FSMA and HACCP certification for bio-based food

HACCP is often described as a paperwork system. It is closer to the opposite: it is the discipline of refusing to accept any control you cannot measure while the food is being made. Everything distinctive about it follows from that one rule.

A critical control point is a measurable parameter with a validated limit

A step in a process qualifies as a critical control point only if three things hold at once. The step must eliminate or reduce a hazard to an acceptable level; there must be a parameter — temperature, time, pH, water activity, flow rate — that responds to loss of control; and that parameter must be readable at a frequency fast enough to isolate the affected product when it drifts. A step that is merely good practice but cannot be monitored in real time is a prerequisite programme, not a control point. This is why sanitation, pest control and supplier approval sit outside the plan’s critical points: they matter, but a failure in them is discovered after the fact.

The critical limit itself is not chosen; it is inherited from microbial death kinetics. For low-acid canned food, the reference is the 12-D botulinum cook — twelve logarithmic reductions of Clostridium botulinum spores, conventionally delivered as three minutes at 121.1 °C with a z-value of 10 °C, so that any equivalent time-temperature combination can be computed rather than argued. Juice processing under the US juice HACCP rule works to a 5-log reduction of the pertinent pathogen. These numbers, not the audit, are what a certificate is ultimately asserting.

Where the control actually lives

Thermal process is the visible control; the quieter ones are compositional. Clostridium botulinum does not grow below about pH 4.6 or below a water activity of roughly 0.93, and US regulation draws its definitional line for low-acid foods at pH 4.6 and water activity 0.85. A product formulated below those thresholds is not being sterilised — it is being made uninhabitable, permanently, by its own composition. That is a far more robust control than a heating step, because it does not depend on an instrument working, and it is the reason so many shelf-stable products are acidified or dried rather than retorted.

The novel ingredient problem

This is where a bio-based ingredient runs into difficulty that has nothing to do with certification scheme choice. A scheduled process rests on a process authority’s determination, and that determination rests on measured thermal resistance — D and z values — for the relevant pathogen in that matrix. A precision-fermented protein, a mycelial biomass, an insect flour or a novel oil has no such data, and resistance is matrix-dependent: fat, sugar and low moisture all protect the organism, sometimes by an order of magnitude.

So the work is a validation study, not a filing. It means inoculated-pack trials or an accepted surrogate — Enterococcus faecium NRRL B-2354 is the usual stand-in for Salmonella in low-moisture processes — run in the real product on the real line. Until that exists, the plan has a critical limit with nothing underneath it, and no auditor can supply the missing number.