considerable. If the effects of additives are genuinely interaction-dependent, then current regulatory frameworks, which establish acceptable daily intakes substance by substance, may systematically underestimate risk. The authors explicitly call for the consideration of mixtures in safety evaluations and for research into the underlying biological mechanisms. In the meantime, their recommendation to consumers is straightforward: limiting unnecessary food additives in the diet remains a prudent choice, and the findings lend fresh scientific weight to dietary guidance that favors minimally processed foods over the additive-laden products that now dominate supermarket shelves.Subject of Research: Associations between chronic exposure to real-life mixtures of food additives and the incidence of hypertension and cardiovascular disease in French adults
Subject of Research: Medicine
Subject of Research: Medicine
Article Title: Food additive mixtures and hypertension and cardiovascular diseases incidence in the NutriNet-Santé prospective cohort
Article References: de La Garanderie, M. P., Hasenböhler, A., Yvroud-Hoyos, P., de Edelenyi, F. S., Bourhis, L., Dechamp, N., Agaësse, C., De Sa, A., Huybrechts, I., Pierre, F., Audebert, M., Coumoul, X., Chassaing, B., Julia, C., Kesse-Guyot, E., Allès, B., Fezeu, L. K., Blacher, J., Hercberg, S., … Touvier, M. (2026). Food additive mixtures and hypertension and cardiovascular diseases incidence in the NutriNet-Santé prospective cohort. BMC Medicine. https://doi.org/10.1186/s12916-026-05132-z
Image Credits: AI Generated
DOI: 10.1186/s12916-026-05132-z
Keywords: Food additives, Additive mixtures, Hypertension, Cardiovascular diseases, Coronary heart disease, Cerebrovascular diseases, Ultra-processed foods, NutriNet-Santé cohort, Prospective cohort, Emulsifiers, Food safety, Non-Negative Matrix Factorisation
Cite Scienmag News
APA MLA Chicago
Frances Kline. (September 8, 2026). Mixed food additives linked to higher hypertension and heart disease risk. Scienmag. https://scienmag.com/mixed-food-additives-linked-to-higher-hypertension-and-heart-disease-risk/
Copy citation Download RIS
Tags: combination of food additivescombined food additive effectsdietary risk factors for cardiovascular diseaseemulsifiers and heart diseaseemulsifiers and heart healthEuropean and US food additive policiesfood additive consumption and health outcomesfood additive health risksfood additive mixtures and hypertensionfood additive research studiesfood additive safety regulationfood additives and cardiovascular diseasefood additives and chronic disease riskfood safety research on additive combinationsimpact of food additive mixturesimpact of food additives on public healthlong-term effects of food additivesprocessed food consumption health impactprospective cohort study on food additivesregulatory assessment of food additive safetyultra-processed foods and cardiovascular healthultra-processed foods and hypertension
