Cereulide in infant formula: a wake-up call for food safety

Woman taking powdered infant formula with scoop from can on light blue background, top view
Cereulide in infant formula: a wake-up call for food safety (Getty/Liudmila Chernetska)

Hidden toxins can slip through even the safest food systems. The recent cereulide incident shows why manufacturers must rethink what they monitor


Hidden food safety risks in modern production – summary

  • Food safety threats can emerge early and persist through processing
  • Heat‑stable toxins may survive after microbes are fully removed
  • New technologies and ingredients introduce unfamiliar and complex safety risks
  • Detailed risk assessments identify realistic microbial threats and toxin pathways
  • Verification experiments ensure mitigation strategies work under real conditions

Food safety is king. Anyone in the industry will tell you that. But if there’s a lesson to learn from the recent cereulide in infant formula case, it’s that food safety threats aren’t always obvious.

New ingredients and processing conditions can bring new food safety risks. Potential pathogens are widespread, and some produce toxins can outlast the organisms that create them.

A thorough risk assessment is therefore essential to understand which threats are relevant and what manufacturers need to monitor in production.

The recent recall of infant formula contaminated with cereulide toxin made headlines globally. Cereulide is produced by some strains of Bacillus cereus. But B. cereus wasn’t present in the formula. Rather, it seems likely that the bacteria grew – and produced the toxin – early in the production chain and were removed during downstream processing. However, the toxin was passed on and, being highly heat stable, it wouldn’t have been destroyed by later heat treatment processes.

Threats outlast microbes

This shows just how complex food safety is today. Like many potentially pathogenic microorganisms, B. cereus can be found almost anywhere – including soil, raw materials and food production facilities. But, as in the cereulide case, although the microorganisms may be long gone, the toxins they produce can still be left behind.

The obvious answer is to test for the presence of the toxin as well as the microorganism. Again, the cereulide case highlights a challenge. B. cereus wouldn’t be expected to grow in formula powder or its ingredients, so testing for cereulide isn’t an obvious step.

And there are many other ubiquitous microorganisms that could contaminate a production chain and produce heat-stable toxins, including Bacillus licheniformis, Staphylococcus aureus and Clostridium botulinum, not to mention numerous fungi that can produce mycotoxins. Continuously testing for all possible toxins would be expensive, if not impossible.

New production technologies, new food safety risks

The challenge is amplified by recent trends in food production. Precision fermentation, circular ingredient streams and biomass-based processes are changing the risk profile for food safety. Fungus-based fermentation often requires temperatures and pH levels that could allow Bacillus species to flourish. High-nutrient substrates can support rapid microbial growth after even small contamination events. And minor changes in temperature, pH or processing time – whether deliberate or accidental – could tip a process into the danger zone for unwanted microbial growth.

So, how can food and ingredient manufacturers maintain food safety? By carrying out a detailed food safety risk assessment before production begins.

Assessing the real risks for food safety

That means mapping every step of the (planned) production process – from raw materials and fermentation conditions to downstream processing and storage. The goal is to understand which microbes are realistic threats, where they might grow and produce toxins, and whether those toxins could reach the final product.

Such an assessment is a multidisciplinary effort. It requires deep knowledge of organism-specific behaviour. For example, different B. cereus strains vary dramatically in their growth rates at low temperatures and how much toxin they produce.

It also takes understanding of the relevant processing technologies – including those used after production streams are cleared of microbes. The choice of, say, how an ingredient is produced could determine whether a toxin is passed on or is discarded in the waste stream.

Modelling tools are invaluable here. They allow multiple pathogen risks to be assessed using realistic growth and inactivation kinetics and process variations. This helps ensure a full assessment of possible risks and avoid any nasty surprises in production.

Data-driven mitigation

A thorough risk assessment helps identify the hazards and critical control points in the production process – moments where conditions could tip in favour of unwanted microbial growth or toxin production. Based on that insight, companies can instigate appropriate monitoring and control measures to prevent it happening or adjust downstream processing options (such as extraction or membrane filtration) to stop toxins reaching the final product.

Verification: an often overlooked step

Model-based risk assessments are the best way to identify potential threats and develop mitigation strategies. The next step – and one that is sometimes overlooked – is to verify the results. This can be done by mimicking production set-ups in the lab and introducing appropriately chosen microbial strains to track growth, toxin production and the passage of toxins through the process. Proper verification gives manufacturers confidence that controls and mitigation measures will actually work under realistic conditions.

Understand where the risk comes from

Modern food safety is about more than just keeping microbes out of a product. Manufacturers have to understand how they behave across entire processes and recognise that toxins can survive long after the organisms are gone.

New ingredients, new technologies and new production conditions are constantly reshaping risk: thorough, science‑based risk assessments can reveal the weak spots and show whether controls are truly robust. The recent infant formula case was unusual, but it highlighted a universal truth: the threats that matter most are often the ones we don’t think to look for.