Probiotics: the CFU count you buy is not the one your customer gets

Probiotic potency is the one specification that is guaranteed to be wrong by the time the product is used. The only question is by how much, and whether that was planned for.

At manufacture versus end of shelf life

Live cultures die steadily. A count taken at blending is the highest number the material will ever show. The figure that matters commercially is the count you can still guarantee on the last day of the stated shelf life, and the two can differ by a factor of two or more.

As with vitamin overages, the industry answer is to overdose at manufacture. A product labelled 10 billion CFU may be built at 20 to 30 billion. That is normal practice — but it means a quote comparing "100 billion CFU/g" against "100 billion CFU/g" may be comparing an at-manufacture figure against an end-of-shelf-life guarantee, which is not a comparison at all.

What drives the die-off rate

• Water activity — the dominant factor. Below about 0.1 is where the good stability data comes from, and any hygroscopic excipient in the blend works against you.
• Temperature — both storage and, critically, the uncontrolled part of the journey.
• Strain — spore-formers such as Bacillus coagulans are dramatically more robust than most Lactobacillus strains, which is often the deciding factor for ambient-stable products.
• Oxygen and packaging — foil versus HDPE changes the curve, as does a desiccant.

What to put on the specification

• Guaranteed CFU at end of shelf life, with the shelf life and storage conditions it assumes. Insist on this figure rather than the at-manufacture count.
• Strain designation, including the deposit number — not just the species. L. rhamnosus is a species; the strain is what the research and the regulatory dossier attach to.
• Enumeration method, and whether the count is plate-based or flow cytometry. Flow cytometry reports active fluorescent units and generally gives higher numbers than plate counts; the two are not interchangeable.
• Water activity on release.
• Carrier and its share of the blend, since a large part of what you are buying by weight is usually maltodextrin or similar.

Ask for the stability curve

A supplier with real data will have a die-off curve at 25°C/60% RH and often accelerated data alongside it. That curve, not the headline CFU, is what tells you whether the material will still meet its claim when it matters.