Food Additives

TL;DR. E numbers are not a warning; they are a filing system. E300 is vitamin C, E160a is beta-carotene, E440 is apple pectin. Every one has been through a safety assessment and has an acceptable daily intake set at typically a hundredth of the level that did nothing in animals. The genuinely contested ones are a short list: certain azo colours with a documented behavioural signal in some children, sulphites in asthmatics, nitrites in cured meat, and emulsifiers, where the microbiome evidence is early and interesting. MSG is one of the most thoroughly exonerated substances in food science and one of the most persistently feared.

1. What the E number system is

"E" means the additive is approved for use across the European Union. The number is a classification code, not a hazard rating. The same substances are used worldwide with different numbering (INS internationally, and no numbering system on US labels, which is why "no E numbers" is a marketing claim that is trivially true for an American product).

Some E numbers you already eat deliberately:

An orange juice listing "E300" is telling you it contains vitamin C.

2. What additives are for

Removing them is not free, and the counterfactual is worth stating.

3. How they are assessed

The process in Chapter 87, applied:

1. Toxicological studies establish a NOAEL.
2. Divide by 100 (or more) to set an ADI.
3. Dietary exposure modelling checks that realistic intakes, including high consumers and children, stay below the ADI.
4. Maximum permitted levels are set per food category.
5. Re-evaluation. EFSA has been systematically re-evaluating all pre-2009 approved additives, a programme that has led to several withdrawals and restrictions.

Additives have been withdrawn when evidence changed: brominated vegetable oil, potassium bromate, several azo dyes, and titanium dioxide (E171), which the EU banned as a food additive in 2022 after EFSA concluded genotoxicity could not be ruled out. Notably, the UK, US, and others did not follow, which illustrates how the same evidence can produce different regulatory judgements under different precautionary standards.

4. The ones with genuine issues

Azo colours and children's behaviour

The Southampton study (McCann and colleagues, Lancet, 2007) gave children mixtures of six azo colours plus sodium benzoate and found small but statistically significant increases in hyperactive behaviour in the general child population, not just in diagnosed ADHD.

The result: the EU requires foods containing these six to carry the warning "may have an adverse effect on activity and attention in children." Most manufacturers reformulated rather than carry the label, which is why these colours are now rare in European food and still common in the US.

The six ("Southampton Six"): E102 tartrazine, E104 quinoline yellow, E110 sunset yellow, E122 carmoisine, E124 ponceau 4R, E129 allura red.

The honest reading: the effect sizes were small, the study used mixtures so individual culprits are unclear, and EFSA's re-evaluation found the evidence insufficient to change the ADIs. The labelling requirement was a precautionary response to a real, if modest, signal. If a child seems sensitive, avoiding these is easy and harmless.

Sulphites (E220 to E228)

Genuinely problematic for a specific group. Sulphur dioxide and sulphites trigger asthma in roughly 3 to 10 percent of asthmatics, sometimes severely.

Where they are: wine (especially white), dried fruit (bright orange apricots, golden raisins), some fruit juices, pickles, processed potato products, sausages, and table grapes shipped with sulphur pads (Chapter 24).

Labelling above 10 mg/kg is mandatory in the EU, UK, US, and Australia, which is why "contains sulphites" appears on wine bottles.

Note that sulphites do not cause the red wine headache, which is more likely related to histamine, tyramine, tannins, congeners, and alcohol itself.

Nitrites and nitrates (E249 to E252)

Used in cured meat to prevent botulism (the primary reason, and a serious one), to preserve colour, and for flavour.

The problem: in the presence of haem iron and amines, and particularly at high cooking temperatures, they form N-nitroso compounds, which are carcinogenic. This is part of the mechanism behind processed meat's IARC Group 1 classification (Chapter 55).

The context that matters: the same nitrate ion in vegetables is beneficial, converting to nitric oxide and lowering blood pressure, because vegetables supply vitamin C and polyphenols that block nitrosation (Chapter 38). Vegetables supply the large majority of dietary nitrate. The ion is not the problem; the company it keeps is.

"Nitrite-free" or "naturally cured" bacon using celery powder is a labelling artefact: celery powder is a concentrated nitrate source, converted to nitrite by added bacterial cultures. The chemistry is identical, and in some products the nitrite level is higher.

Emulsifiers

The most interesting current question in this chapter. Emulsifiers such as carboxymethylcellulose (E466) and polysorbate-80 (E433) keep water and fat mixed and are in an enormous range of processed foods.

The evidence: mouse studies by Chassaing and Gewirtz found these emulsifiers thinned the intestinal mucus layer, allowed bacteria closer to the epithelium, altered the microbiome, and promoted low-grade inflammation and metabolic syndrome. A small randomised controlled feeding study in humans (2022) found CMC altered gut microbiota composition and metabolites and reduced mucus layer integrity in some participants.

How to weigh it: the mechanism is plausible, the human evidence is early and limited, and this is one of the more credible mechanistic candidates for why ultra-processed food might be harmful beyond its nutrient profile (Chapter 58). It is an area to watch rather than a settled harm.

Titanium dioxide (E171)

A white pigment used in sweets, icing, and chewing gum. Banned as a food additive in the EU from 2022 after EFSA concluded that genotoxicity from nanoparticles could not be excluded and no ADI could be set. Still permitted in the US and UK, whose agencies reached different conclusions on the same evidence. A clean example of regulatory divergence under uncertainty.

Phosphates (E338 to E452)

Used widely as acidity regulators and to retain water in processed meat. High phosphate intake is a genuine concern in chronic kidney disease, where phosphate cannot be excreted properly. Critically, additive phosphate is absorbed far more completely (around 90 percent) than the naturally bound phosphate in food (around 50 to 60 percent), so processed food contributes disproportionately. Anyone with kidney disease should be reading labels for phosphate additives, and few are told to.

5. MSG: the case that should be closed

Monosodium glutamate (E621) is the sodium salt of glutamic acid, one of the twenty amino acids, and the compound responsible for umami, identified by Kikunae Ikeda in 1908 from kombu seaweed.

Glutamate is everywhere in food naturally:

Your body does not distinguish the source. Glutamate from a tomato, from parmesan, and from MSG crystals are the same molecule.

"Chinese restaurant syndrome" originated in a 1968 letter to the New England Journal of Medicine speculating about symptoms after Chinese meals. It spread rapidly and was never substantiated.

What the research found: multiple double-blind, placebo-controlled challenge studies have failed to reproduce symptoms when participants do not know whether they received MSG. A large FDA commissioned review (FASEB, 1995) concluded MSG is safe, with a possible transient mild reaction in a small number of sensitive individuals at very high doses (3 g or more) taken without food, which is not how anyone eats.

The regulatory position: MSG is "generally recognised as safe" in the US, and EFSA's 2017 re-evaluation set a group ADI for glutamates, noting that some high consumers may exceed it, which prompted some restrictions on permitted levels.

Why it persisted is now widely acknowledged to involve the racialised framing of the original scare, which attached to Chinese food specifically while parmesan, tomatoes, and yeast extract, which contain far more free glutamate, were never implicated.

The practical upside: MSG contains roughly one third the sodium of table salt by weight and increases perceived saltiness, so partially replacing salt with MSG can reduce sodium intake by 25 to 40 percent in a dish without loss of palatability. Several public health researchers have argued for this as a salt reduction strategy (Chapter 57).

6. Reading a label sensibly

• Ingredients are listed by weight, descending. The first three tell you most of what a product is.
• Additives must be listed by function and then by name or E number: "preservative: sodium benzoate (E211)."
• "Natural flavouring" versus "flavouring": natural means derived from a natural source, and says nothing about the chemistry or the safety. The molecules are often identical.
• "No artificial colours or preservatives" frequently means natural equivalents were used instead, at similar or higher levels.
• "Clean label" is a marketing concept, not a regulatory one, and often means an additive was replaced with a functionally identical ingredient that does not require declaring, such as "cultured dextrose" instead of a named preservative.
• The nutrition panel usually matters more than the additive list. Sugar, salt, saturated fat, and fibre have far larger established effects on health than any approved additive.

7. The bottom line

• E numbers are a filing system, not a warning. E300 is vitamin C, E440 is apple pectin, E101 is riboflavin.
• Additives are assessed with safety factors of 100 or more, re-evaluated periodically, and withdrawn when the evidence changes. The system is not perfect and it is not absent.
• The genuinely contested ones are a short list: the Southampton azo colours (small behavioural signal, EU warning label), sulphites (real asthma trigger), nitrites in cured meat (real, context-dependent), phosphate additives in kidney disease, titanium dioxide (banned in the EU, not elsewhere), and emulsifiers, where the microbiome evidence is early and worth watching.
• The same nitrate ion is protective in vegetables and problematic in cured meat, because of what accompanies it. This is the food matrix principle in its clearest form.
• MSG is exonerated. It is glutamate, which is in breast milk, tomatoes, and parmesan in larger quantities, and blinded challenge studies have consistently failed to reproduce the symptoms. It is also a useful tool for cutting sodium.
• The nutrition panel almost always matters more than the additive list.

Sources and notes

The E number system follows EU Regulation 1333/2008 and the Codex INS. Additive approvals, ADIs, and the ongoing re-evaluation programme follow EFSA's published opinions. The Southampton study on azo colours and children's behaviour is McCann et al., The Lancet, 2007, with the resulting EU warning-label requirement in Regulation 1333/2008 and EFSA's subsequent re-evaluations, which did not change the ADIs. Sulphite asthma prevalence follows Vally and Misso's reviews; labelling thresholds follow EU and FDA rules. Nitrite use in cured meat for botulism control and its role in N-nitroso compound formation follow EFSA's 2017 opinion and IARC Monograph 114; the celery powder labelling issue follows published comparative analyses. Emulsifier effects follow Chassaing et al., Nature, 2015, in mice and Chassaing et al., Gastroenterology, 2022, in humans. Titanium dioxide follows EFSA's 2021 opinion and the EU ban from 2022, with the divergent FDA and FSA positions. Phosphate additive absorption relative to natural phosphate follows nephrology literature and KDIGO guidance. MSG safety follows the FASEB report to the FDA, 1995, EFSA's 2017 re-evaluation, and the blinded challenge literature reviewed by Freeman; glutamate content of foods follows published analyses; the racialised origin of "Chinese restaurant syndrome" traces to Kwok's 1968 NEJM letter and is discussed in subsequent historical reviews. MSG as a sodium reduction tool follows published sensory and public health analyses.

Open questions. Emulsifier effects on the human gut are the most active open question in this chapter, resting on one small feeding trial and a body of mouse work. Whether the Southampton colours matter to any child in ordinary dietary amounts has not been established.

👉 Next: what cooking does.