Tesco pulled a chilled pasta salad off shelves on 25 August 2026 after tests flagged possible Shiga toxin-producing E. coli. It is the third STEC E. coli ready-to-eat food recall in the UK this month, and the pattern should worry every technical manager running a product with a raw produce component. STEC needs only a handful of cells to cause illness, and in a ready-to-eat food there is no cooking step at the customer's end to save you. The Tesco Finest Mozzarella Pearls & Pesto Orzo Salad carried use-by dates of 25, 26 and 27 August, so product was still sitting in fridges when the notice went out. Severe cases progress to haemolytic uraemic syndrome and kidney failure. This is a hazard analysis problem, and it starts with your raw materials.
What happened
Tesco is recalling Tesco Finest Mozzarella Pearls & Pesto Orzo Salad in the 215g pack, use-by dates 25, 26 and 27 August 2026, because of possible contamination with STEC (FSA-PRIN-41-2026, published 25 August 2026). The FSA risk statement lists severe diarrhoea, abdominal pain, vomiting, dehydration and, in serious cases, HUS.
This did not happen in isolation. On 21 August, Summit Foods recalled four filled wrap lines for the same reason (FSA-PRIN-40-2026). On 10 August, A.Vogel recalled Rapunzel bioSnacky Red Clover sprouting seeds after STEC was found in the seed itself (FSA-PRIN-39-2026). Three STEC recalls in a fortnight, all in foods eaten without cooking.
What went wrong
The FSA notice doesn't state the source, but the pattern points to the raw produce or seed component, not the assembly process. STEC lives in the guts of ruminants and reaches crops through contaminated irrigation water, manure or run-off. Once it is on a leaf or a sprouting seed, no chilled assembly process removes it, because there is no heat step downstream.
This is the same failure mode that drove the 2024 O145 outbreak, when contaminated apollo leaf lettuce was linked to 290 cases, two deaths and nine cases of HUS, and Greencore recalled 45 sandwich and salad lines. When your kill step sits on a farm you do not own, your hazard analysis (Codex Principle 1; BRC Issue 9 clause 2.7) has to treat the raw material itself as the point of concern.
The HACCP perspective
For a ready-to-eat salad built from raw produce, the wash step is a genuine control point, and most producers run it as a CCP or OPRP. What matters is being clear about what it controls. Maintaining free chlorine or peracetic acid at the right concentration and contact time controls cross-contamination in the wash water, and that critical limit can be monitored and validated. That is a defensible CCP.
What the wash cannot do is eliminate STEC already present on an incoming leaf, where chlorine delivers only a 1 to 2 log reduction. So the wash is not the kill step for a hazard that arrived on the raw material, and your hazard analysis (Codex Principle 2; BRC clause 2.8) has to be explicit about that. Control of the incoming hazard sits upstream in raw material assurance and agronomic controls: grower standards, water testing, harvest controls and incoming specification.
This is also where validation and verification have to stay separate. Validating the wash proves the sanitiser regime is capable of controlling cross-contamination at the set limits. It does not prove the raw material is safe. Verification then confirms both the wash and the upstream controls are working day to day. Treat the wash as evidence the leaf is clean and you have quietly moved a raw material problem onto a step that was never designed to solve it.
How this could have been prevented
You cannot sterilise a raw leaf, but you can tighten the controls around it. Checks worth running on your own site this week:
- A raw material risk assessment for every produce and seed line, with STEC named as a significant biological hazard wherever there is no kill step downstream (BRC clause 3.5 covers raw material and supplier management).
- Supplier assurance evidence that is current: agronomy standards and irrigation water testing results, not a certificate filed two years ago and forgotten.
- The wash CCP or OPRP validated for what it actually controls, cross-contamination in the wash water, with sanitiser concentration and contact time monitored and recorded rather than assumed.
- A verification or positive-release testing regime for high-risk RTE lines, with a written trigger stating what a presumptive positive stops.
- A traceability and recall route you have tested in the last 12 months, because the use-by dates on this product were still live when the alert landed.
Where SafetyCore fits
SafetyCore is built around the plan, not the production line, and that is where this kind of failure hides. Its Hazard Analysis module makes you score STEC as a biological hazard on every raw produce and seed input, with severity, likelihood and detectability recorded rather than assumed. The CCP Determination step runs the Codex decision tree, so the wash is captured for what it controls and the raw material hazard is not silently assumed away. Prerequisite Programmes are linked back to the risk assessments that justify them, which is where your raw material controls belong.
When an alert like FSA-PRIN-41-2026 lands, a Triggered Review lets you reopen the affected plans, capture the evidence and record what changed, with the Audit Trail timestamping every edit for your next BRC audit. None of this watches your line. It keeps the thinking behind your controls current and defensible.
Final thought
STEC does not care how good your factory is if the hazard walks in on a lettuce leaf. Three recalls in three weeks says the raw material is where the risk lives, and a wash step doing exactly what it was validated to do still will not save a plan that leaned on it to do more.