Overview
A large French prospective cohort study has found that higher exposure to several food coloring additives was associated with a higher incidence of type 2 diabetes. The findings come from the NutriNet-Santé study, which tracked more than 108,000 adults for a median of just over eight years and analyzed detailed dietary records alongside food additive composition databases and laboratory testing.
This study does not prove that coloring additives directly cause diabetes. However, it adds to growing concerns that some food additives may influence long-term metabolic health, either through the foods they are found in, underlying dietary patterns, or possible biological effects on glucose metabolism, inflammation, and the gut microbiome.
Why this study matters
Type 2 diabetes is a major public health problem worldwide. It develops when the body becomes less responsive to insulin, or when the pancreas cannot produce enough insulin to maintain normal blood sugar levels. Diet is one of the most important modifiable risk factors.
For years, researchers have focused mainly on sugar, refined carbohydrates, and ultra-processed foods. This study goes a step further by examining specific food additives, especially coloring agents, that are widely used to make products look more appealing. Food colors are present in a broad range of products, including beverages, desserts, candies, sauces, baked goods, and processed snacks.
How the study was designed
The researchers used data from the French NutriNet-Santé prospective cohort, a long-running study designed to examine the relationships between nutrition, lifestyle, and health outcomes.
Key features of the study included:
Participants: 108,723 adults, 79.2% female, with a mean age of 42.5 years at baseline.
Follow-up period: 2009 to 2023, with a median follow-up of 8.05 years.
Dietary assessment: Repeated 24-hour dietary records, which included brand-specific industrial food products.
Exposure assessment: Cumulative exposure to food additives was estimated using multiple food composition databases and laboratory analyses of food matrices.
Outcome: New cases of type 2 diabetes, identified during follow-up.
Analysis: Multivariable Cox proportional hazards models were used to estimate associations while adjusting for several potential confounders.
The researchers categorized exposure differently depending on how common each additive was in the population. For additives consumed by more than two-thirds of participants, they used sex-specific tertiles. Otherwise, they compared nonexposed/lower-exposed participants with higher-exposed participants based on sex-specific medians.
What the researchers found
During follow-up, 1,131 participants developed type 2 diabetes.
After correcting for multiple comparisons using the false discovery rate method, several food coloring additives were linked to higher diabetes incidence. The strongest associations were observed for the following categories and compounds:
Total food coloring additives: HR 1.38 for higher vs. non/lower consumers
Total caramel: HR 1.43
Plain caramel: HR 1.46
Sulfite ammonia caramel: HR 1.30
Total carotene: HR 1.27
Carotenoids: HR 1.39
β-carotene: HR 1.44
Paprika-capsanthin-capsorubin: HR 1.26
Lutein: HR 1.20
Curcumin: HR 1.49
Cochineal-carminic acid-carmines: HR 1.27
Anthocyanins: HR 1.40
In practical terms, a hazard ratio above 1.0 means a higher rate of new diabetes cases in the more exposed group compared with the reference group. For example, an HR of 1.49 for curcumin indicates a 49% higher hazard of developing type 2 diabetes in the higher-exposure group, after accounting for measured confounders.
Understanding the additives involved
The additives associated with diabetes risk included both natural and synthetic coloring agents.
Caramel colors are widely used in cola drinks, sauces, gravies, and baked goods. They are among the most common coloring agents in the food supply.
Carotenoids and related compounds such as β-carotene, lutein, and paprika-derived pigments are often considered natural colors. They are used to give foods yellow, orange, or red hues.
Curcumin is a yellow pigment derived from turmeric and is used as a colorant in some foods.
Cochineal, carminic acid, and carmines are red colorants derived from insects and are used in some beverages, yogurts, confectionery products, and processed foods.
Anthocyanins are plant-derived pigments responsible for red, purple, and blue colors in fruits and vegetables.
Although some of these compounds are commonly perceived as natural and therefore harmless, this study suggests that their presence in highly processed foods or the overall dietary patterns associated with them may still be relevant to diabetes risk.
Possible explanations
The study was observational, so it cannot establish causality. Several explanations are possible.
First, food coloring additives may be markers of ultra-processed foods. These products often contain more sugar, refined starch, unhealthy fats, and salt, all of which can increase diabetes risk. Even though the researchers adjusted for many factors, residual confounding is still possible.
Second, some additives may have direct or indirect biological effects. Experimental studies have suggested that certain food additives can alter the gut microbiome, influence intestinal barrier function, or affect inflammatory pathways. These mechanisms may play a role in metabolic disease, although more research is needed.
Third, people who consume more additive-containing foods may have broader dietary and lifestyle patterns that are harder to measure completely, such as lower diet quality, less physical activity, or other health-related behaviors.
What this means for consumers
This study does not mean that a single serving of a colored food will cause diabetes. Risk develops over time and depends on the overall diet, body weight, physical activity, genetics, and other factors.
Still, the findings support a practical message: limiting highly processed foods is likely beneficial. A diet centered on minimally processed foods such as vegetables, fruits, legumes, whole grains, nuts, plain dairy products, fish, eggs, and unsweetened beverages is generally associated with better metabolic health.
Consumers can reduce exposure to many food additives by:
Choosing fresh or minimally processed foods more often
Reading ingredient lists on packaged products
Limiting ultra-processed snacks, candies, desserts, and sweetened drinks
Preferring products with shorter ingredient lists
Being cautious with brightly colored processed foods, especially if consumed frequently
Public health and regulatory implications
The authors suggest that these results may warrant reevaluation of some food coloring additives if future studies confirm the associations. Regulatory agencies typically assess the safety of additives based on toxicology, exposure levels, and available evidence. Findings from large cohort studies can complement laboratory data by showing how additives may relate to long-term health in real-world settings.
This is especially important because food additives are often evaluated individually, while consumers are exposed to complex mixtures across many foods. Future work may need to address additive combinations, dose-response relationships, and potential interactions with overall diet quality.
Strengths of the study
This research has several important strengths:
A very large sample size
Long follow-up duration
Repeated dietary assessments rather than a single baseline questionnaire
Detailed data on branded food products
Use of multiple databases and laboratory assays to better estimate additive exposure
Statistical correction for multiple testing
These features increase confidence in the observations, although they do not eliminate the limitations of observational research.
Limitations to keep in mind
There are also limitations. The cohort consisted mainly of volunteers who may be more health-conscious than the general population, so results may not apply equally to all groups. Dietary intake was self-reported, which can lead to measurement error. Additive exposure estimates may also be imperfect, because recipes and formulations can vary over time and by brand.
In addition, observational studies cannot rule out confounding. The associations may reflect not only the additives themselves but also the foods that contain them and the broader dietary patterns linked with those foods.
Bottom line
In this large prospective cohort study, higher exposure to several food coloring additives was associated with a greater incidence of type 2 diabetes. The results were seen for both natural and synthetic colorants, including caramel colors, carotenoids, curcumin, cochineal-derived pigments, and anthocyanins.
The findings do not prove cause and effect, but they raise an important question about the health impact of food additives in modern diets. More research is needed to understand the biological mechanisms and to determine whether reducing exposure to certain colorants can lower diabetes risk.
For now, the most consistent takeaway is familiar but important: a diet built on less processed, more whole foods remains one of the best strategies for protecting metabolic health.
