← Back
A hand in a blue lab glove drips something from a pipette onto a bright pink pudding on a child's plate.

VII · From Lab to Plate

The sweet carries the warning. The children's syrup doesn't.

Pink. Neon orange. Glaring yellow.

Colours your child won't find in any garden. Only on the shelf.

Since 2010, six of these colours have had to carry a warning: may have an adverse effect on attention in children. On the sweet. On the pudding. On the fizzy drink.

And in fever syrup for children? The same colour. No warning.

Same substance. Same child. Different law.

I didn't believe it. Then I read the patient leaflet.

Three findings. One of them turns out differently from what I expected. It's still here. If you only find what you're looking for, you're not researching. You're just proving yourself right.

The warning that stops at the pharmacy door

You've already read it further up: titanium dioxide is banned in food and allowed in medicines. Something has happened since. The 2022 regulation came with a condition — the matter was to be reviewed again within three years. That review is done. On 4 August 2025. Result: E 171 stays in medicines. With no new deadline.

But the colour isn't the sharpest point. I only found that one when I kept digging.

Since 2010, six colours in food law have carried a mandatory sentence on the pack. It's in Annex V of the food additives regulation and reads, word for word: "May have an adverse effect on activity and attention in children." Tartrazine, Quinoline Yellow, Sunset Yellow FCF, Azorubine (Carmoisine), Ponceau 4R, Allura Red AC.

In medicines law, that sentence doesn't apply. The relevant annex from the European Medicines Agency requires exactly one line for the same colours: may cause allergic reactions. Not a word about children. Titanium dioxide and Quinoline Yellow aren't listed there at all — no warning is required for them whatsoever.

A sweet with Sunset Yellow FCF carries the warning. A fever syrup for children with the same colour doesn't.

This isn't a thought experiment. I opened the summary of product characteristics for an over-the-counter paracetamol syrup for children, dated January 2025, section "Other ingredients" ("Sonstige Bestandteile"):

Children's syrup, no prescription, pharmacy

Purified water, Sucrose (sugar), Sodium citrate, Xanthan gum, Citric acid monohydrate, Cream flavour, Methyl parahydroxybenzoate (E 218), Propyl parahydroxybenzoate (E 216), Sunset Yellow FCF (E 110).

One Annex V colour and two parabens. The required warning reads: "may cause allergic reactions."

And now the part that makes this fair, because it proves there's no necessity: there's another children's fever syrup on the same shelf, with no colour, no titanium dioxide, no preservative. It can be done differently. Someone just has to want to.

That's exactly what a German manufacturer from Gießen did. It started taking titanium dioxide out of its products in 2018 and was finished by 2021. The entire cost of the switch is in its own announcement: one of the products is now dark blue instead of light blue.

That was the price. A different colour.

The wrong suspect

A lot gets written about emulsifiers in meat substitutes. Always the same two names: polysorbate 80 and carboxymethylcellulose. I wanted to know how often they're really in there, so I counted — category meat substitutes, Germany, 1,943 products recorded.

What's actually on the shelf

Methylcellulose · E 461 = 515 products · 26.5%

Carrageenan · E 407 = 233 · 12.0%

Xanthan gum · E 415 = 135 · 6.9%

Carboxymethylcellulose · E 466 = 4 · 0.2%

Polysorbate 80 · E 433 = 0 · zero products

The two substances everyone talks about are hardly on the shelf at all.

By far the most common one is methylcellulose. One product in four. So I searched for what it does to the gut flora — in the European study database, systematically.

Zero hits. There isn't a single study in humans on it. The European authority assessed the celluloses as a group in 2018, explicitly by read-across from chemically related substances — not individually.

And here I have to set something straight that you'll read everywhere. The famous study saying emulsifiers destroy the gut flora is from 2015 and was a mouse study. It covered carboxymethylcellulose and polysorbate 80 — exactly the two that hardly turn up here. In 2026 there was a controlled trial with 60 people and five emulsifiers. Result: no differences in the inflammation markers.

I'm adding this because otherwise you're left standing there naked the first time someone contradicts you. Anyone who cites 2015 and leaves out 2026 will get nailed for it — and rightly so.

The finding isn't: "This makes you ill." The finding is: the substance that's in one product in four has never been studied for this. Not refuted. Untested.

One more number that has nothing to do with chemistry. A well-known vegan meat patty (Frikadelle) contains 2.0 g salt per 100 g. According to the official 2025 product monitoring, the average for meat patties made from meat is 1.60 to 1.63 g. To be fair: across all 964 meat substitute products recorded, the average is 1.60 g, which is below sausages and cold cuts. So it's not the category that matters but the individual product. Which is exactly the point.

The Thirteen

Up to here it was about what's in things. Now the opposite — and this is the part nobody can argue away, because it isn't an opinion, it's a table in the law.

Annex II of the EU food additives regulation contains a list of foods in which additives may not be used. Not "should not". May not.

  1. Unprocessed foodstuffs excluding meat preparations
  2. Honey
  3. Non-emulsified oils and fats of animal or vegetable origin
  4. Butter
  5. Pasteurised and sterilised (including UHT) milk unflavoured
  6. Unflavoured live fermented milk products not heat-treated after fermentation
  7. Buttermilk unflavoured
  8. Natural mineral water and spring water
  9. Coffee excluding flavoured instant coffee
  10. Leaf tea unflavoured
  11. Sugars
  12. Dry pasta excluding gluten-free and/or pasta intended for hypoproteic diets
  13. Foods for infants and young children

Around 320 additives are authorised in the EU. In these thirteen categories: none.

For honey, the German regulation even says it word for word: "No substances other than honey may be added to honey." And with water, a single word separates two worlds: natural mineral water may contain nothing added except carbon dioxide. Table water may be blended and may contain additives. Both stand side by side on the shelf.

Not restraint. A ban.

And now the catch I have to spell out, or the whole argument falls over: with these thirteen, a short ingredients list proves nothing at all. It's mandatory. A list only starts to tell you something where something has been processed.

The crucial fault line is in the brackets after number one. Excluding meat preparations. As soon as meat becomes a preparation — marinated, seasoned, salted — the ban no longer applies. Then phosphates, nitrite curing salt and antioxidants are allowed. The same meat. One step in between.

Check it yourself

Don't believe me. Look it up.

Regulation (EC) No 1333/2008 on food additives, Annex II Part A Table 1 and Annex V · Commission Regulation (EU) 2022/63 · European Commission working document of 4 August 2025 on the review of E 171 · European Medicines Agency annex on excipients · Summary of product characteristics ben-u-ron® Saft 40 mg/ml, January 2025 · Manufacturer statement from Pascoe Naturmedizin on the 2018–2021 switch · Count via Open Food Facts, category meat substitutes, Germany · Chassaing et al., Nature 2015 · Wellens et al., Clin Gastroenterol Hepatol 2026 · EFSA 2018 on celluloses · Max Rubner-Institut product monitoring 2025 · German Honey Regulation (Honigverordnung) § 2 · German Mineral and Table Water Regulation (Mineral- und Tafelwasser-Verordnung).

The count comes from an open database maintained by volunteers — not a representative sample. The order of magnitude matches every ingredients list I checked individually.