Azorubine, Carmoisine

Children’s activity and attention: what the mixture trial can tell us

The 2007 Southampton crossover trial included 153 three-year-olds and 144 children aged eight or nine. Children received a placebo drink or one of two drinks containing several colours plus sodium benzoate. Thirty children did not complete the study. Behaviour scores increased in some comparisons, but findings depended on age, mixture and which participants were analysed.

This ingredient was in at least one tested mixture. The study cannot tell us which ingredient, or combination, produced a change. A higher behaviour score is not a new diagnosis of ADHD, and the result does not mean that every child responds.

Southampton trial: colour and sodium-benzoate mixtures. Indexed abstract and EFSA’s subsequent assessment read. The timeline uses the journal issue date. An indexed November 2007 erratum and commentary were identified; their complete texts and the full trial methods remain to be reviewed. Exact effect estimates are not reproduced.

An earlier trial found different results at home and in the clinic

A 2004 study selected 277 preschool children after screening for hyperactivity and allergy-related traits. The challenge combined tartrazine, sunset yellow, carmoisine and Ponceau 4R with sodium benzoate. Parents reported more hyperactive behaviour with the mixture than with placebo; the clinic assessments did not detect that difference. It cannot isolate this ingredient’s effect.

Preschool trial: four colours plus sodium benzoate. Indexed abstract read; mixture membership cross-checked in OEHHA’s trial table. June 2004 is the journal issue month. A 2005 rejoinder and erratum is indexed (PMID 16040891); its complete correction text and the trial’s full methods remain pending.

A later trial found no significant adverse behaviour difference

In a 2013 randomized crossover trial, 130 children in Hong Kong, aged about eight or nine, received a four-colour mixture, sodium benzoate separately, and placebo capsules. Neither active capsule significantly worsened the two behaviour scores compared with placebo. The colour mixture contained E102, E110, E122 and E124.

This is useful contrary evidence, including a separate benzoate comparison. A small, short study cannot rule out a smaller effect or a response in a particular child. The abstract does not replace a full appraisal of the analyses, and the colour mixture still does not isolate one colour.

Hong Kong trial: colour mixture and benzoate tested separately. Indexed abstract read; tested colour membership cross-checked in OEHHA’s trial table. November 2013 is the journal issue month. Full methods, statistical comparisons and declarations remain pending.

What EFSA said about the Southampton study

In 2008, EFSA judged the evidence for a small mixture effect to be limited and inconsistent. It could not identify a responsible individual additive or establish the clinical importance of the changes. EFSA did not consider this study sufficient to change the acceptable daily intakes at that time. Use the separate, additive-specific assessments below for their own dates and scope.

EFSA assessment of the Southampton mixture study. Official EFSA summary and publication identity read. This is a historical study assessment, not a fresh 2026 assessment or a statement of current country permissions.

A 2025 cell study raises a question, without measuring dietary harm

Researchers applied carmoisine and Ponceau 4R directly to cultured human blood cells. Some DNA and chromosome tests changed at higher concentrations; another chromosome-damage test did not. The results were not uniformly positive across all tests.

Cells in a dish do not reproduce digestion, metabolism or the concentration reaching tissues after eating. These results cannot be converted into a claim that a normal serving causes cancer or into a personal safe dose.

Carmoisine and Ponceau 4R in cultured human blood cells. Indexed abstract and publisher-supplied online date read. Full methods and declarations remain pending. Cultured lymphocytes were exposed directly; no dietary serving was tested.

What the UK label warning means

UK food-labelling guidance includes this colour among six that require an activity-and-attention warning on relevant food labels. The wording is “May have an adverse effect on activity and attention in children”. This is a labelling requirement, not a statement that every child reacts or that all uses are banned.

UK food and drink warnings. Check the country section for the separate question of permitted food uses and their conditions.

EatMatrix’s reading

Carmoisine has both mixed-colour behaviour evidence and laboratory cell findings. They answer different questions. Neither identifies a personal dietary threshold or shows that an ordinary serving causes cancer.

A practical next step for parents

NICE does not recommend removing colours and additives as a universal ADHD treatment. If a particular food repeatedly seems linked to behaviour, discuss the pattern with the child’s clinician. A food-and-symptom diary can help; targeted dietary changes should involve a dietitian and the child’s care team. Keep the overall diet varied and nutritionally adequate.

NICE NG87, dietary advice, section 1.6.