Dr Éric Thorin, Ph. D.
Professeur titulaire, Département de Chirurgie, Université de Montréal. Chercheur au centre de recherche de l'Institut de cardiologie de Montréal.
See all articlesOverview
- Food colourings are present in all ultra-processed foods;
- They have no nutritional, taste, or preservative value;
- A recent study shows the pro-diabetic effects (+38%) of so-called “natural” food colourings in the highest consumers compared to the lowest consumers;
- The impact of synthetic colourings could not be assessed because the study is based on products from the French market, where synthetic colourings are rarely used, unlike in Canada;
- It presents, therefore, a “least worst scenario” for Canadians, who are exposed to food colourings that have been proven harmful by numerous scientific studies.
Culinary chefs are familiar with the adage, “we eat with our eyes first.” Although taste, smell, and sight are distinct senses, we know that visual stimuli alter our perception of taste, smell, and flavour. Colour is the most obvious visual cue, but expectations, shaped by cultural associations learned over time and our culinary experiences, are also influenced by other visual cues, such as brightness, uniformity, and shape. These expectations exert cognitive influences that can modify our evaluation of taste and flavour. Professionals in the food industry are well aware of this, hence the use of food colourings in the ultra-processed products they offer. However, these food colourings provide no added value in terms of taste or nutrition. To meet our expectations, they simply make mint appear green when it is naturally colourless, or they give cola-type sodas their caramel hue. Yet these unnecessary food colourings are not without health risks.
This month, a major and methodologically sound study reports a significant association between the consumption of colourings in food and beverages and an increased risk of type 2 diabetes (+38% compared to those with the lowest exposure). This research was conducted using data from the French NutriNet-Santé cohort, which included a sample of 108,723 volunteers who documented their eating habits for 24 hours every two years since 2009. This included collecting specific data associated with the brand names of the food products consumed, following a methodology described in a previous study. The volunteers also reported their physical activity and health information using questionnaires. This information was then cross-referenced with the Open Food Facts and Observatoire de l’alimentation databases.
The volunteers (mean age 42 years, 80% of whom were women) were followed for an average of 8 years, during which time there were 1,131 new cases of type 2 diabetes. Associations with diabetes risk were studied for food colourings consumed by at least 10% of the study population (see Box 1). To distinguish levels of exposure to added food colourings, cohort participants were divided into three groups for each substance, with the top 33% corresponding to the highest level of exposure and the bottom 33% to the lowest level of exposure. For the top tertile, consumption levels were not excessive, as they corresponded to someone regularly consuming a soda, a ready-made meal, and a dessert. However, unsweetened (49.6%) and sweetened (32.2%) beverages were the main contributors to total exposure to food colourings.
Box 1.
Groups of colouring additives analyzed:
1. Colouring additives consumed by at least 10% of study population: curcumin (European [E] code E100), cochineal-carminic acid-carmines (E120), plain caramel (E150a), sulfite ammonia caramel (E150d), carotenoids without specification (E160), β-carotene (E160a), annatto-bixin-norbixin (E160b), paprika-capsanthin-capsorubin (E160c), lutein (E161b), and anthocyanins (E163). These are classified as natural colouring additives.
2. The most consumed natural colouring additives: total caramel (E150, E150a, E150b [caustic sulfite caramel], E150c [ammonia caramel] and E150d), total carotene (E160b, E160b, E160b, E150d E160c, E160d [lycopene], E160e [β-apo-8′-carotenal], and E161b)
3. Total food colouring additives: E100, E101 (riboflavin), E102 (tartrazine), E104 (quinoline yellow), E110 (orange yellow S), E120, E122 (carmoisine), E123 (amaranth), E124 (ponceau 4R), E127 (erythrosine), E129 (allura red AC), E132 (indigo carmine), E133 (brilliant blue FCF), E140 (chlorophylls), E141 (copper complexes of chlorophylls), E150, E150a, E150b, E150c, E150d, E151 (brilliant black BN), E155 (brown HT), E160, E160a, E160b, E160c, E160d, E160e, E161b, E162 (beet red), E163, E170 (calcium carbonate), E171 (titanium dioxide), E172 (iron oxides and hydroxides), E131 (patent blue V), E153 (vegetable carbon) and E175 (gold).
These analyses were then cross-referenced with the occurrence of several pathologies (we only report type 2 diabetes here) over the study period (from 2009 to 2023-2024), correcting for other biases linked to tobacco or alcohol consumption, the socio-demographic profile, the nutritional quality of the diet and other factors. The results obtained (Table 1) are consistent with other studies showing that food colouring additives could alter molecules involved in insulin signalling, and have other effects on blood sugar levels and intestinal microbiota.
| Colouring additives | RR [95 % CI] | Statistical power |
|---|---|---|
| Total food colouring additives | 1.38 [1.17–1.63] | P = 0.0002 |
| Total caramel | 1.43 [1.21–1.67] | P = 0.0002 |
| Plain caramel | 1.46 [1.26–1.70] | P = 0.0002 |
| Sulfite ammonia caramel | 1.30 [1.07–1.59] | P = 0.007 |
| Total carotene | 1.27 [1.08–1.48] | P = 0.007 |
| Carotenoids | 1.39 [1.19–1.62] | P = 0.0002 |
| β-carotene | 1.44 [1.23–1.68] | P = 0.0002 |
| Paprika-capsanthin-capsorubin | 1.26 [1.08–1.46] | P = 0.004 |
| Lutein | 1.20 [1.02–1.40] | P = 0.0002 |
| Curcumin | 1.49 [1.29–1.73] | P = 0.0002 |
| Cochineal-carminic acid-carmines | 1.27 [1.10–1.48] | P = 0.003 |
| Anthocyanins | 1.40 [1.17–1.68] | P = 0.0002 |
Table 1. Associations between exposure to food colourings and the risk of type 2 diabetes. The data show the impact of high consumption compared to low/lower consumption on the risk of developing the disease. These data correspond to the use of the main food colourings in 10% or more of the population. HR: Hazard ratio; CI: confidence interval
There are two important lessons to be learned from these results. The first is the difference in food colouring use between Europe and North America. In Europe, manufacturers use very few synthetic colourings and instead favour natural colourings, which consumers perceive as being safer (a perception reinforced by marketing). Furthermore, some synthetic colourings have shown adverse health effects in preclinical studies and short-term clinical trials, particularly azo dyes (Allura Red, tartrazine, and Sunset Yellow), which are not widely used in Europe compared to Canada. These food colourings, as well as carmoisine, Ponceau 4R, and quinoline yellow, have been associated with behavioural problems in children. Tartrazine induces neurodevelopmental toxicity by targeting mitochondrial function, leading to cell apoptosis in zebrafish. They also produce metabolic changes involving the insulin molecular pathway linked to insulin resistance and diabetes. Other toxic effects have been reported, such as DNA damage and reduced cell survival in culture. The European Food Safety Authority (EFSA) has lowered the permitted levels for quinoline yellow, sunset yellow, and Ponceau 4R, while titanium dioxide (which produces a white colour) has been banned, unlike in Canada. Consequently, the results of this French study almost certainly underestimate the impact of food colourings on the health of Canadians.
The second lesson to be drawn from this study is that even naturally-derived additives used by European manufacturers, in addition to synthetic ones, are not harmless. The results show that curcumin (E100) is associated with a 49% increased risk of type 2 diabetes, beta-carotenes (E160a) with a 44% increased risk, and caramel colourings (E150) with a 43% increased risk. In their discussion of their results, the authors do not offer a clear explanation for this association with natural additives; one possibility is that the physicochemical properties of these products change and generate unexpected side effects outside of their natural structural environment. For example, researchers highlight the effects of processing to understand why beta-carotene is protective and antioxidant when naturally present in carrots, but becomes associated with negative effects (increasing the risk of cancer) when used as an additive.. Concentrated natural products lose their nutritional benefits when removed from their original form and reintroduced into ultra-processed foods—a phenomenon observed for many substances (see the article discussed here).
The “naturalness” argument put forward by the food industry is therefore untenable. It’s often forgotten, but nature provides some very effective poisons! The vast majority of current medications are derived from natural plant or animal products and formed the basis of the pharmacopoeia before the advent of medicinal chemistry in the mid-19th century. Many of these drugs are still irreplaceable today: aspirin, curare (essential during surgeries), digoxin (increases the contractility of the heart in some heart failure patients), morphine, paclitaxel (a cell anti-proliferative drug used to treat certain cancers, which also revolutionized the use of drug-eluting stents during angioplasties to prevent restenosis of coronary arteries treated after balloon dilation), or colchicine (the only treatment for gout and newly indicated in patients with coronary artery disease to combat residual inflammation and the recurrence of events).
In North America, more than half of our daily calories come from ultra-processed foods containing additives. Previous studies have shown that some of these additives, such as emulsifiers, have negative health effects due to their pro-inflammatory properties. The results of the study presented here suggest that food colourings, even those of natural origin, are another class of potentially harmful additives that require stricter regulation. We can only hope that our policymakers will impose stricter regulations to better control the composition of ultra-processed foods, as we recently proposed, including a reduction in the amount of additives in our diets and the complete elimination of the more controversial synthetic additives.