Belgium’s federal public‑health institute Sciensano has identified a critical gap in how infant formula is tested: standard laboratory methods that analyse powdered formula can dramatically understate levels of a heat‑stable bacterial toxin once the powder is mixed into liquid. Measurements run by Sciensano found that cereulide, the toxin produced by Bacillus cereus, can appear at concentrations as much as 75 times higher in reconstituted formula than in the dry powder.
Scientists at Sciensano say the discrepancy stems from the way certain formula ingredients — in particular preparations of arachidonic acid (ARA) — are physically formulated. Some ARA preparations are microencapsulated, effectively enclosing oil droplets in a protective coating. In dry powder form those capsules can “lock” cereulide inside and prevent it from being released during routine testing, but they dissolve when the powder is mixed with water, releasing the toxin into the liquid consumers actually ingest.
The finding has immediate consequences for the recent wave of global recalls by major dairy groups including Nestlé, Danone and Lactalis, many of which trace the suspected contamination to a common ARA supplier. Nestlé has pushed back, saying the ARA used in its recalled products was supplied as a pure oil rather than microencapsulated material and that over 400 ARA oil samples it tested in the last three years contained no cereulide. The company also cautioned that cereulide contamination is uncommon and that testing every conceivable risk in perpetuity is not feasible.
Regulators are already responding. The discovery has prompted Sciensano to revise its laboratory methods and to call on European laboratories to do the same. At the same time the European Food Safety Authority has, for the first time, proposed an acute reference dose for cereulide in infants — 0.014 micrograms per kilogram of body weight — a threshold that, if widely adopted, would tighten acceptable exposure limits and likely accelerate regulatory screening and product withdrawals.
The episode exposes broader weaknesses in food‑safety regimes: many countries still lack harmonised standards for cereulide detection and firms commonly rely on internal test protocols. The supply‑chain dimension — a single ARA supplier serving multiple global brands — highlights how contamination at the ingredient level can cascade into multinational recalls and reputational damage.
For consumers and policymakers the practical lesson is straightforward. Safety assessments must evaluate products in the state in which they are consumed, not only as manufactured. That will require updating laboratory protocols, rethinking the use and formulation of functional ingredients such as microencapsulated oils, and strengthening supplier audits and traceability. The costs of more stringent testing and tighter raw‑material controls will be borne by manufacturers and, ultimately, consumers, but the alternative is persistent blind spots that undermine confidence in a market where the stakes — infant health — are exceptionally high.