The EU has published its residue testing list for 2027, and three things on it are yours

Most enforcement is a surprise. This one is not.

Every year the European Commission publishes a coordinated multiannual control programme setting out exactly which foods member states must sample, which pesticide residues they must test for, and how many samples each country has to take. The current one is Implementing Regulation (EU) 2026/748, adopted on 31 March 2026 and applying from 1 January 2027. It covers 2027, 2028 and 2029.

It is a published list of what official laboratories will be looking for, three years ahead, commodity by commodity and substance by substance.

Almost nobody in food manufacturing reads it.

What it actually is

The programme exists because 30 to 40 products make up the bulk of the European diet, and pesticide use changes enough over three years that those products need monitoring in cycles rather than continuously.

The numbers are set by statistics rather than ambition. The European Food Safety Authority concluded that an exceedance rate above 1% could be estimated with a margin of error of 0.75% by selecting 683 sample units across a minimum of 32 different products, apportioned between member states by population, with a minimum of 12 samples per product per year.

Read the annex carefully, because the 683 is per commodity per year, not per programme. Germany takes 106 of them, France 78, Italy 75, Spain 55, Poland 51. Most of the rest take 12 or 15. Twelve commodities are sampled in 2027, so the real figure is closer to eight thousand samples across the Union that year.

And one line in that table matters more to a UK reader than anything else in the regulation.

Northern Ireland is in the programme

The table includes an entry for the United Kingdom in respect of Northern Ireland, at 12 samples a year.

That is not an oversight. A footnote to Article 1 explains it: under Article 5(4) of the Windsor Framework, read with Section 24 of Annex 2, references to member states in this regulation include the United Kingdom in respect of Northern Ireland.

Great Britain's own plan says the same thing from the other side. It states that Northern Ireland will follow the requirements of the EU control plan, and adds that the two programmes are similar but not identical.

So if you supply Northern Ireland, your product sits inside an EU official control programme with a published sampling list, and the results go to the European Food Safety Authority. And the list it is tested against is not the one applying in Great Britain.

What gets tested, and when

Annex I splits the commodities across three years. Sample in the wrong year and you learn nothing useful, so it is worth knowing which of your raw materials is up when.

Year	Plant products	Animal products
2027	Table grapes, bananas, grapefruits, aubergines, broccoli, melons, cultivated fungi, sweet peppers, wheat grain, virgin olive oil	Bovine fat, chicken eggs
2028	Apples, strawberries, peaches and nectarines, wine, lettuces, head cabbages, tomatoes, spinaches, oat grain, barley grain	Cow's milk, swine fat
2029	Oranges, pears, kiwi fruits, cauliflowers, onions, carrots, potatoes, dried beans, rye grain, brown rice	Poultry fat, bovine liver

Wheat is a 2027 commodity. Oats and barley are 2028. Rye and brown rice are 2029. If you buy cereals, your exposure is spread across all three years and you will want to know which one is live.

Two entries are processed rather than raw, which is unusual for this programme and tells you something about where the Commission expects problems. Virgin olive oil is sampled in 2027, and wine in 2028, with member states required to report the processing factors they used where none is published. Everything else is analysed unprocessed, with narrow exceptions: frozen samples carry a reported processing factor, cereal grain can be substituted with whole grain flour where insufficient grain is available, and rice may be analysed polished rather than husked provided that is reported.

The three that are not pesticides

Here is the part worth stopping on.

Most of the roughly two hundred substances in the list are what you would expect: fungicides, insecticides, herbicides. Three are not. Chlorates, benzalkonium chloride and didecyldimethylammonium chloride are tested in every commodity group, every year, in both plant and animal products, and none of them is a crop protection product in any meaningful sense.

Chlorate arises from chlorinated water. Wash water, flume water, ice, equipment rinse, anywhere chlorine-based treatment meets food.

Benzalkonium chloride and didecyldimethylammonium chloride are quaternary ammonium compounds. They are disinfectants. If your sanitation programme uses quats, they are in your factory by design.

This is the opposite of how residue risk is normally framed. The usual model is that residues arrive with the raw material and your job is supplier approval. For these three, your own process water and your own cleaning chemistry are a plausible source, alongside anything inherited from the grower or the packhouse.

Which changes who owns the problem. A chlorate result is not automatically a supplier issue to be pushed upstream. It may be a question about your water treatment, your changeover rinse, or how long product sits in contact with treated water. A quat result may be a question about rinse efficacy after sanitation, or about lubricant and hand sanitiser carryover, rather than about anything the farm did.

If you have never traced these back through your own process, the fact that they are tested in every commodity group for three consecutive years is a reasonable prompt. The Great Britain plan treats them differently, which is worth knowing before you scope any testing. More on that below.

Organic is not exempt

Annex II is explicit. Samples from products originating from organic farming are taken, where available, in proportion to the market share of those products in each member state, with a minimum of one, and analysed for the same full list of substances.

Organic certification is not a defence and was never meant to be. The programme treats organic product as ordinary product for testing purposes.

Baby food has its own regime inside this one

On top of the 683 samples, every member state must take a separate set aimed at infants and young children, and the set rotates:

• 2027: ten samples of processed cereal-based baby food
• 2028: ten samples of foods for infants and young children other than formulae and processed cereal-based baby food
• 2029: five samples of infant formulae and five of follow-on formulae

These are assessed against the specific maximum residue levels for infant and young child food set in Delegated Regulations (EU) 2016/127 and (EU) 2016/128 and Directive 2006/125/EC, not the general levels. Where a product can be eaten both as sold and reconstituted, results are reported on the product as sold.

If you make anything in this space, the ordinary residue limits are not the limits you are being judged against.

One trap in the detail

Buried in the substance list is a note that will catch laboratories and specifications alike.

Regulation (EU) 2024/2619 changed the residue definition for enforcement purposes across fosetyl, potassium phosphonates and disodium phosphonate. It used to be the sum of fosetyl, phosphonic acid and their salts, expressed as fosetyl. It is now phosphonic acid and its salts, expressed as phosphonic acid.

The control programme states the consequence plainly: findings of fosetyl are not to be considered, and the maximum residue level for its metabolite phosphonic acid applies instead.

So a certificate of analysis reporting fosetyl is reporting against a definition that no longer governs in the EU. If your specifications name fosetyl, they are naming the wrong analyte for that market.

Great Britain has not followed. Its plan lists Fosetyl-Al plainly, with no equivalent note, and points to the Great Britain maximum residue level register for the levels themselves. So the same test result can be read two different ways depending on which market the product is going to, which is exactly the sort of divergence that gets discovered during a customer complaint rather than before one.

Great Britain runs its own version, and it is not the same

Great Britain has its own coordinated multiannual control plan, covering 2026, 2027 and 2028. The Health and Safety Executive delivers it as competent authority on behalf of the administrations, not the Food Standards Agency, which surprises people. It covers 36 foods on the same three-year rolling basis, and the levels come from the Great Britain maximum residue level register rather than the EU annexes.

The scale is the obvious difference. Seventy-one samples per commodity, combined across the competent authorities, against 683 in the EU.

The commodities are not a difference at all. For 2027 and 2028 the two lists are identical, plant and animal, right down to virgin olive oil in 2027 and wine in 2028. The baby food rotation matches for those two years too. If you are comparing the two programmes, do not start with the commodity table, because it will tell you they agree.

Start with the substance list, which is where they part company. Three examples that matter.

Chlorates. The EU tests for them in plant products and animal products. Great Britain tests for them in animal products only. They do not appear in the plant list at all. So a GB manufacturer of fruit, vegetables or cereals is not being sampled for chlorate under the GB plan, and the same product going to the EU or Northern Ireland is.

Copper compounds. Same pattern. In the EU plant list, absent from the GB plant list, present in both animal lists.

Fosetyl. As above. The EU has moved the residue definition to phosphonic acid. The GB plan has not.

Benzalkonium chloride and didecyldimethylammonium chloride are on both plant lists, so the quat half of the earlier point holds in both markets. The chlorate half does not.

None of this means the GB approach is wrong. It means a residue testing schedule built from one programme does not cover you for the other, and that dual-market sites need to work from both documents rather than assuming alignment because the commodity lists look the same.

One more gap. The GB plan runs out at the end of 2028. The EU programme already covers 2029. Until the next GB plan is published there is no GB schedule for that year.

What it finds when it runs

None of this is theoretical, and you can read the results. The Health and Safety Executive publishes quarterly. The report for the first quarter of 2026, covering samples taken between January and the end of March, was published in August 2026.

That quarter sought up to 428 different pesticides. In Great Britain it took 662 samples across 23 foods. Residues were found in 318 of them, and 49 were above the maximum residue level. In Northern Ireland, sampled against the EU programme, 125 samples across 14 foods gave 56 with residues and 5 above the limit.

Read those numbers carefully, because the monitoring programme is wider than the coordinated plan. Surveys carrying an asterisk in the report are the ones the plan requires; the rest are chosen because someone had a reason to look. So the exceedance rate is not a clean estimate of how often food in general is over the limit, and it is not comparable to the 1% the European design was built to detect.

Three findings from that quarter are worth carrying into your own risk assessment.

Eight organic samples contained residues not permitted under the assimilated organic regulation: one dried bean sample with glyphosate, three more with fosetyl-Al, one dried grape sample with cypermethrin, one rice sample with isoprothiolane, and two rye flour samples with chlormequat and mepiquat. Organic is not a residue control, and this is what that looks like in practice rather than in principle.

Copper residues were above the limit in every Great Britain liver sample. The Executive assessed long-term exposure and concluded no health effect was expected, but "all samples" is the kind of result that says something about a commodity rather than about a supplier.

And one sample, of dried brown beans, carried lambda-cyhalothrin above the limit alongside other pesticides. On a combined exposure assessment the Executive concluded the exposure was undesirable and that some people might experience adverse effects, expected to be short-lived and reversible. That was the only such conclusion in the quarter, out of 787 samples across both programmes.

What the list does not tell you

The three-year cycle is reissued every year. This regulation repeals the previous one, though that one continues to apply until 1 September 2027 for samples taken in 2026. Which means 2027 and 2028 have now been scheduled twice, and 2029 may be rescheduled before it arrives. Treat the far year as a plan rather than a commitment. The Great Britain plan is caught by this. It points Northern Ireland at the EU programme and names it, and the one it names is the previous regulation rather than the current one. The signpost is right and the destination has moved, so read it as pointing at whichever EU programme is live rather than at the number printed on the page.

One more thing worth knowing about the direction of travel. Maximum residue levels have historically been set on consumer safety grounds. In 2023 the Commission lowered the levels for two neonicotinoids to the limit of quantification on environmental grounds instead, to avoid European consumption contributing to pollinator decline, and stated that this was the first time it had done so. A residue can now be driven to effectively zero for reasons that have nothing to do with whether it is safe to eat. That is a different risk model, and vulnerability assessments built only around toxicity will not anticipate it.

The Monday morning list

1. Find your raw materials in the commodity table and write down which of the three years each one falls in. Cereals in particular are split across all three.
2. Pull your last two years of residue results for chlorates, benzalkonium chloride and didecyldimethylammonium chloride. If you have never tested for them, that is the answer to the question.
3. For any positive or near-limit result on those three, work backwards through your own process water and sanitation before raising it with the supplier. It may not be theirs.
4. If you sell into both markets, put the two plans side by side on substances rather than commodities. Chlorates and copper compounds are in the EU plant list and not the Great Britain one.
5. Search your raw material specifications for fosetyl. For EU-bound product it should read phosphonic acid, expressed as phosphonic acid. For Great Britain, check the Great Britain maximum residue level register before changing anything.
6. Check what your laboratory actually reports against, and for which market. A certificate naming a superseded residue definition is not evidence of compliance.
7. If you buy organic, remove "certified organic" from your residue risk justification. Both programmes test it the same way as everything else.
8. If you make food for infants or young children, confirm which limits your specifications reference. The general maximum residue levels are not the ones that apply.
9. If you supply Northern Ireland, add the EU programme to your regulatory register. Your product is inside it, and the Great Britain plan says so itself.
10. Read the quarterly monitoring report for the commodities you buy. It is free, it names what was found and in what, and it is the closest thing available to a list of where the problems currently are.

I work in quality in UK food manufacturing. This is information, not advice. Always check the source document before making decisions that matter.