Are Food Dyes Really Unhealthy?
“Quod ali cibus est aliis fuat acre venenum.”
What is food to one is fierce poison to others.1
- attributed to Lucretius in his work De Rerum Natura, 99–55 BC
In April 2025, the Food and Drug Administration (FDA) announced a phase-out of petroleum-based dyes. The document quotes FDA Commissioner Marty Makary and Health and Human Services (HHS) Secretary Robert F. Kennedy (RFK) Jr., urging US manufacturers to replace synthetic dyes with “natural” alternatives to protect our children. According to Makary:
“We have a new epidemic of childhood diabetes, obesity, depression, and ADHD. Given the growing concerns of doctors and parents about the potential role of petroleum-based food dyes, we should not be taking risks and do everything possible to safeguard the health of our children.”
Kennedy and Makary frame food dyes as one possible driver of Americans’ declining health. In their announcement, the FDA references RFK Jr.’s Make America Healthy Again (MAHA) report to support the phase-out of dyes.
In general, I’m skeptical of any claim from RFK Jr. because I strongly disagree with many of his views (vaccinations, mitochondrial dysfunction, etc.). But I do want America to be healthy, and on first read, this link between food dyes and health seems plausible. Anecdotally, I’ve had friends claim ‘allergies’ to Red 40. They said it made them hyperactive and anxious. And I’ve heard the reports of dyes worsening ADHD symptoms.
I hadn’t realized that food dyes are petroleum-based (derived from crude oil). That sounds gross, yet we routinely use petroleum products like Vaseline and Aquaphor, even applying them to our faces and chapped lips. My non-exhaustive search also revealed that synthetic vanilla (vanillin) flavor and the sweetener saccharin are derived from petroleum.
Tolerance for ‘grossness’ is personal. Some people avoid chicken because of what they perceive to be disgusting methods of factory farming. The FDA permits 60 insect fragments per 100g of chocolate. If eating roach bits doesn’t stop me from devouring the occasional Hershey’s bar, should a dash of petrochemical FD&C Red 40?
Should I really be worried about the health effects of food dyes? Does data really show a link to health effects in support of a ban? What are the natural alternatives, and are they safe? Petroleum is naturally occurring as crude oil after all.
How does the FDA regulate our food colorants? Are we really more permissive than Canada and much of Europe, as RFK Jr. claims?
What the MAHA report actually says
Given the FDA references to the MAHA report, I would like to know what it actually says about food dyes. To my surprise, the report focuses more on ultraprocessed foods (UPFs), defining them as “industrially manufactured food products made up of several ingredients (formulations) including sugar, oils, fats, salt, and food substances of no or rare culinary use.”
The report further specifies that substances of no or rare culinary use include preservatives and food dyes. It proceeds to highlight these food additives for contributing to the problem of UPFs. Specific mention of food dyes is limited to the following:
“Certain food colorings, such as [FD&C] red 40, which is present in widely-consumed products have been associated with behavioral issues in children, such as increased hyperactivity and symptoms consistent with ADHD. Additionally, preliminary evidence suggests a possible association between the consumption of food colorings and autism, although further long-term research is necessary to establish a definitive link.”
While the MAHA report primarily focuses on UPFs, RFK Jr. has clearly singled out food dyes as a primary concern. It is food dyes that he has been working to phase out–first through collaboration with manufacturers and now through his influence on the FDA.
Are food dyes really harmful to our health?
There’s plenty of research on UPFs, but much of it is of mixed quality.2 Because UPFs bundle sugar, sweeteners, preservatives, flavors, and colorings, diet studies are confounded; therefore, UPF findings don’t automatically implicate synthetic food dyes. The quality of research is also poor because it’s challenging to research human diets in general.3
RFK Jr. argues that food dyes are harmful, suggesting that he believes they are a primary cause of poor health outcomes related to UPFs. In a CBS News interview, while in his role as HHS Secretary, he cites a “very, very well-documented link…[to] a variety, a grim inventory of diseases”. He notes that comparable products manufactured abroad use vegetable dyes.
Human data on dyes are limited: a few high-quality studies, two meta-analyses, and a systematic review. The most recent meta-analysis, conducted by California’s Office of Environmental Health Hazard Assessment (OEHHA) in 2021, updates an earlier meta-analysis by Nigg et al. from 2012.4
Nigg’s team analyzed studies that examined the effect of food dyes alone on behavior and attention. They found that about 8% of children with ADHD may have worsening symptoms related to food coloring. However, when the analysis excluded synthetic dyes not approved by the FDA (those used only outside the US), the effect on behavior did not persist. This implies that the dyes used in the US are not contributing to worsening ADHD symptoms. Moreover, the effect size (a measure of the severity of or degree of behavior change with the dye) that they found was so small as to be near zero.5
The more recent OEHHA analysis includes two new studies (and a few older, smaller studies) in addition to those reviewed by the Nigg analysis. Even with these new studies, OEHHA reports similar findings.6 However, it excluded a new large negative study (n = 130) due to result reporting issues.7 Frustratingly, the OEHHA also treated non-significant positive associations as evidence of an effect on behavior. This is a stretch given that OEHHA also found that the larger and better-designed studies were not more likely to show statistically significant or large effects. In fact, most studies that showed any substantial behavioral change due to food dyes had small sample sizes (sometimes consisting of only one or two subjects). In this setting, non-significant positives read as noise—not proof of harm.
I notice several other issues across the various studies included in both meta-analyses. Two of the larger studies added sodium benzoate to the synthetic dye doses. This additive has been independently associated with worsened ADHD symptoms and anxiety, clouding the picture.8 All but two studies tested a mix of dyes, making it impossible to identify a single culprit dye. Most study participants were diagnosed with ADHD or elimination diet-responsive children, thus limiting generalizability to all children.
Based on these meta-analyses, it’s unclear that synthetic food coloring used in the US contributes at all to an “epidemic” of ADHD symptoms. Just as likely, other UPF additives are implicated, or none of them are.
A connection does exist between food dyes and genuine allergic or hypersensitivity reactions. Double-blind, placebo-controlled oral food challenge studies of tartrazine show that the dye can induce hypersensitivity reactions, including asthma (as measured by worsening pulmonary function tests) and hives in patients with moderate to severe asthma who also have an aspirin allergy.9 But many foods cause allergies. I wouldn’t insist that the FDA ban peanuts. This is a case for education and transparency, not a nanny state.
What about cancer? Don’t synthetic dyes cause cancer? No human data link any FDA-approved synthetic dyes to cancer.10 First, studies linking petroleum-based dyes to cancer are conducted in rodent models, where dyes are administered at extremely, unimaginably high doses. In particular, lipophilic (or fat-soluble) azo dyes, such as Sudan dyes and Para Red, are carcinogenic in rodents. These dyes are banned in the US for use in food. The more concerning data on these lipophilic, azo dyes is their association with an increased risk of bladder cancer in humans who worked with them in industrial dye processing facilities.11
Finally, the MAHA report draws a link between synthetic food dyes and autism, stating, “preliminary evidence suggests a possible association between the consumption of food colorings and autism.” My review found no supporting evidence. None. I searched major scientific databases myself. I asked ChatGPT 5.0 to help me with my search. Even the book chapter, cited by the MAHA report, to back up this claim, states, “The research does not prove that food coloring actually causes autism spectrum disorder, but there seems to be a link.”12 This hardly proves a link between petroleum-based food dyes and autism.
How are food dyes regulated, and what is the FDA changing?
In their reports, Makary and RFK Jr. indicate that the US is less stringent than Europe and Canada in regulating food dyes. Is that really true? How are these dyes regulated? Are the changes the FDA proposes in line with the policies of other advanced countries? Have other food agencies come to different conclusions regarding the safety of these dyes than I have?
Synthetic dyes used in the US require FDA approval for each specific use case (e.g., food, textiles, cosmetics); each new batch of dye is tested and certified for use. The FDA also determines the recommended weight-based acceptable daily intake (ADI). The FDA requires food labels to name these specific dyes on their ingredient lists.
I was surprised to learn that, unlike synthetic dyes, food dyes considered to be “natural” pigments (or lab-synthesized, bio-identical versions) require FDA approval but are exempt from batch certification. Also, unlike with synthetic dyes, labels need not specify the source of pigment, but may say ‘artificial color’ or ‘color added,’ except in the case of carmine/cochineal. I would like to know the source of “artificial color”. For many of these natural dyes, the FDA also specifies allowable use cases and makes recommendations for ADI, and permissible levels of contaminants.
The FDA allows nine synthetic dyes for use in food products. One, FD&C Red 3, has been slated for ban since the 1990s. 13In fact, because the FDA indicated an intention to ban it, most manufacturers have long since switched from Red 3 to Red 40. Two other synthetic dyes, FD&C Citrus Red 2 and FD&C Orange B, are banned in the EU and Canada and have minimal allowable use cases in the US. Green 3 was never authorized in the EU, but it is approved in Canada. However, there are six synthetic dyes approved for use in foods in the EU and one in Canada that are not approved in the US. In practice, the US has only six synthetic dyes in regular use compared to seven in Canada and twelve in Europe.14
Consumer demand, rather than government regulation (or banning), has been the primary driver of a shift to natural dyes in Canada and Europe. This shift has been particularly evident in frequently used products that are often considered healthy, such as breakfast cereals, yogurts, and boxed macaroni and cheese. In that sense, RFK Jr. is correct that similar products are available in other countries with natural dyes.
But companies in the US have also begun to offer natural alternatives. A few of these brands (e.g., Nestlé and Kraft Heinz) had already made the switch to natural food dyes in US markets even before RFK Jr. came on the scene. Trader Joe’s, the grocery store, only sells products with natural dyes. Given the safety profile of synthetic dyes, allowing this consumer-driven growth that increases choice seems better than forcing a complete changeover.
Furthermore, the European Food Safety Authority (EFSA) has rejected a clear link between synthetic dyes and ADHD. However, it does require manufacturers to include warnings about the potential for ADHD-related effects for certain dyes. Canada more closely regulates the allowable concentration in food (in parts per million) instead of the FDA’s more general ADI. Canada’s Food Inspection Agency also conducts random sampling of foods to spot-check for compliance. On the other hand, both regions allow a wider range of synthetic dyes to be used. While critics suggest the US is far more lax than other countries, the picture is actually more nuanced. It’s not more lax, it's just different.
If RFK Jr. has his way, as it seems he will, synthetic food dyes will soon be significantly (and needlessly) more restricted in the US than in other countries.
The FDA is targeting three dyes for immediate removal: Red 3, Citrus Red 2 (used only on orange peels), and Orange B (used only for sausage casings). These limited-use dyes are low-hanging fruit, making them an easy win for RFK Jr.
In addition to the above, the FDA states that they are:
“Working with industry to eliminate six remaining synthetic dyes—FD&C Green No. 3, FD&C Red No. 40, FD&C Yellow No. 5, FD&C Yellow No. 6, FD&C Blue No. 1, and FD&C Blue No. 2—from the food supply by the end of 2026.”
This step is a significantly more restrictive and concerning step than anything we’ve seen from the EU and Canada.
To maintain adequate food dye options for the food manufacturing industry, the FDA is fast-tracking the review of natural alternatives. These natural dyes include “calcium phosphate (a white mineral), Galdieria extract blue (from red algae), gardenia (flower) blue, butterfly pea flower extract, and other natural alternatives to synthetic food dyes.”
These new dyes will be added to the natural dyes already approved by the FDA. Most of these dyes are considered “generally recognized as safe (GRAS)” because they are derived from plant foods that have been consumed for thousands of years. Today, many are still derived from plant sources, but some are synthesized to create bio-identical molecules, allowing for significantly increased production at a lower cost.
Are “natural” dyes any safer?
Natural isn’t always better. Natural products can be harmless, dangerous, or somewhere in between. To take an extreme example, poison hemlock is a naturally occurring plant. Foxglove contains digitalis; its extract, digoxin, is used to treat heart failure by slowing the heart rate and strengthening contractions, but excessive doses can be fatal.
Natural dyes require FDA approval to prevent inadvertent use of something poisonous. However, once approved, they are exempt from specific labeling and the monitoring inherent in batch testing.
Many natural products are known to be safe, but they can have downsides. Many of them are extremely expensive to produce. Saffron is wildly expensive. Humans must process, on average, 100,000 crocus flowers by hand to obtain one pound of saffron. It can be more expensive, ounce for ounce, than gold. Similarly, beta carotene requires a great deal of vegetables to extract naturally, but chemical synthesis enables us to produce large, inexpensive batches of a bio-identical product. Still other dyes, like carmine, are made from pulverized cochineal beetles15. This might be an issue if you’re vegan, which is why carmine is an exception to the labeling rule.
Even if natural dyes are safe, they’re not always pure. One leading global manufacturer and supplier of natural dyes reported that up to 25% of their botanical ingredients contain deliberate adulterants (such as ground-up brick in paprika) or contaminants (including residual solvents, pesticides, microbes, and heavy metals). While the FDA publishes general guidelines, quality control depends on industry recognition of the importance of self-monitoring.
A 2017 Journal of Food Science committee report urged the food industry to take the lead in setting standards for natural dyes.16 This type of report is reassuring. However, I found little evidence of follow-through. It’s not that I think beet juice is dangerous. But given the FDA's ‘exempt’ status and recent contamination episodes (e.g., Cinnamon with lead), transparency and testing gaps are concerning.
Because oversight for natural dyes is less stringent, we are at the mercy of the food industry to provide transparency on its quality control, processing, and sourcing of natural dyes. Without this transparency, a shift to these products could easily repeat the supplement industry’s problems in the 1990s, when purity and safety were far from guaranteed. Consumers shouldn’t have to worry about ending up with heavy metals, pesticides, or brick powder in their food.
This feels like trading the devil we know for the devil we don’t. We may be trading negligible risks associated with synthetic dyes for opaque but real risks associated with potentially contaminated natural dyes. Large manufacturers may be able to guarantee the safety of their natural dyes. But when the entire food industry switches, can quality assurance keep pace? Will smaller companies cut corners and hide behind lax FDA regulations to sell subpar products?
Are there any benefits to food dyes?
Some might argue that food dyes are unnecessary, given that they don’t add nutritional value or flavor. But food dyes do have benefits.
Colors cue recognition and signal the expected flavor of food. Studies show that when sweets are a deeper red, they taste sweeter even with the same amount of sugar added. Thus, color can enable the development of lower-sugar products. Food dyes also have the potential to make healthy foods more appealing to picky eaters (including autistic children).
Food dyes allow manufacturers to signal taste, standardization, and quality, which benefits producers. Some of the earliest examples of food dye use involved adding annatto to butters and cheeses that varied in color based on the cows' diet, which changed throughout the year. Synthetic dyes offer reliable hues at low quantities and at a comparatively (to most natural dyes) low price.
Finally, and perhaps most importantly, we deserve to enjoy our food. Eating a piece of clear candy or grey-fleshed farmed salmon is much less fun than eating the colored versions. While food dyes can make UPFs more attractive, they can also make healthy foods more enticing and enjoyable. Sometimes I want to make my child a green, lemon-lime chocolate swirl birthday cake because it is celebratory and joyful.
Furthermore, eliminating food dyes from food doesn’t necessarily make it healthier. If anything, it might lead consumers into mistaking UPFs with “natural coloring” for nutritious food, despite the presence of other unhealthy additives. Naturally colored Fruit Loops are still packed with preservatives, heavily processed grains, and added sugars. And we should still be allowed to eat them if we want.
What does this all mean?
Just because food coloring doesn’t add nutritional value, doesn’t mean that it isn’t worth enjoying. Health is a significant value, but it isn’t the only value in life. Many things in life are health-neutral or even unhealthy, yet we indulge in them because they bring us joy. Going on a roller coaster isn’t healthy, but many people enjoy doing it. Most people would agree that candy isn’t particularly healthy, but in moderation, (sour) candy is one of life’s great pleasures.
Synthetic dyes currently approved in the US aren’t clearly linked to disease, despite claims by RFK Jr., MAHA, and Makary’s FDA. Given that eating beautiful foods brings me joy every day and I find no discernible risk of food dyes, I do not plan to stop eating them.
I’ll have my bright green cake and relish eating it too!17
Acknowledgements: Thank you to Alexander Kustov, Karthik Tadepalli, Emma McAleavy, Travis Pew, and Mike Riggs, Steven Adler, and Adam Kroetsch for helpful comments, editing, and discussion.
If you enjoyed the article, please share it around. Feel free to add your comments, questions, and thoughts.
Often translated as “One man’s meat is another man’s poison.”
I’m not a statistician, but as a doctor, I read and judge studies with an eye to how they’re designed and implemented. The best study design in these cases would be a human (not rodent) study that is randomized, double-blind, and placebo-controlled with a crossover design. In that setup, the same people try both the intervention (like a food dye or an ultra-processed food) and a placebo at different times, so each person acts as their own control—making fair comparisons easier. Strong studies also specify dosing, utilize proven rating tools, include the appropriate participants, employ more than one rater (and verify that raters agree), enroll sufficient participants to detect genuine effects, and report errors and conflicts of interest.
My essay isn’t really about UPFs, but they’re peripherally relevant given the context of today’s debates (and MAHA). One of the best UPF studies I’ve seen is here. The authors actually found 20 people to stay in the hospital for a month to facilitate strict diet controls.
The Nigg meta-analysis does implicate UPFs in worsening behavior for children with ADHD, showing that about 30% of children had measurable improvements in behavior when UPFs were eliminated. However, this doesn’t give us anything meaningful to say about food dyes, as these elimination diets also removed sugars, various types of oils, preservatives, artificial flavors, and other additives.
Nigg's effect size for parent ratings of behavior was g≈0.12, (95% CI 0.01–0.23), p = 0.0007 or very small, for teacher ratings of attention, it was slightly higher, g≈0.27, (95% CI 0.07–0.47), p = 0.007 or moderate; both effects lost statistical significance with the exclusion of non-FDA-approved dyes.
One of the new studies was rejected by the European Food Safety Authority (EFSA) for use of sodium benzoate, along with several other experimental design issues, including a lack of standard outcomes measure, heterogeneous evaluations, and failure to report dosages:
McCann, Debbie, et al. “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.” The Lancet 370, no. 9598 (2007): 1560–1567. https://doi.org/10.1016/S0140-6736(07)61306-3
Watson R. European agency rejects links between hyperactivity and food additives. BMJ. 2008 Mar 29;336(7646):687. doi: 10.1136/bmj.39527.401644.DB. Epub 2008 Mar 20. PMID: 18356207; PMCID: PMC2276272.
Lok, K.Y.W. et al. “Food Additives and Behavior in 8- to 9-Year-Old Children in Hong Kong: A Randomized, Double-Blind, Placebo-Controlled Trial.” Journal of Developmental & Behavioral Pediatrics 34, no. 9 (2013): 642–650. https://doi.org/10.1097/DBP.0000000000000005
Leila Hejazi et al. A critical review on sodium benzoate from health effects to analytical methods, Results in Chemistry, Volume 11, 2024, 101798, ISSN 2211-7156, https://doi.org/10.1016/j.rechem.2024.101798.
Ardern KD, Ram FS. Tartrazine exclusion for allergic asthma. Cochrane Database Syst Rev. 2001;2001(4): CD000460. doi: 10.1002/14651858.CD000460. PMID: 11687081; PMCID: PMC6483719.
Oplatowska-Stachowiak, M., & Elliott, C. T. (2016). Food colors: Existing and emerging food safety concerns. Critical Reviews in Food Science and Nutrition, 57(3), 524–548. https://doi.org/10.1080/10408398.2014.889652
Benzidine-Based dyes. International Agency for Research on Cancer. Supplement 7-26.
Although the title of the chapter by Bakthavachalu, Prabasheela, et al. is “Food Color and Autism: A Meta-Analysis,” it is actually a qualitative review, not a meta-analysis (which pools data from independent studies to identify patterns, trends, and more statistically robust conclusions). Furthermore, as a book chapter, it has not undergone rigorous peer review. The primary source of data it references to back its assertion of a link between food dyes and ADHD is the study that the EFSA rejected.
This ban has been protracted because the FDA acknowledges the weak data supporting any link with cancer and Red 3. The ban is based on rat studies (wah wah) showing a link to a hormonal thyroid cancer that is based on a mechanism that doesn’t exist in humans. Additionally, the dose of Red 3 administered to the rats was significantly higher than any conceivable equivalent human exposure.
Data taken from the US FDA, Health Canada, and European Commission databases as well as the OEHHA report previously cited.
United States (FDA) U.S. Food and Drug Administration. “Color Additives in Foods.” FDA, updated July 25, 2024. https://www.fda.gov/food/food-additives-petitions/color-additives-foods
Canada (Health Canada) Health Canada. “List of Permitted Colouring Agents.” Government of Canada, modified June 25, 2024. https://www.canada.ca/en/health-canada/services/food-nutrition/food-safety/food-additives/lists-permitted/3-colouring-agents.html.
European Union (EU)European Commission. “EU Food Additives Database.” European Commission, accessed August 29, 2025. https://webgate.ec.europa.eu/foods_system/main/?sector=FAD&auth=SANCAS.
This insect was a state secret of Spain, allowing the country to maintain a monopoly on its import for over 2,000 years. Cochineal was their second-highest-grossing import after silver for many of those years. It provided the red dye for the robes of cardinals, the Pope, and European royalty, as well as pigment for art. https://humanitiescollaborative.utep.edu/project-blog/cochineal
Simon, J.E., Decker, E.A., Ferruzzi, M.G., Giusti, M.M., Mejia, C.D., Goldschmidt, M. and Talcott, S.T. (2017), Establishing Standards on Colors from Natural Sources. Journal of Food Science, 82: 2539-2553. https://doi.org/10.1111/1750-3841.13927
Some other good, accessible resources that I didn’t cite throughout the paper:
If you don’t want to read all the meta-analyses and studies, the Center for Science in the Pubic Interest has a good summary PDF: Food Dyes a Rainbow of Risks. It’s a little more anti-synthetic dyes than I am, but it reviews most available research. https://www.cspi.org/sites/default/files/media/documents/resource/food-dyes-rainbow-of-risks.pdf
Adam Ragusea’s YouTube episode, “Are food dyes worth banning?” which I found while editing this piece at the recommendation of an editing friend. We make many similar points and a few different points, but overall, I agree with him.
Your post is way too long but very good. And oddly you’ve understated the case. Real scientists and doctorss know it is hopeless to get in disputes with people like Junior. I teach students that chemical reactions completely change the quality of the reactant. To think there’s something inherently wrong with making things from petroleum is simply ignorant. in a chemical reaction, there’s no reason why the products must resemble the reactants. I always say that that is why we call the transcendent character of personal relations, chemistry.