Live Culture Survival Trial: A Worked Example for a Fermented Probiotic Beverage

About this worked example

This is a labelled worked example, not a client case study. The product, the brief and every number on this page are assumptions or calculations from those assumptions. No client data, measured counts or trial results are shown. It sets out how Food Research Lab would plan a live culture survival trial for a refrigerated fermented beverage sold in the US, using methods from published beverage studies^[1][2][3] and FDA guidance on refrigerated foods and label claims.

The nearest FRL lab notes are our development write-ups for a clean-label fermented kefir and a naturally flavoured kombucha. Neither reports culture survival data, so neither is used as evidence here.

The brief: a refrigerated fermented oat drink

The assumed brief is typical for a plant-based probiotic launch. A US brand wants a refrigerated, oat-based fermented drink with added protein and a live-culture count on the label. Before it commits to a co-manufacturer, it wants evidence that the count still holds on the last day of shelf life, that the drink stays inside its pH window, and a filling dose it can write into the production specification.

Table 1. Worked-example brief (all values assumed)

The per-mL figure matters because plate counts are reported per millilitre, not per serving. Converting the claim first avoids a common error when the trial data are read later. Our guide to what CFU means in probiotics explains how counts are tested and declared.

Why live cultures are lost in a fermented beverage

A chilled fermented drink is wet and acidic, and its cultures stay metabolically active. Reviews of probiotic food systems list matrix acidity, oxygen in the product, other lactic acid bacteria and their metabolites among the main causes of viability loss in storage.^[2] Continued acid production after filling, known as post-acidification, adds to that stress.^[4]

The base also matters. In a 2024 study of probiotic milk and oat beverages with added protein, probiotic counts fell sharply in the simulated gastric phase in both matrices, and the authors concluded that the oat matrix sustained probiotic populations less well through digestion than milk.^[1] Base and strain therefore have to be chosen together.

Product, strain and pack

Three inputs are fixed before the trial. The strain must have a documented regulatory basis for food use in the US; FDA’s partial list of microorganisms used in food is a starting point, but FDA states it is not a complete list of GRAS microbial ingredients and is not updated regularly.^[5] Its acid and bile tolerance is screened first, as in published work on fermented soy beverages.^[3] The pack is an oxygen-barrier bottle, because oxygen is one of the listed viability stresses.^[2] If the brand later wants an ambient format, protective carriers and encapsulation become part of the design.^[2]

Designing the live culture survival trial

The trial tests the finished drink in its final pack. It uses three production-representative lots, real-time storage at 4 °C for the full 12 weeks and an accelerated arm at 25 °C for four weeks, with three replicate packs per lot at every pull point (Figure 1). Table 2 sizes the plan.

Table 2. Sample plan (calculated from the assumptions)

A few spare packs per lot are worth adding for retests and sensory checks.

Viability testing and storage conditions

At each pull, the culture is enumerated by plate count on a medium suited to the strain, reported as log₁₀ CFU/mL, alongside pH and a short sensory check. Storage temperature is the main variable: 4 °C represents the cold chain, while 25 °C exposes weak formulas early. The accelerated arm is a screening tool; the label claim is supported by the real-time data. Our article on real-time vs accelerated shelf-life testing explains why.

If the brand also wants evidence of survival through digestion, finished-product samples can be run through a simulated oral, gastric and intestinal model, as in the published milk and oat study, where testing followed seven days of storage at 5 °C.^[1] That is a separate question from survival in the pack and needs its own budget.

Setting the filling dose: a worked calculation, not results

No measured results are shown in this worked example. What can be calculated in advance is the count needed at filling. If the culture loses a given number of log₁₀ units over 12 weeks, the filling count equals the end-of-shelf-life claim multiplied by 10 raised to that loss (Table 3).

Table 3. Filling count needed to meet 1 × 10⁹ CFU per 250 mL at week 12

Source: FRL calculation from the stated assumptions.

In a live project, the real loss would be read from mean log₁₀ CFU/mL for each lot at each pull, and the dose set from the worst lot, not the average. The acceptance rule is agreed before the first pull: for example, every lot at or above 6.6 log₁₀ CFU/mL at week 12.

Formulator’s note. Reading a survival trial is about the lower bound, not the average. One lot that slips below target at week 12 matters more than two that clear it, so fix the acceptance rule before the first pull.

Managing post-acidification and pH drift

Lactic acid bacteria keep producing acid during chilled storage. A review of fermented milks describes this post-acidification as a cause of flavour change, lower consumer acceptance and shorter shelf life, and identifies the starter culture as the main factor, with strain-to-strain differences in acid production and acid tolerance.^[4] The same mechanism applies to a fermented oat drink.

The trial therefore tracks pH against the 4.2 to 4.6 window at every pull. If CFU or pH falls out of specification, the decision tree in Figure 2 applies: raise the dose or change the strain if the count fails; adjust the fermentation end point, the base or the strain, and keep the product cold, if the pH fails. Each change means repeating the affected arm.

From trial to pilot runs and supply

Once CFU and pH hold across all three lots, the strain, dose, fermentation end point and pack go into pilot runs, where shelf life is confirmed on product made the way the factory will make it. Our guide to fermented food scale-up covers what changes between the bench and the plant.

Cold storage is a safety control as well as a viability one. FDA’s guidance on low-acid canned and acidified foods says a product that is stored, distributed and retailed refrigerated is not covered by 21 CFR Part 113 or Part 114,^[6] so the cold chain and its labelling carry the safety case. Any health wording must be a truthful, substantiated structure/function claim for a conventional food;^[7] our article on probiotic claims in the EU and US and our guide to US rules for fermented foods and drinks cover the detail.

The method behind this example is set out in full in our pillar guide, formulating fermented foods and drinks with live cultures.

Takeaways for your probiotic beverage

• Set the dose from the end-of-shelf-life claim, converted to CFU/mL, not from the culture supplier’s count.
• Test the finished drink in its final pack, at two temperatures, across three lots.
• Treat pH and CFU as linked, because post-acidification drives both.^[4]
• Fix the acceptance rule before the first pull, and judge the worst lot, not the average.
• Use accelerated data to screen, and real-time data at 4 °C to support the label.

For brands & formulators

If you are developing a probiotic or fermented drink for the US market, FRL’s Microbiome Solutions team designs the survival trial with you: strain and base selection, the sample plan, the acceptance rule and the filling dose. Our food product development team takes the formula from bench to pilot runs.

We run or arrange shelf-life studies, culture enumeration and other microbiological testing through accredited partner laboratories, and our regulatory and scientific research team reviews label counts and claims. FRL works from Chennai, India, with brands in the US and elsewhere; production runs at your manufacturer or a partner we introduce.

In this series

• Stanford’s fermented-food study explained: what the Wastyk et al. (2021) trial found on fermented foods and the gut microbiome.
• Fermented food scale-up: taking a fermentation from bench to plant.
• Formulating live-culture fermented foods: the pillar guide to strains, bases, dosing and stability.
• US fermented food and drink regulations: alcohol limits, labelling and claims.
• Live culture survival trial: fermented beverage worked example: trial design, filling dose and pH drift. You are here.
• Fermented foods and drinks FAQ: cultures, safety, fizz and shelf life answered.
• Service: Microbiome Solutions at Food Research Lab.

1. Pawlos, M., Szajnar, K., & Znamirowska-Piotrowska, A. (2024). Probiotic milk and oat beverages with increased protein content: Survival of probiotic bacteria under simulated in vitro digestion conditions. Nutrients, 16(21), 3673. https://doi.org/10.3390/nu16213673 (cited for milk vs oat matrices, storage at 5 °C before testing and the loss of viability in the simulated gastric phase).
2. Terpou, A., Papadaki, A., Lappa, I., Kachrimanidou, V., Bosnea, L., & Kopsahelis, N. (2019). Probiotics in food systems: Significance and emerging strategies towards improved viability and delivery of enhanced beneficial value. Nutrients, 11(7), 1591. https://doi.org/10.3390/nu11071591 (cited for the causes of viability loss in foods and protective strategies such as strain selection and immobilisation).
3. Rana, A., Taneja, N. K., Raposo, A., Alarifi, S. N., Teixeira-Lemos, E., Lima, M. J., Gonçalves, J. C., & Dhewa, T. (2024). Exploring prebiotic properties and its probiotic potential of new formulations of soy milk-derived beverages. Frontiers in Microbiology, 15, 1404907. https://doi.org/10.3389/fmicb.2024.1404907 (cited for acid and bile tolerance screening of strains in a plant-based beverage).
4. Guan, Y., Cui, Y., Qu, X., Li, B., & Zhang, L. (2025). Post-acidification of fermented milk and its molecular regulatory mechanism. International Journal of Food Microbiology, 426, 110920. https://doi.org/10.1016/j.ijfoodmicro.2024.110920 (cited for the definition, causes and effects of post-acidification during chilled storage).
5. U.S. Food and Drug Administration. (2018). Microorganisms & microbial-derived ingredients used in food (partial list). https://www.fda.gov/food/generally-recognized-safe-gras/microorganisms-microbial-derived-ingredients-used-food-partial-list (cited for checking a strain’s food-use status and the limits of the list).
6. U.S. Food and Drug Administration. (n.d.). LACF/AF: Precursor questions and flow. https://www.fda.gov/media/89861/download (cited for refrigerated products falling outside 21 CFR Parts 113 and 114).
7. U.S. Food and Drug Administration. (n.d.). Structure/function claims. https://www.fda.gov/food/nutrition-food-labeling-and-critical-foods/structurefunction-claims (cited for structure/function claims on conventional foods).