Why CFU Counts in Probiotics Drop Long Before the Expiration Date

Roughly 1 in 3 probiotic products tested at retail can’t meet their label-claimed CFU count — even when the product is well within its stated shelf life. That’s not a rounding error. That’s a fundamental stability problem that starts at the manufacturing stage and compounds every time the product sits in a warm warehouse, gets shipped cross-country in the summer, or lands on a store shelf under fluorescent lighting.

I’ve spent years watching brands launch high-potency probiotic SKUs with ambitious claims — 50 billion, 100 billion, sometimes 200 billion CFU — without any real stability data to back them up. The shelf life date on the label is, in many of those cases, a guess. A legally protected guess, but a guess.

The “At Manufacture” vs. “At Expiration” Label Problem

The single biggest source of confusion in probiotic labeling is the distinction between CFU “at time of manufacture” (ATM) and CFU “at time of expiration” (ATE). Most consumers assume the number on the label reflects what they’re getting when they open the bottle. That’s rarely true.

When a manufacturer prints “10 billion CFU” on a label and that count is ATM, the product might contain as few as 2–3 billion viable organisms by the time it reaches the consumer. A 70–80% decline in viable count over 18 months of ambient storage isn’t unusual for poorly formulated products. Some strains — particularly Bifidobacterium species — are even more sensitive, with some studies documenting log-scale drops within 6 months at 25°C.

The FDA doesn’t currently require probiotic dietary supplements to specify whether CFU claims are ATM or ATE under 21 CFR Part 111. Brands can legally overstate potency at the point of consumption without violating GMP regulations, as long as the label count was accurate at manufacture. That regulatory gap leaves consumers — and brands that don’t know what questions to ask — in a genuinely difficult position.

What Actually Kills Probiotic Bacteria

Temperature gets all the attention, but the real stability killers are more nuanced than “keep it cold.”

Moisture. Probiotic organisms in finished capsules or tablets are typically in a dormant, desiccated state. The moment water activity (a_w) climbs above 0.3, metabolic activity resumes. And active cells in an oxygen-rich, nutrient-limited environment don’t thrive — they die. Container closure integrity failures, hygroscopic excipients, and inadequate desiccants are all routes to elevated a_w. In our stability testing work at Qalitex, we routinely see moisture-related failures in products packaged in standard HDPE bottles without adequate desiccant or foil induction seals.

Oxygen. Most probiotic strains used in supplements are obligate or facultative anaerobes. Oxygen exposure drives oxidative damage to the cell membrane and DNA. Packaging format matters enormously here — an aluminum foil blister pack versus a standard plastic bottle can mean the difference between 90% viability and 30% viability at the 24-month mark. That’s not a minor formulation variable. That’s the difference between a product that works and one that doesn’t.

Temperature fluctuations. It’s not just average temperature — it’s the cycling. A product that spends two weeks in a shipping container during summer transit, fluctuating between 28°C and 40°C, will experience significantly more cell death than one held at a steady 25°C. This is why ICH Q1A(R2) accelerated stability testing at 40°C/75% relative humidity matters, even for supplements. It’s a stress test that simulates real distribution conditions far more realistically than a controlled refrigerator ever could.

The strain itself. Lactobacillus rhamnosus GG is one of the more robust commercial strains. Bifidobacterium longum is notoriously fragile. Saccharomyces boulardii — a yeast, technically — is far more stable than most bacterial strains. Formulation decisions about which strains to include need to account for their inherent stability profiles, something not all brands think through carefully when they’re chasing label claims.

Why Most Probiotic Stability Programs Fall Short

Here’s what I see in practice: brands run a 3-month accelerated stability study at T=0, T=1, and T=3 months, get numbers they like, and call it done. Then they set a 24-month expiration date based on that compressed window — without real-time confirmation data to back it up.

That approach has two serious problems. First, accelerated conditions don’t always predict real-time behavior accurately for biological products. The Arrhenius equation, which underlies most accelerated stability modeling, was developed for chemical degradation kinetics — not living organisms. Probiotic die-off under heat stress doesn’t follow the same predictable kinetic curves as, say, vitamin C oxidation. The biology is messier.

Second, there’s often no ongoing real-time monitoring. A proper stability program for a probiotic product should include real-time samples pulled at T=0, T=3, T=6, T=9, T=12, T=18, and T=24 months, stored at the intended storage condition — typically 25°C/60% RH for ambient products, 5°C for refrigerated lines. A lot of brands set those samples aside at product launch and then forget about them entirely, until a retailer complaint or a recall forces the issue.

21 CFR Part 111.55 requires manufacturers to establish expiration dates supported by testing. But “supported by testing” leaves enormous room for interpretation, and FDA enforcement on this specific point has been inconsistent. The result is that the industry’s self-imposed standards vary dramatically from brand to brand.

At Qalitex, our stability testing programs for probiotic products follow ICH Q1A(R2) protocols adapted for dietary supplement matrices — including both real-time and accelerated arms, packaging-specific testing using the same container closure system that ships to retail, and enumeration methods calibrated for each strain type. That last point matters more than most people realize: generic aerobic plate count methods systematically undercount many probiotic strains. Strain-specific media and anaerobic incubation conditions can produce counts that are 2–5× higher than what you’d get from a standard USP <61> microbiological assay. If your testing lab isn’t accounting for that, your stability data is telling you the wrong story.

What to Ask Before You Sign a Purchase Order

If you’re a brand working with a contract manufacturer or planning to launch a probiotic product, here’s what to ask before you finalize your formulation and labeling:

Ask for real-time stability data, not just accelerated. Specifically, T=12 and T=18 month real-time data at the intended storage condition. If your contract manufacturer can’t produce this, either they haven’t run it or their product hasn’t been on the market long enough to generate it. Both are concerning.

Confirm whether CFU claims are ATM or ATE. If it’s ATM, ask for the overage factor — how many extra CFU are added at manufacture to compensate for die-off over shelf life. A 10 billion ATE claim with a properly characterized strain and a solid formulation might require a 25–40 billion ATM starting point. If the manufacturer can’t explain their overage rationale with data, that’s a red flag worth pursuing.

Get stability certificates specific to your packaging configuration. Stability data generated on a foil-packed product doesn’t transfer to a bottled product. If your contract manufacturer switches your packaging and doesn’t re-run stability, you’re operating without real data. This comes up more often than you’d think.

Check the enumeration method. Ask specifically: what growth media, what incubation conditions, what temperature, and for how long? Lactobacillus strains require MRS (de Man, Rogosa and Sharpe) agar under anaerobic or microaerophilic conditions. Bifidobacterium strains need RCM or MRS-NNLP with strict anaerobic incubation at 37°C. A lab running aerobic general plate counts on probiotic samples is giving you systematically low numbers — which might sound conservative, but actually makes accurate shelf-life prediction harder.

For EU market entry and European regulatory compliance, Care Europe provides expert consulting from Paris.

What a Credible Stability Program Actually Looks Like

A well-designed probiotic stability program isn’t complicated, but it does require discipline. You need T=0 characterization — viable count, water activity, moisture content, physical appearance — in the actual commercial packaging, stored under both real-time and accelerated conditions. You need time points pulled and tested without gaps. And you need a documented protocol for what happens when a time point falls below your specification limit before expiry.

The brands that consistently deliver what their labels promise treat stability as a continuous quality system function, not a one-time regulatory checkbox. They also tend to be the ones that don’t end up on FDA’s publicly searchable warning letter database — which, incidentally, is worth reviewing periodically. It’s a useful map of exactly which GMP failures regulators are actually citing in the supplement industry right now, and “lack of testing to verify product meets specifications” appears with uncomfortable regularity.

Before you finalize any probiotic product launch, pull your contract manufacturer’s real-time stability data for the specific strain blend and packaging configuration you’re using. If that data doesn’t exist or covers fewer than 12 months at real-time conditions, commission independent third-party stability testing before you set your expiration date and print your label. The cost of a full ICH-aligned stability panel — typically $2,000–$6,000 depending on the number of time points and strains — is a fraction of the cost of a reformulation, a retailer delisting, or an FTC complaint from a consumer who paid a premium for “100 billion CFU” and got fewer than 5 billion by the time they opened the bottle.