When the FDA launched a push to eliminate synthetic food dyes from the American food supply in 2025, the stated goal was healthier food products through the replacement of artificial colorants with natural alternatives. New research from one of the world's largest nutritional epidemiology studies has now complicated that narrative significantly: the natural food color additives being promoted as replacements carry their own documented associations with elevated risk of type 2 diabetes and cancer — in some cases, elevated risks exceeding 40%.
Research published in Diabetes Care and the European Journal of Epidemiology, led by investigators from INRAE, Inserm, EREN, and Cnam using data from the NutriNet-Santé cohort — one of the largest nutritional epidemiology studies in Europe, tracking more than 100,000 French participants over a decade — found that food coloring additives as a whole were associated with a 38% increase in type 2 diabetes risk among those with the highest consumption compared to those with the lowest. Among specific natural colorants, the risk associations were even higher.
As the INRAE and Inserm press release confirmed, and as KFF Health News flagged in its June 17, 2026 Morning Breakout: scientists from several French universities and research institutions identified more than a 40% increased risk of Type 2 diabetes and some cancers when natural food color alternatives are used.
The Specific Natural Colorants Under Scrutiny — and Their Risk Associations
The NutriNet-Santé findings break down by specific additive in ways that directly challenge the common assumption that "natural" origin confers safety. The natural compounds most commonly being proposed as alternatives to synthetic dyes — beta-carotene (from carrots and palm fruit oil), curcumin (from turmeric), and anthocyanins (from berries and red cabbage) — are among those with the most concerning risk associations in the data.
According to the INRAE press release and Food Safety Magazine's analysis:
| Natural Colorant | EU Designation | Source | T2D Risk Association |
| Food colorants overall | — | — | +38% in highest consumers |
| Caramel colors | E150 | Heated sugar | +43% |
| Carotenoid colors | E160 | Plant/animal sources | +39% |
| Beta-carotene (additive) | E160a | Carrots, palm fruit, sweet potato | +44% |
| Ordinary caramel | E150a | Heated sugar | +46% |
| Curcumin | E100 | Turmeric | +49% |
| Anthocyanins | E163 | Berries, red cabbage, red grape | +40% |
For cancer outcomes, the data showed: food coloring additives overall associated with a 14% increase in overall cancer risk, a 21% increase in breast cancer risk, and a 32% increase in post-menopausal breast cancer risk. Individual additives showed additional specific cancer associations not detailed here.
The Beta-Carotene Distinction — Additive Form vs. Whole Food
One of the most important interpretive points in the French research concerns the distinction between beta-carotene consumed as a food additive and beta-carotene naturally present in whole foods like carrots, squash, and sweet potatoes.
As ABC News reported in its analysis of the French studies, the risk associations apply to beta-carotene as a concentrated food additive (E160a), not to beta-carotene naturally present in whole plant foods. This distinction has regulatory and practical significance.
When beta-carotene is extracted and concentrated into a food additive, it is delivered to the body at doses and in a form that does not occur in nature, unaccompanied by the fiber, water, vitamin C, polyphenols, and other bioactive compounds present in whole plant foods that modulate how beta-carotene is processed and affect its biological behavior. This same concern was first documented in the CARET and ATBC trials in the 1990s, which found that beta-carotene supplements (high-dose, concentrated) actually increased lung cancer risk in smokers — a finding that shocked the research community precisely because low dietary beta-carotene intake from whole foods was associated with reduced cancer risk. The additive form replicates the supplement form more closely than the whole food form.
Kantha Shelke, a food scientist at Johns Hopkins University, told ABC News: "The main natural dyes to flag are carmine, also called cochineal extract, and to a lesser extent annatto. The primary concern is allergic reaction rather than toxicity." She noted these represent distinct risks from the metabolic disease associations identified in the French data.
The Policy Timing — Why This Research Matters Now
The NutriNet-Santé findings arrive at a pivotal moment in U.S. food policy. The FDA under the current administration has been removing synthetic food dyes — including Red 3, Red 40, Yellow 5, Yellow 6, and others — from the food supply under the MAHA initiative. Food manufacturers are actively reformulating products with natural alternatives, including beta-carotene, curcumin, anthocyanins, and carmine as replacements.
The French research provides the first large-scale epidemiological evidence that these natural alternatives, when used as concentrated food additives, may carry their own risk associations — at scales that are clinically and epidemiologically meaningful for a population consuming processed foods that use these colorants throughout the day.
The NutriNet-Santé studies do not establish causality. The associations are observational — people who consume more heavily colored processed foods may have other dietary and lifestyle patterns that contribute to disease risk. The researchers account for major confounders, but residual confounding cannot be excluded in observational epidemiology.
However, the same observational limitations apply to the evidence base that originally justified the removal of synthetic dyes. The French data at minimum establishes that the assumption of natural colorants' safety needs rigorous testing — not regulatory default acceptance — before large-scale substitution.
Frequently Asked Questions
What did the French food dye study find?
Research from the NutriNet-Santé cohort (100,000+ French participants tracked for a decade) found that food colorants overall were associated with a 38% higher type 2 diabetes risk among highest consumers. Specific natural colorants showed higher associations: beta-carotene (E160a) +44%, curcumin (E100) +49%, anthocyanins (E163) +40%, caramel colors +43%. Cancer risk associations included +14% overall and +21% for breast cancer.
Are these natural dyes the same as eating turmeric or blueberries?
No. The risk associations apply to concentrated food additive forms (e.g., curcumin E100 added to processed foods), not to natural consumption of whole foods containing these compounds. Beta-carotene in carrots and anthocyanins in blueberries are delivered in the context of fiber, water, and hundreds of other bioactive compounds that modulate how the body processes them.
Does this mean natural food dyes cause diabetes and cancer?
The studies are observational — they identify associations, not causation. People who consume more heavily colored processed foods may have other dietary patterns that contribute to risk. However, the associations are statistically significant, account for major confounders, and appear across multiple specific compounds in one of the largest nutritional epidemiology datasets in the world.
Why does this matter now?
The FDA is actively replacing synthetic dyes with natural alternatives under the MAHA initiative. Many manufacturers are reformulating products with beta-carotene, curcumin, and anthocyanins. The French data challenges the assumption that these natural substitutes are inherently safer than the synthetic dyes they're replacing.
What is the NutriNet-Santé cohort?
NutriNet-Santé is a long-running French nutritional epidemiology study tracking more than 100,000 adult participants since 2009. It is one of the largest dietary-health databases in the world and has produced multiple influential findings on food additives, ultra-processed food, and chronic disease risk.
